This capstone project tries to develop a robust nighttime traffic surveillance system which reuses roadside cameras to capture scenes, automatically analyze the traffic and particularly focuses on nighttime surveillance problem. The system consists of a preprocessing module, responsible for offline configuration for each specific traffic scene and a traffic analyzing module, dealing with real-time detecting and tracking vehicles in the scene. The traffic information obtained from the system during the surveillance includes number of traffic lanes in the scenes, location of vehicles on the lanes and their travel status. The system has been tested on different nighttime scenarios and proven to provide robust performance in both accuracy and processing speed. The purpose of this project is to reduce the lead time of the commuters near signals to reach their destination sooner and also to avoid accidents by introducing TSS (Traffic Surveillance System) which monitors the density of the traffic by giving feedbacks to the signal system to change into the desired triggers(red or green) at variable timings. This system also includes the advanced techniques of night surveillance system which is useful for commuters in the night time, thereby help the passengers to have a safe trip.