Elma Electronic in Wetzikon, Switzerland, has introduced its 3U CompactPCI Serial-based JetKit-3010 single board computing engine that delivers artificial intelligence (AI)-enabled video inspection, analysis, object detection and recognition capabilities to systems developers.
The JetKit-3010 embedded computing kit uses the integrated deep learning capabilities and I/O of the NVIDIA Jetson AGX Xavier general-purpose graphics processing unit (GPGPU) to help designers develop smart and reliable systems for industrial automation, automotive and agriculture applications.
Source: Elma
Prior to the open standards-based JetKit-3010, research and development with AI components required extremely powerful and specialized hardware, much of which is purpose-built and expensive.
The combination of the NVIDIA GPGPU module in a CompactPCI Serial architecture can help develop applications that require pattern recognition, environment recognition or situation analysis.
The JetKit-3010 combines an 8-core ARM processor with 512 NVIDIA CUDA cores and 64 Tensor cores to provide compute density, energy efficiency and AI inferencing capabilities.
The JetKit-3010 can be packaged as a standalone small-form-factor platform or integrated as a multi-function computer board in an existing CompactPCI system. It provides M.2 NVME, HDMI, 4x USB, 2x Gigabit Ethernet, and x8 PCI Express ports.
The display controller supports imaging at 3840-by-2160 resolution at 60 Hz via HDMI, and multi-stream HD video and JPEG support for most encode and decode video standards. A dedicated programmable ARM Cortex A9 audio processor provides a multi-channel, high-definition audio path via the HDMI interface.
For more information, visit the Elma Electronic website.
