Cannot process dataset of 334 images taken at 45° angle

[luismgarciac]

How did you install WebODM (docker, installer, etc.)?

Docker on Ubuntu Linux 22.04 GCP VM

What's your browser and operating system? (Copy/paste the output of https://www.whatismybrowser.com/)

https://whatismybrowser.com/w/GS54EAW [Edge 106 on Windows 10]

What is the problem?

I uploaded the files and set the high quality preset, after some time of processing I get the Cannot process dataset error with the follwing information:

21:57:28 [App ] Build date: Sep 20 2022, 13:07:58 21:57:28 [App ] CPU: Intel(R) Xeon(R) CPU @ 2.20GHz (2 cores) 21:57:28 [App ] RAM: 15.62GB Physical Memory 0B Virtual Memory 21:57:28 [App ] OS: Linux 5.15.0-1019-gcp (x86_64) 21:57:28 [App ] Disk: 21:57:28 [App ] SSE & AVX compatible CPU & OS detected 21:57:28 [App ] Command line: DensifyPointCloud /var/www/data/1e409a4e-6ce4-43ca-8e59-f46116f25e29/opensfm/undistorted/openmvs/scene.mvs --resolution-level 1 --dense-config-file /var/www/data/1e409a4e-6ce4-43ca-8e59-f46116f25e29/opensfm/undistorted/openmvs/Densify.ini --min-resolution 1024 --max-resolution 2048 --max-threads 2 --number-views-fuse 2 --sub-resolution-levels 2 -w /var/www/data/1e409a4e-6ce4-43ca-8e59-f46116f25e29/opensfm/undistorted/openmvs/depthmaps -v 0 --cuda-device -2 --geometric-iters 0 21:57:28 [App ] Find a camera not pointing towards the scene center; the scene will be considered unbounded (no ROI) 21:57:39 [App ] Preparing images for dense reconstruction completed: 334 images (10s532ms) 21:57:42 [App ] Selecting images for dense reconstruction completed: 0 images (3s292ms)

Fused depth-maps 0 (100%, 0ms) 21:57:42 [App ] Densifying point-cloud completed: 0 points (13s847ms) 21:57:42 [App ] MEMORYINFO: { 21:57:42 [App ] VmPeak: 1159976 kB 21:57:42 [App ] VmSize: 1087712 kB 21:57:42 [App ] } ENDINFO [INFO] running "/code/SuperBuild/install/bin/OpenMVS/DensifyPointCloud" --filter-point-cloud -1 -i "/var/www/data/1e409a4e-6ce4-43ca-8e59-f46116f25e29/opensfm/undistorted/openmvs/scene_dense.mvs" -v 0 --cuda-device -2 --geometric-iters 0 21:57:42 [App ] Build date: Sep 20 2022, 13:07:58 21:57:42 [App ] CPU: Intel(R) Xeon(R) CPU @ 2.20GHz (2 cores) 21:57:42 [App ] RAM: 15.62GB Physical Memory 0B Virtual Memory 21:57:42 [App ] OS: Linux 5.15.0-1019-gcp (x86_64) 21:57:42 [App ] Disk: 21:57:42 [App ] SSE & AVX compatible CPU & OS detected 21:57:42 [App ] Command line: DensifyPointCloud --filter-point-cloud -1 -i /var/www/data/1e409a4e-6ce4-43ca-8e59-f46116f25e29/opensfm/undistorted/openmvs/scene_dense.mvs -v 0 --cuda-device -2 --geometric-iters 0 21:57:42 [App ] error: no valid point-cloud for the ROI estimation

===== Dumping Info for Geeks (developers need this to fix bugs) ===== Child returned 1 Traceback (most recent call last): File "/code/stages/odm_app.py", line 88, in execute self.first_stage.run() File "/code/opendm/types.py", line 378, in run self.next_stage.run(outputs) File "/code/opendm/types.py", line 378, in run self.next_stage.run(outputs) File "/code/opendm/types.py", line 378, in run self.next_stage.run(outputs) [Previous line repeated 1 more time] File "/code/opendm/types.py", line 359, in run self.process(self.args, outputs) File "/code/stages/openmvs.py", line 206, in process system.run('"%s" %s' % (context.omvs_densify_path, ' '.join(config + gpu_config + extra_config))) File "/code/opendm/system.py", line 109, in run raise SubprocessException("Child returned {}".format(retcode), retcode) opendm.system.SubprocessException: Child returned 1

===== Done, human-readable information to follow... =====

[ERROR] Uh oh! Processing stopped because of strange values in the reconstruction. This is often a sign that the input data has some issues or the software cannot deal with it. Have you followed best practices for data acquisition? See https://docs.opendronemap.org/flying/

How can we reproduce this? (What steps trigger the problem? What parameters are you using for processing? Include screenshots. If you are having issues processing a dataset, you must include a copy of your dataset uploaded on Dropbox, Google Drive or https://dronedb.app)

I used the following preset: The error I get: The dataset including the console.txt file: https://1drv.ms/u/s!AsZI_CsMoVvfj4djqaDi5uTtoxoNWQ?e=rvLNp2

[pierotofy]

Thanks for sharing the dataset!

Did you use a mission planner to capture the images? Do they have sufficient overlap? You should have at least 70% for all images.

[luismgarciac]

Thanks for sharing the dataset!

Did you use a mission planner to capture the images? Do they have sufficient overlap? You should have at least 70% for all images.

Thanks for looking into this issue, this is the 3rd time it happens, with 3 different datasets, the other datasets where taken at different times of the day, or on multiple days, I figured that was the problem, but this one was taken in a single flight.

I used Map Pilot Pro from Maps Made Easy, it was taken with at a height of 45 m and 85% overlap. Here, are the settings:

[pierotofy]

Mm, strangely this reconstructed on my Windows machine (looks nice too!):

I'm running a test on webodm.net which uses the latest version and is on Linux, we'll see if it processes there.

Note that I didn't resize the images prior to processing (WebODM will do that by default). Have you tried without resizing?

[luismgarciac]

Wonderful, I'll give it a try without resizing and send my results.

Thank you

[pierotofy]

Ok this processed on a docker machine without resizing also. So it might have to do with the resizing. https://wln.s3.webodm.net/432c8dad-46a0-41f5-a3bb-5cf085f15b25/all.zip

[luismgarciac]

It took some time but it was able to process by turning off resizing, as you said.

My result crops the point cloud, how do you keep the all the data as it is shown in your picture?

Thank you

[pierotofy]

Turn off --auto-boundary if you've selected it, or adjust --auto-boundary-distance.