Speaker
Description
We present an integrated approach for high-precision sorting and characterisation of electronic components from E-waste by combining machine vision, X-ray absorption spectroscopy imaging and machine-learning classification. This study demonstrates the effectiveness of optical sorting based on machine vision coupled with classification algorithms such as convolutional neural networks (CNN). This combination allows similar electronic components to be efficiently grouped together, making them easier to recycle. In addition to optical sorting, X-ray absorption spectroscopy is being introduced to overcome the limitations of optical sorting by providing crucial information on the elemental composition of electronic components. The integration of these sorting methods into a single process, supported by the construction of a prototype, demonstrates the relevance of this approach, demonstrating up to 96.9% accuracy. The overall process offers the opportunity not only to group similar electronic components efficiently, but also to significantly enrich the final streams with targeted elements, enabling the recovery of previously lost elements due to their low concentration in electronic waste with elemental enrichments by up to 10,000 for targeted elements. This study opens the door to large-scale industrial application of the process, making it economically viable to recycle many elements of interest.