top of page

The latest news, trends, analysis, interviews and podcasts from the global food and beverage industry

FoodBev Media Logo
Access more as a FoodBev subscriber

Sign up to FoodBev and unlock more insights from the international food and beverage industry. Subscribers have access to webinars, newsletters, publications and more...

Ensuring safety and quality: How vision systems transform the F&B industry
Siân Yates

Siân Yates

8 July 2024

Ensuring safety and quality: How vision systems transform the F&B industry

Vision and inspection technologies are indispensable in the F&B industry, fulfilling critical functions like quality control, product safety and traceability. Whether it’s swiftly scanning a barcode for consumer transactions or meticulously adhering to regulatory standards, vision systems play an integral role across various phases of production, processing and distribution. FoodBev's Siân Yates examines the latest innovations in this essential arena.


Consumer safety is the most important factor in the food and beverage industry. Vision-based inspection equipment ensures rigorous quality control, rapid identification of contaminants and potential safety hazards.


These technologies offer a wide range of advantages. “As a food producer, you want to provide your customers with products of the highest quality,” said Jan Gärtner, product manager for Halcon software at MVTec. “Precise visual inspection is essential to ensure that the products are in the best shape and that no contamination, damage or other defects reach the consumer. These technologies provide valuable support, as they offer high precision, 100% objectivity and operate at very high speeds.”


Nathaniel Hofmann, market product manager for machine vision and measurement systems at German sensor manufacturer Sick, noted that consumer confidence hinges on products meeting stringent quality standards. “As consumers, we frequently base our purchasing decisions on visual appeal, and vision systems help differentiate between high-quality and subpar products,” he commented.


“Vision inspection goes much further, though,” added Jeremy Sacco, senior manager of global content marketing at Cognex, a US-based manufacturer of machine vision systems, software and sensors. “A well-implemented system can detect problems with fill levels, cooking times, toppings and add-ons, and more – all kinds of issues that can lead to unsatisfactory final products.”


© Cognex

Metal detectors, checkweighers and X-ray inspection systems are all essential to any production process, as Phil Brown, managing director at Fortress Technology Europe, told FoodBev. “Establishing the biggest contaminant risks and most cost-efficient critical control points and inspection solutions all help to ensure a robust HACCP-compliant food safety strategy that delivers maximum profitability and prevents costly product waste.”


The rise in legal requirements and increasing certification pressures from major retail chains are also compelling F&B manufacturers to prioritise product quality, packaging integrity and food safety throughout the manufacturing, filling and packaging processes.


“Consequently, it’s imperative for producers to subject each individual product to rigorous quality checks and in-line inspections to meet these standards,” cautioned Dirk Henschke, product manager for X-ray inspection at Heuft Systemtechnik.


Advanced applications


“Machine vision technology can benefit anyone, from Fortune 500 companies to craft brewers, across all segments of the F&B industry,” said Dave Coleman, global product manager at Filtec – a provider of food, beverage and pharmaceutical inspection machines. “It can ensure your product quality, consistency and help reduce downtimes along with protecting from recalls.”



Integration of an inspection system is feasible at virtually any point along any food or beverage processing line. Each sector and product application presents its own unique set of prevalent risks. “Proactively mitigating contaminant risks early in the production process leads to significantly fewer rejects compared to inspecting finished, packaged products,” said Fortress’ Brown.


Brown continued: “Moreover, it empowers processors to pinpoint the source of potential issues. Instances where ingredients are added or multiple mixing, rolling, cutting and baking processes occur pose an elevated risk of contaminants entering the production line.”


“The spectrum ranges from fill level checks, which ensure that the end consumer does not receive less than they have paid for – via the identification of critical defects that threaten the integrity of the packaging and the microbial purity of its contents – to the detection of dangerous foreign bodies such as glass splinters in the product,” said Heuft’s Henschke, adding that these are the cause of costly liability cases and product recalls that damage a company’s reputation.


Henschke highlighted that the beverage sector is benefiting from such technologies due to the growing popularity of circular economy and returnable systems – two major topics in the industry. “In instances where glass bottles are refilled multiple times, both an all-surface empty bottle inspection and a thorough full-container inspection at the end of the line are crucial,” he added.


Heuft is prominent in both domains, offering advanced in-house optical inspection technologies as well as pulsed X-ray technology for challenging scenarios such as detecting minute glass splinters in both empty and filled bottles.


