NRC-GAMMA: A novel benchmark gas meter image dataset

Published on 9 January 2022 at 19:45

National Research Council Canada

Having trouble teaching your old machines new data tricks? That’s about to change, thanks to a new open-source dataset from our Digital Technologies Research Centre’s Data Analytics Centre.

In an increasingly digital world, we’re still surrounded by equipment with gauges, dials and counters that provide visual feedback. Analog equipment like your gas meter is usually read in person – it takes time, and can lead to errors. Advances in artificial intelligence and computer vision will make it possible to use automatic meter reading systems.
We’re introducing a new benchmark dataset of real-life gas meter images. The NRC-GAMMA dataset is the first and largest of its kind to be made open access, including for commercial use. The team used crowdsourcing to annotate over 28,000 meter images. Through a systematic quality control method, programmers coded scripts to validate the image labels. The result? An open-access dataset that can be used to build AI to digitize legacy equipment.
Get the dataset:
Read more:

Add comment


There are no comments yet.