Skip to content

README.md

Latest commit

 

History

History
135 lines (90 loc) · 3.27 KB

README.md

File metadata and controls

135 lines (90 loc) · 3.27 KB

🖥️ Safe-Move Backend

🎓 Part of Graduation Thesis – University of Information Technology – VNU HCM\

This backend module powers the real-time AI-based flood monitoring system for over 600 traffic cameras in Ho Chi Minh City. It is designed using microservices architecture and runs inside Docker containers for scalability and maintainability.

🚀 Features

  • 🌐 FastAPI-based RESTful API server
  • 🧠 Integrates trained AI model for flood detection
  • 📦 Containerized using Docker
  • ⚙️ Microservices-ready architecture
  • 🌍 Receives image data from mobile/web frontend
  • 🧾 Returns flood detection prediction (Flood / No Flood)

🧱 Tech Stack

  • FastAPI – High-performance API framework
  • Python 3.10+ – Backend language
  • TensorFlow/Keras – For loading and running the CNN model
  • OpenCV – Image preprocessing
  • Uvicorn – ASGI server
  • Docker – For containerization

📁 Project Structure

backend/
├── app/
│   ├── main.py               # Entry point
│   ├── model_loader.py       # Load .keras model
│   ├── predictor.py          # Inference logic
│   ├── schemas.py            # Request/response schemas
│   └── utils.py              # Image preprocessing
├── Dockerfile                # Docker instructions
├── requirements.txt          # Python dependencies
└── README.md                 # You are here

🛠️ Installation (Local)

1. Clone the Repository

git clone https://github.com/ManhHoDinh/safe-move-ai.git
cd safe-move-ai/backend

2. Setup Virtual Environment

python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

3. Run the Server

uvicorn app.main:app --reload

Navigate to http://localhost:8000/docs to try the API via Swagger UI.

🐳 Docker Deployment

1. Build Docker Image

docker build -t safe-move-backend .

2. Run Docker Container

docker run -d -p 8000:8000 safe-move-backend

Test API at: http://localhost:8000/docs

🔌 API Endpoints

Method Endpoint Description
POST /predict Upload image and get prediction
GET /health Health check

Example cURL

curl -X POST "http://localhost:8000/predict" \
  -H  "accept: application/json" \
  -H  "Content-Type: multipart/form-data" \
  -F "file=@example.jpg"

📦 Model Loading

  • Place your trained .keras file in app/model/ and update model_loader.py path accordingly.

🔗 Integration

  • Connects with Flutter Mobile App via REST
  • Supports integration with frontend dashboards and admin systems

📚 Reference

Developed by Hồ Đình Mạnh & Lê Thị Bích Hằng – UIT – VNU HCM

📘 Back to Project Overview