Agisoft Metashape
Agisoft Metashape => General => Topic started by: Dmitry Semyonov on September 06, 2019, 07:46:42 PM
-
Pre-release version of Agisoft Metashape 1.6.0 is available for download.
This pre-release is considered as unstable and is not recommended for production use.
Please use the links below for download:
Agisoft Metashape Standard edition
Windows 64 bit (http://download.agisoft.com/metashape_1_6_0_x64.msi)
Windows 32 bit (http://download.agisoft.com/metashape_1_6_0_x86.msi)
Mac OS X (http://download.agisoft.com/metashape_1_6_0.dmg)
Linux 64 bit (http://download.agisoft.com/metashape_1_6_0_amd64.tar.gz)
Agisoft Metashape Professional edition
Windows 64 bit (http://download.agisoft.com/metashape-pro_1_6_0_x64.msi)
Windows 32 bit (http://download.agisoft.com/metashape-pro_1_6_0_x86.msi)
Mac OS X (http://download.agisoft.com/metashape-pro_1_6_0.dmg)
Linux 64 bit (http://download.agisoft.com/metashape-pro_1_6_0_amd64.tar.gz)
Agisoft Metashape Python 3 Module (requires Professional edition license)
Windows 64 bit (http://download.agisoft.com/Metashape-1.6.0-cp35.cp36.cp37-none-win_amd64.whl)
Windows 32 bit (http://download.agisoft.com/Metashape-1.6.0-cp35.cp36.cp37-none-win32.whl)
Mac OS X (http://download.agisoft.com/Metashape-1.6.0-cp35.cp36.cp37-abi3-macosx_10_8_x86_64.whl)
Linux 64 bit (http://download.agisoft.com/Metashape-1.6.0-cp35.cp36.cp37-abi3-linux_x86_64.whl)
Agisoft Metashape Java Library (requires Professional edition license)
Download (http://download.agisoft.com/metashape-java-api-1.6.0.zip)
Metashape 1.6.0 Change Log (http://www.agisoft.com/pdf/metashape_changelog.pdf)
Metashape 1.6.0 Python API Reference (http://www.agisoft.com/pdf/metashape_python_api_1_6_0.pdf)
Please let us know if you have any problems using this version.
Please post problems and/or questions concerning other Metashape version in separate topics.
This topic will be removed when the final 1.6.0 version is released.
Note
Project files created with 1.6.0 version can't be opened with earlier versions.
EDIT:
Metashape 1.6.0 build 9925 was finally released. Thank you for useful feedback.
Please use the links on the main Metashape page for download:
Download Installers (http://www.agisoft.com/downloads/installer/)
-
Dmitry,
Can we have some info about those new features?
• Added normal map and ambient occlusion map generation support.
• Added Sequential and Estimated reference preselection modes.
• Added guided image matching option.
Mak
-
I seem to be unable to export AO & Normal maps generated in Metashape. The export texture option is greyed out? Also generating a Normal map after an AO map made the AO map disappear.
Edit: This may be related to the fact that I was using a Metashape 1.5.4 project. Looks like it works if I create a new empty project and then import the models for baking.
EDIT 2: It's actually totally random. Sometimes I can export textures maps and some times I can't. There's definitely a bug.
EDIT 3: More testing shows that the (imported or decimated) mesh must already have a diffuse map assigned to it for the Export Texture function to work.
Mak
-
Hello Dmitry,
I am looking at processing satellite imagery with RPC metadata in 1.6 pre release...
I have downloaded a pair of Ikonos 1 m pan tif 11bit images over Hobart in Australia....https://www.isprs.org/data/ikonos_hobart/default.aspx
I was able to load the 2 images in Metashape and it did read the RPC metadata for each image. However, the altitude of each satellite image is read in as 238 m which is obviously wrong.
So the produced dense cloud looks good but gives a ground elevation of something like - 6 300 m! see attached screen copy...
Attached are corresponding rpc, hdr and po order metadata files for both images in rar...
-
How to use this "guided image matching" feature ?
-
I seem to be unable to export AO & Normal maps generated in Metashape. The export texture option is greyed out? Also generating a Normal map after an AO map made the AO map disappear.
Edit: This may be related to the fact that I was using a Metashape 1.5.4 project. Looks like it works if I create a new empty project and then import the models for baking.
EDIT 2: It's actually totally random. Sometimes I can export textures maps and some times I can't. There's definitely a bug.
EDIT 3: More testing shows that the (imported or decimated) mesh must already have a diffuse map assigned to it for the Export Texture function to work.
Hello Mak,
Yes, it seems like the Diffuse map is required to enable Export Texture option in the File Menu. I think we could fix it in the next update.
-
How to use this "guided image matching" feature ?
Hello bestable,
Guided matching feature is a new experimental functionality that is designed mostly to produce the excessive matching points set for ultra high-resolution imagery.
-
I am looking at processing satellite imagery with RPC metadata in 1.6 pre release...
I have downloaded a pair of Ikonos 1 m pan tif 11bit images over Hobart in Australia....https://www.isprs.org/data/ikonos_hobart/default.aspx
I was able to load the 2 images in Metashape and it did read the RPC metadata for each image. However, the altitude of each satellite image is read in as 238 m which is obviously wrong.
So the produced dense cloud looks good but gives a ground elevation of something like - 6 300 m! see attached screen copy...
Hello Paul,
Looks like two images are not sufficient for the proper optimization procedure, so if you are planning to use only two RPC images for testing, we do not suggest to optimize cx, cy coefficients. Or to include the ground control point in the optimization procedure.
-
Hi Alexey,
I was able to get a good result downloading and using some 4 GCPS from same web site and now, I get errors on GCPS less than 0.5 m and DEM is now correct ( from 0 m to some 1270 m)....
Still , it would be very helpful to do some blog or tutorial for RPC satellite processing as it is not very clear when to optimize or use certain parameters as f, cx, cy, b1 and b2 whose meaning is not evident in case of satellite push broom sensor....
Also, why is the altitude of 2 images 238 m ?... does not make much sense....
Edit I did a further optimization with f, cx, cy, b1 and b2 and got better results: better than 25 cm error on GCPs and better than 1 m precision on TiePts....
-
Hello Paul,
The altitude information displayed for RPC images in the Reference pane is not related to the satellite height. It corresponds to the origin of the system defined by RPC model.
For optimization of RPC images we suggest to use only cx, cy coefficients.
-
Hi Alexey,
what kind of satellite sensors may be processing in Metashape?
Best Slwomir Krolewicz
-
Awesome can't wait to test it.
-
what kind of satellite sensors may be processing in Metashape?
Hello Slwomir,
Currently Metashape 1.6.0 pre-release requires RPC data for each image. RPC coefficients are loaded automatically if they are present in RPCCoefficient image tag. Alternatively, they can be loaded from supplementary rpc.txt files (if available). To do this, Load satellite RPC data from auxiliary TXT files option should be enabled in Preferences dialog before adding images to the project.
-
Wow AO and Normal map works very well thanks for the add!
Sketchfab model:
https://sketchfab.com/3d-models/pharaoh-statue-b60326b0bcfb494c889bb6f87011d8a2
-
Dear Agisoft,
Can you explain what's • Added Sequential and Estimated reference preselection modes?
Thanks
-
Can I put satellite image and aerial image together in one chunk to align them simultaneously？
-
Can I put satellite image and aerial image together in one chunk to align them simultaneously？
Hello regiser,
Such case might be of interest for some users, however, we have not yet tested it and cannot confirm yet. In case you have any sample data that we can use for such matching, please let me know.
-
I imagine the sequential selection means it will choose pairs which follow each other in the sequence of photos whereas estimated means it will look for large features across the range of photos for pairing. So I imagine sequential will work faster if you have taken the photos in an orderly way (such as from a drone or, as with me, photographing sculpture in a methodical way working my way around it).
But I could be totally wrong! :-\ ;D
Dear Agisoft,
Can you explain what's • Added Sequential and Estimated reference preselection modes?
Thanks
-
I imagine the sequential selection means it will choose pairs which follow each other in the sequence of photos whereas estimated means it will look for large features across the range of photos for pairing. So I imagine sequential will work faster if you have taken the photos in an orderly way (such as from a drone or, as with me, photographing sculpture in a methodical way working my way around it).
But I could be totally wrong! :-\ ;D
Dear Agisoft,
Can you explain what's • Added Sequential and Estimated reference preselection modes?
Thanks
Thanks cbnewham it might make sense, hope Alexey give us an additional information.
-
I see the latest version still uses grey ticks for the "aligned" photos.
Would it be possible to make these red or some other bright colour - it's really not obvious what's been aligned and what hasn't and makes going through the images to select and align any unaligned ones after a preselection much more time consuming.
Thanks!
-
Or, red ticks against those that haven't aligned, which would be even better.
I see the latest version still uses grey ticks for the "aligned" photos.
Would it be possible to make these red or some other bright colour - it's really not obvious what's been aligned and what hasn't and makes going through the images to select and align any unaligned ones after a preselection much more time consuming.
Thanks!
-
Hi, could we get an how-to on the use of diffuse, normal and ao texture generator ? I have not found how to use them after generation and export them.
-
I am also testing RPC image processing using Digital Globe 50 cm pan imagery. This imagery comes with RPC meta information stored in a RPB file instead of normal text file.
Example RPB file attached... It would be great if Metashape could read this format RPB directly avoiding having to manually edit the file to normal RPC txt....
-
Is there a possibility that you could add more baking map types like Bent Normal maps & Height map ?
Cheers
Mak
-
Thank you for update! Now RPB files are read correctly in satellite image processing....
I have tested on a pair of Digital Globe images over Tripoli both in pan and MS.
The alignment is fine but:
- for pan imagery, the dense cloud presents many gaps in coverage see attached capture
- for MS ( 8 band ), the dense does not present great gaps only a regular empty line regularly spaced along model, see 2nd capture
Any ideas on these issues?
Attached is log file of this processing sequence:
1. load pair of pan images in chunk Pan;
2. align pair of pan images;
3. load pair of MS images in chunk MS;
4. align pair of MS images;
5. create dense cloud of MS chunk;
6. create dense cloud of Pan chunk.
PS Last Edit I have checked with last release and though the dense cloud generated from Pan chunk presents less gaps, it still has some significant holes in dense coverage...and regularly spaced empty line of data is still present. See last capture...
PS latest update still does not correct partial gaps in dense cloud coverge arising from holes in depth maps
-
Dear support,
working on this Digital Globe Image processing (previous post), my original project was saved as psx file. So i then saved the psx as psz file and found that in this file the average rms reprojection error for pan chunk Tie Points shot up to some 7 000 pixels from avg rms reprojection error of less than 0.5 pixel in original psx...
In attached screen captures, first corresponds to psx file and second to psz....
Something weird is happening here...
Edit.... I have found that in the psz file, opening the Camera Calibration pane, F is shown as 0 instead of 1 in psx....see 3rd attachment
it seems that PSZ format does not keep the RPC calibration data for each camera as in PSX see 4th screen capture with doc.xml from both PSX and PSZ side by side
Latest edit... latest release does save correctly the psz format with all RPC parameters... Thanks for fix!
-
The baking options are very promising (fast)!
You do need to support custom vertex normals for this to be truely useful though (or at least give an option for auto weighting of choice).
Displacemen/Height baking is a must as well as somebody here previously mentioned.
Great update!
-
I am also testing RPC image processing using Digital Globe 50 cm pan imagery. This imagery comes with RPC meta information stored in a RPB file instead of normal text file.
Example RPB file attached... It would be great if Metashape could read this format RPB directly avoiding having to manually edit the file to normal RPC txt....
Hello Paulo,
I believe in the latest pre-release update the RPB auxiliarry files should be read properly.
As for the gaps, it might be a problem on our side related to the depth maps generation procedure. Are the files that you are using in your example available for public download?
-
Hi Alexey,
thanks for your reply. Yes the RPB input is working fine....
My test was done on Tripoli System-Ready (1B) Stereo, 8-band bundle, 50 cm product, 2015-08-31, downloaded from http://www.digitalglobe.com/samples....
Have you seen my post on problem with psz file saving RPC calibration data?
Just did another test on Tripoli_View-Ready_Stereo_8_Band_Bundle_30cm product from same date. It is not ideal, it has 2A level processing (crude orthorectify on map projection) while System Ready imagery has 1B level and is more appropriate IMHO....
Anyway for this test, I did not find any regularly spaced empty lines on Dense cloud... see attachement...maybe size of images being smaller has impact..
-
Hello really great realease ....
I try the java binding, on windows 64, some small errors in method visibility but it look like some dependent dll are missing in the windows implementation ... is this the case ?
metashape.dll seems to need other dll ...
Any idea where can I found them ?
