Early Warning Landslide Monitoring System

Abstract

Landslide is a serious natural calamity causing hundreds of deaths and damage of properties every year. There are several causes that triggers the landslide in which earthquake and volcanic eruption are the main nature driven causes. Moreover, there are some human activities that cause the destructions, such as, major constructions on hilly area, deforestation and uses of heavy machineries. In India, most of the landslide occurs in monsoon season. The increment in moisture content of the soil increases the driving force of the slope and makes it more prone to slide. In such a case, there is a need of a system that can continuously monitor the landslide prone zone and in case of any emergency, pre-alert the residential people of that area. There are several techniques already existed for this purpose such as DEM based calculation on aerial photography, LIDAR and remote sensing-based technique etc. Most of these techniques are expensive for implementation for a large area and a bit difficult task for early landslide prediction. Therefore, in such condition there exist a need of designing such system which should be least in cost and can alert the people of that area before any unfortunate event. In the present work, a Raspberry Pi 3, Model B+ along with USB webcam-based system has been designed for the continues monitoring of the system based on digital image processing. A soil moisture sensing assembly along with Arduino UNO and YL-69 hygrometer has been fixed on the slope that has been used as the triggering signal for the monitoring system. The prediction of landslide has been done based on the shifting in the position of markers that have been fixed on the slope. The entire technique has been based on MATLAB and Simulink. The detected landslide has been further conveyed to ThingSpeak website and from there to a MATLAB based android application to alert the residential people of that area. In the present work, the experiment has been done on a lab prototype. The concept can be further implemented on an actual landslide slope to check the actual feasibility of the design.