Euresys respond to the trend for more decentralized device-based expertise and processing by offering their recognized Open eVision
image processing libraries for embedded systems.
While imaging embedded systems become smaller, cheaper and less power hungry, the porting of the Open eVision libraries onto these systems offers, where no centralized image processing is needed, the Machine Vision manufacturers and integrators the flexibility of deploying the processing at its immediate point of acquisition.
Beyond running on x86 processors, today’s release of Open eVision 22.04
also runs on ARMv8-A compatible processors running 64-bit Linux, such as Raspberry Pi, Nvidia Jetson, Xilinx Kria K26 SOM, as well as many smart cameras like ADLink Neon-2000, IDS NXT rio, Lucid Triton Edge or Baumer AX Series (and more to come).
All 18 libraries of Open eVision are now available for ARM, including 3D inspection, deep learning, subpixel measurement, alignment, surface inspection, OCR and 1D and 2D code reading. The acceleration of the deep learning functions using GPU will be available soon.
While Open eVision Studio
applications are not available for embedded systems, all versions of Open eVision are compatible and it is possible, for example, to develop and save model files using Open eVision Studio on a PC under Windows and use them with the ARM version on the embedded system. Depending on the capabilities of the chosen target, it is either possible to compile the program on the target platform itself, or perform a cross compilation on a Linux x86 PC prior to sending it to the target.