ESP32CAM Meter Reading Project

My internship project at Kolour Think Tank where I worked on reading a digital utility meter (mainly focused on water meter) using an ESP32-Camera and AWS Rekognition and storing the readings in AWS DynamoDB.

Documentation here alongside setup video guide.

Demo

Video

Overview

Project Architecture

Hardware Requirements

• ESP32-CAM (Most modules should work)
• ESP32-CAM-MB or USB to TTL (For uploading code)

Software

• Frameworks: PlatformIO or Arduino IDE
  • Arduino-Libraries: Arduino-ESP32, ArduinoJSON, ESPAsyncWebServer
• AWS Services: Rekognition, Lambda, API Gateway, S3, DynamoDB

For more details see documentation.

Resources Used

• Documentation and Articles Used

  • AWS Documentation
  • Useful Rekognition Material:
    • Text Detection Rekognition
    • Analyze image from S3
    • Draw boundary boxes w/ Pillow
    • Trigger Rekognition from S3
    • Store Rekogniton result to DynamoDB
  • Creating API for DynamoDB Material:
    • Building Serverless API w/ DynamoDB, API Gateway, Lambda
  • Creating API for Accepting Images:
    • ESP32 Cam To AWS S3 Upload
    • Sharp Documentation
  • ESP32-CAM Code:
    • Taking images, storing w/ SPIFFS, and sending it

• Lambda Layer:

  • Sharp (packaged layer here)
  • Python Pillow (packaged layer here)