Agisoft Metashape
Agisoft Metashape => General => Topic started by: Dmitry Semyonov on October 09, 2016, 12:09:45 AM
-
Pre-release version of Agisoft PhotoScan 1.3.0 is available for download.
This pre-release is considered as unstable and is not recommended for production use.
Please let us know if you have any problems using this version.
Please post problems and/or questions concerning other PhotoScan version in separate topics.
This topic will be removed when the final 1.3.0 version is released.
Note
Project files created with 1.3.0 version can't be opened with earlier versions.
EDIT:
PhotoScan 1.3.0 build 3772 was finally released. Thank you for useful feedback.
Please use the links on the main PhotoScan page for download:
Download Installers (http://www.agisoft.com/downloads/installer/)
-
Hello,
I installed version 1.3.0 o my Windows 10 computer.
After install, i run program and I get message that since 1.2.5, I must register again entering my valid stand alone activation code (which was working well with 1.2.6). I did and got error relating to Rehostable id....see attachment
I am now reinstalling 1.2.6 since 1.3 cannot be activated....
Please help,
Regards,
-
Hi Dmitry.
I have the same problem with registration. After one crash I was not able to register the app anymore. The only way is to completely delete the application files and copy them again to the disk. The registration is then required again. Before the crash, I was able to open a project done with a drone with Rolling Shutter issues. I had troubles to find this new (and very welcome) feature. How do we access it? Is it automatic and user input independent or is there an option to activate and parameters to change anywhere?
Cheers
-
Hello pap1956 and RHenriques,
Have you tried using Remove Rehostable Id as suggested?
-
Alexey,
yes, I have and got error message seen in Console pane....So version is not activated (demo)
-
Hi Alexey
I'm using the MacOSX version and in this case the "Remove Rehostable Id" dialogue does not appear.
I cannot simply enter my serial number because the app says that it's already in use. In another computer I was able to register once but, after a reopen, the app asks for the serial number again and it's no longer possible to activate the app.
Cheers
-
I can sugest the following (when applicable to the OS used), perform a clean installation previously removing the old version folder, run PhotoScan with the administrator privelegies.
-
Hi Alexey,
I uninstalled version 1.2.6 and then proceeded to re install 1.3.0 and still had to re enter my license code and got same error related to Rehostable id....
What could I do?
Best regards,
-
Hello pap1956,
Can you try to start PhotoScan as Administrator and then press Remove hostid button?
-
Alexey,
I tried again running version 1.3.0 as Administrator and get same error:
- in command window I get checkout license error message;
- in GUI, I get message that license has to be activated;
- I enter license code;
- In rehostable window I click rehost id and get same error as previous attemps....
so I really would like to be able to test the PS features!
Best,
-
Hello pap1956,
Please send us an e-mail to support@agisoft.com, for more detailed discussion and investigation.
-
DNG or CR2 RAW files from CANON still don't work properly (100D, 5D MkII, 5D MkIII, and 5D MkIV tested)...
See topic: http://www.agisoft.com/forum/index.php?topic=5659.msg28639#msg28639 (http://www.agisoft.com/forum/index.php?topic=5659.msg28639#msg28639)
-
Hi Alexey
Trashed everything related with the stable version of Photoscan. When I try to input my serial number in the 1.3.0 version i get the message that the "Activation Key is already in use". As I told above, I'm using the MacOS version.
Cheers
-
hello,
i can't run 1.3.0 on osx 10.10.5, is 10.11 required from now on? i am no quite keen in upgrading the os right now..
thanks
-
Hello RHenriques,
Is it possible that you have tried to activate the license on another computer? If not, than could you please specify, if the active user account has full write access rights on the computer? I also suggest to send the e-mail to support@agisoft.com with the link to this topic, so that our support team could perform more detailed investigation.
-
Hello pap1956,
Please send us an e-mail to support@agisoft.com, for more detailed discussion and investigation.
Alexey,
I tried sending my PS 1.3 issues to support@agisoft.com but I got undelivered mail error from server so I sent same message to support@agisoft.ru.
Hope you have received mail.
Best regards,
-
Hello pap1956,
We've got the message to support@agisoft.ru. Hope we'll find the solution soon.
-
still no email notification?
>:(
-
Cool, nice to know a new version is on its way! Can I ask what the main new features of version 1.3 are?
-
;D No. You can not, Thomas. You can just read the release Log file
-
Haha ok thanks, missed that!
-
The 1.3 python reference doc does not contain information on running from the command line. Pretty significant gap. How is the venv sourced? Perhaps a simple hello world example? Is this in the works?
If I am running on a cloud node, how do I enter a temp license key?
Thanks
-
Hello MManning,
To run the script from the command line you need to use -r command line argument, for example:
photoscan.exe -r script.pyIt can be also accompanied by the arguments to be used by the script.
You can get the information about the command line parameters using --help argument.
You can request a 30-day Trial license key for this version via e-mail, by sending a message to support@agisoft.com (just specify, that it should be for 1.3, since the previously issues Trial keys wouldn't fit this version).
-
Hi Alexey,
Would it be possible to provide a portable installation of the beta ? I would like to test v1.3 before its release (and report eventual bugs), but the installation of executables is necessitating administrative privileges and that requires many hierarchical approvals.
-
please fix this issue before releasing: http://www.agisoft.com/forum/index.php?topic=6003.0
-
What is the principle behind the function of Image-based drawing shape?
-
Having a rather large issue with running this release - no matter what I do, I get GL errors...
2016-10-15 16:54:59 renderer_gl.cpp line 399: drawText() failed, error 1280
^^ this error appears over and over again, and eventually, the program crashes.
Windows 8.1
GTX780
i7 4790k
... and...
Windows Server 2012 R2
(no graphics card)
Any ideas? Is there a way to open the .exe with a "-nogl" option? I'm on Standard.
-
Hello ricberw,
Which drivers are installed for your graphic card?
If possible please also submit the crash report with some message to ease the identification of the report.
-
Is it possible to install the trial license for PhotoScan-Pro 1.3 Linux on the command line?
# photoscan.sh -r test.py
Error checking out license
No license for product (-1)
Segmentation fault
Thanks for your help.
-
Hi Alexey
Trashed everything related with the stable version of Photoscan. When I try to input my serial number in the 1.3.0 version i get the message that the "Activation Key is already in use". As I told above, I'm using the MacOS version.
Cheers
Hello RHenriques,
Have you managed to activate the version 1.3 successfully on Mac OS or there is the same issue?
-
Is it possible to install the trial license for PhotoScan-Pro 1.3 Linux on the command line?
Hello MManning,
The activation from the command line can be performed using --activate argument, like in the following example:
./photoscan.sh --activate 11111-22222-33333-44444-55555
-
Having a rather large issue with running this release - no matter what I do, I get GL errors...
2016-10-15 16:54:59 renderer_gl.cpp line 399: drawText() failed, error 1280
^^ this error appears over and over again, and eventually, the program crashes.
Windows 8.1
GTX780
i7 4790k
... and...
Windows Server 2012 R2
(no graphics card)
Any ideas? Is there a way to open the .exe with a "-nogl" option? I'm on Standard.
Hello ricberw,
I suggest to install latest NVidia drivers from their official web-site for your graphic card: http://www.nvidia.com/download/index.aspx?lang=en-us
For the second configuration, you need to have some software rendering device, as you only operate PhotoScan Standard via GUI. Probably you need to use Point Cloud view mode as default, to avoid automatic switching to the mesh view modes.
If the problem still exists, please provide the full Console pane output (from the program start-up).
-
How can I make the point cloud view be the default/not switch away from it?
It seems as though these errors are occurring even when the point cloud view is selected.
Output from the console:
2016-10-18 23:34:49 Agisoft PhotoScan Version: 1.3.0 build 3075 (64 bit)
2016-10-18 23:34:49 OpenGL Vendor: Microsoft Corporation
2016-10-18 23:34:49 OpenGL Renderer: GDI Generic
2016-10-18 23:34:49 OpenGL Version: 1.1.0
2016-10-18 23:34:49 Maximum Texture Size: 1024
2016-10-18 23:34:49 Quad Buffered Stereo: not enabled
2016-10-18 23:34:49 ARB_vertex_buffer_object: not supported
2016-10-18 23:34:49 ARB_texture_non_power_of_two: not supported
2016-10-18 23:34:49 renderer_gl.cpp line 399: drawText() failed, error 1280
2016-10-18 23:34:49 renderer_gl.cpp line 399: drawText() failed, error 1280
2016-10-18 23:34:49 renderer_gl.cpp line 399: drawText() failed, error 1280
2016-10-18 23:34:49 renderer_gl.cpp line 399: drawText() failed, error 1280
2016-10-18 23:34:49 renderer_gl.cpp line 399: drawText() failed, error 1280
(then the drawText() errors show up constantly)
Also included a screenshot of the view with the cameras enabled - if I enable the cameras to be displayed, I get random color blocks showing up on the model view - but if I disable the cameras from view, the random color blocks go away.
As another note, it seems that it only shows drawText() errors when I move the mouse over the model/point cloud view pane.
EDIT: I just turned off "show info" in the "view" menu's show/hide items list, and now my drawText() errors have disappeared! Not sure this will solve the crashing though (since it seems to crash randomly). I'll post again if I see another crash.
-
Hello ricberw,
Do you have the same issue with the version 1.2 on the computer without discrete graphic card?
The default view can be selected in the General tab of PhotoScan Preferences window.
-
I see definitely longer texture blending times and textures aren`t that sharp as in 1.2.6 and muddy in places where 1.2.6 gives pretty sharp results.
-
Hello RouvenMiller,
Can you provide any sample set (images + project file) that can be used for our internal tests? If possible, please send the download link to support@agisoft.com.
In the version 1.3 we have modified Mosaic blending for the texture generation that should have improved the blending quality and minimize visible seams on high zoom level.
-
Hello ricberw,
Do you have the same issue with the version 1.2 on the computer without discrete graphic card?
The default view can be selected in the General tab of PhotoScan Preferences window.
No, it does not - I've been using 1.2.6 for a while now, but the issue was not present there (nor was it present in 1.2.5, which I was using prior to 1.2.6).
It seems to be a problem with drawing any type of text on the model view, as showing cameras does it (the camera names/numbers do not show up in 1.3) and trying to "Show Info" e.g. # of points does it.
-
Hello ricberw,
Ok, thank you for the information, we'll try to reproduce it and fix in one of the next pre-release updates.
-
Hello RouvenMiller,
Can you provide any sample set (images + project file) that can be used for our internal tests? If possible, please send the download link to support@agisoft.com.
In the version 1.3 we have modified Mosaic blending for the texture generation that should have improved the blending quality and minimize visible seams on high zoom level.
Yap, send it to your support. Filesize is 1.5gb so I used google drive.
Seams are less but fine detail in higher zooms are eaten up. I use only focus stacked images so seams are minimal nearly zero. 1.2.6 does such a great job. At least for focus stacked shots 1.3 doesn`t do so well.
regards
-
Hello RouvenMiller,
Thank you for sending the link, we've started to download the data.
-
Hello RouvenMiller,
Thank you for sending the link, we've started to download the data.
Did you guys notice the difference or am I wrong?
-
It is a big problem with 1.3 blending texture mode
v1.3
(http://digitworlds.com/tmp/v1.3.jpg)
v1.2.6
(http://digitworlds.com/tmp/v1.2.6.jpg)
-
Hello RouvenMiller,
Yes, we see the difference that you are referring to.
Also we'll take into account lunar's comment.
-
Hello Alexey,
I work a lot with the Python scripting and I have to say, the new IPython console is really amazing!
I was wondering if there will be, in this version, the possibility of adding a color bar and/or a scale bar with the DEM and the Orthomosaic in index color mode when exporting? Also, now that the index calculator is well established, you should add the color bar to the view (as done for the DEM) when the index transform is enabled.
-
Ruh roh, my old python scripts that import pyside are now broken...
Importing PySide disabled by QtConsole, which has
already imported an incompatible Qt Binding: pyside2
from PySide2.QtGui import QApplication, QDialog, QLayout, QGridLayout, QMessageBox, QGroupBox, QSpinBox, QSlider, QProgressBar, QDial, QDialogButtonBox, QComboBox, QLabel
ImportError: cannot import name 'QApplication'
-
Hello grommit,
In the version 1.3 you need to import PySide2.
Also note that some of the classes has been moved from QtGui to QtWIdgets.
-
Hi Alexey
Problems with the license in the MacOS version continue in the latest beta. I was now able to generate a working version after the first launch. While trying to launch a second time, the app says it need the serial number again and, after the right license number, it says that it cannot write the license file. It's strange because the license file it's already in the folder and it was successfully generated. I've detected that it's due to code written to the app package itself. Sorry to say that I used the following trick but if we substitute the app by a fresh one, with the license file generated, it starts just as expected. A second launch breaks it again. So the trick to use it, by now, is to put a fresh app in the folder every time before launching the app. Please revise this strange behavior.
I've tried to use it in demo mode in my portable and also presented some issues with my graphics card. Version 1.2.6 works fine with the AMD Radeon R9 M370X added to the OpenCL tab. Version 1.3.0, with the same settings, refuses make anything at all due to Open CL issues until we deactivate the GPU.
Cheers
-
Hello RHenriques,
This problem can occur if your account have no read/write access to the PhotoScan installation folder. Are you using the same user account for the application installation each time?
Please check the account permissions in Get Info -> Sharing & Permissions in the folder context menu (https://support.apple.com/kb/PH21997?locale=en_US), they should be "Read & Write".
-
Hello Alexey
Yes, I'm always using the same account and it's my administration account. I'll check later but the folder certainly has read/write privileges. I'll check tonight to confirm if it is a folder privileges issue.
