Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - Steve003

Pages: 1 ... 8 9 [10] 11 12
Hi Alexey,

1) Build Dense Cloud, depth maps quality is defined in the corresponding field ("Quality");
As there are a choice of quality: ultra high, high, medium, low and lowest.
I would like to think that as depth map is the 'new toy' we can play with for better results, ultrahigh will make an ultrahigh depth map and make sure the mesh is great,
ultrahigh dense point cloud (lets call dense point cloud dpc) took 6.5 hours  in photoscan 1.4.5  :o
mesh ultrahigh using 'high' dpc took 6hrs in Metashape  :o

I am frightened to create ultrahigh depth map in dpc 6.5hr then retain its quality with ultrahigh in mesh, a further 6 hr process !  :o :o

or its ultrahigh in dpc and high in mesh 7hrs or high in dpc and ultrahigh in mesh, also 7hrs.

but the former one is taking ultrahigh depth map and despoiling it down to 'high'.

2) Build Mesh (only if both Arbitrary surface type and Depth Maps source options are selected),
We need to alter ‘and’ to ‘or’ because….
In build mesh, which requires that a dense point cloud has been created before being chose, there are two options
a) source: Dense cloud and this then shows arbitrary3D  in the Surface type box
b) source: depth map in which case surface type and the choice of arbitrary is denied, being ‘greyed out’
so depth map and arbitrary are not together possible.

3) Build Tiled Model (only if the Depth Maps source option is selected), depth maps quality is defined in the corresponding field ("Quality").
I cant find a setting to build a tiled model and am unfamiliar with what that is used for.

Sparse cloud selected as a source for the mesh generation is only reasonable if you need to get a rough overview of the reconstruction in short time.

I have tried sparse a couple of times but the result from mesh choosing source sparse sees a model so rough and incomplete its totally unusable for editing the background from, building a mesh based on sparse I just can’t see a use for myself.



Hi Alexei, sending files and project via wetransfer.

Mild filtering is default for the depth maps generation in Metashape and it is the recommended option for the new mesh generation source.

use of :-
sparse cloud

 in dense point cloud first use, saw mesh create a mess with half the wheel missing at inside face.

but recommended workflow suggests depth maps to be generated (either for the dense cloud ro directly for mesh) in order to get detailed processing result.

How does one choose to generate depth maps, I dont see an option in align photos or next stage dense point cloud to tell it to make depth maps. Nothing in manual pdf either.

old mesh (from the dense cloud) generation operation includes the following steps: Build Depth Maps (Time1), Build Dense Cloud (Time2), Build Mesh (Time3).
new mesh (from the depth maps) includes the following: Build Depth Maps (Time1), Build Mesh (Time4).

again same question these steps refer to choosing to make depth map but how ?

Not sure what you mean by Time 1 etc ?


General / speed of depth map I thought it was faster than point cloud
« on: July 05, 2019, 03:00:30 AM »
Metashape 1.5.3
seeing the advert of it being finer detail and faster, 9hrs v 19hr on 730 images with use of depth map,

small wheel and axle of trolley.

Photoscan 1.4.5
create dense point cloud ultrahigh setting aggressive 6.5 hrs
mesh high setting 1 hr

Metashape 1.5.3
create mesh use depth map quality ultrahigh  reprocessing done 2hrs, estimating surface 4 hrs .  so 6hrs  :(when 1.4.5 using cloud 1 hr  :). 6hrs when need pc for other intensive cad work is an issue.

Whilst I will be happy if its a higher quality the 600% more time needed on meshing doesnt tally with the 9hrs  vs 19 hrs on 730 images, I only have 90 images. cuda enabled, 32Gb ddr vengeance pro ram, Intel Core i7 4790 3.6Ghz dual core:-

What settings should I be using to get a faster result than in photoscan 1.4.5 ?

I see also reuse depth in both point cloud and mesh options but unsure of if we tick that or not.

I also see mention of use sparse cloud then select sparse cloud in mesh but so far found sparse to be rough quality so not sure if I will get a super fine detail mesh from such.

Should metashape be starting off with users selecting sparse and moderate on a new project ?


