Motion Multi-object Detection Method under Complex Environment

摘要

An optimization object detection method based on color image symmetrical frame-difference is proposed in order to solve the problem of motion multi-object detection difficult under complex environment. In this paper, firstly the color images distance is defined to calculate the frame-difference between two adjacent images. Then the before and after symmetrical image-distance of three adjacent images in difference frame interval step can be completed respectively. Secondly, an optimization binary method is designed to extract more object pixels. And the before and after object binary results of the adjacent images with the same middle (key frame) image are given respectively. At last, motion multi-object of the key frame image is achieved by the fusion result of logical AND between the before and after object binary results depending on three-frame-adjacent images with the same key frame. Actual color images from traffic surveillance system are used to test, the experimental result shows that the optimization object detection algorithm based on symmetrical frame difference in the proper step can extract motion multi-objects in different movement speed under complex environment and its accuracy and effectiveness of the proposed algorithm are verified.