Yahboom K230 Vision Recognition Module AI Intelligent Camera, 6TOPS, 1GB LPDDR4, 2MP GC2093, 2.4in Touch
Yahboom K230 Vision Recognition Module AI Intelligent Camera, 6TOPS, 1GB LPDDR4, 2MP GC2093, 2.4in Touch
Yahboom
āĻĒāĻŋāĻāĻāĻĒā§āĻ° āĻāĻĒāĻ˛āĻā§āĻ¯āĻ¤āĻž āĻ˛ā§āĻĄ āĻāĻ°āĻž āĻ¯āĻžāĻ¯āĻŧāĻ¨āĻŋ
Overview
The Yahboom K230 vision recognition module (AI intelligent camera) is a high-performance, low-power embedded AI vision processing module for edge computing scenarios. It is based on a RISC-V architecture AloT chip platform and uses a Kendryte K230 with 6TOPS equivalent computing power and 1GB RAM, supporting AI vision tasks such as real-time image recognition, target detection, face recognition, and behavior analysis.
It integrates a 2MP camera (GC2093), 2.4-inch HD touch screen, microphone, RGB light, and WiFi network card, and supports TF cards (a 32G TF card is provided). It can connect to multiple main controllers (STM32, MSPM0, ESP32, Raspberry Pi, AR, micro:bit) via serial communication. Optional mounting brackets and an optional 2DOF electric PTZ enable installation on robots or smart cars.
Key Features
- Performance note: performance can reach 13.7 times that of K210.
- K230 AI computing platform: AloT chip based on RISC-V architecture; third-generation KPU; high precision; fast startup; ultra-low power consumption; developed based on RT-Smart real-time operating system.
- 1GB LPDDR4 large memory.
- Professional hardware configuration: 2.4-inch HD touch screen; ABS case stated as resistant to drop and high temperature; HD high frame rate camera; cooling fan.
- Pre-installed GUI program (based on RTSmart system) with 30+ functions; each function provides detailed descriptions.
- GUI includes: Settings, AI Face, AI Body, AI Scene Applications, LLM Large Model, Camera, Code Recognition, Color Recognition, File Manager, Gallery, Graphics Detection, Hardware Interface Test, Drawing.
- Windows host computer software: supports RTSP real-time video stream monitoring and includes video/image processing tools, data generators, and text tools (Windows only).
- Real-time image transmission: USB data cable / WiFi wireless image transmission.
- Audio: built-in microphone supports audio input and recording; supports external speakers (3525 speaker, GH1.25 2Pin).
- AI visual output result: supports output of target objects and position information through serial port.
- Video tutorials with English subtitles (principle explanations, code explanations, practical demonstrations, experimental phenomena).
For product selection, setup questions, or after-sales help, contact https://rcdrone.top/ or email support@rcdrone.top.
Specifications
| Module | K230 visual recognition module / Yahboom K230 vision module |
| Main chip | Kendryte K230 |
| Architecture | RISC-V (AloT chip platform) |
| KPU / AI compute | 6TOPS equivalent computing power |
| RAM | 1GB (1GB LPDDR4 stated) |
| Main frequency | Highest 1.6GHz |
| Camera | 2MP (GC2093) |
| Display | 2.4-inch HD touch screen |
| Wireless | WiFi network card |
| Storage | TF card support (32G TF card provided) |
| Audio | Microphone; supports external speakers (3525 speaker, GH1.25 2Pin) |
| AI visual processing frame rate | About 85 frames |
| AI visual recognition cases | 40+ AI visual recognition case (table comparison statement) |
| Custom training model | Support |
| System / program | RT-Smart real-time operating system; factory pre-installed GUI program |
| Communication (to main controller) | Serial communication; compatible with MSPM0 / ESP32 / STM32 / Raspberry Pi / AR / micro:bit |
AI Visual Recognition Applications (Examples)
- Image detection: line segment detection, rectangle detection, circle detection, object edge detection
- Color recognition: single color recognition, object counting, multi-color recognition, intelligent line tracking
- Code recognition: barcode recognition, QR code recognition, AprilTag tag recognition, DM code recognition
- Face recognition: face detection, face key point recognition, face orientation detection, face 3D network, gaze direction detection, registered face recognition
- Human feature recognition: human detection, human key point detection, fall detection, palm detection, palm key point classification, gesture recognition, dynamic gesture recognition, rock-paper-scissors game
- More AI visual recognition: OCR character recognition, yolov8n target detection, yolov8n segmentation, license plate recognition, target tracking, autonomous classification learning, garbage classification, road sign recognition
Model Training & Local Deployment (Process Outline)
- Start
- Use K230 to collect image data
- Import the image dataset into the Canaan online platform
- Label the target objects in the dataset images
- Train the model online
- Export the model and code
- Run code, load model, and identify the target object
- End
The training workflow description also states that the platform can provide models in kmodel format and sample code to help complete deployment.
