youyeetoo TC-RV1126B/TC-RV1109B AI Vision Board - Stamp Hole Core Board - 2.0 Tops NPU, 3 Camera Input Supports Buildroot+QT OS

Overview

The youyeetoo TC-RV1126B/TC-RV1109B AI Vision Board is a compact 48 mm × 48 mm stamp hole core module built around the Rockchip RV1126 low-power AI vision processor. Designed for embedded vision applications such as face recognition, gate access control, smart IP cameras, and self-service terminals, this system-on-module integrates a quad-core 32-bit ARM Cortex-A7 CPU running up to 1.5 GHz alongside a dedicated 2.0 TOPS NPU for on-device inference. The 14 nm chip supports 4K H.264/H.265 video encoding and decoding at 30 FPS, multi-level image noise reduction, and 3F-HDR technology for clear, full-color output even in low-light security scenarios.

The stamp hole core board features a 6-layer immersion-gold PCB with up to 172 pins, offering rich interfaces including I2C, SPI, UART, ADC, PWM, GPIO, USB 2.0, SDIO, I2S, MIPI-DSI, MIPI-CSI, CIF, and PHY. It supports up to three camera inputs and runs the Buildroot+QT operating system for fast boot times, small storage footprint, and stable operation with open-source code access. LPDDR4 memory options of 1 GB or 2 GB and eMMC storage options of 8 GB or 16 GB are available.

Key Features

• 2.0 TOPS NPU — Built-in neural network processor delivers 2.0 TOPS computing power at under 10% of the energy a GPU would require for equivalent AI workloads.
• 4K Video Codec — Supports 4K H.264/H.265 encoding and decoding at 30 FPS with multi-channel processing for low-bitrate, low-latency video output.
• 3F-HDR Imaging — Multi-level noise reduction and HDR technology ensure full-color visibility in darkness, meeting real-world security and surveillance demands.
• 172-Pin Scalability — 48 mm × 48 mm 6-layer immersion-gold PCB exposes up to 172 pins across I2C, SPI, UART, USB 2.0, MIPI-CSI, MIPI-DSI, and more.
• Buildroot+QT OS — Lightweight operating system occupies minimal storage, boots quickly, and ships with open-source code for custom development.
• Framework Support — NPU toolchain enables direct conversion and deployment of TensorFlow, PyTorch, Caffe, MXNet, DarkNet, and ONNX models.

Applications

• Face recognition and gesture recognition systems
• Gate access control and smart security installations
• Smart IP cameras and smart doorbells or peephole viewers
• Self-service terminals and smart finance kiosks
• Smart construction site monitoring and smart travel systems

What's in the Box

• 1 × TC-RV1126B/TC-RV1109B stamp hole core board (SOM) / development Board

Compatibility & Support

The AI vision board runs Buildroot+QT OS with open-source code provided for custom development. AI model deployment is supported via the Rockchip NPU toolchain for TensorFlow, PyTorch, Caffe, MXNet, DarkNet, and ONNX frameworks. Tools, drivers, and firmware are available through the youyeetoo documentation wiki at youyeetoo.com/blog/rv1126-hole-boardtkdp1126cor1-143. The board supports FastBoot and TrustZone technology with multiple crypto engines. For technical support and warranty details, contact the youyeetoo support team.

Frequently Asked Questions

What processor does the TC-RV1126B/TC-RV1109B AI vision board use?

It uses the Rockchip RV1126 low-power AI vision processor built on a 14 nm process with a quad-core 32-bit ARM Cortex-A7 architecture running up to 1.5 GHz, integrating NEON and FPU instructions.

How much AI computing power does the built-in NPU provide?

The integrated NPU delivers 2.0 TOPS of computing power while consuming less than 10% of the energy a GPU would require for equivalent AI inference tasks.

Which AI frameworks are supported for model deployment?

The NPU toolchain supports direct conversion and deployment of models from TensorFlow, PyTorch, Caffe, MXNet, DarkNet, and ONNX frameworks.

How many camera inputs does the board support?

The board supports up to three camera inputs through its MIPI-CSI and CIF interfaces, making it suitable for multi-camera vision applications such as smart security and face recognition.

What operating system does the board run?

It runs the Buildroot+QT operating system, which offers a small storage footprint, fast boot times, and stable operation. Source code is open for developers to customize.

What memory and storage options are available?

The board is offered with 1 GB or 2 GB LPDDR4 RAM and 8 GB or 16 GB eMMC storage. Confirm specific configurations at purchase, as options may vary by SKU.

Where can I download drivers, tools, and firmware?

Drivers, tools, and firmware are available on the youyeetoo documentation wiki. Visit the product blog page linked on this page for direct access to all resources.

What is the warranty and support policy?

Warranty terms and technical support are provided by youyeetoo. Contact the support team through the youyeetoo website for assistance with hardware, software, and development questions.