Acquisition of 3-D information using stereo vision system by relative stereo disparity with subpixel resolution

For the acquisition of 3-dimensional information in real space, a stereo vision system is suitable. In the stereo system, the 3D real world position is derived from translation of the coordinates between the cameras and the world. Thus, to use stereo vision, there is a need to construct a precise system which provides kinematic precise translation between the camera and world coordinates, in spite of the intricacy and difficulty. Therefore, much cost and time should be spent to build the system. In this paper, facilely to solve the previous problem, a method which can easily obtain 3D information using reference objects and RSD (relative stereo disparity) is proposed. Instead of direct computation of the position with translation of coordinates, only relative stereo disparity in stereo image pair is used to find the reference depth of objects, and the real 3D position is computed with the initial condition of the reference objects. In computation, subpixel resolution is involved to find the disparity for accuracy. To find the RSD, corresponding points are calculated in subpixel resolution. The results of an experiment show that subpixel resolution is more accurate than 1 pixel resolution