摘要
本文提出一种快速离散Curvelet变换(FDCT)和IHS变换集成的遥感影像融合方法,可获得较传统方法更高质量的融合影像。在融合过程中,通过FDCT获取I分量的多尺度多方向系数集合,采用标准差加权的融合策略,自适应地调整空间细节与光谱信息的权重,从而达到最佳融合空间细节与光谱信息的效果。作者选择QuickBird和WorldView-2全色和多光谱影像进行融合实验,并与基于传统IHS、FDCT的方法进行了比较,采用两种评价模型,选择偏差指数、UIQI等质量指标进行客观量化评价,验证了本文方法的优越性。
In this paper,an integrated fusion method was presented based on IHS Transform and Fast Discrete Curvelet Transform（FDCT）,which could provide fused images with higher quality than the conventional methods.In the procedure of fusion,it got multi-scale and multi-direction coefficient sets of component I by FDCT.According to the spatial details and spectral characteristics showed in the component,the authors applied the standard variance weighted strategy to adjust the weight between spatial details and spectral characteristics self adaptively,so as to merge spatial and spectral information optimally.QuickBird and WorldView-2 PAN and MS images were used to assess the quality of the proposed method.Visual and statistical analyses proved that the proposed method could significantly overcome the color distortion and improve the fusion quality compared to some existed techniques.
出处
《测绘科学》
CSCD
北大核心
2012年第1期121-124,共4页
Science of Surveying and Mapping
基金
国家973计划资助项目(2010CB731800)
国家863计划重点项目(2009AA121404)
国家自然科学基金资助项目(40801165
10978003)
中央高校基本科研业务费专项资金资助(20082130101000046)
