四种用于雷达遥感图像融合的空间分量提取方法

Four methods for spatial component extraction used for SAR image fusion

使用光学多光谱遥感图像对雷达图像进行空间增强,需要从光学多光谱图像中提取其空间分量.提取空间分量的基本方法有两种:主成分分析(PCA)和IHS变换法,PCA法可以对多光谱图像的所有波段进行分析,得到的第一主成分PC-1代表原始多光谱图像的空间信息,而IHS法只能输入多光谱图像的3个波段,经IHS变换得到的Ⅰ分量代表原图像的空间信息;将用两种方法提取的PC-1和Ⅰ分量直接与雷达遥感图像融合,发现PCA法的空间增强效果更为显著,而IHS法的色调保持效果更好.将PC-1和Ⅰ分量分别进行0～255线性拉伸,然后与雷达遥感图像融合,发现二者对雷达图像的空间增强效果都显著增加,而色调保持水平均有所下降,但PCA法的空间增强效果仍然优于IHS法,而IHS法的色调保持水平仍然优于PCA法.该项研究为用户进行雷达图像融合提供了四种可供选择的提取空间分量的方法.

When high-resolution optical multispectral image is used for SAR image fusion, its spatial Component should be extracted first. There are two primary methods for spatial component extraction, principal component analysis （PCA） and intention-hue-saturation （IHS） transform. PCA method transforms the multispectral image into several noncorrelated and independent components and the first component （PC-l） represents the spatial component of the original multispectral image. IHS method transforms only three bands of the original .multispec- tral image into three components of intention, hue and saturation. The spatial components extracted from the optical image respectively using these two methods are directly merged with the SAR image. The result suggests that the spatial component extracted by PCA enhances the spatial features of the SAR image better than that by IHS transform, but distorts the spectral properties of the SAR image compared with that by IHS method does. When the spatial components enhanced by maximum-minimum （0～255） linear stretch are used to fuse with the SAR image, they can both get better spatially enhanced SAR image but with more spectral distortion. But the enhanced spatial component by PCA still spatially enhances the SAR image better than that by IHS method, and the enhanced component by IHS method still preserve the spectral better than that by PCA method. Through the study four methods are proposed for spatial component extraction used for SAR image fusion.