A question about Velodyne Data Label

[iris0329]

Hi, @leeclemnet

Thanks for sharing such good work !
I found in intel-isl/Open3D-PointNet2-Semantic3D project, the author use your work to handle KITTI Velodyne data.
It's a segmentation work, each point has a label to show which class it belongs. But I found velodyne point data only has 4 channels (x, y, z, reflection). So do you have access to the ground truth label of each point?

Thanks for your help !

[slinshady89]

Probably this will help you
http://www.cvlibs.net/datasets/kitti/eval_object.php?obj_benchmark=3d

[xiongyiheng]

Hi, @leeclemnet

Thanks for sharing such good work !
I found in intel-isl/Open3D-PointNet2-Semantic3D project, the author use your work to handle KITTI Velodyne data.
It's a segmentation work, each point has a label to show which class it belongs. But I found velodyne point data only has 4 channels (x, y, z, reflection). So do you have access to the ground truth label of each point?

Thanks for your help !

@iris0329 Hi, have you found out how to solve the issue?