An Intutive Apporach with CNN Model for Landslide Prediction using Satellite Imagery

Landslides are among the most destructive natural disasters, causing severe damage to infrastructure, human life, and the environment. Traditional landslide prediction models rely on geospatial data, handcrafted feature extraction, and statistical techniques, which often suffer from limited accuracy and generalizability. Recent advancements in artificial intelligence (AI), particularly deep learning, have significantly improved spatial feature extraction and predictive capabilities. This study proposes a Hybrid Convolutional Neural Network (HCNN) model for landslide prediction using highresolution satellite imagery. The HCNN model is designed to efficiently capture intricate spatial patterns and multi-temporal correlations, overcoming the limitations of conventional machine learning (ML) approaches such as Random Forest and XGBoost. The proposed HCNN model integrates a probabilistic framework and a hybrid loss function to enhance prediction accuracy. A dataset consisting of 1,034 labeled landslide and non-landslide images was used, with 800 images for training and 234 for testing. The model achieved an accuracy of 98%, precision of 97%, recall of 98%, and an F1-score of 97.50%, demonstrating its robustness and effectiveness in landslide susceptibility mapping. By leveraging deep learning for automated feature extraction and spatial dependency modeling, the HCNN model significantly improves landslide prediction performance across diverse geographic terrains. The findings highlight the potential of AI-driven solutions for real-time and scalable landslide hazard assessment, addressing key challenges in disaster risk management.