Automatic Surveillance using Motion Detection

Abstract — The current video surveillance techniques store the complete video even if there are many idle frames in the video. To go through the complete video is a cumbersome task. The storage requirement of such videos is also huge. So we are attempting to design an application which ignores the idle frames from the video, with effective and real time object detection and video surveillance system with the help of SOBEL operator algorithm, for edge detection, in order to reduce the amount of storage space and remove the redundancy from the video. The aforementioned system can be implemented using a webcam or a CCTV and an efficient algorithm to detect the motion and robustly distinguishes the changes in consecutive frames and ignores the lighting changes. Once the motion is detected, the application will send SOS and activate the alert system and start storing the video. Storing of the video will automatically come to halt when there is a stagnancy in the scene