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 - 2bForgotten

Pages: [1] 2
Hello 2bForgotten,

Looks like you are observing the similar problem as reported before:

Hi Alex,

Thank you so much for the fast response. I found it while you must have been typing  ;)


Hi @Alex Pasumansky,

When trying to refine a mesh, I keep running into this error: Kernel failed: invalid device pointer (17) at line 99.

As I suspected it could have to do with the Nvidia drivers for the two RTX 2080 cards in the workstation, I tried both
- Studio driver v472.12
- Game ready driver v496.13
Both to no avail.

Also tried to run with only 1 RTX 2080. All in vain.

Machine is Windows 10, dual Xeon E5-2678 v3 @2.50 GHz, 128 GB RAM.
Agisoft version is 1.7.5 build 13229 64bit

Last lines in the log file:
2021-10-22 19:25:02 Stage #2 out of 2
2021-10-22 19:25:03 Faces: 4832973, Vertices: 2418981
2021-10-22 19:25:03 Memory required on each device: 57 Mb + 1013 Mb = 1071 Mb
2021-10-22 19:25:03 Subdividing mesh...
2021-10-22 19:25:55 Faces: 15424940, Vertices: 7715781
2021-10-22 19:25:55 Memory required on each device: 258 Mb + 3236 Mb = 3494 Mb
2021-10-22 19:25:56 Using device 'NVIDIA GeForce RTX 2080' in concurrent. (1 times)
2021-10-22 19:25:56 Loading photos...
2021-10-22 19:25:57 Finished processing in 295.559 sec (exit code 0)
2021-10-22 19:25:57 Error: Kernel failed: invalid device pointer (17) at line 99

Please inform me how to solve this error.

Although I searched this forum before posting, I must have overlooked this post that solved the issue by switching to OpenGL.

Kind regards,

General / Re: Your opinion on USGS Agisoft Processing Workflow
« on: March 24, 2020, 01:00:26 PM »
Hello everyone,

I'm new to the forum and also at work on Metashape. I have some experience with photogrammetry - a few years of work on Z / I Imaging. Now I start with Agisoft .... But back to the topic, I practiced the USGS workflow on several data sets - 1.5 ha (75%, 75%) 100 photos, 12 ha (80%, 80%) 530 photos and 12 ha (60%, 40%) 370 photos grid- the same area, GCP measured with accuracy xyz 1cm, pixel 2cm, drone P4RTK. Alignment carried out to obtain errors assumed by USGS, camera parameters - all marked without additional. I recommend this workflow, very good results obtained. At 33 checkpoints, errors no larger than 2cm. The dense cloud finally generates quickly, about 40 minutes on high setting. Dense cloud height accuracy at +/- 2 cm checkpoints. I really recommend.
PS. I want to introduce some modifications of the workflow in order to obtain more accurate image matching from about 0.2 pix


Hi there,

Thanks for sharing!

Some year ago I gave it a try myself. Did get some mixed results on a project with sandy dunes and low vegetation.

But the questions raised in OP regarding the workflow with the gradual selection tool are still open to debate. Have you experimented with that too?


General / Re: Thermal image Ortho with temperature
« on: October 04, 2019, 10:58:36 PM »
Hi there,
There sadly hasn't been any answer to the question raised in OP on workflow / documentation for processing thermal images.

General / Re: Baffling camera alignment PhotoScan 1.2.3 vs MetaShape 1.5.4
« on: September 23, 2019, 04:15:01 PM »
Hello 2bForgotten,

Project files with the alignment results only would be fine already.  At least for now.

Hi Alexey,

It's been done through WeTransfer with link to this thread.


General / Re: Baffling camera alignment PhotoScan 1.2.3 vs MetaShape 1.5.4
« on: September 23, 2019, 10:08:26 AM »
Hello 2bForgotten,

Can you please share the projects processed in 1.2.3 and 1.5.4 with, so that we can study them?

Hi Alexey,

You mean sharing the whole photo set (3Gb)?
Given my poor internet connection, that would take me ages.   :-\
Or do you mean some specific project files?


General / Baffling camera alignment PhotoScan 1.2.3 vs MetaShape 1.5.4
« on: September 20, 2019, 04:19:00 PM »
Hi there,

Because I have some trouble getting camera's well aligned in a project of around 300 camera's (sandy dunes with sparse vegetation as main subject) I am experimenting around to get the best results.

