Real-time moving object tracking in video

Identifying moving objects from a video sequence is a fundamental and critical task in many computer vision applications. After the images are captured they must be processed and then sent to the server. This paper describes an implementation of a real time object tracking system in a video. Real tracking is achieved using a simple and fast motion detection method based on frame substation. The image processing algorithm is explained. Experimental results shown that any kind of moving object can be detected under unconstrained scenes.