(Photo credit: HEUFT)

Exhibition

interpack 2026: HEUFT - look deeper - see more - assess better

2026-05-04 11:52 min Interpack 2026, Image processing and vision systems

HEUFTBurgbrohl, Germany

Making visible what was previously invisible. Detecting dangerous defects and foreign objects even where this was previously impossible. Making more competent decisions about what is good or bad. And only rejecting what really needs to go: with HEUFT’s deep learning AI for in-depth analysis of X-ray and camera images, there is now simply more to be gained when it comes to deeper consumer and brand protection alongside a sustainable reduction in packaging and food waste! The new possibilities will be demonstrated at interpack 2026.

Highlights at the international gathering of the packaging industry include

the deep-learning-enabled HEUFT reflexx A.I. X-ray image processing, which, when detecting glass in glass, makes the invisible visible even in structured, inhomogeneous products such as pickled gherkins and intelligently distinguishes critical from non-critical findings,
the HEUFT eXaminer II XAC uses deep-learning AI in pulsed X-ray inspection to achieve full coverage, sensitivity and assessment power when examining filled food jars, particularly in the challenging task of glass-in-glass detection,
the Multi Colour Image Processing, which now provides complete coverage of the finish areas of wide-mouth jars and even fully reveals the finest cracks on the sealing surface,
the HEUFT reflexx A.I. camera, which, in combination with deep-learning-enabled image processing, now evaluates product markings such as best-before dates more intelligently and verifies them reliably even when they are distorted, smudged or blurred,
the compact HEUFT eXaminer II XS, which achieves full detection reliability with a further reduced false rejection rate in foreign body detection in full containers, such as tins, using pulsed X-ray and HEUFT reflexx A.I.
the lateral laser profile measurement, which detects even the smallest leaks during integrity testing of filled cardboard composite packaging,
state-of-the-art IT solutions, ranging from production data acquisition and line analysis through to brand and recipe management and end-to-end track & trace, ensuring the necessary efficiency, overview and traceability in the filling and packaging process.

Deep learning X-ray image processing

Thanks to deep learning, X-ray image analysis with HEUFT reflexx A.I. now goes even deeper. It reveals what was previously invisible and reliably distinguishes critical foreign objects from non-critical structures and objects.

Whether pickles or red cabbage in a food jar, cereal, or pasta: even with HEUFT’s proprietary, in-house-developed hardware and software for intelligent image processing, high-density foreign objects such as glass splinters or metal fragments were not always clearly detectable in the X-ray image of a random batch of such structured products. The AI, which it has been using for more than ten years to not only identify a wide variety of objects but also to classify and intelligently evaluate them in a multidimensional process, reached its limits with such structured products. This is changing with the latest HEUFT reflexx A.I. version for the intelligent processing of X-ray images: Using deep learning, it makes previously invisible features visible during evaluation and identifies the tiniest foreign bodies and defects even where this was previously impossible: in non-homogeneous product masses with cavities of varying sizes between their individual components – and with irregular structures that absorb X-ray pulses to varying degrees.

Thanks to the new deep learning algorithm, the aluminium fragment among the pickles can now be detected just as reliably as the glass fragment in the jar of red cabbage, the wire in the muesli, or the ring-shaped piece of cable in ring-shaped pasta of the same size. HEUFT reflexx A.I. detects and marks dangerous foreign bodies in real time – whilst reliably distinguishing them from harmless product and packaging structures, so that the false rejection rate in pulsed X-ray inspection continues to approach zero.

Proven image analysis and AI methods have been combined with a multi-layered neural network that delves deeper and thus processes even abstract patterns independently in a meaningful way. The deep-learning-capable HEUFT reflexx A.I. is thus significantly superior to traditional analysis methods such as grey-scale determination, contrast detection and machine learning for the identification and classification of different objects.

What was previously invisible becomes visible, the rate of false rejections of products that are actually uncontaminated is significantly reduced once again – and no more valuable packaging or food is wasted.

Deepened glass-in-glass detection

Deep learning for intelligent X-ray image processing with HEUFT reflexx A.I. deepens the detection reliability of pulsed X-ray inspection using the HEUFT eXaminer II XAC. With a further reduction in the false rejection rate, previously invisible objects become now visible for the first time.

The current X-ray components, which are developed and manufactured in-house, boost the HEUFT eXaminer II XAC's performance and sensitivity when it comes to detecting filled food jars contaminated with glass splinters and other high-density foreign objects. When combined with the deep learning technology for denoising X-ray images, which has been available for some time, the HEUFT eXaminer II XAC can halve the size of objects it can detect at the end of the line. New deep learning algorithms in the HEUFT reflexx A.I X-ray image processing take this a step further. What was previously invisible becomes visible for the first time, and critical defects are distinguished from non-critical ones even more reliably.

This applies to both the double-base inspection and the 360° sidewall inspection, which, when combined, provide complete coverage of the entire container volume. At major food manufacturers such as Carl Kühne KG, appropriately equipped HEUFT eXaminer II XAC systems are already demonstrating their strengths when it comes to getting to the bottom of foreign objects and reliably distinguishing them from harmless product and packaging structures.

Particularly with non-homogeneous products such as gherkins or red cabbage, which contain components that absorb X-rays to varying degrees and have cavities in between, glass-in-glass can be identified even when it is no longer visible to the naked eye or appears extremely similar in shape and size to non-critical product features on the X-ray image.

HEUFT reflexx A.I. thus identifies truly critical contaminants even more reliably – and removes only those food jars that are actually contaminated from circulation. This reduces the false rejection rate and thus prevents packaging and food waste, along with the associated follow-up costs.

To achieve this, HEUFT has combined proven image analysis and AI methods with a multi-layered neural network that delves deeper and thus processes even abstract patterns independently in a meaningful way. The deep-learning-capable HEUFT reflexx A.I. X-ray image processing system is thus significantly superior to both traditional analysis methods such as grey-scale determination or contrast detection, as well as simple machine learning. Even under challenging environmental conditions, it makes previously invisible defects visible and minimises the rate of false rejections

Intelligent wide-neck jar finish inspection

HEUFT’s Multi Colour Image Processing (MCIP) now enables comprehensive inspection of all types of wide-mouth finishes and improves the reliability of finish defect detection.

In-house developed MCIP significantly enhances detection performance during the finish inspection of food jars prior to filling. Even wide-neck openings of varying shapes and diameters can now be inspected with full coverage. Actually, the finest cracks on the surface, which could lead to dangerous glass breakage, are reliably identified during all-surface empty container inspection with HEUFT InLine II systems.

To this end, the systems engineers from the Vulkaneifel region have directly integrated the structured MCIP lighting into their HEUFT reflexx A.I. cameras. From a single perspective, it combines different illumination scenarios such as bright-field and dark-field illumination in transmitted and reflected light. his is done in different colours to spectrally separate the resulting information and to combine the individual colour channels in such a way that a wide variety of features can be identified on the detection images.

They contain fewer interfering structures and produce a reflection ring on the mouth sealing surface that is twice as large as before. Where this ring is interrupted or blurred, there are notches and cracks, which are intelligently detected in real time by the HEUFT reflexx A.I. image processing system and automatically flagged as critical defects.

At HEUFT’s interpack stand, the new possibilities for wide-neck finish inspection will be demonstrated at a dedicated Enlightenment station.

Adaptive code verification

With new deep learning, Original Character Verification (OCV) of product markings such as the best-before date (BBD) is now possible even in difficult cases.

Achieve reliable Original Character Recognition (OCR) even with declining coding quality and verify the content of product labels even when they are misaligned, distorted, smudged or blurred: Deep learning-enabled HEUFT reflexx A.I. image processing makes it possible. Even in such cases, reliable OCV of alphanumeric codes such as best-before dates, batch and lot numbers is now achievable. To ensure this even when, for example, the print quality of the inkjet coder deteriorates over time, the print image suffers and the best-before date is barely legible to the naked eye, countless corresponding images have been trained into a deep learning model.

HEUFT reflexx A.I. now not only separates and centres individual characters more precisely, but also assesses their various appearances more competently than ever before. And it does so at line speeds of 1,200 packages per minute and above. Creeping printing errors and display problems such as misaligned ink dots, distorted, smudged and blurred letters, numbers and lines of text are reliably detected and correctly assessed – thanks to prior training on a wide variety of appearances. The inkjet printer does not require immediate maintenance, and the relevant filling and packaging line does not need to be stopped straight away. Instead, maintenance can be postponed until regular stops to ensure full efficiency and productivity. Excessive high false rejection rates due to declining coding quality are prevented sustainably.

Full detection reliability in the tightest of spaces

Versatile pulsed X-ray technology and HEUFT reflexx A.I. X-ray image processing open up new possibilities for gentle and precise foreign object detection at the end of the line using the compact HEUFT eXaminer II XS – and improve detection reliability.

