MVTec Software GmbH has successfully tested HALCON, the standard machine vision software, on DragonBoard using the board's native 64-bit instruction set. The tests were performed on a DragonBoard using a quad-core ARM Cortex A53 processor. Compared to 32-bit ARMv7 code running on the same hardware, the 64-bit operating mode of the ARM Cortex A53 processor can speed up vision algorithms such as shape-based matching by as much as 40 percent. The availability of HALCON Embedded for DragonBoard will enable high-performance machine vision applications to take full advantage of the 64-bit ARMv8 architecture. MVTec offers to port HALCON Embedded to customer hardware on request, taking into account the customers' specific requirements.
64-bit architectures such as ARM Cortex A53 are increasingly becoming standard. Due to the many benefits of this technology, a wide range of newly developed smartphones are equipped with modern 64-bit processors. The use of DragonBoard is just as widespread. Due to its compact dimensions and low energy consumption, it is predestined for use in mobile environments. The board is therefore used for embedded systems not only in cell phones but also in all conceivable smart devices such as cameras and handhelds — for personal computing as well as in company production and logistics processes.
Faster machine vision saves money
"By porting HALCON to DragonBoard, we take into account the growing importance and widespread use of 64-bit processors. Customers can address us with their very specific requirements. We then provide them with a solution that is tailored to their exact needs. Companies will thus be able to significantly speed up their machine vision processes. Certain applications can also impressively increase their performance on mobile devices. Examples include smartphone-based identification of components, which simplifies processes such as the maintenance of industrial plants and thereby cuts costs," explains Johannes Hiltner, Product Manager HALCON at MVTec Software GmbH.
