Could you provide the code on how to generate depth？

[Sugar55888]

Thank you!

[cwchenwang]

Hi, what kind of depth do you want to generate?

[Sugar55888]

@cwchenwang Hi, I want to generate the depth of depth completion and stereo depth estimation. Can you provide some instructions? Thank you!

[cwchenwang]

The code is provided in the depth-methods folder and you can directly use that.

[Sugar55888]

Thank you! I will check it out.

[Sugar55888]

@cwchenwang Do you have any tips for using colmap on the kitti dataset? Because I found that colmap cannot reconstruct good results on some kitti sequences. Do we need to manually select sequences with rich features?

[cwchenwang]

We don't use colmap on kitti. We just use the ground truth camera and convert it to colmap format.

[Sugar55888]

We don't use colmap on kitti. We just use the ground truth camera and convert it to colmap format.

Get it! Thank you!

[Sugar55888]

@cwchenwang Hi! Can I ask you another question? How did you convert the ground truth camera to colmap format? I still have question about getting the points3d.txt in colmap. Because points3d.txt seems to need to be obtained through feature matching of colmap.

[cwchenwang]

We didn't use points3d.txt because NeRF models don't require that.