AI-Based Quality Control in Concrete Block and Paving block Production – Practical Experience with the KBH QC Imager
Ensuring consistently high product quality is one of the central challenges in concrete block and paving block production. Fluctuations in raw materials, changes in the concrete mix, wear on tools, or altered machine settings all have visible and measurable effects. At the same time, market demands for quality continue to rise. To meet these expectations, production faces increasing pressure toward higher levels of automation and reproducibility. The ongoing shortage of skilled labor further intensifies this pressure. The insights in this article are based on experience gathered by KBH in its own concrete block production as well as from customer projects.
Traditionally, quality control in many concrete plants is still performed predominantly manually through visual inspection or, at best, automated height measurement. However, this approach is inherently subjective, dependent on individual operators, and therefore limited in reproducibility. Against this backdrop, camera-based and AI-supported systems are gaining increasing importance. With the KBH QC Imager, a system is available that automates and objectifies visual and physical quality inspection and can be retrofitted into existing production lines.
The retrofit capability is independent of machine type and can therefore be standardized across different block machine types. Interface modules for any control system are available or close to completion. KBH itself operates very different control systems in its plants, and integration has proven to be remarkably straightforward.
A key question in automated quality control is the optimal installation point within the production line. In principle, both the wet side and the dry side are possible. Quality control on the wet side has the advantage that process deviations can be detected early and machine settings can be adjusted immediately. At this stage, the product is not yet cured, and a few specific defect types may only be partially visible. However, the advantages of wet-side installation far outweigh the limitations – particularly with regard to preventing defects by enabling immediate machine adjustment.
High-resolution industrial cameras und AI-based image analysis
The KBH QC Imager combines high-resolution industrial cameras with AI-based image analysis and an intuitive user interface. The system detects surface defects, edge damage, dimensional deviations, height variations, and irregularities in the production pattern. Additionally, the vertical sides of the concrete product can be included in the measurement if desired. Monitoring the two vertical side flanks of a layer is particularly cost-effective and beneficial, as it is sufficient to detect concrete blocks produced with excess moisture (the familiar “bellies”).
In combination with board weighing, the density of the block layer can be determined with high precision. Through the user interface, defect classes can be defined, thresholds set, and new products taught very quickly. Interaction between the operator and the system is intuitive and simple, increasing acceptance and ease of use.
Limited installation space at KBH demonstrated that the Imager’s compact design makes it possible to find a suitable mounting position in nearly all scenarios. Additionally, results show that positioning the system as close as possible to the press position significantly increases overall responsiveness. At KBH’s Lachen plant, the QC Imager is installed between the feed box arms and, using a specially designed housing and a mobile carriage, is directly docked onto the face mix device. When the face mix device opens, the QC Imager moves with it. The proximity to the compaction point allows immediate corrective action when necessary.
The KBH production team was also convinced by the fact that the system comes pre-loaded with a highly capable AI model. Teaching additional features is also fast when unique characteristics need to be added.
Consistently positive experiences with the QC Imager
A key component of modern quality control concepts is the automated removal or replacement of defective products. The QC Imager can be combined with various automation levels—ranging from manual decisions supported by automated detection to fully automated sorting on the dry side, typically using KBH robotics. This allows manufacturers to gradually increase their level of automation and autonomy while aligning investments with their operational needs. KBH Maschinenbau has developed various approaches to meet these requirements as well.
Beyond improving product quality, the QC Imager opens new possibilities for handling process and quality data. The recorded images and information can be stored, analyzed, and used long-term for process optimization.
In summary, KBH’s in-house experience with the QC Imager in concrete block and paving block production has been consistently positive. The measurable contribution to efficiency and quality assurance is significant. Crucially, the first step was not maximum automation, but rather the intelligent integration into existing production concepts and immediate quality feedback at the block machine. The KBH QC Imager is designed as a tool to support human operators and sustainably improve industrial processes.
