‘Deep learning’ boosting sorting Down Under

Monash University in Australia is pioneering a detection solution that could advance the recycling of building and construction waste.

The work envisages the latest robotics and ‘deep learning’ automation sorting waste more accurately and efficiently. This includes the recognition of materials by artificial intelligence (AI).

PhD candidate Diani Sirimewan, a civil engineering expert, is leading the R&D project. She points out that less manual sorting increases the safety of workers who are frequently dealing with contaminated or hazardous materials. Another benefit of smarter sorting is identifying potentially toxic components more easily, she says.

Sirimewan believes her research is the first to capture detailed images of dense construction and demolition waste inside bins on construction sites. This enables her team to build significantly advanced models capable of detecting and recognising waste buried within other rubbish.

Monash University is undertaking a simulation with robotic arms. The researchers hope their findings will spark new investment in robotics and automated solutions.

‘Our deep learning models showed the remarkable ability to recognise the composition of construction and demolition waste streams, including the identification of contaminants,’ Sirimewan reports. She believes it’s ‘exciting’ how better-quality recycling using this kind of AI technology could significantly reduce the volume of waste sent to landfill.

More details about the R&D project can be found here.