Food For Thought: How Machine Vision Is Revolutionising The Food And Beverage Industry

Wayne Goh, Head of ASEAN/ANZ, Cognex, shares his insights on harvesting the full benefits from your vision systems within a smart factory ecosystem.

As an industry in the fast-moving consumer goods (FMCG) category, the food and beverage sector is not only incredibly complex—with many players, systems and value chains involved—but also ripe with opportunities for innovation. Whatever the innovation, however, the enduring need for quality, efficiency, and most importantly, safety, still remain.

One of the key drivers of this innovation is the fourth industrial revolution, or Industry 4.0. The new industrial age has brought about the introduction and propagation of the Internet of Things (IoT), which have given rise to Smart Factories where physical machine systems are able to exchange data—or communicate—and automatically turn insights from this data into action.

This is where machine vision comes into play. According to the Automated Imaging Association (AIA), machine vision encompasses all industrial and non-industrial applications in which a combination of hardware and software provide operational guidance to devices in the execution of their functions based on the capture and processing of images. In the context of food and beverage, these image-based technologies can play an integral role in improving the efficiency, safety and quality across the different stages of the production, processing and distribution process.

One of the early applications of machine vision in the food and beverage sector was for muffin inspection in the 1980s. Using a camera with a resolution of just 30 x 32 pixels, the vision system was used to reject oversized products that would have jammed up the packaging machines. Since then, machine vision technologies have advanced dramatically, not only in terms of camera technology, but also in computer processing power, software and illumination techniques.

A key conflict, however, in the procurement and implementation of machine vision solutions across the food and beverage sector is the fundamental reality of the industry: the low-margin of operations, contrasted with the high stakes of human consumption, high volume of production, and the need for incredibly stringent levels of safety and quality. This poses a tug-of-war challenge of sorts for machine vision systems to be kept affordable, yet also capable of delivering top quality solutions for the production line without fail.

Therein, however, also lies the opportunity for vision technologies to play a powerful role in the industry and automate the processes that demand reliability and repeatability with zero room for error. No other single aspect of the production line captures more information, and can be more valuable in assessing products and finding defects. Nor can any other piece of equipment be more useful in collecting data that can help direct operations and optimise the productivity of other robots and complementary pieces of equipment. Machine vision can be an industry-wide game changer that can provide a high rate of return on initial investments.

 Envisioning Food Product Safety

With the huge gamut of operations and processes within the food and beverage industry, from production and processing, to packaging and distribution, machine vision can greatly enhance the safety and quality of food and beverage products across different stages of the value chain.

Removing human subjectivity from the equation, machine vision systems analyse patterns and examine data from objects that are scanned within its set field of view. This input of data is then compared with pre-existing trained information in its database—as defined and input by humans—and unbiased conclusions are drawn about the object being scanned. And all these calculations are completed and conclusions made within less than a second.

At this level of speed, consistency and objectivity, machine vision technologies can be used to inspect a product’s colour, the degree of ripeness, spoilage or damage of raw ingredients, or even whether the product is under or overcooked, and sort them accordingly during the production and processing stage.

With a level of detail that goes beyond the ability of human eyes, vision systems can capture hyperspectral images of an object in different wavelength ranges. Combined, these images can provide a greater depth of information and can be particularly useful in areas of applications where ingredients or substances need to be identified and discerned, and are not recognisable through a standard colour or monochrome picture.

Detecting Packaging Defects

Another key aspect that contributes to the safety and quality of food and beverage products is ensuring that the packaging is free from defects that could potentially devalue and degrade the food encased within.

Machine vision inspection systems can perform the crucial role of identifying imperfections in packaging that could cause product recalls, segregate the defective packages, and have the defects rectified. This can also lead to a reduction in wastage of materials for producers.

As part of the packaging of a product, machine systems can also identify products that have been mislabelled—whether due to machine or human error—and have them removed from the production line. Misaligned or wrinkled labels can also be identified and corrected to ensure the standards of the products that reach the consumer.

A case in point is New Zealand’s leading ice cream company, Tip Top. The company made a decision to remove barcodes from the lid of its packaging of its popular ice cream tubs. This, however, introduced a challenging food safety concern: how to ensure the correct packaging is being used for the ice cream in production? Using the Cognex In-Sight system with pattern matching features, the Tip Top packaging line was able to correctly identify and match the right lid to the correct products without the use of barcodes, helping the brand maintain a high level of production without a compromise in quality.

Product Tracking And Tracing

Even so, packaging is not only important at the end-user phase when consumers come into direct contact with the finished product. The level of safety and identification of intermediate products is also crucial when many players are involved in the production of a single end-product. Here’s where vision technologies can come into play in the tracking, tracing and data management of intermediate products as they make their way from one producer to another, or even within a single producer.

An example is Korean food producer, CJ Food System. The company focuses primarily on the distribution of food materials for business, contract food services and the supply of raw material commodities to the food processing industry. With the production of both raw and subsidiary food materials as well as finished food products, the producer needed to ensure reliable tracking during the high-speed distribution processes.

Using Cognex’s In-Sight vision systems, CJ Food System was able to ensure the code reading and verification of their feed products from the mills, with the cameras guiding the right products to be stacked for distribution.

Reaping The Most From Vision Technologies

It is an open secret that food and beverage brands, and the industry as a whole, can be highly susceptible to an erosion of consumer trust as a result of product safety incidents. All it takes is just one bad apple to cause a rippling effect that can taint not just the reputation of one brand, but a category of products, or even an entire nation of manufacturers—as in the case of the milk formula scandal, for example. The high stakes environment of the industry can also be attributed to the ever-tightening regulatory requirements and heightening industry standards.

Given the robustness and strict level of operations required, the adoption of machine vision technologies can greatly improve the productivity, efficiency and safety across the food and beverage production and manufacturing process. When planning for the integration of vision solutions into the production line, it is crucial to take note of the environmental conditions where the systems will be installed. Depending on factors such as humidity, substances and materials used for example, a solution must be carefully selected to meet the regulatory requirement for its particular purpose, and customisations made for the piece of equipment to meet its intended use. Non-enclosed solutions and glass materials which would work for other industries, might not meet the special requirements of the food and beverage industry.

Apart from the increases in productivity, better quality control, and the freeing up of manpower for work in other value-added tasks, the true intrinsic value that machine vision systems can deliver to food and beverage producers and manufacturers can best be seen in the context of a smart factory or smart manufacturing line with a network of connected smart equipment.

With the sheer amount of data and measurements vision systems are capable of generating, information collected can be fed back into a connected manufacturing network where actionable insights and conclusions can be drawn and communicated to other pieces of smart equipment within the ecosystem, and processes can be automatically executed by these other machines—all happening passively without the need for active human intervention.

Ultimately, when adopting any new technologies such as vision systems, it’s important to consider the needs of your business, your existing infrastructure and system compatibilities, as well as how you can future-proof your investment. The key to that is ensuring flexibility and ease of change within your existing systems, be it in terms of the networking protocols within a smart factory ecosystem, or even in terms of tailoring the physical manufacturing environment to harvest the full benefits from your vision systems.