Considering the GPU problem, it occurs when the dense cloud building is triggered. The app says "Can't Initialize OpenCL". It's working fine without the GPU activated and it's also working fine in the 1.2.6 version.
Cheers
-
Hello RHenriques,
We'll check the issue on a similar configuration for Mac OS X and if it can be reproduced, will fix it in the next update.
-
Hello Alexey
In the new 1.3 pre-release I am able to run the following steps in interactive mode:
root@li641-212:/opt/test# photoscan.sh -i
Agisoft PhotoScan Python Interactive Console (Version 1.3.0 build 3075)
Type exit() and press Enter to exit
>>> import PhotoScan
>>> doc = PhotoScan.app.document
>>> chunk = PhotoScan.app.document.addChunk()
>>> chunk.addPhotos(["/scratch/age_43.png", "/scratch/page_44.png"])
but when I put those steps into a test.sh file and run them I receive a seg fault:
root@li641-212:/opt/test# photoscan.sh -r test.sh
Segmentation fault
Can you advise how I might diagnose the issue?
Thank you for your help
Maurice
-
Hello Maurice,
Script should be saved with .py extenstion.
Also please re-install the pre-release version, as we have updated it to the build 3149.
-
Hi Alexey
Just tested the privileges and the issue remains. It's impossible to keep the license working on a second app launch. I'm only testing this version in this computer with this license. All Sharing&Permissions options are set to Read/Write. It's strange because the first app launch with the generated license file works just fine. I hope that you can nail the issue.
Cheers
-
I notice that with latest pre release, I can no longer have two windows with different photos opened.
I select one photo and opens as a tab over model window. I send it to Other Tab. I now have two windows side by side, Model and Photo.
I select Model view and try to open other photo side by side with previous photo, but it always opens in the Photo tab...
Hope this can be fixed,
-
Hello RHenriques,
Thanks for additional information, we'll try to reproduce that. Could you also specify the Mac OS X version you are using?
Hello pap1956,
When the photo is opened in Photo view mode, you can right-click on the photo label and use Move to the Other tab. It works in both directions and you can have multiple photos opened in left and right frame.
-
Hello Alexey,
sorry but that is what I do. I have a first photo opened in Other Tab. But when I open the second photo it always opens in the same tab as first one .... and replaces first one....
Best,
-
Hello pap1956,
You can choose "Open in the new tab" option in the context menu of the Photos pane, I think that's where the confusion arises, as some mechanics really changed in the version 1.3.
-
Ok,
now I see how to do it!
Thanks for your time!
-
Just tested the privileges and the issue remains. It's impossible to keep the license working on a second app launch. I'm only testing this version in this computer with this license. All Sharing&Permissions options are set to Read/Write. It's strange because the first app launch with the generated license file works just fine. I hope that you can nail the issue.
Hello RHenriques,
I have an additional question, how you are running PhotoScan? From DMG directly, from some folder where the application file is copied to, using shortcut, from terminal?
-
Hi Alexey
I'm using the latest MacOS Sierra.
Concerning how I run Photoscan, it's "from some folder where the application file is copied to". Running by simply "double-click" in the application icon.
Cheers
-
Could you guys provide some explanation for the Render Photos feature?
Thanks!
-
"Update Orthomosaic" does not seem to work in this release.
After I create my shapes and select which photos to use, Pressing udate orthomosaic just removes my image selections I have chosen...
Any work around for this?
(edit) The issue stemmed from changing projections and trying to rebuild the mosaic. Old shapes just needed to be deleted to solve the issue.
-
Please can you add support for the 3DConnexion space mouse? Thanks.
-
Please can you add support for the 3DConnexion space mouse? Thanks.
+1
Appreciate not a pre-release issue but would greatly appreciate this when possible.
Dave
-
Could you guys provide some explanation for the Render Photos feature?
Hello wojtek,
At the moment this feature helps to generate lenticular images either by fixing the existing images, or by adding intermediate renders. Later we might extend the functionality for circular image acquisition scenario.
-
"Update Orthomosaic" does not seem to work in this release.
After I create my shapes and select which photos to use, Pressing udate orthomosaic just removes my image selections I have chosen...
Any work around for this?
(edit) The issue stemmed from changing projections and trying to rebuild the mosaic. Old shapes just needed to be deleted to solve the issue.
Hello SuperNewb,
So is the problem solved or can be reproduced in 1.3? The shapes have their coordinate system stored in the project, so if you georeference the model in another coordinate system after drawing the shape, there could be problems, as the shape and the model would be in the different coordinate systems.
-
"Update Orthomosaic" does not seem to work in this release.
After I create my shapes and select which photos to use, Pressing udate orthomosaic just removes my image selections I have chosen...
Any work around for this?
(edit) The issue stemmed from changing projections and trying to rebuild the mosaic. Old shapes just needed to be deleted to solve the issue.
Hello SuperNewb,
So is the problem solved or can be reproduced in 1.3? The shapes have their coordinate system stored in the project, so if you georeference the model in another coordinate system after drawing the shape, there could be problems, as the shape and the model would be in the different coordinate systems.
Yes it was in fact caused by my changing coordinate system. Thanks
-
Just a small thing, when using "Gradual Selection" in the previous versions, the focus returned to the model window, so you could do a gradual selection, then hit "del" to remove points. In this version the focus doesn't return so you have to click each time. If you have point selection mode on, this deselects your gradual selection.
Thanks!
-
Some feedback:
-Depth reconstruction seems longer and more unstable in the new version. I've had two projects fail (they had high overlap) but completed without an issue on the 1.2.6 build.
-Read-Only option should be something we can toggle on/off. In my case where i flick between multiple machines constantly i did find projects that were locked when I did not want them to be.
-New theme buttons could use some coloring so that it's easier to distinguish them (and to tell if DC is present for example, like in classic theme)
-
Hello wojtek,
What kind of issues were there with the depth maps reconstruction? Can you provide the log file related tot he operation?
-
Hello wojtek,
What kind of issues were there with the depth maps reconstruction? Can you provide the log file related tot he operation?
I'll try to reproduce this for you between projects (it takes several hours before it crashes).
-
Hello wojtek,
Thank you in advance. Some additional information would be also helpful, like computer hardware specifications (CPU, GPU, RAM, OS), number/resolution of images, processing settings used, chunk info for effective overlap value. Memory consumption peak for the operation wold be also useful to know.
Also please check that you are using build 3149.
-
There you go, I've let it run overnight.
Log and Chunk info in the attachment.
Windows 8, AMD 7990 and R290. 256GB of ram. Images are 42mpix
-
Hello wojtek,
Thank you for additional information.
According to the log there are issues related to OpenCL, we'll try to reproduce it on similar GPU. For now I can suggest to update AMD drivers.
-
The File become heavy and Dens point cloud take time to load and when i move/Rotate the 3d model it take a time to load the dens cloud again!
Mohammed
-
Hello Mohammed,
Is the behavior the same if you switch "Enable Performance Optimization" option in the Advanced tab of PhotoScan preferences?
-
Hi Alexey,
Thanks it works fine now!
Mohammed
-
Hello,
I've updated the drivers and gave it another go, same issue:
-
Hi Alexey and Dmitry,
What is the anticipated release date for stable v1.3? Thanks!
VG
-
Dear Alexey: I kindly suggest you to update the Potree converter to 1.7 in the comming new photoscan 1.3 release. The 1.6 converter that you are using has some documented bugs that cause some clouds not to be rendered when a projected reference system is being used.
Thanks in advance.
Best regards.
José
-
Hi,
I see in update paper "ARA thermal image"....
there are also some type of support for thermal camera on UAV??
Thanks
-
What is the anticipated release date for stable v1.3? Thanks!
I think it will be in the late December.
I kindly suggest you to update the Potree converter to 1.7 in the comming new photoscan 1.3 release. The 1.6 converter that you are using has some documented bugs that cause some clouds not to be rendered when a projected reference system is being used.
We'll try to update it in the next 1.3 pre-release updates.
I see in update paper "ARA thermal image"....
there are also some type of support for thermal camera on UAV??
ARA thermal image support allows to interpret the image data automatically according to the format specification. So it's mostly related to the visual representation of ARA data and processing results basing on it.
-
Please, can you share the format specification...?
-
ARA format is a thermal data format from AscTec (http://wiki.asctec.de/pages/viewpage.action?pageId=5275945).
-
According to the 1.3.0 Python Reference exportPoints has changed and now requires format to be of type PhotoScan.PointsFormat, in contrast to the old behavior where exportPoints expected a string.
What exactly does the new exportPoints do when format is not specified explicitly? I assume that, like the old exportPoints, it tries to guess the format from the file ending of path. What happens when format does not match the file ending? I believe that the old behavior was buggy (should I still report this?) and produced broken files when the file ending was txt or xyz.
-
Hello Jan,
If the format is not not specified, then PhotoScan will parse the export path and use the extension to identify the format to be used. In case the format is specified but doesn't match the extension, PhotoScan will export the data according to the format use (PointsFormat), but save it under the user-defined path, without changing anything in it.
In the current 1.3 pre-release TXT extension is not interpreted as XYZ output, so for such extension you need to specify PointsFormatXYZ explicitly.
If you can reproduce the broken or incomplete export for XYZ format, please let me know.
-
Hello Alexey
thank you for your reply.
If the format is not not specified, then PhotoScan will parse the export path and use the extension to identify the format to be used. In case the format is specified but doesn't match the extension, PhotoScan will export the data according to the format use (PointsFormat), but save it under the user-defined path, without changing anything in it.
Great, that's pretty much what one would expect. I don't know about your documentation policy, but that'd be useful information in the reference.
In the current 1.3 pre-release TXT extension is not interpreted as XYZ output, so for such extension you need to specify PointsFormatXYZ explicitly.
If you can reproduce the broken or incomplete export for XYZ format, please let me know.
I'll try to find the time to put together an MWE and get it to you. By the way, if memory serves me right, the format was/is called TXT in 1.2.6 somewhere in the GUI's export dialog.
-
Hi Alexey,
Version 1.3 got Algorithmic improvements for dense cloud generation in case of large overlap.
What are the improvements? Because I have projects with an overlap of over 6.
The stage of depth map generation goes fast, but the actually dense cloud generation takes a long time.
-
Hello Alexey - I used the rolling shutter function (worked perfectly) and I started another project and I receive a license error notification box. The release mostly still works except some functions are shut out. I desperately have to complete another project. I am using a windows based system. Is there a new pre-release version that does not have this issue. I downloaded the pre-release version from the link in the first thread note. Is there a location to download the latest pe-release version?
Thank you,
Jim
-
Hello Jim,
What problem do you have with the activation? Maybe you can drop an e-mail to support@agisoft.com describing the issue and providing the screenshots and order number, so that we could also check the license key?
-
What are the improvements? Because I have projects with an overlap of over 6.
The stage of depth map generation goes fast, but the actually dense cloud generation takes a long time.
Hello PK_2509,
The performance of the dense cloud generation (second step of Build Dense Cloud stage) should be improved in the version 1.3 for the projects with the excessive overlap. What is the Effective Overlap value from the chunk info for your project?
Have you compared the processing time for this stage between the versions 1.2 and 1.3?
-
Hi Alexey,
I use PhotoScan for my thesis, so I was not sure if Version 1.3 is stable.
Therefore I haven't compared the versions. I will process my Projekt again in Version 1.3.
The Effective Overlap in my Project is 6.45.
Edit: One other thing, my Propertys says i have checked the Adaptive camera model fitting, but i don´t.
The change comes when I optimized my project with markes. I have a precalibration for my cameras and also set they fix at the Camera Calibration section, so is this a bug? (see picture two)
-
Hello PK_2509,
If you run the test on 1.3, it would be helpful to save the processing log as well (you can enable it's redirection to the external file in the Preferences).
-
Hello Hello PK_2509,
I've tried optimization for the chunk where all the cameras have fixed calibration and the Adaptive camera model fitting flag remain the same, maybe some other steps should be followed to reproduce the issue?
-
Hello Alexey,
at the moment i can't reproduce the issue. I will try to find the Probelm.
Also I will upload the processing log, if the project is reprocessed.
-
Hi,
I have a problem when running script from command line:
PhotoScan.app.gpu_mask doesn't work (AttributeError: 'PhotoScan.Application' object has no attribute 'gpu_mask')
It works when running from the interface, and the GPU is used in command line just fine (depending on the option set through the interface before running the script)
Thank you so much for the command line and floating license functionality, Regards,
-
Hello BenjaminG,
We have slightly updated Python API, but haven't yet updated the reference file.
It should be gpu_mask_opencl now.
-
Hello,
what is this "image based shape drawing support" feature?
How does it work?
Thank you
Mathias
-
Hello Mathias,
It means that now you can also draw a shapes in Photo view, providing that the corresponding camera is aligned. In the version 1.2 it was only possible to draw shapes in Model and Ortho view.
If you also have "attach markers to the shape vertices" flag on in the Preferences, you'll be able to adjust the projections of each vertex of such shape (like regular markers) to achieve better accuracy of the shape positioning.
-
Hi Alexey,
I think I've found the issue for the adaptive camera model fitting change.
After the alignment one camera was not aligned. So I aligned the single camera afterwards and then adaptive camera model fitting changes to yes.
-
Hello PK_2509,
When you run the Align Selected Cameras procedure PhotoScan will refine the interior parameters of camera orientation, unless the Fix Calibration flag is turned on for the calibration group.
-
Hi Alexey,
I turned the Fix Calibration on, so the interior parameters of the camera orientation doesn't change. Only the Info says yes.
There is no issue with my project or something else, I only noticed it.
-
im using Phantom 4 for collect data and Pix4Dcapture for Flight Plan.
whether using pix4Dcapture on a Phantom 4, photos can be processed data on Agisoft?
