Research On Camera Calibration Method Of Binocular Visionfor A Mobile Robot

At the present,mobile robot vision technology is a hot content in the field of robotics research.Giving mobile robot visual ability and making it can use "eye" to find a target, evades obstacles and so on as humans,are many scientific research worker's long-cherished wish. The robot orientation navigation by the vision dependent on the ability that machine vision extract three-dimensional information from the image plane. However, setting up reasonable camera imaging model is the basic of visual function realization. How to calculate the related parameter of model is the camera calibration technology.Camera calibration is the first step and the foundation of that the computer vision finishes converting information from 2d image information to 3d space. And the precision of camera calibration affects the computer vision follow-up work results directly. For the specific requirements of camera calibration based on mobile robot, such as high precision, good real-time performance and robust and flexible and convenient operation, this paper studies some key technology problems of the binocular stereo camera calibration. The main work as follows:1.According to point extracting work based on a camera calibration template image characteristics, basing on the Harris corner detection algorith,this paper Puts forward a pixel level and improved Harris corner detection methods.2.Basing on a deep research on the theoretical basis of camera calibration technique, this paper analyzes the advantages and disadvantages of all kinds of existing camera calibration model and algorithm and creates a better adaptability of the camera calibration model. This model fully considers the distortion parameters influence and can further improve the precision of the camera calibration.At the same time,this paper introduces the classic Zhang's plane calibration algorithm and finishes the corresponding experimental evaluation.3.In the calibration , traditional optimization algorithm is insufficient, such as sensitivity for the initial value, easiness to getting into the local advantages, difficulty to accept gathered sex and robustness and so on. After the study of intelligent optimization methods, this paper for the first time will now introduce popular quantum particle swarm algorithm into camera calibration work,and put forward a camera calibration algorithm based on intelligent optimization.When compared to the previous calibration algorithm, this calibration method is rapid and effective, and no complex set of initial value, and can get a higher standard for accuracy.4.Finally, according to the research results, this paper compiles a set of binocular stereo vision camera calibration software besed on the MATLAB GUI.The software only need to adjust the parameters that can be achieved to finish different type of camera calibration activities.Interface with all Chinese interactive thought weaves and becomes.It is easy to use for beginners. According to understand.