Thanks a lot,
Kind regards
Pierre Drap
-
Some really good additions in this release, thanks Agisoft :)
The Metashape Pro scripts aren't working with version 6.
https://github.com/agisoft-llc/metashape-scripts
Will out-of-core for meshing with depth maps be added in this release?
-
Some really good additions in this release, thanks Agisoft :)
The Metashape Pro scripts aren't working with version 6.
https://github.com/agisoft-llc/metashape-scripts
Will out-of-core for meshing with depth maps be added in this release?
The api naming has been updated, all you need to do is rename some things to make it work again.
-
Some really good additions in this release, thanks Agisoft :)
The Metashape Pro scripts aren't working with version 6.
https://github.com/agisoft-llc/metashape-scripts
Will out-of-core for meshing with depth maps be added in this release?
The api naming has been updated, all you need to do is rename some things to make it work again.
Thanks wojtek, it worked. :)
I changed compatible_major_version = "1.5" to compatible_major_version = "1.6"
-
What does the Publish results option do when checked in the Cloud processing dialog?
-
What does the Publish results option do when checked in the Cloud processing dialog?
Hello Bill,
Agisoft Cloud account page has been updated, now it is possible to display 3D content (point clouds, mesh and tiled models) of the georeferenced projects in the cloud from the personal account page.
Currently the visualization is only accessible for the account page owner, but we are planning to implement the option to share the data via link and also to allow 2D results (DEM and orthomosaics) visualization.
-
Hi, I see in the changelog for Version 1.6.0th that it is possible to export normal maps with .OBJ format, yet I see no such option in any menu items or during export ... am I missing something?
-
Hi, I see in the changelog for Version 1.6.0th that it is possible to export normal maps with .OBJ format, yet I see no such option in any menu items or during export ... am I missing something?
It's done automatically provided that you've created it ;)
-
Hi,
I've seen in API changelog " Added Model.getActiveTexture() and Model.setActiveTexture()" methods.
I would like to switch programmatically between the different textures in order to make a capture of the textured model view for each of them.
Here's my code:
tex = Metashape.app.document.chunk.model.textures
num = 0
for i in tex:
Metashape.app.document.chunk.model.setActiveTexture(i,Metashape.Model.TextureType(i.type))
Capture = Metashape.app.captureModelView(width = 1114, height = 763, transparent = (True), hide_items = (True), source = Metashape.DataSource.ModelData, mode = Metashape.ModelViewMode.TexturedModelView)
Cap = Metashape.Image(Capture.width, Capture.height, "RGBA", "U8")
Capture.save('C:/Users/.../Desktop/num{}.png'.format(num))
num +=1
But unfortunately it dosen't change the texture type of the model...
Any ideas please?
-
FYI using the latest build in the python API I get very low numbers of tie points after "match photo + align cameras", no matter which settings I use, and I often get complete crashes during the "align cameras step" :)
Moving back to 1.5.5 now, but will test the next pre release of 1.6 to see.
-
What does the Publish results option do when checked in the Cloud processing dialog?
Hello Bill,
Agisoft Cloud account page has been updated, now it is possible to display 3D content (point clouds, mesh and tiled models) of the georeferenced projects in the cloud from the personal account page.
Currently the visualization is only accessible for the account page owner, but we are planning to implement the option to share the data via link and also to allow 2D results (DEM and orthomosaics) visualization.
Hi Alexey,
Sounds great! that will be very useful. 8)
-
Hi Alexey,
What is ''Guided image matching'' in Align photo stage mean please?
great update BTW.
Mohamed
-
I try the java binding, on windows 64, some small errors in method visibility but it look like some dependent dll are missing in the windows implementation ... is this the case ?
metashape.dll seems to need other dll ...
Any idea where can I found them ?
Hello Pierre,
Please check the pre-release update. Now Java bindings should include all the required DLLs.
-
Hi Alexey,
What is ''Guided image matching'' in Align photo stage mean please?
great update BTW.
Mohamed
I just tried Guided image match and got long time process and miss alignment many photos? So what dose it do?
Mohamed
-
Hello Mohamed,
Guided Matching option is recommended for the high resolution images, like 50-100+ MPix images in order to provide excessive tie points that could be missed by the feature point detector.
For common consumer grade cameras the feature would not give considerable effect and would take longer time, compared to the common workflow.
-
FYI using the latest build in the python API I get very low numbers of tie points after "match photo + align cameras", no matter which settings I use, and I often get complete crashes during the "align cameras step" :)
Moving back to 1.5.5 now, but will test the next pre release of 1.6 to see.
Hello Stephan,
Can you provide the code lines that you are using and the processing logs related to the crashed operations?
If you are running the matching and alignment multiple times for the same dataset, are you removing the previously acquired tie points and resetting the previous alignment?
-
Hello,
I was wondering about the new options in the align photos stage : source - estimated - sequential.
I think sequential speaks for itself but what do the others ?
Thank you
-
Team,
What's the ETA of the next 1.6 pre-release build?
Cheers
Mak
-
Hello Mak,
Next pre-release update is expected by the end of this week, or maybe during the next week.
-
I was wondering about the new options in the align photos stage : source - estimated - sequential.
I think sequential speaks for itself but what do the others ?
Sequential pre-selection is trying to match the images with a bunch of their neighbors according to the camera labels.
Estimated preselection is based on the estimated camera locations after initial alignment (for example, performed using rough accuracy settings).
One of the possible scenarios of use - when aligning the video sequence frames run Sequential preselection at first and then Estimated preselection, if there are intersecting parts of the image acquisition path.
-
Hello Mak,
Next pre-release update is expected by the end of this week, or maybe during the next week.
Great. Patiently waiting.
Cheers
Mak
-
I'm now encountering full system crashes/lock-ups with 1.6 Build 9217 during the "Estimating camera locations..." (which is 100% CPU bound) part of the alignement process on some data-sets (those who seems the more intensive to process). Never had Photoscan/Metashape hard lock a PC like that before. First time it happened I got a BSOD. All other attempts resulted in a hard lock (only solution is to hard reset the system).
Only thing that pop-ed up in the Event Viewer is this (when the BSOD happened):
Log Name: Application
Source: Application Error
Date: 21-Oct-19 12:11:26 AM
Event ID: 1000
Task Category: (100)
Level: Error
Keywords: Classic
User: N/A
Computer: DESKTOP-4612L03
Description:
Faulting application name: metashape.exe, version: 1.0.0.1, time stamp: 0x5d90ea7f
Faulting module name: ntdll.dll, version: 10.0.18362.418, time stamp: 0x99ca0526
Exception code: 0xc0000005
Fault offset: 0x000000000009fab0
Faulting process id: 0x26f0
Faulting application start time: 0x01d58791bf92b983
Faulting application path: C:\Program Files\Agisoft\Metashape\metashape.exe
Faulting module path: C:\WINDOWS\SYSTEM32\ntdll.dll
Report Id: 6f7e4ac4-b51e-48cc-8a9a-0dc3d14c95c9
Faulting package full name:
Faulting package-relative application ID:
Event Xml:
<Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">
<System>
<Provider Name="Application Error" />
<EventID Qualifiers="0">1000</EventID>
<Level>2</Level>
<Task>100</Task>
<Keywords>0x80000000000000</Keywords>
<TimeCreated SystemTime="2019-10-20T22:11:26.605443100Z" />
<EventRecordID>38721</EventRecordID>
<Channel>Application</Channel>
<Computer>DESKTOP-4612L03</Computer>
<Security />
</System>
<EventData>
<Data>metashape.exe</Data>
<Data>1.0.0.1</Data>
<Data>5d90ea7f</Data>
<Data>ntdll.dll</Data>
<Data>10.0.18362.418</Data>
<Data>99ca0526</Data>
<Data>c0000005</Data>
<Data>000000000009fab0</Data>
<Data>26f0</Data>
<Data>01d58791bf92b983</Data>
<Data>C:\Program Files\Agisoft\Metashape\metashape.exe</Data>
<Data>C:\WINDOWS\SYSTEM32\ntdll.dll</Data>
<Data>6f7e4ac4-b51e-48cc-8a9a-0dc3d14c95c9</Data>
<Data>
</Data>
<Data>
</Data>
</EventData>
</Event>
Uninstalled 1.6 and re-installed 1.5.5 & everything seems to work fine.
EDIT: Happening on 1.5.5...🤔 Must be a hardware issue on my side...
Lowered my CPU clock & no more crashes since... Looks like it's on its last leg..
Cheers
MAK
-
Dmitry, Alexey,
Added GPU-based texturing generation support for NVIDIA GPUs.
Any chance/hope that OpenCL support (AMD GPUs & Intel iGPUs soon to be released Intel dGPUs) will be added?
Cheers
Mak
-
Thanks for the revised UV parameterization!
It seems that dense meshes no longer have really high overlap ratio although i do get much lower fill ratio in return. Are there any new settings we can use via python for buildUV?
-
I'm guessing that this will come back.. but the Calculate Vertex Colors option is no longer available when generating meshes via the Depth Maps.
Mak
-
Dmitry, Alexey,
Added GPU-based texturing generation support for NVIDIA GPUs.
Any chance/hope that OpenCL support (AMD GPUs & Intel iGPUs soon to be released Intel dGPUs) will be added?
Hello Mak,
GPU texturing relies on Vulkan API which is supported by all GPU vendors. But we encountered some problems in AMD driver implementation - when they will be resolved we will be able to confirm GPU texturing on AMD. Intel GPUs will be tested after that.
As for the vertex colors and volumetric masks - they are not gone forever and will be back for out-of-core mesh implementation in the next pre-release update.
-
Alexey,
Glad to know that it's Vulkan based & not CUDA. What kind of issue did you encounter on AMD GPUs? Also, does the current CPU fallback provide the same end result (but slower) that the GPU path?
One more question: the out of core mesh generation seems to rely a lot less on GPU compute after the depth maps have been generated, right?
Haven't test a whole lot but I'm happy to see that the final mesh output is now better with less of the unwanted polys everywhere.
Edit:
Trimming seems to be a bit strange:
1.6 9217:
(https://i.postimg.cc/63XRKFpK/Annotation-2019-10-21-190413.jpg)
1.6 9397:
(https://i.postimg.cc/fLNHP24t/Annotation-2019-10-21-190446.jpg)
(also "main/mesh_trimming_radius: 30" doesn't seem to have any effect at all now)
BTW there's a typo in the console log it says Outlier instead of Outliner during texture generation.
Cheers
Mak
-
I am curious to know what Fit additional corrections option do in Optimize Cameras dialog?....
What are these corrections and when should we use this option?
-
Hello Mak,
Looks like the trimmed areas have poor coverage by the depth maps. Do you have any trimming tweaks enabled?
The GPU utilization for the depth maps method should be during the similar steps as in the version 1.5 - similar parts of the method are GPU-accelerated.
As for the possible difference between CPU and GPU implementation results for the texture generation - they would not be identical, however, should look very similar and it shouldn't be possible to distinguish the difference in general case.
-
Alexey,
No trimming tweaks were applied. If you try the Doll sample you will see similar results (holes) where there's poor depth map coverage (under the skirt) and mesh_trimming_radius 30 doesn't seem to have any effect when applied in this build. (obviously 1.5.0 handled does cases perfectly without any tweaks...but you already know that 😉). Visibility Consistent Mesh Generation, Metashape up to v 1.5.1 ( & to a lesser degree Metashape 1.5.3+ with mesh_trimming_radius set to "30") correctly remeshed those holes/gaps. So there must be a way to fix this.
Yeah I'm seeing some GPU utilization but what appears in the console has changed quite a bit so it was harder to figure out what was happening.
Out-of-core meshing is perfectly working though. So big congrats on this achievement! Never went over 14GB of Ram usage on data-sets (with 150+/200+ 24MP images) that would fail before.
Cheers
MAK
Cheers
-
I am curious to know what Fit additional corrections option do in Optimize Cameras dialog?....
What are these corrections and when should we use this option?
Hello Paulo,
Fit additional corrections option enables additional calibration coefficients that allows a better fit of the lens distortion model. This is especially important for wide angle lens with large radial distortions, e.g. DJI Phantom 4 RTK camera.
This option is mainly intended to be used in cases when highly accurate reference data is available (RTK/PPK) and it can help to achieve higher accuracy when no GCPs are available.
-
Just an update on Metashape v1.6 processing of satellite images...
With latest release, I still find big gaps in depth maps and dense cloud on Tripoli sample image dataset....
-
Hey Paulo,
But you only have 2 images there?
-
Just trying this new version.
I used a model I've previously built on High. This time I tried Ultra and it worked with 2,000,000 faces, but when I upped this to 6,000,000 faces to produce a more detailed mesh I got "Assertion Failure at line 118" after processing for a couple of hours (BTW, I selected "reuse depth maps" on that second build).
