Supervised Learning and IoT Sensor-Based Landslide Detection and Early Warning System: Design, Implementation, and Evaluation

DOI

10.2991/978-94-6239-791-0_27How to use a DOI?

Keywords

Landslide Detection; U-Net Segmentation; IoT Sensor; Fusion; Early Warning System; Satellite Imagery

Abstract

Landslides present a significant natural disaster that leads to considerable damage to physical infrastructure and human lives, especially in hilly areas. In this work, a novel landslide detection and early warning system is designed that uses the combination of U-Net satellite image segmentation and IoT sensors’ monitoring. The U-Net architecture was used for segmentation based on the Landslide4Sense data and multispectral Sentinel-2 satellite images with the application of the loss function that contains Binary Cross-Entropy and Dice loss functions. Experimental results on the test set show that our approach yields pixelwise accuracy of 98.3%, F1-score of 75.6%, and Intersection over Union of 60.9%. The inconsistency in these values can be explained by the strong class imbalance between landslide and non-landslide classes. The IoT subsystem consists of various types of sensors including soil moisture sensors, rain sensors, vibrational and tilt sensors integrated by Arduino Mega 2560. These sensors provide data that can be used to detect landslides at ground level and to create the final warning message in collaboration with satellite data using fusion that consists of three types of warnings: HIGH RISK, MONITOR, and GROUND ALERT. Successful operation was tested with a prototype, and results have shown good accuracy of land detection with decreased numbers of false detections.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Ayush Ingrulkar
AU  - Nishant Pawar
AU  - Vishwa Shende
AU  - Aditya Ghadge
AU  - Ramesh Lavhe
AU  - Ashok Kalal
PY  - 2026
DA  - 2026/10/09
TI  - Supervised Learning and IoT Sensor-Based Landslide Detection and Early Warning System: Design, Implementation, and Evaluation
BT  - Proceedings of the International Conference on Artificial Intelligence and Machine Learning for Agro-Ecosystems (ICAIML-AE 2026)
PB  - Atlantis Press
SP  - 370
EP  - 381
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6239-791-0_27
DO  - 10.2991/978-94-6239-791-0_27
ID  - Ingrulkar2026
ER  -