/MultiCamCalib

Primary LanguageC++Apache License 2.0Apache-2.0

Multi Camera Calibration

  1. 'modules/camera_calibration/app/camera_calibration.cpp' is for calculating extrinsic parameter of each individual cameras.
  2. 'modules/camera_calibration/app/HandEyeCalibration.cpp' is the implmentation of multi-camera hand-eye calibration algorithm.

Usage

  1. Pre-calibrate the intrinsic parameters of each camera with arbitary open-sourced methods, e.g. rosrun camera_calibration cameracalibrator.py ...
  2. Use modified config file (e.g parameter/AGV_calib/extrinsic_17023550.xml), change the input path, pattern size, number of corners, number of images to be extrinsically calibrated and intrinsics. Select the options of using pre-computed camera model or not, and moving cameras or calibration target.
  3. Run: ./camera_calibration settingFilePath SavePath, rename the TrajectoryByCV.txt as you wish.
  4. Formulate the captured trajectory from tracking system as timestamp tx ty tz qx qy qz qw into file TrajectoryByGT.txt: 
     double timestamp; // second
 double tx ty tz; // give the position
 double qx qy qz qw; // give the orientation in quaternion format
  5. Modify the config file (e.g parameter/multi_extrinsic_handeye.yaml). 
     int NumOfMeasures; // How many measurements will be used for calibration
 bool UseMultiCam; // Use our proposed method or existing single hand-eye/robot-world algorithms
 bool HandeyeSolver; // Type of Solver: Shah[0] Li[1]
  6. Run: ./handeye_camera_calibration settingFilePath GTFilePath_1 TrajectoryFile_1 GTFilePath_2 TrajectoryFile_2 ...

Video

Experiments Video

Datasets

For convenience we provide a zip file for testing:

 'AGV_Calib_for_test/CAM_ID/resize' are for computing the camera trajectories individually. please refer to Usage 2. 
 'AGV_Calib_for_test/CAM_ID/gt_CAM_ID' are fomulated captured trajectory from tracking system. please refer to Usage 4.
 'AGV_Calib_for_test/handeye' contains the prepared GTFiles and TrajectoryFiles, please refer to Usage 5 and 6.

They can be downloaded from the Datasets

Related Papers

[1] Yifu Wang*, Wenqing Jiang*, Kun Huang, Sören Schwertfeger, Laurent Kneip. "Accurate calibration of multi-perspective cameras from a generalization of the hand-eye constraint", 2022 IEEE International conference on robotics and automation (ICRA).

@article{wang2022accurate,
  author={Wang, Yifu and Jiang, Wenqing and Huang, Kun and Schwertfeger, S{\"o}ren and Kneip, Laurent},
  booktitle={2022 International Conference on Robotics and Automation (ICRA)}, 
  title={Accurate Calibration of Multi-Perspective Cameras from a Generalization of the Hand-Eye Constraint}, 
  year={2022},
  pages={1244-1250},
  doi={10.1109/ICRA46639.2022.9811577}}
}

If you use MultiCamCalib for your academic research, please cite our related papers.