In the tightest of spaces, the slim turnkey solution for pulsed X-ray inspection of tins, doy packs, squeeze bottles, stand-up pouches or carton packaging achieves full detection reliability: high-density foreign objects such as metal particles or hard plastic fragments in the product are identified gently and precisely. To this end, the modular system can be flexibly equipped with one or two sidewall X-ray flashes. This ensures that the inspection always covers the entire fill volume and, in conjunction with HEUFT reflexx A.I., increases the precision of foreign body detection.

Using deep learning, the intelligent image processing system denoises the X-ray images and simultaneously removes artefacts. This halves the size of foreign objects that can be reliably detected. For precise base inspection of carton packs filled with liquid products, the HEUFT reflexx A.I .X-ray image processing provides an ‘unfolded’ view in which even particularly small foreign objects lying flat on the bottom become visible.

During pulsed X-ray inspection of particularly tall filled containers using a single detection unit, its optional angled orientation enables the entire container volume to be covered; foreign bodies are thus detected not only when they lie on the bottom, but also anywhere else within the packaging. A full-field image converter image converter ensures increased sensitivity and guarantees that every single X-ray pulse covers a significantly larger container area than before.

The HEUFT SPECTRUM II control unit of the compact HEUFT eXaminer II XS, to which many other detection functions – including those for the precise verification of product marking – can be linked, is highly automated. For instance, for fast and easy product changes, the height and alignment of the upper X-ray flash module automatically adjust to the changed container format. The HEUFT NaVi user guidance offers the operator comprehensive audio-visual step-by-step assistance.

All this makes the compact HEUFT eXaminer II XS a genuine turnkey solution for comprehensive foreign object detection at the end of the line.

On the trail of leaks

With lateral laser profile measurement, technology, HEUFT’s full-container check systems examiner the leak-tightness of freshly filled beverage cartons and other composite packaging.
Unbreakable, opaque, extremely lightweight, easy to stack, transport and store, and increasingly recyclable: steam-flushed and hot-filled beverage cartons offer many advantages. Even when unchilled, liquid contents such as fruit juice or UHT milk remain fresh and palatable for a long time. Provided, of course, that they are stable and leak-proof. HEUFT’s laser profile measurement technology for flexible packaging checks that and more immediately after filling these block-shaped packages – contact-free, using clever sensor technology and intelligent optics!

Even micro-leaks are detected in-line by the precise profile measurement of each individual package. This is because leaky cartons draw in air, thereby altering their dimensions. This is registered by several optical sensors that inspect them with millimetre precision. Tiny deviations are thus uncovered – and with them, leaks that are invisible to the naked eye. An X-ray fill level check optimises the result by detecting product loss within the beverage carton caused by packaging leaks.

In the same operation, the presence, integrity, colour and position of the closures are also checked. The intelligent HEUFT reflexx A.I. camera, when mounted in a top-down configuration, detects cap colours that do not match the product just as reliably as it does closures that are not in the correct position and could perforate the coated hole when unscrewed. Sealing foil that protrudes beyond the lid is also reliably identified. The same applies, for example, to carton flaps that are not correctly glued or are sticking out.

At the same time, the smart camera – developed and manufactured in-house – is used to check whether, for example, the best-before date is actually present, legible and correct in terms of content. This is now even possible with deteriorating coding quality, such as characters that have shifted, become distorted or blurred.

Incorrectly coded, unsecured and leaking packaging is clearly identified and reliably removed from the production line – using proven sorting systems such as the HEUFT DELTA-FW flow rejector. This eliminates the risk of incorrectly labelled composite carton packaging with prematurely spoiled contents entering the retail market, as well as the risk of contamination from leaking contents.

Effective line analysis

The HEUFT CyBridge.profiler provides comprehensive production data acquisition (PDA) and line analysis for improved efficiency and production quality.

The HEUFT CyBridge.profiler records operational and quality data in real time. As a lean software solution, it logs, for example, test container programmes with which HEUFT inspection systems check their own detection performance. Via the HEUFT CyBridge.hub server platform, which is available both as a physical rack and as a virtual version for integration into an existing server environment, comprehensive PDA and line analysis is achieved, incorporating key performance indicators from HEUFT inspectors on a line as well as those from machines by other manufacturers.

This enables those responsible for efficiency and quality management to obtain an up-to-date overview of the status at any time. The new, clear interface – featuring attractively presented key figures, notifications, statistics, diagrams and graphics – can now be displayed accurately and clearly on all standard web browsers and mobile devices such as tablets and smartphones. HEUFT CyBridge.profiler users decide for themselves how, for example, consumption meters or HACCP data are best displayed, so that they can monitor and evaluate the performance of individual machines, Overall Equipment Effectiveness (OEE) and overall production quality.

