zhijianglu / RCLC

LiDAR-Camera Calibration for non-repeative scanning Livox LiDAR

Livox-LiDAR-Camera Calibrator

In this repository, we aim to build a high-precision automatic calibration tool for Livox-LiDAR-Camera system using a printed chessboard.

Paper

For non-commercial research use. Please cite our Optics Express paper when use it, and it can be downloaded here:

  @article{RCLC,
  author = {Zhengchao Lai and Yue Wang and Shangwei Guo and Xiantong Meng and Jun Li and Wenhao Li and Shaokun Han}
  number = {10},
  pages = {16242--16263},
  publisher = {OSA},
  title = {Laser reflectance feature assisted accurate extrinsic calibration for non-repetitive scanning LiDAR and camera systems},
  volume = {30},
  month = {May},
  year = {2022},
  url = {http://opg.optica.org/oe/abstract.cfm?URI=oe-30-10-16242},
  doi = {10.1364/OE.453449}
  }

Calibration results

• Grid fitting process:

  

• Reproject results: point cloud to image ( more than 100m distance, toward pixel-wise align precision ):

  

• Reproject results: image pixel map to point clouds:

  

Requirements

• PCL (>1.7)
• Eigen3(3.3.4)
• OpenCV (>3.0)
• ceres

Usage

1. Configure data_root_path to be the data path in file config_real.yaml.

2. build project:

   mkdir build && cd build

   cmake .. && make

3. Segment the chessboard from pointcloud.

   ./BoardSegmentation

4. Start calibrate and show the results.

   ./Calibrate

Realworld Datasets

Indoor and outdoor calibration data for MID-40 and Zed2 system can be downloded at GoogleDrive

Simulated Datasets

The complete code of simulation tool has been uploaded to Livox_Cam_Simulator. Some result as shown in the following figures.

• The scan model of Livox LiDAR:

  

• The zed camera combined with Livox LiDAR:

  

• The Gazebo scene:

  

• The rviz visulation:

  

Point clouds with reflectance intensity which mapped according to the color of the materials: