Detection and tracking of moving objects by fuzzy textures

Video surveillance plays a prominent role in law enforcement, personal safety, traffic control, resource planning and security of assets, etc. The need for such systems is increasing every day, with a number of surveillance cameras deployed in public places to analyze moving objects. Automatic video surveillance system can enforce the security in the monitored area without requiring the continuous attention of human operators. For such systems, moving object detection and tracking in dynamic backgrounds such as waving trees, oceans etc... is still a challenging task. Noise in the foreground due to dynamic background condition should be removed effectively for efficient tracking of the moving objects. In this paper, objects present in dynamic backgrounds are detected using a simple and robust fuzzy texture based analysis of the video sequences. Then the detected moving object is tracked using the mean shift tracking algorithm to describe the object´s direction of motion. The proposed method effectively detects the moving objects and describes the object´s motion.