图切割支持的融合颜色和梯度特征的实时背景减除方法  

A Fusing Color and Gradient Features Approach to Real-time Background Subtraction using Graph Cuts

作　　者：邓宇[1] 李振波[1] 李华[1] 

机构地区：[1]中国科学院计算技术研究所智能信息处理重点实验室

出　　处：《计算机辅助设计与图形学学报》2006年第11期1741-1747,共7页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家科技攻关计划课题奥运科技专项(2001BA904B08);国家重点基础研究发展规划项目(2004CB318000);国家自然科学基金重点项目(60533090)

摘　　要：通过融合图像的颜色和梯度特征,实现了一种实时背景减除方法·首先融合颜色和梯度特征建立新的能量函数;然后基于图切割算法最小化能量函数,并对前景/背景进行分割;最后使用光流验证前景区域的真实性,并更新背景模型·对不同场景的实验结果表明:该方法可以实时地检测出视频序列中的运动物体,结果准确、有效·Based on the fusion of color and gradient features, this paper implements a novel approach to real-time background subtraction. Firstly, an energy function is defined based on the fusion of color and gradient features. Secondly, the graph cuts based algorithm is employed to minimize energy function and segment the foreground. Finally, average optical flow is used to make inference about the validity of foreground regions, background models are then updated. The experimental results of different real scenes show that the proposed approach can produce real-time detection and promising results.

关 键 词：背景减除 颜色和梯度特征 能量函数 图切割 实时检测 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]