Modular Heuft systems integrate various detection technologies into a single device. For instance, optical identification can pinpoint anomalies like mould or transparent film in bottles, while X-ray technology detects glass splinters or metal fragments.


“Meanwhile, in the food filling and packaging process, vision inspection technologies are used for label inspection, best before date and barcode verification or for checking closure security and integrity, among other things,” Henschke commented. “However, vision technology is not yet a sufficient solution for finding the smallest foreign bodies in non-transparent packaging or products such as tin cans or baby food.”


A deeper look


According to Sick’s Hofmann, there are two significant advancements currently driving the vision inspection space: the increasing prevalence of AI image processing, and the growing accessibility and user-friendliness of 3D imaging technologies.


“AI holds vast potential across industries, with applications ranging from AI-controlled toothbrushes to everyday smartphones. However, its impact is particularly profound in the food industry,” Hofmann said. “Deep learning classification proves invaluable in distinguishing between organically diverse yet identical products, while anomaly detection allows for the identification of deviations from known quality standards. For instance, training a vision system to differentiate between various types of apples and assess their quality ensures consumers receive consistently high-quality products."


He explained that 3D inspection is particularly beneficial for items produced in trays or molds, while

advanced technologies such as time-of-flight (TOF) and laser profile sensors offer enhanced capabilities.


Hofmann continued: “Smart tools integrated into camera devices eliminate the need for additional

PCs and cabling, streamlining installation and deployment processes. This not only simplifies operations but also enhances repeatability, ensuring consistent product quality for consumers.”


The company has developed Sick Nova image processing software, incorporating AI and 3D image processing technologies. This software offers a unified and intuitive development environment, ensuring consistency across 2D, 3D and AI processing. The interface remains consistent across different hardware, enabling smooth scalability of applications and allowing users to transition between technologies seamlessly while maintaining a familiar development process.


The introduction of the Sick Inspector 8xx family of devices showcases advancements in sensor technology and on-camera processing chips, enabling efficient utilisation of AI capabilities at high speeds. The Nova software supports a wide range of 2D cameras, from VGA to 12MP resolution.


“Employing Nova’s 3D tool with Sick 3D cameras like the Visionary T Mini (a TOF device) and the Ruler laser triangulation sensor represents progress in on-device image processing options,” Hofmann explained. “With the recent launch of the Ruler 3000 family, we can now cover items ranging from 26mm wide up to 1.6m with a single model, expanding our capabilities.”


Meanwhile, Heuft’s in-house hardware and software for real-time image processing have classified detected objects to differentiate harmless deviations from critical defects. With its latest innovation – Heuft reflexx A.I. – the team can employ deep learning to detect foreign objects that were previously challenging or impossible to identify. Launched at Anuga FoodTec 2024 back in March, reflexx A.I. leverages over two decades of experience in pulsed X-ray technology and promises precise detection

with minimal radiation exposure.


Reflexx A.I. can detect even the smallest high-density foreign objects, previously undetectable in heterogeneous X-ray images. This technology distinguishes contaminants from harmless deviations, reducing false rejections and preventing productivity losses and waste.


The company recently unveiled the new Heuft InLine II IX empty bottle inspector and enhanced its Heuft eXaminer II XOS for detecting foreign objects in full containers. These systems are now equipped with the latest version of reflexx A.I.



“Recent advancements in the food sector have led to the application of multi-layered neural networks for intelligent X-ray image analysis,” Henschke explained. “With the implementation of reflexx A.I., the eXaminer II XAC can now identify minute foreign objects like aluminium fragments in pickles or tiny glass splinters in jars of red cabbage, which may not be visible to the naked eye. Similarly, the compact eXaminer II XS can detect objects like ring-shaped wire in pasta packets.”


Integrated architecture


MVTec’s Halcon is positioned as a versatile solution, offering an integrated development environment and flexible architecture. It integrates advanced deep learning capabilities, offering manufacturers a tool for image analysis and recognition. By leveraging neural networks, Halcon enables precise defect detection, part classification and quality control, ensuring products meet rigorous standards.


Gärtner told FoodBev: “Our aim is to maintain and strengthen our position as a technology leader in machine vision software. That is why we have our own research department. Our colleagues here research the latest trends in machine vision. For example, the department actively participates in leading conferences worldwide and is respected for its R&D contributions.”


