Brought to you by Antonis Porichis from University of Essex
Our everyday lives are increasingly flooded by all sorts of electronic devices from smartphones and tablets to toothbrushes and scales. Once they reach their end-of-lifetime these are disposed as electronic waste (WEEE). Handling this ever-expanding waste stream presents a critical environmental and safety challenge.
Among the most pressing issues is the risk of fire caused by lithium-ion batteries that are improperly disposed of within WEEE waste streams. These batteries, if not detected early, can lead to devastating fires in recycling facilities, causing property damage, operational disruptions, and risks to workers.
X-ray technology can help see through the device’s casing while it moves on a recycling facility conveyor belt. Yet, a crucial challenge remains; is it possible to detect the presence of a battery among the vastly diverse hardware components within an electronic device and do so fast enough to remove the device in time?
Using Object Tracking AI for Battery Detection
AI has shown impressive capabilities in terms of speed and accuracy of visual identification tasks. AI models are designed to analyse images and identify specific items, even in highly cluttered scenes. Such models are what power our smartphone apps’ capabilities to seamlessly blend the digital and physical world with funny face filters and animated effects in live video. AI algorithms are able to locate the user’s face in real-time so that the animations are overlayed at the correct position.
In recycling facilities, where various electronic devices are mixed in large quantities, the challenge lies in identifying potentially dangerous lithium-ion batteries within the device casings and distinguishing them from other, harmless hardware components. AI models with similar architectures to those making social media apps fun to use, can be adapted to recognise these batteries in X-ray scans, and be able to process multiple images per second.
GRINNER introduces novel, high-resolution X-ray technology to examine the internal structure of WEEE waste, providing detailed images that reveal components not visible from the outside. AI models are then trained to recognise the unique characteristics of batteries, learning from the acquired X-ray datasets that include numerous examples of electronic waste with and without batteries. Once trained, these models can quickly process new X-ray images and flag potential battery locations with high accuracy and speed, allowing operators to take action before batteries are accidentally shredded or crushed, preventing fires and ensuring safer recycling processes.
Enhancing Safety and Efficiency with AI
The integration of AI for battery localisation in WEEE recycling offers several important advantages. First and foremost, it greatly improves safety by mitigating the risk of fires caused by undetected lithium-ion batteries. By automatically identifying and isolating these batteries before they reach the shredding stage, AI helps protect facility workers and reduce the likelihood of accidents that could lead to significant damage or injury.
Additionally, AI models can dramatically increase the efficiency of recycling operations. Manually identifying batteries within mixed electronic waste is a slow and error-prone task, especially when dealing with large volumes of material. AI systems, by contrast, can process and analyse images in real time, allowing facilities to handle more waste with greater accuracy. This streamlines the entire recycling process, reducing bottlenecks and improving throughput.
Conclusion
As AI models continue to advance, their application in niche industrial contexts, such as WEEE recycling, is proving to be both innovative and essential. By combining AI with X-ray imaging, recycling facilities can localise and detect hidden batteries within electronic waste, significantly reducing the risk of fires and improving operational safety and efficiency. The adoption of these technologies marks an important step toward a safer, more sustainable future for the recycling industry, where critical materials are recovered with minimal risk to both people and the environment.