Meshroom: Kinect camera pipeline

[TheBricktop]

Do Alicevision could have support for RGBD cameras for sfm ?

[jgwinner]

I've got the modern Intel D415 depth camera, which is similar to the Kinect. There are a number of RGB depth cameras out there now, so I'd be interested in this topic as well.

So giving this a bump up.

== John ==

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[stale[bot]]

This issue is closed due to inactivity. Feel free to re-open if new information is available.

[JonathanLehner]

?

[fabiencastan]

I'm sorry I didn't answer earlier. That's of course an interesting topic, but it's not on our roadmap for now. But we could be happy to support any initiative on this topic.
Re-open the issue if someone is interested to work on it.

[natowi]

Well, there is https://github.com/alicevision/AliceVision/pull/489

[fabiencastan]

Yes, that's a first step in this direction: getting the kinect depth maps and the SfM in the same coordinate system (to be able to combine them in the dense part of the pipeline).

[jgwinner]

Just jumping in: I hope "Kinect" means "RGBD in general" as the Kinect is of course pretty obsolete and hard to find. The D415 and others are easily available.

    == John ==

[fabiencastan]

@jgwinner @TheBricktop @JonathanLehner

Is there a standard file format supported by all RGBD sensors with a standard way to retrieve focal length and principal point (to be able to convert the depth values into 3D points)?
Do you have some sample datasets to share?

What do you expect from RGBD sensors? Get scale information in SfM and improve robustness to weakly textured surfaces?

[natowi]

There does not seem to be one standard format, but there are libraries for depth cameras like https://github.com/AmmarkoV/RGBDAcquisition