feat(T03): OCR-Erfassung via OpenRouter Qwen2.5-VL + OCR UI with drag-and-drop

This commit is contained in:
2026-07-17 00:45:14 +02:00
parent b128ea6b39
commit 149ac04dc1
20 changed files with 2348 additions and 182 deletions
+1
View File
@@ -149,6 +149,7 @@
"id": "T03",
"title": "OCR-Erfassung (ZB I/II) via OpenRouter Vision + OCR UI",
"category": "ocr",
"status": "completed",
"description": "Komplettes OCR-Modul: OCRResult Model, ocr_service mit OpenRouter Qwen2.5-VL Integration, OCR router (upload, get results, apply to vehicle). Async OCR processing via Redis Queue. Prompt-Engineering für ZB I/II: structured JSON output (brand, model, vin, first_registration, mileage, power_kw, fuel_type). Confidence-Score Berechnung. Manual Review bei confidence < 0.7. Response-Caching für identische Scans. Frontend: OCR Upload page (drag-and-drop), OCR Results view mit side-by-side Original Scan + Extracted Data, Apply-to-Vehicle Button.",
"assigned_subagent": "implementation_engineer",
"dependencies": ["T01", "T02"],