基于背景差和多层背景的目标持续检测研究  

Object Detection Based on Background Subtraction and Multiple Layer

作　　者：崔浩[1] 张骏[1] 张通[1] 

机构地区：[1]西北工业大学自动化学院,西安710072

出　　处：《机电一体化》2011年第1期26-29,共4页Mechatronics

摘　　要：传统的目标检测方法只能检测出运动的前景目标,而进入监控场景后停止或移动缓慢的目标将被更新到背景里。文章提出了一种基于背景差和多层背景模型的前景目标持续检测方法,将背景分为动态背景和参考背景两层,采用改进的Surendra算法快速提取背景;目标检测中,采用动态背景检测运动前景,通过动态背景与参考背景之差检测出静止前景。对PETS2001数据库视频进行实验,结果表明该方法不受目标运动状态的影响,可以对目标进行持续检测。Most existing object detecting method detect only moving object but stationary object will updated in background. This paper presents a novel object detecting method, in which detect object with multiple layer and classify foreground divide into two classes, moving object and stationary object. The background is divided into two layers, reference background and dynamic background. Extended Surendra algorithm is used, Characteristic of the proposed method is to detect stationary object by analyzing the difference of the two layers, and detect moving object by background subtraction. The experiment results show that the proposed method can detect moving object and stationary object.

关 键 词：智能视频监控 背景差 多层背景 静止前景 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]