The software combines machine vision and deep learning


MVTec Software GmbH plans to release a new version (22.11) of its machine vision toolkit HALCON on November 22, 2022. The new functionality includes 3D gripping point detection, which detects surfaces on objects suitable for gripping by suction, such as in bin picking applications. This feature eliminates the need to train the software on object surfaces. The company has also added a new data type, memory block, which stores and transfers binary data in HALCON and then processes it with other applications, increasing HALCON’s compatibility with machine vision communication protocols. The toolkit also has an option to encrypt serialized data, improving security. Additionally, the new release sheds light on the “black box” of deep learning through Guided GradCam. This feature provides a heat map highlighting the regions of an image that the deep learning network used in its decision making. HALCON 22.11 supports HAILO AI acceleration hardware to perform deep learning inference. Finally, users can obtain a HALCON license via a network, allowing developers to share a predefined number of licenses using a network connection.

To learn more:

Contact: MVTec Software GmbH

Headquarter: Munich, Germany

Product: HALCON 22.11

Main characteristics: Detects surfaces on objects suitable for vacuum gripping, Encrypts serialized data, Heat map highlights regions of an image that the deep learning network used in its decision making.

What MVTec says:

Show more information about HALCON 22.11.

Share your vision-related news by contacting Chris McLoone, Editor-in-Chief, Vision Systems Design.


Comments are closed.