MVTec has published version 21.05 of its HALCON machine vision software, which includes a new Generic Shape Matching feature and improvements in the Deep Learning Framework, subpixel bar code reading, optical character recognition (OCR), and the IDE itself.
The Shape Matching Feature allows combines HALCON’s shape-matching functions into a single set of easy-to-use operators.
Support for the NVIDIA TensorRT inference engine has been added to the HALCON Deep Learning Framework.
The integrated subpixel bar code reader has been optimized for low-resolution codes, improving the decode rate for these workloads by up to 50 percent.
The deep learning-based OCR function handles large images and different worde lengths more efficiently without parameter tuning. It also includes additional character support.
The IDE window docking feature has been optimized with capabilities like “Auto-hide” of widgets and simplified grouping and organization of windows.