Skip to content
SIH Buddyby Ganeev Singh
Dev
All problem statements
SIH26162Strong pickacceptance 4/5

AI-Based Detection and Classification of Industrial Fires and Persistent Thermal Sources Using NASA FIRMS, OSM & Satellite Data

National Technical Research Organisation (NTRO) · Miscellaneous · Software

A tractable, well-scoped geospatial problem with all data free and linked — lean on the temporal-persistence signal to separate flares from fires, and be honest that FIRMS resolution limits facility-level attribution.

Data: firms.modaps.eosdis.nasa.gov/map

What it actually is

Satellites detect thermal hotspots from space, but systems like NASA FIRMS just report a hotspot without saying what it is — an industrial fire, a gas flare, crop burning, a wildfire or mining. The ask is a system that classifies these thermal anomalies by integrating the fire data with land-cover, industrial-facility databases and satellite imagery, so an industrial fire can be told apart from a forest fire.

What to build

A geospatial classification system that ingests NASA FIRMS thermal anomaly detections and, for each, pulls contextual layers — OpenStreetMap and industrial-facility locations, land-cover classification, and satellite imagery of the location — then classifies the anomaly into categories such as industrial fire, gas flare, thermal power plant, agricultural burning, mining or wildfire, distinguishing persistent thermal sources like flares from transient accidental fires by their temporal behaviour, and presenting the classified anomalies as an overlay on a GIS map with the ability to monitor persistent sources over time.

Smallest thing that wins the room

Load a region's FIRMS detections and show the system classifying them — a persistent hotspot at a known refinery location labelled as a gas flare, a cluster over farmland in season labelled agricultural burning, and a hotspot in forest labelled wildfire — each justified by its land-cover and facility context on the map.

How crowded this one gets

A guess, projected from the 2025 statements — the last year where both the submission counts and the winners were published.

Moderate150–340 teams expectedroughly 1 in 125–289 wins it

Quieter than 37% of the 226 · #143 of 226 by expected field

A normal-sized field. Your idea has to be good, not miraculous.

Why: defence, intelligence and space bodies drew small fields.

This is a guess, not a fact

Nobody has published 2026’s numbers yet. This is an analysed estimate from last year’s pattern, so please do not take it as the truth — check the live counter on the SIH portal before you decide anything. The range covers the middle half of likely outcomes, so one statement in two lands outside it. Entry closes at 500 ideas per statement, so no range goes past that — a statement that reaches the cap fills and shuts rather than drawing an unlimited crowd. The model reads only three things a team can see before choosing — software or hardware, the theme, and what kind of body posted it — and those explain about a quarter of the variation in last year’s field sizes (R² 0.25 on held-out statements). Trust the band more than the number, and the ordering more than either. It cannot see how good your idea is, which is the part that actually decides it.

The scores

The number is the shorthand. The line under it is the reason.

What you will be writing

  • NASA FIRMS thermal anomaly ingestion
  • OpenStreetMap industrial facility data (Overpass)
  • Land-cover classification (ESA WorldCover)
  • Temporal persistence analysis for flare vs fire
  • Sentinel-2 imagery context
  • PostGIS + Leaflet GIS overlay
  • Thermal remote sensing
  • Industrial monitoring
  • Geospatial classification

Prior art to read before you start

thermal anomaly classification · industrial fire detection from satellite · persistent thermal source monitoring

Analysed by Claude Opus. Every score above is a judgment call with its reasoning attached — kindly cross-check this against the official statement on the SIH portal before your team commits to it.