Image Detection And Tracking Of Moving Object

The visual information, especially the dynamic visual information, occupies most of the composition among the environment information perceived by human beings. Therefore, the dynamic visual information has become an important research field of the computer vision in the perceiving environment. The image detection and tracking of moving object is one of the most important projects in the research field of applies vision and moving image coding. And, it has wide applications in many fields. The image detection and tracking of moving object is a process of detecting-recognition-tracking for the target of the image sequence, which combines the methods both of image detection and image tracking based on the dynamic image analysis.This article chooses the moving object as a target. A method is put forward that the image detection and tracking for moving target are realized under the condition of a simply still background through comparison and study of the algorithm of the target detection and tracking. The major work described is as follows:1. In respect of the target detection with static background, the current theories of target detection and extraction algorithms are summarized and the instant difference method and background subtraction are laid stress on the analysis.2. Based on GMM adaptive Background Estimation method is mainly analyzed and the updating parameters of the method are discussed. What's more, during the implementation of Adaptive Background Subtraction method, an improved scheme on updating background for specific application is put forward.. The experimental results show that the improved scheme can fairly meet the need of system and improve the accuracy of detected result.3. On the research of detection of shadow of moving targets, beginning with the optical characteristic of the shadow of moving targets, the paper proposes the algorithm of detection of shadow of moving targets by combining with two color spaces(the brightness and the chroma). The experiments show that the proposed methods have an ability to suppress the shadow of moving targets.4. A preliminary investigation is made for moving targets tracking. We present an algorithm of tracking moving targets based on Kalman filtering, which accomplishes targets tracking by targets detection, motion estimation and targets matching three modules. The target detection module involves verifying the existence of moving targets in image sequence and describing the shape feature of present target such as the location of target and the size of its enclosing rectangle. Motion estimation module is to predict the location of the tracked target by means of Kalman filter and ascertain the scope for searching the tracked target. The target matching module is to search for one-to-one correspondence relationship of the moving targets in every frame of image sequence and ascertain the moving locus of the target.