The background modeling methods based on traditional Kalman filter theory can not solve the smear problem on background model resulted from very slow movement target.To solve this problem

an improved adaptive background modeling algorithm combining Kalman filter theory with dynamic region reconstruction is proposed.The main ideas of Kalman filter theory and the processes and effect of the improved algorithm is elaborated.Compared with the traditional Kalman background modeling method

the improved algorithm can solve the smear problem and eliminate the background noise much better with a small increase of computational complexity.The improved algorithm is proved to have good practicability and robustness through simulation of image sequences.