
Sipeed MaixCAM RISCV AI Camera
Processor: SOPHGO SG2002
USD · Volume pricing available
Санаттар:
Certifications
Global Shipping 7-14 Days
DHL, FedEx, UPS, Air & Sea
Genuine Products
Authorized distribution
Export License Ready
ECCN, CO, CI, PL, EUS docs
24/7 Tech Support
Pre & post sales
Export Compliance
EAR99AI confidence 85% · NLR — No license required. Standard end-use statement recommended.
Buyer Tier Benefits
Tier discounts stack with volume pricing. Automatically applied after your first order.
Buy the Sipeed MaixCAM to quickly launch AI projects on edge devices
This is a powerful camera with artificial intelligence based on a RISC-V processor and a built-in neural accelerator (NPU). It allows you to run neural networks directly on board without connecting to cloud servers. The device is ideal for recognizing objects, faces, gestures, as well as audio analytics and AIoT tasks. MaixCAM is a turnkey solution for prototyping and mass production.
- Built-in 1 TOPS NPU to run YOLOv5, YOLOv8 and Mobilenetv2.
- 1 GHz RISC-V C906 hybrid processor with optional ARM A53.
- Camera with a resolution of up to 5 MP and a 2.3-inch IPS touch display.
- Built-in WiFi 6 and Bluetooth 5.4 modules for wireless data transfer.
- MaixPy ecosystem and online platform for training models without code.
How the Sipeed MaixCAM vision camera works
The device is based on the SOPHGO SG2002 SoC with three cores: a main 1 GHz RISC-V (or ARM A53), a small 700 MHz RISC-V for RTOS and a low-power 8051. The built-in 1 TOPS NPU accelerates the execution of neural networks. The camera captures video and transmits it to the NPU for real-time analysis.
Algorithms recognize objects, faces, gestures and other visual patterns. The results are displayed on the touch display or transmitted via WiFi and Bluetooth. Development is carried out in Python through the MaixPy package with hardware acceleration. For quick deployment, the MaixHub online platform with pre-trained models is available.
Technical characteristics and functionality
The SOPHGO SG2002 SoC includes 1 GHz RISC-V C906 (or ARM A53), 700 MHz RISC-V C906 and 25-300 MHz 8051. The neural processor has 1 TOPS INT8 performance and supports BF16. Models Mobilenetv2, YOLOv5, YOLOv8 and others are supported. The amount of RAM is 256 MB DDR3. The camera supports sensors up to 5 MP, and is officially compatible with the 4 MP GC4653 and OS04A10.
The display has a diagonal of 2.3 inches, a resolution of 552×368 pixels and touch control. The video codec supports H.264/H.265/MJPEG with 2K@30fps recording. Built-in WiFi 6, BLE 5.4, USB Type-C, 14-pin GPIO and a microSD slot. Audio output includes a built-in speaker amplifier up to 1W and a built-in microphone. Power is provided via USB-C.
About the manufacturer Sipeed
Sipeed was founded in 2015 in Shenzhen. It is a national high-tech enterprise specializing in RISC-V and AI solutions for edge devices. Sipeed develops open hardware platforms, AI cameras, modules and systems for the Internet of Things. The company's products are used in education, industry and scientific research. Sipeed is actively developing the MaixPy ecosystem and providing tools for rapid AI implementation.
Choose the Sipeed MaixCAM camera in the COPTERPARTS store. Your project will gain powerful computer vision without complex code. Order a ready-made AI solution for your tasks today.
- Processor: SOPHGO SG2002
- Main Core: 1 GHz RISC-V C906 or Arm Cortex-A53 (runs Linux)
- Secondary core: 700 MHz (runs RTOS)
- NPU performance: 1 TOPS
- RAM: 256 MB DDR3
- Storage: TF card slot
- Display: 2.3-inch IPS touch screen, resolution 552x368 pixels
- Camera interface: MIPI CSI on 4 lines
- Supported cameras: GC4653 (4 MP), OS04A10 (4 MP, color night photography and HDR)
- Video codec: hardware H.264, H.265, MJPEG
- Maximum video resolution: up to 2K (2880×1620) at 30 fps
- Network: Wi-Fi 6
- Wireless interfaces: Bluetooth 5.4
- Ports: USB 2.0 Type-C
- Optional: 2 GPIO connectors for expansion
- Optional: IAP global shutter, 256x192 thermal imager, Ethernet