How Automated Bottle Screening Is Reducing Environmental Impact

“Reduce, Reuse, Recycle.” You may have heard these words before as part of the waste management hierarchy, a tool that is commonly used in sustainability efforts. What’s often misunderstood is that “the 3Rs” are meant to be followed in that specific order. Thai Beverage Recycle’s (TBR) plants are looking to increase the efficiency of current bottle screening processes, resulting in more glass bottles being reused instead of unnecessarily recycled. With glass currently accounting for 3% of Thailand’s waste composition, this is a number that can easily be reduced. While some bottles must be recycled due to contamination or damage, reusing bottles instead of recycling them decreases costs as well as creates less waste. 

 

In collaboration with ThaiBev, researchers from Carnegie Mellon University Thailand (CMKL University) are developing an approach that is able to screen glass bottles efficiently and classify them automatically. The main objective of this project is to implement a system that can successfully screen the bottles and sort them into the following categories: reusable, reusable only after cleaning, and non-reusable. Specific criteria are established for each category and the bottles are sorted accordingly. With the current bottle screening process being the primary bottleneck of the reutilization process and restricting output, a large number of bottles are being unnecessarily recycled instead of reused.

Carnegie Mellon University Thailand researchers have developed a multi-view data-driven approach to address the problem. Using computational photography and artificial intelligence, a system has been developed that uses multiple views of the bottle to achieve a more accurate analysis. Paired with deep neural networks, this system is able to process information with an extremely high degree of accuracy, similar to the way a human brain would. Unlike a human, the system is also extremely efficient. After testing a variety of different camera and lighting setups, researchers discovered a final hardware system that was able to accurately assess the reusability of the bottles paired with the software.

 


With a suitable hardware setup established, researchers have begun collecting large amounts of data to train AI systems to detect defective bottles. With accurate and efficient detection systems in place, less bottles will be unnecessarily recycled and reused instead. Paired with a faster screening process, the number of bottles unnecessarily produced to meet demands will be reduced. 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.

In hopes of developing more advanced technology and reducing their environmental impact, TBR has been working with Carnegie Mellon University Thailand to conduct research and development. From artificial intelligence and machine learning applications, to innovations in cybersecurity and entertainment, Carnegie Mellon University Thailand is collaborating with industry leaders to help the nation achieve Thailand’s 4.0 initiative. With ongoing research projects in a variety of fields, Carnegie Mellon University Thailand is confident that its research can contribute to the development of Thailand and the ASEAN community. To learn more about Carnegie Mellon University Thailand’s graduate programs or research projects, visit here. To learn more about ongoing research collaborations between CMKL University and ThaiBev, visit here.


Blog Post written by:
Michael Beddie
Intern