Exploration of Glacial Landforms by Object-Based Image Analysis and Spectral Parameters of Digital Elevation Model

Glacial landforms are a significant element of landscape in many regions Earth. The increasing availability high-resolution digital elevation models (DEMs) provides an opportunity to develop automated methods glacial exploration and classification. In this study, we aimed to: 1) identify based on DEM datasets; 2) determine relevant geomorphometric spectral parameters object-based features for the mapping landforms; 3) accurate workflow landform classification DEM. developed methodology included extraction secondary from DEM, feature selection with Boruta algorithm, image analysis, random forest supervised We applied three study sites: one Svalbard two Poland. It allowed identification six categories landforms: till plains, end moraines, hummocky outwash/glaciolacustrine valleys, kettle holes. majority represented calculated 2-D Fourier analysis. highest performance exhibited up 96% overall accuracy regard groundtruth dataset. This showed that can be identified using novel image-processing complete herein solution transparent generation thematic maps may reproducible transferrable various worldwide.