search

sshaoshuai / PointRCNN

PointRCNN: 3D Object Proposal Generation and Detection from Point Cloud, CVPR 2019.
MIT License
1.71k stars 426 forks source link

Visualization #43

Open zjx99 opened 5 years ago

zjx99 commented 5 years ago

Hi, shaoshuai, I am wondering how to show the detection results with 3D bbxes and show the detected vehicles in a distinguish color. Did you use any library? Thanks.

sshaoshuai commented 5 years ago

Hi， I use mayavi for all the visualization of point cloud, which is a python library.

zjx99 commented 5 years ago

Hi, shaoshuai, You pointed out that this work takes lidar as the only input, so I am quite confused about images included in testing folder. What are they used for in testing stage? If they are not used in the testing scheme itself, then it should work when I put my own lidar data which has been transfered into bin files, is that right? Thanks

sshaoshuai commented 5 years ago

Yes, the code just uses the image_shape to crop the lidar points within the image. You could delete related codes to apply on the full areas of lidar scenes in your dataset.

zjx99 commented 5 years ago

Hi, I am still struggling with showing 3d boxes on lidar points directly.
Here is an example in the output file: Car -1 -1 -1.7723 753.6551 163.8756 814.0421 204.0525 1.5462 (height) 1.6426 (width) 3.9845 (length) 7.0502 (x) 1.2010 (y) 29.7814 (z) -1.5398 (theta) 1.6346 As (x,y,z) is in camera coordinate system, I have no idea how to transfer it to Velodyne coordinate. Is there any transfermation matrices needed? Also, why is the first nubmer behind "car" "-1" rather than float from 0 to 1 ?

Thanks

iris0329 commented 5 years ago

You can refer to visual code of frustum-pointnet .

https://github.com/sshaoshuai/PointRCNN/issues/16#issuecomment-507593480

chenxyyy commented 4 years ago

You can refer to visual code here https://github.com/kuixu/kitti_object_vis.git