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.%基于传统Kalman滤波器理论的背景建模方法,不能很好地解决目标缓慢运动导致背景模型出现拖影的问题.针对该问题,提出了一种结合Kalman滤波器理论与动态区域重构的自适应背景建模改进算法,介绍了Kalman滤波器理论主要思想和改进算法的方法流程与效果.与传统的Kalman背景建模相比,该方法在增加少量计算复杂度的前提下,较好地解决了目标缓慢运动导致背景模型出现拖影的问题,同时也能较好地消除背景噪声.通过对图像序列的仿真实验证明:该方法具有很好的实用性与鲁棒性.
展开▼