文献类型：期刊文献

中文题名：一种基于多属性单值中智集相关度量测的视频运动目标检测方法

英文题名：A Method for Visual Foreground Detection Using the Correlation Coefficient between Multi-Criteria Single Valued Neutrosophic Multisets

作者：胡珂立[1];范恩[1];叶军[2];樊长兴[1];沈士根[1];谷宇章[3]

机构：[1]绍兴文理学院计算机科学与工程系;[2]绍兴文理学院电子与信息工程系;[3]中国科学院上海微系统与信息技术研究所

年份：2018

卷号：31

期号：5

起止页码：738

中文期刊名：传感技术学报

外文期刊名：Chinese Journal of Sensors and Actuators

收录：CSTPCD、、北大核心2017、Scopus、CSCD2017_2018、北大核心、CSCD

基金：国家自然科学基金项目(61603258;61703280);浙江省公益技术应用研究项目(2016C31082)

语种：中文

中文关键词：运动目标检测;中智集;单值中智集;相关度量测

外文关键词：foreground detection;neutrosophic set;single valued neutrosophic set;correlation coefficient

中文摘要：视频运动目标检测是视觉传感器数据分析的主要任务之一,其主要用于视频监控。鉴于效率,一些基本的背景模型经常被用于该项任务。然而,当视频图像中存在由于天气、背景扰动等因素带来的噪声时,基于该类模型的动目标检测效果往往会受到严重影响。基于中智集理论提出了一种改进的运动目标检测方法。首先,利用基础背景模型计算得到背景差图像序列;接着提出面向中智集理论的真(Truth)、不确定(Indeterminacy)、假(Falsity)量测,将各背景差图像转换到中智集范畴;最后,综合利用多周期单值中智集相关度量测和一般单值中智集相关度量测强化图像中的运动目标区域,并利用Otsu方法确定最优分割阈值。通过对现实视频序列测试,实验结果表明本文方法能够克服恶劣天气、背景扰动等不良因素,鲁棒完成运动目标提取。

外文摘要：Visual foreground detection is one of the most important tasks for the data analysis of the visual sensor,and it is always applied in video surveillance. Several basic background models are often employed due to their high efficiency. However,their results will be highly disturbed when there exists noisy information generated by the bad weather,dynamic background etc. We utilize the theory of NS（ Neutrosophic Set） to propose an improved method for foreground detection. First,a sequence of background subtraction images is calculated. Second,such a sequence is represented in the NS domain via three membership subsets T,I,and F. Finally,the correlation coefficient between multiperiod or normal single valued neutrosophic multisets is employed to highlight the moving objects on the image plane.Then the Otsu＇s method is employed to determine an optimized value for segmentation. Experiments are conducted on a variety of real-world video sequences,and the experimental results demonstrate that the proposed approach can extract the moving objects robustly,even when there exists challenges like bad weather or dynamic background.