首页 | 官方网站   微博 | 高级检索  
     

基于灰度值限定信息熵的图像检索方法
引用本文:胡潭高,张锦水,潘耀忠,朱文泉.基于灰度值限定信息熵的图像检索方法[J].遥感技术与应用,2007,22(4):543-548.
作者姓名:胡潭高  张锦水  潘耀忠  朱文泉
作者单位:(北京师范大学资源学院资源技术与工程研究所,环境演变与自然灾害教育部重点实验室,北京 100875)
基金项目:教育部跨世纪优秀人才培养计划;国家高技术研究发展计划(863计划)
摘    要:颜色直方图图像检索方法是一种重要的基于颜色特征图像检索方法。在分析了基于颜色直方图信息熵进行图像检索方法的基础上,提出了一种通过灰度值对信息熵进行限定的遥感图像检索算法。该方法在一定程度上解决了传统的基于颜色直方图信息熵方法进行图像检索时,由于熵的对称特性造成图像误检索的不足。在实验精度评价中,分别采用直方图分析法和相关系数法,从定性与定量两个角度对检索精度进行了评价。实验结果显示,运用该方法进行检索,具有较高的检索精度,相关系数在0.95以上,在实际应用中具有良好的可行性。

关 键 词:相似度  信息熵  颜色直方图  灰度值  
文章编号:1004-0323(2007)04-0543-06
收稿时间:2007-02-08
修稿时间:2007-02-082007-07-02

Image Retrieval Based on Limited Hue Value on Information Entropy
HU Tan-gao,ZHANG Jin-shui,PAN Yao-zhong,ZHU Wen-quan.Image Retrieval Based on Limited Hue Value on Information Entropy[J].Remote Sensing Technology and Application,2007,22(4):543-548.
Authors:HU Tan-gao  ZHANG Jin-shui  PAN Yao-zhong  ZHU Wen-quan
Affiliation:(College of Resources Science and Technology at Beijing Normal University,Key Laboratory of Environmental Change and Natural Disaster of the Ministry of Education of China at Beijing Normal University,Beijing100875,China)
Abstract:A new image retrieval algorithm based on color histogram information entropy is presented.Because the traditional method has not considered the symmetrical character of entropy,so it is easy to create mistakes.By using a limited hue value on information entropy the new method can solve this problem well.As to evaluate the accuracy,we used histogram analysis and correlation coefficients,from qualitative and quantitative of two views to make sure accuracy is dependable.The results show that use this method to search for a similar image is possible and have a high accuracy,the correlation coefficients is above 0.95,and in practical application this method also has a good possibility.
Keywords:Similarity  Information entropy  Color histogram  Hue value
本文献已被 CNKI 维普 万方数据 等数据库收录!
点击此处可从《遥感技术与应用》浏览原始摘要信息
点击此处可从《遥感技术与应用》下载全文
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-23

京公网安备 11010802026262号