Moving Target Detection And Tracking In Video Sequences,

Visual information, especially dynamic visual information, takes the most part of environment information that is perceived by human beings. Therefore, perceiving the dynamic visual information in the environment has become an important research field in computer vision. Moving objects detection and tracking is one of the most important issues in applied vision and moving image coding and has wide applications in many fields.The research interest of moving objects detection and tracking is video sequence, i.e. image sequence. Object detection in videos involves verifying the presence of an object in image sequences and possibly locating it precisely. Object tracking is to monitor an object's spatial and temporal changes during a video sequence, including its presence, position, size, shape, etc. These two processes are closely related because tracking usually starts with detecting objects, while detecting an object repeatedly in subsequent image sequence is often necessary to verify tracking. Due to changes in illumination of the scene, background perturbations, shadows, vibration of the camera and occlusions between moving objects, precise detection and tracking of moving objects are still a challenge field of research.Based on the review and analysis of current methods of detecting and tracking moving objects, we lay an emphasis on the research of moving objects detection and tracking in image sequence obtained by a stationary camera. The major work implemented in this paper is presented as follows:(1) An edge-based algorithm of detecting moving objects is proposed. For a video sequence, moving change regions are achieved firstly by means of converting difference image to binary image using adaptive thresholding and morphological filtering. Then, with the edge information of current frame, the edges of moving objects are generated. Finally, moving objects are detected according to connective areas obtained by region filling.(2) According to the information between video sequence and constraint of average intensity and image energy, an improved algorithm of background reconstruction based on block is presented. Experimental results show that the improved algorithm combines the advantages of the original algorithm, but is faster in reconstructed speed.(3) Based on the analysis of familiar methods of detecting moving objects, we present a framework for detecting moving objects based on difference image. Detection of moving objects is done by three steps: change detection, morphological filtering and connected component labeling. A great deal of experiments in the case of different difference show that this framework is effective.(4) A preliminary investigation is made into moving objects tracking. We present an algorithm of tracking moving objects based on Kalman filtering, which accomplishes object tracking by object detection, motion estimation and object matching. Object detection involves verifying the presence of moving objects in image sequence and describing the shape feature of present object such as the location of object and the size of its enclosing rectangle. Motion estimation is to predict the location of the tracked object by means of Kalman filter in subsequent frames and ascertain the scope for searching the tracked object. Object matching is to establish one-to-one correspondence between moving objects over frames and the track of every moving object.