改进的小波变换在多源遥感图像融合中的应用

Multi-source remote sensing image fusion based on improved wavelet transform method

在分析方向可调的小波变换与不可分离小波变换的基础上,以全色SPOT和多光谱TM遥感影像为数据源,结合可调小波变换与不可分离小波变换的优点,借助Matlab与ERDASIMAGINE等工具,实现了改进的方向可调的不可分离框架小波变换的遥感影像融合处理。该方法与传统的融合方法相比,具有不可分离小波变换和方向可调性等优点,对图像的多分辨率信息的刻画比较精细,既能保持多光谱图像的光谱信息,又能保持全色图像的空间分辨率特性,从而提高了对影像的解译能力。

SPOT panchromatic images and multi-spectrum TM remote sensing images were fused by an improved wavelet transform adopting the advantages of direction adjusted wavelet transform and inseparable frame wavelet transform with the aid of Matlab and ERDAS IMAGINE. Compared with traditional fusion methods, this method can maintain both the spectrum information of the multi-spectrum images and the spatial resolution characteristic of the full-color images, therefore enhance the ability of phantom interpretation.