Vision-based track and trace improves product integrity
Vision-based track-and-trace systems, which are also used in the pharmaceutical and medical device industries, automotive manufacturing, and general packaging, represent big profit potential. According to a report from MarketandMarkets, the track-and-trace solutions market is projected to reach $3.93 billion by 2023 at a CAGR of 18.9%. Machine vision original equipment manufacturers (OEMs) and integrators are capturing this market with a variety of solutions.
Food for thought
The food and beverage packaging industry is a top adopter of machine vision track-and-trace systems. This is driven by recalls such as romaine lettuce and food products containing allergens.
“Manufacturers need lot traceability to determine the point of origin of all ingredients, as well as where they are sent,” said Pierantonio Boriero, director of product management, Matrox Imaging. “Version control details the historical view of the recipe for each finished product lot to provide a traceable trail back to the original ingredients, their source, and the precise date and time they were added to the lot.”
One traditional use for machine vision is verifying the contents inside food packaging. This is a method i4 Solutions, LLC, has perfected over the last dozen years. The company’s automated package inspection system ensures contents are placed in the correct package. The vision system identifies the packaging, whether it contains a product code number, a photograph of the food, or stylized text indicating the flavor of the contents.
Machine vision has also been used “to capture information that shows the lot and source it came from and where it is at in the manufacturing process — for example, whether it’s just entering the plant or going to the shipping department,” said Brian Durand, president of i4 Solutions. “If a recall does happen, we can minimize its impact by knowing what was in that package and where it came from.”
Improvements include support for higher resolution cameras operating at higher inspection rates and interfacing with fixed barcode scanners from other manufacturers. The company has added technology to help prevent operator setup problems.
Even with the improvements, a big selling point for remains its ease of use, which is another growing requirement among customers. “With a little bit of training, an operator can easily configure a new product type or accommodate a design change,” Durand said.
Keeping medical devices on track
Like food packaging, pharmaceutical product manufacturing focuses on product integrity. Tracking and tracing products is vital to regulating and assuring quality products. This is especially true in cases where ingredients in a single pill might originate in multiple countries; all of them with unique regulatory codes.
When it comes to medications, the U.S. Drug Supply Chain Security Act involves product tracing, verification, and identification, with requirements for an electronic, interoperable system to trace products at the package level by 2023.
On the medical device side, the market for surgical instrument tracking is projected to grow at 18.6% over the next seven years. This has been buoyed by increasing incidences of instrument misplacement and hospital-acquired infections, according to Grand View Research.
Machine vision is often used to track unique device identifiers (UDIs). The unique device identification system tags medical devices through their distribution and end use using codes that are readable by both humans and machines. UDIs includes information such as manufacturing dates, batch and lot codes, serial numbers, and expiration dates to ensure traceability and verify all items contained within a kit are correct. But traceability is more than just track and trace; it also includes control and chain of custody, said Bradley Weber, product marketing manager at Datalogic.
Smaller components are starting to be tracked. “Every screw in a joint replacement kit, for example, will eventually have a UDI,” Weber said. “Without disrupting a surface finish, you can use the newest laser markers to print a code so small that you can barely make it out with the eye. Obviously, you have to have readers at higher resolution to be able to read those marks and track the items all the way into the body.”
Implementing advanced technologies
Machine vision companies are refining their products to keep up with increasing demand for track-and-trace solutions. Datalogic has developed an advanced optical character recognition (OCR) tool to read less-than-perfect characters, including those with a low gradient or low contrast between the letters and background. Weber sees more improvements to advanced OCR using deep learning. However, he cautioned it is an assistive technology rather than a replacement technology.
Boriero said Matrox sees the demand for track and trace manifesting as add-ons to larger vision systems. “Customers are looking for greater integration, with platforms and hardware components like smart cameras that can do more than simply read code,” Boriero said. “The technology is there. Now it is a question of augmenting the functionality of existing tools and leveraging these to create new track-and-trace solutions.”
Matrox’s existing tools are designed to address the challenges of reading dot-matrix text produced by ink-jet printers. They’re also designed to use geometric features to locate and read text in images where solid characters are separated both from the background and from one another.
“The market is poised to exploit technological advances,” Boriero said. “Customers want comprehensive, cost-effective solutions that provide reassurance of product integrity and safety.”
Meeting the consumer’s needs
Whether they want to know about food allergens or a shipment from a retailer, consumers are driving the demand for better track-and-trace solutions. Datalogic’s Weber cites Amazon as an example of the future of track and trace. “If you receive an item that is damaged, you have to go through all of the trouble of putting the object back in the box, taping the box shut, and sending an email to get a return label,” Weber said. “And because we are humans, we want it right, and we want it now.”
That’s why more track-and-trace solutions are becoming imaging-based. “Instead of just reading a bar code, machine vision allows us to look at the integrity of the box or for special graphics, such as the hazmat warning or ‘fragile’ symbol,” Weber said. “That tells Amazon to handle the box in a special way to keep it facing up at all times while automation is handling it.”
Imaging-based systems will eventually allow a consumer to see a picture of an object when it comes from the manufacturer into the warehouse before it is boxed. “And if that object appears damaged or the box has a smashed corner or open lid, your package can be routed for a second inspection to ensure that it isn’t damaged when it arrives on your doorstep,” Weber said. Time will tell if big retailers see a cost benefit in adopting such practices.
As consumers and global regulatory bodies increase their calls for complete transparency in the handling of products, machine vision companies can provide reliable, cost-effective, and easy-to-use track-and-trace solutions.
This article originally appeared in Vision Online. AIA is a part of the Association for Advancing Automation (A3), a CFE Media content partner. Edited by Chris Vavra, production editor, CFE Media, firstname.lastname@example.org.