Kneron introduces KNEO Pi for edge AI development

Kneron announced the launch of KNEO Pi, a development platform designed for edge AI applications. The platform features a proprietary KDP-series Neural Processing Unit (NPU), an open software ecosystem, and industrial-grade reliability to support real-time AI inference and deployment.
KNEO Pi’s NPU delivers up to 4 eTOPS of INT8 performance, using 2W of power for tasks like YOLOv5 object detection at over 30 FPS, compared to 2–3 FPS on Raspberry Pi’s CPU mode. It supports parallel processing of up to four camera streams for applications like facial recognition and vehicle detection, with local processing for zero-latency inference.

The platform includes onboard storage with SD card support, a 40-pin GPIO compatible with Raspberry Pi sensors and displays, and a pre-installed KNEO Pi OS based on ArchLinux ARM, with NPU drivers, media codec libraries, and AI model libraries for quick deployment.
KNEO Pi operates in temperatures from -20°C to 70°C and features certified electromagnetic shielding for industrial environments. Use cases include smart security systems with sub-0.3 second facial recognition, industrial inspection with 99.2% accuracy for defect analysis, and educational projects with Python/C++ compatibility and open-source templates.
The plat...