As background subtraction is a common computer vision task; we analyze the usual pixel-level approach. We develop an efficient adaptive algorithm system based on pixel comparison. Recursive comparison of pixel is used between the present frame and the reference frame. This algorithm is implemented using image processing in MatLAB Environment and we work on making it a real-time applicable tool for various possible applications. Thus an attempt to build a video system for real-time detection and tracking of motion which has the ability to minimize both false detections and missed detections, interfaced with a hardware unit based on microcontroller, communicating serially with the computer system as a control unit panel prototype. It is capable of processing 320times240 video at 28 fps, excluding post processing.