Python workflow for UAV-derived DSM coregistration and automated crevasse mapping by Uslular et al.

This software archive contains the unified Python workflow used as supplementary material for the submitted manuscript (to Earth Science Informatics) “Automated Crevasse Mapping Using UAV-Derived Digital Surface Models: A Case Study from Shoesmith Glacier, Horseshoe Island, Antarctic Peninsula” by Uslular et al. The workflow supports (1) rasterisation of a glacier or ice-outline polygon to the grid of a reference digital surface model (DSM), (2) optional Nuth and Kaab coregistration of a moving DSM to a reference DSM using a stable-terrain mask, and (3) automated detection of linear crevasse depressions through NaN-aware Gaussian detrending, multi-scale and multi-orientation directional black top-hat filtering, hysteresis thresholding, morphological cleaning, and component-level shape filtering. The script additionally estimates mapped crevasse area, morphology-based depth, and raster-integrated volume, and writes processing parameters and summary statistics to JSON files. All inputs, outputs, and processing parameters are supplied through a command-line interface. The archive includes the Python script, a recommended Conda/Mamba environment, usage documentation, and an open-source licence. Input UAV DSMs and vector/raster masks are not included and can be freely provided by the authors.