Also, this version has the out-of-core processing. I presume that is using Disk to offload things to? There don't seem to be any settings for this?
Good work - looking forward to the next release.
-
I used a model I've previously built on High. This time I tried Ultra and it worked with 2,000,000 faces, but when I upped this to 6,000,000 faces to produce a more detailed mesh I got "Assertion Failure at line 118" after processing for a couple of hours (BTW, I selected "reuse depth maps" on that second build).
Also, this version has the out-of-core processing. I presume that is using Disk to offload things to? There don't seem to be any settings for this?
Hello cbnewham,
Do you have the related processing log that you can share?
The out-of-core mesh generation works for PSX projects, providing that Fine-Level Task Subdivision is enabled in the Advanced preferences tab.
Currently the temporary data is saved in the project.files folder. Also at the moment there's a tweak, that allows to define the temp folder for the mesh generation.
-
Alexey,
No trimming tweaks were applied. If you try the Doll sample you will see similar results (holes) where there's poor depth map coverage (under the skirt) and mesh_trimming_radius 30 doesn't seem to have any effect when applied in this build. (obviously 1.5.0 handled does cases perfectly without any tweaks...but you already know that 😉). Visibility Consistent Mesh Generation, Metashape up to v 1.5.1 ( & to a lesser degree Metashape 1.5.3+ with mesh_trimming_radius set to "30") correctly remeshed those holes/gaps. So there must be a way to fix this.
Yeah I'm seeing some GPU utilization but what appears in the console has changed quite a bit so it was harder to figure out what was happening.
Out-of-core meshing is perfectly working though. So big congrats on this achievement! Never went over 14GB of Ram usage on data-sets (with 150+/200+ 24MP images) that would fail before.
Cheers
MAK
Cheers
Alexey,
After further testing it unfortunately does look like the main/mesh_trimming_radius tweak doesn't have any effect anymore.
Example the Doll: With and Without main/mesh_trimming_radius tweak (set to 0 or 30 or 90) the mesh is exactly the same with exactly the same number of polys and vertices:
(https://i.postimg.cc/DzVDqvwT/Annotation-2019-10-22-160627.jpg)
Cheers
Mak
-
Hello Mak,
You can disable trimming of the out-of-core method by using main/mesh_tgv_ooc_trimming_enabled tweak and setting its value to False.
If you want to switch off out-of-core mesh generation method and use the one from the latest 1.5 release, then you need to set to False another tweak: main/mesh_tgv_ooc_enabled
-
Alexey,
Interesting. Just tested with the doll and it did indeed close the holes (also went crazy on top but this was sort of expected given that the mesh was already poorly reconstructed.)
(https://i.postimg.cc/fyPJvxf2/Annotation-2019-10-22-165737.jpg)
I'm now testing on an other project to see how it turns out.
Maybe there's a better balance to find with the out-of-core method trimming and make it less aggressive (instead of having to turn it off/False) ?
Will post my finding once they are processed.
There's no way I'm going back to non-out-of-core method 😮😊. I've been waiting for it for years and its finally delivered (and is a god send for some projects). Also the reconstructed mesh details is the best so far. Only thing left to fix/tweak is those holes/vertices issues.
UPDATE:
Looks a lot better with main/mesh_tgv_ooc_trimming_enabled "FALSE" ! And without any messy polys running inside the mesh or all over the place like in 1.5.2+.
(https://i.postimg.cc/DwcTPXsd/Annotation-2019-10-22-182725.jpg)
Cheers
Mak
-
Hi Alexey,
I set the number of faces to 4,000,000 and it didn't generate the error again. I will post the log if it happens again.
Thanks!
-
Just an update on Metashape v1.6 processing of satellite images...
With latest release, I still find big gaps in depth maps and dense cloud on Tripoli sample image dataset....
Hello Paulo,
Looks like we were able to reproduce similar problems and will implement the fix in the next pre-release update.
-
Great, Alexey,
Thanks for the update
Just to confirm that RPC satellite processing of Tripoli Digital Globe data sample no longer encounters gaps in dense cloud processing... Thanks for the good work! :D
-
Another issue with the new out-of-core mesh generation is that there is some strange inconsistencies in the mesh topology for no apparent reason at all. It varies every time when a mesh is generated even with the
same depth maps correction meant "using the same data set & generating new the depth maps":
Extremely low polycount in some places:
(https://i.postimg.cc/ZqSt2pLw/Annotation-2019-10-25-021720.jpg)
Regenerated the mesh and:
(https://i.postimg.cc/CL5FsGvD/Annotation-2019-10-25-021613.jpg)
MAK
-
Is there any way to force more UV Island stitching? (at the expense of extra deformation?)
The new uv parameterization is very nice but dense meshes get broken down a bit too much.
-
I was wondering about the new options in the align photos stage : source - estimated - sequential.
I think sequential speaks for itself but what do the others ?
Sequential pre-selection is trying to match the images with a bunch of their neighbors according to the camera labels.
Estimated preselection is based on the estimated camera locations after initial alignment (for example, performed using rough accuracy settings).
One of the possible scenarios of use - when aligning the video sequence frames run Sequential preselection at first and then Estimated preselection, if there are intersecting parts of the image acquisition path.
I've given this a try.
Sequenial Preselection certainly is much quicker. 6 minutes for 594 cameras compared with 81 minutes for "Estimated". Both runs reset alignment.
Running the alignment a second time with Estimated after first doing a Sequential didn't seem to do anything. (Both done at High accuracy)
Is the idea that the Estimated (second alignment) should improve the accuracy?
Can you explain the steps?
-
Hello BjFaeTorphins,
Using Estimated preselection option after quick Sequential alignment run may be reasonable, for example, if the frames are acquired from the vertical zig-zag survey. Sequential preselection option would not give you tie points for the images from the different flight lines, therefore further processing may result in the artifacts (like stairs) on the surface in the areas of the overlap between the flight lines.
-
Hello BjFaeTorphins,
Using Estimated preselection option after quick Sequential alignment run may be reasonable, for example, if the frames are acquired from the vertical zig-zag survey. Sequential preselection option would not give you tie points for the images from the different flight lines, therefore further processing may result in the artifacts (like stairs) on the surface in the areas of the overlap between the flight lines.
After alignment with Sequence when I try a second alignment using Estimated I see the following message...
"Warning: Can't resume matching without keypoints"
What am I doing wrong?
-
Hello BjFaeTorphins,
Have you added new images to the chunk before re-starting the alignment operation?
If you are to use Estimated Preselection in order to find new matching points between the images that have not been matched, you should also enable "Reset Current Alignment" option.
-
...
-
Alexey,
Added GPU-based texture generation support for AMD GPUs
Which drivers should be used ? Because it doesn't seem to work (everything is done on the CPU still) on the latest 19.9.2 WHQL drivers or the latest 19.11.1.
Cheers
MAK
-
Hello Mak,
Which OS and GPU model you are using?
-
Alexey,
Windows 10 64Bit / Radeon VII
MAK
-
Hello Mak,
Are there any GPU-related lines in the log?
Currently there are some limitations:
- no masked photos should be present,
- only 8-bit images on input,
- no ghosting filter applied.
-
Alexey,
I had ghosting filter enabled..Now it does work when it's turned OFF (those limitations should be added in the change log IMO).
GPU blending is up to 8X faster based on the short test I did which is insane! 😮
Once those limitations are lifted it will be pretty awesome. (Ghosting Filter and masking are super important to me)
CPU (Ghosting Filter On)
(https://i.postimg.cc/DfLBpdbB/Annotation-2019-11-06-153245.jpg)
GPU (Ghosting Filter Off)
(https://i.postimg.cc/ydmBN3kB/Annotation-2019-11-06-153316.jpg)
Cheers
MAK
-
The new point cloud confidence option is very useful for filtering noise and simplifying classifying. :)
(http://www.digital-mapping.net/forums/Metashape/cloud_confidence.jpg)
Out-of-core is excellent too! 8)
Thanks Agisoft!
-
With build 9487 I'm getting:
Assertion 239107090 failed at line 7122!
when trying to construct the mesh from depth maps. This occurs both on Ultra and High quality settings.
I attach the log file.
-
Hi cbnewham
Try to update your nvidia GPU with latest driver, my config show 10020/8000
2019-11-11 14:48:45 driver/runtime CUDA: 10000/8000
-
With build 9487 I'm getting:
Assertion 239107090 failed at line 7122!
when trying to construct the mesh from depth maps. This occurs both on Ultra and High quality settings.
Hello cbnewham,
This issue will be fixed in the next version update.
-
Could you please provide more information on the additional calibration coefficients? What additional coefficients are used compared to the standard Brown model and how are the equations different?
Thank you
I am curious to know what Fit additional corrections option do in Optimize Cameras dialog?....
What are these corrections and when should we use this option?
Hello Paulo,
Fit additional corrections option enables additional calibration coefficients that allows a better fit of the lens distortion model. This is especially important for wide angle lens with large radial distortions, e.g. DJI Phantom 4 RTK camera.
This option is mainly intended to be used in cases when highly accurate reference data is available (RTK/PPK) and it can help to achieve higher accuracy when no GCPs are available.
-
How to interpret confidence range?
The color scale bar for confidence goes from 1 to 100, but the Select Confidence Range has a Min of 0 and a Max of 255. Also the color legend goes from 1 to 5, 5 to 10, and then makes a jump from 10 to 100.
(http://www.digital-mapping.net/forums/Metashape/confidence_range.jpg)
-
Hello bisenberger,
The legend shows non-linear scale, since the most of confidence variations that are of interest are observed on 1-5 range. On the legend 255 number is pretty close to 100, actually.
-
Hello bisenberger,
The legend shows non-linear scale, since the most of confidence variations that are of interest are observed on 1-5 range. On the legend 255 number is pretty close to 100, actually.
Thanks Alexey,
Filtering out 1 to 5 makes a big difference in the quality of the point cloud. :)
-
The new release looks so good, with several new settings in alignment and empower the processing.
I got two questions about the new release:
(1) I find that camera calibration part set cx,cy as one tick together, and then omitted the p3 p4 parameters, is there any relationship to the new fit additional adjust function? If that is adding p3p4 in, would this add extra error?
(2) if the estimated alignment function should be used after using "source" method and then do not "reset alignment"?
Thanks for answering this!
-
With build 9487 I'm getting:
Assertion 239107090 failed at line 7122!
when trying to construct the mesh from depth maps. This occurs both on Ultra and High quality settings.
Hello cbnewham,
This issue will be fixed in the next version update.
Alexey,
Good to know. Just got hit by it when processing on Ultra High finished.
Will the next update also re-enable vertex colors ?
Also, in addition to the low density poly patches randomly scattering the meshes (since the out-of-core mesh update) which I reported https://www.agisoft.com/forum/index.php?topic=11331.msg51488#msg51488 (https://www.agisoft.com/forum/index.php?topic=11331.msg51488#msg51488) I'm also seeing some other anomalies like seen below:
(https://i.postimg.cc/hj2hh7hk/Annotation-2019-11-14-162044.jpg)
(https://i.postimg.cc/gcBnWRCR/Annotation-2019-11-14-025302.jpg)
(https://i.postimg.cc/zXb3XvNQ/Annotation-2019-11-14-025331.jpg)
It happens with main/mesh_tgv_ooc_trimming_enabled True & False.
It would also be good if trimming was dialed a little bit :)
main/mesh_tgv_ooc_trimming_enabled FALSE
(https://i.postimg.cc/13Y89Fnb/Annotation-2019-11-13-204133.jpg)
main/mesh_tgv_ooc_trimming_enabled TRUE
(https://i.postimg.cc/QM9HVMcC/Annotation-2019-11-13-204110.jpg)
Cheers
MAK
-
This I would really like to know as well!
Could you please provide more information on the additional calibration coefficients? What additional coefficients are used compared to the standard Brown model and how are the equations different?
Thank you
I am curious to know what Fit additional corrections option do in Optimize Cameras dialog?....
What are these corrections and when should we use this option?
Hello Paulo,
Fit additional corrections option enables additional calibration coefficients that allows a better fit of the lens distortion model. This is especially important for wide angle lens with large radial distortions, e.g. DJI Phantom 4 RTK camera.
This option is mainly intended to be used in cases when highly accurate reference data is available (RTK/PPK) and it can help to achieve higher accuracy when no GCPs are available.
-
Pre-release version has been updated.
Vertex colors calculation has been enabled for the depth maps based mesh (new option in Tools Menu - Colorize Vertices should also allow to apply the colors from the images to the vertices of already generated models). Also the "anomalies" reported by Mak should be fixed (small inverted polygons).
-
Alexey,
Awesome work. Will try it asap. Is there any changes to GPU texturing? Are masks and ghosting filter still not supported?
