When looking at artificial intelligence (AI) and its uses on the manufacturing floor, it’s natural to first think of robotic automation and data collection—but there’s more to AI than meets the eye. Namely, vision AI.
“When you think of automation, a lot of people jump to the final step of full automation, no human in the loop. But to actually get there can be pretty expensive and pretty complex,” says Ed Goffin, vice president of product marketing at Pleora Technologies Inc. “And when you go on a shop floor, there is still a large reliance on human decisions and processes. With AI, there’s the opportunity, especially around decision support, to give those humans some tools so that their decisions are always consistent, reliable and traceable.”
Pleora wants to help manufacturers get the most out of AI, which is why the Ottawa-based company began development of its camera-based app suite Vaira. To help beta test the technology, the company turned to Dairy Distillery.
“[Vaira] is a turnkey solution with machine vision camera, lighting, edge processing and then behind that a suite of applications around visual inspection processes, traceability processes, assembly and checklists,” Goffin says. “And these apps come as a series of templates so that a quality manager can train their own workflow and train their own AI model.”
The distillery briefly transitioned into producing hand sanitizer during the COVID-19 pandemic. But when it returned to making the Vodkow cream liquor it was known for, Pleora was presented with a challenge.
Dairy Distillery is a small operation, producing about 1,500 bottles per day, while competing against global players with massive marketing budgets. The distillery is known for its uniquely shaped bottle—fashioned after a traditional milk bottle to reflect the dairy byproduct that serves as the product’s base—and eye-catching labeling.
The bottle has three independent brand elements. The main label and cap sticker are placed by a machine. A round, transparent emblem is then affixed by a human, who must align the sticker with branding on the other two labels. The manufacturer had been using a machine for this process, but was consistently seeing a high failure rate. While other steps in the process had been automated, it was easier and more economical to revert to a manual process for emblem placement.
“At one point, they had automated that full process. But the labeling machine was struggling with putting on the emblems, with something like a 50% error rate. So then they started using a human to put on the emblems, who was doing better, but you still wouldn’t catch the error until the products were on store shelves,” Goffin confides.
Over a long shift, the placement of the emblem often drifted from bottle to bottle. Production would have to be halted whenever employees found askew emblems during the packing phase. Dairy Distillery estimates that it would take about five minutes per bottle to remove and replace labels.
As it ramped up sales, it was vital for the distillery to maintain consistency in its labeling, lest customers question the quality of the product. If a poorly labeled item were to reach the store, it would reflect poorly on the brand. Retailers could also reject the product, resulting in shipping costs and possible waste.
The distillery is now using Vaira in order to improve consistency. With Vaira, an operator places the bottle under a camera, letting the app identify the brand elements on the machine-placed labels. On the display screen, the operator sees a real-time image of the bottle along with a virtual grid that guides the placement of the emblem.
The app can be trained for different brand elements, depending on the product. It is especially helpful for new operators and seasonal staff as they learn branding for the different product lines, according to Pleora.
Next, the distillery trained a Vaira visual inspection app for quality control (QC) checks. Generally, at the end of production, the distillery staff would manually inspect bottles before packaging to ensure brand integrity. The inspection app quickly shows if the brand labels are aligned within a certain tolerance with a fast pass or fail assessment. Prior to implementing Vaira, if an operator or packaging employee suspected a labeling error, production would stop while employees determined if the labeling was within tolerance.
The app recognizes the type of bottle and the required brand elements to verify everything is on the package and aligned, saving time for the production staff. The QC check also helps the distillery gather data around its processes to help pinpoint the cause of errors when they happen. This could be human error, an improperly setup machine or an automation issue. Then, they’re alerted and can address issues before they produce hundreds of faulty products.
With Vaira’s app-based approach, Dairy Distillery is able to more easily and cost effectively migrate toward the next stage in AI implementation, while collecting a database of “good” and “bad” images that can be used to develop and train machine vision algorithms for an inline inspection application.
“Pleora’s system is simple and easy to use, and helps remove ambiguity and stress for employees. As a QC tool, the inspection system helps increase our confidence in both our manual and automated processes,” says David Geros, chief operating officer of Dairy Distillery.
Utilizing the Vaira app suite for digitization, the distillery aims to expand its product offering and grow revenue without adding more workers. That means operators are finding and resolving errors earlier in their processes, as well as adapting the apps for their specific requirements.
By digitizing manual and paper-based packaging and shipping processes, the distillery cuts errors and time spent on such tasks. To optimize returns, Dairy Distillery uses Vaira to gather better data on how a product left a facility, ensuring improved product handling. By capturing a digital image of products as they leave the facility, and connecting those images with inventory and shipping records, the distillery can spot trends around damage and returns so they can improve processes.
“As the operator is doing their final visual inspection, the system is also capturing images along the way,” Goffin says. “Then they can enter operator notes into the app, and it’s all saved together so that if there’s an issue down the road with the product—and there’s a serial code associated with it—then they can just pull up that traceability report and say ‘this is what the product looked like when it left.’”
The data the distillery gathers around product returns, along with the data from visual inspection tools, also helps guide future automation investments. If Dairy Distillery shifts to inline inspection, the company can leverage digitized data to build AI and machine vision algorithms. Or it may find there are digitization opportunities in other areas, such as quality reporting or work instructions, that are easier to address and provide more immediate operational benefits.
“I think people are starting to recognize that there is an opportunity to start an automation process by using AI and vision to just give humans the decision support,” Goffin says. “And when you do that, then you can start to gather the data and the images, and even the employee buy-in. That then helps you take the next step toward automating a process.”
For more information about Pleora Technologies, visit pleora.com or call 613-270-0625. For more information about Dairy Distillery, visit dairydistillery.com or call 613-256-6136.
Connect With Us
