Dynamic Probe Window Based Optimization for Surveillance in Home Security System

Tracking of a moving object is very important for video surveillance in a real time scenario. The proposed algorithm uses dynamic probe window based approach&combines the conventional edge based and frame differencing approach to achieve better algorithmic time complexity as well as improved results. First it computes the edge map of two consecutive frames with the help of first order differential sobel operator due to its noise resistant attributes and applies the frame differencing method between the two consecutive edge maps. Apart from the above optimization, our method doesn't differentiate between the scenario when motion occurs and when it doesn't, that is, almost same computation overhead is required even if motion is not there so it reduces the time complexity of the algorithm when no motion is detected. The effectiveness of the proposed motion detection algorithm is demonstrated in a real time environment and the evaluation results are reported.