An Improved Background Update Algorithm and Application in Traffic Engineering
Shi Song Zhu, Haiyan Zhang, Cui Yun Zhang, Hong Jin Zhu · Applied Mechanics and Materials · 2014
The statistical histogram algorithm is one of the common background update algorithms. Considering the statistical histogram algorithm disadvantages of large complex computation, the paper proposes an improved method named fixed window statistical background updating algorithm. The algorithm sets a window which contains N records for each pixel point firstly, every record records value in the permissible range of error and the number of pixel point occurrence, and then counting pixel values for every record through statistical histogram algorithm. When the window is full, the new pixel values will replace the smaller counting, while the pixel which counting larger will be adaptively updated by adaptive background updating formula. The experimental result shows that this method has a high real-time performance and robustness in background updating.