Visual Object Tracking Based On Mean-shift And Particle-Kalman Filter

Due to the increasing of video surveillance system requirements, Intelligent video surveillance system has become challenging topic in computer vision research field. There are four key steps in intelligent video surveillance system, i.e. object detection, object classification, object tracking, and object analysis. Among these steps, object tracking is considered as crucial and significant task in intelligent video surveillance system. Object tracking is considered as difficult task because of several problems such as illumination variation, tracking non-rigid object, non-linear motion, occlusion, and requirement of real time implementation. Therefore it is necessary to build a visual object tracking algorithm which can overcome these problems.Every single algorithm in visual object tracking always has both strengths and drawbacks. Therefore, utilizing only one single algorithm for tracking usually is considered as inefficient because every single algorithm has limitations. Based on this reason, in this thesis a tracking algorithm which combines mean-shift and particle-Kalman filter is proposed. In the proposed method, mean-shift is used as master tracker when the target object is not occluded. When occlusion is occurred or the mean-shift tracking result is not convincing, particle-Kalman filter will act as master tracker to improve the tracking results. Experimental shows that the proposed method can work well in dealing with tracking problems such as non-rigid object, occlusion, non-linear motion, and more suitable for real time implementation compare to the conventional algorithms.