For decades, Pallet Track® enterprise resource planning software has helped pallet companies manage their operations, inventory, billing, accounting and more. The next generation of Pallet Track is coming. And it uses IP cameras connected via Ethernet cables to gather data from images based on AI modeling.
Alan Miceli, director of automation systems at 48forty Solutions and the manager of the Pallet Track software division, said, “We’re in the development stage and proof of concept right now. We have developed an AI camera-based system that takes images and converts that into usable data that can seamlessly interact with Pallet Track software. This system is currently being tested in multiple 48forty facilities right now.”
Just like the Pallet Track software, the Pallet Track AI Vision System will be made available to customers very soon. Miceli added, “We have done some similar things to what other suppliers have done, but we are using some different technology that significantly reduces costs.”
Pallet Track is a division of 48forty Solutions and has offered solutions to the pallet industry for more than two decades.
Pallet Track Adopts Artificial Intelligence Technology
Miceli noted, “We are now able to track production in real-time that was difficult to track automatically.” This approach uses IP cameras, similar to what companies use as security cameras, that are connected via Ethernet to a central hub. Images from the cameras are fed into an AI model that translates the images into usable data. This information is then put into Pallet Track and can be accessed anywhere via a smartphone, tablet or laptop.
The initial tests use images from a camera above a bandsaw dismantler to track production. Miceli explained, “We’re able to count the pallets going through a bandsaw dismantler. This allows us to automatically calculate production per hour and pay the workers piece rate.”
Miceli added, “With the camera system, it takes production tracking to a whole new level. We don’t need scanners or barcodes. We don’t need to validate it because we’re using the vision to identify that an operator just took a pallet apart.”
This technology can be used to track production on other machinery too, such as saw lines, repair stations, nailing machines and more. All the data goes into a switch on a network the same as a Pallet Track kiosk, so that it doesn’t take any bandwidth away from the facility.
Is This A Better Approach?
Miceli believes the approach being deployed by 48forty is better than other methods. A big advantage is that the IP camera/Ethernet approach eliminates the need to use cellular technology and cameras powered by lithium ion batteries. He explained, “We are not using cameras that run on lithium ion batteries, which can be a fire hazard. Sort of like what the airlines are checking for now on planes.”
The Pallet Track Vision System uses IP cameras that are tied into the infrastructure and are wired into the electricity just like any other piece of machinery. The vision system communicates with the kiosk, basically like a scanner where you would scan a barcode and validate it. This new approach uses camera footage to validate what is going on and sends that information to the kiosk and the Pallet Track software. This new approach basically replaces bar code data collection with cameras.
Miceli noted, “Since we don’t use cellular phones and lithium ion batteries, you don’t need to replace the batteries or pay for a costly monthly cell phone plan for each device.”
Why is data capture so critical? The reason is that data will drive future automation projects and technologies. Companies with better data will develop better processes and monitor trouble spots more easily. “Data is the key to driving efficiency in pallet operations today and into the future,” suggested Miceli.
Ever Evolving Technology and Capabilities
What about the accuracy? Is the new Pallet Track Vision System battle-tested? Miceli commented, “We have developed an AI model that we are testing now in pallet facilities. We believe our approach will be as accurate as anything else out there because the camera data can be connected to information in the Pallet Track software. This approach provides a more accurate picture of what is happening in real time.”
Miceli clarified, “Cameras can only track and monitor what they see. So, unless cameras are used on the sides as well. It can be difficult to monitor the thickness of boards. But if we use cameras to monitor a saw line and know the thickness of a particular run based on what is in the Pallet Track software, we can provide really accurate yield data, better than just using cameras alone.”
While a dashboard can be great to allow for remote monitoring, seamless integration with your Enterprise Resource Planning (ERP) software is important. Miceli pointed out, “Dashboards are nice. But if that data doesn’t seamlessly flow into your ERP, the process creates extra work for your team. Our approach seamlessly integrates AI data into Pallet Track. This eliminates data re-entry or human errors.”
Pallet Track’s Vision System fully integrates with the Mill Manager™ software. This allows users to calculate compensation, adjust inventory, manage production schedules, generate reporting per man-hour, confirm processes and integrate into your accounting and payroll systems. This integration adds real value to your vision system, optimizing operator performance and identifying true production time per employee.
