This project is significant for the cornerstone project selection editor, as I have been engaged in the Internet of Things and edge computing fields, primarily working with microcontrollers/embedded ARM processors that are not designed for AI-capable computing devices.Although these devices are functionally stable, they struggle with data collection, remote monitoring, and intelligent management; moreover, in today’s AI-centric environment, they seem less impressive. However, I recently discovered an open-source project on GitHub that leverages modern AI technology and low-cost hardware to empower these traditional devices with AI capabilities. While it still adheres to traditional embedded product development concepts, it certainly opens a window of opportunity.It is worth noting that similar products are already available on the market, and the applications I am aware of include remote water and electricity meter reading, which is one of the values of this project. Additionally, using a camera for reading may allow for further expansions.
AI on the Edge Device is an open-source project aimed at connecting traditional non-digital metering devices (such as water meters, electricity meters, gas meters, etc.) to the digital world through edge computing technology. The core of the project is a small device based on the ESP32-CAM, which can automatically read the readings of traditional metering devices and digitize them for remote monitoring and data analysis.
🛠️ Core Features
• TensorFlow Lite Integration:
The project integrates TensorFlow Lite, a lightweight machine learning framework, to run AI models on the device for image recognition and data processing.
• Image Processing:
The device can automatically identify the reading area (ROI) of the metering device and perform image processing to ensure the accuracy of the readings.
• Web Interface Control:
Through the built-in web interface, users can easily configure and monitor the device, view real-time data and historical records.
• Multiple Data Transmission Methods:
Supports MQTT protocol, InfluxDB, and REST API, allowing users to transmit data to various cloud platforms or local servers for remote monitoring and analysis.
• OTA Updates:
The device supports firmware updates via Wi-Fi, allowing users to easily upgrade device functionality without manual intervention.
💡 Project Highlights
• Low Cost: The entire project is based on the ESP32-CAM, an inexpensive yet powerful microcontroller, along with some simple peripheral devices, totaling less than 10 euros.
• Easy to Use: The project provides detailed documentation and tutorials, making it accessible even for beginners.
• Highly Integrated: The device supports multiple data transmission methods and can seamlessly integrate with smart home systems like Home Assistant for smarter home automation management.
Practical Application Cases
Water meter digitization can be achieved by installing the AI on the Edge Device on the water meter, allowing the device to automatically read the water meter’s readings and transmit them to the cloud or local server. Users can monitor water usage anytime and anywhere through their mobile phones or computers, enabling remote monitoring and data analysis.
Similarly, for electricity meters, the device can automatically read the electricity meter’s readings and send the data to the smart home system via MQTT protocol. Users can monitor electricity usage in real-time, optimize energy consumption, and reduce electricity costs.
Original project link:https://github.com/jomjol/AI-on-the-edge-device