EVT offers a broad spectrum of ARM boards with MIPI sensors for compact, powerful embedded vision systems – custom-made for each customer.
In addition, EyeVision offers deep learning algorithms for these embedded vision platforms. Therefore, solutions such as precise surface inspection and number plate recognition are possible.
The MIPI sensors are directly from the EVT hardware development and are available in different resolutions and models. Those can be combined with different ARM boards, such as the Raspberry Pi 4 or Zero, NVIDIA or the Pine board.
The small form factor is especially an advantage, when there is only constricted room. Exact adjustments are as easy to realise as complete and complex housing modifications.
The EyeVision deep learning algorithms run on those embedded vision systems. The software has ready-to-use commands in the graphical user interface with drag-and-drop programming for surface inspections and number plate reading. Therefore, complex applications can be solved with only a few mouse clicks.
The surfaces can be inspected for cracks, scratches, rills, grout, inadequate machining surface, etc. without complex configurations. The different error types are classified and are no problem for the Deep Learning algorithms. Therefore, the Deep Learning AddOn can be integrated very quickly into the production process.
The operation system is Linux and together with the EyeVision software the embedded vision systems can also be integrated into the SPS with the Profinet interface. Also, other communication interfaces are available such as Modbus, TCP/IP or UDP. In addition, the OPC UA interface allows an integration into the SCADA system.