Camera AI DevicesIn Stock & Shipping

EltroNerd VisionAI-Cam 4K NPU Board

High-speed smart computer vision board with dual-core RISC-V K210 processor, MIPI camera, and YOLO object classifier.

⚡ Dual RISC-V K210 AI Processor ⚡ 5MP CSI Camera Module ⚡ 60 FPS Face Detection ⚡ WebUSB Live Stream
Unit Price$59.00
Test in Cloud IDE
// SILICON ARCHITECTURESTATUS: READY
$ eltronerd-mcu-info --device=camera-ai-devices
> Cores: Dual-Core LX7 @ 240MHz
> Memory: 512KB SRAM + 8MB PSRAM
> Flash Interface: 16MB QSPI Flash
WebUSB Bootloader v3.2100% DRIVERLESS
// DATASHEET & HARDWARE SPECS

Detailed Technical Specifications

Processor & Architecture

Vision ProcessorKendryte K210 Dual-Core 64-bit RISC-V @ 400MHz
KPU Convolution Core0.8 TOPS Hardware KPU AI Vision Accelerator
Audio APU Core8-Mic Array Hardware Audio Processor
Clock Frequency400 MHz

Camera & Video Engine

Image Sensor5 MP OV5640 / 2 MP OV2640 MIPI-CSI Camera
Frame Rate60 FPS (VGA), 30 FPS (720p / 1080p)
Night VisionDual 850nm IR LEDs + IR-Cut Filter Switch
CodecsHardware JPEG Encoder, H.264 Video Streamer

Memory & Storage

SRAM8 MB On-Chip High-Speed SRAM
Flash16 MB NOR Flash
Card SlotMicroSD Slot for Frame Capture & Recording

Interfaces & Output

Display InterfaceMCU LCD Interface (supports 2.4" / 3.5" TFT)
WirelessESP32 Wi-Fi + Bluetooth Co-Processor
USB InterfaceUSB-C WebUSB Live MJPEG Video Streamer

Power Specs

Operating Power1.2W during 60FPS YOLOv8 classification
Supply Voltage5V DC via USB-C
// KEY ADVANTAGES

Why Engineers Choose EltroNerd VisionAI-Cam 4K NPU Board

01

60 FPS Face & YOLO Object Recognition

Hardware KPU processes neural vision pipelines at 60 FPS without overloading the main CPU.

02

WebUSB Browser Live Stream

View live camera feed and detection bounding boxes directly inside the EltroNerd Cloud IDE console.

03

Dual IR LEDs for Night Vision

Automatic IR-Cut filter switching for 24/7 security monitoring and license plate capture.

// QUICK START FIRMWARE SDK

Initialization & Code Sample

main.cpp (EltroNerd VisionAI-Cam 4K NPU Board)ESP-IDF / FreeRTOS v5.2
// EltroNerd VisionAI-Cam Object Detection Sample
#include <K210_Vision.h>

CameraSensor cam;
KPU_ObjectClassifier kpu;

void setup() {
  Serial.begin(115200);
  Serial.println("[EltroNerd] VisionAI-Cam Initializing...");

  cam.begin(OV5640, RESOLUTION_VGA);
  kpu.loadModel("yolov8_tiny_quant.kmodel");
  
  Serial.println("Camera CSI Engine Active @ 60 FPS");
}

void loop() {
  Image frame = cam.capture();
  DetectionResult res = kpu.run(frame);
  
  if (res.count > 0) {
    Serial.printf("Detected %s [Confidence: %.2f%%]\n", res.label, res.confidence * 100);
  }
}

Start Building With EltroNerd VisionAI-Cam 4K NPU Board Today

Flash firmware directly from browser over WebUSB with zero driver installation.

Launch Cloud Testbed