EDIT: Looks like GPU texture generation works with Masks & ghosting filter. Awesome! Just did a quick test. it's was 7x faster than on the CPU. 😮
Cheers
Ike & Tina
-
Select by point is not working how I would expect it to.
Build 9397.
I create the sparse could and see that I have sections which are mis-aligned.
I select some points and press "Filter photos by point". I correctly get a dozen cameras. I reset them and re-align. All OK.
I then see some other points that are not right and select those and follow the same procedure. In the "details" list of the photos I get a list of several cameras. However, if I Ctr-A them or even select them individually and do "Reset alignment", I am told it's going to reset far more cameras - in fact, I believe the set of cameras to be aligned each time is the summation of previous selections and it is not being reset when a new "filter by point" is done.
In one case I had filtered only 7 cameras (and the list showed only 7 cameras) but the reset said I was resetting 90! If I carry out the realignment it is clearly realigning 90 due to the time it takes.
If I remove filters, select one camera and do "Reset alignment" then it correctly says I am resetting one camera. If I then do select by point as above it all works again as expected.
I wish there was also a way to see what's not aligned clearly in the photo display - something I've requested before. Grey ticks on a dark grey background are next to useless. Perhaps consider allowing us to change the colours for various parts of Metashape? A way to sort in details by "disabled camera" would also be of great assistance.
Thanks - these are small issues in an otherwise fantastic program.
-
Select by point is not working how I would expect it to.
Build 9397.
I create the sparse could and see that I have sections which are mis-aligned.
I select some points and press "Filter photos by point". I correctly get a dozen cameras. I reset them and re-align. All OK.
I then see some other points that are not right and select those and follow the same procedure. In the "details" list of the photos I get a list of several cameras. However, if I Ctr-A them or even select them individually and do "Reset alignment", I am told it's going to reset far more cameras - in fact, I believe the set of cameras to be aligned each time is the summation of previous selections and it is not being reset when a new "filter by point" is done.
...
I've had the same bug that has been in every version since 1.3? The problem is that the left hand list of photos retains the previous selection when you use the filter or even just select a range of photos in the "Photos" window.
I work around it by looking for a photo from the filtered list in the left hand list, manually clicking on it in the left hand list, and then performing the mouse drag selection in the filtered list. This always gets the correct photos for realignment.
Note: I did re-report this bug recently so it may be you have a slightly different issue.
https://www.agisoft.com/forum/index.php?topic=11531.0
-
Hello cbnewham and mrc,
We'll try to fix it before version 1.6.0 is officially released. It appears that Remove Cameras is working properly from the Photos pane (applied to the selection in the pane only), while Reset Alignment is not considering the pane's selection.
-
Hi Alexey,
I'm not sure if this is a bug, when I export depth map using GUI, if I specify the image route individually, it will return error. If I don't specify route but remain it as {filename}, and select folder in the end, it would work.
See attached screen snippings.
Thanks!
-
Awesome work. Will try it asap. Is there any changes to GPU texturing? Are masks and ghosting filter still not supported?
EDIT: Looks like GPU texture generation works with Masks & ghosting filter. Awesome! Just did a quick test. it's was 7x faster than on the CPU.
Yes, currently no limitation to masked photos or ghosting filter, also the GPU acceleration is support for the tiled model texture generation sub-step.
The texture atlas size should be, however, a multiple of 4. There are still some other limitations, but mainly for quite uncommon cases related to big number of channels (higher than 6) in the source images.
-
Not sure if this is a bug or feature, but using the 'convert images' on a set of fisheye images, the output seems still to be a fisheye rather than rectilinear image.
Is it just that the conversion removes distortions to make a 'perfect' fisheye, or should it be removing the whole fisheye effect plus distortions to make a 'perfect' rectilinear image?
I thought that in previous versions i had been able to use what used to be called 'undistort' to remove the fisheye effect, but that doesn't seem to now be the case, although i can see some subtle differences in the output, still fisheye though!
Images are set to fisheye in 'camera calibration'. There is a single calibration group; just 29 images in the project using nikon d810 and nikon 16mm full frame fisheye, and the reconstruction is very good so i believe the coefficients are right. see attached.
sadly can't share the input images, but if i get a chance will try to recreate the issue with a set safe to share.
-
Hey Alexey, thanks for adding support for GPU based Texturing. Are you also planning GPU support for the Densification (not Depth Matching) soon? This also takes really long time for our data sets. :)
Thanks,
Khedar
-
I'm getting a crash running build Standard 9617. There is no error in the log file. I'm running an align of multiple chunks in batch mode. It seems to happen during the 2nd chunk. last entry is...
2019-12-01 15:20:18 Selecting pairs...
2019-12-01 15:20:18 Found 2 GPUs in 0.001 sec (CUDA: 0.001 sec, OpenCL: 0 sec)
2019-12-01 15:20:18 Using device: GeForce GTX 1080 Ti, 28 compute units, free memory: 9233/11264 MB, compute capability 6.1
2019-12-01 15:20:18 driver/runtime CUDA: 10010/8000
2019-12-01 15:20:18 max work group size 1024
2019-12-01 15:20:18 max work item sizes [1024, 1024, 64]
2019-12-01 15:20:18 Using device: GeForce GTX 1080 Ti, 28 compute units, free memory: 9320/11264 MB, compute capability 6.1
2019-12-01 15:20:18 driver/runtime CUDA: 10010/8000
2019-12-01 15:20:18 max work group size 1024
2019-12-01 15:20:18 max work item sizes [1024, 1024, 64]
-
Not sure if this is a bug or feature, but using the 'convert images' on a set of fisheye images, the output seems still to be a fisheye rather than rectilinear image.
Hello James,
The undistortion operation result corresponds to the selected camera type, so if the camera type is Fisheye, then you should get an "ideal" fisheye image.
Are you also planning GPU support for the Densification (not Depth Matching) soon? This also takes really long time for our data sets.
Hello Khedar,
Currently we are not planning (at least in short-term future) GPU acceleration of other major workflow steps.
I'm getting a crash running build Standard 9617. There is no error in the log file. I'm running an align of multiple chunks in batch mode. It seems to happen during the 2nd chunk.
Hello BjFaeTorphins,
Was the crash reporter window shown for you?
Also which OS version and which driver version you are using?
-
I'm getting a crash running build Standard 9617. There is no error in the log file. I'm running an align of multiple chunks in batch mode. It seems to happen during the 2nd chunk.
Hello BjFaeTorphins,
Was the crash reporter window shown for you?
Also which OS version and which driver version you are using?
Metashape exited silently. no error dialogue or message written to logfile. Also loses setting to save a logfile and it's location.
-
Hello BjFaeTorphins,
If you are using Windows, can you please check the Windows Event Viewer -> Windows Logs -> Application section and check, if there are any lines marked with the red exclamation sign related to metashape.exe? If there are any, please check the Faulting Module Name line and the Exception Code.
-
Hello BjFaeTorphins,
If you are using Windows, can you please check the Windows Event Viewer -> Windows Logs -> Application section and check, if there are any lines marked with the red exclamation sign related to metashape.exe? If there are any, please check the Faulting Module Name line and the Exception Code.
Yes I'm using windows and right enough there are eventvwr records...
Faulting application name: metashape.exe, version: 1.0.0.1, time stamp: 0x5dda841b
Faulting module name: ucrtbase.dll, version: 10.0.18362.387, time stamp: 0x4361b720
Exception code: 0xc0000409
Fault offset: 0x000000000006db8e
Faulting process id: 0x1564
Faulting application start time: 0x01d5a94dd7919c83
Faulting application path: C:\Program Files\Agisoft\Metashape\metashape.exe
Faulting module path: C:\WINDOWS\System32\ucrtbase.dll
Report Id: 10c82760-251f-42af-a201-fa87adbe2e0b
Faulting package full name:
Faulting package-relative application ID:
-
Hello BjFaeTorphins,
If the crash is reproducible, can you please confirm if Metashape still crashes, when you start it as administrator (using Run as Administrator command)?
-
Hello BjFaeTorphins,
If the crash is reproducible, can you please confirm if Metashape still crashes, when you start it as administrator (using Run as Administrator command)?
Just to mention this is not running batch mode. it's just doing an alignment interactively.
Yes run i tried run as administrator which then failed silently and got the following in the event log...
Faulting application name: metashape.exe, version: 1.0.0.1, time stamp: 0x5dda841b
Faulting module name: ucrtbase.dll, version: 10.0.18362.387, time stamp: 0x4361b720
Exception code: 0xc0000409
Fault offset: 0x000000000006db8e
Faulting process id: 0x54f0
Faulting application start time: 0x01d5aa744f95857b
Faulting application path: C:\Program Files\Agisoft\Metashape\metashape.exe
Faulting module path: C:\WINDOWS\System32\ucrtbase.dll
Report Id: 6776fd24-291c-424d-90dc-355f05591a61
Faulting package full name:
Faulting package-relative application ID:
-
Alxexey,
main/mesh_tgv_ooc_trimming_enabled is no longer available in build 9617. Has it been replaced by another tweak ??
Mak
-
Hello Mak,
You can try different Interpolation options now. Disabled means max trimming, Extrapolated - no trimming.
-
Alexey,
I didn't notice the interpolation feature now being available for the Depthmaps workflow!
Thanks
Mak
-
Hello,
in latest pre release, we can get variance calculations for markers during bundle adjustment (Optimize cameras with Estimate tie point variance option ticked).
However when exporting Markers Reference CSV, the Save Variance option is greyed out while for Cameras Reference Export it is available...
-
Hello Paul,
Thank you for letting us know, we'll fix that in the next update.
Meanwhile, if you need to save the contents of the Reference pane tab to the file, you can select all lines and use CTRL+C to copy the data to the clipboard and then CTRL+V to paste it to the text editing application.
-
For the Mac OS X users I would like to mention that in the latest 1.6.0 update (build 9760) some changes related to the activation system have been introduced that hopefully should solve the problems with the loss of the activation due to the Mac OS X version upgrade/update and annoying "can't write license" problem.
In 1.6.0 version now for the first time activation on the given machine you'll be asked to input the admin password, as this first time activation requires elevated privileges. The "rehostable id" folder will be created in /Library/Application Support/Reprise/ directory, and the license file will be put by default to /Library/Application Support/Agisoft/Metashape Pro/ folder. Both rehostable and license folders are created with read access for all users, so it shouldn't be the problem to use them by other non-admin accounts on the same OS.
In the older versions the default location of "rehostable id" was in /var/tmp/ folder which got wiped with every Mac OS X version update.
If version 1.5.5 is already activated on the computer, then 1.6.0 would properly read the old rehostable location (and also would look for .lic file in the application bundle contents), but I would rather suggest to deactivate the license and re-activate it with the same key to test the new approach.
In case the version 1.6.0 has been activated using the new approach, but you would like to use older version on the same machine under the same license, you should run older version from the Terminal using the following command that sets up the environment variable to the new rehostable directory location:
AGISOFT_REHOST_PATH=/Library/Application\ Support/Reprise /Applications/MetashapePro1.5.app/Contents/MacOS/MetashapePro
Any feedback regarding this new Mac OS X activation is appreciated.
-
FYI the latest AMD Radeon Adrenaline 2020 19.12.2 WHQL drivers crash the whole system when generating Depth Maps. Reverting to the previous driver release "fixes" it.
(Running Win 10 1909+ Radeon VII)
Mak
-
Hello Mak,
Thanks for letting me know, we'll try to check that.
-
Alexey,
Today's 19.12.3 drivers seem to work: https://www.amd.com/en/support/kb/release-notes/rn-rad-win-19-12-3
One odd thing though is that since 19.12.2 Windows task manager no longer shows the Compute 2 graph (its now only Compute 0, 1 & 3).
Cheers
Mak
-
Unfortunately even the latest AMD drivers are highly unstable. The crash the GPU during intensive compute task wimeyher it's in Metashape or another application (rendering ik Blender for example).
Mak
-
I am excited about the improvements in the Python API version 1.6.0. Can you please explain the meaning of the following parameters (they are parameters for buildDepthMaps, buildDenseCloud, etc.:
workitem_size_cameras
max_workgroup_size
Is it possible that if my PC has sufficient memory, I can increase these values for faster processing?
-
Alexey,
Once 1.6 release you should really communicate about the depth maps meshing workflow and absolutely make it clear that this is the one to use if a project doesn't require the generation of a dense cloud. The speed gains now that Metashape is fully out of core, are insane. Just did a comparison on a small data-set (77x 24Mp photos of a cliff wall) using Utlra Settings:
- 10 Hours using the old Dense Cloud workflow (because I originally couldn't do it using the depth maps workflow prior to v1.6)
- 2h30 Using the Depth Maps workflow in v 1.6
EDIT: Version 1.6.0 build 9852
Reduced VRAM consumption in Solid model view mode.
