南京林业大学学报(自然科学版) ›› 2010, Vol. 34 ›› Issue (01): 69-72.doi: 10.3969/j.jssn.1000-2006.2010.01.015

• 研究论文 • 上一篇    下一篇

基于ETM+遥感图像的图像融合试验及评价方法

孟京辉1,陆元昌1*,刘刚1,2,王懿祥1,3   

  1. 1.中国林业科学研究院资源信息研究所,北京100091;2.华中农业大学园艺林学学院,湖北武汉430070;江林学院环境科技学院,浙江临安311300)
  • 出版日期:2010-02-09 发布日期:2010-02-09
  • 基金资助:
    收稿日期:2009-03-17修回日期:2009-10-25基金项目:“十一五”国家科技支撑计划(2006BAD23B01);国家自然科学基金项目(30671679)作者简介:孟京辉(1980—),博士生。*陆元昌(通信作者),研究员。Email: ylu@caf.ac.cn。引文格式:孟京辉,陆元昌,刘刚,等. 基于ETM+遥感图像的图像融合试验及评价方法[J]. 南京林业大学学报:自然科学版,2010,34(1):69-72.

Trial of image fusion methods and quality assessment for ETM+image

MENG Jinghui1, LU Yuanchang1*, LIU Gang1,2, WANG Yixiang1,3   

  1. 1.Institute of Forest Resource Information Techniques, CAF, Beijing 100091, China; 2.College of Horticulture and Forestry Science, Huazhong Agricultural University, Wuhan, 430070, China; 2.School of Environmental Technology, Zhejiang Forestry College,Lin’
  • Online:2010-02-09 Published:2010-02-09

摘要: 为了对比分析不同方法在融合ETM+遥感图像的高空间分辨率全色波段和多光谱波段的效果,选取基于彩色技术的IHS和RGB变换、基于图像变换技术的PC变换、Gramschmidt变换和基于算数技术的Brovey变换4种融合方法,对研究区域ETM+图像进行了融合,并采用衡量信息量的信息熵、标准差及衡量光谱保真能力的偏差指数和相关系数进行了融合效果评价。融合结果表明:依Brovey、IHS、PC、GramSchmit方法融合后的图像,其光谱保真程度逐渐降低。Brovey变换最大限度保持了原始图像的光谱信息,而空间信息的详细程度较差;IHS变换空间融合后的图像空间细节信息最大,但光谱保真能力差;PC变换和Gramschmidt变换后融合图像的光谱保真和空间信息详细程度介于Brovey和IHS之间;Brovey和IHS有较好的融合效果。

Abstract: In order to evaluate different fusion methods in merging pan band and multispectral band for ETM+ image, fusion methods such as color based technique, nominal technique and image transformation were used and tested. For the quality assessment of fusion methods, quantitative analysis was done using the following criteria: standard variation, information entropy, bias index, correlation coefficient. The result showed that there was an obvious improvement of the spatial resolution for all fusion methods. For the spectral fidelity, it was decreasing following such an order: Brovey, IHS, PC and GramSchmit. Brovey kept the maximum spectral information of the original image but the details of spatial information were the lowest. In contrast, IHS had the highest spatial information but the spectral fidelity was not encouraging. PC and GramSchmit were in the middle level concerned with the detail of spatial information and spectral fidelity. Brovey and IHS were the two best fusion methods for this study and the results made significant sense for work related with remote sensing in the study area.

中图分类号: