Classical and quantum machine-learning framework for detecting and mapping natural hydrogen prospectivity in Kazakhstan using Sentinel-2 satellite imagery.

GRASS addon to download Sentinel-2 scenes by S2 scene ID using eodag.

We are using Sentinel-2 satellite imagery and a specialized U-Net deep learning model to detect changes in landscapes before and after flood events. Using the OMBRIA dataset, the model reliably identifies flooded areas to support disaster management and response efforts.

Extension to read from Earth Observation data archives

Hybrid ML + GIS pipeline for wildfire vegetation risk: U-Net segmentation on Sentinel-2 with Dynamic World labels, fused with line-distance and slope to produce a tunable risk raster. Validated across three external California AOIs.

Data analysis and object detection to assess Hurricane Maria's impact using NDVI analysis and YOLO modeling for disaster relief planning.

Multimodal Accessibility and Place Profiling Engine for arbitrary Regions

An analysis of seasonal NDVI changes using the imagery from Sentinel-2 mission. Includes plots and stats for insights into vegetation health.