Solid Model View is finally usable after all these years! Perfectly smooth even with super dense meshes (+15M polys)!
Mak
-
build 9855 hard crashes during image alignment (detecting points) when GPU is enabled.
-
I can confirmed that. I've sent a crash report.
Mak
-
Thanks for reporting.
Seems to be OpenCL related bug, we'll check it and try to fix as soon as possible.
-
Alexey,
Same happens with the previous 9852 Build.
Metahshape crashes as soon as an OpenCL process in invoked (alignement, depth-maps generation, mesh generation etc..)
Mak
-
We have updated the pre-release build to 9862 - the issue with OpenCL should be fixed now.
-
Alexey,
Was just testing it & can confirm that it's fixed.
Time for you guys/gals to enjoy the holidays!
Cheers
Mak
-
Solid Model View is finally usable after all these years! Perfectly smooth even with super dense meshes (+15M polys)!
Hello Mak,
In the next 1.6.0 update it would be possible to define the max amount of VRAM available for model view rendering via Advanced preferences tab.
For example, for solid view mode 1 GB VRAM would allow for smooth rendering of approx. 30 M faces.
-
Alexey,
I've just processed a 40M mesh & everything was perfectly smooth :) Good to know that this would be tweak-able soon.
Is there now a limit to how dense mesh can be ?
Cheers
Mak
-
Merging Dense clouds Problem
I have divided my project into 2 chunks to calculate dense cloud with Point Confidence... However when I merge the 2 chunks, I lose the point confidence in resulting merged dense cloud....
-
Using build 9862 i generated mesh from depth maps which has worked pretty good. Image 1 shows a small piece of the shaded view.
I then tried building texture the result of which is shown in image 2.
What am I doing wrong?
-
BjFaeTorphins,
Can you post your system specs & project settings ?
Mak
-
BjFaeTorphins,
Can you post your system specs & project settings ?
Mak
2019-12-28 07:24:56 BuildTexture: mapping mode = Generic, page count = 1, texture size = 4096, texture type = Diffuse map, source data = Images, blending mode = Mosaic, fill holes = 1, ghosting filter = 1, enable_gpu = 1, relaxed_precision = 1
2019-12-28 07:24:56 Parameterizing texture atlas...
2019-12-28 07:24:56 Model size: 44373248 faces, 22195119 vertices
2019-12-28 07:26:59 Agisoft Metashape Standard Version: 1.6.0 build 9862 (64 bit)
2019-12-28 07:26:59 Platform: Windows
2019-12-28 07:26:59 CPU: Intel(R) Core(TM) i9-7980XE CPU @ 2.60GHz (desktop)
2019-12-28 07:26:59 CPU family: 6 model: 85 signature: 50654h
2019-12-28 07:26:59 RAM: 127.7 GB
2019-12-28 07:27:00 OpenGL Vendor: NVIDIA Corporation
2019-12-28 07:27:00 OpenGL Renderer: GeForce GTX 1080 Ti/PCIe/SSE2
2019-12-28 07:27:00 OpenGL Version: 4.6.0 NVIDIA 441.41
2019-12-28 07:27:00 Maximum Texture Size: 32768
2019-12-28 07:27:00 Quad Buffered Stereo: not enabled
2019-12-28 13:07:24 Found 2 GPUs in 0.001 sec (CUDA: 0 sec, OpenCL: 0.001 sec)
2019-12-28 13:07:24 Using device: GeForce GTX 1080 Ti, 28 compute units, free memory: 9225/11264 MB, compute capability 6.1
2019-12-28 13:07:24 driver/runtime CUDA: 10020/8000
2019-12-28 13:07:24 max work group size 1024
2019-12-28 13:07:24 max work item sizes [1024, 1024, 64]
-
Are you using the latest Nvidia Drivers?
What are the specs of your photos (JPG/PNG..16/24/32bit) ? Could be an issue the new GPU accelerate texture generation. One way to find out would be to disable it (Alexey could probably share the tweak to manually turn it off).
Mak
-
Are you using the latest Nvidia Drivers?
What are the specs of your photos (JPG/PNG..16/24/32bit) ? Could be an issue the new GPU accelerate texture generation. One way to find out would be to disable it (Alexey could probably share the tweak to manually turn it off).
Mak
version 441.41 released 26 November so not that old. There is a newer build released on 10 December which i'll download.
-
I've had a few failures when building mesh from depth maps. Here are a couple of example error messages...
2019-12-28 14:22:00 level part done in 172.15 s (53% loading + 0% blocks + 0% sorting + 1% iters (0% queueing) + 46% saving)
2019-12-28 14:22:02 Peak memory used: 5.64 GB at 2019-12-28 12:04:39
2019-12-28 14:22:02 deleting all temporary files...
2019-12-28 14:33:48 done in 705.696 s
2019-12-28 14:33:48 Finished processing in 25429.5 sec (exit code 0)
2019-12-28 14:33:48 Error: Can't rename file or directory: Access is denied (5): H:/2019/2019-11-17 HMS/2019-11-26 AM/Gangway.files/11/0/model.tmp/leaves/u/pre
and...
2019-12-28 16:41:56 level part done in 136.006 s (49% loading + 0% blocks + 0% sorting + 1% iters (0% queueing) + 50% saving)
2019-12-28 16:41:58 Peak memory used: 5.10 GB at 2019-12-28 15:40:19
2019-12-28 16:41:58 deleting all temporary files...
2019-12-28 16:49:59 done in 481.096 s
2019-12-28 16:49:59 Finished processing in 14239.8 sec (exit code 0)
2019-12-28 16:49:59 Error: Can't rename file or directory: Access is denied (5): H:/2019/2019-11-17 HMS/2019-11-18 AM/AA Guns to Casemates 333mS.files/13/0/model.tmp/leaves/u/pre
-
Alexey,
I'm still getting those random topology issues on meshes (low poly patches & inverted polygons) as reported a few weeks ago while generating meshes via depth maps. Reconstruction detail & noise reduction have been greatly improved and this is now the last oddity I'm seing. 1.6 is shaping up to be a huge update.
(https://i.postimg.cc/8zKCsZ8G/Annotation-2019-12-28-161814.jpg)
(https://i.postimg.cc/yND8R0BZ/Annotation-2019-12-28-161852.jpg)
(https://i.postimg.cc/y8YdVwbn/Annotation-2019-12-28-161921.jpg)
It happens on all quality settings (LOW/MED/HIGH/ULTRA-HIGH)
(https://i.postimg.cc/7hP73N4w/Annotation-2019-12-28-195731.jpg)
Cheers
Mak
-
Hello BjFaeTorphins,
Is it the latest pre-release build that you are using?
Is the project data stored locally or the path leads to the network/external drive?
-
Hello Mak,
We'll try to reproduce the issue, but if you can share the project (in PSZ format) with the depth maps related to the problematic area that we can re-use to observe the issue, it would be helpful.
-
Hello BjFaeTorphins,
Is it the latest pre-release build that you are using?
Is the project data stored locally or the path leads to the network/external drive?
yes Build 9862.
it was on the NAS in the case of the two errors i logged above.
I did try moving the project files to the local SSD drive and leaving the photos on the NAS and so far that has been working. I've had some successful mesh builds so i've not been able to isolate what combination of things always cause the error.
-
Are you using the latest Nvidia Drivers?
What are the specs of your photos (JPG/PNG..16/24/32bit) ? Could be an issue the new GPU accelerate texture generation. One way to find out would be to disable it (Alexey could probably share the tweak to manually turn it off).
Mak
version 441.41 released 26 November so not that old. There is a newer build released on 10 December which i'll download.
I've updated to the latest 441.66 20th December drivers and still get the issue.
I note if i go back into the build texture item on the workflow menu that the mapping mode is set to Keep UV, I've tried setting this to Generic and re-running the texture build but the same happens.
-
Hello Mak,
We'll try to reproduce the issue, but if you can share the project (in PSZ format) with the depth maps related to the problematic area that we can re-use to observe the issue, it would be helpful.
Alexey,
I've sent you WEtransfer link by mail.
Cheers
Mak
-
I just noticed that there are an excessive number of handles when running and alignment of 8000 images. 7000 out of 8000 aligned first time. it was when selecting all images and running the alignment for the 2nd time to add the unaligned images that i noticed the handles. So far this has not caused a problem but i have seen windows become unresponsive when a process does not clean up handles.
-
Hello BjFaeTorphins,
Is it the latest pre-release build that you are using?
Is the project data stored locally or the path leads to the network/external drive?
yes Build 9862.
it was on the NAS in the case of the two errors i logged above.
I did try moving the project files to the local SSD drive and leaving the photos on the NAS and so far that has been working. I've had some successful mesh builds so i've not been able to isolate what combination of things always cause the error.
Alexey,
I got the "Can't rename file or directory: Access is denied" running a mesh from depth maps on the local SSD drive so the NAS vs local drive theory is not the answer.
-
Hello BjFaeTorphins,
I think that something messed up with the access rights - for some reason the created files cannot be removed or rewritten. Do you observe any similar behavior when trying to export anything from Metashape to the new file or to the file with existing filename?
-
Hello Alexey,
Aligned photos with 9852 version
Build mesh - deph map 9862 version
I had these errors:
2019-12-24 13:18:52 checking for missing images...Checking for missing images...
2019-12-24 13:18:54 done in 2.637 sec
2019-12-24 13:18:54 Finished processing in 2.637 sec (exit code 1)
2019-12-24 13:18:54 BuildModel: quality = Ultra high, depth filtering = Mild, source data = Depth maps, surface type = Arbitrary, face count = High, volumetric masking = 0, OOC version, interpolation = Enabled, vertex colors = 1
2019-12-24 13:18:54 Generating depth maps...
2019-12-24 13:18:54 Initializing...
2019-12-24 13:18:54 sorting point cloud... done in 1.812 sec
2019-12-24 13:18:56 processing matches... done in 1.642 sec
2019-12-24 13:18:58 initializing...
2019-12-24 13:19:16 selected 9902 cameras from 10354 in 17.673 sec
2019-12-24 13:19:16 groups: 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 100 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 100
2019-12-24 13:19:16 29013 of 9902 used (293.001%)
2019-12-24 13:19:16 scheduled 100 depth map groups
2019-12-24 13:19:16 saved depth map partition in 0.303 sec
2019-12-24 13:19:28 loaded depth map partition in 12.22 sec
2019-12-24 13:19:28 Initializing...
2019-12-24 13:19:28 Found 3 GPUs in 0.002 sec (CUDA: 0.001 sec, OpenCL: 0.001 sec)
2019-12-24 13:19:29 Using device: GeForce RTX 2070 SUPER, 40 compute units, free memory: 6711/8192 MB, compute capability 7.5
2019-12-24 13:19:29 driver/runtime CUDA: 10020/8000
2019-12-24 13:19:29 max work group size 1024
2019-12-24 13:19:29 max work item sizes [1024, 1024, 64]
2019-12-24 13:19:29 Using device: GeForce RTX 2070 SUPER, 40 compute units, free memory: 6711/8192 MB, compute capability 7.5
2019-12-24 13:19:29 driver/runtime CUDA: 10020/8000
2019-12-24 13:19:29 max work group size 1024
2019-12-24 13:19:29 max work item sizes [1024, 1024, 64]
2019-12-24 13:19:30 Using device: GeForce RTX 2080 Ti, 68 compute units, free memory: 9133/11264 MB, compute capability 7.5
2019-12-24 13:19:30 driver/runtime CUDA: 10020/8000
2019-12-24 13:19:30 max work group size 1024
2019-12-24 13:19:30 max work item sizes [1024, 1024, 64]
2019-12-24 13:19:30 Using CUDA device 'GeForce RTX 2080 Ti' in concurrent. (2 times)
2019-12-24 13:19:33 Loading photos...
2019-12-24 13:21:51 loaded photos in 138.19 seconds
2019-12-24 13:21:51 Generating depth maps...
2019-12-24 13:21:52 [GPU] estimating 9188x5090x1024 disparity using 1532x1280x8u tiles
2019-12-24 13:21:52 [GPU] estimating 8058x4928x2624 disparity using 1343x1280x8u tiles
...
2019-12-30 08:48:39 Error: Unsupported depth image
...
2019-12-30 15:04:51 Error: Null image
...
2019-12-30 15:16:20 Error: Too small depth map!
...
2019-12-31 19:56:42 Finished processing in 90728.3 sec (exit code 0)
2019-12-31 19:56:42 Error: Broken data order in file D:/10445 clean.files/0/0/model.tmp/cubes/1/12.data at index=354263496
But in the meantime
Thanks to devolp this beautiful software!
