Object detection and tracking in video using particle filter

Deployment of effective surveillance and security measures is important in these days. The system must be able to provide access and track movement of different types of vehicles and people entering the secured premises, to avoid any mishap from happening.The paper proposes a system that recognizes the car with 3 different features namely license plate, logo and colour of the car. Existing systems perform recognition mainly by using license plate alone. Addition of features will increase the security of the system. Initially car region is extracted using frame subtraction method. On the extracted car region, License plate search and logo identification is being performed. Average colour of the car forms the third feature that helps in classification of cars. Finally with the extracted features, classification of cars into two categories is performed i.e. Authenticated and Non Authenticated The spatial segmentation and the temporal segmentation yields the moving objects. However, in practice, a moving object may suddenly cease motion or moves very slowly during several frames, which results in its corresponding intensity differences to be insignificant. Object in video are tracked and detected using particle filter. The particle filter is a Bayesian sequential importance sampling technique. It consists of essentially two steps: prediction and update. The paper analyzes applying of particle filter for tracking the object. The approach can further be combined with the training model developed using features for detecting and tracking cars in real time.