A subsidiary of the International Society of Automation
Cancel
Search Sponsored By Sponsor

Simplify Food and Beverage Inspection with Machine Vision

  • July 07, 2010
  • Feature

Summary

Inspecting the packaging, labeling and overall integrity of food and beverage products is pivotal for ensuring customer safety and satisfaction. If a food product’s packaging has been compromised—for instance, if a cap is not properly sealed—pathogens may be introduced to the product during shipping, presenting a health concern for consumers. Consequences of an improperly labeled product range from poor aesthetics to misrepresented foods—which can merely inconvenience a customer or, in some scenarios, jeopardize a consumer with severe food allergies. Food integrity affects nearly every aspect of the consumer’s experience: from the product’s freshness to its consistency in taste, texture and nutrition.Machine vision provides a superior inspection solution for these applications due to its sophisticated analysis capabilities. While alternative sensors such as photoelectrics can deliver presence/absence information—detecting that a bottle is capped, for instance, but not whether the cap has been properly sealed—vision provides a valuable snapshot of the full application in process. Thus, it is able to analyze multiple features of interest at once, such as label presence and placement. A vision sensor includes hardware, consisting of a lens, light, camera, and a processor, and software—encompassing the programming and image algorithms used to control, analyze and monitor the inspection. In food and beverage applications, the sensor can be susceptible to caustic washdowns, with pressures up to 1200 psi. Off-the-shelf vision sensors traditionally have housings made of plastic or nickel-plated aluminum, which may be destroyed due to the high washdown pressure or corrode from prolonged exposure to harsh cleaning chemicals. Previously, a common solution for this challenge involved installing the vision hardware within a protective enclosure designed to withstand washdown conditions. This can be a costly, time-consuming process due to the cost of the enclosure, additional installation time and efforts required to ensure all equipment meets food and beverage industry regulations. The more efficient, cost-effective solution: a vision sensor designed specifically to resist exposure to washdown conditions while delivering reliable, repeatable inspection results. With stainless steel housings featuring a curved, hygienic design, these new vision sensors promote longevity and cleanliness—defending against corrosion and contamination in food and beverage manufacturing areas. Their sealed construction, which satisfies NEMA 6P and IP68 standards, wards off moisture ingress even under high-pressure caustic washdowns. The durable, sanitary design makes the sensors inherently suitable for food and beverage applications with no additional enclosures or equipment certifications required. When this rugged hardware design is combined with software containing a comprehensive set of vision tools, a vision system can tackle the widest range of food and beverage packaging, labeling and product integrity inspections. Applying Vision in Food and Beverage ManufacturingA machine vision inspection has a relatively simple purpose: to input a snapshot of the application, evaluate this snapshot based upon a set of given tolerances and output a “pass” or “fail” signal. This sequence is otherwise known as acquisition, analysis, and determination. One or more analysis tools within the vision system software determine whether a product passes or fails the inspection. A vision tool may also be used to perform the logical operations required to activate the vision sensor’s output, which can in turn activate an external device that diverts rejected parts from the line. To practically apply a vision solution in food and beverage applications, the sensor is taught to identify a known “good” reference part, such as a properly filled and capped milk bottle, from a “bad” part. The system accomplishes this by comparing the application image the vision sensor acquires to the stored “good” part, then rejecting those products that demonstrate inconsistencies. For instance, in a bottle filling application, a fill tube may become loose and fall into a bottle—contaminating product and interrupting the filling process. A vision system that detects this abnormality can output a “fail” signal that stops the line, minimizing downtime by allowing the problem to be readily addressed. Color vision sensors add color differentiation to this vision system’s capabilities, so that the sensor can identify the hue variations that could mean the difference between a “good” and “bad” product. If each bottle cap in a container must be of a certain color, the sensor can verify that each colored cap is in its proper location and that all container slots are filled. A similar arrangement allows the vision sensor to inspect a box of chocolate candies—verifying that each candy, differentiated by hue, is in its designated paper nest and that all candies are intact. To verify label presence and position on a jar or bottle, a vision sensor measures the distance from the top of the label to the neck of the container in two locations, thus confirming proper alignment and label height. Simultaneously, the sensor measures the distance from each end of the label to its associated container side. This ensures the label is within its proper horizontal alignment. A comprehensive yet efficient process, this vision inspection can save manufacturers many wasted hours and dollars due to returned or recalled product.Recently, vision technology has been applied for confirming date/lot codes on food and beverage items. While these codes have been used in the pharmaceutical industry for years, the food and beverage industry is now using this technology more often to assist in product recalls. Vision sensors ensure the proper date/lot code has been placed on each package before shipping, so that if a recall is needed, the product can be traced back to its initial batch. Date/lot code verification simplifies identification of the faulty product and also saves manufacturers time and resources. Rather than requiring a recall of truckloads of product, the recall can be isolated to only the loads containing the faulty batch—which may only be a few pallets.To assure food integrity, vision tools examine products for consistent color, shape, size and quality. Color vision tools allow vision sensors to inspect for matching hue and intensity, identifying the fat percentage in meat before processing or grading. Inspection and analysis tools may also be used to verify bacon strips are cut to the same size and length, or that granola bars are all of the same width and length without breakage. This quality assurance promotes customer safety and satisfaction while enhancing the user’s bottom line. Learn More

Did you enjoy this great article?

Check out our free e-newsletters to read more great articles..

Subscribe