Abstract:
It is difficult to realize the automatic mapping of mountain geomorphology by geographic grid statistical classification or clustering analysis. In practice, semi-quantitative or manual synthesis methods are generally used to draw geomorphic maps, and the key of mountain geomorphology research is the construction of geomorphic spatial characteristic variables and the integration of spatial grid synthesis, based on which are proposed new geomorphic feature indexes such as uplift height and cutting depth, fold abundance and terrain fragmentation. Meanwhile, based on the traditional statistical ISO clustering results, a secondary aggregation method is constructed by using spatial "core particle" and "texture index". The research shows that: ① the newly proposed geomorphic index can express the degree of upward uplift and downward cutting of mountains with specific values, and measure the degree of fragmentation in horizontal and vertical directions; ② mountain geomorphology is the result of multiple cluster analyses and spatial texture re-aggregation. The geomorphic types in line with the cognitive system can be classified by using nuclear particles and texture. The whole land of China is classified into 6 major regions, 33 sub-major regions, which is consistent with the previous mapping results. The method of twice aggregation classification of geomorphic types realizes the computer automatic classification and mapping of geomorphology, which is suitable for the spatial classification of geoscience categories with texture characteristics.