01KOI2
Introduction
Product Details
KOI2 is an AI camera module based on the K210 chip from KittenBot Technology, which aims to combine advanced AI technology and high-performance embedded systems to provide strong support for visual recognition and intelligent AI applications. It comes with hardware such as a screen, camera, speaker, microphone, and buttons, which can fully unleash the capabilities of AI applications. It has a front and rear flip camera, a plastic case with Lego structural holes, providing a richer and more convenient building form, meeting the classroom and project needs of educational scenarios. We provide friendly Kittenblock plug-ins and Makecode plug-ins, which combine KOI2 with the main control board to use graphical programming to explore the new world of AI.
Features
- Complete AI functionality to meet the needs of AI projects
- Flippable camera for greater flexibility
- Seamless integration of graphical and code programming, suitable for both beginners and makers
- Ability to independently train machine models
Applications
- Training institutions carrying out artificial intelligence teaching
- School artificial intelligence project teaching and project demonstration
- Personal AI works creation
Parameter Details
| Dimensions | 5.5 x 4 x 2.5 cm |
|---|---|
| Net Weight | 27.5g |
| Processor | K210 |
| Interface | Type-C |
| Operating Voltage | 5v |
| Output Current | 200mA |
| Onboard Resources | 640x480 RGB Camera, Audio Input / Output, SD Card Slot, 1.3" 240x240 IPS Screen Display, Programming 2 Buttons, 1 Reset Button, WIFI (ESP32 C3) |
| Programming Platform | Makecode, Kittenblock, Arduino IDE |
| Communication Method | UART/USB (Instruction Communication) |
| Supported Hosts | Micro:bit, Future Board, Future Board Lite, Arduino, Raspberry Pi |
Interface and Structure Guidelines
Function List
| Function | Detail | Case |
|---|---|---|
| Screen direction control | Basic function | |
| Screen character display | Basic function | |
| On-board AB buttons | Basic function | |
| Recording and playback | Basic function | |
| Photo save | Basic function | |
| MQTT | Internet of Things IOT | Remote data upload |
| Color block tracking | Track color blocks of specific colors. Built-in several colors, and custom learning colors are also available | Color block object handling |
| Line tracking | Track lines of specific colors. Built-in several colors, and custom learning colors are also available | Line tracking car |
| Face detection | Attribute, gender, smile, glasses, open mouth, number of people | Face tracking cloud platform |
| Mask recognition | Detect whether a mask is worn | Mask protection detection |
| Feature classification | Machine learning feature classification, unlimited number of classifications | Rock-paper-scissors robot |
| Garbage classification device | Specific object recognition | |
| QR code | Common QR code recognition | Book lending system |
| Barcode | Common barcode recognition | Unmanned supermarket scanning |
| Letter recognition A~F | Only applicable to the supporting letter cards provided by KittenBot | |
| Number recognition 0~9 | Only applicable to the supporting number cards provided by KittenBot | Intelligent arithmetic |
| Traffic sign recognition | Only applicable to the supporting traffic sign cards provided by KittenBot | AI traffic competition |
| Common object recognition | 20 common objects. | |
| Plane, ship, bicycle, bus, car, motorcycle, train | ||
| Bird, horse, cow, sheep, human, cat, dog | ||
| Bottle, chair, table, potted plant, sofa, TV | Object recognition experience | |
| Self-training model | Self-training models are used for letter, number, traffic sign, and common object recognition, which is different from feature classification and can recognize the position and size of objects. It requires self-photographing, calibration, and training, which is suitable for advanced players. | Self-training specific object tracking |
Differences between KOI2 and KOI1 Appearance
Startup screen:
KOI2
KOI
- USB port KOI2's USB port is a Type-C interface KOI's USB interface is a Micro USB interface
KOI2 vs. KOI Functionality Comparison
Since its launch, KittenBot's KOI artificial intelligence module has been widely loved by students and teachers for its powerful functionality and educational design. In 2024, KittenBot upgraded the KOI firmware and optimized its hardware, renaming it KOI2. To meet the needs of certain competitions and creative projects, the new firmware adds mask recognition, letter recognition, number recognition, traffic sign recognition, general recognition of 20 objects, and self-training models. (Due to limited internal space, functions with poor performance have been removed.) KOI's overall frame rate has been increased, and its overall framework has been optimized, with more reasonable communication and memory allocation.
| KOI | KOI2 |
|---|---|
| Recording and playback | Recording and playback (consistent) |
| Screen orientation control | Screen orientation control (consistent) |
| Onboard AB buttons | Onboard AB buttons (consistent) |
| Screen character display | Screen character display (consistent) |
| Photo storage | Photo storage (consistent) |
| MQTT | MQTT (consistent) |
| Color block tracking | Color block tracking (frame rate optimization, built-in colors added) |
| Line tracking | Line tracking (frame rate optimization) |
| Face detection | Face detection (4 new attributes added: gender, smile, glasses, open mouth, number of people) |
| Feature classification (40 image limit) | Feature classification (optimized, no limit on number of categories) |
| Apriltag | Apriltag (recognition optimization) |
| QR code | QR code (recognition optimization) |
| Barcode | Barcode (recognition optimization) |
| Geometric shape tracking | Geometric shape tracking (removed) |
| Cloud speech synthesis | Speech synthesis (removed) |
| Cloud face recognition | Cloud face recognition (removed) |
| Speech recognition (tone) | Speech recognition (removed) |
| Mask recognition (new) | |
| Letter recognition A~F (new) | |
| Number recognition 0~9 (new) | |
| Traffic sign recognition (new) | |
| General recognition of 20 objects (new) | |
| Self-training model (new) |