search

Fyusion / LLFF

Code release for Local Light Field Fusion at SIGGRAPH 2019
https://fyusion.com/LLFF
GNU General Public License v3.0
1.54k stars 246 forks source link

How to get depth boundaries of poses_bounds.npy? #72

Open LEEJUHUI6635 opened 2 years ago

LEEJUHUI6635 commented 2 years ago

Hello. I have a problem in the process of understanding how to obtain the boundaries of depth in pose_bounds.npy file. image I thought that pts_arr is the point of world coordinates, poses[:3, 3:4, :] is the translation term of camera to world coordinates, and poses[:3, 2:3, :] is the rotation term of z values in c2w matrix. I think that the task of finding the depth boundaries is to find the z value at the camera coordinate point. To do, I think we can take the following formula and obtain the z value, but I can't understand the formula of the code. image

If you explain the formula for me, I think that it will be helpful for my research. Thank you for reading my question.

Cerf-Volant425 commented 1 year ago

sorry to bother you, I was wondering whether you find the way to obtain the boundaries of depth in pose_bounds.npy file, thanks in advance.

LEEJUHUI6635 commented 1 year ago

Sorry for late response. I just understood that through the formula in the picture attached above, the z_value of the 3D points can be obtained, and min_depth and max_depth can be obtained by scaling the min and max values of the z_value. Thank you.

Cerf-Volant425 commented 1 year ago

Thank you so much, it can be very helpful.