K-L变换是将图像的多维信息进行数学正交矩阵变换,从而将原始图像的多个指标简化成几个少数综合指标的一种方法。多极化遥感图像经过K-L变换后,保留了原始图像辐射信息主要成分,从而降低图像的维数,节省图像存储空间,大大提高图像检索和查询效率。 K-L transform is a method that transforms the multidimensional information of image into mathematical orthogonal matrix, thus simplifying the multiple indicators of the original image into several minority comprehensive indicators. After K-L transform, the multi-polarization remote sensing image retains the main components of the original image radiance information, thus reducing the image dimension, saving image storage space and greatly improving the efficiency of image retrieval and query.