Identification of running Vehicle using Vision Based Holistic Property of Registration Technique for Traffic Surveillance in India

The fundamental step of moving object in any image is visualize in different vision-systems including automated visual surveillance and conventional systems. Mostly in India, traffic system related to small geographic areas are more complex to handle. So our objective is to control the congestions and monitor the traffic activities by detecting the vehicle by using vision-based holistic registration technique, which will be helpful to predict the traffic flow. Many researchers apart from our country are doing more and more research and experiments to control and monitor the existing surveillance system. Some existing algorithm for vision-based holistic properties registration are discussed in this paper and we use this hybrid technique to develop this surveillance system based on monocular image sequences. In this paper we proposed a working model of such surveillance framework for detection and identifying the high speed vehicles and traffic flow measurement. The method is natural which is based on the establishment of interfaces between regions and vehicles, while moving through the image sequencing. Due to Background subtraction technique it improves the adaptive background mixture model and makes the system learn faster and more accurately. For the experiments and results Real-life traffic image sequences have been taken measure the effectiveness of the proposed algorithm.