Deep Learning Framework for Landslide Severity Prediction and Susceptibility Mapping

Landslides are a natural hazard that is unpredictable, but we can prevent them. The Landslide Susceptibility Index reduces the uncertainty of living with landslides significantly. Planning and managing landslide-prone areas critical. Using most optimistic deep neural network techniques, proposed work classifies analyses severity landslide. selected experimental study area Kerala’s Idukki district. A total 3363 points were considered for this experiment using historic landslide points, field surveys, literature searches. primary triggering factors slope degree, aspect, elevation (altitude), normalized difference vegetation index (NDVI), distance from road, lithology, rainfall considered. susceptibility map was generated Arc geographic information system (GIS) tool all frequency ratio method, Shannon entropy Relative effect fuzzy logic method. new Deep Neural Network (DNN) framework has been developed multiclass classification prediction zones as low, moderate, high, very high. Existing works only uses statistical methods, used DNN to predict at four different level even semi data over accurately. training learning model Sentinel Satellite images survey. label which Among method Entropy accurate methods achieves 99.16%, accuracy. based 97.08% accuracy, relative 92.72% Fuzzy 86.60%