Water Leak Detection System 💧🌿

A smart IoT solution designed to detect water leaks in garden irrigation systems using machine learning. This project integrates embedded hardware, real-time data streaming, and a modern web dashboard to prevent water waste and protect plants.

🚀 Project Overview

This system monitors water flow and soil moisture data from ESP32-based sensors. It uses a custom neural network (LeakNet), based on the paper "Inherently Interpretable Time Series Classification via Multiple Instance Learning", to analyze time-series data and detect anomalies indicative of leaks. Users can monitor their garden's status and receive alerts through email.

🏗️ System Architecture

The project follows a microservices architecture:

1. Edge Devices (ESP32): Collects sensor data and transmits it via MQTT/HTTP. Supports BLE provisioning.
2. Data Ingestion Service: A Node.js service that consumes raw sensor data, validates it using Protobuf, and queues it in RabbitMQ.
3. LeakNet (ML Engine): A Python-based service that processes data streams, runs inference using a PyTorch model, and flags potential leaks.
4. Web Dashboard: A user-friendly interface for real-time monitoring, historical data visualization, and system management.

🛠️ Technology Stack

Frontend (Web Dashboard)

• Framework: Next.js 15 (App Router)
• UI Library: React 19
• Styling: Tailwind CSS v4
• Visualization: React-Chartjs-2
• State/Comms: Socket.io for real-time updates

Backend Services

• Runtime: Node.js
• Messaging: RabbitMQ
• Serialization: Protocol Buffers (Protobuf)
• Database: MongoDB (via Mongoose)

Machine Learning (LeakNet)

• Framework: PyTorch & Lightning
• Data Processing: Pandas, NumPy, Scikit-learn
• Optimization: Optuna
• Time Series: TSAug (Time Series Augmentation)

Firmware

• Platform: ESP32
• SDK: ESP-IDF
• Features: BLE Provisioning, WiFi connectivity

📂 Project Structure

leak-detection-system/
├── esp32/                  # Firmware code for ESP32 microcontrollers
├── leaknet/                # Python ML project (Training, Inference, Notebooks)
└── web/                    # Monorepo for Web & Node.js services
    ├── apps/
    │   ├── website/                # Next.js Dashboard
    │   ├── data-ingestion-service/ # Node.js Data Receiver
    │   ├── inference-service/      # Service bridging ML & Backend
    │   └── notification-service/   # Alerting system
    └── packages/           # Shared libraries (DB, UI, Config)

🔗 Related Repositories

• Firmware (ESP32): leak-detection-system-esp32
• Machine Learning (LeakNet): leaknet

🚦 Getting Started

Prerequisites

• Docker & Docker Compose

Installation

1. Clone the repository:

   git clone https://github.com/nikas-belogolov/leak-detection-system.git
   cd leak-detection-system/web

2. Build and run the entire stack:

   docker compose build
   docker compose up -d

The system will be available at http://localhost:3000

👨‍💻 Author

Nikas Belogolov

---

This project is part of my portfolio demonstrating full-stack IoT development capabilities.