GitHub

AI on the Edge Device: Digitizing Your non-digital meters with an ESP32-CAM

What is the project about?

This project is about using an ESP32-CAM, a low-cost microcontroller with a camera, to digitize analog meters (water, gas, electricity) by applying AI and image processing on the device itself (edge computing). It reads the meter values and makes them available digitally.

What problem does it solve?

It solves the problem of manually reading analog meters. It allows users to convert their existing, non-digital meters into "smart" meters without replacing them, enabling remote monitoring and data logging.

What are the features of the project?

  • TensorFlow Lite (TFLite) Integration: Uses TFLite for on-device AI processing.
  • Inline Image Processing: Performs image alignment, feature detection, and Region of Interest (ROI) extraction.
  • Small and Affordable: Uses a compact and inexpensive ESP32-CAM.
  • Integrated Camera and Illumination: Includes a camera and lighting for capturing meter readings.
  • Web Interface: Provides a web interface for configuration and control.
  • OTA Updates: Supports Over-The-Air updates via the web interface.
  • Home Assistant Integration: Seamlessly integrates with Home Assistant.
  • Data Storage and Forwarding: Supports InfluxDB (versions 1 and 2) and MQTT for data storage and transmission.
  • REST API: Offers a REST API for accessing meter data.

What are the technologies used in the project?

  • Hardware: ESP32-CAM (ESP32 microcontroller with camera)
  • Software/Frameworks:
    • C++ (main programming language)
    • TensorFlow Lite (for AI model inference)
    • Web Server (for the user interface and API)
    • MQTT (for data communication)
    • InfluxDB (for time-series data storage)
    • REST API

What are the benefits of the project?

  • Cost-Effective: Uses inexpensive hardware.
  • Non-Invasive: Doesn't require replacing existing meters.
  • Remote Monitoring: Enables remote reading of meter values.
  • Data Logging: Allows for tracking consumption over time.
  • Automation: Facilitates integration with smart home systems (e.g., Home Assistant).
  • Edge Computing: Performs AI processing locally, reducing reliance on cloud services and improving privacy/latency.

What are the use cases of the project?

  • Smart Water Metering: Monitoring water usage.
  • Smart Electricity Metering: Tracking electricity consumption.
  • Smart Gas Metering: Monitoring gas usage.
  • Home Automation: Integrating meter readings into smart home systems.
  • Data Analysis: Collecting consumption data for analysis and optimization.
  • Remote Monitoring of any analog Gauge: It could be used to monitor any analog gauge.
AI-on-the-edge-device screenshot