search

kujason / avod

Code for 3D object detection for autonomous driving
MIT License
939 stars 347 forks source link

anchor generation according to plane #145

Closed yzl96 closed 4 years ago

yzl96 commented 5 years ago

I see your code: a, b, c, d = ground_plane all_x = before_sub[:, 0] all_z = before_sub[:, 1] all_y = -(a all_x + c all_z + d) / b

it looks like you just put the anchor center to the ground, but in your paper, the 3D anchor center seems to off the ground?

kujason commented 5 years ago

In the KITTI dataset, the convention of the box centroids are the bottom centre of the 3D box.