MVTec Software GmbH, a leading international machine vision software provider, will release version 23.05 of its standard machine vision software HALCON on May 23, 2023. The focus of the new version is deep learning methods. The main feature is Deep Counting, a deep learning-based method that allows for stable and reliable counting of large numbers of objects. In addition, the new HALCON version incorporates improvements to the deep learning technology 3D gripper point detection as well as Deep OCR training. With HALCON 23.05, the underlying deep learning networks, which have been pre-trained with industry-relevant images for users’ own applications, can now be further optimized. This makes the recognition rate of Deep OCR applications more stable and also helps applications using 3D gripper point detection technology to detect suitable gripping surfaces more reliably. In addition, there are many other beneficial improvements, for example, it is now easier to integrate external code into HALCON.

“We have seen a significant increase in customer interest in integrating deep learning methods into their own solutions. This is what guided us when developing the new HALCON version, and the result is new deep learning technologies and further developments that help customers achieve even more precise results,” explains Jan Grötner, Product Manager for HALCON at MVTec.

Deep Counting

Starting with HALCON 23.05, customers can use the “Deep Counting” function, which allows fast and reliable counting and position detection of a large number of objects. This deep learning-based technology has clear advantages over existing machine vision methods: The function can be deployed very quickly, because only a few objects need to be labeled and trained, and both steps can be easily done in HALCON. The technology provides reliable results even for objects made of highly reflective amorphous materials. Deep Counting allows a large number of objects to be counted, such as glass bottles, tree trunks, or food products.

Training of Deep OCR

Deep OCR reads text very stably and is not even affected by the orientation and font. This technology first detects the relevant text in the image and then reads it. With HALCON 23.05, it is now also possible to fine-tune the text detection by retraining the pre-trained network with application-specific images. This produces more stable results and opens up new possibilities. For example: detecting text in arbitrary print types or previously unseen character types, and improving reading in low-contrast, noisy environments.

Training for 3D gripper point detection

The 3D gripper point detection reliably detects surfaces on any object that are suitable for gripping with suction. In HALCON 23.05, it is now possible to retrain the pre-trained model with your own application-specific image data. This results in a more stable recognition of grippable surfaces. The necessary annotations can be done easily and efficiently with the MVTec Deep Learning Tool.

Easy extension interface

With the help of HALCON extension packages, external programming languages ​​can be integrated. The integration of external code in HALCON 23.05 is even easier. With the new Easy Extensions Interface, users can now use their own functions written in .NET code in HDevelop and HDevEngine in just a few steps. Even data types and HALCON operators known from the HALCON/.NET language interface can be used. Customer benefit: HALCON can now cover functions that go beyond pure image processing. This increases HALCON’s flexibility and application possibilities.
Related product recommendations:
ABB PM860
ABB PM866K02 3BSE050199R1
ABB PM862K02
ABB PM803F
ABB PM856AK01 3BSE066490R1
ABB PM891K01 3BSE053241R1
ABB PM860 3BSE018110R1
ABB PM856AK01 3BSE066490R1
ABB PM866AK02
ABB 3BDH000606R1 PM875-2
ABB PM858K01
ABB PM825 3BSE010796R1
ABB PM864AK01 3BSE018161R1
ABB PM856K01
ABB PM862K01
ABB PM851AK01
Siemens PS16 6DD1602-0AE0
MCGS TPC7062HI
ABB PTQ-PDPMV1
AB 1775-SR
VIPA CPU 317-2AJ12
ABB PC D231 3BHE025541R0101
ABB DSQC345D
ABB 3BHE020455R0001 PPD103 B01
ABB 3BHE041576R3011 PPD517 A3011
ABB 3BHE023784R1023 PPD113 B01-10-150000
ABB PPD115A02 3BHE017628R0002
ABB PPD113B03-26-100100 3BHE023584R2625
PPE091A101 3BHE044481R0101
PU512V2 3BUR001401R1
PPD512A10-454000 3BHE040375R103E
PPD539A102 3BHE039770R0102
PU516A 3BSE032402R1
PU517 3BSC980050R46
RC527 3BSE008154R1
ABB SA801F 3BDH000011R1
ABB HIEE450964R0001
ABB S-113H 3BHB018008R0003
ABB SC520 3BSE003816R1
ABB SC520M 3BSE016237R1
ABB SC540 3BSE006096R1
SC610  3BSE001552R1
SCYC51010 58052515G
SCYC51020 58052582H
SCYC51040 58052680E
More……