PARADAPARADA

PARADA is my capstone at Lyceum of the Philippines University — Batangas. A camera at each zone gate sends frames to a Python vision service (OpenCV finds the plate, EasyOCR reads it), which reports an observation to the Express API. The API is the only authority on what that means: it resolves the vehicle, opens or closes the session, and updates the zone's count in one transaction. Server-Sent Events push the new number to the driver app (React Native + Expo) and the admin console (Next.js). Slots exist for layout only — the screen never pretends to know more than the gate saw. Phases 0–14 are done, including a synthetic OCR benchmark I published with its failures: 77.4% exact reads on 1,150 generated plates — 100% in clean, dark and noisy conditions, far lower when a plate is half hidden — and no false plates on 100 empty scenes. Deployment is under way.