Implementation of Motion detection and Tracking based on Mathematical Morpohology

Security is becoming the primary concern of society. Having a security system is therefore becoming a requirement. Video surveillance plays a vital role in security systems. Consider a video stream taken by a fixed camera to monitor motion-restricted area. The system works on a real-time video,from the camera, which is accessed frame to frame. The proposed system uses mathematical morphology for implementation. The system is implemented using a generic tool for Computer Vision applications, EmguCV package in OpenCV. The system uses absolute differencing for motion detection. Each video frame goes through image filtering, for noise removal. For handling illumination changes edge feature, contour extraction is used, to avoid spurious movements appropriate threshold is selected for respective area to be surveillance. Once motion due to target object is detected, alert is generated continuously until the alert is not de-activated. After the motion is detected, the object causing motion is tracked indefinitely in live video. Even if the subject leaves the area, it's last location is bounded and seen on the system. It's movements are tracked and respective frames are stored on hard disk drive.