While it’s definitely important to keep the heavy metals in discarded batteries out of the environment, the sorting of all of the different types of batteries that arrive at a recycling depot could no doubt get extremely tedious. It’s the type of job that often goes to a machine. Well, such a machine has been invented. Called the Optisort, it can recognize about 2,000 types of batteries, and is currently being used to sort one third of those recycled in the UK.
The machine was first conceived by Claes Strannegård, an Artificial Intelligence (AI) researcher with Sweden’s University of Gothenburg and Chalmers University of Technology. He got the idea when he was sorting his own garbage, and proceeded to get in touch with Swedish recycling company Renova. In a collaboration between the company and the universities, the system was created.
NEW ATLAS NEEDS YOUR SUPPORT
Upgrade to a Plus subscription today, and read the site without ads.
It's just US$19 a year.UPGRADE NOW
Batteries are fed into the Optisort on a conveyor belt, where each one is photographed by the machine’s camera. Each image is then compared to a database of existing shots of different types of batteries, until a match is made. Based on the battery’s chemical content, a jet of compressed air is then used to direct it into a designated bin.
Up to ten batteries can be processed per second. Because the machine’s AI system allows it to learn from what it “sees,” it can even recognize damaged or dirty batteries.
Although the system is designed primarily with the environment in mind, it also makes allowances for commerce. It keeps a record of the composition, brand and model of each battery received, so that the raw materials can be sold by the recycler.
Optisort machines are now being produced by the Gothenburg-based Optisort company. One is currently in use at Renova, while another is handling approximately 33 percent of the UK’s recycled batteries at England’s G&P Batteries.
More information is available in the video below.
Source: University of Gothenburg