Detection and recognition of batteries on X-Ray images of waste electrical and electronic equipment using deep learning

The trend of increased use lithium-ion batteries, challenges the cost-effectiveness and safety manual battery separation during end-of-life treatment Waste Electric Electronic Equipment (WEEE). Therefore, need for novel techniques to separate sort batteries from WEEE is increasingly important. For this reason, presented research investigates potential facilitate development extraction sorting by examining technical feasibility predicting presence, location, type inside electronic devices with a deep learning object detection network using X-Ray images internal structure WEEE. To determine required X-ray imaging parameters, 532 were arbitrarily collected recycling facility. From each product, two Transmission (XRT) captured at different source configurations. Results obtained limited dataset are promising, demonstrating 91% true positive rate only 6% false classifying battery-containing devices. Moreover, precision 89% recall 81% demonstrated detection, an average 85% 76% distinguish amongst following six technologies: cylindrical nickel-metal hydride or nickel-cadmium, alkaline, zinc-carbon, lithium-ion, pouch button cell batteries. These results demonstrate on XRT-generated both automated sorting, regardless condition shape products.