· AI Vision Processor RV1126B
· Deployment of Lightweight AI models
· Built-in 12MP ISP
· 4K High-Frame-Rate Video Decoding
· Simultaneous Input From 4 cameras
· Compact Gold-Finger Package
· Robust Network Capabilities
· Rich Expansion Interfaces
Debian
Buildroot+QT
Ubuntu
Powered by Rockchip quad-core AI vision processor RV1126B, it integrates 3 TOPS NPU and supports mainstream large AI models and deep learning frameworks. It features a built-in 12MP ISP, an 8MP AI-ISP, and incorporates multiple algorithm accelerators. The SoM supports 12MP video encoding, 4K video decoding, and simultaneous input from 4 cameras.
AI Processor RV1126B
Lightweight AI models
Built-in 12MP ISP
4K Video Decoding
Input From 4 cameras
Gold-Finger Package
Robust Network Capabilities
Rich Expansion Interfaces
Integrated with a 3 TOPS NPU supporting INT4/INT8/INT16/FP16 hybrid quantization, it maintains compatibility with mainstream deep learning frameworks including TensorFlow, TensorFlow Lite, PyTorch, Caffe, and ONNX. This enables intelligent data processing, voice recognition, and image analysis capabilities, meeting the demands of most edge AI applications in terminal devices.
CORE-1126BJD4 |
ICORE-1126BJJD4 |
||
Basic Specifications |
SOC |
Rockchip RV1126B |
Rockchip RV1126BJ |
CPU |
Quad-core 64-bit ARM Cortex-A53, integrated with NEON and FPU |
||
NPU |
3TOPS NPU, supporting INT4/INT8/INT16/FP16 mixed-precision computing |
||
ISP |
12MP ISP, 8MP AI-ISP, integrated with multiple algorithm accelerators including HDR, 3A, LSC, 3DNR, 2DNR, sharpening, defogging, fisheye correction, gamma correction, feature point detection, etc. |
||
Encoding/ Decoding |
Encoding: 12M@30fps H.264 Decoding: 4K@30fps H.264 |
||
RAM |
LPDDR4 (1GB/2GB/4GB Optional) |
||
Storage |
eMMC (8GB/16GB/32GB Optional) |
||
Power |
5V (voltage tolerance ±5%) |
||
OS |
Debian12, Buildroot+QT, Ubuntu |
||
Software support |
Supports private deployment of lightweight large language models and multimodal large models with up to 2B parameters under the Transformer architecture, such as small-sized large AI models including the Qwen series, Gemma2-2B, Phi2, InternLM2, MiniCPM series, TinyLLAMA, and RWKV7. Supports deep learning frameworks such as TensorFlow, TensorFlow Lite, PyTorch, Caffe, and ONNX. |
||
Interface |
Gold finger (SODIMM 260P standard connector, 0.5mm pitch) |
||
Size |
69.9mm × 33.96mm |
||
Weight |
≈10g |
||
Environment |
Operating Temperature: -20℃ ~ 60℃, Storage Humidity: 10%~90%RH (non-condensing) |
Operating Temperature: -40℃ ~ 85℃, Storage Humidity: 10%~90%RH (non-condensing) |
|
Interface Specifications |
Internet |
One Gigabit Ethernet port can be expanded via the GMAC interface, with support for network acceleration features including TSO (TCP Segmentation Offload) and USO (UDP Segmentation Offload) Dual-band 2.4GHz/5GHz WiFi and Bluetooth can be expanded via the SDIO3.0 interface, and the interface also supports 4G/3G network expansion |
|
Video input |
2 × MIPI-CSI/LVDS/sub LVDS DPHY (1×4lanes or 2×2lanes) 1 × DVP (BT.601/BT.656/BT.1120) Support 4 cameras input at the same time: 2 × MIPI CSI/LVDS/sub LVDS + 1 × DVP |
||
Video output |
1 × MIPI-DSI (1080P@60Hz, 1×4lanes) Supports BT.656/BT.1120 interfaces Supports MCU/RGB LCD interfaces, with a maximum of 24 bits |
||
Audio |
1 × SAI (4T/4R), 1 × SAI (1T/1R), 1 × SAI (1T/3R), supporting I2S/TDM/PCM audio protocols PDM and Audio Codec support a sampling rate of up to 192KHz |
||
USB |
1 × USB2.0 HOST, 1 × USB3.0 DRD |
||
Other interfaces |
2 × SPI, 8 × UART, 6 × I2C, 2 × CAN, 28 × PWM, GPIOs |
Firefly team, with over 20 years of experience in product design, research and development, and production, provides you with services such as hardware, software, complete machine customization, and OEM server.