一种构建多光谱全色波段与小波变换相结合的遥感影像融合方法  

A Remote Sensing Image Fusion Method Based on Multi-spectral Panchromatic Image and Wavelet Transform

作　　者：刘林[1] 田庆久[1] 

机构地区：[1]南京大学国际地球系统科学研究所,南京210093

出　　处：《遥感信息》2008年第6期12-17,共6页Remote Sensing Information

基　　金：国家科技支撑项目(2006BAK30B01)

摘　　要：对于多光谱和全色波段遥感影像融合而言,影像间光谱响应范围的差异是造成融合结果光谱畸变的重要原因。方法通过对遥感影像成像过程中的参数进行合理地近似,将多光谱影像有机地结合为光谱响应范围与全色波段接近的、低分辨率的"多光谱全色波段"。利用特定的小波融合法则,将全色波段中适量的空间信息融入到"多光谱全色波段"中,并根据构建"多光谱全色波段"时各多光谱影像所占的权重,将融入的空间信息逐像元地分解到各个多光谱波段,从而较好地削弱了融合影像间光谱响应范围的差异所造成的光谱畸变。通过对IKONOS遥感影像进行融合实验,验证了本文方法较传统方法具有更高的性能。The difference between remote sensing image spectral response is the main reason that causes distortions while merging multi- spectral and panchromatic images. A novel algorithm is proposed to deal with the problem. In the algorithm, multi-spectral images are combined into a new ＇multi-spectral panchromatic＇ image that is of similar spectral response to that of panchromatic image, by simula- ting some physical characteristics of the imaging system. Accurate detail information is extracted from panchromatic image, and divided into each multi-spectral image depending on the weight of each image in the＇multi-spectral panchromatic＇using wavelet transform. According to the traditional algorithms, the algorithm is good at preserving the spectral characteristics of image objects. And that is demonstrated by the fusion experiment between Ikonos images.

关 键 词：多光谱 遥感 数据融合 小波交换 全色 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]