Skip to main content

CanMV K230

What is CanMV?

intro1

The CanMV open source project is officially created and maintained by Canaan. It is an open source project that ported MicroPython to Canaan's edge computing AI chip K210/K230 (a 64-bit dual-core RISC-V with hardware FPU and convolution accelerator). The project aims to create a low-cost, scalable, Python-driven AI artificial intelligence module.

In other words, if you are familiar with Micropython or Python, you can complete a large number of machine vision, machine hearing and deep learning applications through the CanMV project in a very short time.

CanMV K230 Development Board

01Studio CanMV K230 AI development board is based on Canaan Technology's edge computing chip K230 (RSIC-V architecture, 64-bit dual-core) and the CanMV open source project.

intro2

Hardware

intro3

Product Parameters

Product Parameters
K230● CPU1: RISC-V , 1.6GHz, Support RVV 1.0
● CPU0: RISC-V , 800MHz
Neural NetworkKPU(6TOPS Equivalence),Support INT8 & INT16
RAM● 1GBytes(LPDDR4)
● 2GBytes(LPDDR4)
StorageMicroSD(512G max)
Camera● GC2093(Standard) / OV5647
● Support 3 interfaces(3 x 2 lane CSI)
Display● MIPI LCD(1x4 lane DSI), up to 1920x1080
● HDMI Display, up to1920x1080
● CanMV IDE, up to 1920x1080
Network● 2.4G WiFi(onboard antenna)
● Ethernet(An external USB to Ethernet cable is required)
Audio Output3.5mm Audio port (dual channel)
Audio InputMicrophone
USBx1 USB 2.0 HOST
KEYx2 (Programmable Key,Reset Key)
LEDx2 (Programmable LED,Power LED)
GPIO40Pin GPIO Pin Header(compatible with Raspberry Pi)
UART/I2CXH-1.25mm-4P
UART Debug● CPU1(UART3)
● CPU0(UART0)
TYPE-CIDE connection development,code debugging,filetransferpowersupplyall-in-one
Power Supply5V @ 1A
Appearance Parameters
Size85 x 56 x 15mm (PCB Size)
Weight31g

GPIO PinOut

intro3

Size

intro

K230 Chip Parameters

intro

K230 Chip Parameters
CPU● CPU1: RISC-V , 1.6GHz, 32KB I-cache, 32KB D-cache, 256KB L2 Cache, 128bit RVV 1.0
● CPU0: RISC-V , 800MHz, 32KB I-cache, 32KB D-cache, 128KB L2 Cache
KPU6TOPS Equivalence,Support INT8 & INT16
Typical network performance:
Resnet50 ≥ 85fps @ INT8;Mobilenet_v2 ≥ 670fps @ INT8;YOLO V5s ≥ 38fps @ INT8
DPU3D structured light depth engine, supports up to 1920x1080
VPUH.264 and H.265 video codec, suports up to 4096x4096Encoder
Encoder performance:4K@20fps
Decoder performance:4K@40fps
JEPG codec:supports up to 8K(8192x8192)
Image InputSupport 3 Interfaces MIPI CSI input:1x4 lane+1x2 lane or 3x2 lane
Display Output1 x MIPI DSI (1x4lane or 1x2lane), up to 1920x1080
Peripheral● 5 x UART
● 5 x I2C
● 1 x I2S
● 6 x PWM
● 64 x GPIO + 8 x PMU GPIO
● 2 x USB 2.0 OTG
● 2 x SDxC: SD3.0, EMMC 5.0
● 3 x SPI: 1 x OSPI + 2 x QSPI
● Timer / RTC / WDT