MaixPy - MaixPy

MaixPy Banner

</div>
<h1><span>MaixPy (v4)</span></h1>
<h3>极速落地 AI 视觉、听觉应用</h3>

快速开始 🚀📖 API 参考 📚 硬件平台：MaixCAM/MaixCAM2 📷 开源代码 ⭐️ 应用商店 📦

English | 中文

MaixCAM 系列

喜欢 MaixPy 请给 MaixPy 开源项目点个 Star ⭐️ 以鼓励我们开发更多功能。

简易的 API 设计， 10 行代码进行 AI 图像识别

from maix import camera, display, image, nn

classifier = nn.Classifier(model="/root/models/mobilenetv2.mud")
cam = camera.Camera(classifier.input_width(), classifier.input_height(), classifier.input_format())
disp = display.Display()

while 1:
    img = cam.read()
    res = classifier.classify(img)
    max_idx, max_prob = res[0]
    msg = f"{max_prob:5.2f}: {classifier.labels[max_idx]}"
    img.draw_string(10, 10, msg, image.COLOR_RED)
    disp.show(img)

硬件外设控制，不在话下

串口收发：

from maix import uart

devices = uart.list_devices()

serial = uart.UART(devices[0], 115200)
serial.write_str("hello world")
print("received:", serial.read(timeout = 2000))

I2C 收发：

from maix import i2c

devices = i2c.list_devices()
dev1 = i2c.I2C(devices[0], freq=100000)
slaves = dev1.scan()
print("find slaves:", slaves)
dev1.writeto(0x12, b'hello')
print("received:", dev1.readfrom(0x12, 5))

便捷的 MaixVision 工作站

简化开发环境，让开发更简单快速

在线 AI 训练平台 MaixHub

无需 AI 基础和昂贵的训练设备，一键训练模型，一键部署到 MaixCAM

MaixHub

性能强劲 MaixCAM 硬件平台

目前有MaixCAM / MaixCAM-Pro 和 MaixCAM2 两代硬件产品，从性能、配件、外观提供不同选择

后文有性能对比

MaixCAM2

        <div class="mask"></div>
    </a>
</div>
<div class="flex flex-row w-full justify-between">
    <div class="flex_center flex-row justify-start w-1/2">
        <a href="https://wiki.sipeed.com/hardware/zh/maixcam/maixcam.html" target="_blank" class="flex_center card_item mask_wrapper item2">

            <div class="item_name pt-8">MaixCAM</div>
            <div class="mask"></div>
        </a>
    </div>
    <div class="flex_center flex-row justify-end w-1/2">
        <a href="https://wiki.sipeed.com/maixcam-pro" target="_blank" class="flex_center card_item mask_wrapper item3">

            <div class="item_name pt-8">MaixCAM-Pro</div>
            <div class="mask"></div>
        </a>
    </div>
</div>

部分功能展示

以下为部分功能简介，更多到社区找到更多

基于 MaixPy 提供的丰富 API 可以创造出更多新功能

(LLM/VLM 相关只 MaixCAM2 支持)

OpenCV + OpenMV

        <p class="feature">OpenCV + OpenMV</p>
        <p class="description">支持 OpenCV， 兼容 OpenMV</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">

MaixCDK

        <p class="feature">C++版本</p>
        <p class="description"><a href="https://github.com/sipeed/MaixCDK">MaixCDK</a> C++版本的SDK，与MaixPy的API相同, 商业友好</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">

