Logistics industry looking to improve with machine vision
Logistics operations must be able to process massive databases of SKUs and information and enhancements in machine vision can help improve efficiency.
E-commerce has forced the logistics industry to take inventory of how it applies automation and machine vision. Where distribution operations once shipped comparatively uniform goods in relatively uniform packaging, they now must process boxes, parcels, polybags, and jiffy bags, as well as ship-in-own-container products that come in countless colors, sizes, and geometries. Additionally, as more products become available online every year and as consumers increasingly expect those products to be customized and delivered within days, logistics operations must be able to process massive databases of SKUs and information.
Expectations for customization and fast delivery often compete against each other in the warehouse, said Bradley Weber, product market manager at Datalogic.
“Customers want the ultimate in product flexibility, and they also want it delivered in two days — a challenge indeed,” he said. “Datalogic helps make this a reality by delivering logistics solutions with higher read rates, greater accuracy, and optimized pattern recognition capabilities that enable manufacturers and distributors to customize on the fly. Such solutions are becoming increasingly important for enterprises that rely on vision technology to track massive inventories of subcomponents so they can quickly bring them all together for last-minute assembly into a final product.”
For decades, bar code scanners were the workhorse of the track and trace industry, and they continue to evolve to keep pace with changes in distribution. But the need to collect more data more quickly from more packages has prompted more advanced vision technology to become a fixture in many distribution operations.
“If you look at Cognex over the past 5 to 7 years, logistics has been one of our major growth engines,” said Ben Carey, senior manager, Logistic Vision Products at Cognex Corp. “That’s stemmed from our ID business. When we first went into logistics, the industry was predominantly using laser-based sensing technology to read bar codes. Initially, competition thought we were crazy for trying to do this with area scan cameras — the same type of technology that’s on your smartphone — because there were several technical challenges to overcome, such as depth of field, lighting, speed, etc.”
Overnight (delivery) growth
Cell phone technology progressed to offer not only higher resolution and smaller form factors for pixels and chips, but also the improved processing power and lower costs required to provide a strong foundation to become an industrial solution. The potential that embedded imaging held for the changing logistics industry was clear.
“The step from analog to digital gave the industry increased performance along with some accountability and process optimization,” Carey said. “Whenever, for example, a laser-based solution fails, you don’t have anything you can post-analyze. If there wasn’t a label on the product, you have no further input other than it was either read or it wasn’t read. So, being able to do area scan imaging of a bar code means that if you don’t read it you still have an image available for investigation.”
One basic benefit of post-analysis is redundancy. “If a parcel containing valuable goods enters a distribution center and its barcode label gets ripped off or damaged, then you don’t know where to tie it back to or where it should go,” Weber said. “It becomes lost money. Now you have to ship it back, or you have auction it off.”
What many companies are trying to do, Bradley added, is to capture images of a package either when it enters the facility with a barcode label intact, or even when the label is applied. “If I store all that data together and the label gets separated in the chain of custody, I still have a better chance of tying that package’s information back together,” he said.
Scanning, imaging and beyond
Despite imaging’s comparative differences to laser-based bar code scanners, the two remain complementary rather than competitive solutions. Lasers often have an advantage when reading codes at a distance, in motion or applied somewhere on a large package, such as a refrigerator box. But the technology is migrating toward imagers, said Caleb Zwar, industry marketing manager for Datalogic, and imagers are one step away from machine vision. The solution you select comes down to what you want to do with the image data. He compared the choice to decide between an iPad, a laptop, or a desktop. Both can surf the internet, access web content, or live network. But while the iPad is streamlined for consuming content, a laptop may be better for doing homework or managing the family budget.
“It’s the same thing with 1-D or 2-D barcode ID readers,” Weber said. “Can you do optical character recognition [OCR] with them? Maybe a little bit. But if you want to do advanced OCR, you have to use more advanced vision technology to train the font, apply different font libraries, or to read fonts of different colors or types on the same package. You need a machine vision system with programmability built into the system to do that.”
As users compare the relative versatility of scanning, imaging, and vision at the system level, component advances are also injecting greater versatility into the mix – and smartphone technology continues to be among the sources for innovation. The high dynamic range (HDR) feature on phone cameras, for example, captures a series of images to form a composite that highlights more distant objects and darkens those that are closer. Applied to a conveyor line with, say, a small package closely following a tall one down the conveyor, HDR can enable clearer images of the space between the two packages where a bar code might otherwise be hidden. The drawback: HDR historically hasn’t handled moving targets well.
Liquid lens autofocus technology has also seen increasing use in logistics applications. By applying electrical charges to change the shape of the interface between oil and water, these lenses help overcome the depth of field challenges when imaging packages on a conveyor line. As the travel speed of package conveyors has increased, liquid lens technology has evolved to keep pace.
“We’ve been working for the past five years to develop a liquid lens technology that could adjust frame-by-frame,” Carey said. “This enables you to dynamically focus the camera in all situations, which gives you pretty much indefinite depth of field to capture one bar code that’s really close to the camera and a subsequent one 3-4 meters away.”
Capturing parcel dimensions is also becoming important for logistics as shipping packages come in an increasing range of shapes, surfaces, and proportions, and as distribution operations employ dimensional-based pricing models. One product leverages symbolic light technology (SLT), in which a light pattern is captured in a single frame before a 3-D decoding algorithm finds individual symbols in a pattern to deliver sub-millimeter depth precision. The company’s 3D-A1000 Dimensioning System handles high-speed motion, uneven transitions, curved conveyance, and colorful target surfaces.
Coming next down the conveyor
E-commerce has driven the evolution and growth of vision technology over the past decade, but future growth may come from a surprising quarter: the conventional brick-and-mortar sector. While online stores have knocked large retailers on their heels, the incumbents are awakening to the one major advantage they have in this fight. With a nationwide network of local stores, they own the last mile to the customer, making them less reliant on third-party shipping services.
“Brick-and-mortar retailers don’t have the automation infrastructure that e-commerce companies have, but that’s something that they can change,” Carey said. “What we’re seeing now is the reawakening of the retail market. In terms of general automation, one technology that’s coming up quite often is automatic storage and retrieval services (ASRS). They’re certainly looking at similar solutions that e-commerce has been leveraging to ask themselves, ‘What are they doing that we can’t do? Maybe it’s time to put pen to checkbook.”
Dan McCarthy is contributing editor for AIA. 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.