MVTec Updates Machine Vision Software for Faster AI Processing

Industrial machine vision systems increasingly rely on both deep learning and rule-based algorithms to meet throughput and accuracy requirements in automated inspection. Performance constraints often limit deployment in high-speed production environments. In this context, MVTec Software GmbH has announced the release of HALCON 26.05, scheduled for May 20, 2026, with updates focused on improving processing speed and system efficiency.

Performance improvements across AI and rule-based methods
HALCON 26.05 introduces optimizations for both deep learning and rule-based image processing workflows. These improvements target faster execution times in applications such as quality inspection, object detection, and code reading, where latency directly impacts production throughput.

The updated deep learning-based object detection supports faster inference while maintaining detection accuracy across varying object sizes, including small or irregularly shaped items. Integrated data augmentation techniques improve robustness against environmental variations such as lighting changes, rotations, and partial occlusions.

Enhancing inspection reliability with new image processing functions
A new rectification function enables reading of 2D data codes on curved or deformed surfaces. This is relevant for industries such as packaging and electronics, where markings may not be applied on flat geometries.

The Shape Matching workflow has also been updated with automated contour optimization. This feature removes unstable or misleading contours from training samples, resulting in more consistent matching results and reduced processing overhead during runtime.

These additions improve reliability in rule-based inspection tasks while supporting higher processing speeds in automated production lines.

Development workflow updates with HDevelopEVO
Alongside the HALCON release, a preview version of the HDevelopEVO development environment is introduced. The update enables integration of scripts created in HDevelopEVO into custom applications through a .NET interface, supporting more flexible deployment of machine vision solutions.

The environment also expands support for multimodal interaction with large language models through visual prompting. Developers can incorporate image data directly into prompts, enabling more interactive workflows when configuring or testing vision algorithms.

Implications for industrial automation systems
The updates in HALCON 26.05 reflect ongoing efforts to balance accuracy and execution speed in machine vision applications. Faster inference and optimized rule-based processing support higher inspection throughput, which is critical in sectors such as manufacturing, logistics, and electronics.

By combining performance improvements with development workflow enhancements, the release supports integration of machine vision into broader automation architectures and digital supply chain environments, where processing efficiency and adaptability are key requirements.

Edited by Aishwarya Mambet, Induportals Editor, with AI assistance.

www.mvtec.com