Development of landslide forecasting system using deep learning

Landslide is one of the most dangerous natural hazard affecting the Himalayan region. The seriousness and effect of landslide disasters are gradually increasing both in terms of magnitude and frequency. Thus, landslide disaster mitigation is one of the top priorities for Indian government. Landslide occurrence should be predicted in advance to effectively mitigate the damage near the landslide area. Therefore, this study proposes a WSN-AI based landslide displacement forecasting architecture that can generate landslide early warning at-least 24 hours ahead of time. landslide displacement forecast AI algorithms are developed using the real-time data i.e., rainfall and displacement data; acquired from the WSN system installed at an actual landslide location. The key objectives of the proposed system is to provide a reliable early warning system that can predict and forecast landslide disaster as well as provide early warning even when the WSN system failed to provide real-time landslide data because of network failure at the location. This has been made possible by integrating weather forecast models of forecast agencies with the developed WSN-AI forecasting model. The result outcomes have proven the efficacy of the model by providing precise landslide displacement forecast of an actual landslide location.