So I looked back in the archives to find PhotoScan 1.2.3 and installed it back.
I loaded the photo's (which come without GPS info) and got really nice alignment (40'000 - 4'000). Min/max error 0.184/0.677.
Project was then saved.

So I decided out of curiosity to just open the saved project in MetaShape 1.5.4.
To my utter suprise the camera errors reported now are terrible: min/max error 0.797/2.722

That is without having done anything to the saved project with just the camera's aligned. I cannot believe my eyes!

Does anyone know what's going on here?
I do know about the reports from the community about poor alignment performance in MS, that's what made me revert to PS in the first place.

Looking forward to your opinions,


General / Re: combining infrared and rgb camera
« on: June 28, 2019, 08:19:55 PM »
Hi Alexey,

Thank you for your answers.

One final question though about the combined RGB and IR camera.
Given that they will fire at the same time in the exact same direction. There still would be a different field of view for the RGB compared to the IR.
Would Metashape be able to handle that? Or would one have to create two different workflows for both image sets?

I read around on the internet, but haven't found a conclusive answer to that question,


The issue looks to be related to the alignment problems, like if there are several subsets of images aligned independently within the same project.

Hi Alexey,

No alignment problems in my cameras and no subsets. Just one set and all camera's did align.
Someone on Facebook suggested alignment problems too, so I added extra markers in all respective foto's but to no avail.

I would be happy is there would be a solution to this phenomenon though.


General / Re: Is this GCP Workflow correct?
« on: June 25, 2019, 01:09:05 PM »
Disregard... I figured it out.

Hi jazzyj,

What a pitty you didn't leave your workflow for others like me to learn from, but instead deleted it :(


General / Re: combining infrared and rgb camera
« on: June 25, 2019, 12:08:40 PM »
Hello 2bForgotten,

Since the time the topic has been started in Metashape Pro the multi-camera system approach has been implemented, that allows to process the data captured from multiple sensors that are fixed relative to each other. It could be both RGB cameras looking in different directions and different range sensors (RGB + NIR, for example).

The main requirements for such approach are: properly synchronized sensors and absolutely fixed relative orientation of the sensors in such multi-camera system rig.

Hi Alexey,

Could you please tell me if when using IR camera, the radiometric info will still be available in the ortho when created with Metashape?


Hi there,

Posted my question on the same subject (or problem ;)) here

No really satisfying answers up till now  :-\

On Facebook I posted the same question and there someone suggested to use MeshLab to manually edit and flatten the mesh before creating DEM.
I haven't tried it yet.


General / Re: combining infrared and rgb camera
« on: June 19, 2019, 04:24:18 PM »
great, i am also interested in this topic.

I am also trying to combine my RGB chunk with my NIR chunk.

I've so far build the spare cloud for both chunk separately and then I merge the chunk according to the points.. it is still processing and I'll update if it works.

I hope to hear from you guys who are trying as well.


Hi there,
Would like to know about your experience so far. I suppose the processing has finished yet. ;)

You'd better compare the original image with the orthomosaic, instead of the vertical view of the point cloud.
José Gonçalves

Hi there,

Yes it seems that the lack of contrast might be the culprit.
Is there a way to accomodate for it so that the DEM I want to create won't show those disturbances that I now see in the dense point cloud?
I now have included a transection of the area so you can see what my problem is.


Hi there,

Quite novice to Metashape and trying some features I ran into this problem:

I have a small project with sandy structure. Pictures are taken with drone and quite good quality camera (Sony Alpha) (70% and 80% overlap).

When I try to build a dense cloud, a rather unstructured/distorted surface emerges from pictures that show more or less flat sandy surface. The distortions are both above and under the actual surface.
The DEM will be equally rough when generated, whereas I know and see on the photo's that the surface should be nice and flat.

I tried cleaning and shaving the sparse cloud and camera optimizations but it doesn't seem to help.

Is there something I am missing or is the standard workflow (from manual. I tried both aggressive and mild depth filtering) inappropriate? Or is there a way to edit or accomodate for the phenomenon?

image_1 = dense cloud
image_2 = one original photo


Pages: [1] 2