一种基于模糊分类的运动目标检测算法  

A New Method for Moving Objects Segmentation Based On Fuzzy Clustering

作　　者：任明艺[1] 李晓峰[1] 李在铭[1] 

机构地区：[1]电子科技大学通信与信息工程学院,成都610054

出　　处：《信号处理》2009年第3期435-438,共4页Journal of Signal Processing

基　　金：国家863高技术计划(2004AA823120);国家自然科学基金(10376005)联合资助

摘　　要：本文提出了一种基于模糊分类的运动目标检测算法。首先对序列图像进行差分运算,并计算差分图像的局域二阶矩,然后使用模糊分类的Powell快速算法提取运动目标。实验结果证明了本算法的有效性和可靠性。A new method for moving objects segmentation, which based on fuzzy clustering, is proposed in this paper. First the temporal difference transform is performed on the image sequences, the local variance image calculated, and then a Powell algorism based on fuzzy clustering is calculated to segment the moving objects. The experiment simulation shows the validity and reliability.

关 键 词：运动目标 检测 模糊分类 

分 类 号：TN911.73[电子电信—通信与信息系统]