Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

mov to image sequence metadata #2329

Open
mirelloo opened this issue Feb 28, 2024 · 3 comments
Open

mov to image sequence metadata #2329

mirelloo opened this issue Feb 28, 2024 · 3 comments

Comments

@mirelloo
Copy link

hello,

I'm using the camera track + photogrammetry pipeline to solve plate's 3D cameras. it s working great but I'm running into some precision issues which might be due to some lack of metadata. I'm using an iphone 13 to both take videos of the environment and the actual plate, then convert those MOV files to image sequences, but in the FFMPG conversion I'm losing critical metadata information about camera and lens so Meshroom is left with best guesses,

I'm looking to find a solution or workaround and my questions to the group are:

  • What metadata fields are needed for meshroom to retrieve the correct camera from the DB? I can use exif tool to inject those to my image sequence.
  • is there a way I can edit/define the camera intrinsics manually inside meshroom?
  • can meshroom use MOV files ( with the correct metadata ) without the need to convert to image sequences?

thank you for your help!

Stefano

@skinkie
Copy link

skinkie commented Feb 29, 2024

For a similar problem with DJI drone video's I have modified a python script that extracts GPS location from subtitles and creates annotated exif annoated image sequences out of them. I have then used it in COLMAP, because that stuff supports GPS priors. I think the ultimate problem here is that image sequences are proven technology and any other container format (while maybe technically superiour) requires special handling, which handled for the denormalised 'individual images' case. I think you can define the intrinsics manually.

@FlachyJoe
Copy link

ffmpeg seems able to copy the metadata:
https://stackoverflow.com/questions/70261246/ffmpeg-metadata-copy-all

@mirelloo
Copy link
Author

mirelloo commented Mar 2, 2024

thank you for all the pointers.

@FlachyJoe I tryed using the -movflags use_metadata_tags in ffmpg flag but didn t work video to image for me.

@skinkie i think I can just plain override the intrinsics in the camera init node but wanted to see if adding the correct metadata I was going to fetch the correct info from the db rather than inserting those manually, and not being 100% sure if my calculations are correct.

I ll try and use exif tool to add these two flags maybe that all meshroom needs for the DB:

Make : Apple
Camera Model Name : iPhone 13 mini

will keep you posted once I know more

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

3 participants