Ensuring quality control with NextGuard Pack Division software

Ensuring quality control with NextGuard Pack Division software

Importance of the topic

Detecting missing items in packaged goods is a frequent challenge in high-speed food production, especially when individual items are light relative to total package mass and production variability masks the absence of components. Traditional weight-based controls (checkweighers) and bulk X-ray mass estimation can fail to reject packages with missing parts because the missing mass may fall within normal batch-to-batch variation. Software that visually identifies and assesses individual product stacks inside a package therefore fills a critical gap in product verification and brand protection, improving food safety, customer satisfaction, and regulatory compliance.

Objectives and overview of the application note

This application note presents Pack Division, a new software module for the Thermo Scientific NextGuard X-ray Inspection System. The software is intended to detect missing whole or partial items arranged in stacks within a package (both circular and rectangular product shapes). Key goals are to (1) reliably identify missing items that would not be detected by package mass alone, (2) integrate with existing NextGuard contaminant detection and mass estimation tools, and (3) provide an adjustable, user-friendly configuration for a variety of product sizes and stack arrangements.

Methodology and algorithmic approach

Pack Division uses gray-scale X-ray images to segment the package into a grid of stacks and to estimate the number of items in each stack based on pixel-level intensity and geometry. The main configurable parameters are presented below:

• Show Pack Location — visual overlay to locate the package and individual stacks during setup.
• Threshold — gray-scale cutoff to distinguish product pixels from the background and delineate stack boundaries.
• Rows and columns — number of stacks arranged in the package grid; the algorithm spaces stacks optimally within the package rectangle.
• Length and width/diameter — expected object size in pixels for either rectangular or circular products; used for accurate counting.
• Items per stack — expected maximum items per stack (derived from a complete-package image).
• Stack minimum — the minimum acceptable number of items per stack to avoid rejection; supports fractional settings (e.g., 2.5 to reject stacks of two or fewer items).

The tool analyzes deviations between stacks within the same package rather than relying solely on absolute mass. This internal comparison makes the method robust to product density variability and reduces the need for frequent recalibrations. The software can detect partial missing items (down to fractions of a unit in tested cases) by treating each stack as the comparison unit: a quarter-missing item in a stack corresponds to a large relative change within that stack even when it is negligible relative to the whole package mass.

Used instrumentation

The application note describes Pack Division as a feature of the Thermo Scientific NextGuard X-ray Inspection System, specifically referenced alongside the NextGuard C330 family. The system combines X-ray imaging for grayscale analysis with the NextGuard user interface (HMI) for recipe and parameter configuration. Pack Division is distributed as part of the Product Verification option in the NextGuard software suite.

Main results and discussion

Pack Division was validated in a production example with stacked cookies. Key empirical findings and illustrative numbers from the application note include:

• Example acceptance range on a checkweigher: 270 g ± 10 g (acceptance 260–280 g).
• Observed product weight variance: 260–280 g; cookie individual weight variance: 7–8 g.
• Full package measured at 276 g (accepted by checkweigher).
• Package with one cookie missing measured at 269 g — still within checkweigher acceptance and thus would pass despite missing product.

Using Pack Division, the system identified the missing cookie by evaluating stack-level grayscale and mass distribution. Tests reported detection of a missing full cookie, half-cookie, and—in favorable conditions—even down to one quarter of a cookie. The detection precision depends on operational parameters including conveyor speed, package orientation, the pixel calibration of expected dimensions, and stack minimum thresholds; the software estimation may vary by approximately ±0.1 items in some cases, so threshold selection is important.

Benefits and practical applications

Pack Division extends NextGuard capabilities in the following practical ways:

• Detects missing or partially missing products that escape mass-based controls, reducing false-acceptance of incomplete packages.
• Adapts to circular and rectangular product geometries as well as a range of stack counts and arrangements.
• Operates robustly under product density variability by comparing relative differences between stacks instead of absolute mass alone.
• Integrates with other NextGuard features (contaminant detection, mass estimation) to provide a consolidated product verification workflow on a single platform.
• User-friendly HMI and visual aids (show pack location) simplify commissioning and recipe management for multiple product formats.

Future trends and opportunities

Potential directions for further development and application include:

• Machine-learning enhancements to improve segmentation and to adapt automatically to new product shapes or deformation patterns.
• Real-time multi-modal fusion combining X-ray grayscale analysis with inline vision or hyperspectral sensors to increase discrimination between product defects and true missing items.
• Automated recipe libraries and rapid self-calibration procedures to reduce setup time for multi-format production lines.
• Extension to high-mix, low-volume environments through dynamic grid detection rather than fixed rows/columns.
• Statistical process control integration to feed stack-level rejection data into manufacturing analytics for root-cause analysis and yield improvement.

Conclusion

Pack Division provides a practical solution for a common failure mode in packaged-product inspection: lightweight or partial missing items that are not detected by package mass alone. By subdividing the package into stacks and comparing internal grayscale/mass signatures, the software reliably identifies missing whole and partial items in both circular and rectangular configurations. The approach reduces reliance on absolute mass thresholds, improves robustness to density variability, and integrates into existing NextGuard inspection workflows, offering measurable gains in quality assurance for food manufacturers and other pack-based industries.

References

• Thermo Fisher Scientific. Application note: Ensuring quality control with NextGuard Pack Division software. PPA-AN1462-EN, 6/2025.