Moving object detection and locating based on region shrinking algorithm

Segmentation and tracking of moving object are the important tasks in computer vision, in which the object detection and locating are the essential steps and take a great impact on the effect of segmentation and tracking. A new approach of moving object detection and locating is proposed in the Paper. In a binary difference image the region with high density of motion pixels is approached by region shrinking to implement moving object detection and locating. Under the situation of more than one object in an image power transformation is adopted to enhance objects at different position and then the region shrinking is used to detect and locate the objects. The feature rectangle of an object is derived, which is able to be used in further tracking or recognition. The method needs no prior knowledge of the number of moving object and it is robust to noise. The experiment result in the Paper shows the effectiveness of the algorithm.