The Application Of Remote Sensing Image Fusion Research

With the development of remote sensing technology more and moremultispectral, multiresolution and multitemporal image data of the saxme scene comefrom different sensors, the process of fusing different remote sensing data, extractingvaluable information applied to resource investigation, environment monitoring andmany field becomes imminent.This paper mainly focuses on the research of algorithms of remote sensingimages: the methods of polynomial, orthogonal polynomials, weighted orthogonalpolynomials, and surface fitting based on spline function are adopted to spatialcoregistration of SPOT panchromatic and Radarsat images. Then the speeds andprecision of these methods are contrasted; The principal component transformmethod for fusing several correlative Synthetic Aperture Radar (SAR) images basedon maximizing the Signal-to-Noise (SNR) is discussed, then a new method forclassifications' fusion of SAR image based on edge detection is provided. Thismethod is simple and manageable, Which not only preserves the edge of the image,avoiding the problem of edge blurring in image classification, but also be effectivein region division; The methods of Intensity-Hue-Saturation (IHS) transform,wavelet transform and adjusted wavelet transform for the fusion of SPOTpanchromatic and multispectral imagery are also discussed and implemented, finallywe provide the experimental results and contrast the reserving spectral characteristicof panchromatic image and spatial characteristic of multispectral image after usingthese fusing methods, and obtain useful results.