• About Us
  • News Center
  • Kendryte K210
    Overview
    Solution
    Developer Community
    AI Products
    Specifications
    Consultation
    Kendryte K210

    K210 is the most powerful edge computing chip, designs for both visual and semantic recognition, widely used in various scenarios.

    Processor Architecture

    RISC-V Dual Core 64bit, with FPU

    1 TOPS

    Face detection

    60 fps

    Power consumption

    300mW

    Buy now

    Feature 1
    High Performance High Efficiency

    With high computing power at 1TOPS, K210 consumes only 0.3W while other typical devices consume 1W. K210 comes with SRAM and offline database, which allows complete the offline data processing and storage on the device.

    Power consumption of K210
    0.3W

    Typical devices power consumption
    <1W

    Feature 2
    Widely used in various scenarios

    K210 is capable of the face detection and face recognition, image recognition and image classification, and anti-spoof.



    Processed by local device

    Developer Community
    Join Canaan Developer Community

    You may get help and support from the
    community forum.

    Learn more

    Solutions
    Smart Home

    Smart Home

    Capable of storing up to 10,000 face images and 100,000 unlock records offline, our access control can process images then convert into structuring, and protects privacy.

    Smart Industrial Park

    Smart Industrial Park

    Our access control system enables visitor verification in 100 milliseconds with no delay.

    Smart Energy Management

    Smart Energy Management

    By converting the existing meter into smart meter, our device can improve efficiency, reduce labor costs, and provide useful data for analysis and monitoring purposes.

    Specifications

    Processor:RISC-V Dual Core 64bit, with FPU
    Image Recognition :QVGA@60fps/VGA@30fps
    Voice Recognition:Microphone arrays

    Advanced encryption hardware accelerator(AES)
    Read - only memory(OTP)SHA256

    Scenario power consumption<1w
    Power consumption of Chip <300mW

    Operating system: FreeRTOS
    Network model: TinyYOLOv2(after pruned)
    Deep learning framework: TensorFlow/Keras/Darknet
    Peripherals: FPIOA、UART、GPIO、SPI etc

    流量代码 其他js代码

    Cooperation Consultation

    Business Cooperation and Product Consultation

    Submit

    Welcome to contact us E-mail:sales@canaan.io

    Welcome to contact us E-mail:salesAI@canaan-creative.com