BagAlert

AI-powered surveillance system using object detection to monitor belongings and detect disturbances.

Table of Contents

Table of Contents

1. About The Project
   • Key Features
2. Built With
3. Getting Started
   • Prerequisites
   • Installation
4. More Information

About The Project

BagAlert is an AI-powered security system designed to monitor your belongings and detect disturbances. It utilizes object detection via a camera and the aitrios-rpi-application-module-library to identify and track items. If an object is moved or goes missing, the system triggers an alert. The system also has an RFID trigger to arm the system.

Key Features

• Object Detection: Uses the IMX500 and SSDMobileNetV2FPNLite320x320 model to identify objects in the camera's view.
• Disturbance Detection: Calculates Intersection over Union (IoU) to determine if protected objects have been moved.
• Real-time Alerts: Sends alerts via WebSocket to a web-based frontend when disturbances are detected.
• Web Interface: Provides a user interface to view the camera feed, activate/deactivate protection, and view recent alerts.
• RFID Trigger: Simulates an RFID trigger to activate protection (can be integrated with actual RFID hardware).

Built With

• Backend:
  • Python 3
  • Flask
  • FastAPI
  • OpenCV
  • modlib (aitrios-rpi-application-module-library)
  • uvicorn
• Frontend:
  • React
  • Material UI
  • Socket.IO Client
• Hardware (intended):
  • Raspberry Pi
  • IMX500 Camera
  • ESP32 (for RFID)

Getting Started

To get started with BagAlert, follow the instructions below.

Prerequisites

• Raspberry Pi with a camera (IMX500 recommended)
• Python 3.6+
• Node.js and npm (for the frontend)
• Access to a mobile hotspot/internet connection for the Raspberry Pi and your development machine to communicate

Installation

1. Clone the repository:

   git clone https://github.com/jalliet/bagalert.git
   cd bagalert

2. Backend Setup:

   • Run the backend_setup.sh script:

     ./backend_setup.sh

     This script performs the following actions:
     • Installs ModLib and its dependencies.
     • Creates and activates a Python virtual environment.
     • Installs required Python packages using install_requirements.sh.
     • Starts the camera server (camera/camera_server.py).

3. Install ModLib:

   ./install_modlib.sh

   This script handles the installation of the aitrios-rpi-application-module-library and its dependencies.

4. Frontend Setup:

   • Navigate to the frontend directory:

     cd frontend

   • Install npm packages:

     npm install --force

   • Start the frontend:

     npm run start

   • The frontend will be accessible at http://localhost:3000.

5. ESP32 Setup (RFID - Optional):

   • Open the rfid/rfid.ino file in the Arduino IDE.
   • Modify the ssid, password, and mqtt_server variables with your network and Raspberry Pi's IP address.
   • Upload the code to your ESP32.

6. MQTT Broker Setup (If using MQTT):

   • Ensure an MQTT broker is running and accessible to both the Raspberry Pi and the ESP32.
   • Modify the mqtt_publisher.py and mqtt_suscriber.py files with the correct broker IP address.

Running on Raspberry Pi

Follow these steps to run BagAlert on your Raspberry Pi:

1. Navigate to Project Directory:

   cd bagalert

2. Setup Instructions:

   First Time Setup

   If you haven't installed the requirements yet:

   # Make scripts executable
   chmod +x backend_setup.sh frontend_setup.sh install_requirements.sh
   
   # Install dependencies and set up the environment
   ./install_requirements.sh

   For All Runs (Including First Time)

   # Make scripts executable (if not done already)
   chmod +x backend_setup.sh frontend_setup.sh

3. Start Backend Server:

   # In your first terminal window
   ./backend_setup.sh

4. Launch Frontend Server:

   # Open a new terminal window
   cd bagalert
   ./frontend_setup.sh

5. Start MQTT Subscriber:

   # Open a new terminal window
   cd bagalert/mqtt
   python mqtt_subscriber.py

The system should now be up and running! You can access the web interface at http://localhost:3000 and start monitoring your belongings.

System Architecture

BagAlert uses a modular architecture with three main components:

1. Main Application (main.py):

   • System initialization and coordination
   • Service management
   • Event handling
   • Graceful shutdown procedures

2. Camera Service (camera_service.py):

   • Camera feed management
   • Object detection and tracking
   • Protection state management
   • Web interface and WebSocket streaming
   • Alert generation

3. RFID Service (rfid_service.py):

   • MQTT client management
   • RFID event processing
   • User session handling
   • Service communication

Key Features

• Inter-service Communication: Event-driven architecture ensuring smooth communication between components
• Configuration Management: Flexible configuration through external files and environment variables
• Logging and Monitoring: Comprehensive activity tracking and status reporting
• Web Interface: Real-time status display, live camera feed, and alert notifications

More info

• DevPost
• Hackster.io