(5)
Synaps Vision
Synaps Vision Overview:
Synaps Vision is an advanced Smart Parking and Automated License Plate Recognition (ALPR) system specifically engineered for the Thai market. By combining state-of-the-art YOLOv8 object tracking with a hybrid PaddleOCR and Typhoon OCR pipeline, it delivers unparalleled accuracy in diverse lighting conditions while seamlessly integrating with a cloud-based tiered fee engine.
Key Features
Computer Vision Pipeline • YOLOv8
Hybrid OCR Engine • PaddleOCR & Typhoon
Real-time HUD Overlay • OpenCV & Pillow
Enterprise Cloud Sync • Supabase
Year
2026
Role
AI / Computer Vision Engineer Systems Engineer Tech Lead
Objectives:
Achieve 99.5% average accuracy in license plate bounding box detection.
Implement a resilient hybrid OCR pipeline utilizing both local and cloud APIs.
Develop an advanced image preprocessing engine for optimal optical clarity.
Integrate a robust, real-time tiered fee engine supporting complex parking tariff logic.
Design a Material Design-inspired Interactive HUD with full Thai Unicode support.
Scope of Work:
The development of Synaps Vision focused on creating a highly accurate, end-to-end ALPR pipeline and robust business logic synchronization:
Detection & Tracking: Deploying YOLOv8 Nano with Dual-Stage Tracking and Image Quality Assessment (IQA) for extracting the sharpest frames.
Preprocessing Engine: Engineering a Smart Inner Plate Cropper using Dual-Polarity Otsu Thresholding, Sobel Operators, and LAB CLAHE Normalization.
Cognitive Parsing: Implementing a hybrid PaddleOCR and Typhoon OCR architecture fortified with Regex filtering and Fuzzy Logic for Thai province matching.
Business Logic & Sync: Synchronizing state asynchronously with Supabase to evaluate tiered parking tariffs, grace periods, and daily caps.
Automation & Visual Status: Developing a rich terminal console interface and a real-time OpenCV overlay with background threading for automated gate cycling.
A highly scalable, production-ready ALPR system that dramatically elevates the parking management experience through state-of-the-art computer vision and robust cloud integration.
- Get in touch.
- Get in touch.
- Get in touch.
- Get in touch.
- Get in touch.
- Get in touch.
