Lsy-1997 / kitti-velo2cam

lidar to camera projection of data collected and calibrated by yourself, a GUI tool that can be used to mannually adjust extrinsic after calibration.

Lidar to camera projection

Support KITTI and ROS data

Compare to Original Code

Add comprehensable annotation and add reflectance projection.

Intro

中文博客

This is a Python implementation of how to project point cloud from Velodyne coordinates into the left color image with KITTI data set, or use data collected and calibrated by yourself.

Dependices

matplotlib == 3.1.3
numpy == 1.18.1

Also tested with

matplotlib == 3.4.3
numpy == 1.23.5

Usage

KITTI dataset

Download KITTI dataset and place proj_velo2cam.py in the root path.

python3 proj_velo2cam.py

ROS record data

You are assumed knowing how to use ROS(robot operating system), and you have record a rosbag of image and point cloud, and you also got a calibration parameter files.

Use extract_img_pc_from_rosbag.py to extract image in rosbag:

python extract_img_pc_from_rosbag.py

Use pcl_ros package to extract point cloud, and use rosbag play to play the bag:

rosbag play <your_ros.bag>
rosrun pcl_ros pointcloud_to_pcd input:=/points_raw

Quick demo

KITTI dataset

Just clone the whole repo and run proj_velo2cam.py. By default, run with frame 000007 with path below:

.
├── data_object_image_2
│   └── testing
│       ├── image_2
│       │   └── 000007.png
│       └── projection
│           └── 000007.png
├── data_object_velodyne
│   └── testing
│       └── velodyne
│           └── 000007.bin
├── proj_velo2cam.py
├── readme.md
└── testing
    └── calib
        └── 000007.txt

Projection Result

ROS data

todo: