提出了一种基于模型切换的背景建模方法(M SBM ).该方法以嫡图像为纽带, 实现了不同精细程度的背景模型在空间上的自适应选取和在时间上的自适应切换.对于亮度分布复杂度高的背景区域采用精细的模型以保证运动目标检测的精度,反之采用简单的模型以降低计算量

.通过模型结构自适应结合参数自适应, 很好地兼顾了检测精度和计算代价.墓于高斯混合模型和时间平均模型的双模型切换式运动目标检测算法被用于实验研究, 结果表明这种算法的检测效果和单独采用高斯混合模型的检测效果相当, 而计算速度却比后者提高很多-Proposed a model of background modeling method switching (M SBM) Based on the method of entropy image as a link, to achieve different degrees of background model in the fine spatial adaptive selection and adaptive switching in time. For high brightness distribution complex background area using sophisticated models to ensure the accuracy of the moving target detection, whereas a simple model to reduce the amount of computation

By combining adaptive parameter adaptive model structure, a good balance between detection accuracy and computational cost dual model tomb on Gaussian mixture model and the time-averaged model switched moving target detection algorithm is used for experimental research results show that the detection algorithm using Gaussian mixture model and a separate testing results have been very, and improve computing speed than many of the latter