串口模块

        <p class="feature">作为串口模块</p>
        <p class="description">其它 MCU 通过串口命令控制</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/app_store.mp4"></video>
        <p class="feature">APP商店</p>
        <p class="description">将您的APP分享给社区，并一键安装<a href="https://maixhub.com/app">APPs</a>。</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/classifier.mp4"></video>
        <p class="feature">AI 分类</p>
        <p class="description">识别物体类别</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/detector.mp4"></video>
        <p class="feature">AI 对象检测</p>
        <p class="description">识别物体类别和坐标</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/face_recognition.mp4"></video>
        <p class="feature">AI 人脸识别</p>
        <p class="description">识别不同人脸特征，表情识别等</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/maixcam_face_landmarks.mp4"></video>
        <p class="feature">AI 人脸关键点</p>
        <p class="description">检测人脸关键点，面部特征/动作识别，AI 换脸</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/keypoints.mp4"></video>
        <p class="feature">AI 人体关键点检测</p>
        <p class="description">姿态识别、体感游戏</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/hands_landmarks.mp4"></video>
        <p class="feature">AI 手部关键点</p>
        <p class="description">检测手部关键点，手势识别</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/self_learn_classifier.mp4"></video>
        <p class="feature">AI 自学习分类器</p>
        <p class="description">无需在PC上训练，在设备上瞬间学习任意物体</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/self_learn_tracker.mp4"></video>
        <p class="feature">AI 自学习检测器</p>
        <p class="description">无需在PC上训练，在设备上瞬间学习任意物体</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/tracker.mp4"></video>
        <p class="feature">AI 物体轨迹跟踪</p>
        <p class="description">轨迹追踪，流量统计</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/streaming.mp4"></video>
        <p class="feature">AI 监控，串流</p>
        <p class="description">安防监控，可串流，甚至可以向直播平台 比如 Bilibili.com 直播</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/find_blobs.mp4"></video>
        <p class="feature">查找颜色</p>
        <p class="description">查找颜色斑点</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/line_track.mp4"></video>
        <p class="feature">巡线</p>
        <p class="description">小车巡线，物流搬运</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/qr_apriltag.mp4"></video>
        <p class="feature">QR码和AprilTag</p>
        <p class="description">识别QR码和AprilTag</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/ocr.mp4"></video>
        <p class="feature">OCR</p>
        <p class="description">识别图片中的字符，旧物数字化</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/depth_anything_v2.mp4"></video>
        <p class="feature">单目深度估计</p>
        <p class="description">单目摄像头估计深度</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/asr.mp4"></video>
        <p class="feature">语音识别</p>
        <p class="description">实时连续语音识别</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/tts.mp4"></video>
        <p class="feature">语音合成</p>
        <p class="description">TTS生成语音，支持多种语言</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/llm.mp4"></video>
        <p class="feature">离线大语言模型LLM</p>
        <p class="description">玩全离线跑大语言模型LLM</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/vlm.mp4"></video>
        <p class="feature">离线视觉大语言模型VLM</p>
        <p class="description">玩全离线跑视觉大语言模型VLM</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/desktop_monitor.mp4"></video>
        <p class="feature">桌面监视器</p>
        <p class="description">监视PC信息，如CPU，内存，网络等。</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">

天气站

        <p class="feature">天气站</p>
        <p class="description">监视天气信息，如温度，湿度等。</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">

热红外摄像头

        <p class="feature">热红外摄像头</p>
        <p class="description">选配摄像头，温度图像获取/测量</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">

HDMI 捕获视频

        <p class="feature">HDMI 捕获视频</p>
        <p class="description">选配，通过 HDMI 捕获图像，作为服务器监控（KVM）和远程控制、外接 AI、推流设备等</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/video_play.mp4"></video>
        <p class="feature">大屏视频播放</p>
        <p class="description">多种规格屏幕选择(2.3" 2.4" 5" 7"等), 硬件解码支持</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/microscope.mp4"></video>
        <p class="feature">显微镜</p>
        <p class="description">搭配1/8"大底传感器 + 显微镜头 = 数字显微镜</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">

高速识别

        <p class="feature">高速识别</p>
        <p class="description">搭配全局摄像头，高速运动物体也能准确识别</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/time_lapse.mp4"></video>
        <p class="feature">延时摄影</p>
        <p class="description">搭配1/8"大底传感器实现全天候延时摄影</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/sky.mp4"></video>
        <p class="feature">天文摄影</p>
        <p class="description">搭配1/8"大底传感器+高倍镜头实现天文摄影，支持长曝光模式和RAW 图输出</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/gyroflow.mp4"></video>
        <p class="feature">陀螺仪增稳</p>
        <p class="description">板载陀螺仪，支持导出 gyroflow 防抖格式，DIY 摄影</p>
    </div>
    <div>
    </div>
</div>
<div class="feature_item">
    <div class="img_video">
        <video playsinline controls autoplay loop muted preload src="/static/video/tof.mp4"></video>
        <p class="feature">TOF 配件深度测量</p>
        <p class="description">搭配 TOF 模块实现精准深度测量</p>
    </div>
    <div>
    </div>
</div>