I can report back that after trying further steps as it seems KIS has a part to play in this, I looked at all ways possible of excluding Agisoft from the attentions of KIS !
I found two places to add the agisoft folder, so ignoring anything agisoft ! just incase adding just the .exe wasnt enough. This must be affecting anyone with KIS, I have the latest version
I then launched Metashape AND IT RAN !!!!  :) :) :) :)
see attached image of settings in KIS

I ask you alter the instruction in the install section of the pdf manual for 1.5, see attached....give steps to adding agisoft folder to all possible exclusion settings in KIS.
 Contact KIS  there might be other settings I havent found yet.
Anyone with raid setup would have had the hot power offs I had to do, wreck the pc requiring hours of setting up raid config again. I know as I used to run raid and hot power off was a death blow.
I ask that Agisoft liaise with KIS and other virus progs on this., as turning off check for updates simply isn't enough. I have never had such lockup oof a PC with other progs, even ctrl alt del showed nothing.


Hi, I must get this to run as 1.4.5 on ultrahigh densepoint cloud took all night to make a cloud, and on high mesh was able to show bolt thread and tyre number.
ran it again on high and high and that detail was lost. I need something better than high but not that takes 7 hours for dense point cloud.
looking at promo pics metashape has more definition than photoscan., I must get it to run having paid for it.

 I back up PC again, then add latest windows updates, then install latest GTX970 driver.

1) I unplug internet and exit Kaspersky internet security (KIS).
2) install metashape,
3) reboot pc and  turn off KIS again.
4) launch metashape, it runs !  :)
5) close metashape, turn on KIS,
6) launch metashape, black dos window, so with KIS on it fails, PC locked up, try all ways, get total black screen, hot power off PC  :'(
7) reboot PC turn off KIS, launch metashape but this time despite KIS being off, I get just that empty dos screen, WHY IS THAT ???  failure gets remembered and applied thereafter to every launch, PC locked up again  :'( get jet black screen exiting metashape via taskbar.
8) powerbutton off PC, reboot PC, uninstall metashape std, reboot pc , turn off KIS, launch photoscan 1.4.5 oh heck, now a black dos box for that as well. unable to shutdown PC, power button it off. what the heck is this agisoft prog doing to me.
9) reboot, install metashape again and reboot pc.
10) turn off KIS, so we are now at step 3) so it should run, launch metashape BLACK DOS BOX PC FROZEN AGAIN.  so as said the pc somehow now has remembered it did this and does it again. The only way to remove that memory is to revert PC to a backup made just prior to all this.

This just should not be !!!!!!!  :'( :'( :'( :'(

after another reboot I can get 1.4.5 to run...phew !!!

I see your text about black dos screen, to try running application in ANGLE compatibility mode (metashape.exe --opengl angle) I right click the exe and go to compatability tab but see no ANGLE option.  I alter name of exe to what you have written but of course it then asks to be opened by  a program offering photoshop and others.
What in more detail do I have to do with ANGLE compatibility mode (metashape.exe --opengl angle)

If I can get metashape to run I can turn off check of updates but I am in a catch 22 situation I need it to run to do that which will then run it !

no option but to revert PC back to prior install.
Install metashape, unplug internet turn off KIS Metashape LAUNCHES. now I can turn off check for updates, it launches again after shutting metashape down.
reboot pc, KIS is on, launch metashape, it fails and locks up PC, reboot and turn KIS off, launch metashape, it still fails, locks up pc.

so that setting made no difference and now KIS ON or OFF makes no difference.

why is it that metashape wont run and photoscan will ? In all my 30 yrs of being with PCs I have never had this issue. Its doing something that its predecessors normally dont do, that photoscan also had a moment there is worrying.  I run many progs but never had this.


Posted 24june, still floundering, hope for some guidance here please, I now have a tyre wall with fine raised number on it formed with a 1mm high raised edge round the numbers.
mesh is 12,566,203 and if I choose decimate, surely that will be 1,256,620    10% of the 12,566,203 but the number appearing in the box when choosing decimate is 300,000 and I have seen offered by default 200,000  I recall trying for 300,000 incase 200,000 was too low. so last value used gets offered again.
Just what number should we be using to retain the 1mm raised beaded edge to tyre wall numbers for example, or the thread on a 12mm diameter bolt on the vintage tyre hub ?
do we have to look at the face count and divide by 10 then use our number ?

