Research And Achievement On Binocular Vision Calibration And Stereo Matching

With the development of the robot technique and the vision technique, application of the robot is getting more and more widely, Computer vision system is an important component in a robot. Eighty percents of human messages come from eyes. Machine vision robot with the outside world as a window, get the three dimension world scenery, such as the gesture, shape, movement and depth information. Then put those messages to the computer (robot's brain), the messages are processed and treated to finish operating the robot by computer.This paper is about the research of binocular vision camera calibration and stereo matching. First, two-dimension digital image grabbing, basic principle of CCD camera and video acquisition card, hardware connection of CCD camera and video acquisition card. And software design of image grabbing are introduced.The binocular stereo vision is the method that observe the same scenery between two different angle, in order to get the pictures on the different visual angle. By the image geometry theory, calculate pixel's warp on two images, finally get the three dimensional information of the environment. Second, the article research contents are including capture of images, camera calibration, feature detection and stereo matching. The paper just finish a whole robot binocular stereo vision system according to this sequence. In the camera calibration, on the basis of some typical camera calibration methods, adopt RAC camera calibration method, stereo matching is that search the imaging position of the same target point in two different images, choosing the match character, confirm the match rule, realize the match arithmetic is the most important step in the carry out of stereo matching, in the stereo matching, research region relating and character matching algorithm.The paper used the MATLAB, VC++ to realize the arithmetic.