Inspiration Manual land-record verification is slow and error-prone, especially for handwritten, faded, and multilingual documents. We wanted to make the process faster, safer, and more transparent. What it does The system digitizes land records using OCR, extracts important fields, validates them against Patta and Registration data, checks GIS boundaries, detects mismatches or fraud, and sends risky cases to officers for review. How we built it We used React, TypeScript, Vite, Tailwind CSS, Node.js, Express, tRPC, Drizzle ORM, OCR, NER, GIS, and risk-scoring concepts. Challenges we ran into Poor-quality and handwritten documents Multilingual OCR Limited access to real land-record data Data privacy concerns Accurate fraud and boundary detection Accomplishments we're proud of Built a working land-record validation dashboard Added handwritten document mockup downloads Added OCR testing support Added GIS parcel verification Added risk scoring and officer review Added SDG alignment for SDGs 9, 10, 11, 16, and 17 What we learned We learned that OCR alone is not enough. Reliable land-record validation needs multiple data sources, GIS checks, privacy protection, explainable results, and human officer review. What's next Integrate real OCR and government datasets Support more Indian languages Add PostGIS and satellite verification Generate QR-verifiable certificates Add SMS and WhatsApp notifications Pilot the system with a revenue department
Built With
- artificial-intelligence
- digital-governance
- e-governance
- fraud-detection
- gis
- handwritten-ocr
- land-records
- machine-learning
- node.js
- ocr
- react
- sdg
- smart-india-hackathon
- typescript
Log in or sign up for Devpost to join the conversation.