Video surveillance has lengthy been in use to display protection-sensitive regions which include banks, branch stores, highways, crowded public locations, and borders. The develop in computing electricity, availability of great ability garage gadgets and excessive speed network infrastructure paved the way for less expensive, multi sensor video surveillance systems. Traditionally, the video outputs are processed on line by using human operators and are stored to tapes for later use handiest after a forensic event. The growth in the variety of cameras in conventional surveillance structures overloaded each the human operators and the garage devices with excessive volumes of facts and made it infeasible to make certain proper tracking of sensitive areas for lengthy instances. To filter redundant statistics generated using an array of cameras, and boom the reaction time to forensic occasions, helping the human operators with the identity of important events in the video using the use of “clever” video surveillance systems has come to be an essential requirement. The making of video surveillance structures “smart” requires fast, dependable and strong algorithms for moving item detection, type, monitoring and pastime analysis.

Keywords : video surveillance, object tracking, mean square error