While foreign material recalls do not occur that often in dairy food products, any reports of potential metal, plastic, wood, rubber, or glass contamination is alarming to consumers.

These recalls create product and financial losses for the companies involved and erode brand loyalty and consumer trust. According to the International Food Information Council (IFIC) annual food and health survey (https://ific.org/research/2025-food-health-survey), confidence in the safety of the U.S. food supply has fallen dramatically across nearly all demographic groups compared to 2024.

The decline is most notable among Gen Z, those with higher household incomes, men and Asian Americans. Among respondents in the survey, 41% indicated that they are either not too confident or not confident at all in the safety of the U.S. food supply. However, the good news is that detection technologies are advancing and offering companies new ways to detect problems and take corrective actions before they reach the consumer.

Foreign material contamination can occur via incoming raw ingredients, improper material handling, machine wear and tear, packaging materials or through human error. Magnet technologies are commonly used in liquid and powder operations.

Kristi Peterson, chief marketing officer at Magnattack Global in Burnsville, Minn., explains: “It is important to match your magnet strength with your product and process throughput. We work with clients to provide customized solutions for their testing needs. Many of our inline magnets are designed to be self-cleaning and able to withstand high temperatures and vibration. For entrances and transition areas, Magnattack Global has developed a Magstride cleanable mat that contains 200 magnets that can ensure that metal fragments or metallic dust is not inadvertently tracked into a facility by maintenance personnel or third-party contractors.” shares

Eric Garr, regional sales manager at Toronto-based Fortress Technology Inc., adds: “Most processors rely on end-of-line inspection to meet FDA regulations and retailer codes of practice. Yet, the most effective strategy is to have multistage inspections to catch contaminants early on. Dairy products like cheese that are higher in moisture and contain salt can sometimes be conductive and cause the metal detectors to react like there is metal in the product.

“One way to overcome these phenomena is to use wide spectrum metal detection operating frequencies simultaneously to improve sensitivity to address the variations in product density and avoid the risk of false positives,” Garr continues.

However, metal detectors do fail and require testing to validate that they are working correctly. GFSI and HACCP production standards require that food metal detectors be tested with ferrous (i.e., steel & iron), non-ferrous (i.e., copper, aluminum & brass) and non-magnetic stainless-steel samples to verify detection accuracy. Depending on company policy and the type of product being produced, detectors may be tested at shift changes, product changeovers and even hourly. Human error can occur when performing the physical test or by improper documentation of the results.

Besides being time consuming, the testing of the detector requires physical contact with the product flow, which may require a stoppage of production. To address this concern, Fortress Technology has developed a testing technology called Halo that replicates the signal disturbance of the standard test pieces without having to physically pass the test piece through the metal detector.

This allows for the system to be verified for sensitivity without disturbing the product and production line. It is designed to complement manual testing and reduce the amount of labor required to generate a robust audit trail for food safety compliance. In addition, it reduces the number of physical breaches to the production system, which further lowers the food safety contamination risk.

Way beyond just metal

Not all foreign materials are metal. Other detection methods may need to be utilized in combination with magnets and metal detectors to help detect items like glass, plastic, wood or rubber. For example, microwave technology is typically used to detect low density foreign materials like wood, soft plastics or fruit stone fragments in homogeneous liquid and semi-liquid products like sauces and purees. Antennas are positioned around the liquid, so the microwaves penetrate and inspect the entire area of the product. Variations in the electric field indicate a contaminant, so when coupled with Artificial Intelligence (AI) algorithms, these technologies have the ability to identify which type of contaminant is present through 3D reconstruction. Microwave technologies are less energy intensive and do not heat the product, which also helps maintain finished product integrity.

Because metal, ceramic, glass and high-density plastic contaminants come in a variety of densities and sizes, use of hygienic X-ray systems to detect differences in density between product and potential contaminants are becoming more frequently used in food production operations. Both 2-D X-ray, 3-D X-ray, and 3-D X-ray with CT scan technologies like those used at hospitals are all available for use. 2D X-ray systems analyze average differences in density, while 3-D X-ray systems have greater image clarity and can pick up more subtle differences in shape even within layers of packaging. Computer-generated digital CT scans can also be created based upon multiple 3-D X-ray measurements from different angles.

With the use of AI algorithms to analyze the image results, inline automated reject devices can be integrated post testing to keep the production line moving smoothly. Images are then stored for food safety auditing purposes. However, due to the amount of analysis required to generate the images, these types of systems may have slower throughput rates, require additional energy usage and be more expensive than traditional monitoring systems. It is important for dairy manufacturers to work closely with their chosen supplier to ensure that line speed and sensitivity are appropriate for the density of the product being inspected. In addition, electronic sensors wear out over time, so additional verification of efficacy is required. X-ray technologies also require additional employee safety training and documentation to adhere to radiation exposure standards.

How to detect defects

Product quality defects can also arise from inconsistent raw ingredients, which can impact the color or texture of the finished product; for example, variegate appearance in ice cream. Improper filling of containers in the case of milk and cultured dairy products or defective packaging or waxing of cheeses may also cause product quality and safety concerns. Visual detection technologies in combination with AI may be able to detect these concerns before the products reach the consumer.

“Our full AI Vision platform enables processors to detect defects related to color, texture, shape, surface anomalies, packaging integrity, fill level and contamination indicators,” says James Nolan, Sr. Director of AI Vision Solutions at Belgium-based Robovision. “The software platform allows QA (quality assurance) and engineering teams to train, deploy, update and scale multiple inspection applications across different lines and facilities, while integrating with programmable logic controllers and automation systems to trigger real-time actions. Human inspectors miss subtle defects and struggle with consistency at higher speeds. AI Vision helps provide the repeatability and precision needed for today’s quality expectations.”

The Netherlands-based, visionplatform.ai, also uses visual detection to identify defects in products or packaging. Its platform processes images from different angles and stores the data onsite with an AI server so that clients can build and control their own datasets making them both secure and shareable across multisite operations. Upfront investment costs are minimized because its system uses images from existing camera surveillance systems. Some operations use this detection technology in combination with label recognition for traceability purposes.

Lastly, foreign material contamination and food safety concerns can also originate from humans. With the use of AI models, visual technologies can now be used to monitor and analyze people behaviors as it relates to use or movement of tools, improper use of hygienic personal protective equipment or unauthorized access to certain locations, which can be sources of accidental contamination or even intentional theft or adulteration. According to Koen de Jong, CEO and founder of visionplatform.ai, “We are in the process of expanding our offerings to include enhanced facial-analysis capabilities, and early releases of Vision Language Models with AI, which are being designed to understand and generate language based on visual inputs that helps describe the images, answers questions about them and allows users to search for outlying images more quickly based on textual queries. This ability to identify mistakes and take corrective actions with employees or visitors will help companies maintain control of their operations more efficiently and ensure product quality and safety.”

With the availability of so many different types of detection systems, dairy processors can now choose a combination of technologies to fit their specific product and operational needs, while ensuring that dairy foods continue to maintain the highest standards of quality and safety for consumers everywhere. DF