← workDomain
Forest & Governance
AI-driven tools and research for forest ecosystem management and community forest governance, spanning species mapping, biomass estimation, and local-language AI agents that support forest-dependent communities.
9applications
35technologies
Applications
applicationshipped
Monocular metrology · instance segmentation
Metric tree diameter from a single RGB frame — trunk segmentation plus pinhole projection, no reference object or specialised hardware.
- 0.953
- Mask mAP50
- 130 FPS
- Inference
stack
- YOLOv8s-seg
- Grounding DINO
- SAM 2
- Pinhole projection
- +1
~/work/monodbhDataset↗open→ applicationshipped
Field geodata capture · spatial modelling
Geospatial data platform: an offline-first Android capture app and a web platform, running the inventory and boundary mapping pipelines.
- 500K+
- Field data points
- 7+
- States deployed
stack
- Android · offline-first
- Kalman-filtered GPS
- Geofencing
- Sentinel-1 / 2
- Species distribution models
- Web platform
- +3
~/work/ncountLaunch↗open→ applicationin build
Biomass estimation · carbon stocks
Above-ground biomass from DBH, species, and height via species-specific allometric equations.
- DBH · spp · H
- Inputs
- Allometric
- Method
stack
- Allometric modelling
- MonoDBH
- BarkVisionAI
~/work/agb-estimationopen→
applicationshipped
Blockchain marketplace · forest produce
Blockchain-powered marketplace app digitising the forest-produce trade for women-led producer companies — from KYC onboarding to on-chain receipts.
- Blockchain
- Ledger
- QR · PDF · print
- Receipts
stack
- Android
- Blockchain ledger
- QR verification
- Bluetooth thermal printing
- PDF receipts
- +2
~/work/arthasakhiGet app↗open→ applicationshipped
Species recognition · research dataset
CNN bark-image species ID and large-scale benchmark dataset, published in Nature Scientific Data.
- 156,001
- Images
- Nature Sci. Data
- Venue
stack
- CNN
- ResNet50
- Transfer learning
- Image classification
- +1
~/work/barkvisionaiPaper↗open→ applicationin build
Research tooling · RAG + interviews
Van Gyan knowledge RAG and Van Samvad interview chatbot for community forest research.
- 2
- Components
- JFM · FRA · PESA
- Focus
stack
- LangGraph
- RAG
- Vector DB
- Semantic search
- +1
~/work/collaborative-forest-governanceopen→
applicationshipped
Multi-sensor fusion · species distribution
Species distribution modelling on fused SAR + optical satellite stacks — landscape-scale vegetation layers, deployed across four states.
- 4
- States deployed
- S1 + S2
- Sensor fusion
stack
- Sentinel-1 SAR
- Sentinel-2
- Random Forest
- SDM
- Raster pipelines
- +2
~/work/inventory-mappingopen→
applicationshipped
Forest rights · claim workflow
Claim filing and tracking system for Community Forest Resource rights under the Forest Rights Act — prefilled claims, stage-by-stage tracking, and offline use.
- CFR · FRA 2006
- Claims
- JH · AP
- Deployments
stack
- Claim workflow engine
- Prefilled templates
- Real-time notifications
- Document repository
- Offline-first
- +2
~/work/ratifiLaunch↗open→ applicationin build
Speech pipeline · LLM synthesis
Indic-language voice agent — multi-speaker transcription and diarisation fused with geospatial field data, answered back as speech.
- 20
- Speakers diarised
- Speech
- Output
stack
- Whisper
- Sarvam STT / TTS
- pyannote
- Claude
- ECS Fargate
- +2
~/work/vansaathi-commforLaunch↗open→ applicationin build
Knowledge RAG · literature discovery
Vector search over forest-governance literature, field reports, and policy documents.
- RAG
- Mode
- Collab tools
- Parent
stack
- RAG
- Semantic search
- Vector DB
- TTS
- +1
~/work/van-gyanopen→
applicationin build
Multilingual interview agent
Multilingual interview chatbot for stakeholders in JFM, FRA, PESA, and related institutions.
- Speech + text
- Interface
- Collab tools
- Parent
stack
- LangGraph
- Sarvam AI
- Whisper
- RAG
- +1
~/work/van-samvadopen→