-
I don't know if this has been proposed, but I think it would be helpful to be able to create multiple point clouds within the same chunk. That would help in cases where you would like to reproduce certain areas of a project in high accuracy while others in lower accuracies (ie ground data)
-
Hi Alexey, unrelated to my recent bug post for 1.2.6, I noticed that 1.3.0 appears to have an issue with processing photos at the edge of chunks. See this image:
(https://dl.dropboxusercontent.com/u/13848478/gaps.jpg)
In version 1.2.6, there is not a gap between the upper two chunks when the data is processed. In 1.3.0 data is lost at the edges.
Mark
-
Hello Alexey,
I've been using v 1.3 for a few days now and I'm loving the command line feature!
Are there any plans to introduce Photoscan as a dynamic library (.so) so we can seamlessly use import and use the Photoscan module with an IDE? It would greatly improve the python scripting experience, particularly for users interested only in headless mode.
Cheers,
Venkat
-
Locks up upon opening a pretty large .psx file - 2000 photos... File worked fine with 1.2 anyone else have a problem?
-
I'm having an issue with markers.
I am scanning small objects and include a scale bar positioned near the object in some photos. My final model never includes these scale bars, I crop them out in the alignment stage as they often cause noise.
In previous versions, my workflow was to finish up to the model stage, then place markers on the scale bar in the first photo, find a second photo with the scale bar and place the markers again (the line comes up to show the 2d line detected and I find the intersect on that line and set the marker). I then create a scale bar and give it a value.
In this version, even though my scale bar is not within my model, the system acts as if it knows the 3D point and places the marker directly from the first photo. i don't have to find the intersect on a second photo. The only problem is that this 3D point is wrong, and even if I drag the marker to the right point in the other photos, my scale is way out in the final model.
Is there a way to return to the old behaviour?
Thanks
Michael
-
Actually, further to above, it seems like some of the tie points I am deleting are not actually getting deleted. I generally mask everything other than my main subject, but the last few objects I have processed with this new version have reconstructed objects in the masked areas, even though I deleted tie points in those masked areas in the sparse cloud before dense reconstruction. Definitely something odd going on!
Edit--Actually not sure what caused above as regenerated and the other points didn't show up. Will see what happens on the next scan.
-
Are there any plans to introduce Photoscan as a dynamic library (.so) so we can seamlessly use import and use the Photoscan module with an IDE? It would greatly improve the python scripting experience, particularly for users interested only in headless mode.
Hello Venkat,
At the moment we are not planning to supply an external library or SDK that allows to access PhotoScan algorithms.
-
I'm having an issue with markers.
Hello Michael,
If you need to reproduce the marker placement behavior with the equipolar lines, you need to do the following:
- Add new marker in the Workspace pane by the corresponding button,
- Open image in the Photo view and put marker projection using Place Marker option in the context menu instead of using Create Marker tool,
- After that you'll have the marker with the single projection only and will see the equipolar lines on another images.
In the version 1.3 PhotoScan is always using Guided Marker placement approach, even if only sparse cloud is present (in the previous version it worked only starting from the dense cloud).
-
I don't know if this has been proposed, but I think it would be helpful to be able to create multiple point clouds within the same chunk. That would help in cases where you would like to reproduce certain areas of a project in high accuracy while others in lower accuracies (ie ground data)
Hello cadm8,
It's better to put the suggest to the Feature Requests forum section:
http://www.agisoft.com/forum/index.php?board=11.0
At the moment we are not planning to implement the support for multiple instances of the sparse cloud, dense cloud, mesh and etc. for the same chunk, as it may lead to the confusion.
-
Hi Alexey, unrelated to my recent bug post for 1.2.6, I noticed that 1.3.0 appears to have an issue with processing photos at the edge of chunks. See this image:
In version 1.2.6, there is not a gap between the upper two chunks when the data is processed. In 1.3.0 data is lost at the edges.
Hello Mark,
Could you provide some additional details to support@agisoft.com regarding the issue. Probably the screenshots from both versions, so that the difference can be clearly observed.
-
Hello
with the command line feature of v1.3 it would be nice to submit the PS model build to a GPU node on a cluster. the issue is that when you submit a job to a cluster you dont know which GPU node will be assigned the job. Since I cant install the license on every GPU node, I get a license error as the result of my run. ( the PS install itself is in a container which is executed on the node). Does anyone know of options to allow for this type of processing?
Thx!
-
Dear Alexey,
-I tried to update the new version in my laptop but when i put the license the software give me that '' Activation Key is already in use'' what dose that mean? in previous version it works fine on my PC and laptop now i can only open it on PC.
So is that mean the new floating license work on one computer not PC and Laptop?
Mohammed,
-
After importing an existing camera calibration via the GUI, PhotoScan automatically initiates building of the coarse point cloud. The Python API does not seem to offer this possibility (or does it?), and so, after importing via chunk.importCameras(camera_path), no dense cloud can be constructed. How would you work this using Python?
On a related note -- can you reset the calibration of specific cameras and then recalibrate them via chunk.alignCameras(recal_cams)?
-
Hello Jan Philip Göpfert,
After importing cameras via Python you need to use buildPoints() function. Using alignCameras() will discard imported camera positions and will estimate them from scratch. Actually, it's not a version 1.3 feature, it was present in the older versions as well.
-
Hello tornado2011,
I think that you have a node-locked license (since floating license policy has been only introduced recently with the new activation system), so one license key can be only used on a single machine. If you need to transfer the license, you need to deactivate the license first.
-
In the Model view front faces are transparent and back faces are visible in Shaded, Solid, Wireframe, and Textured modes.
(http://www.digital-mapping.net/forums/Photoscan/2016/PS13_faces.jpg)
-
Hello bisenberger,
Do you observe it on any mesh generated in 1.3?
-
Hi Alexey,
Yes.
The project was created with version 1.2.6. I ran the Build Mesh in version 1.3.0 on the point cloud created in version 1.2.6.
-
Hello bisenberger,
Can you send any sample project where the issue can be reproduced to support@agisoft.com? Please also specify the reconstruction settings used.
-
Troubles with open CL on macbook pro
Hy I have a Macbook pro mid 2014 with Intel iris pro as GPU.. This GPU is correctly recognized by photoscan, but when I start dense cloud I obtain this error :
BuildDenseCloud: quality = Medium, depth filtering = Aggressive
2016-12-08 19:23:34 Using device: Iris Pro, 40 compute units, 1536 MB global memory, OpenCL 1.2
2016-12-08 19:23:34 max work group size 512
2016-12-08 19:23:34 max work item sizes [512, 512, 512]
2016-12-08 19:23:34 max mem alloc size 384 MB
2016-12-08 19:23:34 work group size multiple 32, warp/wavefront size 32
2016-12-08 19:23:34 Initializing...
2016-12-08 19:23:34 Device 1
2016-12-08 19:23:34 <program source>:20035:23: warning: comparison of integers of different signs: 'size_t' (aka 'unsigned long') and 'int'
2016-12-08 19:23:34 if (get_global_id(0) < width && get_global_id(1) > crow0 && get_global_id(1) - crow0 < nrows) {
2016-12-08 19:23:34 ~~~~~~~~~~~~~~~~ ^ ~~~~~
2016-12-08 19:23:34 <program source>:20035:51: warning: comparison of integers of different signs: 'size_t' (aka 'unsigned long') and 'int'
....... [omiss] ......
~~~~~~~~~~~~~~~~ ^ ~~~~~
2016-12-08 19:23:34 <program source>:120193:52: warning: comparison of integers of different signs: 'size_t' (aka 'unsigned long') and 'int'
2016-12-08 19:23:34 if (get_global_id(0) >= ncols || get_global_id(1) >= nrows || disp[(y - row0) * ncols + (x - col0)] == FLT_MAX) return;
2016-12-08 19:23:34 ~~~~~~~~~~~~~~~~ ^ ~~~~~
2016-12-08 19:23:34 <program source>:120202:7: warning: no previous prototype for function 'atomic_cmpxchg_f32'
2016-12-08 19:23:34 float atomic_cmpxchg_f32(volatile __global float *p, float cmp, float val) {
2016-12-08 19:23:34 ^
2016-12-08 19:23:34
2016-12-08 19:23:34
2016-12-08 19:23:34 stereo.cpp line 196: error CL_BUILD_PROGRAM_FAILURE
2016-12-08 19:23:34 Finished processing in 0.076811 sec (exit code 0)
2016-12-08 19:23:34 Error: Can't initialize OpenCL
-
Hi Alexey,
Sample project link sent.
I have the issue with all the projects I have opened in PS 1.3. Including one done from scratch in PS 1.3.
Also having wierd display issues with Dense point cloud.
Here is the info on my GPU's:
(http://www.digital-mapping.net/forums/Photoscan/2016/PS13_GPU.jpg)
(http://www.digital-mapping.net/forums/Photoscan/2016/graphic_drivers.jpg)
-
I made a completely new test project with PS 1.3. Same issue with display and the texture is really messed up as you can see in this screen grab.
(http://www.digital-mapping.net/forums/Photoscan/2016/ps31_texture.jpg)
I'll send a link to this one to support too.
-
Hello mattcalavera,
We'll check this issue, but generally we do not recommend to use integrated graphics, like Intel HD, Intel Iris for processing.
-
Hello bisenberger,
We've downloaded both projects, but do not see the issues that you are referring to. Which of your graphic cards is used as a primary OpenGL rendering device?
-
This is what it shows for the AMD RADEON SETTINGS:
(http://www.digital-mapping.net/forums/Photoscan/2016/PS13_primary_GPU.jpg)
Update: Just got an email from support. They were able to reproduce the issue with the polygons on AMD Radeon graphic card.
-
Hello bisenberger,
Yes, we were able to reproduce this issue with the mesh faces on Radeon card.
As for the texture, it's a new mosaicing method. Seems that the dark spots are influenced by the photos taken from sides of the building (379-381). We'll also look at this issue and will try to improve the method.
-
Hello mattcalavera,
We'll check this issue, but generally we do not recommend to use integrated graphics, like Intel HD, Intel Iris for processing.
Thanks Alexey,
With previous version (1.2.6) The use of this graphic card gave me huge speed up! It works so well that I can't imagine not to use it in The future, also because it is the only graphic card in my portable Mac.....
Thanks
-
Hello mattcalavera,
We'll check the compatibility with OpenCL on Mac anyway. But generally, according to our experience, the speed up of the intgrated graphics is usually not coniderable.
-
Hello colleagues,
I have some problem in version 1.3 with Python API.
The problem was with call chunk.exportOrthomosaic('ortho2.tif').
I can open project in GUI and make export with no errors.
And the call of exportModel() do model export to OBJ without problems.
Script say:
ExportOrthomosaic
generating 1 x 1 tiles
Traceback (most recent call last):
File "temp.py", line 46, in <module>
chunk.exportOrthomosaic('ortho2.tif')
RuntimeError: Can't save file: ortho2.tif
There is my script:
import PhotoScan
project_path=""
output_path=""
crs = PhotoScan.CoordinateSystem("EPSG::3857")
doc = PhotoScan.app.document
doc.open("test.psx")
chunk = PhotoScan.app.document.addChunk()
chunk.addPhotos(["6360_.jpg", "6361_.jpg", "6362_.jpg", "6363_.jpg", "6364_.jpg"])
camera = chunk.cameras[0]
camera.photo.meta["Exif/FocalLength"]
chunk.matchPhotos(accuracy=PhotoScan.HighAccuracy, preselection=PhotoScan.GenericPreselection)
print ("project_path ====> ", project_path)
doc.save(project_path, chunks = doc.chunks)
doc.open(project_path)
chunk = doc.chunk
chunk.alignCameras()
chunk.buildDenseCloud(quality=PhotoScan.MediumQuality)
chunk.buildModel(surface=PhotoScan.Arbitrary, interpolation=PhotoScan.EnabledInterpolation)
chunk.buildUV(mapping=PhotoScan.GenericMapping)
chunk.buildTexture(blending=PhotoScan.MosaicBlending, size=4096)
# Save project file
doc.save(project_path, chunks = doc.chunks)
doc.open(project_path)
chunk = doc.chunk
chunk.buildOrthomosaic(surface=PhotoScan.ModelData, blending=PhotoScan.MosaicBlending, color_correction=False)
chunk.buildDem(source=PhotoScan.DenseCloudData)
doc.save(project_path, chunks = doc.chunks)
chunk.exportModel('model.obj')
chunk.exportOrthomosaic('ortho2.tif')
#chunk.exportOrthomosaic("ortho2.TIF", raster_transform=PhotoScan.RasterTransformNone, write_kml=False, write_world=False, write_alpha=True, tiff_compression=PhotoScan.TiffCompressionDeflate)
doc.save()
Thanks
-
Hello andrey.d,
It seems that you are not using full export path, so PhotoScan may be trying to save the export data to the system folder (where PhotoScan executive is placed) and has no rights for that.
-
Hi, the progress float return (Callable[[float], None]) for the buildDenseCloud has a glitch; it goes 2 times from 0 to 100 (I'm running from cli):
- Loading photos... progress goes from 0% to 100% (it should be from 0 to ~2%)
- Reconstructing depth... progress does from 0% to 70% (it should be from ~2% to 70%)
- Generating dense point cloud... progress goes from 70% to 100%
The effect is visible on the progress bar too when using the interface
-
Hello andrey.d,
It seems that you are not using full export path, so PhotoScan may be trying to save the export data to the system folder (where PhotoScan executive is placed) and has no rights for that.
Dear Alexey,
I also use
chunk.exportOrthomosaic("C:\\Users\\Andrew\\Documents\\agisoft_test\\ortho2.tif")with the similar error.
After this error, I new file ortho2.tif was created with zero size.
I test it on Agisoft Professional 1.3.0 build 3149 (64bit)
My OS - Windows 8.1
Andrew
-
Hello Andrew,
I was able to reproduce the issue, so we'll see how to fix that.
