Baking Lines Benefit from Vision Inspection (Snack Food & Wholesale Bakery)
This interview originally appeared in Snack Food & Wholesale Bakery, authored by Jenni Spinner. Click here to access the article on the publication's website.
Baking professionals are increasingly tasked with doing more with their equipment—and there is ever-increasing pressure to do it faster, more efficiently, more cost-effectively, and yield better quality. According to Yuegang Zhao, chief commercial officer with KPM Analytics, vision inspection technology can help move bakery operations in the right direction.
Jenni Spinner: First of all, I was wondering if you could tell us a little bit about some of the ways in which snack and bakery producers might incorporate vision and inspection technology and their processing and packaging environments.
Yuegang Zhao: KPM is a leader in quality inspection solutions for food manufacturers of all kinds. We as a company started in 2015 and over the years we’ve acquired companies with different technologies, but vision inspection technologies are at the center of it all. We are 100% focused on food and within that majority of the applications goes into the bakery and snack food industry.
Whether it's measuring the height of a dough before baking to evaluate the proofing process, looking at the color of a biscuit that's right out of oven, or looking at the number of sesame seeds deposited on top of a hamburger bun, there really is a wide variety of applications for vision inspection technology.
Now one thing with our vision inspection system is we also have automatic rejection capabilities that can be integrated into the overall system. If the system detects an aspect of the product that does not meet the quality standard, the system can automatically place that product into separate bins or conveyors to remove that product from the processing line.
We can achieve total 360-degree inspection of a product to achieve a comprehensive view of the product, its volume, its height as well as its color, presence of ingredients, and other customizable measurements. Our technologies truly capture every visual aspect of quality inspection.
JS: How has the technology evolved in recent years? Maybe you could talk about a little bit about what benefits and capabilities have been integrated or improved in vision and detection technology.
YZ: Vision inspection technologies have been used in the baking industry for decades, but in recent years, there have been several advancements in the technology that's really made the vision more powerful and accessible. First is the camera technology. There are much higher resolution cameras available in vision systems today that can capture products traveling at higher speeds with remarkable quality. Then there are advancements in different lighting techniques can help systems achieve more.
We detect things that's very difficult to do in the past. For example, there has been a growing trend among fast food chains to produce a hamburger bun that's noticeably shiny on the surface. Naturally, this shiny surface will impact the color measurement from the vision system. But with the right lighting structure, software, and camera setup, we can help companies achieve these measurements and satisfy customer demands.
Artificial intelligence is another emerging aspect surrounding vision inspection technology. AI makes product feature detections much simpler and easier for users. For example, let's say we’re measuring pizza. Then pizza has multiple objects like toppings, textures, colors, etc. In the past, you’d have to teach the software to detect individual ingredients like pepperoni, mushrooms, cheese distribution, and many others. The user would have to layer each of these measurements to complete the overall product measurement—this requires a lot of computing power and complexity. Now, with artificial intelligence, you can just train the software to immediately detect an acceptable pizza of with all these toppings. Over time, the software will start to really classify what's the high-quality pizza and what's the low-quality pizza. That just makes the user experience much, much simpler.
JS: Related to that, what are some of the challenges when a producer is integrating such equipment, either on a new line or an existing line? I.e., the equipment has to play well with the ovens and the other technology. What are some of the things that they should be looking for, and maybe the questions they should be asking when they're specifying new vision and detection gear?
YZ: There are a few fundamental questions we ask our bakery customers who are interested in implementing vision inspection technology into their operations. First, what are their overall objectives of using vision now? Sometimes, this answer is not always so simple. You may have the production manager say “We need a vision, but is it vision for reducing waste or is it vision for improving quality or is a vision for a process improvement?” Our customers really need data and if their objective is not clear, how do you measure success? For example, if you put a vision system with automatic rejection in the production line but you have not addressed other product quality- or consistency-related needs of the operation, what is this system really going to generate? It's going to create a lot waste because immediately it will see bad quality product and it will reject it. For example, if one baking operation has 20% of product being rejected because the operation did not address other quality tolerances of their operations, the customer may panic because the rejection margin is too high.
Our most successful customers take the time to understand how to achieve the ideal product and get to the route causes of what may be inhibiting their ability to achieve this ideal product consistently. This is why we also offer offline vision inspection technologies that can help companies establish a baseline for quality, then implement an over-line vision system to help deploy these standards on the production line before turning on the rejection system. Otherwise, the rejection system becomes a waste generator in the eyes of the production manager
The quality assurance and production teams at a bakery plant need to work together to establish agreed upon objectives for using vision inspection technology improve their process, their overall quality, and ultimately drive revenue for the organization.
JS: Did you want to talk a little bit more in depth about the specific solutions that KPM Analytics offers snack and bakery professionals?
YZ: We have a whole series of product lines for vision inspection. First, we have a line of benchtop vision inspection technologies for use in QC labs. The QC managers simply takes samples from the baking lab or the production line to measure product color, size, shape, etc. It essentially automates manual product measurement tools like calipers and color meters, significantly improving the productivity and establishing a more reliable baseline for product quality. Then, we also have an overline system that performs these same functions without disrupting the product flow. The camera system simply sits on top of the conveyor and measures products passing underneath, providing 100% inspection. With an overline system, users can get a lot more data and do statistical analysis and establish quality boundaries.
Then we have an inline system. These are integrated within the bakery’s conveyor system and include multiple cameras. All systems are customized to the bakery’s unique needs, including top camera for two- and three-dimensional measurements, a bottom camera to analyze the bottom of the product, and rejection systems to automatically remove out-of-spec products from the line.
Inline systems ensure only acceptable products proceed to the packaging machines so that we would call this inline system. There are many more customizations that can be made to accommodate various conveyor widths, product flow speeds, and then the size of the product.
Some applications may need higher resolution cameras to accurately detect small features. Or if it's a bigger product, you may get by with lower resolution for some of the products we have only two dimensions, like tortillas, where you are most interested in consistent size, color, toast marks, etc. But for operations like hamburger buns you’ll need more measurements, including top, bottom, and 3D, to achieve complete analysis. There’s a lot of options and a lot of different type of rejection systems we can use for different products.
Then, we also have a vision-related product for automatic lane balancing, or a “laner” as we call it. The laner is specifically designed for bagels and the bread that is before the packaging machine. We can optimize the flow of the product going into the packaging machine so that it does not jam the packaging machine.
KPM truly offers the complete package of vision inspection solutions to suit the needs and demands of bakery and snack food producers.
JS: You have a lot of options for people in our audience, but before I let you go, I was going to ask you to look into your crystal ball and based on your conversations and your expertise, are there any features that you think bakery and snack producers are looking for that might be added or enhanced in future and inspection and detection equipment?
YS: We do see more and more people like to incorporate other technologies into the whole vision inspection system. For example, in the baking and snack industry, moisture content is a major factor in product quality. In-line moisture analyzers can help the operator make sure their oven is balanced and also whether they have the right to moisture at the final inspection point.
We offer online NIR moisture sensor series and thermal imaging cameras that can be incorporated with our vision inspection system to simultaneously measure these important quality aspects. We also offer applications to detect unwanted foreign materials at virtually any processing phase.
Hyperspectral imaging is another emerging trend for vision inspection. Hyperspectral imaging analyzes products beyond visual wavelengths and into near infrared, so that the user can see the spectrum reflected from different material and identify challenging foreign materials like wood, paper, and others. Incorporating different techniques technologies into this overall vision inspection platform, that's what we see as a trend more and more in adopted by the industry.