To this end, the enhanced BDE tool calculates all relevant key figures in accordance with OMAC standards, based on the Weihenstephan Standards co-developed by HEUFT as an industry partner, and presents them individually. Different counters can be displayed in a single diagram to, for example, compare the number of containers checked during in-line inspection with the number subsequently rejected. Weak points and faults are thus identified immediately. Their causes can then be rectified before the efficiency of entire lines is compromised. Unplanned downtime is completely avoided or at least significantly reduced.

From logic to buffer analysis, from downtime monitoring to batch reports and audit trail logs: with its modular design, the HEUFT CyBridge.profiler can be tailored precisely to the specific needs of the respective bottler and the individual requirements of the responsible line managers – and expanded retrospectively if necessary. For example, to provide a visualisation of the entire plant with real-time status and event displays, or a central archive for backing up the configuration data and log entries of all connected devices to facilitate easy resets.

For comprehensive reporting of key performance indicators, the HEUFT CyBridge.profiler generates professional reports that can be printed at any time, sent electronically and – to fulfil basic documentation requirements – archived long-term. For the analysis of line efficiency, such a report then contains, for example, clearly presented infographics that compare the net operating time of individual machines with the total operating time of the respective line. Key performance indicators can be freely selected, defined and further calculated to generate even highly specific analyses and reports. The necessary inspection quality is ensured, amongst other things, by the logging of self-test programmes from HEUFT inspectors, the monitoring of HACCP limit values and also the newly integrated module for recipe and product changeover monitoring – an ideal complement to the HEUFT CyBrigde.reciper, which enables comprehensive recipe management and centralised product changeovers.

reflexx AI camera. (Photo credit: HEUFT)

End-to-end track & trace

Smart detection and verification, precise product tracking and secure data management in real time: HEUFT technologies such as these are not only the key to top-level in-line quality inspection. They also enable track & trace of the entire filling and packaging process. HEUFT will be showcasing what’s on offer at interpack!

With network-compatible quality control and inspection systems, smart recognition modules and digital technologies, HEUFT enables effective track & trace: with an accuracy of up to 0.5 millimetres, devices from the HEUFT SPECTRUM II generation determine the current position of every single product or packaging material and track it continuously in real time via a ‘shift register’ – if only to ensure that every single item is inspected and that any identified as faulty is actually diverted.

This naturally also applies to all network-compatible HEUFT devices equipped with the new HEUFT reflexx A.I. camera. In addition to a wide range of inspection tasks, it also performs OCR functions and even verifies the content of packaging codes. Adaptive LED lighting is integrated into the in-house developed colour sensor camera, as is image processing using artificial intelligence. It handles the reading and verification of GS1-128 barcodes and 2D Data Matrix codes autonomously and in real time, ensuring that the origin, authenticity, transport and process data of the respective product stored within them can always be clearly assigned to a specific packaging unit and supplemented.

The secure transmission and long-term archiving of all information contained in the codes is ensured by the real-time data management of network-compatible HEUFT systems. All key figures, operational and quality data obtained during inline inspection and line analysis using the HEUFT CyBridge.profiler (see above) are also incorporated – right down to the individual detection images generated during empty and full container inspection. From audit trails to unique identification and seamless traceability, serialisation and aggregation of a wide variety of products and packaging materials, and from proof of origin to complete documentation of the filling and packaging process, the enhanced BDE tool (see above), together with HEUFT reflexx A.I.-cameras, generates big data for traceable production. The current position of every single container can be determined precisely at any time, regardless of whether it is still on the line, in the tray or already fully packed on the pallet.

Among other things, this also enables targeted recall management: in the event of a recall, all quality control and inspection results are documented in detail and permanently, and the batch to be withdrawn from circulation can be narrowed down so precisely that brand protection does not result in food waste. The latter is also prevented by proactive pool and resource management, which enables end-to-end track & trace at HEUFT level, for example by ensuring that only food jars suitable for use are filled, and that non-marketable packaging is clearly identified, reliably tracked and accurately rejected in good time, even before the valuable product is added.

HEUFT@interpack2026: stand A45, Hall 13

www.heuft.com

PETnology's Resource Guide

comPETence center

The comPETence center provides your organisation with a dynamic, cost effective way to promote your products and services.

Find out more