| At present, there are many application of detection and tracking of moving objects in video surveillance system, for instance, all day video surveillance is adopted by supermarket and bank and so on. The thesis is focused on detection and tracking of moving objects for auto video surveillance and alarm system.Firstly, detecting of moving objects is researched in the thesis. According to the characteristics that video sequences are influenced by factors such as light irradiation and how to select the best threshold, improved background difference is presented and coparative analysis of experimental methods by traditional background difference and improved background difference.Secondly, tracking of moving objects is researched in the thesis.in the aspect of tracking moving objects, least-square curve-fitting is adopted to delineate the trajectory of series of barycenter, thus the information of moving object is obtained and the trend of motion is predicted by analysing the curve. The fitting results show the information of the moving objects.Finally, a general design frame of auto video surveillance and alarm system is presented and functions of detecting, tracking and alarm are designed and realized. Effective detecting and tracking of moving objects and real-time alarm can be done under condition of stationary camera. |