Meanwhile you need to use image_format argument:
chunk.exportOrthomosaic("D:\\temp.tif", image_format=PhotoScan.ImageFormatTIFF)
-
Hello Andrew,
I was able to reproduce the issue, so we'll see how to fix that.
Meanwhile you need to use image_format argument:
chunk.exportOrthomosaic("D:\\temp.tif", image_format=PhotoScan.ImageFormatTIFF)
Thank you
It work
-
I have version 1.2.6 installed and running on my computer (windows) and have been trying to install version 1.3 to test.
I cannot seem to be able to get it to activate.
I get "error checking out license -1" in a separate window when opening.
I've tried uninstalling older version, running the program in administrator etc without any luck.
Any ideas on what to check or try to activate??
-
Hello ARF,
In the next version new activation procedure is implemented that requires online activation (or sending offline activation request to us).
So you need to enter you license key to activation dialog while the internet connection is enabled. Let me know if you have any issues with online activation.
-
Hi Alexey,
I was connected to the internet while trying to activate. Is there something else that needs to be done?
I did see these error below in a window when opening the program.
Error checking out license
No license for product (-1)
qt.network.ssl: QSslSocket: cannot resolve TLSv1_1_client_method
qt.network.ssl: QSslSocket: cannot resolve TLSv1_2_client_method
qt.network.ssl: QSslSocket: cannot resolve TLSv1_1_server_method
qt.network.ssl: QSslSocket: cannot resolve TLSv1_2_server_method
qt.network.ssl: QSslSocket: cannot resolve SSL_select_next_proto
qt.network.ssl: QSslSocket: cannot resolve SSL_CTX_set_next_proto_select_cb
qt.network.ssl: QSslSocket: cannot resolve SSL_get0_next_proto_negotiated
-
Hello ARF,
And what happen after you enter the activation key?
The first two lines means that the product is not activated, while QSslSocket lines means that SSL is not installed, but it's just a warning message related to Check for Updates function.
-
Thanks for quick reply.
Once the code is entered i get this error:
Error checking out license
No license for product (-1)
-
Hello ARF,
Just to confirm, have you pressed OK button after the key is entered? In case it is not active (as it may happen sometimes in the build 3149), you need to switch radio button in the activation dialog to Continue in Demo mode and back again.
-
Hi,
yes I've pressed the ok button and then get a small window with "unexpected error occurred while activating..."
-
Hello ARF,
In the next version new activation procedure is implemented that requires online activation (or sending offline activation request to us).
Does the online activation work with a corporate proxy ?
Is the offline version necessitating admin privileges ?
-
Hello ARF,
Could you please check that antivirus/firewall is not blocking PhotoScan connection?
Also could you please try to activate the license from the command line using --activate argument, as it should contain additional information, like:
photoscan --activate 12345-12345-12345-12345-12345
-
Hello JRM,
In case the proxy is used for the Internet connection, the following system environment variable should be set up (for the current build 3149):
HTTP_PROXY=hostname:portThe properties of the proxy in Windows can be checked in Control Panel -> Network and Internet -> Internet Options -> Connections -> LAN Settings -> "Use a proxy server for your LAN".
In case the proxy is requiring the autentification, then additional variable should be used:
HTTP_PROXY_CREDENTIALS=user:passwordAfter setting up these variables it is recommended to reboot and re-run PhotoScan.
As for offline activation, usually it doesn't require administrator privileges, providing that current used has access write rights to Program Data folder and activation request file is not saved to the system folder that requires special permissions.
-
Hi Alex,
I've checked to confirm no proxy is being used.
I've turned off virus and firewall during activation.
I've tried the command line activation but get "syntax" error. I'm not familiar with using the command line so might not be doing it correctly but I just entered "photoscan --activate 12345-12345-12345-12345-12345" with my code ofcourse.
-
Hello ARF,
You need to enter the command from the OS terminal, not from PhotoScan console pane itself.
On Windows you need to open terminal using cmd, then browse to PhotoScan Pro installation directory and run the command form there.
-
I've run CMD as adminstrator and got the following error:
C:\Program Files\Agisoft\PhotoScan Pro>photoscan --activate and correct code
Error requesting activation
Read error from network (-105)No such file or directory (errno: 2)
-
with the command line feature of v1.3 it would be nice to submit the PS model build to a GPU node on a cluster. the issue is that when you submit a job to a cluster you dont know which GPU node will be assigned the job. Since I cant install the license on every GPU node, I get a license error as the result of my run. ( the PS install itself is in a container which is executed on the node). Does anyone know of options to allow for this type of processing?
Hello MManning,
Seems that I've missed your post, can you provide additional details regarding the issue? How you are starting the processing?
-
I've run CMD as adminstrator and got the following error:
C:\Program Files\Agisoft\PhotoScan Pro>photoscan --activate and correct code
Error requesting activation
Read error from network (-105)No such file or directory (errno: 2)
Hello ARF,
Seems strange, I still assume that it may be related to the proxy/antivirus/firewall settings, but if you are sure it is not the case, I can suggest to perform offline activation, to do that please execute the following command from CMD:
C:\Program Files\Agisoft\PhotoScan Pro>photoscanphotoscan --activate-offline <key> <path/request_file.act>path to the activation request file (*.act) should be accessible with read/write rights, so it's better not to save to the Program Files folder.
Then send the *.act file to support@agisoft.com (don't post it here) and will send you back the license file to be put to PhotoScan installation directory.
-
Shift+scroll wheel in model view changes camera size, nice. :)
-
Photoscan stuck at the end of processing Dense Cloud. Check the time and date in console against time in bottom right of screen capture to see how long it has been in this state.
(http://www.digital-mapping.net/forums/Photoscan/2016/PS13_dense_cloud.jpg)
-
Hello bisenberger,
It seems that some sub-part of the depth maps estimation has been sent to GPU and was never solved, while all the rest has been calculated by PhotoScan.
I can suggest to update graphic card drivers. If you are using integrated graphics, like Intel HD, it might be reasonable to switch them off.
-
I am testing 1.3 again after going back to 1.26.
I am scanning fist sized skulls and bones, and my workflow usually involves cleaning up the mounts and supports that I use to photograph the skulls on in the sparse cloud, and then building the dense cloud.
In 1.26, this works well. I get the sparse cloud cleaned up so that just the skull is visible, then create a dense cloud and the skull is reconstructed.
In 1.3, this doesn't work. I get the sparse cloud clean but the dense cloud recreates all of the supports and other objects that I edited out in the sparse cloud. These are significantly more difficult to remove in the sparse cloud (almost impossible...) In 1.3, there are also a lot of odd black objects created around the edges of my model and also in space - this sometimes happens in 1.26 but is many times worse in 1.3
Both processed with depth filtering on mild, high accuracy paired alignment, medium dense cloud.
See screenshots for examples.
-
Hello bisenberger,
It seems that some sub-part of the depth maps estimation has been sent to GPU and was never solved, while all the rest has been calculated by PhotoScan.
I can suggest to update graphic card drivers. If you are using integrated graphics, like Intel HD, it might be reasonable to switch them off.
Graphic card drivers are up to date. Will look into disabling Intel HD although it is not listed under the PhotoScan Preferences GPU tab.
-
I am having problems generatin reports, when the process gets to 80% it locks and wont finish. i cannot cancell the process and have to "kill" agisoft in the task manager.
does anuone have that problem to? and how do i fix it?
in advance thank you.
/Soren
-
Hello Soren,
I can suggest to update to the build 3149 of the version 1.3.
This issue is the known bug in the build version that you are using.
-
Thank you for the quick reply! :-) I will try that
-
Hello bisenberger,
Can you then provide the complete log (console output) from the operation that was stuck in the end of depth maps generation stage?
-
I am testing 1.3 again after going back to 1.26.
I am scanning fist sized skulls and bones, and my workflow usually involves cleaning up the mounts and supports that I use to photograph the skulls on in the sparse cloud, and then building the dense cloud.
In 1.26, this works well. I get the sparse cloud cleaned up so that just the skull is visible, then create a dense cloud and the skull is reconstructed.
In 1.3, this doesn't work. I get the sparse cloud clean but the dense cloud recreates all of the supports and other objects that I edited out in the sparse cloud. These are significantly more difficult to remove in the sparse cloud (almost impossible...) In 1.3, there are also a lot of odd black objects created around the edges of my model and also in space - this sometimes happens in 1.26 but is many times worse in 1.3
Hello mrc,
The sparse cloud is used only to estimate the scene depth and overlapping image pairs, the dense cloud generation stage then begins with the depth maps generation for the overlapping images.
Looking on the screenshots I assume that in the version 1.2.6 the masks were applied to the images, while in the version 1.3 there are no masks that covers all the background elements.
-
Hi Alexey,
I am fairly certain the masking did not cover the mounts in the 1.26 reconstruction. Unfortunately I no longer have that reconstruction but I'll reinstall 1.26 and leave it to process overnight and see what happens.
I generally only mask the backgrounds and not the mounts, so it would have been unusual for me to mask them out in this model.
I have since processed another model and there are lots of artefacts (mostly black chunks at margins and in space) in 1.3 that are not there in 1.26.
It is sort of like in 1.26 if there was something in some of the photos and not others it is not reconstructed in the dense cloud, but in 1.3 it is reconstructed. The mounts in the screenshot were only there for about 1/4 of the photos and then the object was flipped over and a new mount was placed.
Thanks
Michael
-
Hi Alexey,
I ran the reconstruction overnight. The images are not masked, except for the black background. I have attached 2 of the images below.
The result is shown in the screenshots. In 1.26 the dense cloud has none of the stands and props. In 1.3, they are all there along with lots of artefacts.
Please see attached.
Thanks
Michael
-
Hello,
I prepare csv log file for my dataset.
It look like this
IMG_2883.JPG,45.3594005,26.712383,431.32,255.82,11.31,-1.95
IMG_2884.JPG,45.3593339,26.7119093,434.13,262.28,2.12,-1.96
IMG_2885.JPG,45.3592974,26.7114327,432.92,267.82,0.48,-2.77And call loadReference() with it. And coordinats added as well.
But by default coordinates in csv file assumed as local coordinates.
How I can set other CS, e.g. EPSG:4326?
Part of code below:
crs = PhotoScan.CoordinateSystem("EPSG::3857")
doc = PhotoScan.app.document
doc.open(project_path)
chunk = PhotoScan.app.document.addChunk()
chunk.addPhotos(images_list)
camera = chunk.cameras[0]
chunk.loadReference(out_csv, format=PhotoScan.ReferenceFormatCSV, columns="nxyzabc", delimiter=",", group_delimiters=False, skip_rows=0)
doc.save(project_path, chunks = doc.chunks)
Tanks
-
Hello Michael,
That seems quite strange. Can you send any small sample dataset (project with alignment + images) that can be used to reproduce the issue?
-
Hello andrey.d,
You can select different coordinate system for the chunk after (or before) loading the data from CSV:
crs = PhotoScan.CoordinateSystem("EPSG::4326")
chunk.crs = crs
or use it directly:
chunk.crs = PhotoScan.CoordinateSystem("EPSG::3857")
-
It is work. Great!
-
Hello bisenberger,
Can you then provide the complete log (console output) from the operation that was stuck in the end of depth maps generation stage?
Here's a link Alexey:
http://www.digital-mapping.net/forums/Photoscan/2016/log_2016-12-08.zip
I did some other stuff after it got stuck, so it shouldn't be the last thing in the log.
-
PS120_split_in_chunks_dialog.py not working with version 1.3. :(
http://wiki.agisoft.com/wiki/Python
-
Hello bisenberger,
I'm posting the updated script for Split in Blocks, hope it works now. If not - let me know.
The main differences are that now PySide2 module should be imported instead of PySide and some interface elements have been moved to QtWidgets from QtGui class. But most of the changes can be found in Python API reference in the first post.
-
Hello bisenberger,
Can you then provide the complete log (console output) from the operation that was stuck in the end of depth maps generation stage?
Here's a link Alexey:
http://www.digital-mapping.net/forums/Photoscan/2016/log_2016-12-08.zip
I did some other stuff after it got stuck, so it shouldn't be the last thing in the log.
Thanks for the log.
It looks like a critical error within OpenCL appeared ("Unable to recover OpenCL context!" message), but I think that it should be solved already in the next update of 1.3.
-
Good news on the error being resolved in the next update. Also thanks for updating the split in blocks script. I'll let you know if I have a problem.
-
Ran into this issue batch processing dense point cloud:
2016-12-28 12:40:58 ocl_engine.cpp line 433: clEnqueueNDRangeKernel failed, CL_MEM_OBJECT_ALLOCATION_FAILURE
2016-12-28 12:40:58 GPU processing failed, switching to CPU mode
2016-12-28 12:40:58 Unable to recover OpenCL context!
2016-12-28 12:40:58 [CPU] estimating 2018x8417x320 disparity using 1009x1053x8u tiles, offset -28
In [1]: import PhotoScan
Link to log file:
http://www.digital-mapping.net/forums/Photoscan/2016/log_2016-12-08b.zip (http://www.digital-mapping.net/forums/Photoscan/2016/log_2016-12-08b.zip)
-
I prefer the previous mouse cursor for marker placement. It came to a point at the actual pick location and did not obscure the pick location. The new cursor surrounds the pick location with a circle and tics marks making it difficult to see the actual pick location.
-
Now it's even more difficult to use this latest beta because of registration issues. When I try to enter my key the app shows the app just shows the message from the included image.
I'musing administrator user.
Cheers
-
Hello RHenriques,
Are you using Sierra? The issue seems to be similar to the one described here:
https://eclecticlight.co/2016/06/16/macos-sierra-will-break-many-installers-and-updaters/
So it is suggested to move the application from the unpacked dmg first before running it.
