Closed Bavar2142 closed 8 years ago
That is usually the case.
How much overlap are these photos?
About 20 to 30%. That was just a trial run though and was a part of a single track. V2 as it were will be at a lower altitude in a grid.
Hi I am having the same issue I have tried 1) Two very similar images 2) The same image twice Same error images attached
`
[INFO] Initializing OpenDroneMap app - Thu Jul 14 09:10:44 2016
[INFO] Running ODM Load Dataset Cell
[DEBUG] Loading dataset from: /home/ubuntu/TEMP/images
[WARNING] Could not find ccd_width in file. Use --force-ccd or edit the sensor_data.json file to manually input ccd width
[DEBUG] Loaded G0021343.JPG | camera: gopro hero4 silver | dimensions: 4000 x 3000 | focal: 3.0 | ccd: None
[WARNING] Could not find ccd_width in file. Use --force-ccd or edit the sensor_data.json file to manually input ccd width
[DEBUG] Loaded G0021342.JPG | camera: gopro hero4 silver | dimensions: 4000 x 3000 | focal: 3.0 | ccd: None
[INFO] Found 2 usable images
[INFO] Running ODM Load Dataset Cell - Finished
[INFO] Running ODM Resize Cell
[DEBUG] Resizing dataset to: /home/ubuntu/TEMP/images_resize
[DEBUG] Resized G0021343.JPG | dimensions: (768, 1024, 3)
[DEBUG] Resized G0021342.JPG | dimensions: (768, 1024, 3)
[INFO] Resized 2 images
[INFO] Running ODM Resize Cell - Finished
[INFO] Running ODM OpenSfM Cell
[DEBUG] running PYTHONPATH=/home/ubuntu/git/OpenDroneMap/SuperBuild/install/lib/python2.7/dist-packages /home/ubuntu/git/OpenDroneMap/SuperBuild/src/opensfm/bin/run_all /home/ubuntu/TEMP/opensfm
2016-07-14 09:10:45,481 Extracting focal lengths for image G0021343.JPG
2016-07-14 09:10:45,500 Extracting focal lengths for image G0021342.JPG
INFO:main:Extracting ROOT_HAHOG features for image G0021343.JPG
INFO:main:Extracting ROOT_HAHOG features for image G0021342.JPG
DEBUG:opensfm.features:Found 4862 points in 1.44987201691s
DEBUG:opensfm.features:Found 4862 points in 1.45249009132s
Matching 1 image pairs
Matching G0021343.JPG - 1 / 2
Matching G0021342.JPG - 2 / 2
G0021342.JPG - G0021343.JPG has 4862 candidate matches
Robust matching time : 0.0281529426575s
Full matching 4725 / 4862, time: 1.30093002319s
2016-07-14 09:10:49,584 reading features
2016-07-14 09:10:49,608 Merging features onto tracks
2016-07-14 09:10:49,653 Good tracks: 4725
images 2
nonfisheye images 2
0 partial reconstructions in total.
Traceback (most recent call last):
File "/home/ubuntu/git/OpenDroneMap/SuperBuild/src/opensfm/bin/mesh", line 23, in
An error occurred. Check stdout above or the logs.
I don't know what the absolute minimum number of images you can have to run ODM (at the recommended 60-75% overlap) but 2 images is too few. Even if you get a reconstruction in OpenSfM with 2 images, you would have to change
--pmvs-minImageNum <positive integer>
Each 3D point must be visible in at least minImageNum
images for being reconstructed. 3 is suggested in
general.
to 2 to generate a dense point cloud.
My minumum recommendation is 50 images, given 60% overlap minimum. You may still run into problems though.
And you really need to take different images. ;)
The rerun worked beautifully with 100 is images.
Hello all, ran a small roughly 20 image dataset and had the following error. Was it because the data set was too small?
Traceback (most recent call last): File "/home/c/OpenDroneMap/src/OpenSfM/bin/mesh", line 23, in
reconstructions = data.load_reconstruction()
File "/home/c/OpenDroneMap/src/OpenSfM/opensfm/dataset.py", line 291, in load_reconstruction
with open(self.__reconstruction_file(filename)) as fin:
IOError: [Errno 2] No such file or directory: 'opensfm/reconstruction.json'
b
quitting cause: PYTHONPATH=/home/c/OpenDroneMap/lib/python2.7/dist-packages "/home/c/OpenDroneMap/src/OpenSfM/bin/run_all" opensfm returned with code 256.