🚗 Vehicle License Plate Recognition System

This project is a real-time system for automatic vehicle license plate detection and recognition, based on deep learning, OCR, and microservice architecture.

---

🏗️ System Architecture Overview

The system is composed of several microservices connected via gRPC and WebSocket, and supports real-time video processing, object tracking, and license plate recognition.

---

🔁 Workflow

1. Camera Input
   A video stream is captured from a surveillance camera and sent to the Gateway Server.

2. Gateway Server

   • Splits the stream into individual frames.
   • Sends frames via gRPC to:
     • Vehicle Detection Server
     • Plate Detection Server

3. Vehicle Detection Server (YOLOv8)

   • Uses YOLOv8 to detect vehicles in each frame.
   • Sends vehicle bounding boxes to the Tracking Server.

4. Tracking Server (DeepSORT)

   • Tracks vehicles across frames using DeepSORT.
   • Outputs results with tracking IDs to the Result Aggregator Server.

5. Plate Detection Server (YOLOv8)

   • Detects license plate areas in frames.
   • Sends plate regions to the OCR Server.

6. OCR Server (PaddleOCR)

   • Extracts characters from license plates.
   • Sends recognized plate numbers to the Result Aggregator Server.

7. Result Aggregator Server

   • Merges vehicle tracking data and recognized plate text into one final result per frame.
   • Sends aggregated results to the Gateway Server.

8. Backend Server (FastAPI)

   • Receives results from Gateway via WebSocket.
   • Saves results to MongoDB.
   • Handles frontend queries (e.g., plate search, logs).

9. Frontend

   • Displays real-time video with vehicle and license plate overlays.
   • Allows search/filtering of historical vehicle data.

---

🧰 Technologies Used

Functionality	Tool/Library
Object Detection	YOLOv8 (Ultralytics)
Vehicle Tracking	DeepSORT
License Plate Recognition	PaddleOCR
Communication	gRPC, WebSocket
Backend API	FastAPI
Database	MongoDB
Frontend	React.js

---

📦 Microservices List

• Gateway Server: Streams video and routes frames.
• Vehicle Detection Server: YOLOv8-based vehicle detection.
• Plate Detection Server: YOLOv8-based license plate detection.
• Tracking Server: DeepSORT vehicle tracker.
• OCR Server: PaddleOCR for text recognition.
• Result Aggregator Server: Combines all output into unified results.
• Backend Server: API and WebSocket hub.
• Frontend: Real-time dashboard and search interface.

---

🚀 Deployment Instructions

1. Install dependencies:

   pip install torch==2.5.1 torchvision==0.20.1 torchaudio==2.5.1 --index-url https://download.pytorch.org/whl/cu118
   pip install paddlepaddle-gpu==2.6.2 -i https://www.paddlepaddle.org.cn/packages/stable/cu118/
   pip install -r requirements.txt

2. Start services in this order:

   • Gateway Server
   • YOLO Servers (vehicle and plate detection)
   • Tracking Server
   • OCR Server
   • Result Aggregator Server
   • Backend (FastAPI)
   • Frontend (React)

3. Test the system with sample video:

   Run the following script to simulate a live camera feed using a sample .mp4 video:

   python test_video_stream.py

   This script reads the video frame-by-frame and streams it into the system via the Gateway Server.

   Once the system is running, access the dashboard to:

   • View real-time video with bounding boxes and recognized license plates.
   • Monitor vehicle tracking and OCR output.
   • Search license plate logs and view historical data.

---

🖼️ Sample Results

Below are example outputs from the system:

🔹 Live Video Stream with Bounding Boxes

• Green box: Vehicle detection (YOLOv8 + DeepSORT ID)
• Blue box: License plate region
• Plate number text (PaddleOCR)

🔹 License Plate Log Example

🔹 Search License Plate

---

📜 License

This project is for educational purposes only. Commercial use is not permitted without authorization.