QR and OCR Metal Surface Text Analyzer

Заказчик: AI | Опубликовано: 28.09.2025

I need a Windows-ready prototype that, given a product image, returns JSON with: - metal_text (embossed first-line text on the metal surface) - qr_above_text (single line printed text just above the QR label) - qr_data (decoded QR payload) - status ("OK" if metal_text == qr_above_text (case-insensitive) or metal_text is substring of qr_data; otherwise "NOT OK") - qr_all (array of detected QR objects with rect info) Requirements: 1. Prefer local/offline solution (no paid cloud). Recommended stack: Python (OpenCV, pytesseract, pyzbar) OR VB.NET + Tesseract/ZXing. Hybrid (Python script + VB.NET caller) is OK. 2. Must handle glare/reflection, small rotation, and different image sizes. Use preprocessing (CLAHE, denoise, adaptive threshold, morphological ops). 3. Deliverables: - `ocr_qr_pipeline.py` (or compiled .exe) — CLI: accepts image path, prints JSON to stdout. - Simple VB.NET demo that saves PB1 image to disk, calls script, parses JSON, and shows formatted ChatGPT-style text: Metal First Line: <value> QR Above Text: <value> QR Code Data: <value> Status: <OK/NOT OK> - README with install steps for Windows, Tesseract path, Python deps, and run instructions. - 7 days bugfix support after delivery. 4. Trial (paid): I will provide 1 sample image. Deliver expected JSON for that image within 6 hours. Trial pay: ₹800. If you refuse trial, do not apply. 5. Acceptance: script must correctly decode provided sample image and run locally on Windows with clear README. JSON schema example (exact): { "metal_text": "F5D08714", "qr_above_text": "F5D09714", "qr_data": "720-4107F5D09714|S/N:152L6A7033|...", "qr_all": [{"data":"720-4107F5D09714...", "type":"QRCODE", "rect":{"x":100,"y":200,"w":300,"h":300}}], "status": "NOT OK" } How to apply (required): - Portfolio (GitHub / repo / demo) - Short answers to screening Qs (below) - Confirm: "I accept paid trial ₹800" and earliest start time Screening Qs (paste these as required answers): 1) Have you done embossed/metal OCR before? Link/sample. 2) Which tools will you use? (OpenCV/Tesseract/pyzbar/ZXing) 3) Can you run locally on Windows and be called from VB.NET? (yes/no) 4) Earliest start & ETA for prototype. Budget suggestion: Prototype ₹2,000–₹4,000 e). Use platform escrow/milestones.