3-D Map Reconstruction of Robot Based on Stereo Vision in Unknown Outdoor Environment

摘要

Binocular stereo vision measurement system is proposed to achieve three-dimensional reconstruction on the scene around of the robot. Two key technologies of the robot stereo vision measurement system are presented. One is to extract the suitable image features using grey correction and edge profile tracking. The other is to obtain the disparity image of measured object and to complete the image match. At first the corresponding relationship of the features above is established. And the two image map points in the two CCD image plane system of the same point in the world coordinate system are corresponding. Then the 3-D environment of robot is recovered. Two unknown indoor environment are used to experiment. The result shows that the three-dimensional profile measurement method is easy to implement, less demanding on hardware, high precision and high stability. It can be applied to the 3-D scene reconstruction on moon surface inspecting robot.