An Efficient Algorithm for Calculating Optical Flow Parameters in Computer Vision Systems

The solution to the problem of constructing the velocity field of a moving object in computer (machine) vision systems using monocular methods for determining the proper motion of a video camera is considered. The estimation of optical flow parameters is obtained on the basis of minimization of the integrand expression of the energy functional (by Horn-Schunck) with the introduction of a regularizing component by Tikhonov. Unlike the traditional procedure for minimizing the energy functional, which requires information about the optical flow values for the entire scanning area and a large amount of computational costs, the proposed method requires optical flow values determined only at a specific point in the image area. The solution is obtained in a simple analytical form, easily implemented from a computational point of view in on-board computers. A numerical experiment was carried out to illustrate the possibility of effective practical application of the proposed approach.