PFU (Kahoku City, Ishikawa Prefecture, President Izumi Nagahori) is developing an "automatic resource waste AI sorter" that utilizes the technology of its own products. From the perspective of the United Nations Sustainable Development Goals (SDGs), in addition to efficient recycling, China's waste plastic import ban policy and the increase in household waste due to corona damage are urgent resources at home and abroad. We will contribute to dealing with garbage.
Select the color and shape of the bottle by AI.To date, the recognition rate for the three types of glass bottles has improved to 99.1%.
Currently, we are working on labor saving by machines in order to improve work efficiency and ensure business continuity by solving the labor shortage at waste disposal facilities. The sorting machine aims to be introduced within 2022.
Even among recyclable waste, glass bottles for beverages are classified into brown such as beer, transparent such as sake, and other colors such as wine, and need to be sorted at the time of recycling. Sorting at the treatment facility is done manually, but the on-site workers are aging and the working environment is not good due to long standing work.
The company has been discussing the possibility of waste treatment business since the fall of 2019, and started a demonstration experiment in February 2008. "There are other recycled products such as cans and PET bottles, but for the time being, we will focus on glass bottles that have not been technically established," said Noboru Tabata, General Manager of the Second Sales Department of the Edge Products Division. Bottles are sorted by combining the optical technology and recognition technology cultivated in the development.
A prototype installed in the head office.Grab and sort bottles with a vacuum suction hand through a recognition camera
The position of the center of gravity is read from the color coding and contour of the glass bottle, and sorted by the vacuum suction hand. If it is a four-goal bottle, it is one bottle per second, which enables a processing speed close to the sorting speed of humans. For practical use, it will be more compact and self-supporting than the prototype in order to support various lines in the processing facility, and it will be possible to install it afterwards without major line modification.
Masanobu Motoe of the Document Imaging Division, who is involved in the development, said, "(As of July 2009) the average recognition rate of three types of glass bottles is 99.1%. I want to raise this to 99.9%. Especially difficult is beverages. It's a selection of bottles for cosmetics and bottles for cosmetics. I want to make the best use of my company's technological capabilities to differentiate myself from other companies. " In the future, we will first build up a track record in domestic processing facilities and are considering global expansion.
Nikkan Kogyo Shimbun December 16, 2021