I try for 10% and the result is bad. see attached image.

manual says decimate is for pdf and google earth so reading about now decimate mesh in youtube tutorials clearly is wrong for a requirement to get this into a CAD prog to reverse engineer.
However 12.5k will hardly rotate or zoom in photoscan on a dedicated CAD pc.

what percent is advisable to retain detail for export to CAD and surface/curve work such as Rhino3D ?

Also I have chosen high for mesh generation stage after choosing high for dense point cloud, mesh just reached 100% generation and now its decimating it, I dont recall that happening a few projects ago, is that normal ?


General / Re: Just what is the purpose of the region ?
« on: July 01, 2019, 11:48:21 PM »
Hi Alexey,
understood, so to get it nicely positioned to the object is a good thing.

I have turned on grid, then rotated object so its baseboard it sits on is at grid level and squared up to grid.
 I am using numpad 7 ortho top, 1 and 3 for front and side views to do so.
Now to adjust the region box to fit near to object.  However the region box ignores the fact that I am in orthographic view, where it should just be a singular box shape of 4 sides, and persists in having 6 sides making sizing it very difficult, I cant tell if the line for the box edge is the one nearest me or the one furthest away.
should not the region box also show itself as ortho in ortho view ? and only as perspective when in perspective view ?



General / Just what is the purpose of the region ?
« on: July 01, 2019, 07:31:34 PM »
 I cant see mention of what the region does (its purpose) in the manual, I see how to move and resize and rotate it, but what is it for exactly ?
I would like to think it tells agisoft where to focus its attention for processing.
I see mention of it indicating orientation, not sure though if I experience that !
Could it not be used to tell agisoft to ignore all points outside of it when creating meshes ?
i.e. another version of masking ?


just now I posted images saying I still cant see it, when I look again I can.



nice article but he doesnt seem to cover my scenario of object not on turntable with me moving around the object then moving it and repeating the walkaround.

 I am writing up a step by step for different shooting scenarios, which I will then post here, hopefully it can be made a sticky so its always seen at top of forum, as other forums do. I have sent it to you for approval but I am seeking clarification on method 1 now, see your emails.

Initially focusing on METHOD D for the above situation.  I think it also needs altering  as using your method 2 it mentioned the requirement for a shot of the background without the object, that's not what was done or was possible in my sessions.

I have tried to Incorporate what David Cockey in another thread wrote.

I frequently use Mask from Model in situations where the background is moving relative to the object by a small amount, such as a boat tied to a dock or an object which was bumped while being photographed. I align photos, create a dense point cloud and build an initial mesh without masking. Masks are then created using the initial mesh. The masked photos are re-aligned. The masks are then removed and a new dense point cloud is created and mesh built.

So I now have my Method D as follows, but I am unsure of step 10 onwards.

Photos taken moving around  the object (object not on turntable) . Background may move during the overall photoshoot.
For object deliberately moved after one photo sequence, e.g to photo its underside, or items in background change.
1st sequence object upright, 2nd sequence object placed on its side, 3rd sequence object placed upside down.

1 ) Create chunk per sequence
2) Align Photos in each chunk.
3) reduce the bounding box  in each chunk  to fit just your model. (use Model>Show/Hide Items>show region as well as region resize tools)
4) create rough model (from the lowest quality dense cloud, for example) for the limited area of interest ,
5) create a mesh.
6) roughly cut unwanted polygons that are not related to the object.
7) Then use Import Masks -> From Model option. This operation would apply mask to every aligned camera according to the rough mesh model.
8 Repeat Import Masks -> From Model option for each chunk,
9) re align the masked photos
10) remove the masks (HOW IS THAT DONE ?)
11) then merge chunks
12) then re-align the complete image set from scratch using Apply Masks to Key Points option.
13)  After that build dense cloud or mesh.

I did have...
10) then merge chunks
11) then re-align the complete image set from scratch using Apply Masks to Key Points option.
12)  After that build dense cloud or mesh.

Can someone correct these steps for me.

and tell me what to add to step 10 as I am not sure how masks are removed or even why !

