AI paves the way for efficient textile-to-textile recycling

Could this state-of-the-art tech help keep clothes out of landfills?

Spotted: Despite increasing regulations around clothing circularity, masses of garments are sent to dumps every year. That includes piles in Chile’s Atacama desert, which are now so big they can be seen from space and are mostly made up of clothes imported from countries thousands of miles away. US-based Refiberd wants to revolutionise how we deal with textile waste, making these kinds of dumpsites a thing of the past.

Garments are often not recycled because they’re made up of several different components and fabric blends that all need to be treated differently. Refiberd uses a combination of generative AI and state-of-the-art hyperspectral imaging, which detects how light reflects off or is absorbed by different materials, to massively speed up clothes sorting, so that garments are processed correctly for recycling.

A clothing item passes along the conveyor belt and the camera quickly measures its hyperspectral data, which is processed by Refiberd’s machine learning model to reveal a garment’s material composition. The technology is extremely accurate, identifying the various fibres in blended and layered fabrics and even spotting trace materials.

According to the women-led company, its technology could help to divert up to 70 per cent of textile waste to recyclers, instead of it ending up in landfills. Refiberd has already been recognised by the H&M Foundation, winning a Global Change Award and prize money that is helping it scale and commercialise the technology faster.

Springwise has spotted other ways innovators are tackling our textile waste problem, including a social-impact repair service and a polycotton recycling process.

Written By: Matilda Cox