Sipeed Maix II M2 Dock AI Linux Vision Dev Board w/ 1.3-Inch IPS Screen

Sipeed provides the MaixPy3 SDK, a Python 3 environment optimized for the board's hardware, ideal for rapid AI application development. For lower-level control, the libmaix C/C++ SDK offers direct access to the NPU and peripherals. You develop on a PC, cross-compile your code, and transfer it to the board via SD card or network. The board runs a lightweight Linux distribution, allowing you to also install standard Python packages via pip.

The dock has a dedicated connector for the compatible SP2305 camera module. Physically attach the camera ribbon cable, ensuring proper orientation. In your MaixPy3 Python code, you initialize the camera sensor using the `sensor` module, set the resolution and frame rate, and capture frames. These image buffers can be passed directly to the loaded AI model (converted from ONNX or trained via MaixHub) for inference using the board's hardware NPU accelerator, returning detection or classification results.

Yes, the board's design separates the Allwinner V831 core module from the extension dock. You can detach the core module, which contains the SoC, memory, and essential power circuitry. You would then design a custom carrier PCB that provides the necessary 3.3V power supply via the module's pin headers and routes the required interfaces like MIPI-CSI for camera, RGB for display, I2C, SPI, and GPIOs to your peripherals. This is intended for productization and space-constrained embedded vision applications.

The board is powered via the onboard USB Type-C port. A standard 5V/2A USB power adapter or a capable USB port on a computer is sufficient. The Allwinner V831 SoC and its peripherals operate at low voltage, with on-board regulators providing the necessary core voltages. For battery-powered applications, you can provide a stable 5V input to the Type-C port from a battery management circuit. Monitor total system current, especially when the screen, camera, and NPU are active simultaneously.