Skip to content
SIH Buddyby Ganeev Singh
Dev
All problem statements
SIH26070Worth consideringacceptance 3/5

To develop an Artificial Intelligence (AI) / Machine Learning (ML) based system for identification, classification, and prediction of different tropical cyclone patterns using multi-source satellite data.

Ministry of Earth Sciences (MoES) · Smart Education · Software

Excellent labelled data and a genuinely dramatic demo, but the description is just the title and it bundles three problems — pick intensity classification, do it against published baselines, and treat track prediction as a stretch rather than a claim.

What it actually is

Forecasters judge how strong a cyclone is largely by looking at its satellite picture and matching the cloud pattern to a reference scheme, which is expert work and somewhat subjective. The ask is a system that reads the satellite imagery itself and identifies the cyclone, classifies its pattern and intensity, and says where it is going.

What to build

A three-stage system over satellite imagery: detection and centre-fixing locating a cyclonic system and its centre in a full-disk image; a classification stage assigning intensity and cloud pattern type against an established grading scheme, trained on imagery paired with best-track intensity so the labels come from an authoritative archive rather than your own reading; and a short-range track and intensity projection compared honestly against both a persistence baseline and the operational forecast for the same case, presented as a case viewer where a stored historical cyclone can be stepped through frame by frame with the model's classification and projected track shown against what actually happened.

Smallest thing that wins the room

Step through a real historical cyclone frame by frame, showing your intensity classification tracking the best-track record it never saw, and the projected path against the track the storm actually took.

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.

Moderate150–360 teams expectedroughly 1 in 130–301 wins it

Quieter than 30% of the 226 · #158 of 226 by expected field

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

Why: central ministry statements sat below the average.

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

  • CNN intensity regression on geostationary IR imagery
  • IBTrACS best-track intensity labels
  • INSAT-3D / Himawari full-disk imagery ingestion
  • cyclone centre-fixing via spiral pattern matching
  • persistence and climatology baseline comparison
  • case replay viewer with track overlay
  • Tropical meteorology
  • Satellite image analysis
  • Severe weather forecasting

Prior art to read before you start

cyclone intensity estimation from satellite imagery · storm track prediction and validation · pattern classification against expert grading scheme

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.