ClusterODM - Distributed Split/Merge Directory & File naming

General Help
1

Question: I’m using ClusterODM - Distributed Split/Merge for stitching large datasets.

These datasets are stored by individual flights (folder F_01, F_02, F_03, etc).
For each flight, the file names are reset. i.e. “/F_01/DSC00001.JPG”, “/F_02/DSC00001.JPG”, “/F_03/DSC00001.JPG”, etc…
So executing the cmd in PyODM, the file array will look something like this:

['/Path/F_01/DSC00001.JPG'', '/Path/F_01/DSC00001.JPG', '/Path/F_01/DSC00001.JPG', 
'/Path/F_02/DSC00001.JPG', '/Path/F_02/DSC00002.JPG', '/Path/F_02/DSC00003.JPG', 
'/Path/F_03/DSC00001.JPG', '/Path/F_03/DSC00002.JPG', '/Path/F_03/DSC00003.JPG']

Will this type of naming convention effect the process? Having identical file names even though the path is different?

Dean

2

Identical filenames currently will override each other (so processing will start, but images with duplicate filenames will cause troubles). There’s also the possibility of race conditions if parallel uploads are employed (the default) that could corrupt the files.

I’ve opened a bug report Cannot upload files with the same basename from different directories · Issue #9 · OpenDroneMap/PyODM · GitHub

Thanks for reporting it!

Contributions to fix this are welcome.

1 Like
3

@spifftek70 would you be interested in contributing to fix this?

4

Hi @pierotofy. I work with @zachaller who already submitted a fix here. :slight_smile:

2 Likes