一种用于土地利用调查的颜色保真的遥感图像融合方法

A RS Image Fusion Method for Retaining Color for Land-use Investigation

快速获取土地利用/土地覆盖信息是土地利用/土地覆盖变化研究的重要内容,RS是快速获取土地利用/土地覆盖现势信息的重要手段之一,由于不同类型的卫星传感器获取同一地区的遥感数据日益增多,常规的遥感图像融合方法如HSI变换法、Brovey变换法和PCA变换法进行融合时,存在不同程度的颜色失真现象,即融合后图像的色彩与多光谱图像存在较大差异,从而影响了地物的识别。分析了Brovey变换法存在颜色失真的原因,提出了一种低通比值融合法,该融合法首先对高几何分辨率的全色波段进行低能滤波,尔后将低分辨率多光谱图像与全色波段图像相乘,再除以滤波后的全色波段图像,便得到融合图像。从辐照的角度证明了该低通比值融合法具备理论基础,并从目视评价、定量分析、分类精度证实了该低通比值融合法优于Brovey变换法,该低通比值融合法是一种能较好地保全低分辨率多光谱图像颜色的融合方法。

Remote sensing technology is often applied to fleetly investigate land use/cover change; the fused production might acquire detailed land use/cover content from multi-source RS data in less cost by the means of merging low resolution multi-spectral data with high resolution panchromatic data, and obtain the synthetical information derived from both of the multi-spectral and panchromatic data. These conventional fusion methods including IHS transform, Brovey transform and principal components transform could merger two optical image data of different resolutions-a high spatial resolution panchromatic image and a low spatial resolution but multi-spectral image.But their production＇s colors differ from that of the low spatial resolution multi-spectral image. We analyzed the Brovey transform, and then proposed a improved method called panchromatic image low-pass filter based ratio fusion （Low-pass RF）, that production＇s color is accordant with that of the multi-spectral image. We discussed its theory mechanism at terms of radiancy. The results of visual interpretation and quantitative analysis to gray variance index showed that the Low-pass RF transform have a capability of retaining color from multi-spectral image. Classification accuracy of the fused image by Low-pass RF transform is 82.45%, more 5.69% higher than the Brovey image. Therefore, Low-pass RF transform is more excellent than Brovey transform.