Manchester AI expert helps local SME develop the technology to battle battery waste

A partnership between University of Manchester academics and Lion Vision, a North West-based Artificial Intelligence (AI) specialist, has made a breakthrough with successful launch of a product poised to revolutionise the waste and recycling industry. 

Research from Material Focus, the not-for-profit organisation funded by the waste electrical and electronic equipment (WEEE), found that “batteries that have not been removed from unwanted electricals cause more than 700 fires annually in refuse collection vehicles (RCVs) and at household waste recycling centres (HWRCs).” Batteries are also thought to cause an estimated 48% of all waste fires in the UK each year, with the cost to the UK thought to be in the region of £158 million annually. 

The team of entrepreneurs behind Lion Vision, along with the University, successfully applied to the Knowledge Transfer Partnerships (KTP) programme run by Innovate UK and was given a grant of more than £125,000 to assist in the quest to deliver a battery detection system. They partnered with Professor Hujun Yin, Professor of Artificial Intelligence in the School of Engineering, to bring their concept to life. 

The new technology has now been proven to reduce the existential threat of lithium-ion batteries and the environmental impact they pose within society and waste streams globally. The system combines advanced vision systems with innovative machine-learning techniques to detect, visualise and extract lithium-ion batteries and other hazardous items from the waste stream, using real-time analytics to identify where the flammable batteries are and how they should be removed. 

As waste passes underneath it, the Lion Vision system can analyse more than half a million images in a 24-hour window and detect more than 600 cylinder batteries per hour. While the system is currently focused on detecting cylinder batteries, it can be programmed to detect more than 40 battery subtypes and other hazardous objects such as vapes. 

The detection system is now in place at a range of sites across the UK, most notably at SWEEEP in Kent which processes 100 tons of waste electrical and electronic equipment (WEEE) per day. Typically, amongst this waste, the Lion Vision system is detecting more than 4500-cylinder batteries daily. 

Hujun Yin, Professor of Artificial Intelligence, based in the Department of Electrical and Electronic Engineering said, “My work in AI and vision systems has often given me insight into challenges that society faces, and this project was no exception. While policy change and progress should be pursued, we cannot underestimate the environmental damage that is being caused by lithium-ion batteries. It is our responsibility to find engineering solutions to these problems. I have no doubt that the system created by the partnership and the team at Lion Vision will have a significant impact on the waste industry.” 

Today’s news is an example of a University of Manchester Knowledge Exchange (KE) project, which match businesses with researchers, in order to increase the company’s economic growth. Manchester’s KE programmes are delivered by the University’s Business Engagement and Knowledge Exchange Team and can support companies at any stage of their project — from applying for funding, to project planning and evaluation. Its team of experts deliver opportunities through innovative and supportive schemes: Impact Acceleration Accounts and Knowledge Transfer Partnerships. 

Contact collaborate@manchester.ac.uk to discuss Knowledge Exchange further. 

---- 

Professor Hujun Yin's main research interests include AI, machine learning, deep learning, image recognition, and data analytics. Recent projects focus on developing deep learning-based vision systems for recycling industries, advanced machine learning for multispectral image analysis for early detection of plant viral infection, and data-driven surrogate models in engineering designs. He was a Turing Fellow of the ATI (the Alan Turing Institute) 2018-2023, a senior member of the IEEE since 2003, and a member of the EPSRC Peer Review College. He has been the Chair of the IEEE CIS UK and Ireland Chapter since 2023. He leads a team of 12 researchers working in a wide range of vision and machine learning challenges with strong emphasis on real-world medical, sustainable and industrial applications. 

Read recent papers: 

• Feature-Enhanced Representation with Transformers for Multi-View Stereo 
• High-Frequency Channel Attention and Contrastive Learning for Image Super-Resolution 
• A Divide-and-Conquer Machine Learning Approach for Modelling Turbulent Flows 
• A Machine Learning Based Approach for Selection of SF6 Alternatives 
• DRLFluent: A distributed co-simulation framework coupling deep reinforcement learning with Ansys-Fluent on high-performance computing systems 
• Manifold-enhanced CycleGAN for facial expression synthesis 

To discuss this research or potential partnerships, contact Professor Yin at hujun.yin@manchester.ac.uk.