Programmable RISC-V-based AI Reference Camera Supports MicroPython, OpenMV
July 18, 2019
The 28 nm Kendryte K210 SoC at the heart of the camera leverages dual 64-bit RISC-V CPU cores running at 400 MHz, while the OmniVision OV7740 image sensor leverages OmniPixel3-HS technology.
M5STACK has released the M5StickV K210 RISC-V AI camera, a programmable system that simplifies the integration of machine learning, image, and speech recognition through support for OpenMV and MicroPython programming. The 28 nm Kendryte K210 SoC at the heart of the camera leverages dual 64-bit RISC-V CPU cores running at 400 MHz, while the OmniVision OV7740 image sensor leverages OmniPixel3-HS technology.
In addition to the RISC-V CPUs, the M5StickV’s Kendryte K210 SoC includes dual-independent double precision floating point units (FPUs), a neural network processor for convolutional neural network (CNN) acceleration, an embedded audio processing unit (APU) with 512-point FFTs, and a field-programmable I/O array.
Other features of the K210 SoC include:
- 8 MiB on-chip SRAM
- DMA controller
- SPI, I2C, UART, I2S, RTC, PWM, and Timer support
- AES, SHA256 accelerator
The M5StickV camera module itself includes a MEMS microphone, speaker with built-in I2S Class-D DAC, IPS screen, 6-axis IMU, 200 mAH Li-po battery, optional wireless communication, and an enclosure.
The reference system is available without wireless connectivity for $26.50. For more information visit m5stack.com/products/stickv.