IJARCCE - Computer and Communication Engineering

Abstract

Surveillance cameras are video cameras used for the purpose of observing an area. Sensitive areas need to be observed continuously for security reasons. The security of these areas shall not be depending on the size of an area but, shall be considered on the high priority of security. Before few decades this task was done by the humans, but now it is not possible for human to keep watch day and night (24x7) on such areas. We all know that there is massive increase in theft in either in one or more ways. Human can easily be distracted and, a small distraction is sufficient for thieves to break the security. System is proposed and presented using our own data set captured in laboratory datasets. The proposed system works on QVGA (Quarter Video Graphics Array) resolution at which most CCTV (closed circuit television) cameras operate and uses a simple mathematical model for detection of human and object. Proposed work is a combination of base works as 1: Real time video feed. 2: Discover frame from the video after fixed time interval. 3: Dual time background subtraction algorithm which periodically updates two sets of background. 4: Deriving a reference image. 5: Maintaining a Buffered Background image/frame. 6: Blob detection. 7: Motion Tracking. 8: Object tracking. 9: Detection and maintaining log of and Alarm signalling. An algorithm for tracking of abandoned objects even under occlusion is also proposed.