Segmentation of lines and arcs and its application for depth recovery

In this paper we describe an advanced segmentation approach for stereo images improving the computation of depth compared to the commonly used straight line segmentation. Using a straight line-circular arc approximation of chain coded lines, the number of primitives is reduced significantly. This approximation lowers the computational effort as well as the frequency of erroneous matches. Starting with matched pairs of primitives, a disparity image is computed containing the initial disparity values for a subsequent block matching algorithm. The output of this algorithm is the partially dense depth image of one aspect of the object. We describe the result of a parallel implementation using object-oriented programming techniques. In segmentation as well as in matching we evaluate color information to improve accuracy and reliability of the depth values. The algorithms are part of a system computing depth from monocular image sequences. Taking a sequence of different views by a camera mounted to a robot hand, each two consecutive images are considered as a stereo image. The depth images computed from these stereo images are fused to one complete depth map of the object surface. The results show substantial improvements in comparison to a monochrome system with respect to speed, accuracy, and completeness.