四川大学学报（自然科学版）2011，Vol.48Issue(3)：566-570,5.DOI:10.3969/j.issn.0490-6756.2011.03.016
基于块背景建模的运动目标检测
Moving object detection using background model based on image blocks
摘要
Abstract
Moving object detection is a very important research topic of Computer Vision.It is the prerequisite of the following-up processing such as object identification, object tracking and scene understanding.This paper presents a block-based Mixture of Gaussian background modeling method, improves the shortcomings of the traditional Mixture of Gaussian background modeling which are slow and sensitive to non-stationary scenes.The new method builds the background model for scenes base on the mean and variance of image block.The results show that compared with the Mixture of Gaussian background modeling, the proposed approach can ensure the accuracy of extracted target and greatly improve the processing speed.关键词
运动目标检测/混合高斯模型/时空背景建模/分块背景建模Key words
moving object detection/ mixture of gaussians/ spatiotemporal background model/ blockbased background model分类
信息技术与安全科学引用本文复制引用
余孟泽,刘正熙,骆健,杨丹..基于块背景建模的运动目标检测[J].四川大学学报（自然科学版）,2011,48(3):566-570,5.基金项目
国家863计划课题(2009AA01Z332) （2009AA01Z332）