Happy new Year !!!
-
Hello eltama,
Do you have these errors during the depth maps generation step or already during the mesh reconstruction (all depth maps are already generated)?
-
I checked reuse deph map (ultra setting) and yes I have errors on both steps, now I had installed the latest version 9925, and build mesh without reuse depth maps with high quality.
I have some duplicated images, now deleted, runs as admin.
I noticed a strange behavior, even if I am admin, I can’t disable read only mode on files.
Best regards
-
Hello BjFaeTorphins,
I think that something messed up with the access rights - for some reason the created files cannot be removed or rewritten. Do you observe any similar behavior when trying to export anything from Metashape to the new file or to the file with existing filename?
Alexey
After the error happens the model.tmp directory mentioned in the error message is gone. I've not seen any permission errors and the project file can be saved after the error.
-
Are you using the latest Nvidia Drivers?
What are the specs of your photos (JPG/PNG..16/24/32bit) ? Could be an issue the new GPU accelerate texture generation. One way to find out would be to disable it (Alexey could probably share the tweak to manually turn it off).
Mak
version 441.41 released 26 November so not that old. There is a newer build released on 10 December which i'll download.
I've updated to the latest 441.66 20th December drivers and still get the issue.
I note if i go back into the build texture item on the workflow menu that the mapping mode is set to Keep UV, I've tried setting this to Generic and re-running the texture build but the same happens.
Alexey
I've updated to the 9925 build and am still getting this weird looking almost greyscale texture.
-
Thanks for the wonderful update !
it's wonderfull to be able to work with a perfectly smooth canvas with a 70 millons triangles model
:D
however i have a problem concerning textured models as you can see on the picture not all the textures are displayed, why (unchecking VBO support is solving the issue) ? (last update, latest drivers 40000Mb vram allocated)
-
Are you using the latest Nvidia Drivers?
What are the specs of your photos (JPG/PNG..16/24/32bit) ? Could be an issue the new GPU accelerate texture generation. One way to find out would be to disable it (Alexey could probably share the tweak to manually turn it off).
Mak
version 441.41 released 26 November so not that old. There is a newer build released on 10 December which i'll download.
I've updated to the latest 441.66 20th December drivers and still get the issue.
I note if i go back into the build texture item on the workflow menu that the mapping mode is set to Keep UV, I've tried setting this to Generic and re-running the texture build but the same happens.
Alexey
I've updated to the 9925 build and am still getting this weird looking almost greyscale texture.
Alexey
I went back to a model that i built using v1.3.5 and re did the texture process. See the two attached images.
-
I checked reuse deph map (ultra setting) and yes I have errors on both steps, now I had installed the latest version 9925, and build mesh without reuse depth maps with high quality.
I have some duplicated images, now deleted, runs as admin.
I noticed a strange behavior, even if I am admin, I can’t disable read only mode on files.
Best regards
Hello Alexey
I had this error, build mesh from a new deph map (didn't check reuse map) my setting: quality high, face count high:
...
2020-01-02 23:31:24 finished treetop part #1013/1205: 433 treetop nodes, 73.5762 KB registry
2020-01-02 23:31:24 Creating treetop for part #1014/1205...
2020-01-02 23:31:31 finished treetop part #1014/1205: 433 treetop nodes, 73.5762 KB registry
2020-01-02 23:31:31 Creating treetop for part #1015/1205...
2020-01-02 23:31:39 Peak memory used: 17.92 GB at 2020-01-02 06:40:07
2020-01-02 23:31:39 deleting all temporary files...
2020-01-02 23:32:01 done in 22.855 s
2020-01-02 23:32:01 Finished processing in 171202 sec (exit code 0)
2020-01-02 23:32:01 Error: Assertion 239115764 failed at line 69!
-
Hello eltama,
Can you please confirm that the same issue is observed in the release version 9925?
-
Hello BjFaeTorphins,
What is the input data format - file type, bit depth? Was the texture blending performed on GPU?
-
however i have a problem concerning textured models as you can see on the picture not all the textures are displayed, why (unchecking VBO support is solving the issue) ? (last update, latest drivers 40000Mb vram allocated)
Hello Iluvathar,
Can you please specify the GPU model and OS version that you are using? Additionally the texture atlas size, number of pages and bit depth.
Also please confirm that you are using 40 000 Mb of VRAM.
-
Hello BjFaeTorphins,
What is the input data format - file type, bit depth? Was the texture blending performed on GPU?
TIF files, 4000x3000 pixels 32bit LZW Compression. I tried to attached an example but at 17Mb its to big. It's the same camera and settings I've been using for a long time. and still works wiht v1.3.5 and v1.5.5. A GoPro recording 4k video from which I extract frames.
How can I tell if the texture blending was performed on GPU? I used the 9925 build and the two builds before hd the same problem.
-
Alexey
The problem I reported about silent crashes during mesh from depth maps is still happening with build 9925 but I think I understand the circumstances under which it happens.
With reference to the attached screenshot which shows one of the GPU tabs from Task Manager. I've highlighted the "Shared GPU memory" number which is 29.7 out of 63.8GB. The other GPU is also using ~30GB so a total of 60Gb out of 64Gb of Shared GPU memory is used at this point.
An instance of Metashape has just finished mesh from depth maps and it has used this 60Gb of shared GPU memory and when it finished it has not released the memory. I've started up a 2nd instance of Metashape to process another different model and it starts to use GPU shared memory and shortly after the attached screen shot was taken the 2nd process failed silently.
To summarise: I think there is a memory leak in the mesh from depth maps process where it does not release shared GPU memory and eventually if Metashape is not restarted it either hits a handled error and produces the Not enough GPU memory or just fails silently.
HTH
Brian
-
Hello Brian,
Are these 32 bit floating point data type or 32 bit integer?
If the issue is reproducible, please provide the corresponding log from the Console pane related to the texturing operation. Please also try to disable GPUs in the Preferences dialog and re-build the texture using Keep UV option. Let me know if the result is the same or different (correct).
-
Hello Brian,
Are these 32 bit floating point data type or 32 bit integer?
If the issue is reproducible, please provide the corresponding log from the Console pane related to the texturing operation. Please also try to disable GPUs in the Preferences dialog and re-build the texture using Keep UV option. Let me know if the result is the same or different (correct).
Alexey
Thanks for the reply.
I don't know if it is 32bit floating point or integer. Windows reports it as bit depth 32. I'm using the same tool to extract TIFs from video as i have been for a couple of years without problems. It's based on the FFmpeg library.
I disabled the GPUs and re-run the texture and that has fixed the problem. i.e the texture looks right again.
-
Hello Brian,
Can you please send any sample image in this format to support@agisoft.com or as a download link via PM to me?
-
Hello Brian,
Thank you for sending the sample image. It appears that the presence of the alpha-channel in the images is causing the problem with the GPU-based texture blending.
We''\ll try to implement the fix for the issue in the next version update, but meanwhile you should either disable GPU for the texture blending process or remove alpha channel from the source images.
-
Hello eltama,
Can you please confirm that the same issue is observed in the release version 9925?
Hello Alexey
of course, I make new alignement in 9925 in highest mode without error, now I'm waiting for build mesh from deph map (no resume) in highest mode too, images are 61MP.
-
however i have a problem concerning textured models as you can see on the picture not all the textures are displayed, why (unchecking VBO support is solving the issue) ? (last update, latest drivers 40000Mb vram allocated)
Hello Iluvathar,
Can you please specify the GPU model and OS version that you are using? Additionally the texture atlas size, number of pages and bit depth.
Also please confirm that you are using 40 000 Mb of VRAM.
Hello Alexey
The GPU model is a quadro RTX 8000, I use to build 16384 textures in one atlas, i also 8192x2, 4 bands, uint8,
The OS is windows 10
i also worked on an other computer, with a gtx1080, same issue with 16384 or 8192 VBO enabled,
note : when vbo is disabled the textures are coming back to normal
note 2 : when the model has lower complexity textures are displaying just fine with vbo enabled (less than 10-20mio poly and 16384 resolution textures)
-
Hello BjFaeTorphins,
I think that something messed up with the access rights - for some reason the created files cannot be removed or rewritten. Do you observe any similar behavior when trying to export anything from Metashape to the new file or to the file with existing filename?
Alexey
After the error happens the model.tmp directory mentioned in the error message is gone. I've not seen any permission errors and the project file can be saved after the error.
Hi Alexey
I think I've figured out what is causing these errors.
I'm using a Synology NAS to store my models. I have a fast SSD drive on my desktop computer which is continuously synchronised to a Work in Progress folder on the NAS. When I want to do work on a model i copy the files for the model to be worked on from the main folder structure on the NAS to the Work in Progress folder on the NAS which is then automatically synchronised to the SSD drive on the desktop. The software that does this synchronisation between the NAS and the desktop is called Synology Drive. I've been disabling this software when doing mesh from depth maps for about a week now and not had any more errors. Whereas, I was getting the file errors at least 50% of the time previously.
I have been using this arrangement for quite some time (since Photoscan v1.3.5) and not had the file errors before. I suspect that it is something to do with the very high number and size of the temporary files that v1.6 mesh from depth maps out-of-core creates.
I can live with this work-around but it would be more convenient if i didn't need to remember to switch off Synology Drive while running mesh from depth maps. Perhaps introducing a re-try of the file operation if i fails.
HTH
Brian.
-
however i have a problem concerning textured models as you can see on the picture not all the textures are displayed, why (unchecking VBO support is solving the issue) ? (last update, latest drivers 40000Mb vram allocated)
Hello Iluvathar,
Can you please specify the GPU model and OS version that you are using? Additionally the texture atlas size, number of pages and bit depth.
Also please confirm that you are using 40 000 Mb of VRAM.
Hello Alexey
The GPU model is a quadro RTX 8000, I use to build 16384 textures in one atlas, i also 8192x2, 4 bands, uint8,
The OS is windows 10
i also worked on an other computer, with a gtx1080, same issue with 16384 or 8192 VBO enabled,
note : when vbo is disabled the textures are coming back to normal
note 2 : when the model has lower complexity textures are displaying just fine with vbo enabled (less than 10-20mio poly and 16384 resolution textures)
Hello Iluvathar,
Thank you for providing the additional information.
The issue may be related to the unstable driver work with high VBO values. For 70 M poly model 10 GB will be sufficient for textured view display mode, so I can suggest to reduce the amount of dedicated VRAM for VBO option and try to re-open the model. If the textured view is already on, please switch to point cloud and back to textured and see, if it helps.
-
however i have a problem concerning textured models as you can see on the picture not all the textures are displayed, why (unchecking VBO support is solving the issue) ? (last update, latest drivers 40000Mb vram allocated)
Hello Iluvathar,
Can you please specify the GPU model and OS version that you are using? Additionally the texture atlas size, number of pages and bit depth.
Also please confirm that you are using 40 000 Mb of VRAM.
Hello Alexey
The GPU model is a quadro RTX 8000, I use to build 16384 textures in one atlas, i also 8192x2, 4 bands, uint8,
The OS is windows 10
i also worked on an other computer, with a gtx1080, same issue with 16384 or 8192 VBO enabled,
note : when vbo is disabled the textures are coming back to normal
note 2 : when the model has lower complexity textures are displaying just fine with vbo enabled (less than 10-20mio poly and 16384 resolution textures)
Hello Iluvathar,
Thank you for providing the additional information.
The issue may be related to the unstable driver work with high VBO values. For 70 M poly model 10 GB will be sufficient for textured view display mode, so I can suggest to reduce the amount of dedicated VRAM for VBO option and try to re-open the model. If the textured view is already on, please switch to point cloud and back to textured and see, if it helps.
Hello Alexey,
It seems that reducing the amount of VRAM doesn't change anything because it uses only the vram it needs to display the model, if i choose to allow the software to use 48000mb of vram, not all the memory is used to display just one model (it's quite fantastic to be able to show all the aligned chunks smotthly and it uses massively vram depending on the number of models (achaeologics levels sometime 10 models, up to 70-80 mio poly at the same time, severals models with 10-20 millions poly textured without any issues with vbo enabled)
i believe it's as you sayed, due to the unstable driver :/
-
Build normal map texture option is missing. I have made diffuse and occlusion map, should i not be able to do a normal map then?
Running 1.6.0.9925
-
It seems that reducing the amount of VRAM doesn't change anything because it uses only the vram it needs to display the model, if i choose to allow the software to use 48000mb of vram, not all the memory is used to display just one model (it's quite fantastic to be able to show all the aligned chunks smotthly and it uses massively vram depending on the number of models (achaeologics levels sometime 10 models, up to 70-80 mio poly at the same time, severals models with 10-20 millions poly textured without any issues with vbo enabled)
i believe it's as you sayed, due to the unstable driver :/
Hello Iluvathar,
We will try to optimize the textured view mode for the next version update in order to handle higher polygonal models better providing that sufficient VBO RAM limit is defined.
