What It Studies
Probable mosquito breeding-prone micro-areas using SAR-readable wetness, seasonal surface change, vegetation and land-surface structure, terrain and human-exposure signals.
Saptrishi Research
Building NISAR-ready public-health intelligence.
One of Saptrishi’s prime research projects is a NISAR-ready GeoAI model for community medicine, designed to help teams prioritize mosquito habitat inspection during monsoon, cloudy and post-rainfall windows.
Prime Research Project
NISAR-Ready GeoAI for Community Medicine
A NISAR-aligned mosquito habitat and human-exposure surveillance model for targeted vector-control planning.
Project paper for community medicine review
NISAR-Ready Vector Intelligence for Community Medicine
Prepared by Abhishek Dubey, this flagship research concept is designed to evolve with NISAR-style L-band and S-band SAR inputs while keeping the operational goal simple: help field teams decide where to inspect first after rainfall, complaints or seasonal vector-control alerts.
Prototype output
28/100 habitat
39/100 exposure
Localized hotspot reached 68/100 due to wetness, low slope and proximity to buildings.
Why It Matters
Community medicine and municipal teams often know a ward is at risk, but still need to decide which lanes, drains, plots or peri-domestic spaces to inspect first.
How It Helps
The model separates habitat suitability from human-exposure priority, so field action can focus on breeding-prone conditions closer to likely human habitation.
Prototype Output
From NISAR-ready context to inspection-priority surface.
These prototype visuals show the decision workflow: satellite context is translated into adjusted exposure surfaces and local inspection points. The NISAR direction strengthens this by bringing all-weather SAR monitoring into the model roadmap.
These maps demonstrate the inspection-priority workflow. They are not presented as NISAR-derived disease maps and do not confirm larvae, adult mosquitoes or disease transmission without field validation.
Concept Understanding
See the idea behind the research.
This concept video introduces the thinking behind the work: using AI, NISAR-ready geospatial signals and field-oriented workflows to make public-health inspection more targeted. The page below expands that idea into the community medicine vector-surveillance model.
- AI supports prioritization, not final public-health decisions.
- NISAR-style SAR can support inspection planning through cloud cover.
- Dual-band radar can strengthen wetness and land-surface change signals.
- Human proximity makes a habitat signal more operationally relevant.
Research Basis
How NISAR aids this research direction.
NISAR is valuable for this model because Synthetic Aperture Radar can observe land surfaces day or night and through cloud cover. For vector-control planning, that matters during monsoon and post-rainfall periods, when wetness, vegetation structure and surface-change signals can guide smarter inspection planning.
NISAR L-band SARdeeper landscape response for vegetation structure, wetlands and persistent moisture patterns
NISAR S-band SARsurface-level response that can complement wet terrain and built-up contrast
All-weather imagingsupports monitoring when cloud cover limits ordinary satellite views
Repeat observationhelps compare rainfall windows, seasonal change and hotspot persistence
Polarimetric dataadds richer surface information for future model calibration
Field-priority layerconverts SAR-informed signals into inspection cards, not disease claims
Validation Roadmap
From prototype to field-ready evidence.
The model is intentionally positioned as a NISAR-ready inspection-priority tool today. Field validation is required before it can be described as a disease early-warning model.
Entomological validation
Inspect high, moderate, watch and low score points for stagnant water, larvae, pupae, adult mosquito evidence and habitat type.
Household and container survey
Compare Aedes-oriented outputs with House Index, Container Index and Breteau Index where field surveys are available.
Disease data linkage
Use ward or locality-level case data with rainfall and temperature lags before calling the system an early-warning model.
Scientific and ethical guardrails
- This is not a confirmed disease prediction model.
- The map should trigger field inspection, not automatic chemical application.
- Larvicide or insecticide decisions should follow official public-health protocol.
- If disease case data is added later, it should be aggregated and privacy-safe.
Research Internships
Research interns can apply to work on Saptrishi projects.
We welcome serious students and early researchers who want exposure to applied AI, public-health data, geospatial analysis and field-oriented product thinking. Interns may contribute to literature review, dataset preparation, validation planning, map interpretation, documentation and prototype testing.
Good fit for
- Medical, public health or community medicine students
- Engineering, data science or GIS learners
- Students interested in AI for social impact
- Applicants who can write clearly and work with evidence
Research Application
Apply for research internship or collaboration.
Share your background and the area you want to work on. This form is for research internships, academic discussions and field-collaboration inquiries, not student test enrollment.
Interested in academic or field collaboration?
We are open to community medicine review, pilot planning, research internships and collaboration with public-health teams, medical colleges and municipal vector-control units.