Renesas RZ/V2L Evaluation Board Kit (RTK9754L23S01000BE) for AI Vision

Design AI-powered vision solutions with the RZ/V2L Evaluation Kit, combining DRP-AI acceleration, Cortex-A55 CPU, DRP-AI accelerator, Mali-G31 GPU, MIPI camera, and SMARC v2.1 SOM support.

The Renesas RZ/V2L Evaluation Board Kit (RTK9754L23S01000BE) is a complete platform for AI-powered vision application development, combining the RZ/V2L module (SOM), carrier board, and a MIPI CSI camera module. Designed for seamless evaluation, the carrier board supports multiple SMARC v2.1 modules, including RZ/G2L, RZ/G2LC, RZ/G2UL, and RZ/V2L, enabling flexible prototyping across different platforms.

At its core, the RZ/V2L features a 1.2 GHz Arm® Cortex®-A55 CPU, a DRP-AI accelerator, and a Mali-G31 GPU, providing efficient real-time AI inference and image processing. DRP-AI includes AI-MAC and DRP cores, enabling high-performance vision tasks such as color correction, noise reduction, and object detection without the need for an external ISP. The board supports 16-bit DDR3L/DDR4 memory, a 3D graphics engine, and H.264 video codec, delivering both computational and graphical capabilities for AI vision applications.

Thanks to the power-efficient DRP-AI, heat dissipation solutions like fans or heat sinks are unnecessary, reducing system complexity and cost. The kit is ideal for consumer electronics, industrial equipment, retail POS systems, and a wide range of camera-based AI applications, allowing rapid prototyping and evaluation of advanced vision solutions.

Buy Now ►

Quick Start Guide ► About Renesas ►

AI Model Deployer

Renesas AI model deployer is a no-code, intuitive interface that streamlines the entire Vision AI development pipeline from model selection and training to deployment on Renesas MPUs and MCUs powered by NVIDIA TAO. It enables embedded developers to quickly build, optimize, and deploy AI models with guided workflows, hardware-aware tooling, and optional deep customization via Jupyter notebooks. The tool includes computer vision pipelines for object detection/image classification including dataset conversion, model training, model evaluation, inference and deployment. The QSG covers the following modules:

Guidance for how to set up Renesas AI model deployer interface, and RZ/V2L-EVKIT board for model deployment.
End-to-end tutorial from model selection to dataset curation to training and evaluation.
Exporting the model, compiling for DRP-AI and deploying onto RZ/V2L-EVKIT board.

The QSG is based on Detectnetv2 demo, the GUI and the Jupyter notebook support various others, such as mobilenetv2 image classification on RA8 MCU.

High Level Architecture

Data management is performed locally using desktop data management. The toolkit comprises two workflows: MPU workflow and MCU workflow. In the MPU workflow, once the model is trained on the toolkit, it is exported via ONNX and compiled by DRP AI TVM for deployment on the board. In the MCU workflow, following model training on the toolkit, the model is exported to ONNX, translated to TFLite for quantization, and subsequently exported in .c format for MCU deployment. Real-time inference is achieved through web socket streaming from the board to the UI on MPU and via Segger Jlink for MCU.

Kit Contents

RZ/V2L SMARC module board
RZ SMARC series carrier board
MIPI camera module.

Features

Device: RZ/V2L
- Cortex-A55 Dual, Cortex-M33
- BGA551pin, 15mmSq body, 0.5mm pitch
Module Board Function
- DDR4 SDRAM: 2GB × 1pc
- QSPI flash memory: 512Mb × 1pc
- eMMC memory: 64GB × 1pc
- The microSD card slot is implemented and used as an eSD for boot.
- 5-output clock oscillator 5P35023 implemented
- PMIC power supply RAA215300 implemented
Carrier Board Function
- The FFC/FPC connector is mounted as standard for connection to high-speed serial interface for camera module.
- The Micro-HDMI connector via DSI/HDMI conversion module is mounted as standard for connection to high-speed serial interface for digital video module.
- The Micro-AB receptacle (ch0: USB2.0 OTG) and A receptacle (ch1: USB2.0 Host) are respectively mounted as standard for connection to USB interface.
- The RJ45 connector is mounted as standard for software development and evaluation using Ethernet.
- The audio codec is mounted as standard for advance development of audio system. The audio jack is implemented for connection to audio interface.
- The CAN connector is implemented for connection to CAN-Bus interface.
- The Micro-AB receptacles are implemented for connection to asynchronous serial port interface.
- The microSD card slot and two sockets for PMOD are implemented as an interface for RZ/V2L peripheral functions.
- For power supply, a mounted USB Type-C receptacle supports the USB PD standard.
MIPI Camera Module
- MIPI Camera Module (MIPI CSI) is included. Image recognition processing can be used with images input with MIPI camera.