In an ideal world, technology’s role would be to make all of our lives easier. The ads that suddenly pop up on our smart TVs and gaming systems might make us think otherwise, but some technology can improve our lives if we work hard at it. for [Cian]which meant training a neural network to read gas meters so that he didn’t have to read them himself.
There are two fundamental issues here. [Cian]gas companies have not upgraded their technology to modern remotely readable meters. And second, the meters are hidden deep inside the gas company and cannot be read by gas employees. [Cian]basement. The latter fact would require digging moria-like depths to reach the meter, so the solution here was to put the Raspberry Pi in this location instead. Pointing a camera at the meter isn’t powerful enough to identify the numbers on its own, so a neural network was trained to take accurate readings on the dial. And finally, since the machine is already connected to the network, [Cian] If you set it to automatically notify the gas company of your readings, the gas company will be completely out of the loop.
To automate such tedious tasks, a Raspberry Pi with computer vision tools like OpenCV is a fairly mature platform for such light machine learning work. We’ve seen license plate readers and neighborhood traffic surveys built on top of these platforms, automating human labor and making our lives easier, one single-board computer at a time.
