Unsupervised deep learning for environmental risk monitoring: Landslide detection from multi-resolution remote sensing imagery

Science of Remote Sensing 2026 Open access

Read full text
Authors: Hejar Shahabi, Saeid Homayouni, Omid Ghorbanzadeh, Didier Perret, Bernard Giroux, Jimmy Poulin
DOI: 10.1016/j.srs.2026.100375
Cited by: 1
References: 46

Abstract

This study assesses three self-supervised learning (SSL) models including SimCLR, SwAV, and DINOv2 for landslide segmentation using Sentinel-2 and Landsat 8 imagery. The evaluation leverages the Landslide4Sense dataset and a custom Canadian dataset encompassing Yukon, the Northwest Territories, British Columbia, and Northern Quebec. Embedding analysis showed SSL models effectively distinguished landslide features, with DINOv2 yielding high similarity (0.56–0.842) for landslide images and low/negative scores (

Fields of Study & Subjects

Landslides and related hazards Remote-Sensing Image Classification Synthetic Aperture Radar (SAR) Applications and Techniques

Publication details

Journal / source
Science of Remote Sensing
Publisher
Elsevier BV
Volume
13
Pages
100375–100375
Publication date
29 January 2026
Language
EN