改进的经验统计地形校正模型及其应用

Improved statistic-empirical topographic correction model and its application

在地形起伏较大的地区,不同地表部位所接收到的太阳辐射会有所不同,使得对地物的识别和分类变得更加困难。为了提高遥感图像的分类精度,就必须对图像上的地形信息进行校正。传统的经验统计地形校正模型是一种基于统计学的经验校正模型,它将坡面像元接收的辐射通过对应的线性回归关系旋转到水平位置,从而达到校正的目的。但是,如果坡面角度不同(即坡度有差异),上述校正方法必将对校正精度产生一定的影响。为此,该文通过坡度分级对原经验统计地形校正模型进行了改进:首先对研究区进行坡度分级,然后通过对应的线性回归关系将不同的坡度级别分别旋转到水平位置。以ASTER图像为例,分别用原模型和改进模型进行地形校正的结果表明:改进的经验统计地形校正模型既能够有效去除地形的影响,使得同类地物在同一图像上波谱信息一致;同时又能很好地保持地物本身的波谱特征,校正效果更好。

In the relief area, solar radiation is different in different parts of the surface, resulting in difficulty in identification and classification of features. Topographic correction thus becomes a must for improving the accuracy of remote sensing image classification. The traditional statistic -empirical terrain correction model is an empirical correction model based on statistics, which rotates solar radiation received by pixels on the slope to a horizontal position by the corresponding linear regression relationship so as to achieve the purpose of correction. However, the different angles of the slope, i. e. , slope differences, have a certain impact on its correction accuracy. This paper improves the original model by slope grading： first of all, the study area is classified into different slope grades, then different slope grades are rotated to a horizontal position by the corresponding linear regression relationship, and the original model and the improved model are used to correct the study area based on ASTER image. The results show that the improved statistic - empirical terrain correction model can both remove the influence of terrain, making the same features have the same spectral information in an image, and keep the spectral characteristics of the feature itself, thus obtaining better correction effects.