基于灰色绝对关联度边缘检测的多源遥感影像加权IHS融合  

Weighted IHS Fusion of Multi-source Remotely-Sensed Image Based on Edge Detection Using Grey Absolute Correlation Degree

作　　者：袁林山[1] 杜培军[1] 王莉[1] 张华鹏[1] 

机构地区：[1]中国矿业大学地理信息与遥感科学系,江苏徐州221008

出　　处：《地理与地理信息科学》2008年第3期11-15,共5页Geography and Geo-Information Science

基　　金：教育部新世纪优秀人才支持计划资助项目(NCET-06-0476);江苏省自然科学基金创新人才青年学术带头人项目(BK2006505);江苏省高等学校"青蓝工程"中青年学术带头人培养计划资助项目;中国矿业大学科技基金项目(2005B018);江苏省研究生创新计划项目(CX07B_138z)

摘　　要：鉴于IHS变换的融合影像存在光谱扭曲问题,而灰色关联度对图像边缘检测具有局限性,该文将基于灰色绝对关联度的灰色关联分析与IHS变换相结合进行影像融合。采用灰色绝对关联度检测高分辨率影像的边缘点和非边缘点,基于边缘信息确定高分辨率影像和IHS变换后的亮度分量在组合中的权值,最后通过IHS反变换实现影像加权融合。应用该算法分别对多源遥感影像进行融合处理,在提高空间分辨率的同时能够很好地保持多光谱影像的光谱信息,优于传统的IHS融合方法和基于灰色关联度的IHS融合算法。An improved fusion algorithm for multi-source remote sensing images with high spatial resolution and multi-spectral capacity is proposed based on traditional IHS fusion and grey correlation analysis. Firstly,grey absolute correlation degree is used to discriminate non-edge pixels and edge pixels in high resolution images,by which the weight of intensity component is identified in order to combine it with high resolution image. Therefore, image fusion is achieved using reverse IHS transform. The proposed method is applied to multi-spectral IRMSS data and CCD data from CBERS, and multi-spectral Landsat TM data and SPOT panchromatic data respectively. The experimental results prove that the fusion method proposed in the paper can efficiently preserve spectral information of the original nmlti-spectral images while enhancing spatial resolution greatly. By comparison and analysis, the improved fusion algorithm is better than traditional IHS fusion and fusion method based on grey relational analysis and IHS transform.

关 键 词：影像融合 灰色绝对关联度 边缘检测 IHS融合 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]