-
Hi Alexey
I did it from a folder into the Applications folder. I just tried to enter as root user and the same happened.
I'm going to try to ihnibit gatekeeper to see is there's a difference.
Cheers
PS:No luck, even with Gatekeeper turned off.
-
It's not important but i guess you should know.
Clear log in console pane not work in classic theme. :D
-
Hi
On build 3520 when adding photos using chunk.addPhotos()only the first photo of the list is added (I see that there is a new argument to add photos, but changing it doesn't change the result)
Thanks
-
A small but annoying bug; PhotoScan icons are not appearing, just generic windows icon. I am running win 10 and PS 1.3.0 3520 (64bit)
Bug was in Win 10, not PS.
-
No GPU is found anymore!
With PS 1.3.0 3520 the settings window for the GPUs is empty and console says that my graphic cards do not support SPIR.
I am on MacBookPro mid 2012 with a NVIDIA GeForce GT 650M 512 MB.
-
Hello Airgo,
You might need to install CUDA drivers, as they are not packed with the standard graphic card drivers for Mac OS X:
http://www.nvidia.com/object/mac-driver-archive.html
-
Hello Alexey,
thank you. And should I work with the NVIDIA Web Driver or with the Default OS X Graphics Driver?
Which one will deliver better (faster) results?
-
Hello Airgo,
Actually not sure, if it does matter. If you are able to see the GPU device in PhotoScan preferences with OS X default driver, then it's fine.
-
Hello Alexey,
I don't understand clearly this window so shall i check on GPU drivers to accelerate the process.
Thank you,,
-
Hello tornado2011,
In the GPU devices list you need to choose the devices that you wish to be used in the processing. The check box in the bottom defines whether the CPU should be used in parallel to GPU (the check box is on) or the processing should be perforemed on GPU only. Note that even if you uncheck the CPU option, there still will be some activity on CPU, as it controls GPU jobs.
In the latest update GPU is used at two processing steps: building depth maps and matching photos.
-
Hello Alexej,
I see the NVIDIA GPU in both settings, Max OS X oder NVIDIA Web driver, so it seems really not to make any difference (maybe I will drive some tests for speed).
But what is gone is the Intel HD Graphics 4000. Can it be reactivated as well?
-
Hello Airgo,
For Intel HD you need to update drivers (OpenCL), if it's possible on Mac OS X.
-
Hi Alexey,
Thanks for your quick answer.
I also face an issue while i put some GCPs on one of my recent projects i made Optimization, after that i decided to put some additional markers but when i place the marker on the specific area it jumped couple of pixels and i have to correct the position again manually.
Thank you,
Mohammed
-
Hello,
thanks for this new beta release, it seems that the key finding and matching use GPU now, the gain in time is considerable, but I noticed some differences in the result quality comparing to the pure CPU method.
Point color seems to be not correct and it is very difficult (sometime impossible) to have all the images matched with GPU when there is no problem with CPU, so the new method seems to be more demanding about shooting quality.
If this issue is not solvable, you could consider to give the possibility to use GPU only in dense reconstruction in case of difficulty with GPU matching.
Thanks.
Pascal
-
Hi all,
yes it seems that with latest release, the matching algorithm does not find same number of Tie points as in previous release.
I have a data set of 69 photos on which I performed matching with two latest releases and number of matches varies quite a bit.
Here is result of previous release Matching:
Point Cloud
Puntos 87,459 de 121,950
RMS error de reproyección 0.135979 (0.75202 pix)
Error de reproyección máximo 1.25603 (7.50969 pix)
Tamaño promedio de puntos característicos 4.92251 pix
Superposición efectiva 3.02671
Parámetros de orientación
Precisión Alta
Pre-procesar emparejamiento de imágenes Referencia
Puntos claves por foto 40,000
Puntos de enlace por foto 0
Restricción de máscara activa No
Adaptativo ajuste del modelo de cámara No
Tiempo búsqueda de puntos homólogos 5 minutos 7 segundos
Tiempo de orientación 42 segundos
Parámetros de optimización
Parámetros f, b1, b2, cx, cy, k1-k3, p1, p2
Fit rolling shutter No
Tiempo de optimización 1 segundo
and here is matching result for latest release:
Point Cloud
Points 61,346 of 90,573
RMS reprojection error 0.109892 (0.660971 pix)
Max reprojection error 0.453098 (6.862 pix)
Mean key point size 5.06373 pix
Effective overlap 2.87035
Alignment parameters
Accuracy High
Generic preselection Yes
Reference preselection Yes
Key point limit 40,000
Tie point limit 0
Adaptive camera model fitting No
Matching time 2 minutes 26 seconds
Alignment time 22 seconds
Optimization parameters
Parameters f, b1, b2, cx, cy, k1-k3, p1, p2
Fit rolling shutter No
Optimization time 1 seconds
Matching time is much faster but with lower number of points...
The question is what is the impact of alignment precision?
-
I prefer the previous mouse cursor for marker placement. It came to a point at the actual pick location and did not obscure the pick location. The new cursor surrounds the pick location with a circle and tics marks making it difficult to see the actual pick location.
Maybe just get rid of the circle with hatches and leave the center tic.
-
I did it from a folder into the Applications folder. I just tried to enter as root user and the same happened.
I'm going to try to ihnibit gatekeeper to see is there's a difference.
Cheers
PS:No luck, even with Gatekeeper turned off.
Hello RHenriques,
Do you get similar errors all the time, with those strange paths?
-
Hi Alexey
Yes, same thing. I tested everything. The same error occur all the time. Sierra duplicates Photoscan to that folder that appears in the path but the folder seems not be accessible for write privileges, even for the root user. I guess that the possible solution is to generate a .lic file outside the application. Maybe using an invisible writable folder in a shared folder or in the user's home folder.
Cheers
-
Hello tornado2011,
In the GPU devices list you need to choose the devices that you wish to be used in the processing. The check box in the bottom defines whether the CPU should be used in parallel to GPU (the check box is on) or the processing should be perforemed on GPU only. Note that even if you uncheck the CPU option, there still will be some activity on CPU, as it controls GPU jobs.
In the latest update GPU is used at two processing steps: building depth maps and matching photos.
Hi Alexey,
You say the GPU is now used for matching photos. The GPU's in my computer are inactive during this process. I have two AMD graphic cards with the latest drivers as indicated in an earlier post.
(http://www.digital-mapping.net/forums/Photoscan/2016/ps31_matching.jpg)
(http://www.digital-mapping.net/forums/Photoscan/2016/PS13_GPU.jpg)
-
Hello bisenberger,
It seems that you are using previous pre-release version. The latest one has build number 3520.
-
Hello bisenberger,
It seems that you are using previous pre-release version. The latest one has build number 3520.
Hi Alexy,
Version 3520 is the version I'm running.
(http://www.digital-mapping.net/forums/Photoscan/2016/ps31_matching2.jpg)
The display problem I reported previously has been fixed in this version. The GPU's not being used with the AMD GPU's are with build 3520.
-
Hello bisenberger,
GPU tab in the latest build should look different from the screenshot you've provided earlier.
Could you check once again that you are using latest build when opening GPU tab and processing the data?
-
GPU Nvidia Work fine on Build 3520
(http://i68.tinypic.com/2qi0oec.jpg)
-
Hello bisenberger,
GPU tab in the latest build should look different from the screenshot you've provided earlier.
Could you check once again that you are using latest build when opening GPU tab and processing the data?
Hi Alexy, I just linked to an image in a previous post to show my GPU's (it was from the earlier build). This is from build 3520. Um.... they are not checked let me try that.
(http://www.digital-mapping.net/forums/Photoscan/2016/ps31_matching_GPU.jpg)
-
I received this error message after checking the boxes for the GPU's under preferences.
(http://www.digital-mapping.net/forums/Photoscan/2016/ps31_matching3.jpg)
I copied the following from the console:
2017-01-03 12:01:18 AlignPhotos: accuracy = Medium, preselection = reference, keypoint limit = 40000, tiepoint limit = 4000, adaptive fitting = 0
2017-01-03 12:01:18 Detecting points...
2017-01-03 12:01:19 Using device: Pitcairn, 16 compute units, 1024 MB global memory, OpenCL 2.0
2017-01-03 12:01:19 max work group size 256
2017-01-03 12:01:19 max work item sizes [256, 256, 256]
2017-01-03 12:01:19 max mem alloc size 576 MB
2017-01-03 12:01:19 wavefront width 64
2017-01-03 12:01:19 Using device: Tahiti, 32 compute units, 3072 MB global memory, OpenCL 2.0
2017-01-03 12:01:19 max work group size 256
2017-01-03 12:01:19 max work item sizes [256, 256, 256]
2017-01-03 12:01:19 max mem alloc size 2112 MB
2017-01-03 12:01:19 wavefront width 64
2017-01-03 12:01:20 photo 264: 22945 points
2017-01-03 12:01:20 photo 267: 19902 points
2017-01-03 12:01:20 photo 268: 20923 points
2017-01-03 12:01:21 photo 269: 25710 points
2017-01-03 12:01:21 photo 270: 21050 points
2017-01-03 12:01:21 Device 1
2017-01-03 12:01:21 Program build log:
2017-01-03 12:01:21 Error: Undeclared function index 1025
2017-01-03 12:01:21
2017-01-03 12:01:21
2017-01-03 12:01:22 photo 271: 23660 points
2017-01-03 12:01:23 Device 1
2017-01-03 12:01:23 Program build log:
2017-01-03 12:01:23 Error: Undeclared function index 1025
2017-01-03 12:01:23
2017-01-03 12:01:23
2017-01-03 12:01:23 Finished processing in 5.384 sec (exit code 0)
2017-01-03 12:01:23 Error: CL_BUILD_PROGRAM_FAILURE at line 520
-
GPU Nvidia Work fine on Build 3520
Cool! Look forward to seeing my AMD's doing that.
-
Hello bisenberger,
We'll try to reproduce this issue on AMD cards.
-
Hello bisenberger,
We'll try to reproduce this issue on AMD cards.
Thanks Alexey,
Some cool stuff happening in this version, as usual.
-
I noticed after installation of latest version processing was slow. When I checked it had unselected the GPU in the settings.
I'm assuming as the GPU settings look different and there isn't an option to choose the number of cores that agisoft sorts this automatically?
-
I have a project that I tried batch processing two chunks, make dense cloud, settings medium/aggressive. The process never finishes. My GPU and CPU go idle. PS 1.3.0 build 3520. As indicated in previous post I have two AMD GPU's.
Compare the time in the consul compared to the time on the taskbar.
(http://www.digital-mapping.net/forums/Photoscan/2016/ps13_stalling.jpg)
-
Hello bisenberger,
Is it reproducible on any project? Does it happen only in batch mode or also by using the regular Workflow.
-
Also getting:
Error: CL_BUILD_PROGRAM_FAILURE at line 520
With ATI GPU enabled in the new build.
-
Hello bisenberger,
Is it reproducible on any project? Does it happen only in batch mode or also by using the regular Workflow.
Hi Alexey,
The only thing that I've had similar to this is the one I reported on Reply #153 on: December 22, 2016, 03:29:11 PM in this thread.
-
Hello bisenberger,
Is it reproducible on any project? Does it happen only in batch mode or also by using the regular Workflow.
Hi Alexey,
The only thing that I've had similar to this is the one I reported on Reply #153 on: December 22, 2016, 03:29:11 PM in this thread.
so the issue is not reproducible even on the same project and just happens sometimes?
P.S. also the build 3520 should contain the mosaic blending improvements (the issues observed in shed texturing project should be fixed now). as well as the mesh visualization issues in shaded/solid/wireframe modes on ATI/AMD cards.
-
Also getting:
Error: CL_BUILD_PROGRAM_FAILURE at line 520
With ATI GPU enabled in the new build.
Hello mrc,
The issue for ATI/AMD cards should be fixed in the next version update.
-
Hello bisenberger,
Is it reproducible on any project? Does it happen only in batch mode or also by using the regular Workflow.
Hi Alexey,
The only thing that I've had similar to this is the one I reported on Reply #153 on: December 22, 2016, 03:29:11 PM in this thread.
so the issue is not reproducible even on the same project and just happens sometimes?
P.S. also the build 3520 should contain the mosaic blending improvements (the issues observed in shed texturing project should be fixed now). as well as the mesh visualization issues in shaded/solid/wireframe modes on ATI/AMD cards.
Hi Alexey,
It happened twice on the same project. I ended up going back to version 1.2.6 to finish processing the project. I will try to reproduce later.
I noticed the fixes to aforementioned issues, thanks. :)
-
Hello bisenberger,
Is it reproducible on any project? Does it happen only in batch mode or also by using the regular Workflow.
Hi Alexey,
The only thing that I've had similar to this is the one I reported on Reply #153 on: December 22, 2016, 03:29:11 PM in this thread.
so the issue is not reproducible even on the same project and just happens sometimes?
P.S. also the build 3520 should contain the mosaic blending improvements (the issues observed in shed texturing project should be fixed now). as well as the mesh visualization issues in shaded/solid/wireframe modes on ATI/AMD cards.
It is repeatable. I have tried 2 more times to batch create a dense cloud with the project with the same results. The project has 17 chunks so it would be a major pain to try to upload. I can process the project with version 1.2.6 without issue.
-
Hello bisenberger,
And it happens on the same chunk or on a random one?
-
I get this error during the alignment stage (1.3.3520)
2017-01-11 11:35:40 AlignPhotos: accuracy = High, preselection = generic, reference, keypoint limit = 60000, tiepoint limit = 6000, adaptive fitting = 1
2017-01-11 11:35:40 Detecting points...
2017-01-11 11:35:41 Using device: Pitcairn, 16 compute units, 1024 MB global memory, OpenCL 2.0
2017-01-11 11:35:41 max work group size 256
2017-01-11 11:35:41 max work item sizes [256, 256, 256]
2017-01-11 11:35:41 max mem alloc size 576 MB
2017-01-11 11:35:41 wavefront width 64
2017-01-11 11:35:42 photo 0: 59996 points
2017-01-11 11:35:43 Device 1
2017-01-11 11:35:43 Program build log:
2017-01-11 11:35:43 Error: Undeclared function index 1025
2017-01-11 11:35:43
2017-01-11 11:35:43
2017-01-11 11:35:43 photo 2: 59992 points
2017-01-11 11:35:43 Finished processing in 2.602 sec (exit code 0)
2017-01-11 11:35:43 Error: CL_BUILD_PROGRAM_FAILURE at line 520
-
Hello lunar,
This is a known issue for ATI/AMD cards related to GPU-based image matching. Will be fixed in the next update.
-
after reinstall - everything work fine
-
lunar - are your GPUs still ticked under preferences now that it is working?
-
sorry i just saw that i reinstall to 1.3.3149
in 1.3.3520 GPU OpenCl -> Pitcairn check "on"
-
When I try to activate the PS 1.3 I get the following error messages:
2017-01-11 15:19:06 Error requesting activation
2017-01-11 15:19:06 The activation key you supplied has already been used to activate a license.
2017-01-11 15:19:06 Please check the key and ensure that you have entered it correctly.
2017-01-11 15:19:06 Details: No license for product (-1)
2017-01-11 15:19:06 Activation key already used (-1005)
I am on a Mac.
-
Hello Airgo,
Have you already used this key for the version 1.3 activation? Probably with the previous pre-release?
The license file may be moved to some other location where PhotoScan can no longer find it. You can try to install the previous pre-release version and if it appears activated, then deactivate it.
-
Hello Alexey,
yes I have used this key for the previous pre-release. But activation did not work there due to problems writing the activation file.
And I have a new SSD in my computer with a clean install. So there is no old stuff on the SSD. Where should the license file be stored?
Unfortunately I do not have the the previous pre-release to deactivate it.
Another problem: I wanted to align photos for a spherical panorama and get this errors:
2017-01-11 15:26:23 AlignPhotos: accuracy = High, preselection = generic, reference, keypoint limit = 40000, tiepoint limit = 4000, adaptive fitting = 1
2017-01-11 15:26:23 Detecting points...
2017-01-11 15:26:23 Using device: GeForce GT 650M, 2 compute units, 511 MB global memory, CUDA 3.0
2017-01-11 15:26:23 max work group size 1024
2017-01-11 15:26:23 max work item sizes [1024, 1024, 64]
2017-01-11 15:26:23 HD Graphics 4000: no SPIR support
2017-01-11 15:26:23 GeForce GT 650M: no SPIR support
2017-01-11 15:26:25 photo 0: 39936 points
2017-01-11 15:26:25 Finished processing in 2.03725 sec (exit code 0)
2017-01-11 15:26:25 Error: out of memory at line 177
PS did quit the operation.
Thanks.
-
Hello Airgo,
Can you please send the order number or purchase data and reseller name to support@agisoft.com, so that we could check the license status?
IntelHD is currently not supported for image matching, you need to disable it in the GPU tab, as for the GeForce GT 650M, do you have latest drivers installed?
-
Hello Alexey,
the Intel HD does not show up at all in my GPU list so I can not disable it. A Driver update is not available for Mac.
The Geforce GT 650M does have the latest driver installed.
-
Hello Airgo,
Ok, thanks for the information. Actually, SPIR support line should be even displayed in the console for NVIDIA cards, as for them PhotoScan is using CUDA now.
But as for the spherical data processing, we have already found the issue and will also include the fix in the next update.
-
Hi Alexey
It seems that Airgo problem with the license is the same that I'm facing. Do you have any successful story of version 1.3 beta installation in MacOS Sierra?
Cheers
-
Hello
I am trying to duplicate the processing of images in batch mode that my researcher does via the UI. She follows this path:
- load images
- build dense cloud
- build mesh (buildModel)
- build texture
- build orthomosaic[/right][/right]
However when I hit the buildTexture step in my batch process I get this error:
RuntimeError: Missing uv coordinates
In the example in the Photoscan 1.3 docs it show that buildUV step is before buildTexture. Is this a requirement for buildTexture? My researcher says that she does not perform a buildUV-type step in the UI that she is aware of.
Any ideas on what option I might be missing on my build steps?
Thanks
-
I'm trying to compare PS version 1.2.6 and 1.3.0.
I Use a workspace of 416 photos from inspire 1|X5
in version 1.2.6 all 416 photos correctly aligned
in version 1.3.0 (with the same alignment settings) only 211 photos aligned, and nothing change also if I use maximum alignment settings (Highest accuracy, etc)
Why?
Thanks,
Matt
-
Hello Matt,
Can you provide the processing logs from both tries and also the report files generated from both versions to support@agisoft.com?
-
Hello
I am trying to duplicate the processing of images in batch mode that my researcher does via the UI. She follows this path:
- load images
- build dense cloud
- build mesh (buildModel)
- build texture
- build orthomosaic[/right][/right]
However when I hit the buildTexture step in my batch process I get this error:
RuntimeError: Missing uv coordinates
In the example in the Photoscan 1.3 docs it show that buildUV step is before buildTexture. Is this a requirement for buildTexture? My researcher says that she does not perform a buildUV-type step in the UI that she is aware of.
Any ideas on what option I might be missing on my build steps?
Thanks
Hello MManning,
Do you mean Python scripting?
If so, then the GUI Align Photos option consist of two Python operations: .matchPhotos() and .alignCameras(). Similar is with the GUI Build Texture option - in Python it is split to .buildUV() and .buildTexture().
-
Hi Alexey
thanks for the response. When I add in the buildUV step before the buildTexture:
doc = PhotoScan.app.document
doc.open("...")
chunk = doc.chunk
chunk.buildUV(mapping=PhotoScan.GenericMapping)
chunk.buildTexture( color_correction=True )
chunk.buildOrthomosaic( surface=PhotoScan.DataSource.ModelData )
doc.save()
I get an error :
'blending textures... *********************************** frames quality estimated in 1.23499 sec', '*********************************** done in 13.5811 sec', 'postprocessing atlas... done in 0.011684 sec', 'BuildOrthomosaic: projection type = Planar, blending mode = Mesh, surface = Mosaic, resolution = 0', 'Please save project in PSX format before processing', 'Traceback (most recent call last):', ' File "/global/home/users/mmanning/actest/commandscript.sh", line 10, in <module>', ' chunk.buildOrthomosaic( surface=PhotoScan.DataSource.ModelData ) ', 'RuntimeError: Empty frame path'
Im not sure what to do with this, I see only one reference to a PSX in the 1.3 ref documentation. Do I need to export the model into a PSX then run build texture? Is the "Empty frame path" message related or something different?
Thank you for your assistance, still learning Photoscan.
Maurice
-
Hello MManning,
According to the log the texture is built. But you need to save the project in PSX format to build orthomosaic (it's not connected anyhow to the texture).
-
Hello Matt,
Can you provide the processing logs from both tries and also the report files generated from both versions to support@agisoft.com?
Alexey, mail sent....
-
Hello MManning,
According to the log the texture is built. But you need to save the project in PSX format to build orthomosaic (it's not connected anyhow to the texture).
ok, after reading the docs I understand the difference between a psz and psx a little bit better however Im not clear, when using the python sdk, how to save to each format.
- It seems from the photoscan python api document that doc.save() seems to create a .psz by default, is that true?
- if I just name the file with the project extension ( doc.save("project.psx" ) will photoscan save to the correct (project vs archive) format?
- is Model.saveTexture("project.psx") preferred way to save saving before building the orthomosaic?
- is it true that archive (psz) format is the best format for holding the full set of information (images, settings, etc) for long term?
Again, thanks for your help and patience.
Maurice
-
Hello Maurice,
If you pass to doc.save() function the path to the .psx project, it will be saved in psx format. For example: doc.save("D://projects/new_project.psx")
So the function is checking the extension in the provided path to define the save format.
Model.saveTexture("project.psx") is meaningless, you need to put doc.save(some_psx_path) line before generating the orthomosaic.
PSZ - is a single archive file, while PSX is the folder structure. PSZ cannot contain orthomosaic, individually orthorectified images, DEM and tiled model. Also when PSZ file is opened it is fully loaded to the memory, whereas when working with PSX format PhotoScan will load only required project element.
-
Hello Matt,
Can you provide the processing logs from both tries and also the report files generated from both versions to support@agisoft.com?
Alexey, mail sent....
Hello Alexey, this night I made another test, with version 1.3.0 disabling GPU, and yes without GPU it align correctly all the photos, so I think the problem is with the align algoritm that use the GPU..... I sent a mail with the three reports and logs ( 1.2.6 with GPU, 1.3.0 with GPU, and 1.3.0 Without GPU)....
Let me know...
thanks,
Matt
-
Hello Matt,
Thanks, we are checking it.
-
Hello Matt,
We have requested the sample dataset, but it seems that it's no longer needed, as we were able to reproduce the problem on multi-GPU setup.
Could you please also check the alignment on your system with only one of two GPU devices one, just to be sure that it works properly in your case?
-
It seems that Airgo problem with the license is the same that I'm facing. Do you have any successful story of version 1.3 beta installation in MacOS Sierra?
Hello RHenriques,
It was a different issue, that was related to the hardware change.
We've tried to reproduce your problem on MacOS Sierra (clean install), but it worked as expected for current pre-release build 3520 - both for Applications and random installation directory, so I assume that there might be some settings in OS that moves the app bundle to AppTranslocation folder.
Aren't you running PhotoScan from .dmg? You need to copy PhotoScan Pro application bundle from the dmg somewhere before running it.
-
Hello all,
we have updated pre-release build to 3620.
solved issues mentioned in this thread:
- crosshair cursor optionality in the Appearance tab - http://www.agisoft.com/forum/index.php?topic=5976.msg30583#msg30583
- clear log button issue in the Console pane - http://www.agisoft.com/forum/index.php?topic=5976.msg30610#msg30610
- addPhotos() issue for Python scripting - http://www.agisoft.com/forum/index.php?topic=5976.msg30612#msg30612
- image matching issue on AMD cards (OpenCL) - http://www.agisoft.com/forum/index.php?topic=5976.msg30651#msg30651
- image matching issue on multi-GPU layout - http://www.agisoft.com/forum/index.php?topic=5976.msg30780#msg30780
-
Hello,
ist there a way to change the font size of the console?
In 1.3 it is much bigger than in 1.2
-
Hello all,
we have updated pre-release build to 3620.
So far everything perfect with this new version, excellent work and support.
Thanks
-
In a quick test the extraction of points is much better in version 1.3, What I do not understand is because the processing speed values are so different between versions.
v1.2.6
Device 1 performance: 141.395 million samples/sec (CPU)
Device 2 performance: 423.937 million samples/sec (GeForce GTX TITAN X)
Total performance: 565.332 million samples/sec
Finished processing in 60.611 sec
v1.3.0
Device 1: 17.0262 million samples/sec (CPU)
Device 2: 274.403 million samples/sec (GeForce GTX TITAN X)
Total performance: 291.43 million samples/sec
Finished processing in 42.773 sec
v1.2.6 = 565.332 million samples/sec vs v1.3.0 = 291.43 million samples/sec
v1.2.6 real time = 60.611 sec vs v1.3.0 real time = 42.773 sec
Apparently better in version 1.2.6 but the truth is that version 1.3.0 is much faster, but apparently numerically slower
Because now the performance numerically shows slower, but in time is faster?
-
Hello Diego,
The values of million samples per second cannot be compared between 1.3 and older versions now.
-
ist there a way to change the font size of the console?
In 1.3 it is much bigger than in 1.2
If you are on Mac OS X, you can use CMD + "minus" and CMD + SHIFT + "plus".
-
Thank you.
For the sake of completeness: Bigger font size is just CMD+"PLUS"
-
Hello Diego,
The values of million samples per second cannot be compared between 1.3 and older versions now.
Hello Alexey,
Thanks for the clarification. Excellent support
-
Hello ,
I am testing 1.3.0 versions , and at least in the previous and last version , viewport doesn't work.
No grid , no models , points or anything.
Here is a screenshot of the last version https://www.dropbox.com/s/g6ogupqosylf63p/photoscan.PNG?dl=0
Viewport stays white-blank regardless what i do (align,dense cloud , load previous project etc)
Probably some gpu/drivers issue?
As it can been seen at the screenshot , gpu is nvidia gtx 660m (latest drivers) (optimus intel/nvidia) and is detected by the program and is used just fine during align/dense cloud.
Thanks in advanced.
-
Hello mphxths,
What happens when you try to open a model or point cloud in the model view for some existing project?
-
Hello mphxths,
What happens when you try to open a model or point cloud in the model view for some existing project?
Hello ,
Nothing happens. Check following screenshot https://www.dropbox.com/s/rangsuuqhubwm0p/photoscan2.PNG?dl=0
Project is loaded , but viewport stays empty. Selecting dense cloud , mesh (shaded,solid,wireframe) doesn't change anything in viewport.
Tried with 4-5 different project , same result.
Thanks.
-
Thanks for the fixes. :)
I'm noticing strange behavior with manually assigning class to a dense point cloud. After I assign a class and hit save the points revert back to what class they were previously. This also happens if I run Build DEM after assigning a class, the DEM is built with the previously assigned class.
I initially ran Classify Ground Points and I'm trying to manually clean up the results.
-
Hello bisenberger,
Are you working with older project or the dense cloud has been generated recently in 1.3?
If you hit Update Dense Cloud statistics from the Workspace pane - does it fix the issue?
-
Hi Alexey,
I think it was originally a version 1.2.6 project, but I think the dense cloud was built in 1.3.
