In attempts to optimize Thai Beverage Recycle’s (TBR) recycling process, a team of researchers from CMKL University and Carnegie Mellon University have developed an automated bottle screening process in collaboration with ThaiBev. Previously relying on the inefficient task of manual sorting, automated bottle processing is set to increase efficiency considerably. Using computational photography, artificial intelligence, and deep neural networks, a system has been developed that uses multiple-views of the bottle to achieve a highly accurate analysis. With accurate and efficient detection systems in place, more bottles will be processed and less bottles will be unnecessarily disposed of. The end result of the project is increased output resulting in a more efficient, cost-effective, and environmentally friendly process. In other words, everyone benefits – manufacturers, consumers, and future inhabitants of our planet. With climate change projections looking increasingly grim, the fate of our future is in our hands. It’s our responsibility to do everything within our power to mitigate the impact.
"What really excites me is taking the basic research of how we design pattern recognition and AI algorithms to look for defects in bottles," Savvides said.
The bottles meet a system equipped with four cameras that snap images of the bottles’ neck, bottom, and two sides. At the end of the conveyer belt, all bottles will be categorized by AI in real time. In addition, ThaiBev’s researchers gathered data on a wide variety of bottles from each of the categories and sent to Savvides and his team in Pittsburgh to train the AI algorithm model. The system mainly focuses on information about texture and minor details of bottle and filter it into ‘Reusable,’ ‘Clean before reuse,’ and ‘Non-reusable’ categories. The project is still in progress, but so far, the results have been positive.
This project is not only an opportunity to collaborate between CMKL University and Carnegie Mellon University, but it is a great chance to improve AI technology and potentially adapt into different factories.
