Machine Vision MV Tec expands deep-learning-based anomaly detection in Halcon 22.05

Source: Press release

Innovative anomaly detection extends the deep learning product range of MV Tec Software: The company will release the new version (22.05) of its Halcon machine vision software on May 25, 2022. The highlight is the new technology “Global Context Anomaly Detection”, which is available in Halcon 22.05 in this form as a world`s first.

Related Companies

Companies using the new version of Halcon machine vision software benefit from more efficient production.
Companies using the new version of Halcon machine vision software benefit from more efficient production.
(Source: MV Tec)

The new release of MV Tec’s Halcon software features an expansion of the long-established anomaly detection technology, taking deep learning to a new level. The new release also includes significant expansions such as new features as well as new improvements to the software’s core technologies. As a result, the software now enables the practical implementation of software solutions for even more demanding applications across a wide range of industries. Companies using this machine vision software benefit from more efficient production, especially in application areas like quality assurance, the developers say.

“The new technology provides our customers with brand-new possibilities, such as for inspection activities. We’ve also added a deep-learning-based training option to the Deep OCR feature,” explains Mario Bohnacker, Technical Product Manager for Halcon at MV Tec Software.

By detecting logical anomalies in images, Halcon 22.05 opens up new application areas and represents a further development of the deep learning technology of anomaly detection. Until now, it has only been possible to detect structural anomalies strictly on a local level. According to the company, the new Global Context Anomaly Detection feature is currently the only technology that can understand the logical content of the entire image. Like the existing anomaly detection in Halcon, Global Context Anomaly Detection requires only good images for training. The training data does not need to be labeled. The technology can thus detect completely new anomaly variants, such as missing, deformed, or incorrectly arranged components of an assembly, for example.

Using Halcon’s Deep OCR, users can address OCR applications for a variety of application areas. The release of version 22.05 now expands this technology to include a training function that allows users to perform individualised training based on their own application dataset. This makes it possible to handle even the most complex applications, such as reading text with poor contrast. As a further benefit, it is possible to train special characters and print styles that are very rarely used. Ultimately, training for Deep OCR improves performance and user-friendliness and makes applications more robust.

Halcon supports various standards for evaluating the print quality of 1D and 2D codes. This ensures that all readers will have no trouble reading the printed code in actual practice. Version 22.05 brings further improvements to the print quality inspection (PQI) of bar codes and data codes, making the determination of the module grid for the print quality inspection of the ECC200 code much more robust. Moreover, the PQI of 2D data codes is now up to 150 percent faster. And finally, the user-friendliness of the PQI of 2D data codes has been improved through the addition of a new method for calculating the evaluations.

Halcon 22.05 offers still more improvements — for example, a new operator that helps to optimise image contrast locally. Another new operator permits image smoothing with randomly shaped regions.

Subscribe to the newsletter now

Don't Miss out on Our Best Content

By clicking on „Subscribe to Newsletter“ I agree to the processing and use of my data according to the consent form (please expand for details) and accept the Terms of Use. For more information, please see our Privacy Policy.

Unfold for details of your consent

(ID:48207492)