Automatic Landform Recognition from the Perspective of Watershed Spatial Structure Based on Digital Elevation Models

Landform recognition is one of the most significant aspects geomorphology research, which essential tool for landform classification and understanding geomorphological processes. Watershed object-based a new spot in field recognition. However, relevant studies, quantitative description watershed generally focused on overall terrain features watershed, ignored spatial structure topological relationship, internal mechanism watershed. For first time, we proposed an effective method from perspective structure, separated previous studies that invariably used indices or texture derivatives. The slope spectrum was herein to solve uncertainty issue determination area. Complex network P–N terrain, are two methodologies describe relationship were adopted simulate Then, 13 were, respectively, derived kinds structures. With advanced machine learning algorithm (LightGBM), experiment results showed good comprehensive performances. accuracy achieved 91.67% Kappa coefficient 0.90. By comparing with using derivatives, it better performance robustness. It noted that, terms loess ridge hill, can achieve higher accuracy, may indicate more than methods alleviating confusion landforms whose morphologies complex similar. In addition, LightGBM suitable method, since manifestation their combination other by contrast. Overall, out provided insights recognition; experiments show valuable great potential as well being