Gis based model for monitoring and predition of landslide susceptibility

Abstract

Landslide disasters tend to occur suddenly at any point in time and causes huge damages to human life and resources. Constant monitoring of mountainous regions and an efficient prediction system is a necessity for saving many lives. Uttarkashi district of the Uttarakhand state has been chosen as the region of study as the place tends to receive frequent landslides. Past data of landslides and its causes has been collected and a model for analyzing and predicting landslide susceptibility is been proposed. Dynamic maps are created with the use of QGIS(open source) software. Real time data of rainfall levels must be captured by the sensors installed at the locations. The system has been trained to predict future possibility of any occurrence of landslide by applying machine learning techniques. K-means clustering algorithm is used for creating clusters defining different rainfall levels and ID3 decision tree learning classification is applied to predict alert level in a susceptible area. Alerts are generated on appearance of any risks. System administrators are able to view alerts in the map and perform other related queries.