首页 | 本学科首页   官方微博 | 高级检索  
     检索      

综合视觉特征度量与SVM的遥感图像检索方法
作者姓名:赵理君  唐家奎  于新菊  王春磊  张成雯
作者单位:1. 中国科学院海岸带环境过程重点实验室,山东省海岸带环境过程重点实验室,中国科学院烟台海岸带研究所, 山东 烟台 264003; 2. 中国科学院研究生院, 北京 100049; 3. 河北联合大学, 河北 唐山 063009
基金项目:国家自然科学基金(40801124)、山东省中青年科学家科研奖励基金(2010BSA06013)、中国科学院创新团队国际合作伙伴计划 、中国科学院数字地球重点实验室开放基金(2011LDE015)和中国科学院研究生院院长基金资助 
摘    要:基于支持向量机(SVM)的相关反馈算法在反馈过程中只利用SVM的分类器,反馈结果排序会出现一定错误, 提出一种改进的相关反馈策略,将图像的视觉特征度量函数和SVM分类器函数进行线性加权,作为相关反馈中的相似性度量准则. 实验表明,改进策略能够优化遥感图像检索排序结果,提高检索的精度.

关 键 词:遥感图像检索    支持向量机    相关反馈    相似性度量
收稿时间:2012-05-02
修稿时间:2012-05-29

Remote sensing image retrieval method integrating feature similarity measurement and SVM
Authors:ZHAO Li-Jun  TANG Jia-Kui  YU Xin-Ju  WANG Chun-Lei  ZHANG Cheng-Wen
Institution:1. Key Laboratory of Coastal Zone Environmental Processes, Chinese Academy of Sciences; Shandong Provincial Key Laboratory of Coastal Zone Environmental Processes, Yantai Institute of Coastal Zone Research, Chinese Academy of Sciences, Yantai 264003, Shandong, China; 2. Graduate University, Chinese Academy of Science, Beijing 100049, China; 3. Hebei United University, Tangshan 063009, Hebei, China
Abstract:The support vector machine (SVM)-based relevance feedback algorithm has been used in common image retrieval, but not widely applied to remote sensing images. Traditional algorithm only uses SVM classifiers, resulting in some wrong ranking sequences of retrieval results. An improved relevance feedback strategy is proposed, and it modifies the similarity measurement criterion using a weighted linear combination of feature similarity measurement and SVM classifier. Experimental results show that the proposed method improves the ranking sequence and accuracy of retrieval results.
Keywords:remote sensing image retrieval                                                                                                                        support vector machine                                                                                                                        relevance feedback                                                                                                                        similarity measurement
点击此处可从《》浏览原始摘要信息
点击此处可从《》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号