Abstract: In order to overcome the deficiency of updating the background model and removing the shadow when the background subtraction method is applied to detect moving objects，an improved algorithm based on the single-Gaussian background model is proposed.The experimental results indicate that the algorithm can remove the shadow and extract the moving object effectively，which lays a solid foundation for the following work such as body recognization，behavior analysis and so on.

摘要： 针对背景减除法应用于运动目标检测中的背景模型更新和阴影消除问题，提出了一种改进的单高斯背景模型估计算法和快速的阴影消除方法。实验结果表明，算法在有效去除阴影的同时能够较好地提取运动目标，为后续的人体识别、行为分析等工作奠定了良好的基础。