Detection and Tracking for Moving Objects Using Feature Extraction Algorithm in Video

Video surveillance is basically used for analysis of object behaviors. It helps to detect as well as to track moving objects like car, bike, bus etc. Detection for moving object is a very challenging for any video surveillance system. In this paper, a framework is designed for moving vehicle detection and tracking that performs background subtraction, optical flow separately in video scene. We also developed a new method that is the combination of these two methods which can be used for both, single and multiple moving objects. Frame differencing method is used on proposed algorithm that performs feature extraction for tracking of moving objects in video. A comparison was also made between proposed algorithm with previous method such as background subtraction and optical flow by using three parameters. These parameters are called false alarm rate, precision and accuracy. Proposed algorithm gives better result in comparison of these methods. It also increases the value of occlusion rate for moving objects in video.