Update Dense Cloud statistics did not fix it.
-
Hello bisenberger,
Can the issue be reproduced on any project with the dense cloud? If so, then please specify the steps we should follow. Also whether the issue is related to PSX or PSZ project type (or both)?
-
Hi Alexy,
I built a project from scratch with 1.3 build 3620. Here are the steps:
Added cameras
Aligned images
Built mesh with sparse cloud
Placed markers
Renamed markers
Set project to UTM zone 14, WGS84
Imported control points
Ran optimize
Build dense cloud
Ran auto classify ground points
Selected points and reclassified
Saved project (PSX)
When I saved the project the points I manually reclassified reverted back to what they were previously.
Also worth noting, the mesh created from the sparse cloud appeared black, top and bottom.
-
Hello,
I have next error "RuntimeError: Grid size overflow" on call BuildOrthomosaic()
I use with PhotoScan 1.3 linux 64b
Here is my code:
...
doc = PhotoScan.app.document
doc.open(project_path)
chunk = PhotoScan.app.document.addChunk()
chunk.addPhotos(images_list)
camera = chunk.cameras[0]
chunk.loadReference(out_csv, format=PhotoScan.ReferenceFormatCSV, columns="nxyzabc", delimiter=",", group_delimiters=False, skip_rows=0)
chunk.crs = PhotoScan.CoordinateSystem("EPSG::4326")
chunk.matchPhotos(accuracy=PhotoScan.HighAccuracy, preselection=PhotoScan.GenericPreselection)
chunk.alignCameras()
chunk.buildDenseCloud(quality=PhotoScan.MediumQuality)
chunk.buildModel(surface=PhotoScan.Arbitrary, interpolation=PhotoScan.EnabledInterpolation)
chunk.buildUV(mapping=PhotoScan.GenericMapping)
chunk.buildTexture(blending=PhotoScan.MosaicBlending, size=4096)
chunk.buildOrthomosaic(surface=PhotoScan.ModelData, blending=PhotoScan.MosaicBlending, color_correction=False)
...
Full text of error:
BuildOrthomosaic: projection type = Geographic, projection = WGS 84, surface = Mesh, blending mode = Mosaic, resolution = 0
tessellating mesh... done (277124 -> 277124 faces)
Traceback (most recent call last):
File "agisoft_test.py", line 37, in <module>
chunk.buildOrthomosaic(surface=PhotoScan.ModelData, blending=PhotoScan.MosaicBlending, color_correction=False)
RuntimeError: Grid size overflow
I have read about this error in context of the wrong sequence of lat/lng here
http://www.agisoft.com/forum/index.php?topic=6027.0
I check reversed value of lat/lng and orthomap was build with no error but with wrong coordinates.
There is log from GUI console:
2017-01-19 01:54:31 Agisoft PhotoScan Professional Version: 1.3.0 build 3620 (64 bit)
2017-01-19 01:54:31 Platform: Linux
2017-01-19 01:54:32 OpenGL Vendor: NVIDIA Corporation
2017-01-19 01:54:32 OpenGL Renderer: GeForce GT 755M/PCIe/SSE2
2017-01-19 01:54:32 OpenGL Version: 4.5.0 NVIDIA 367.57
2017-01-19 01:54:32 Maximum Texture Size: 16384
2017-01-19 01:54:32 Quad Buffered Stereo: not enabled
2017-01-19 01:54:32 ARB_vertex_buffer_object: supported
2017-01-19 01:54:32 ARB_texture_non_power_of_two: supported
2017-01-19 01:54:37 LoadProject
2017-01-19 01:54:37 Loading project...
2017-01-19 01:54:37 loaded project in 0.032689 sec
2017-01-19 01:54:37 Finished processing in 0.03325 sec (exit code 1)
2017-01-19 01:59:53 Estimating resolution...
2017-01-19 01:59:53 Finished processing in 6.9e-05 sec (exit code 1)
2017-01-19 01:59:57 Estimating resolution...
2017-01-19 01:59:57 Finished processing in 5.3e-05 sec (exit code 1)
2017-01-19 02:00:10 BuildOrthomosaic: projection type = Geographic, projection = WGS 84, surface = Mesh, blending mode = Mosaic, resolution = 0
2017-01-19 02:00:10 Analyzing mesh...
2017-01-19 02:00:10 tessellating mesh... done (277124 -> 277124 faces)
2017-01-19 02:00:10 Finished processing in 0.211809 sec (exit code 0)
2017-01-19 02:00:10 Error: Grid size overflow
2017-01-19 02:00:31 Estimating resolution...
2017-01-19 02:00:31 Finished processing in 5.1e-05 sec (exit code 1)
-
Hello bisenberger,
We will try to reproduce the issue, although quick tests worked as expected.
-
Hello andrey.d,
Have you tried to build orthomosaic basing on the same data but using the GUI option? Also in the model view do you see that the coordinate axis orientation is correct?
-
Hi Alexey,
Here are some screen grabs showing some of the issues I'm having.
This one shoes some strange color artifacts on the dense point cloud.
Note that one of the graphic cards I'm using has changed. Previously I had two AMD cards. I replaced the slower one with a faster nVIDIA card.
(http://www.digital-mapping.net/forums/Photoscan/2017/PS13_dense_cloud_color.jpg)
This shows a mesh created from a sparse point cloud. It is completely black.
(http://www.digital-mapping.net/forums/Photoscan/2017/PS13_mesh_sparse_cloud.jpg)
This one shows an area where I manually reclassified some points.
(http://www.digital-mapping.net/forums/Photoscan/2017/PS13_re-classify.jpg)
and this shows the results of saving the project immediately after reclassifying the points.
(http://www.digital-mapping.net/forums/Photoscan/2017/PS13_re-classify_save.jpg)
nVIDIA driver version 376.33
AMD Radeon driver version 16.12.2
Besides swapping a graphic card I recently added another 32GB of ram to my system.
(http://www.digital-mapping.net/forums/Photoscan/2017/PS13_system.jpg)
-
Hello bisenberger,
And which graphic card is used as OpenGL rendering device? Is the behavior the same if you switch to another card?
Also can you try to remove some part of the dense cloud manually and save - would the removed part be restored after saving?
-
Hi Alexy,
By rendering device, do mean which card is the monitor plugged into?
I deleted a patch from the dense point cloud and saved the project. The points came back when I saved.
-
Hello bisenberger,
I think we were able to reproduce the issue with PSZ file format.
-
Hello bisenberger,
I think we were able to reproduce the issue with PSZ file format.
The format I'm using is PSX.
If I open the file up in version 1.2.6 the colors for the point cloud are really messed up.
-
Hello bisenberger,
At the moment we weren't able to reproduce the problem with PSX, but there's an issue like you've described with PSZ.
As for the colors, the internal representation of the dense cloud has been changed in 1.3 to allow multispectral and HDR colors for the point clouds, so if you are opening PSX project in the older version, it may be interpreted incorrectly. The compatible file format for the version 1.2 is only PSZ in Save As dialog.
-
Thanks for the info Alexey. Let me know if there is anything I can do to help track down the issue.
-
Hello Matt,
We have requested the sample dataset, but it seems that it's no longer needed, as we were able to reproduce the problem on multi-GPU setup.
Could you please also check the alignment on your system with only one of two GPU devices one, just to be sure that it works properly in your case?
Alexey Thanks, i'll check and let you know
-
Hello andrey.d,
Have you tried to build orthomosaic basing on the same data but using the GUI option? Also in the model view do you see that the coordinate axis orientation is correct?
Hello Alexey,
You advise me to try test my dataset in GUI mode.
A did it and have the similar error: Grid size overflow
And again if I change lat and long the map was built with no error but with wrong coordinates (placed in Saudi Arabia instead Bulgaria).
Please, find csv-files with coordinates attached. And psz-file with the project at this link https://drive.google.com/open?id=0B5NgaXpGOtOUQjlmT01TYllFU3M
2017-01-20 00:52:48 BuildTexture: blending mode = Generic, texture count = 1, blending mode = Mosaic, texture size = 4096
2017-01-20 00:52:48 Parameterizing texture atlas...
2017-01-20 00:57:30 Packing 2001 charts...
2017-01-20 00:57:35 Blending textures...
2017-01-20 00:57:35 estimating quality... ***************************** done in 16.6054 sec
2017-01-20 00:57:52 blending textures... ***************************** done in 37.9517 sec
2017-01-20 00:58:30 postprocessing atlas... done in 0.10269 sec
2017-01-20 00:58:38 Saving project...
2017-01-20 00:58:38 saved project in 0.273669 sec
2017-01-20 00:58:39 BuildTiledModel: source data = Dense cloud, pixel size = 0, tile size = 256
2017-01-20 00:58:47 Error: 2017-01-20 00:58:47 Saving project...
Grid size overflow
2017-01-20 00:58:47 saved project in 0.001245 sec
2017-01-20 00:58:47 BuildOrthomosaic: projection type = Geographic, surface = Mesh, blending mode = Mosaic, resolution = 0
2017-01-20 00:58:47 Analyzing mesh...
2017-01-20 00:58:47 tessellating mesh... done (275364 -> 275364 faces)
2017-01-20 00:58:47 Error: Grid size overflow2017-01-20 00:58:47 Saving project...
2017-01-20 00:58:47 saved project in 0.000891 sec
2017-01-20 00:58:47 Finished processing in 1211.58 sec (exit code 1)
-
There seems to be a bug when importing cameras from xml, setting the key and tie points always ignores the settings and defaults to 0,0. :(
-
Hello andrey.d,
For me it looks like the camera orientation angles loaded to the Reference pane are incorrect - so the model orientation in space in also incorrect. I've input very large number as "camera accuracy (deg)" to the Reference pane settings and then Updated the Reference pane - the orientation of the model after that looked valid.
-
Hello andrey.d,
For me it looks like the camera orientation angles loaded to the Reference pane is incorrect - so the model orientation in space in also incorrect. I've input very large number as "camera accuracy (deg)" to the Reference pane settings and then Updated the Reference pane - the orientation of the model after that looked valid.
I have image_name, longitude, latitude, altitude, yaw, pitch, roll
Could you give right format of csv file in my case?
-
Hello andrey.d,
I can suggest to skip loading yaw, pitch and roll, or to use the following lines to lower the accuracy of the angles after you load the CSV:
chunk.camera_rotation_accuracy = (1E5, 1E5, 1E5)
chunk.updateTransform()
-
Just grabed 1.3.0 do test it out and things are definitely broken. My photos are simply not aligning correctly at all (this is with OpenCL enabled. (in some cases the point cloud is sliced in half an translated to a different axis...wth) I also noticed that a big chuck of the photos are ignored in the process. Just retest in 1.23 and everything works fine. will do some testing without OpenCL now.
I'm attaching the Log and some screens of my second run where the "slicing wasn't as severe as the first time where literally the whole point cloud was perfectly slice in half vertically and shifted on the X axis)
-
And here's the log attached.
-
Just ran the alignment without OpenCL and things aren't better. A handful of cameras are not correctly aligned at all and a big chuck of the model is in the air..but this time only 2 of the 509 photos were dismissed (way less than with OpenCL). The second screenshot is how it looks in 1.2.6 (all cameras should be at ground level etc..)
-
Ok, so I uninstalled 1.3.0 and re-installed 1.2.6. Opened the same project, hit align without changing anything and everything was re-aligned perfectly. Something definitely fishy in 1.3.0.
-
Hello Mak11,
Can you provide the sample dataset were you are observing the problem to support@agisoft.com?
Also please specify if you are using exactly the same parameters for alignment.
-
i`ve got the same problem, with openCl poor align (1/3 of photos), turning off gfx card all photos aligned
-
Hello lunar,
Could you specify the 1.3 build number and GPU model used?
-
1.3.3620 - old HD7800 pitcairn 1024MB - this maybe a problem - very old card
-
Hello Mak11,
Can you provide the sample dataset were you are observing the problem to support@agisoft.com?
Also please specify if you are using exactly the same parameters for alignment.
Hello Alexey,
Yes as far I as could see I was using exactly the same parameters for alignement. I initially though that it was an OpenCL bug but the CPU only run was messed too (but a lot better than the OpenCL run) so alignment was incorrect in both modes while it's perfect every time I run it in 1.2.6.
The dataset is over 10GB of DNG photos :-( I don't think I could upload the whole thing in less than a week with my current internet connection.
-
Hello Mak11,
Maybe the same issue can be reproduced on the subset of images for JPGs, for example?
-
Alexey,
I'll try this out this week-end and let you know.
-
Thank you Mak11,
Maybe you can also send the report files generated from 1.2.6 and 1.3.0 projects, just for case we can find something unusual there and try to reproduce the problem on similar dataset.
-
How do I grab the report files? is it just the log (like the one I posted above?) or do I have to do anything else?
-
Sorry, thought you were using Professional edition.
-
I was first using the Professional 30Day Trial but then bought the Standard Edition :-/
-
Just did another alignement test in 1.3.0 with the CPU only with the same batch of photos but this time in JPG (4.5GB of data now instead of 10+GB in DNG) and the results are even worst. Only 183 of the 504 cameras were aligned...ran the same test in 1.2.6 and everything was perfectly fine.
-
Hello Mak11,
Can you post the screenshot of Chunk Info dialog for both versions? What happen if you try to reset the alignment for all cameras and then use Align Selected option? Also is the result also unstable if you run the alignment with the Adaptive fitting disabled?
-
Argh! I'll have to re-rerun it on 1.3.0 then ? Cause I keep switching between 1.2.6 and 1.3.0 by un-installing & re-installing each version...this is going to take some time..:-). I'm re-doing a clean run now using the JPGs on 1.2.6 and will make sure to grab as much info this time before trying again in 1.3.0.
