Research And Application On Binocular Vision System Of The Mobile Robot

Vision technology of mobile robot is one of the hot topics in robot field recently. Just like humans, vision is one of the most important sources to know environment. It can supply the information of navigation for robot. This paper has studied the technology of binocular stereo vision of the mobile robot. The aim of the research is to recognize the interesting objects and provide the information of navigation from the color images captured by the two cameras on the robot.Color image is the basic information for mobile robot to get the environment around it. It is the main problem for mobile robot to recognize object from the color image rapidly and correctly. First of all, the HSI color model is selected by comparing with some popular color space models (RGB, YUV and HSI). A method of using double threshold of H and S value to segment the object from the background is used. Then, the shape of object is recognized using the method of Fourier descriptor, which improves veracity of recognition.This paper focuses on the components of binocular stereo vision: camera calibration and 3D information calculation. In camera calibration, a volume template-based method is proposed for the calibration of the intrinsic and extrinsic parameters of camera, which adopts an ideal pin-hole model. Taking the transform matrix, it works out the intrinsic and extrinsic parameters of camera. In the inertia coordinates, position relation of two cameras is worked out with stereo calibration object. Experimental results show that method needs no iteration and has a high operation speed. It needs to put the template only once for the field calibration, parameters of cameras could be calibrated. In resume of coordinates, after the camera calibration and object recognition, the coordinates of the point can easily be calculated. It is the resolution of some projection equations. This can be achieved by least square algorithm.