“…We emphasize that the zoning of interventions within erosions was only possible with the vectorization of the drainage lines through visual inspection. Although the current trend of applying machine learning algorithms dominates publications on the spatialization of erosive processes (Ghorbanzadeh et al, 2020) or the detection of ravine edges (Li et al, 2021), here we highlight the importance of human expertise on RS images analysis. In this sense, our algebra of selection and exclusion of false positives can be useful as an attribute for achieving more robust models (Minella et al, 2010), or even for integrating variables in GIS, such as distance to roads, lithology, slope curvatures, and topographic indices (Zabihi et al, 2018;Pourghasemi et al, 2020;Amiri and Pourghasemi, 2020), especially regarding the locations suitable for palisades, whose studies are still scarce.…”