Abstract: An algorithm for moving objects detection based on Gaussian mixture model and three-frame differencing is proposed to make up the deficiency of the background subtraction method based on mixture model of Gaussian.First，for improving the sensitivity，the three-frame differencing is used to rapidly detect the possible moving objects.Second，a threshold is used to judge if the moving objects exist that can ease the complexity of the calculations.Third，in order to increase the convergence speed，the model sets different update rate between moving objects area and background area.The experimental results indicate that the proposed method possesses faster processing speed and better effect.It can be used in real-time video surveillance system.

Key words: moving objects detection, Gaussian mixture model, three-frame differencing

摘要： 针对混合高斯模型背景法的不足，提出了一种将混合高斯模型背景法与三帧差分法相结合的运动目标检测算法。利用三帧法快速检测出变化区域，提高了算法的灵敏度；引入目标是否存在的判决阈值，减低了算法的运算量；对目标区域和背景区域进行不同的混合高斯背景模型的更新策略，提高了模型的收敛速度。实验结果表明，改进的方法与混合高斯模型背景法相比其处理速度快，效果更好，适用于实时视频监控系统。

关键词: 运动目标检测, 混合高斯模型, 三帧差分法