Target Tracking Based On Color Video Image

With the development of computer technology and digital signal processing theory, moving target tracking technology has become one of the most important research topics in the fields of image processing, computer vision and security monitoring. The application of visual moving target tracking technologies are of great significance in military, industrial, security surveillance, intelligent transportation, and other fields.Nowadays, most moving target tracking methods are still based on the gray-scale image. However, it is sensitive to illumination changes, object occlusion and complex background, which brings certain difficulties for gray-scale moving target tracking. Compared with the gray-scale images, color images can carry much more visual information, and have richer information stages. Target tracking based on color video image will further enrich the contents of the tracking and improve the robustness and accuracy of tracking.In recent years, with the expanding applications of the color image, a large number of researchers are committed to more in-depth study of the color image processing. This paper carried out researches on the target tracking system based on color video image under this background. The main research work and innovations in this paper are as follows:1. Research on the related theories and key technologies for color video image target tracking. Summary and analyze on the basis of the existing color video image processing technology, combined the color moving object detection, color object segmentation and color target tracking together to build a complete target.tracking system based on color video images. So as to better solve the existence problem of low tracking efficiency and poor robustness in color target tracking system.2. In the color moving target detection aspect, a novel detection algorithm based on background subtraction and frame difference is proposed, which makes full use of the color information and is insensitive to the light changes. The proposed algorithm directly extract target which can effectively preserve the color information; involve background subtraction method and frame difference method to improve the accuracy of target detection, ensuring the robustness of detection, and prepare for the follow-up procedure of target segmentation and target tracking. 3. As to color image segmentation, proposed an adaptive threshold image segmentation algorithm for color object segmentation. The proposed algorithm searches target in specific region to ensure the accuracy and speed of the segmentation. Meanwhile, identifies the boundary of the target area by chain code. Use adaptive threshold while segment, to solve the problems in traditional fixed threshold, like undetected or false detection issue.4. Focus on the application of Bayesian filtering in video target tracking algorithm. Propose a target tracking algorithm based on weighted color histogram by combining Kalman filter and particle filter. The proposed algorithm use adaptive tracking strategy to integrate the advantages of both Kalman filter and particle filter. The algorithm proposed an improved Kalman filter tracking method based on background subtraction, while a particle filter tracking algorithm use key area weighted color histogram is used to guarantee the sensitiveness of the target’s color feature.