基于PCA变换的多元遥感影像融合方法研究

基于PCA变换的融合方法是用于多波段遥感图像的第一主分量具有相同均值和方差的高分辨率遥感图像直接代替多波段遥感图像的第一主分量,这样会损失部分多波段遥感图像第一主分量中反映光谱特性的有用信息,使得融合结果图像的光谱分辨率受到较大影响。因此,在PCA变换融合算法基础上引入小波包变换融合算法,保留多波段遥感图像第一主分量中一些反映光谱特性的有用信息,提高融合后遥感图像的效果。

Fusion Method of Multi-source Remote-sensing Image Based on PCA Transform

The fusion method based on PCA transform is to use high resolution remote sensing image which has similar value and variance to the first principal component of multiband remote sensing image to replace it. However, this method will lose some useful information which reflects the spectral characteristics in the first principal component of multiband remote sensing image. What＇s more, the spectral resolution of the result image fused will be affected greatly. Therefore, this paper brings in the fusion algorithm of wavelet packet transform based on the fusion algorithm of PCA transform. Some useful information which reflects the spectral characteristics in the first principal component of multi - band remote sensing image will be kept, and the effect of the remote sensing image fused will be improved.