🌊 Automated Water Quality Monitoring System

An intelligent foam detection system that monitors water quality in real-time using YOLOv8 computer vision and automated email notifications.

##Overview

This project implements an automated water quality monitoring solution that:

• Detects foam presence in live video streams using YOLOv8 object detection
• Monitors water flow continuously with OpenCV-based analysis
• Sends automated alerts via email when foam is detected or streams are offline
• Prevents duplicate processing with intelligent video tracking

✨ Key Features

• Detection: YOLOv8 model trained for 25 epochs with optimized 224×224 resolution
• Smart Notifications: Automated email system with 24-hour interval monitoring
• Stream Validation: Real-time video stream availability checking
• Water Flow Analysis: OpenCV-based flow detection with 10% pixel threshold
• Duplicate Prevention: Tracks processed videos to avoid redundant analysis
• Continuous Monitoring: 24/7 automated surveillance with configurable intervals

🛠️ Technology Stack

• Deep Learning: YOLOv8 (Ultralytics)
• Computer Vision: OpenCV
• Email Service: SMTP with Gmail integration
• Cloud Platform: Google Colab
• Storage: Google Drive integration
• Languages: Python 3.x

📋 Requirements

ultralytics
opencv-python
google-colab
smtplib (built-in)
subprocess (built-in)

🚀 Installation & Setup

1. Clone the repository

   git clone https://github.com/yourusername/foam-detection.git
   cd foam-detection

2. Install dependencies

   pip install -U ultralytics opencv-python

3. Setup Google Drive (if using Colab)

   from google.colab import drive
   drive.mount('/content/drive')

4. Configure email credentials

   # Update in the script
   TO_EMAIL = 'your-email@gmail.com'
   from_email = 'sender-email@gmail.com'
   from_password = 'your-app-password'  # Gmail App Password

🔧 Configuration

Model Training

!yolo task=detect mode=train model=yolov8m.pt data=data.yaml epochs=25 imgsz=224 plots=True

Key Parameters

• Detection Confidence: 0.25
• Image Size: 224×224 pixels
• Monitoring Interval: 24 hours
• Email Cooldown: 24 hours
• Water Flow Threshold: 10% pixel activity

📊 System Architecture

Video Stream → YOLOv8 Detection → Water Flow Analysis → Email Notification
     ↓              ↓                    ↓                    ↓
Stream Check → Foam Detection → Flow Validation → Alert System

🎮 Usage

Training the Model

# Train YOLOv8 on your foam dataset
!yolo task=detect mode=train model=yolov8m.pt data=data.yaml epochs=25 imgsz=224

Running Detection

# Single prediction
!yolo task=detect mode=predict model=runs/detect/train6/weights/best.pt conf=0.25 source=your_video.mp4

# Continuous monitoring
python foam_detection.py

Email Notifications

The system automatically sends different types of alerts:

• ✅ "No Foam Detected" - Water quality is good
• ⚠️ "Foam Detected" - Foam present in water
• 🔴 "No Live Stream" - Video stream unavailable
• 💧 "No Water Flow Detected" - No water movement detected

📁 Project Structure

foam-detection/
├── foam_detection.py          # Main detection script
├── data/
│   ├── train/                # Training images
│   ├── val/                  # Validation images
│   ├── test/                 # Test videos
│   └── data.yaml             # Dataset configuration
├── runs/
│   └── detect/
│       └── train6/
│           └── weights/
│               └── best.pt   # Trained model weights
├── last_email_time.txt       # Email tracking
├── processed_videos.txt      # Video processing log
└── README.md

🔍 Detection Process

1. Stream Validation: Checks if video stream is accessible
2. Water Flow Analysis: Validates presence of water movement
3. YOLOv8 Detection: Runs foam detection on video frames
4. Result Processing: Analyzes detection outputs
5. Notification: Sends appropriate email alerts
6. Logging: Records processed videos to prevent duplicates

📈 Performance Metrics

• Model: YOLOv8 Medium (yolov8m.pt)
• Training Epochs: 25
• Input Resolution: 224×224
• Confidence Threshold: 0.25
• Processing: Real-time video analysis
• Monitoring: 24/7 continuous surveillance

Educational implementation demonstrating fundamental clustering concepts in machine learning.