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 and MakeCode extension, 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 Function Detail Sample Basic function Screen direction control Screen character display On-board AB buttons Recording and playback Screenshot / IoT MQTT / Color tracking Track color blocks of specific colors. Built-in several colors, and custom learning colors are also available Color blob object tracking Line tracking Track lines of specific colors. Built-in several colors, and custom learning colors are also available Visual line following Face detection Attribute, gender, smile, glasses, open mouth, number of people, Detect whether a mask is worn Face tracking robot Classification Machine learning feature classification, unlimited number of classifications Rock-paper-scissors robot Garbage classification device Code Scan QR code Barcode Book lending system Card image recognition Letter A~F Number 0~9 Traffic sign Plane, ship, bicycle, bus, car, motorcycle, train Only applicable to the supporting letter cards provided by KittenBot Object recognition Plane, ship, bicycle, bus, car, motorcycle, train, Bird, horse, cow, sheep, human, cat, dog, Bottle, chair, table, potted plant, sofa, TV / Self-training model Self-training can recognize any object, and 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. / Parameters MCU k210 Flash 16MB Overall dimension 58*40*24mm Main control supported micro:bit, Meowbit, Arduino series and any microcontroller with serial port function Display 1.3 "IPS screen (240x240) Camera RGB (640x480 ) Audio input / output 48K single channel sample rate output 8K sample rate microphone input Wireless communication WiFi (Based on Esp8285) Data output mode UART / WiFi / USB Button 2 (programmable) SD card TF Card Memory Card support ≤16G Power supply voltage 3.3~5V Programming environment KOI2 is a serial communication module that, in theory, can be utilized on any platform. Below are the programming platforms for which we have already provided sample cases.： Kittenblock MakeCode MakeCode extension: github:Kittenbot/pxt-koi2 Includes 1x KOI 2 AI Module 1x Type-C Cable 1x PH2.0 4PIN to DuPont Cable (Default) Learn & Documents KittenBot KOI 2 AI Module
