摘要
提出一种基于局部特征融合的目标再识别方法.通过视觉传感网采集视频图像,针对场景内的目标,首先结合颜色和形体特征,对目标进行区域分割,进一步设计特征描述子,对关键区域的颜色和形体信息进行提取,基于此判断目标间的相似度,实现目标再识别.实验表明了所提方法的有效性.
An object re-identification method based on fusion of local features was proposed.Video image was obtained through visual sensor network and objects in scenes were concerned.Firstly,ob-ject was divided into multiple regions based on color feature and shape feature.Then,feature descrip-tors were designed to extract color and shape information of key regions.Finally,object re-identifica-tion was implemented by evaluating the similarity among the objects.The effectiveness of the pro-posed method was verified by experiments.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2015年第S1期192-195,共4页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(61273352
61175111
60805038)
山东省自然科学基金资助项目(ZR2011FM008)
关键词
目标识别
区域分割
局部特征
再识别
视觉传感网
objcect recognition
region segmentation
local feature
re-identification
visual sensor network
