Raspberry Pis might be the most well-known single-board computers (SBCs) on the market, but that doesn’t mean they are always the best choice. That is especially true now that supply shortages are limiting availability and driving up prices. If you’re developing for AI vision applications, the new Avnet RZBoard V2L could be the Raspberry Pi alternative that you’ve been looking for.
Avnet is Hackster’s parent company. They distribute semiconductor products and provide many solutions for technology companies. They also design and manufacture a variety of hardware products. Among those products is the brand new Avnet RZBoard V2L.
The Avnet RZBoard V2L is an SBC that mirrors the form factor of the current Raspberry Pi 4 Model B. But while it may look the same—and even has a pin header compatible with Raspberry Pi HATs—it packs a lot more power thanks to the Renesas RZ/V2L system-on-chip (SoC).
Renesas developed the RZ/V2L SoC for AI applications. It contains a high-performance 1.2GHz dual-core ARM Cortex-A55 CPU, an additional 200MHz ARM Cortex-M33, an on-chip DRP-AI accelerator, on-chip H.264 1080p encoding/decoding, and a MALI GT31 3D-GPU running at 500MHz with image scaling unit.
To complement the Renesas RZ/V2L, the Avnet RZBoard V2L includes 2GB of DDR4 RAM, 32GB eMMC, and 16MB QSPI. Connectivity includes Gigabit Ethernet, 801.11ac WiFi, Bluetooth 5, two USB 2.0 Host ports, and one USB 2.0 OTG port. An onboard CANFD transceiver, MIPI-DSI and HDMI, and MIPI-CSI camera interface allow for peripheral connections.
All of that hardware sounds impressive, but what can you do with it? The RZBoard V2L’s primary purpose is vision-based artificial intelligence and machine learning. Connect a USB webcam or standard CSI camera and you can immediately take advantage of the DRP-AI acceleration. That provides 1 TOPS/W and is capable of Tiny YOLOv2 object-detection at a blazing 28FPS. While it works with Tiny YOLOv2 right out of the box, you can use other neural networks and machine learning models.
Outputting the results with a GUI is also easy, thanks to the HDMI and MIPI-DSI video options. With the Avnet RZBoard V2L, you can setup a complete vision system with just three other components: an MIPI-CSI camera or USB webcam, an MIPI-DSI or HDMI display, and a 5V USB-C power supply.
If you do require other hardware for your application, you have many options. The Avnet RZBoard V2L is fully compatible with the Pi HAT expansion boards already common in the community, as well as MikroE Click boards. More than 1200 Click boards are available today and are orderable from Avnet. MikroE has a library of open source code available for these boards.