Georeferencing frames extracted from a video


As some other users, I would like to be able to process a drone video with webODM. I had some success extracting frames with Sharp Frame Extractor, but it would be nice if I could georeference the project, to get a correct scale and orientation. I tried to approaches:

  • Adding GPS info to the frames using the video subtitle (using exiftool and some scripting)
  • Using the GCP Interface

None of those worked, I’m getting some errors during processing but without a clear diagnosis. I suspect that the “subtitle” solution is limited because the GPS coordinates are inaccurate in the subtitles (too few significant digits in this human-readable form).

Anyone had success with adding location info to a video-based project ?

Thanks !

1 Like

You’re on the cutting edge here! It seems like Parrot and DJI may embed geolocation information into the metadata of video streams, but accessing that data and dumping it out into the individual frames is quite a challenge.

We did read metadata like this in ODMax for GoPro Max which may be similar:

Take a look at whether the upstream libraries support parsing that metadata (likely encoded in the audio so it’s synchronized with the video).

1 Like

Thanks for this suggestion. If I understand correctly, exiftool provides the GoPro metadata reading that is used in ODMax. I tried to see what exiftool can make about GPS coordinates in DJI video, and it seems that there is only a custom config file dji.config that parses the subtitles, but no actual decoding of the audio video or data track which presumably contains more information…

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.