Method of Moving Object Detection from Mobile Vision System

Abstract

The task of detecting mobile objects is important in the development of video surveillance systems, optical navigation systems, etc. Currently, in connection with processes and the organization of product development, in which the most effective means are being used, technical vision systems are increasingly introduced. The most topical direction in this area is the development associated with the visual perception of moving objects. Due to the existence of the possibility of wide application of the method in the vision systems of robots, there is considerable interest in the detection of objects moving in space. In this paper, a stationary vision system is analyzed, with an object moving relative to it entering its field of view. In this case, the task of object selection is narrowed down to analyzing the sequence of images and identifying changes. However, the situation is different in the case of a dynamic vision system, when not only the movement of the object under study takes place, but the system that conducts observation also moves relative to the environment. Since dynamic vision systems can be used in mobile surveillance systems, they are of the greatest interest for study. The article investigates the method of analyzing the two-dimensional motion of an object and suggests a method for selecting objects moving in three-dimensional space, as well as a mathematical model for describing this method. For this, it is necessary to define an additional variable characterizing the depth of the points of the scene, in contrast to the method of two-dimensional motion of the object. The proposed method allows determining the spatial coordinates of dynamic objects from the moving bases of the vision systems.