Mapping Flood Damage From Space

Inspiration

Floods can damage roads, buildings, and bridges while leaving entire communities isolated. During a disaster, getting reliable information about affected areas can be difficult, especially when ground access is limited. This inspired us to build Mapping Flood Damage from Space, a project that uses satellite imagery and geospatial analysis to help assess potential flood impacts.

What it does

Our project provides an interactive dashboard for preliminary flood-impact assessment. Users can define an Area of Interest using latitude and longitude coordinates and specify an event date.

The system analyses satellite-derived changes, maps potentially affected buildings and roads using OpenStreetMap data, and evaluates road-network connectivity to identify potentially isolated settlements. The dashboard presents the results through interactive map layers, impact metrics, validation indicators, and a structured Situation Report.

How we built it

We built a geospatial processing pipeline using Python and an interactive web dashboard. The workflow integrates multi-temporal Sentinel-1 Synthetic Aperture Radar imagery, optional Sentinel-2 optical imagery, OpenStreetMap infrastructure data, and Copernicus Digital Elevation Model data.

The pipeline connects satellite data acquisition, change detection, infrastructure impact estimation, and settlement connectivity analysis. The dashboard brings these outputs together so users can explore the geographical distribution of potential impacts in one place.

Challenges we ran into

One major challenge was working with satellite data that varies in availability and quality across locations and dates. Optical imagery can be affected by cloud cover, while radar changes can result from factors other than flooding.

Integrating different geospatial datasets also required careful handling of coordinates, acquisition dates, and infrastructure information. Another challenge was distinguishing pipeline completion from scientifically validated results. We therefore made validation status and analytical limitations visible in the dashboard rather than treating every detected change as confirmed damage.

Accomplishments that we're proud of

We are proud to have brought satellite-based change detection, infrastructure impact estimation, and settlement connectivity analysis into a single interactive dashboard.

The system includes coordinate-based AOI inputs, map-based visualization, impact summaries, and a Situation Report interface. We also incorporated validation indicators that communicate when detected radar changes have not been confirmed as flood damage.

What we learned

We learned that meaningful flood assessment involves much more than detecting changes in satellite imagery. Data quality, temporal consistency, coordinate systems, infrastructure completeness, and independent validation all influence the reliability of the results.

We also learned the importance of communicating uncertainty clearly. A detected change is a signal for further assessment, not automatic proof of damage.

What's next for Mapping Flood Damage From Space

Our next steps are to strengthen independent validation, improve handling of missing or unavailable source data, and test the pipeline across more geographical areas and event dates.

We also aim to improve the assessment of road and bridge connectivity, enhance the reliability of settlement-isolation estimates, and make the Situation Report more useful for communicating preliminary findings.

Our long-term goal is to build a geographically adaptable and transparent flood-assessment system that helps turn satellite observations into useful information for disaster-impact analysis.

Built With

copernicus-dem
css
data
fastapi
flood-mapping
gdal
geopandas
geospatial-analysis
html
javascript
leaflet.js
numpy
openstreetmap
python
rasterio
remote-sensing
rest-api
sentinel-1-sar
sentinel-2
shapely

Submitted to

Multimodal AI Hackathon 2026

Created by

Contributed to the development of the Flood Damage Mapping from Space project by supporting research, geospatial data analysis, system testing, and dashboard evaluation. Helped review the pipeline outputs, identify potential issues, and improve the presentation of flood-impact estimates and settlement connectivity results.

Priyanshu Kumar
I led the development and integration of Flood Damage Mapping from Space, focusing on building a coordinate-driven geospatial pipeline for preliminary flood-impact assessment.

My work involved designing the overall workflow that connects satellite data acquisition, radar-based change detection, infrastructure impact analysis, and road-network connectivity assessment. I worked on integrating multi-temporal Sentinel-1 SAR imagery, optional Sentinel-2 optical data, OpenStreetMap infrastructure data, and Copernicus DEM into a unified analysis system.

I contributed to developing the interactive dashboard, which allows users to define an Area of Interest through latitude and longitude coordinates, configure event parameters, visualize spatial results, and explore different map layers. I also worked on presenting key impact metrics, pipeline status, validation indicators, and a structured Situation Report.

Another important part of my contribution was evaluating the reliability of the analysis workflow. This included examining satellite acquisition dates, interpreting radar-change outputs, reviewing infrastructure impact estimates, and identifying cases where detected changes could not be treated as confirmed flood damage. I focused on making limitations and validation failures visible instead of presenting preliminary estimates as verified results.

I also worked on integrating road-network connectivity analysis to identify potentially isolated settlements and on improving the system's handling of geospatial data and pipeline outputs.

Through this project, I gained practical experience in remote sensing, geospatial processing, data integration, full-stack development, and the challenges of building reliable disaster-assessment systems using real-world data.

Harsh KUMAR