Optional Mounting (As Stated)
- Fixed bracket: suitable for fixed viewing angle, or installation on a PTZ or robot car with angle adjustment function.
- Angle adjustable bracket: flexibly adjust the pitch angle; suitable for installation on most robot cars.
- Heightened angle adjustable bracket: flexibly adjust the pitch angle or module height (1.2~9.5cm); suitable for scenarios requiring a high viewing angle.
- 2DOF electric PTZ: for flexibly tracking dynamic objects; suitable for robots, cars, desktops, etc.
- The above brackets are equipped with a cooling fan.
What's Included
- 32G TF card (provided)
Manuals / Downloads
- Yahboom K230 vision module tutorial
- Kendryte official resource page (K230)
- GitHub: kendryte/k230_rtos_sdk (boards) (includes a k230_canmv_yahboom folder as shown)
Details
A compact edgeâAI camera module built around the Kendryte K230, combining a 2MP camera, 2.4âinch touch display, and onboard connectivity for fast prototyping.
Key platform highlights include 6TOPSâclass AI performance, 1GB LPDDR4 memory, and a ready-to-use software ecosystem for vision projects.
Mounting options range from fixed and angleâadjustable brackets to a 2DOF electric PTZ for robotics, smart cars, and pan/tilt vision demos.
A quick comparison helps evaluate K230 performance and typical vision throughput versus other popular vision modules.
Development support spans common embedded AI and vision workflows, with guidance for code/firmware resources and model deployment.
Built-in demos cover practical tasks like object detection, color recognition, barcode/QR scanning, and human/gesture recognition.
A structured workflow supports collecting datasets, training models, and deploying them locally for on-device inference.
An LLM integration interface enables API-based voice/text interaction features when connected to supported online services.
The pre-installed GUI groups AI features, camera tools, and device utilities into a touchscreen-friendly menu.
Source code and tutorials are provided to help adapt the built-in demos into your own applications.
Windows software adds RTSP monitoring plus video/image utilities for development, testing, and stream review.
English-subtitled tutorials and onboard audio input support make it easier to learn features and build interactive projects.
External speaker support lets recorded audio be played back through a simple GH1.25 2âpin connection.
Stream camera output to a PC over USB or WiâFi for monitoring, debugging, and rapid iteration.
The 2.4âinch capacitive touch screen supports interactive demos such as touch tracking, drawing, and snapshot capture.
Built for AIoT networking, the module supports WiâFi/AP modes and RTSP streaming for remote monitoring and interaction.
A full hardware summary includes onboard WiâFi, microphone, TF card support, and expansion interfaces for controller integration.
Support for common platforms like Raspberry Pi, Jetson, STM32, ESP32, Pico, and micro:bit helps integrate the module into a wide range of projects.
MicroPython support and the CanMV development environment provide an accessible workflow for coding and testing vision projects on the K230 module.
The upgraded cv_lite library lists higher processing frame rates than OpenMV across several 480x640 image recognition tasks.
The Yahboom K230 vision recognition camera module supports common AI tasks like road sign and license plate recognition, face tracking, and QR/tag following.
The K230 camera moduleâs labeled connectors include a TypeâC interface, TF card slot, touch screen interface, and GPIO for cleaner wiring and expansion.
The K230 vision recognition module combines an onboard camera, USB port, and a 12-pin GPIO header with labeled pin mapping for easy wiring.
The Kendryte K230 vision module combines a 2.4-inch capacitive touch screen, about 6TOPS AI compute, WiFi, and a serial/GPIO interface for flexible integration.
The K210 vision recognition module pairs a 2.0-inch capacitive touch screen with a serial port interface in a compact 73.4Ã21.1Ã7.4 mm form factor.
The ESP32-S3 WiFi camera module supports up to 30fps recognition with WiFi hotspot mode plus serial and I2C interfaces for easy integration.
The Yahboom K230 study page provides a tutorial link and course outline covering environment setup, interfaces, and example programs.
The Yahboom K230 vision recognition module combines a 2.4-inch touchscreen with an AI camera and supports a wide range of recognition and tracking functions.
The K230 vision module comes with organized tutorial and code folders covering image processing, face recognition, and basic networking.
Yahboom K230 vision module resources include model training materials, example code with comments, and downloadable 3D model files for integration.
The K230 visual recognition module includes fixed and adjustable angle bracket options with a 150° adjustable range (15° up, 135° down) for flexible mounting.
The K230 vision camera mounts on a 2DOF electric PTZ bracket with up to 150° tilt and 1.2â9.5 cm height adjustment for flexible aiming.
The K230 vision recognition module pairs a 2.4-inch capacitive touch display with 1GB LPDDR4 memory and a 6TOPS KPU for AI workloads.
The kit includes the core K230 vision camera module with a 2.4-inch touch display, plus cables and multiple optional bracket and PTZ mounting accessories.
The Yahboom K230 vision recognition module connects via a USB cable and includes a compact touchscreen enclosure for easy setup.
