基于加权核范数与3D全变分的背景减除  

Background Subtraction Based on Weighted Nuclear Norm and 3D Total Variation

作　　者：班颖 邵泽军[1] 牛玉玲[1] BAN Ying;SHAO Zejun;NIU Yuling(School of Architecture,Yanching Institute of Technology,Langfang Hebei 065201,China)

机构地区：[1]燕京理工学院建筑学院,河北廊坊065201

出　　处：《信息与电脑》2023年第4期17-20,47,共5页Information & Computer

基　　金：燕京理工学院2020年度科教研一般项目数字图像处理教学辅助软件的开发(项目编号:2020YITSRF102)。

摘　　要：针对鲁棒主成分分析模型(Robust Principal Component Analysis,RPCA)一般将前景看作背景中存在的异常像素点,从而使得在复杂背景中前景检测精度下降的问题,提出一种基于加权核范数与3D全变分(3D-TV)的背景减除模型。该模型以RPCA为基础,利用加权核范数来约束背景的低秩性,考虑了不同奇异值对秩函数的影响,使其更接近实际背景的秩;然后利用3D-TV来约束前景的稀疏性,考虑了目标在时空上的连续性,有效抑制了复杂背景对前景提取造成的干扰。实验结果表明,与其他4种算法对比,所提模型的F值基本上是最优的,且能准确地分离图像中的背景和前景。In view of the fact that Robust Principal Component Analysis(RPCA)generally regard the foreground as abnormal pixels in the background,which makes the foreground detection precision decrease in the complex background,a new background subtraction model was proposed based on weighted nuclear norm and 3D Total Variation(3D-TV).Based on the RPCA,the weighted nuclear norm was used to constrain the low rank of the background,which considered the influence of different singular values on the rank function,making it closer to the rank of the actual background.Then the moving objects were regularized by 3D Total Variation.It considered the spatio-temporal continuity of moving foreground and effectively suppressed the interference of bad background.The experimental results show that,compared with the other 4 algorithms,the F values of the proposed model are mostly optimal.Therefore,the proposed algorithm can accurately separate the background and foreground in the image.

关 键 词：背景减除 鲁棒主成分分析(RPCA) 加权核范数 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]