地形校正的目的是消除太阳光照对不规则地面地物辐射值的影响。这种影响会使相似植被类型地物的辐射值发生很大的变化。因此,在地形复杂的地区,地形校正是影像预处理的一个重要步骤。传统的基于单像素的地形校正方法,虽然减小了辐射值的变化,但在太阳入射角低的地区常常出现校正过度的情况。针对这种误差进行分析,提出一种考虑了空间相关性的校正算法,并且利用鄂西地区的Landsat7卫星影像进行的试验证明,该算法优于传统的地形校正模型。
The topographic correction refers to the compensation of the different solar illuminations due to the irregular shape of the terrain. The effect causes a high variation in the reflectance response for similar vegetation types. Therefore, the process of topographic normalization may be critical in areas of rough terrain, as a preliminary step to the multispectral image classification. The conventional pixel-based topographic correction methods usually overestimate the radiation in low sun incidence angle, although reducing the variation in averages. A spatial contextual approach is proposed after analyzing these errors. The experiments using a Landsat TM image and DEM in west of Hubei Province have proved that the results of the contextual method are better than that of the conventional correction methods.