“The core focus of the new HALCON version is speed. Our goal is for machine vision applications to operate not only reliably and accurately, but also quickly across all industrial sectors and in demanding automation scenarios. For HALCON 26.05, we have improved the speed and performance of both deep learning methods and rule-based procedures,” explains Jan Gärtner, HALCON Product Manager at MVTec.

For example, a new rectification function has been integrated into HALCON in the new version. This enables the fast and reliable reading of 2D data codes on curved or deformed surfaces. Furthermore, automated contour optimization via Shape Matching is now possible. This optimization eliminates unstable or misleading contours from sample images, resulting in faster, more stable, and more accurate matching. MVTec is also expanding its capabilities in the area of ​​AI-based object detection. The new generation of object detection based on deep learning in HALCON 26.05 enables significantly faster inference with high detection accuracy. Detection also works reliably with small objects and objects of highly variable sizes. Integrated data augmentation techniques also improve resilience to changes in lighting, rotations, and obscurations.

A new preview version of MVTec's HDevelopEVO integrated development environment is also available as part of HALCON 26.05. With this new version, it is now possible to integrate scripts created in HDevelopEVO into custom applications via the .NET interface. Furthermore, LLM's multimodal support has been enhanced with visual prompting. Developers can now integrate image data directly into the AI ​​wizard prompts.

More information