DeepLandforms: A Deep Learning Computer Vision Toolset Applied to a Prime Use Case for Mapping Planetary Skylights

Abstract Thematic map creation is a meticulous process that requires several steps to be accomplished regardless of the type produced, from data collection, through exploitation and publication in print, image, GIS format. Examples are geolithological, geomorphological maps which most highest time‐consuming tasks those related discretization single objects. Introducing also interpretative biases because different experience mappers identifying set unique characteristics describe In this setting, Deep Learning Computer Vision techniques could play key role but lack availability complete tools specific for planetary mapping. The aim work develop comprehensive ready‐to‐use landforms mapping based on validated methodologies open‐source libraries. We present DeepLandforms , first pre‐release toolset landform using includes all components preparation, model training, monitoring, inference. DeepLandforms, users have full access workflow control over processes involved, granting customization capabilities. order validate applicability our tool, we results achieved science case sinkhole‐like Mars, as example can lead us into multiple diverse future applications.