谁在用 MaixPy？

AI 算法工程师：轻松将你的 AI 模型部署到嵌入式设备。

易用的 API 访问 NPU，开源量化工具，详细的 AI 模型的文档。

STEM：想要教学生 AI 和嵌入式开发的老师。

易用的 API，PC 工具，在线 AI 训练服务等，让你专注于教授 AI，而不是硬件和复杂的软件开发。

创客: 想要制作一些酷炫的项目，但不想把事件浪费在太复杂的硬件和软件。

Python 和 C++ 丰富简易 API，快速上手，甚至可以在几分钟内完成你的 DIY 项目。

工程师: 想要做一些项目，但希望尽快有原型和落地。

Python 和 C++ 丰富 API，高效稳定易使用，助力快速出原型及直接落地项目。

学生：想要学习 AI，嵌入式开发。

提供丰富文档和教程和开源代码，帮助你找到学习路线，并逐步成长。从简单的 Python 编程到视觉，AI，音频，Linux，RTOS等。

企业: 想要开发 AI 视觉产品，但没有时间或工程师来开发复杂的嵌入式系统。

使用 MaixPy 甚至图形编程来，用更少的员工和时间。例如，向生产线添加 AI QA 系统，或为办公室添加一个 AI 安全监控器。

竞赛者: 想要赢得比赛的比赛人们。

MaixPy 集成了许多功能，易于使用，加快你作品产出速度，助力有限时间内赢得比赛，国内常见比赛都有很多同学使用 MaixPy 赢得比赛。

性能对比

这里列出比较重要的软硬件性能参数供选型参考。

特征	Maix-I K210	MaixCAM	MaixCAM2
CPU	400MHz RISC-V x2	1GHz RISC-V(Linux) 700MHz RISC-V(RTOS) 25~300MHz 8051(Low Power)	1.2GHz A53 x2(Linux) RISC-V 32bit E907(RTT)
内存	6MB SRAM	256MB DDR3	1GB / 4GB LPDDR4
NPU	0.25Tops@INT8 official says 1T but...	1Tops@INT8	3.2Tops@INT8 12.8Tops@INT4
Encoder	❌	2880x1620@30fps H.254/H.265/JPEG	3840*2160@30fps H.254/H.265/JPEG
Decoder	❌	2880x1620@30fps H.264/JPEG	1080p@60fps H.264/JPEG
屏幕	2.4" 320x240	2.3" 552x368(MaixCAM) 2.4" 640x480(MaixCAM-Pro) 5" 1280x720 7" 1280x800 10“ 1280x800	2.4" 640x480
触摸屏	❌	2.3" 552x368/2.4" 640x480	2.4" 640x480
摄像头	30W	500W(5M)	800W(8M)
AI ISP	❌	❌	✅
WiFi	2.4G	WiFi6 2.4G/5G	WiFi6 2.4G/5G
BLE	❌	BLE5.4	BLE5.4
USB	❌	USB2.0	USB2.0
以太网	❌	100M(选配)	100M(板载FPC, 可外接转RJ45模块)
SD 卡接口	SPI	SDIO	SDIO
操作系统	RTOS	Linux(BuildRoot) + RTOS	Linux(Ubuntu) + RTT
编程语言	C / C++ / MicroPython	C / C++ / Python3	C / C++ / Python3
软件包	MaixPy-v1	MaixCDK + MaixPy v4 + opencv + numpy + ... 纯Python包或者手动交叉编译	MaixCDK + MaixPy v4 + opencv + numpy + scipy + ... 大量AArch64预编译包直接安装，支持板上编译和交叉编译
PC 软件	MaixPy IDE	MaixVision Workstation	MaixVision Workstation
文档	⭐️⭐️⭐️⭐️	🌟🌟🌟🌟🌟	🌟🌟🌟🌟🌟
在线 AI 训练	⭐️⭐️⭐️	🌟🌟🌟🌟🌟	🌟🌟🌟🌟🌟
官方应用	⭐️	🌟🌟🌟🌟🌟	🌟🌟🌟🌟🌟
易用性	⭐️⭐️⭐️⭐️	🌟🌟🌟🌟🌟	🌟🌟🌟🌟🌟
AI 分类(224x224)	MobileNetv1 50fps MobileNetv2 ❌ Resnet ❌	MobileNetv2 130fps Resnet18 62fps Resnet50 28fps	MobileNetv2 1218fps Resnet50 200fps
