Rapid mapping of landslides in the Western Ghats (India) triggered by 2018 extreme monsoon rainfall using a deep learning approach

Abstract Rainfall-induced landslide inventories can be compiled using remote sensing and topographical data, gathered either traditional or semi-automatic supervised methods. In this study, we used the PlanetScope imagery deep learning convolution neural networks (CNNs) to map 2018 rainfall-induced landslides in Kodagu district of Karnataka state Western Ghats India. We a fourfold cross-validation (CV) select training testing data remove any random results model. Topographic slope was as auxiliary information increase performance The resulting inventory map, created with spectral information, reduces false positives, which helps distinguish areas from other similar features such barren lands riverbeds. However, while including did not true overall accuracy higher compared only train mean accuracies correctly classified values were 65.5% when optical increased 78% use data. methodology presented research applied landslide-prone regions, support hazard mitigation regions.