The machine, launched in 2013 as a foreign object detection system, has now been designed to inspect packaged products for missing pieces or components using customizable vision software.
Product verification software allows customers to use a pull down menu to select and sequence image processing and analysis algorithms.
This enables size, shape, angle, count and position-based inspections in minutes, rather than waiting weeks for custom software.
The standard system processor is six times faster than the first generation and an optional 0.4mm-pixel detector is available that offers 3.2 times better resolution for improved metal detection in lightweight, homogeneous products such as powders or cheese.
Product verification inspections
Bob Ries, lead product manager, metal detection and X-ray inspection at Thermo Fisher Scientific, said the software can address product verification inspections such as counting, checking size, finding cracks or determining presence/absence.
“This is important for customers wanting to check quality inside a closed package they can only see into with X-ray,” he told FoodQualityNews.com.
“Because they are using X-ray to find foreign objects, adding this quality inspection only requires purchasing a software option and not a completely new system.
“Once customers can see inside their packages they often ask about other inspections they might implement. These quality inspections are not related to food safety but are nonetheless important to protect the brand.”
Targeting less dense contaminants
Ries said algorithms have been added targeting the detection of less dense contaminants such as glass or rocks in complex, textured products such as nuts or tea bag boxes.
Optional mass estimation software is designed to utilize the density information in the X-ray image to estimate weight and detect under/over filled products.
The firm anticipates the need to add new image processing and analysis tools to the framework that will allow new inspection problems to be solved.
NextGuard’s features have been driven by a changing industry landscape, including evolving HACCP standards, increased use of metalized film packaging and various wet or semi-frozen foods (cheese, ice cream, bakery products) that do not lend themselves to metal detection due to product issues.
Ries said generally X-ray challenges fall into two classes.
“First and foremost they are associated with density. Thick, dense products can easily mask small, lightweight contaminants such as plastics,” he said.
“Current approaches to this problem are to inspect the product flow further upstream where the total density is less, pass the X-ray beam through the product in a different orientation to reduce this effect or use materials that are better suited to X-ray detection.
“The second challenge is to design software algorithms that can locate random shape/size contaminants in complex images. This is referred to the “anything anywhere” problem. The typical solution to this is to offer a wide variety of algorithms that all run on an image to raise the probability of detecting a hard to spot problem.”
Machines are installed in 10-plus countries for a variety of packaged product applications, said Ries.
“Many customers have turned their metal detectors off and now rely solely on X-ray. Metallized film and foil packaging continue to be big drivers as well as detection of plastics and rocks,” he said.
“A few years ago to get capabilities like inspection, mass estimation, high resolution detectors or no-lead curtains you would have had to purchase a mid- to high-end system. With the improvements we have made to NextGuard, this is no longer the case.”