基于帧间差分和背景差分的运动目标检测算法

Moving object detection algorithm based on frame difference and background subtraction

针对运动目标检测领域中帧间差分法和背景差分法的缺陷,提出一种将两种方法融合在一起的新算法。首先,该算法在采用混合高斯建立背景模型时对方差更新作了修改,使得模型与真实背景更接近。其次,用连续三帧差分代替两帧差分,采取自适应差分阈值的方法。最后,将两种差分的结果融合并作形态学处理提取目标。实验结果表明,本文算法能有效抑制噪声和空洞,适应性强、检测效果良好。

In order to overcome the defects of frame difference and background subtraction algorithms in moving object detection,a new method that integrates the two algorithms is proposed.First,when the Gaussian mixture model is employed to build the background model,the update of the variance is modified,so the model is more close to the real background.Then,two consecutive frame differencing is replaced by three consecutive frame differencing,and the differencing threshold is adaptive.Finally,the results of the two differences are integrated and the target is extracted after morphological process.Experimental results show that the method can effectively suppress noise and the ＂ hole＂,the method is more adaptable and has good detection results.