An adaptive background subtraction approach based on frame differences in video surveillance

In the past decades, the insertion of cameras for the aim of surveillance is increased. Hence, a huge amount of data is produced by cameras. It is impossible to categorize and store all data. Therefore, algorithms that automatically process big data and track objects of interest are needed. Many methods are based on the opinion that the movement of objects causes differences in frames of a video, and the background would remain motionless during the video. Continuous dynamic behavior in the background deteriorates object detection performance.On the other hand, an excellent background extraction model can help to gain beneficent foreground detection results. The target of this paper is to model an algorithm that provides the pure background from video sequences. The idea of the proposed approach is to extract the background of complex and crowded scenes by using the differences of two consecutive frames’ pixels. Our experimental results show that the proposed approach provides significant performance in comparison with some previous techniques.