一种基于ViBe的自适应运动目标检测算法

An adaptive moving target detection algorithm based on ViBe

针对视觉背景提取算法(ViBe)不能快速适应背景复杂度变化比较大以及检测出的运动目标容易产生伪影的问题,融合样本集标准差与颜色畸变的概念,提出了一种基于ViBe的自适应运动目标检测算法。首先利用前m帧视频序列对应的各个像素点的均值构建背景模型;然后将样本集的选取范围由8邻域扩展到24邻域;最后用自适应阈值和颜色畸变阈值双重限制代替原来的半径阈值R和匹配阈值T。实验表明,改进的算法可以更快的消除背景模型中的鬼影,可以准确的检测出动态背景下的运动目标,并且F 1-measure最高提升了10%,具有更高的准确性。

In view of the problem that the visual background extraction algorithm(ViBe)can't quickly adapt to the large background complexity changes and the detected moving objects are prone to generate artifacts,this paper proposes an adaptive moving object detection algorithm based on vibe,which integrates the concept of standard deviation of sample set and color distortion.Firstly,the background model is constructed by using the average value of each pixel corresponding to the pre frame video sequence;secondly,the selection range of the sample set is extended from 8 neighborhood to 24 neighborhood;finally,the original radius threshold and matching threshold are replaced by the double limit of adaptive threshold and color distortion threshold.Experimental results show that the improved algorithm can eliminate the ghost in the background model faster,and can accurately detect the moving target in the dynamic background,and the F1 increase of 10%,with higher accuracy.