At the end of last year, Filtec showcased its latest machine vision technology at Pack Expo Las Vegas. Among the featured products were the company’s Remote Vision Inspection Modules, offering a range of inspection applications, including date and lot code verification, label presence, closure integrity, colour consistency and fill level checks. The modules integrate with Filtec’s Intellect inspection platform, delivering flexible quality assurance solutions for containers on fill lines of all sizes.


The company also unveiled its latest photon, vision and remote vision fill level inspection machines, showcasing a 360-degree fill and closure solution. Meanwhile, its Empty Can Inspection System is designed for high-speed lines processing up to 2,400 cans per minute, using vision technology to identify defects in empty cans Featuring a dual circular top lighting station, it inspects each can’s flange, surface and bottom to uphold product integrity.


Best practice


For its efforts, Fortress developed a Retail Spec Combination system last year that combines checkweighing and metal detection, and includes every retail code of practice (COP) specification pre-programmed into the machine. “This not only saves space, but also prioritises food safety,” said Brown. “The latest version now includes new features – for example, a walk-through COP guide and easy-to-follow animated metal detector test screens.”


Fortress’ inspection systems are modular, providing users with greater flexibility. They enable the integration of electronics, upgrading of metal detectors with newer software, and configuration with both legacy upstream and downstream equipment.



“We have noted a significant push from multi-national retailers for processors to adopt X-ray technology,” Brown highlighted. “The decision to invest in either X-ray or metal detection equipment hinges on the specific requirements of manufacturers, guided by a thorough risk analysis of potential contaminants in their production line. While both technologies offer distinct advantages, they should not replace good manufacturing practices. Instead, food contaminant detection units should complement the process based on individual needs and risks.”


He explained: “Metal detectors are inherently application-specific pieces of equipment, tailored to suit various types of applications such as conveyor belt, gravity/free fall and pipeline/pumped systems. In contrast, X-ray systems are available only in conveyor and pipe formats. Therefore, the suitability of an X-ray system depends on the food application and packaging used, determining whether it is the appropriate choice.”


Sustainability is key


Vision systems serve various purposes that evolve over time. One notable trend where vision-based inspection is crucial is sustainability, particularly in packaging. Manufacturers across the F&B spectrum are constantly seeking more environmentally friendly packaging solutions, such as fibreboard and paper-based alternatives to plastic.


“It’s a positive trend for sustainability, but these new types of packaging can pose challenges for inspection,” warned Cognex’s Sacco. “Recycled materials can have surface irregularities and/or less consistent backgrounds, making automated inspection more difficult. Paper-based packaging isn’t always as uniform as plastic. Fortunately, thanks to AI, image-based inspection systems have gotten very good at adapting to variable backgrounds and products to accurately identify the main areas of concern.”


For such scenarios and beyond, Cognex has been developing AI-powered machine vision systems.


Unlike many other manufacturing processes characterised by identical products or parts, food items often exhibit natural variations from one to the next. Products like the Cognex In-Sight 2800 utilise edge-learning AI technology to rapidly learn to differentiate between good and defective products.


In-Sight 2800 Vision System is a solution for error detection that combines AI with traditional vision tools. Features include unlimited classification outputs and advanced optical character recognition. Easy to integrate and deploy, it enhances operational efficiency without the need for programming.


“Our emphasis on AI is also simplifying the set-up and operation of automated inspection systems,” Sacco added. “Edge learning technology allows for rapid training on new tasks by anyone with sample products, eliminating the need for experienced vision engineers or programmers. This facilitates smooth changeovers between production lines and reduces downtime. Manufacturers dealing with multiple product lines or experiencing rapid growth can particularly benefit from the quick implementation afforded by these advancements.”


Vision inspection technologies are indispensable for ensuring product quality and safety in the F&B industry. These systems offer precise defect detection and part classification, enhancing operational efficiency and meeting stringent quality standards. Key advancements like AI image processing and 3D imaging drive continuous innovation in vision inspection, enabling higher levels of accuracy and reliability.


Manufacturers should embrace, rather than fear, this technology, as Filtec’s Coleman concluded: “Many advancements have occurred in the last decade or more...Challenge your suppliers to provide comprehensive insights into the technology’s capabilities and the maintenance requirements necessary for its effective operation. With proper training and user-friendly interfaces, you can ensure a successful implementation of machine vision that will serve you well for years to come.”

Related posts
Fortress Technology unveils Icon X-Ray System to enhance food safety
Technology

Fortress Technology unveils Icon X-Ray System to enhance food safety

bottom of page