Space Targeting Systems Based On Computer Vision Research

In this paper, binocular stereo vision is applied to acquire coordinate of objects, which imitate the man's feeling distance to know 3D world. In this method two or more cameras are used to take pictures to the same objects from different position, then disparity of feature point is calculated to get Z coordinate. This research aims at testing the feasibility of harvesting totnatoes using stereo vision. The system is programmed with Visual C++6.O. Before the object extracted, image is preprocessed with median filter, then 2D histogram is used to segment the tomato. After the area is labeled, area features and sphericity features of the objects are extracted. By means of area match, the corresponding feature point in right image is found. The Z coordinate of the feature point can be gotten. A simple and effective camera labeling method is offered, through which the inner parameter of the camera is gotten, then the coordinate nexus between image and world is founded, and the disparity of the objects can be gotten. Large experiment data prove that the optimal distance of the two cameras is 200mm and the distance between objects and camera is 250mm 450mm. By this means, the error can be limited about 16mm. These formula were used to get (x, y, z) coordinate of the objects. X1, Xr -coordinate of feature point in left and right image; Y, -coordinate of feature point in left and right image; ?Xr isparity of the object;...