A robust algorithm for the digital stereo matching

The stereo matching is a computer vision process, able to solve the problem of the correspondences in a pair of stereoscopic images. This function consists of identifying features (generally points, lines or structures) on the left and right images which correspond to the same physical object in the scene. Its main use is the computing 3-D coordinates of objects belonging to a scene. In this paper, an algorithm for the stereo matching is shown. This algorithm is characterized by a simple and robust method of identifying a pair of homologous points of the same object in stereoscopic images, taking into account three different-size tests in opportune neighbourhood of pixels