In this article, I’m going to show you how you can create an Analog meter reader with computer vision. This can be a very good alternative when for reasons that can be costs, time restrictions, or anything that doesn’t allow you to have a digital meter.

Why is this solution very useful? For many reasons, and I’m going to give you three of them:

  1. You can have reading meters in real-time, and you can access remotely with a phone or a computer
  2. The information can be saved automatically with date and time so that you can have stats and precise reading for any hour/day/month/year
  3. You can get a warning in realtime to your phone or email

water meter remotely with computer vision

When to use Analog meter reader

It is not always possible to change the old analog meter but it is increasingly important to have consumption under control. Even with an old water meter, gas meter, or other type of supplies you can have real-time warnings in case of problems or be constantly updated on consumption even remotely. To do all this you can create a program with computer vision that sends the data to a server from which we can consult the values on a simple management panel.

Here are some ideas on how to make it

The reader of the analog meter is easily achievable because all you need is a webcam, hardware such as raspberry pi, and computer vision software for processing the information. Here is an example scheme

analog water meter reader scheme

OpenCV Software for Analog meter reader

There are several tools for reading numbers, some more specific and dedicated to the Analog meter reader, or in general recognition of the text with Text Recognition (OCR) with Tesseract. If you are experienced you can also build your image classification to identify the numbers and integrate it into your OpenCV project.

If you want to have an even more precise reading and identify the numbers in the circular dial you can do it simply with OpenCV, it is simply geometric operations. For example, you can identify the circle in an image, divide the circle into portions, and by establishing in which intersection the hand is located you will also get the corresponding number.

analog meter and computer vision

Control panel

Now there is the simpler part. Once the information processed by the reading software with OpenCV has been received, it is sufficient to send them to a simple hosting from which to access and read everything comfortably from a control panel.

If you want to exaggerate, save the reading data at regular intervals and generate graphs with the history of consumption.