CNG Photo Data Automation Portal

I run a small transport business with five Tata vehicles. After every CNG refill, my drivers already snap a photo of both the dispenser display and the vehicle’s odometer. What I need now is a streamlined web portal that lets them upload those images directly—one vehicle per login or page—so I no longer have to dig through WhatsApp chats.

Core workflow
• Driver opens the portal on a phone, selects the vehicle, uploads the refill photo(s), and hits submit.
• Behind the scenes your code performs automatic data extraction (OCR/AI vision) to pick up three items from each image: CNG filled amount, current meter reading, and the date / time shown or embedded in the file.
• The extracted numbers are saved to a database, matched to the correct vehicle, and used to calculate the day’s distance-travelled-per-kg (or rupee) average.
• Every midnight the system produces an Excel file that lists, for each vehicle and for the fleet total, the CNG added, kilometres driven, and resulting average. A running monthly sheet is fine if that simplifies things.

Deliverables
1. Responsive web portal (simple UI, mobile-friendly).
2. Server-side pipeline for automatic text/number detection (e.g., Python with Tesseract/OpenCV or an equivalent cloud vision API).
3. Database schema with five vehicle tables or a single table with a vehicle ID field.
4. Scheduler that generates and emails / stores a daily Excel report.
5. Brief setup guide so I can host it on my own VPS later.

Acceptance criteria
• Minimum 90 % accuracy on clearly shot dispenser and odometer photos.
• Excel report ready by 7 a.m. IST each day.
• Average calculation matches a manual sample within ±1 %.
• Portal tested on Android Chrome and desktop Chrome/Edge.

If any part of this flow needs tweaking—say, adding an admin dashboard or exporting in Google Sheets instead of Excel—let me know; I’m flexible as long as the data lands in a neat daily report without manual typing.