I have gone for separate chunks as when I put all the pics into one chunk I ended up with two models on the same gridded baseboard, the right way up and the upside down one and couldnt then place a region around it as it was not one model. If the background had been masked then that wouldnt have happened, but this method is about making the mask, a catch22 situation !


Its taken me ages to mask up two photos of those in chunk 1. No way could I ever mask up my 2000 this  way.

I have imported the two masks for chunk 1 using the correct {name}.bmp approach, I see the two masks in mask view, I then run align photos selecting the settings as per the blog on this, i.e accuracy high generic preselection tick key point 40,000  tie point limit 10,000 apply masks to tie points adaptive camera tick and hit ok.

How do I know this has worked, I still have just 2 of 15 photos showing a mask in mask view.  :'(


Hi Alexey,
attached a screen grab of my 1.4.5 magic wand area, unlike the video (also attached) it has no options there.

I would at the moment prefer photoshop but not sure how I bring the mask in and use it for 'apply masks to tie points'.

my object got placed in a different position 7 times so I presume I need 7 masks one per chunk, and use the tie points method 7 times ?


General / Re: Intelligent scissors are not Intelligent
« on: June 24, 2019, 10:25:00 PM »
Hi Alexey,
I shall send via wetransfer a video showing how despite many crtl clicks, the scissors get distracted from the edge on a gridded board. I wish that video was in English !

I need instead to create a mask using photoshop and then use that in 'apply masks to tie points' though I dont know how to bring the mask in for the tie point command to use it.

How is that done ?

Mak11, I am simply separating out problems I need help on, if they were all stuffed into one thread the initial post would be too lengthy people dont read what looks long, also the title would simply be help needed with this lot, and would not feature keywords to be useful for anyone searching on a similar problem they are having and they wouldnt want to try and find an answer in a thread full of various answers and problems. Also to keep adding new problems into the initial thread would not see folk realise it had new content so it would be overlooked. I wish all was well and everything worked but instead it isnt and I need to find out whats wrong, what to do etc and it is best such are separated out, as is normal practice on forums. I normally get told keep the problems as separate posts to help others navigate and search on such !

I have encountered problems I cant find answers to on youtube or in existing forum threads, only then do I post, and need to get operational now, rather than post one a week as if rationed and take 2 months to get operational.


General / Re: Automatic mask in PhotoShop
« on: June 24, 2019, 09:55:10 PM »
Hi David,
Is there not going to be a catch22 situation, unable to use the model to chop away accurately at the baseboard where the model touches it, due to a snowdrift effect going on, because the photos needed masking prior to use to deter Agisoft from creating a poor definition at that object/baseboard junction ?

in other words or maybe saying it just as badly, to get a model  with well defined object/baseboard junction so as to edit mesh away and create the masks, one needs well masked photos and to use agisoft to mask all the photos in one fell swoop with this method we need the model already created from well masked photos.

I have several chunks, each one where the object was repositioned on the gridded table beyond which is a concrete floor and buildings, flor and buildings which appear occasionally in some of the shots, gridded table in all.

I am wondering how I remove all those backgrounds as fast as possible.

If I generate model first, surely the existence of the backgrounds will cause inaccuracies, poor edges where it touches grid etc, but to do as your method and delete background away from model  which I also saw on youtube and liked the idea of, I have to create an accurate model, so catch22.


General / photo masking tools none work where is tolerance options ?
« on: June 24, 2019, 09:29:34 PM »

I am having terrible experience trying to mask an image, intelligent scissors get sidetracked by lines on my grid table. it hugged the object then when it came to a grid box line it lost the edge trace and went round my inch box on the grid, then eventually I ended up with one straight red dashed line after having gone a few inches.
tried paintbrush but got lots of peculiar bits.
tried magic wand but not any better.  he has standard prog and uses tolerance values in magic wand options but I have Photoscan 1.4.5 and those options are missing.

I wish to follow the apply masks to tie points method, can I use photoshop instead to mask the image or two required and how do I bring this in and use it in the tie points process ? I could have masked 20 in the time it took to not even manage to mask using the agisoft tools.


Pages: 1 ... 8 9 [10] 11 12