Implementation of real time moving object detection using background subtraction in FPGA

Background subtraction is one of the techniques used in video surveillance system for detecting moving objects in a video. In this paper we propose a background subtraction method which gives output even when the camera is shaking. There are many challenges that we have to consider for developing a robust background subtraction method mainly illumination variation. Here the algorithm works in such a way that the input frames are compared and compensated with reference frame then separating the foreground object with respect to background. Background subtraction can be built very efficiently on an FPGA which can process 640×480 video sequence. Experimental result shows that it is possible to detect the moving object in a very fast manner. It is implemented in Digilent Atlys Spartan-6 FPGA development board can solve various problems like complex computation, data transmission, cost of hardware resources etc. The real time video is captured by using VmodCAM.