AI 检测仅推理部分 / [包含前后处理(Python)] / [双缓冲模式(Python)]	YOLOv2: 224x224: 15fps	YOLOv5s: 224x224: 100fps 320x256 70fps 640x640: 15fps YOLOv8n: 640x640: 23fps YOLO11n: 224x224: 175fps 320x224: 120fps 320x320: 95fps 640x640: 23fps	YOLOv5s: 224x224: 495fps 320x256: 400fps 640x480: 106fps / 73fps / 103fps 640x640: 80fps YOLO11n: 224x224: 1214fps 640x480: 168fps / 77fps / 143fps 640x640: 113fps / 56fps / 98fps YOLO11s: 640x480: 87fps / 53fps / 83fps 640x640: 62fps / 39fps / 59fps YOLO11l: 640x640: 19fps / 16fps / 19fps
大模型	❌	❌	Qwen/DeepSeek 0.5B(fftf: 640ms, 9 tokens/s) Qwen/DeepSeek 1.5B(fftf: 1610ms, 4 tokens/s) VLM(InterVL 1B) 更多模型
OpenMV 典型算法	测试图像参考 Benchmark APP	测试图像参考 Benchmark APP 测试日期: 2025.8.22，更新可能会有优化	测试图像参考 Benchmark APP 测试日期: 2025.8.22，更新可能会有优化
二值化	灰度 320x240: 7.4ms (135fps) 灰度 640x480: ❌ RGB 320x240: 11.3ms (88.5fps) RGB 640x480: ❌	灰度 320x240: 3.1ms (326fps) 灰度 640x480: 11ms (90fps) RGB 320x240: 13.2ms (75fps) RGB 640x480: 52.8ms (18fps)	灰度 320x240: 1.3ms (799fps) 灰度 640x480: 4.8ms (206fps) RGB 320x240: 3.4ms (294fps) RGB 640x480: 13.3ms (75fps)
找色块	320x240: 8.8ms (114fps) 640x480: ❌	320x240: 7ms (143fps) 640x480: 20ms (50fps)	320x240: 3.7ms (271fps) 640x480: 11.1ms (89fps)
单通道直方图	320x240: 7.7ms (130fps) 640x480: ❌	320x240: 10.9ms (91fps) 640x480: 42.8ms (23fps)	320x240: 1.5ms (661fps) 640x480: 5.9ms (168fps)
二维码	320x240: 130.8ms (7.6fps) 640x480: ❌	640x480: 136.9ms (7fps) NPU 加速： 320x240: 22.1ms (45fps) 640x480: 57.6ms (17fps)	640x480: 57.9ms (17fps) NPU 加速： 320x240: 9.2ms (109fps) 640x480: 23.2ms (43fps)
OpenCV 典型算法		测试图像参考 Benchmark APP 测试日期: 2025.8.22，更新可能会有优化	测试图像参考 Benchmark APP 测试日期: 2025.8.22，更新可能会有优化
二值化	❌	灰度 320x240: 2.2ms (463fps) 灰度 640x480: 7.1ms (140fps)	灰度 320x240: 0.1ms (8174fps) 灰度 640x480: 0.3ms (2959fps)
灰度图自适应二值化	❌	320x240: 5.8ms (171fps) 640x480: 21.3ms (46fps)	320x240: 1.6ms (608fps) 640x480: 6.3ms (159fps)
单通道直方图	❌	320x240: 1ms (1000fps) 640x480: 6.2ms (160fps)	320x240: 0.4ms (2308fps) 640x480: 1.7ms (604fps)
轮廓提取	❌	320x240: 2.8ms (351fps) 640x480: 8.6ms (116fps)	320x240: 0.4ms (2286fps) 640x480: 1.4ms (692fps)

Maix 生态

社区

社区	地址
文档	MaixPy 文档
应用商店	maixhub.com/app
项目分享	maixhub.com/share
GitHub	GitHub搜索 `MaixCAM` 或者 `MaixPy`
Bilibili	B站搜索 `MaixCAM` 或者 `MaixPy`
讨论	maixhub.com/discussion
MaixPy issues	github.com/sipeed/MaixPy/issues
Telegram	t.me/maixpy
QQ 群	862340358