Research On Technologies For Calibration Of Camera And 2D Laser Rangefinder

Camera and Laser Rangefinder(LRF)are widely utilized in various applications and systems,such as Robotics,road survey and mapping vehicle,and autonomous driving,etc.LRF can obtain the target depth and 3D information and it has the advantage of measuring distance and high precision,but its obtain the LIDAR data resolution is usually low.The image data is acquired from the camera with a high resolution data,large amount of information,etc.,but can not be effective and accurate depth perception and 3D information about distant targets.Combination of two types of sensors can greatly improve the accuracy and stability of the traditional method.Extrinsic calibration is crucial and basically the first step to fuse image and laser data.The caliration of camera and 2D laser range finder are studied in this paper.In the process of two sensor calibration,the camera calibration is also very important.In this paper,the chief research works are as follows:(1)Two algorithms of camera self-calibration are presented based on active vision.These two algorithms were completed using reference camera and single camera self-calibration.Using the pure translation motion constraints and compute the infinite homography from scene depths,so that the computed infinite homography is readily decomposed for camera calibration.Experimental results demonstrate that the method is both practical and accurate.The proposed self-calibration method has great potentials to solve online camera calibration problem,which has important applications especially for multi-camera and zooming camera systems.(2)Two algorithms of the caliration of camera and 2D laser range finder are presented based on checkerboard.Firstly,A novel minimal solution based method for extrinsic calibration of a camera and a 2D LRF by introducing a virtual trihedron from 3 input chessboard planes are presented in this paper.A perspective-three-plane(P3P)problem from virtual trihedron to compute the extrinsic parameters are formulated.A novel geometric constraint are also presented in this paper,it called Restricted Area on Plane Image(RAPI),and to derive the unique real solution from up to 8 solutions to the P3 P problem,which makes the minimal solution based method more practical and flexible.The results demonstrate that the proposed method outperforms existing methods with identical input image and LIDAR data.The algorithm effectively simplifies the calibration process,while reducing the checkerboard input number true solution can be obtained accurately.Secondly,an algorithm of the caliration of camera and 2D laser range finder based on particle swarm optimization(PSO)are presented.In view of the existing algorithm of reference in a minimum of 5 input poor optimization results with the traditional method,using the advantages of particle swarm optimization algorithm for calibration algorithm.It effectively solved the least five input reference algorithm problem of poor results.(3)Present a novel folding pattern for extrinsic calibration of camera and 2D Laser Rangefinder(LRF).Using folded calibration plate point line constraint conditions to calibrate the external parameters.The results demonstrate the proposed method is both accurate and practical.It effectively solves the existing calibration problem such that complicated process,larger plate specification is not easy to carry and estimate of the initial value in the process of calibration greatly influenced the result of the last calibration and so on.