Fairlife: AI vision inspection for 100% automated packaging inspection of thousands of yogurt containers per day
- Ryosuke Murai
- Oct 24
- 3 min read
For dairy and other food manufacturers, improper lid installation and misprinted labels are serious issues that directly affect the brand and safety. Fairlife (a US premium dairy brand funded and owned by Coca-Cola), which we will introduce here, introduced Elementary's AI vision inspection system for the quality inspection of yogurt containers , automating the detection of defective labels and lids 100% , reducing defective products to almost zero. Moreover, this system was implemented and operational in just a few days , solving all at once the problems that conventional rule-based inspections could not address.
#Fairlife and background to its introduction
Fairlife is known as a dairy brand that focuses on nutritional value and quality, and is rapidly expanding in the US market as part of the Coca-Cola Group. On a newly launched yogurt line, there were frequent issues with lids not being properly attached, misaligned printing, and missing characters. Manual visual inspection alone was prone to oversights. In particular , conventional rule-based inspections failed to properly determine pass/fail status if the assumption that "printing should be in a certain position" was misplaced, resulting in unnecessary trouble on-site.
#The effectiveness of implementation in numbers
Automatic inspection of thousands of yogurt containers per day
Label inspection automation rate: 100%
Completed in just a few days
Almost zero missing lids and printing defects
#Overview of the AI solution introduced
The Elementary AI vision inspection system adopted by Fairlife combines cameras and sensors to photograph yogurt containers from multiple angles, and the AI instantly determines whether the lid should be attached and whether the label should be printed.
AI can determine that "it's OK even if the marking position is slightly off from the correct position." It can also detect deviations in the printing position, which are difficult to address with rule-based methods, and only detects errors that fall outside the required range.
Easy to integrate into production lines . The compact equipment can be installed and operational in just a few days. It checks for the presence or absence of lids and the readability of printed text in real time, automatically rejecting defective products.
Accumulating data makes it easy to track down the cause. Recording what kind of defects occurred and when helps improve the process.
#A story for getting results with AI
(1) Flexibility that surpasses conventional rule-based inspection
The rule-based method could not deal with this problem, as it would make a false judgment if the position of the printed code was slightly shifted. However, the AI learns the entire image and allows for "misalignment within an acceptable range," detecting only the slightest defects or faults.
(2) Covering mass production lines with short-term operation
While many AI systems are thought to require long preparation times, Elementary can be set up in just a few days. It was piloted in the midst of mass production and was able to quickly improve the defect rate.
#Suggestions for small and medium-sized enterprises
AI can be introduced without large investments or long-term preparations Examples of installation and operation in a short period of time are feasible even for small and medium-sized factories.
AI vision is more effective in processes with greater deviations and variations AI can supplement minute defects and errors that would be easily missed by humans or rule-based methods, improving the accuracy of quality assurance. It can also flexibly respond to ever-changing designs.
Reducing defect rates directly impacts brand value and costs. Defective products on the market damage a brand, but AI inspection can eliminate defects early, significantly reducing the risk of recalls and complaints, leading to cost savings.
The Fairlife example shows that even in the dairy industry, where quality requirements are strict, it is possible to build an AI inspection line in just a few days, automate 100% label inspection , and virtually eliminate packaging errors. Even small and medium-sized Japanese food manufacturers and factories that feel they have reached the limits of manual inspection can use this type of AI vision inspection to improve the quality of their mass production lines while reducing costs.
Quest AI provides efficient support based on overseas case studies for companies looking to quickly implement AI. Why not start by trying out AI in some processes and consider ways to improve product quality and brand image?
[Note: What is Elementary?] Elementary is a company that provides AI vision inspection solutions for the food and manufacturing industries. It can flexibly handle subtle misalignments and complex defects that traditional rule-based inspections struggle with, and can be implemented in a short period of time (just a few days). Its unique feature is that it combines cameras and AI algorithms to accurately assess label and print quality, package attachment, and more, maintaining high accuracy even on mass production lines.