-
Hello Mak11,
After installing 1.2 you can temporary move the folder from Program Files to another location and then install 1.3, thus having both versions side by side.
-
Last tests for the day.
504 Photos JPG with 1.2.6 on CPU = perfect alignment
504 Photos JPG with 1.3.0 (on CPU..it's even worst with OpenCL) = totally messed alignment (but strangely this time more cameras were aligned compared to he previous 1.3.0 test//)
Exactly the same settings used on both (will try on Monday with adaptive fitting off
Attachement of the 1.26 run below & 1.30 run on the next message.
-
Logs of the 1.3.0 run.
-
Hi Alexey
I was finally able to install and register this latest beta. Problem solved.
However I noticed that my two GPU's are not appearing into the OpenCL TAB. Version 1.2.6 are using them without any problem. My Machine is a MacPro Octacore with two AMD Radeon HD FirePro 500 GPU's.
Cheers
-
I'm posting the updated script for Split in Blocks, hope it works now. If not - let me know.
I'm not finding the new script OR its still not working. Does anyone have a copy of this handy?
Thanks.
-
I'm posting the updated script for Split in Blocks, hope it works now. If not - let me know.
I'm not finding the new script OR its still not working. Does anyone have a copy of this handy?
Thanks.
Re: Agisoft PhotoScan 1.3.0 pre-release
« Reply #170 on: December 27, 2016, 05:51:00 AM »
Here's the link to Alexey's script:
http://www.agisoft.com/forum/index.php?action=dlattach;topic=5976.0;attach=6749
-
Hello Alexey,
I just did another run in 1.3.0 with Adaptive Camera fitting OFF and the results are even worst. see attachments. Still can't figure out why it works perfectly in 1.2.6 but goes all wacky in 1.3.0 for no apparent reason at all.
-
Hello Mak11,
Then having sample dataset where the same behavior is reproducible would be helpful. I think JPGs would also work.
-
Hello Alexey,
I've sent you PM with a link to the JPG dataset.
-
Hello bisenberger,
At the moment we weren't able to reproduce the problem with PSX, but there's an issue like you've described with PSZ.
As for the colors, the internal representation of the dense cloud has been changed in 1.3 to allow multispectral and HDR colors for the point clouds, so if you are opening PSX project in the older version, it may be interpreted incorrectly. The compatible file format for the version 1.2 is only PSZ in Save As dialog.
When I try to open a version 1.3 PSZ file with version 1.2.6 I get this message. :(
(http://www.digital-mapping.net/forums/Photoscan/2017/PS13_PSZ.jpg)
-
Hello bisenberger,
Project files are not backward compatible, you need to use Save As -> PhotoScan 1.2.0 format when saving the PSZ project from the newest version.
-
Is there a way to turn off the auto-placement of markers (blue flag)? In my test, it is doing more harm than good.
-
Hello maddin,
You can use Place marker option to put a projection of existing marker on photo.
In the version 1.3 there would be a possibility to choose New Marker option from the Place Marker menu, thus you do not need to create new marker instance first.
-
You can use Place marker option to put a projection of existing marker on photo.
In the version 1.3 there would be a possibility to choose New Marker option from the Place Marker menu, thus you do not need to create new marker instance first.
Not sure I explained well what my issue is:
In 1.3, I place a marker on one photo. Automatically, PS creates (blue flagged) markers in all the other photos. Unfortunately, these blue markers are often in wrong locations. In version 1.26, placing one marker in a photo would create epipolar lines in the other photos; then, upon placement of the next marker, all other photos would get a grey marker suggestion, which is fine. Now in 1.3, that grey suggestion stage seems to be replaced. Since the blue markers are in wrong locations, I get very bad refinement. Is there a way to prevent those blue markers?
-
Hello maddin,
If you do not wish marker projections to be created automatically (blue flags) using available surface (point cloud, mesh, etc.) you need to you Place marker option in Photo view instead of Create Marker.
-
Dear Aleksey,
Could I manage mem, virtual mem and cpu cores by script.
I see examples for cpu cores setting, but it don't work in last build of 1.3.
Thanks.
-
Hello andrey.d,
There's no setting to manage the memory or CPU usage in PhotoScan 1.3.
The CPU cores option that was available in the version 1.2 was related to the depth maps calculation procedure only when at least one GPU device was turned on. The setting defined how many CPU cores would be used for the depth maps calculation, other cores could be used for the GPU tasks management.
-
Hi Alexey, about my previous message:
"However I noticed that my two GPU's are not appearing into the OpenCL TAB. Version 1.2.6 are using them without any problem. My Machine is a MacPro Octacore with two AMD Radeon HD FirePro 500 GPU's."
Aren't you supporting these GPU's anymore or still a work in progress?
Cheers
-
Hello RHenriques,
I think we'll provide special installer for Mac users that have AMD cards. Currently there are some issues with OpenCL drivers on Mac OS X.
-
Great Alexey. Some others, like me, invested a lot in these machines to use with PhotoScan. I'm very satisfied with it and it's not that old. That would be awesome. Without the drivers this version feels a lot slower than version 1.2.6, mainly when using mesh or texture view.
Cheers
-
Hello RHenriques,
Sure, we understand the issue and we'll be finding a proper solution for it. For some reason AMD drivers for Mac OS X do not include SPIR support.
-
Hello Alexey,
Exactly how it works multichannel support for point cloud colors, format export?
Thanks.
-
I see a new build was released yesterday. :)
Updated to version 1.3.0 build 3745.
« Last Edit: January 31, 2017, 01:30:29 PM by Dmitry Semyonov »
Will start checking it out tomorrow.
-
thats nice :) i will also start to test the new version.
@Alexey Pasumansky
did you have a new target release date for a stable 1.3.0 version ?
-
@Alexey Pasumansky
did you have a new target release date for a stable 1.3.0 version ?
I think the end of this week or beginning of the next one.
-
Hello Aleksey,
I have some problems with using dataset without reference with python script.
I need result in EPSG::3857 (it placed in virtual place near (0, 0) point and one pixel equal to one point in EPSG::3857 crs- but it okey for me.
In GUI mode all work fine. BUT when I do it in a script I receive huge map (about 1M x 1M pixels)
The question is how I can get from agisoft optimal size or resolution for export.
This is part of my code:
crs = PhotoScan.CoordinateSystem("EPSG::3857")
doc = PhotoScan.app.document
doc.open(project_path)
chunk = doc.chunk
chunk.addPhotos(images_list)
chunk.matchPhotos(accuracy=PhotoScan.MediumAccuracy, preselection=PhotoScan.NoPreselection)
chunk.alignCameras()
chunk.buildModel(surface=PhotoScan.HeightField, interpolation=PhotoScan.EnabledInterpolation)
chunk.crs = crs
chunk.buildOrthomosaic(surface=PhotoScan.ModelData, blending=PhotoScan.MosaicBlending, color_correction=False)
chunk.exportOrthomosaic(args.ortho_path, image_format=PhotoScan.ImageFormatTIFF, raster_transform=PhotoScan.RasterTransformNone, write_kml=False, write_world=False, write_alpha=True, tiff_compression=PhotoScan.TiffCompressionJPEG, jpeg_quality=90, tiff_big=True)
-
Hello andrey.d,
Actually, it doesn't seem to be related to the version 1.3 - you have set the projected coordinate system, so by default PhotoScan will be building the orthomosaic in geographic coordinates, but since you have no reference information that model is arbitrary oriented causing incorrect scene estimation.
-
Hello andrey.d,
Actually, it doesn't seem to be related to the version 1.3 - you have set the projected coordinate system, so by default PhotoScan will be building the orthomosaic in geographic coordinates, but since you have no reference information that model is arbitrary oriented causing incorrect scene estimation.
Thank you, Aleksey.
The answer was in using Local CRS as you write here:
http://www.agisoft.com/forum/index.php?topic=2718.0
-
Latest version of 1.3 seems to be working well for me. Just one thing I reported before that is really annoying.
When using the "Gradual Selection" tool in this version, you need to click on the model window once after gradual selection and before pressing delete. If you left a selection tool on, this deselects the "gradual selection".
Previous versions returned focus to the model window after the gradual selection dialog is closed. I run gradual selection sometimes 15 times on each model - if you could make this work again I would be very happy.
Thank you.
-
Hello mrc,
We'll check that, but as an alternative option you can just click on Delete button on the Toolbar.
Also if you are in Navigation mode the selection wouldn't be discarded once you click in the Model view.
-
Hello,
i get the following errors.
I tested different graphic drivers, but nothing works. Only calculation without gpu works.
greetings
andreas
-
Hello Andreas,
Are there any other processes that loads your GPU working in parallel?
We have tested GTX 1080 and haven't encountered such issue before.
-
Apart from Windows nothing runs parallel. The funny is, at the first start I had no problems, everything ran. I have also tested the latest driver, but the GPU did not show up under the settings. I then installed an older driver and the GPU was back, but can not be used somehow. What driver version did you use?
-
I have tried again a driver version and it runs great. The version 368.81 for Nvidia GeForce GTX 1080 seems to run stable.
greetings
Andreas
-
Hi,
since Version 3772 i can't draw Polylines on a Orthofoto to Change Pictures.
The Konsole Shows: Error: Can't transform coordinates.
I work with local Coords, the Orthofoto is build with planar -> local view.
The Orthofoto appears, but i can't Change the Pictures anymore....
-
Hello Dieter,
Can you rebuild the orthomosaic for the same or similar project using Planar projection plane and check if the issue is still there? (before that remove current orthomosaic and shapes folder from the chunk).
-
Hello,
What is the new function ?
== > Added WIRIS thermal image format support.
Is there a thermal image alignment option is the new version ?
Thanks
-
Hello madekwe,
Thermal data from WIRIS sensor comes as single channel floating point TIFF. Upon loading such data PhotoScan prompts the user, if he wish to apply the temperature scale (via Raster Calculator functionality) for proper visualization. For other processing steps, like alignment and dense cloud generation, images will be treated just as single channel dataset.
-
Hi -
on Mac, I get this error for writing the license file.
Can't write license file. Please make sure that you have write access to /private/var/folders/f0/cm47pk9s31sclhww2txk9_fc0000gq/T/AppTranslocation/C61F044A-A2C2-432E-9E5D-4F611F066A4B/d/PhotoScanPro.app/Contents/MacOS folder.
I restarted and rebuilt permissions, but no joy.
thanks,
Michael Ashley
-
Hello Michael,
It seems that your Mac OS X is moving the application to the temporary location.
Please make sure that you have copied PhotoScanPro application bundle from unpacked DMG image to some other location and are not running it directly from the archive.
Also I can suggest to run the activation from the terminal using --activate argument.
-
Sometime projects created with PS 1.2 open in write protection mode.
Why is this and how can I switch off the write protection mode?
-
In the latest build the button to switch the display of contour lines is missing an the (in 1.2) created contours are visible even in the point cloud (that's cool).
-
Hello Airgo,
Maybe they were not closed properly or you do not have write permissions to the project folder.
But please check that you have version 1.3.0 build 3772 installed. Also you can manually remove .lock file from the project.files folder, if PhotoScan doesn't suggest an option to open the read-only project for editing anyway.
-
In the latest build the button to switch the display of contour lines is missing an the (in 1.2) created contours are visible even in the point cloud (that's cool).
In the version 1.3 created contours are now moved to shape layer "contours" and can be treated as regular shapes. The elevation value is kept in the corresponding shape attribute.
To hide the contours, just disable their layer in the Workspace pane.
-
Hello Alexey,
great!
But to change the color of the contours I have to rebuild them in 1.3?
-
Hy Alexey,
despite the older, the new release of 1.3.0 (downloaded from download section) doesn't recognise INTEL IRIS Pro of my mac book pro. The older builds recognise it.
Why???
-
Hi - the application is in the Applications folder on the Mac, where it belongs. I copied it from the DMG, ejected the DMG, and restarted the machine.
How do I do the console activation command? In Photoscan?
thanks,
Michael
-
But to change the color of the contours I have to rebuild them in 1.3?
Hello Airgo,
I think they would be opened as a contours layer even if generated in the older version. So you can change their color in the layer properties.
-
despite the older, the new release of 1.3.0 (downloaded from download section) doesn't recognise INTEL IRIS Pro of my mac book pro. The older builds recognise it.
Why???
By older build do you mean one of 1.3 pre-releases or version 1.2?
-
Hi - the application is in the Applications folder on the Mac, where it belongs. I copied it from the DMG, ejected the DMG, and restarted the machine.
How do I do the console activation command? In Photoscan?
Hello Michael,
You need to do it from the Terminal (you can find it from Launchpad), then go to the directory where PhotoScanPro application bundle is located, then use the following command:
PhotoScanPro.app/Contents/MacOS/PhotoScanPro --activate 11111-22222-33333-44444-55555Or you can just input the full path to the application bundle in the beginning of the command.
-
despite the older, the new release of 1.3.0 (downloaded from download section) doesn't recognise INTEL IRIS Pro of my mac book pro. The older builds recognise it.
Why???
By older build do you mean one of 1.3 pre-releases or version 1.2?
Hy Alexiei, I mean one of pre-release build of 1.3
(I add also a bug report on volume calculation when using custom level)
Thanks
-
Hi - the application is in the Applications folder on the Mac, where it belongs. I copied it from the DMG, ejected the DMG, and restarted the machine.
How do I do the console activation command? In Photoscan?
Hello Michael,
You need to do it from the Terminal (you can find it from Launchpad), then go to the directory where PhotoScanPro application bundle is located, then use the following command:
PhotoScanPro.app/Contents/MacOS/PhotoScanPro --activate 11111-22222-33333-44444-55555Or you can just input the full path to the application bundle in the beginning of the command.
WORKED! Here's the full path /Applications/Photoscan/PhotoScanPro.app/Contents/MacOS/PhotoScanPro thanks!