-
[/quote]
Hello Iluvathar,
We will try to optimize the textured view mode for the next version update in order to handle higher polygonal models better providing that sufficient VBO RAM limit is defined.
[/quote]
It would be amazing !
Thanks a lot
-
Build normal map texture option is missing. I have made diffuse and occlusion map, should i not be able to do a normal map then?
Running 1.6.0.9925
Thank you for reporting. We'll check that.
-
Build normal map texture option is missing. I have made diffuse and occlusion map, should i not be able to do a normal map then?
Running 1.6.0.9925
Thank you for reporting. We'll check that.
Alexey,
Generating Normal maps is working just fine on my side with 1.6.0.9925.
Mak
-
Hello Kjellis85,
Yes, normals texture maps generation should be available in the version 1.6.0 build 9925. It requires at least two mesh models in the active chunk, as a different mesh model should be used as a source for the Normal texture map generation. Using the same model will give plain output which cannot be used any further.
-
Hello Kjellis85,
Yes, normals texture maps generation should be available in the version 1.6.0 build 9925. It requires at least two mesh models in the active chunk, as a different mesh model should be used as a source for the Normal texture map generation. Using the same model will give plain output which cannot be used any further.
Ah, this makes much more sense! But I could not find this described in the manual https://www.agisoft.com/pdf/metashape-pro_1_6_en.pdf
-
Alexey,
The latest AMD Radeon 20.1.1 still crash the whole system during heavy OpenCL workloads like depth maps generation (issue since 19.12.2). Maybe you guys can get in touch with them about this as this doesn't seem to be a priority at all for them (most bug fixes are gaming related lately and they usually take more than 6 months to fix DirectCompute/OpenCL related bugs) .
Mak
-
Hello Mak,
Please specify, which GPU model and OS version you are using? So that we could try to reproduce the issue on similar configuration first.
-
Alexey,
Windows 10 v1909
Radeon VII
All Radeon Adrenalin 2020 driver releases (19.12.2/19.12.3/20.1.1) are prone to crash the display driver & hang the whole system during heavy OpenCL compute workloads. Happens in Adobe Lightroom, Blender when using Cycles & Radeon Pro Render & Metashape (usually when generating depth maps on Ultra High or Aligning cameras). Also as I noted in a previous post the "Compute 2" graph has disappeared from the Windows task manager (it now only reports Compute 0, 1 & 3). Something is definitely wrong on the compute side of the drivers.
Cheers
Mak
-
Hello eltama,
Can you please confirm that the same issue is observed in the release version 9925?
Hello Alexey
of course, I make new alignement in 9925 in highest mode without error, now I'm waiting for build mesh from deph map (no resume) in highest mode too, images are 61MP.
Hello Alexey
I did new alignement on highest mode on 9925, build mesh in ultra high is fine but after build deph map there was an error:
2019-12-31 23:58:39 BuildModel: quality = High, depth filtering = Mild, source data = Depth maps, surface type = Arbitrary, face count = High, volumetric masking = 0, OOC version, interpolation = Enabled, vertex colors = 1
2019-12-31 23:58:40 Generating depth maps...
2019-12-31 23:58:40 Initializing...
2019-12-31 23:58:40 sorting point cloud... done in 0.044 sec
2019-12-31 23:58:40 processing matches... done in 1.525 sec
2019-12-31 23:58:41 initializing...
2019-12-31 23:58:53 selected 9819 cameras from 10272 in 12.149 sec
2019-12-31 23:58:54 groups: 98 98 98 98 98 98 98 98 98 98 98 99 98 98 98 98 98 99 98 98 98 98 99 98 99 98 98 98 98 98 99 98 98 98 98 98 99 98 98 98 98 98 99 98 98 98 98 99 98 99 98 98 98 98 98 99 98 98 98 98 98 99 98 98 98 98 98 99 98 98 98 98 99 98 99 98 98 98 98 98 99 98 98 98 98 98 99 98 98 98 98 98 99 98 98 98 98 99 98 99
2019-12-31 23:58:54 29050 of 9819 used (295.855%)
2019-12-31 23:58:54 scheduled 100 depth map groups
2019-12-31 23:58:54 saved depth map partition in 0.297 sec
2019-12-31 23:59:02 loaded depth map partition in 7.641 sec
2019-12-31 23:59:02 Initializing...
2019-12-31 23:59:02 Found 3 GPUs in 0.029 sec (CUDA: 0.028 sec, OpenCL: 0.001 sec)
2019-12-31 23:59:10 Using device: GeForce RTX 2070 SUPER, 40 compute units, free memory: 6711/8192 MB, compute capability 7.5
2019-12-31 23:59:10 driver/runtime CUDA: 10020/8000
2019-12-31 23:59:10 max work group size 1024
2019-12-31 23:59:10 max work item sizes [1024, 1024, 64]
2019-12-31 23:59:10 got device properties in 0.007 sec, free memory in 8.736 sec
2019-12-31 23:59:11 Using device: GeForce RTX 2070 SUPER, 40 compute units, free memory: 6711/8192 MB, compute capability 7.5
2019-12-31 23:59:11 driver/runtime CUDA: 10020/8000
2019-12-31 23:59:11 max work group size 1024
2019-12-31 23:59:11 max work item sizes [1024, 1024, 64]
2019-12-31 23:59:11 Using device: GeForce RTX 2080 Ti, 68 compute units, free memory: 9133/11264 MB, compute capability 7.5
2019-12-31 23:59:11 driver/runtime CUDA: 10020/8000
2019-12-31 23:59:11 max work group size 1024
2019-12-31 23:59:11 max work item sizes [1024, 1024, 64]
2019-12-31 23:59:11 Using CUDA device 'GeForce RTX 2070 SUPER' in concurrent. (2 times)
2019-12-31 23:59:11 Using CUDA device 'GeForce RTX 2070 SUPER' in concurrent. (2 times)
2019-12-31 23:59:11 Using CUDA device 'GeForce RTX 2080 Ti' in concurrent. (2 times)
2019-12-31 23:59:12 Loading photos...
2020-01-01 00:00:42 loaded photos in 89.377 seconds
2020-01-01 00:00:42 Generating depth maps...
2020-01-01 00:00:43 [GPU] estimating 3754x3093x1056 disparity using 1252x1088x8u tiles
2020-01-01 00:00:43 [GPU] estimating 3415x3255x864 disparity using 1139x1088x8u tiles
...
2020-01-09 09:35:26 Depth reconstruction devices performance:
2020-01-09 09:35:26 - 49% done by GeForce RTX 2080 Ti
2020-01-09 09:35:26 - 28% done by GeForce RTX 2070 SUPER
2020-01-09 09:35:26 - 23% done by GeForce RTX 2070 SUPER
2020-01-09 09:35:26 Total time: 4930.66 seconds
2020-01-09 09:35:26
2020-01-09 09:35:26 98 depth maps generated in 5007.46 sec
2020-01-09 09:35:30 saved depth maps block in 3.577 sec
2020-01-09 09:48:04 loaded depth map partition in 750.023 sec
2020-01-09 09:48:04 Building model...
2020-01-09 09:48:04 Compression level: 1
2020-01-09 09:48:04 Preparing depth maps...
2020-01-09 09:48:04 scheduled 100 depth map groups (9772 cameras)
2020-01-09 09:48:04 saved camera partition in 0.005 sec
2020-01-09 09:48:04 loaded camera partition in 0.002 sec
2020-01-09 09:55:11 saved group #1/100: done in 426.708 s, 98 cameras, 14063 MB data, 207.102 KB registry
2020-01-09 09:55:11 loaded camera partition in 0.001 sec
2020-01-09 10:02:26 saved group #2/100: done in 434.326 s, 98 cameras, 14561.7 MB data, 207.102 KB registry
2020-01-09 10:02:26 loaded camera partition in 0.001 sec
2020-01-09 10:09:10 saved group #3/100: done in 404.787 s, 98 cameras, 4404.62 MB data, 207.102 KB registry
2020-01-09 10:09:10 loaded camera partition in 0.001 sec
2020-01-09 10:16:01 saved group #4/100: done in 410.263 s, 98 cameras, 6907.82 MB data, 207.102 KB registry
2020-01-09 10:16:01 loaded camera partition in 0.002 sec
2020-01-09 10:23:04 saved group #5/100: done in 423.715 s, 98 cameras, 7168.78 MB data, 207.102 KB registry
2020-01-09 10:23:04 loaded camera partition in 0.001 sec
2020-01-09 10:30:12 saved group #6/100: done in 427.896 s, 98 cameras, 10572.1 MB data, 207.102 KB registry
2020-01-09 10:30:12 loaded camera partition in 0.001 sec
2020-01-09 10:37:17 saved group #7/100: done in 424.942 s, 98 cameras, 9835.96 MB data, 207.102 KB registry
2020-01-09 10:37:17 loaded camera partition in 0.001 sec
2020-01-09 10:44:34 saved group #8/100: done in 436.523 s, 98 cameras, 12499.9 MB data, 207.102 KB registry
2020-01-09 10:44:34 loaded camera partition in 0.001 sec
depth maps done
2020-01-09 09:35:26 98 depth maps generated in 5007.46 sec
2020-01-09 09:35:30 saved depth maps block in 3.577 sec
2020-01-09 09:48:04 loaded depth map partition in 750.023 sec
2020-01-09 09:48:04 Building model...
2020-01-09 09:48:04 Compression level: 1
2020-01-09 09:48:04 Preparing depth maps...
2020-01-09 09:48:04 scheduled 100 depth map groups (9772 cameras)
2020-01-09 09:48:04 saved camera partition in 0.005 sec
2020-01-09 09:48:04 loaded camera partition in 0.002 sec
2020-01-09 09:55:11 saved group #1/100: done in 426.708 s, 98 cameras, 14063 MB data, 207.102 KB registry
2020-01-09 09:55:11 loaded camera partition in 0.001 sec
...
2020-01-09 19:29:15 saved group #82/100: done in 422.132 s, 98 cameras, 5447.74 MB data, 207.102 KB registry
2020-01-09 19:29:15 loaded camera partition in 0.001 sec
2020-01-09 19:36:19 saved group #83/100: done in 424.036 s, 98 cameras, 6125.03 MB data, 207.102 KB registry
2020-01-09 19:36:19 loaded camera partition in 0.001 sec
2020-01-09 19:43:32 saved group #84/100: done in 432.451 s, 98 cameras, 9505.05 MB data, 207.102 KB registry
2020-01-09 19:43:32 loaded camera partition in 0.001 sec
2020-01-09 19:48:47 Peak memory used: 69.97 GB at 2020-01-09 11:51:13
2020-01-09 19:48:47 deleting all temporary files...
2020-01-09 19:49:01 done in 14.298 s
2020-01-09 19:49:01 Finished processing in 473553 sec (exit code 0)
2020-01-09 19:49:01 Error: Too small depth map!
Now I try to generate deph maps on Higt setting.
I attach the task manager data, In a ultra high deph map generation it seems a low GPU work due to a low vram usage on rtx 2070 s (8 GB) only ~ 2.5 GB compared to on 2080 ti (11GB) ~ 5 GB.
2020-01-09 09:35:26 - 49% done by GeForce RTX 2080 Ti
2020-01-09 09:35:26 - 28% done by GeForce RTX 2070 SUPER
2020-01-09 09:35:26 - 23% done by GeForce RTX 2070 SUPER
in a high deph map generation I noticed a full use of GPU computing and more VRAM usage into the weaker GPU
I haven't the log but it was ~ like this:
- 40% done by GeForce RTX 2080 Ti
- 32% done by GeForce RTX 2070 SUPER
- 28% done by GeForce RTX 2070 SUPER
Best regards
Elmar
-
I have a Radeon VII and a Nvidia Titan XP in a Ryzen 9 3950X system. If I enable the use of the Radeon VII in the settings (1.6.0) I get a system crash within 5 minutes requiring me to hard reset. No logs reveal any issues, Power to the Radeon VII is two 8 pin cables on separate busses on the PSU. Running burn in tests do not reveal any power issues with both GPU's and the CPU at full load. This is the only app I use that will crash within minutes of using the Radeon VII. I assume it's a driver issue, but I cannot be sure, any help is appreciated.
-
I have a Radeon VII and a Nvidia Titan XP in a Ryzen 9 3950X system. If I enable the use of the Radeon VII in the settings (1.6.0) I get a system crash within 5 minutes requiring me to hard reset. No logs reveal any issues, Power to the Radeon VII is two 8 pin cables on separate busses on the PSU. Running burn in tests do not reveal any power issues with both GPU's and the CPU at full load. This is the only app I use that will crash within minutes of using the Radeon VII. I assume it's a driver issue, but I cannot be sure, any help is appreciated.
