Real-Time Spill Detection System Developed by Team Aura++

Description

A real-time intelligent system that detects liquid spills and garbage using deep learning, notifies maintenance teams instantly through a connected dashboard and task management system.

Built for industrial and public safety, this solution ensures quick response and reduced hazard risk.

---

Table of Contents

• Features
• Project Structure
• Setup Instructions
  • Prerequisites
  • Installation
• Tech Stack
• Testing
• Notes
• Demo
• Contributing
• License

---

Features

• Real-time Detection – Instantly identifies liquid spills and garbage using a custom-trained deep learning model
• Admin Dashboard – Centralized interface to monitor detections, alerts, and system activity in real-time
• Crew Management – Smart task assignment and tracking system for cleaning crew, ensuring timely resolution
• Firebase Integration – Secure authentication and real-time data synchronization using Firebase
• Scalable Architecture – Modular MERN stack integrated with Firebase for high performance and scalability

---

UI Images :

Admin Dashboard

Crew Dashboard

App UI

Project Structure

Aura_plus_plus_hackgenx/
├── backend/               # Node.js + Express backend with ML integration
├── admin-dashboard/
├── crew-dashboard/        # Frontend interface for cleaning crew
├── model_training/        # Google Colab notebook for training
├── realtime-detection/    # Real-time detection via webcam(OPEN CV)/CCTV (Python)
└── README.md

---

Setup Instructions

Prerequisites (with Links)

• Node.js
• npm
• Python 3.10
• Git
• Firebase Console
• Google Colab

Installation

1. Clone the Repository

git clone https://github.com/darshanbagade/Aura_plus_plus_hackgenx.git
cd Aura_plus_plus_hackgenx

2. Train the Model Using Google Colab

• Open model_training/spill_detection_colab.ipynb in Google Colab
• Upload your spill dataset or link via Google Drive

Install dependencies:

!pip install tensorflow keras opencv-python matplotlib

Train and save the model:

model.save('spill_detection_model.h5')

• Download spill_detection_model.h5 and place it in:
  backend/models/spill_detection_model.h5

---

4. Open ML Detection Model

Navigate to the backend:

cd realtime-detection

Install all dependencies

pip install opencv-python requests python-dotenv firebase-admin ultralytics

Run the Object Detection Model

python detection.py

Add your Firebase Admin SDK JSON to:

backend/config/firebaseServiceAccountKey.json

Start the backend server:

npm start

---

4. Backend Setup

Navigate to the backend:

cd backend
npm install

Add your Firebase Admin SDK JSON to:

backend/config/firebaseServiceAccountKey.json

Start the backend server:

npm start

---

5. Frontend Setup (Admin Dashboard)

Navigate to the frontend:

cd ../admin-dashboard
npm install

Update API base URL in src/config.js:

export const API_BASE_URL = 'http://localhost:3000';

Run the frontend development server:

npm run dev

---

5. Firebase Configuration

• Go to Firebase Console
• Create a new project or use an existing one
• Enable Firestore Database and Firebase Authentication
• Download the Admin SDK credentials and place them in:

backend/config/firebaseServiceAccountKey.json

Firebase init code in backend/index.js:

const admin = require('firebase-admin');
const serviceAccount = require('./config/firebaseServiceAccountKey.json');

admin.initializeApp({
  credential: admin.credential.cert(serviceAccount),
  databaseURL: 'https://<your-project-id>.firebaseio.com'
});

---

Tech Stack

Layer	Technology
Frontend	React.js (JSX)
Backend	Node.js, Express.js
ML Model	Yolo, OpenCV
Database	Firebase Firestore
Mobile App	React Native, Expo
Auth	Firebase Auth
Training	Google Colab

---

Notes

• Use .env files to manage sensitive keys and secrets
• Place trained model in backend/models/spill_detection_model.h5
• Log model loading steps in backend for debugging