BeagleBone AI supercharges machine learning and computer vision on Raspberry Pi-style board

Unveiled this week, the single-board computer houses four dedicated chips originally designed to help self-driving cars “see” the world around them.

The newly revealed BeagleBone AI is a board aimed at developers interested in experimenting with machine-learning and computer vision.

Unveiled this week, the computer houses four dedicated chips originally designed to help self-driving cars “see” the world around them.

Image: BeagleBoard.org Foundation

The board’s Texas Instruments (TI) Embedded Vision Engine (EVE) chips offer up to 8x the performance per watt when running calculations for computer-vision models compared to running on an Arm Cortex A15-based CPU. This optimized hardware is accessible to developers via the TI Deep Learning OpenCL (Open Computing Language) API.

Combined with the board’s two TI C66x digital-signal-processor (DSP) cores, the BeagleBoard.org Foundation say the BeagleBone AI board will be able to automate tasks in industrial, commercial and home settings.

It’s been possible to experiment with running trained machine-learning models on low-cost, single-board computers like the Raspberry Pi for a while, although getting good results generally requires investing in additional hardware. Google offers its Artificial Intelligence Yourself kits for carrying out speech and image recognition on the Pi and is readying a USB stick that will turbo charge the Pi’s machine-learning capabilities.

SEE: Raspberry Pi: The inside story of the $35 computer

Premium boards have also started shipping with custom chips for running trained machine-learning models, such as the HiKey 960 with its dedicated Neural Processing Unit.

The BeagleBoard.org Foundation pitch the BeagleBone AI as falling somewhere between a low-cost single-board computer like the Pi and more powerful industrial computers.

While there’s no price or release date yet for the board, the BeagleBoard X15, which offers similar specs, sells for about €240 or $300.

The BeagleBone AI is based around a TI AM5729 SoC, which offers a dual-core Arm Cortex A15-based CPU, paired with a dual-core Arm Cortex M4 CPUs that serve as an image-processing unit. Also included is a dual-core PowerVR SGX544 3D GPU, Vivante GC320 Core 2D accelerator, 1GB of DDR3 RAM, 16GB on-board eMMC flash with a high-speed interface, Gigabit Ethernet, “high-speed” Wi-Fi, one USB Type-A port, and USB Type-C port for power.

It also offers four Programmable Realtime Unit cores and has mechanical and header compatibility with the BeagleBone Black.