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

Add more features to the COLMAP similar to automatic reconstruction #3136

Open
abrahamezzeddine opened this issue May 10, 2024 · 1 comment

Comments

@abrahamezzeddine
Copy link

abrahamezzeddine commented May 10, 2024

Is your feature request related to a problem? Please describe.
Not a problem, but it could improve the quality of the COLMAP functionality.

Describe the solution you'd like
The ability to have the same parameter suggestions as COLMAP GUI Automatic Reconstruction. For example, we are able to determine the quality of the COLMAP reconstruction (low, medium, high and extreme). It is also possible to add a vocab tree for sequential images to support lots of images. As of now, you can only trigger sequential (but no possibility to add optional vocab tree) or vocab tree (but no sequential).

Describe alternatives you've considered
Using COLMAP directly, but then it does not create the Nerfstudio data that is needed such as transforms.json.

Additional context

@ishipachev
Copy link

Using COLMAP directly, but then it does not create the Nerfstudio data that is needed such as transforms.json.

How to do it with COLMAP reconstruction done directly with COLMAP:

  • You get trasnforms.json by running ns-process-data with --skip-colmap flat and setting path to your exported colmap model with --colmap-model-path parameter. By doing so you will get your data prepared in nerfstudio format including transoforms.json file.
  • After in ns-train you can use colmap data loader, instead of default nerfstudio-data

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

No branches or pull requests

2 participants