背景建模是实现运动目标检测与跟踪任务的关键技术之一, 背景模型的鲁棒性问题受到普遍关注. 本文针对背景建模所依赖的不同信息特征, 从实际应用和样本集形态两个方面分析了背景模型的鲁棒性需求. 根据不同信息的描述和处理的特点综述了背景建模的典型算法, 并考察其对鲁棒性需求的处理策略. 然后就不同层次信息的描述及其鲁棒性, 比较了典型背景建模系统, 并分析了背景建模技术的发展趋势.-Background modeling is a key technology to achieve moving target detection and tracking tasks, robustness issues of general concern background model paper for background modeling features rely on different information, from practical application form and two sample sets analyze the robustness of the demand for background model. according to the characteristics described and summarized the different information processing typical background modeling algorithm and investigate its handling strategy for robust demand. then describe the different levels of information and Lu stick resistance, the system compares the typical background modeling and analyzes the development trend of background modeling technology.