Try to roll back to the 19.11.1 drivers if you are running the latest Adrenalin 2020 packages (19.12.1/19.12.2/20.1.1).
Mak
-
Try to roll back to the 19.11.1 drivers if you are running the latest Adrenalin 2020 packages (19.12.1/19.12.2/20.1.1).
Mak
That appears to have fixed the issue, thank you so much!
-
Try to roll back to the 19.11.1 drivers if you are running the latest Adrenalin 2020 packages (19.12.1/19.12.2/20.1.1).
Mak
That appears to have fixed the issue, thank you so much!
Glad to know that this worked. As I reported earlier in this thread, the AMD Adrenalin 2020 drivers have buggy/broken OpenCL/Compute support (at least on Radeon VII GPUs).
Mak
-
Does Agisoft have a recommended driver for use with AMD cards? I know these drivers are approved my Microsoft but testing by third party software makers would be helpful considering the cost of the software and GPU. I bought the Radeon VII specifically for use with photogrammetry.
-
Hello BjFaeTorphins,
I think that something messed up with the access rights - for some reason the created files cannot be removed or rewritten. Do you observe any similar behavior when trying to export anything from Metashape to the new file or to the file with existing filename?
Alexey
After the error happens the model.tmp directory mentioned in the error message is gone. I've not seen any permission errors and the project file can be saved after the error.
Hi Alexey
I think I've figured out what is causing these errors.
I'm using a Synology NAS to store my models. I have a fast SSD drive on my desktop computer which is continuously synchronised to a Work in Progress folder on the NAS. When I want to do work on a model i copy the files for the model to be worked on from the main folder structure on the NAS to the Work in Progress folder on the NAS which is then automatically synchronised to the SSD drive on the desktop. The software that does this synchronisation between the NAS and the desktop is called Synology Drive. I've been disabling this software when doing mesh from depth maps for about a week now and not had any more errors. Whereas, I was getting the file errors at least 50% of the time previously.
I have been using this arrangement for quite some time (since Photoscan v1.3.5) and not had the file errors before. I suspect that it is something to do with the very high number and size of the temporary files that v1.6 mesh from depth maps out-of-core creates.
I can live with this work-around but it would be more convenient if i didn't need to remember to switch off Synology Drive while running mesh from depth maps. Perhaps introducing a re-try of the file operation if i fails.
HTH
Brian.
Hi Alexey
I've changed my mind about the circumstances of these errors.
I'm now quite sure its when .psx and .files are stored on the NAS mounted shared drive. I checked and all of the errors that i've seen happen when those files are stored on the NAS and does not happen on my local SSD drive even if Synology Drive synchronisation is enabled.
HTH
-
Hello eltama,
Do you have these errors during the depth maps generation step or already during the mesh reconstruction (all depth maps are already generated)?
Just had the same error on Metashape 1.6.0 build 9925 at the end of mesh generation.
This project was stored on a folder that dropbox has access to. Maybe there is a similar problem with the access rights.
-
Try to roll back to the 19.11.1 drivers if you are running the latest Adrenalin 2020 packages (19.12.1/19.12.2/20.1.1).
Mak
That appears to have fixed the issue, thank you so much!
Glad to know that this worked. As I reported earlier in this thread, the AMD Adrenalin 2020 drivers have buggy/broken OpenCL/Compute support (at least on Radeon VII GPUs).
Mak
Driver update, drivers 20.1.2 cause a system lockup during the dense cloud creation requiring a hard restart with the Radeon VII. Would not recommend them,
-
Try to roll back to the 19.11.1 drivers if you are running the latest Adrenalin 2020 packages (19.12.1/19.12.2/20.1.1).
Mak
That appears to have fixed the issue, thank you so much!
Glad to know that this worked. As I reported earlier in this thread, the AMD Adrenalin 2020 drivers have buggy/broken OpenCL/Compute support (at least on Radeon VII GPUs).
Mak
Driver update, drivers 20.1.2 cause a system lockup during the dense cloud creation requiring a hard restart with the Radeon VII. Would not recommend them,
Thanks for the heads up. I didn't bother testing them as I was fairly certain that AMD didn't fix it (as I said in another post:it often takes them more than 6 months to fix compute related issues unfortunately).
Mak
-
20.1.3 is out today
-
Hello eltama,
Do you have these errors during the depth maps generation step or already during the mesh reconstruction (all depth maps are already generated)?
Just had the same error on Metashape 1.6.0 build 9925 at the end of mesh generation.
This project was stored on a folder that dropbox has access to. Maybe there is a similar problem with the access rights.
I have this error even in build 10009:
...
2020-01-21 19:33:17 Devices performance:
2020-01-21 19:33:17 - 2% done by [GPU 1] with timings: 315 s = 1% load hists + 67% load cams + 4% calcs
2020-01-21 19:33:17 - 49% done by [GPU 2] with timings: 315 s = 0% load hists + 45% load cams + 34% calcs
2020-01-21 19:33:17 - 49% done by [GPU 3] with timings: 315 s = 0% load hists + 48% load cams + 30% calcs
2020-01-21 19:33:27 finished group #36/60: 8625 cameras, 381.065 s working (0% files opening + 28% IO + 65% unpacking + 3% queueing)
2020-01-21 19:33:27 Initing hists in leaves group #37/60: [517; 528)...
2020-01-21 19:33:41 total cubes: 108995332 cubes, 1663.14 MB, 124.888 MB/s
2020-01-21 19:34:01 Data: 2494.71 MB hists, 39.9879 MB pyramid
2020-01-21 19:34:01 Found 3 GPUs in 0.002 sec (CUDA: 0.001 sec, OpenCL: 0.001 sec)
2020-01-21 19:34:02 [GPU 1] Using device: GeForce RTX 2070 SUPER, 40 compute units, free memory: 6711/8192 MB, compute capability 7.5
2020-01-21 19:34:02 driver/runtime CUDA: 10020/8000
2020-01-21 19:34:02 max work group size 1024
2020-01-21 19:34:02 max work item sizes [1024, 1024, 64]
2020-01-21 19:34:02 [GPU 2] Using device: GeForce RTX 2070 SUPER, 40 compute units, free memory: 6711/8192 MB, compute capability 7.5
2020-01-21 19:34:02 driver/runtime CUDA: 10020/8000
2020-01-21 19:34:02 max work group size 1024
2020-01-21 19:34:02 max work item sizes [1024, 1024, 64]
2020-01-21 19:34:02 [GPU 3] Using device: GeForce RTX 2080 Ti, 68 compute units, free memory: 9158/11264 MB, compute capability 7.5
2020-01-21 19:34:02 driver/runtime CUDA: 10020/8000
2020-01-21 19:34:02 max work group size 1024
2020-01-21 19:34:02 max work item sizes [1024, 1024, 64]
2020-01-21 19:36:55 [GPU 2] Exception!
2020-01-21 19:36:55 [GPU 3] Exception!
2020-01-21 19:36:55 [GPU 1] Exception!
2020-01-21 19:36:56 Peak memory used: 24.71 GB at 2020-01-21 15:58:16
2020-01-21 19:36:56 deleting all temporary files...
2020-01-21 19:37:14 done in 18.331 s
2020-01-21 19:37:14 Finished processing in 32796.9 sec (exit code 0)
2020-01-21 19:37:14 Error: Internal error 239166172! Info: {0x0, 0x0-2406x1584, 2406x1584}
Now I try to build mesh from Dense points
-
20.1.3 is out today
Tried the new drivers, the system does not lock up anymore but the software stalls and won't continue during the dense cloud creation. Back to 19.11.1
-
Hello BjFaeTorphins,
I think that something messed up with the access rights - for some reason the created files cannot be removed or rewritten. Do you observe any similar behavior when trying to export anything from Metashape to the new file or to the file with existing filename?
Alexey
After the error happens the model.tmp directory mentioned in the error message is gone. I've not seen any permission errors and the project file can be saved after the error.
Hi Alexey
I think I've figured out what is causing these errors.
I'm using a Synology NAS to store my models. I have a fast SSD drive on my desktop computer which is continuously synchronised to a Work in Progress folder on the NAS. When I want to do work on a model i copy the files for the model to be worked on from the main folder structure on the NAS to the Work in Progress folder on the NAS which is then automatically synchronised to the SSD drive on the desktop. The software that does this synchronisation between the NAS and the desktop is called Synology Drive. I've been disabling this software when doing mesh from depth maps for about a week now and not had any more errors. Whereas, I was getting the file errors at least 50% of the time previously.
I have been using this arrangement for quite some time (since Photoscan v1.3.5) and not had the file errors before. I suspect that it is something to do with the very high number and size of the temporary files that v1.6 mesh from depth maps out-of-core creates.
I can live with this work-around but it would be more convenient if i didn't need to remember to switch off Synology Drive while running mesh from depth maps. Perhaps introducing a re-try of the file operation if i fails.
HTH
Brian.
Hi Alexey
I've changed my mind about the circumstances of these errors.
I'm now quite sure its when .psx and .files are stored on the NAS mounted shared drive. I checked and all of the errors that i've seen happen when those files are stored on the NAS and does not happen on my local SSD drive even if Synology Drive synchronisation is enabled.
HTH
Alexey
This problem still happens with 1.6.1 build 10009.
-
Hi,
I've seen in API changelog " Added Model.getActiveTexture() and Model.setActiveTexture()" methods.
I would like to switch programmatically between the different textures in order to make a capture of the textured model view for each of them.
Here's my code:
tex = Metashape.app.document.chunk.model.textures
num = 0
for i in tex:
Metashape.app.document.chunk.model.setActiveTexture(i,Metashape.Model.TextureType(i.type))
Capture = Metashape.app.captureModelView(width = 1114, height = 763, transparent = (True), hide_items = (True), source = Metashape.DataSource.ModelData, mode = Metashape.ModelViewMode.TexturedModelView)
Cap = Metashape.Image(Capture.width, Capture.height, "RGBA", "U8")
Capture.save('C:/Users/.../Desktop/num{}.png'.format(num))
num +=1
But unfortunately it dosen't change the texture type of the model...
Any ideas please?
Is there a solution?
Thank you!
-
thanks for the update, texture issue is solved !
-
Hello Brian,
Thank you for sending the sample image. It appears that the presence of the alpha-channel in the images is causing the problem with the GPU-based texture blending.
We''\ll try to implement the fix for the issue in the next version update, but meanwhile you should either disable GPU for the texture blending process or remove alpha channel from the source images.
I can confirm that this issue is fixed in v1.6.1 build 10009
-
Hello mcstieg,
Set Active Texture command is designed to switch the textures of the same type within the same model. In such way the multiple textures per model can exist (do not mix with multi-page texture atlas), similar to several instances of the dense cloud, for example, in the same chunk.
Unfortunately, there's no way to switch the display mode of the Model view at the moment or define Normal/Occlusion/Diffuse texture type in the CaptureModelView command.
-
20.1.3 is out today
Tried the new drivers, the system does not lock up anymore but the software stalls and won't continue during the dense cloud creation. Back to 19.11.1
The latest driver version 20.2.1 is also broken. Stick to 19.11.1
Mak
-
The latest driver version 20.2.1 is also broken. Stick to 19.11.1
Mak
The latest Radeon Pro SSG drivers work with my Radeon VII, and do not crash. Version 20.Q1.
-
Hi,
I like a lot of the new features in 1.6, especially the basemap (for quick sanity check on orthomosaics) and the code preview in the console.
However, it's been very frustrating having some of the parameters completely change how they work, without any sort of warning. Please keep this in mind with future API changes, for example 'mergeChunks' no longer takes chunks as input, but the chunk index, so it would be good if the function raised an Exception in 1.6 if a list of chunks was given in the first parameter (since clearly the user is thinking of 1.5 API).
The new API python manual also seems to be lacking some instructions that are critical for using the changed parameters.
-
Yes I would have to say i agree with all the changes that is really confuse me..
When I run to align the photos, there are three to choose from the list.
It would be nice if there is going to be update new version, I prefer to keep the same as previous..
just sayin'
-
Hello PROBERT1968,
Source option for the Reference preselection is identical to the Reference preselection from the version 1.5. Other reference preselection options are described in the manual (page 25 - https://www.agisoft.com/pdf/metashape-pro_1_6_en.pdf):
In the Estimated preselection mode the align takes into account the calculated values of the external orientation of the camera. That is, if you have a project with an adjustment performed, the values that were obtained will be taken into account when you run it again.
When using Sequential preselection mode the correspondence between the photos is determined according to the sequence of photos (the sequence number of the image) it is worth noting that with this adjustment, the first with the last picture will also be compared.