基于自适应背景的实时运动物体检测

Real-time motion detection based on adaptive background

为了能够在监控系统中实时的检测出运动物体,提出了一种能够自适应获取并更新背景的运动检测方法。它是利用连续帧差分,当前帧与背景图像差分来获取运动信息,同时选择性的更新背景。实验证明该算法运算量小,鲁棒性好,算法中通过直方图分析来获取合适的阈值,提高了系统的自动化程度。

An efficient method for motion detection was presented here.With this method,a stable background can be obtained and kept to cope with variations on environmental changing conditions.At the same time the final moving object can be gained.In the difference image,an adaptive thresholding was used to distinguish background and foreground to improve the level of self-adaptive.Finally,the experiments has been done to prove the method is quick and robust in detecting the moving object even in the outside environment.