Lessons learned working on xBat+

Ailbhe O'Connell has been working on the project for the last 12 months, scanning devices and gaining experience with complex datasets.  She shares her thoughts so far:

"Working on the XBAT+ project has provided me with valuable hands-on experience in developing the image datasets that support AI-based battery detection in WEEE recycling. My role has included x-raying and photographing WEEE devices, systematically naming and organising the image files, documenting x-ray acquisition parameters and device information, and using Roboflow to annotate images for AI model training. 

Throughout this process, I learned that creating a reliable dataset depends on careful attention to detail. Minor variations in device positioning, image quality, file naming or annotation can all have an impact on the final dataset. Additionally, working with such a wide variety of WEEE items also highlighted the fact that there is rarely a one-size-fits-all approach, often requiring adjustments to imaging parameters and scanning procedures to achieve the best results. It has been rewarding to see how these small but important steps come together to produce high-quality data that underpins the development of reliable computer vision systems, ultimately contributing to safer and more efficient WEEE recycling."