Object-based video streaming technique with application to intelligent transportation systems

With rapid development of the Internet, traffic surveillance systems using Internet video streaming techniques are becoming mature. This paper presents a vision-based intelligent transportation system that can perform real-time traffic monitoring remotely. A background-registration method is used to dynamically adapt the background and separate vehicles from the background with an adaptive threshold. To utilize the bandwidth efficiently, important vehicles can, in real time, be segmented, encoded, and transmitted with higher quality and higher frame rate than those of the background. According to the traffic flow and mean speed, traffic events such as traffic jams and speeding can be detected. Automatic monitoring of traffic events would be useful in reducing the workload of human operators and providing road information. The experimental results show that the real time traffic surveillance system is indeed effective and efficient.