Advanced Motion Detection Technique using Running Average Discrete Cosine Transform for Video Surveillance Application

Object detection is always the first important step in video surveillance applications. This paper presents an automated moving object detection technique using background subtraction method. First, a background module generated from the video sequence effectively. Furthermore, the moving objects are detected by comparing current and background frame. Many background subtraction approaches are available but most of them are not illumination sensitive. The proposed method updated the current status of the background with sudden illumination change background in frequency domain using modified running average discrete cosine transform (RA-DCT). In order to produce higher accuracy for the motion detection, the proposed method also further includes a noise filter after which the binary motion detection mask can be computed. Experimental results demonstrate that the proposed method is faster and efficient as compared to the other existing methods.