search

mileyan / pseudo_lidar

(CVPR 2019) Pseudo-LiDAR from Visual Depth Estimation: Bridging the Gap in 3D Object Detection for Autonomous Driving
https://mileyan.github.io/pseudo_lidar/
MIT License
975 stars 216 forks source link

Calibration data for test data #5

Closed zsfVishnu closed 5 years ago

zsfVishnu commented 5 years ago

1) Are you using the calibration parameters for test data too, like how you used for point cloud generation? < generate_lidar > 2) Also , could you give an idea as to what the 'Tr_velo_to_cam' parameter is ? <kitti_util >

Thank You.

mileyan commented 5 years ago

Hi,

  1. Yes. We use the calibration files inside the testing folder to generate test point clouds. You can simply modify the calib_dir.

  2. This is the projection matrix from velodyne to camera coordinate.

zsfVishnu commented 5 years ago

@mileyan , thank you for your reply. I understand that the calibration file is generated for individual images of Kitti. Is it possible for an alternative to these calibration files,or a way for generating these for our custom data as well. Thank You

mileyan commented 5 years ago

Yes, you can generate the calibration files for your custom data. You can try to use this tool http://www.cvlibs.net/software/calibration/ .

Haawron commented 3 years ago

Yes, you can generate the calibration files for your custom data. You can try to use this tool http://www.cvlibs.net/software/calibration/ .

Hi, that link seems to be dead. Did you keep that tool on your local and could you share? I have searched for a while but i cannot find any breakthroughs...

EDIT: How can I get P2, R0_rect and Tr_velo_to_cam matrices?

SoulProficiency commented 6 months ago

Yes, you can generate the calibration files for your custom data. You can try to use this tool http://www.cvlibs.net/software/calibration/ .

Hi, that link seems to be dead. Did you keep that tool on your local and could you share? I have searched for a while but i cannot find any breakthroughs...

EDIT: How can I get P2, R0_rect and Tr_velo_to_cam matrices?

have you slove this problem？