Like any other real AI-based software, the Pallet Track Vision program is based on a model developed by our programmers. The better the model, the better the results. The sky is really the limit in what the vision system can monitor.
Miceli predicted, “Eventually the software will be able to grade pallets. This operator produced so many As, so many Bs. I believe this quality monitoring with cameras is especially a big benefit for companies that repair CHEP or PECO pallets because the rental companies don’t want a barcode put on their pallets in the repair process.”
Is there still a use case for barcodes? Yes, it all depends on how many cameras are deployed to monitor quality. Miceli said, “But there’s still going to be a use for barcodes. I feel the AI cameras, at least initially, are more for tracking production than monitoring how many of each grade are produced. I think a combination is probably best for a repair line. The repairer labels his pallet with a barcode and a grading system grades and transfers the pallet to the appropriate stacker” “This technology is available right now by incorporating our grading system into your repair line.”
In the not too distant future, this technology will be able to detect pallet quality once the models are developed. “We can do the same thing as the pallet inspection solutions offered by other suppliers. The big difference is that our solution will be a lot cheaper.” It may involve many IP cameras in a configuration where each camera costs initially $100 per camera and monthly software fee versus an inspection machine that can cost a half a million dollars or more. Miceli stated, “The secret for everyone is their AI models. And we think our technology across our network will help us develop very good models.”
There are two primary technologies to detect various measurable actions, such as boards produced or pallet types: laser scanners and sensors or cameras using vision systems. Some technologies deploy both because each methodology does certain tasks better than others. Miceli explained, “Keep in mind that cameras can only see what a human eye could see. It can’t see the interior of the pallet stack to determine quality.” Even when a human operator is looking at a pallet up close, he/she can still miss things.
While Miceli said that you can develop a basic model that works well in a lot of pallet plants, there are still some conditions that may require special customization. Miceli warned, “You would have to build a model that is tailored to each manufacturing environment because things, such as, your lighting conditions or vantage points, really matter.”
Artificial intelligence is still pretty basic in what it can handle right now. But the technology is improving at an exponential rate. Miceli clarified, “AI is still pretty basic in its ability to differentiate conditions. I mean it’s getting there, but it can’t really compare to human intelligence where I know that the lighting is different and that’s why this color shade is a little bit off as opposed to if you shine a bright light on it; we know that we don’t even think about it.”
Why ERPs and Costs Matter…
The Pallet Track approach sees vision systems as a critical data collection tool to improve data gathering for ERPs, not a replacement for those computer systems. Miceli warned, “If pallet companies get rid of their ERP system and just put in vision telematics, they are going to go backwards from a technology perspective and data capabilities. Vision telematics can’t manage sales orders, accounting or business management functions.”
A big advantage of vision systems is the ability to have real-time accountability based on what is going on in the production environment. Miceli said, “Cameras add accountability, which is critical to drive better performance.”
A key benefit of the Pallet Track approach is the total system costs. Miceli explained, “Our approach using Ethernet cables and IP cameras cost less all the way around.”
Miceli suggested, “We’re doing AI technology similar to the competition at half the cost.” The Pallet Track AI Vision System is only available to pallet companies that use Pallet Track ERP software. It is not being sold as a stand-alone product. Miceli commented, “What we offer is more than just a camera system. It is a fully functioning ERP with data coming from IP cameras that are run through an AI model.”
Pallet Grading at a Fraction of the Cost
The Pallet Track team recently developed an autonomous grading system for any brand, sort and repair lines. The first system shipped and should be installed by the time this article is published. Miceli said, “We combined our Pallet AI with high resolution scanners to accurately pre-sort pallets from a sort line.”
Pallets are tipped and fed onto the line. Each pallet is scanned by the system to identify its type and grade. Pallets graded AA, A, and B are directed to the stackers, while broken pallets are sent to the repair line. Pallets with odd sizes are automatically removed from the line, eliminating the need to sort them out before loading the tipper.
Miceli explained, “Everything is graded without human intervention. The accuracy is just amazing, and the scan time allows for about 900 pallets per hour. Pallet Track offers a complete scanning section that can be incorporated into any brand repair/sort lines.”
For more information on Pallet Track’s new AI Vision System, visit www.pallettrack.com or call (631) 244-0069.
