v4.4.12

New Features

• i18n now support load from yaml dict files.
• Camera APP support record video(mp4 format), you can find video in /maixapp/share/video directory of device.

Optimization

• Camera(GC4653) support 60FPS mode now, and now fefault 60FPS(only support when resolution <= 720p).

60FPS mode's color has some different with 30FPS mode, if you want to use 30FPS mode, please set fps=30 in camera.Camera().

• AI model(NPU) support dual_buff mode and set it to default, now runing faster twice than before.

dual buff mode enabled by default, if you want to disable it, set arg dual_buff = False when initilize model like `nn.YOLOv5(model="“, dual_buff=False)

• APP bootup/switch faster now.
• YOLOv5 and YOLOv8 object coordinate now ensure all value >= 0.
• disp.set_backlight(0) now will close backlight now.

BUG Fix

• Setting install runtime success but always show not install bug.
• Face recoginizer run fail bug.
• image.cv2image and image.image2cv memory leak bug
• RTSP optimize, now you can play it with ffplay rtsp://10.228.104.1:8554/live or mpv rtsp://10.228.104.1:8554/live

Downloads

github release assets (recommend!)	Sourceforge	Pypi.org
Assets below ↓↓↓↓↓↓↓↓↓ 点击下方 Assets

v4.4.10

New features

• Add YOLOv8-seg object segmentation support.
  
• Add YOLOv8-face model.
• Add more examples like network(mqtt/http/socket/websocket/flask)
• Add support for no screen MaixCAM, now you can use scripts in examples/tools to connect wifi, install runtime, set auto start app etc.
• Open all APPs source code at MaixPy/projects and MaixCDK/projects.
• Add RTSP APP and YOLOv8-seg APP and face-tracking(gimbal) APP and audio record APP and self-learn-classifier APP.
• Add maix.app.switch_app function to switch app.
• JPEG and PNG now support convert to other format by image.to_format() function.
• Audio support wav record and play.
• image.draw_string support \r\n \n chars to wrap line and utf-8 charactors auto linewrap now.

Optimization

• Optimize opencv and image convertion API
• Optimize RTSP demo
• Optimize API doc and its layout
• Limit YOLOv8 and YOLOv5 bbox output value >=0
• Optimzie image.get_histogram
• Increase the driving capability of pwm
• YOLOv8 now support all number of keypoint, not only 17 points

BUG fix

• Fix bug that RTSP streaming can't play by old vlc and ffplay
• Fix camera buffer issues

Attention

To use MaixPy v4.4.10, you MUST upgrade system!!! The old system can not use camera due to the change of driver.
注意：要使用 MaixPy v4.4.10， 必须 更新系统！！！ 用旧的系统由于驱动更新会无法初始化摄像头。

Downloads

github release assets (recommend!)	Sourceforge	Pypi.org
Assets below ↓↓↓↓↓↓↓↓↓ 点击下方 Assets

v4.3.2

MaixPy

• New features

  • Add YOLOv8 support.
  • Add 17 human body keypoints detection support(YOLOv8-Pose).
  • Add self learn classifier support(learn and classify anything).
  • Add 2-axis gimbal project.
  • Add SPI support(soft Simulate), add spi st7789 demo.
  • Add ADC support.
  • image.draw_string support auto linefeed and \r\n \n now.
  • Add time.ticks_s() and time.ticks_ms() time.ticks_us() to calculate time interval instead of time.time().
  • Support set camera to grayscale mode.
  • Audio support wav format.

• Bug fix

  • Fix time.sleep stuck thread bug.
  • Repalce time.time_ms() with time.ticks_ms() to avoid datetime change bug.
  • Fix camera's first read() call failed when capture black scenes.
  • Fix disp.show() and image.send_to_maixvision func cannot send fisrt image error.

APP

• Add Human Pose to detect human body 17 keypoints, support multiple people.
• Add Self Learn Classifier, you can learn anything and recognize it in a minute.

System

• Add SPI driver.
• Fix maixvision auto disconcnect after first connection for a while.

sha256sum maixcam-2024-06-20-maixpy-v4.3.2.img.xz:
4a6bba1242aa38287fa899d8ee65ac08a7ad46dd52b025e5c2e222e8e199b545

Downloads

github release assets (recommend!)	Sourceforge	Pypi.org
Assets below ↓↓↓↓↓↓↓↓↓ 点击下方 Assets

v4.2.1

MaixPy

New features

• Add RTMP support
• Add face detector support
• Add face recognize support
• Add soft simulate I2C/SPI support
• Add pinmap module

Bug fix

• Camera reopen failed and compress jpeg error

APP

• setttings add wifi scan QR code
• Add RTMP live streaming APP(streaming to bilibili.com or youtube twitch etc.
• Add face recognizer APP, support multi face recognize

System

• Add boot up splash
• Remove always fsck disk, so bootup faster, if you encounter boot up failed with disk files error, tell us or just reburn system image to resolve.
• Add bitbang soft i2c spi driver

sha256sum maixcam-2024-06-04-maixpy-v4.2.1.img.xz: 7eb7d3d1e7488b49dab239fdcb9b498dc8e19e01abb8c75b01fa36e93fdc2f83

Downloads

github release assets (recommend!)	Sourceforge	Pypi.org
Assets below ↓↓↓↓↓↓↓↓↓ 点击下方 Assets

v4.2.0

optimize show

v4.1.2

MaixPy

New features

• Add audio record play
• Add video record
• Add PWM support
• Add RTSP support
• Add image.to_jpeg(quality) API

Bug fix

• Compress image error when second open program
• UART find_devices can not get anything bug
• MaixVision code completion error

System

• Support PWM
• Auto resize space to fit TF card on boot
• set DNS by default in /boot/resolv.conf， remove it if you want to use auto DNS server by gateway

v4.0.11

MaixPy

New features

• Add WiFi API
• display.show's fit arg now default to FIT_CONTAIN

Bug fix

• yolov5 load one label always failed issue

System

• Update APP Store support MaixHub scan download install.
• Update Settings APP:
  • support backlight setting
  • support auto start APP setting
  • optimize Power setting
  • optimize WiFi connect
• Classifier APP now display as squre
• Optimize system start speed, now faster
• Add error hint when bootup init display failed

v4.0.10

First release

attention, this is a beta version, some API may have changes in ther future

other download links: https://sourceforge.net/projects/maixpy/files/v4.0.10/