search

TRAILab / CaDDN

Categorical Depth Distribution Network for Monocular 3D Object Detection (CVPR 2021 Oral)
Apache License 2.0
359 stars 62 forks source link

Question about the grid_to_lidar_unproject function #61

Closed taylover-pei closed 2 years ago

taylover-pei commented 2 years ago

I have read your paper and it is really a great job.

Since I'm a new one in this field, I have a question about the code.

In the below code, what does the ' Calculate grid to LiDAR unprojection for each plane' actually means? In my opinion, x_size/y_size/z_size means one point in the grid represents x_size meters in the lidar coordinates, is that right? So what's the relationship between x_size/y_size/z_size and the unproject tensor? https://github.com/TRAILab/CaDDN/blob/5a96b37f16b3c29dd2509507b1cdfdff5d53c558/pcdet/models/backbones_3d/f2v/frustum_grid_generator.py#L45

Thank you very much. Looking forward to your reply!

codyreading commented 2 years ago

Yup thats correct, this transformation converts coordinates in the voxel grid (indices that range from 0 to the maximum shape of the grid) to real world coordinates (in the LiDAR frame in metres). The unproject tensor represents that transformation, as it scales multiplies the voxel coordinates by the voxel size and adds the origin of the coordinate frame (pc_min).