数学形态学在高分辨率图像土地覆盖分类中的应用

高分辨率影像可以清楚地显示树木房屋等地物细节,但是用基于像元的分类很难将这些细碎的地物融合为通常的土地覆盖分类图。首先对高分辨率影像进行有监分类,再根据有监分类的结果对不同地类进行二值化,运用数学形态学对各地类二值图进行处理,从而得到地类的边界图。通过多组实验表明,数学形态学运用于高分辨率图像土地覆盖分类既保留了图像的地类特征,也快速得到了不同地类的边界。

The Application of Mathematical Morphology in High - Resolution Image of Land Cover Classifi- cation

The high- resolution image can shows the details of trees and houses clearly, but the classification based on the pixel cannot fuse the finely surface features and make the land use classification map. The first step of this method is super- vised classification of the high - resolution image, and then considering the result of the supervised classification, we make the classification to the binarization. The binarization are processed by the mathematical morphology and get the boundary figure. Experiments show that the method can use the mathematical morphology keep the class features of the image, and get the accurate boundary rapidly.