Research-driven projects spanning satellite radar, deep learning, and spatial analytics, from landslide early-warning to marine-pollution monitoring.
A flagship programme developing a large-scale landslide hazard-assessment scheme based on time-series analysis of satellite radar interferometry (InSAR). The approach detects slow-moving ground deformation across mountainous terrain to support early-warning systems and disaster-risk reduction, built on methods developed with leading international research partners.
Glacier-dynamics research and Glacial Lake Outburst Flood (GLOF) risk assessment across High Mountain Asia, including logistic-regression risk models and a glacial-lake mapping approach using a remote-sensing Geo-Foundation Model, developed with international collaborators and published in a leading Earth-observation journal.
Detection of marine oil spills from synthetic-aperture-radar (SAR) imagery using deep learning, the subject of an invited tutorial at APSAR 2025, Japan. Enables rapid, automated monitoring of marine-pollution events over large ocean areas.
Deep learning on Sentinel-1 SAR imagery for automated ship detection using the YOLO-v8 model, enabling maritime monitoring and surveillance from freely available satellite radar data, without dependence on optical conditions or daylight.
A spatio-temporal study of annual NO₂ concentration and population exposure, quantifying air-quality patterns and the human exposure burden across a region, demonstrating our spatial-analytics and public-health application capabilities using satellite atmospheric data.