search

ram-lab / plycal

Extrinsic calibration of the camera and LiDAR via polygon plane
GNU General Public License v3.0
137 stars 24 forks source link

readme

Extrinsic Calibration of the camera and LiDAR via polygon plane

This project includes the extrinsic calibration library and gui tool for calibrating the extrinsic parameter(6-DoF rigid-body transformation) between the camera and LiDAR. This method needs convex polygon plane as calibration object and the camera intrinsic parameter(K, D) are needed (if images are not undistorted).

note: current implemention only support rectangular plane

Dependency list

Build

we test on Ubuntu 16.04, if you already install ros kinect and then you only need to compile and install ceres-solver. 

$ mkdir build& cd build
$ cmake ..                          # build library and QT GUI
$ cmake -DBUILD_PlyCal_TEST=True .. #build test tools(only for debug usage)
$ make

Test

We have tested the current qt-based tool with RSLidar-16, RSlidar-32, Rslidar-mems by rectangular plane. And we provide test data and config file under ./datadirectory which was collected with RSLidar-16 and usb webcam.

Usage

中文使用说明

Before using the gui tool:

  1. Calibrate camera and put intrinsic parameter D, K at config file or set by gui.
  2. Collect synced image(png,jpg,jpeg) and pointcloud (pcd), put them into some_place/image_orig and some_place/pointcloud directory, respectively. Name the file like 0.jpg, 1.jpg, ..., n.jpg.
    Demo

Reference

For the method you can read the paper: 

@INPROCEEDINGS{8665256, 
author={Q. {Liao} and Z. {Chen} and Y. {Liu} and Z. {Wang} and M. {Liu}}, 
booktitle={2018 IEEE International Conference on Robotics and Biomimetics (ROBIO)}, 
title={Extrinsic Calibration of Lidar and Camera with Polygon}, 
year={2018}, 
volume={}, 
number={}, 
pages={200-205}, 
keywords={calibration;cameras;feature extraction;image fusion;image sensors;micromechanical devices;optical radar;sensor fusion;stereo image processing;heterogeneous exteroceptive sensors;heterogeneous sensory systems;multisensor information;polygon board;t6/32-beam Lidar;extrinsic calibration;MEMS-Lidar;point-cloud;laser range finder;2D feature space;3D feature space;Calibration;Cameras;Laser radar;Three-dimensional displays;Sensors;Image edge detection;Two dimensional displays}, 
doi={10.1109/ROBIO.2018.8665256}, 
ISSN={}, 
month={Dec},}