基于像素的背景建模方法速度较快但不能很好地描述背景运动，光流能准确描述物体运动但计算量大，难以满足实时的要求．提出一种结合基于像素的背景建模方法速度快以及光流描述物体运动准确优点的背景建模和目标检测方法．具体来说，为静止背景建立传统基于像素的灰度背景模型，为运动背景建立光流背景模型，通过２种背景模型的有效结合快速准确地实现目标检测．实验结果表明，提出的方法建模速度与基于像素背景建模方法相当，同时，又有光流准确描述背景运动的优点，综合性能超越上述２种方法．-Faster but not a good descr iption of the background motion background modeling method based on the speed of the pixel, the optical flow can accurately describe the movement of objects but computationally intensive, difficult to meet real-time requirements. Proposed background modeling method based on a combination of speed and a fast pixel optical flow motion of an object accurately describe the advantages of background modeling and object detection methods. Specifically, the establishment of the traditional pixel-based gray background model to establish the background model optical flow motion background, quickly and accurately achieved through the effective combination of two kinds of background model of target detection is still background. Experimental results show that the proposed method of modeling the speed of the pixel-based background modeling method is quite the same time, another advantage of accurately describe the background of the optical flow motion, the overall perform