How did you calculate or where did you download depth images for kitti sequence?

halajun / VDO_SLAM

VDO-SLAM: A Visual Dynamic Object-aware SLAM System

Other

747 stars 115 forks source link

How did you calculate or where did you download depth images for kitti sequence? #4

Closed andrey1908 closed 4 years ago

andrey1908 commented 4 years ago

Hi, I want to try running your algorithm on some other kitti sequences, but I could not find where to download depth images or how to calculate them from velodyne points. Could you please tell how did you get depth images for kitti sequence?

halajun commented 4 years ago

Hi @andrey1908 . We generate the depth map in two ways depending on the input image sequence. If the input is stereo images, we calculate the depth(or disparity) map using off-the-shelf stereo depth estimation method for the KITTI sequence. If the input is monocular image, we compute the depth map using off-the-shelf learning-based single image depth estimation method. Please see the VDO-SLAM paper for the specific methods we applied if you would like to know more details. Hope this may help.