针对视觉背景提取(visual background extractor,ViBe)算法在运动目标检测过程中容易受到噪声干扰的问题,将两帧差分法融入ViBe的前景检测阶段,提出一种融合两帧差分信息的改进ViBe算法(ViBe with two-frame differencing,ViBe-TD).首先,设计单阈值形ViBe(single-threshold form of ViBe,S-ViBe)检测,为信息融合做准备;其次,基于逻辑斯蒂(logistic)回归模型,实现像素点上两帧差分和S-ViBe检测信息的融合;最后,综合两类检测信息完成前景像素点的判定.实验结果表明,ViBe-TD算法在4种不同场景视频上的检测效果达到了 0.932的平均精确率,0.785的平均召回率以及0.842的平均F1 值.与原算法相比,ViBe-TD算法的各项指标平均有0.158的提高,具有良好的检测效果.
Abstract
Aiming at the problems of noise interference in the moving target detection by the visual background extractor(ViBe)al-gorithm,the two-frame differencing method was incorporated into the foreground detection phase of ViBe,and an improved ViBe algo-rithm(ViBe with two-frame differencing,ViBe-TD)that incorporates two-frame differencing information was proposed.Firstly,the sin-gle-threshold form of ViBe(S-ViBe)detection was designed to prepare for the information fusion.Secondly,the fusion of two-frame difference and S-ViBe detection information on pixel points was achieved based on logistic regression model.Finally,the two types of detection information were combined to complete the determination of foreground pixel points.The experimental results show that the ViBe-TD algorithm achieves an average precision of 0.932,an average recall of 0.785 and an average F1 value of 0.842 for the de-tection on four different scene videos.Compared with the original algorithm,the ViBe-TD algorithm shows an average improvement of 0.158 in each metric and has a good detection effect.
维普
万方数据