Google Coral Edge's TPU (Tensor Processing Acceleration Unit), with up to 4 TOPS (trillion operations per second), ensures significantly faster image and video data processing. Compared to an application using simple USB cameras without a TPU at approximately 6 frames per second, the TPU accelerates to 30 frames per second, making it five times faster.
Artificial Intelligence at the Intelligent Edge is gaining increasing importance in industrial automation. Kontron's new compact AI platform consists of an M.2 module with the Google Coral accelerator chip for the TensorFlow Lite software ecosystem in a Kontron 2.5" pITX SBC with an NXP i.MX8M processor. The TPU supports small, low-power applications and requires only 1 watt for 2 TOPS, and 2 watts for 4 TOPS.
All current TensorFlow Lite applications, as well as pre-trained models, can be downloaded free of charge. This enables the rapid and easy development of custom neural networks with AI/ML/DL applications, and therefore a reduced time to market.
Kontron's solution can also be flexibly customized to meet the needs of specific applications. The platform is suitable for use in harsh environments with extended temperature ranges from -40° to +85° Celsius and is compatible with Yocto Linux using the latest kernel.
