Advanced Moving Object Detection and Tracking for Video Surveillance

Abstract

Moving object detection is a very crucial and challenging task in computer vision applications such as surveillance, vehicle and human tracking. Background subtraction is a preliminary technique widely used for the moving object detection. In this paper, an advanced automated moving object detection technique using background subtraction is proposed. The method uses running average wavelet transform (RAWT) for accurate registration of background from the video sequence. Furthermore, the moving objects are detected by comparing current and background frame. In order to produce higher accuracy for the object detection, the proposed method also further includes post-processing filter operation after which the binary object detection mask can be obtained. After moving object detection, tracking is performed. Experimental results demonstrate that the proposed method is faster and efficient as compared to the other state-of-the-art existing methods.