Comparison of moving object detection algorithms

At present, object detection methods widely used are background subtraction and Frame difference. The core of background subtraction is background modeling. There are several commonly used background modeling algorithms, such as Single Gauss background modelling (SG), Mixed of Gaussian (MOG) background modelling, and Running Average (RA) background modelling. In the paper, firstly the background subtraction based on these three different background modelings, and the frame difference algorithm are systematic studied. Furthermore, the performance of all the algorithms is compared. Based on the comparison, a new object detection algorithm fused MOG and RA is proposed. This method effectively overcomes the detection failures, which are caused by illustrate sudden change in video detection, The experimental results prove effectiveness of the proposed method.