Last edited 4 years ago

STM32MP1 artificial intelligence expansion packages

The artificial intelligence expansion package contains Linux AI frameworks to enable AI application examples that can be run on STM32MP1 Series devices.
This package consists in an OpenEmbedded meta layer, named meta-st-stm32mpu-ai, to be added on top of the STM32MP1 Distribution Package. It brings a complete and coherent easy-to-build / install environment to take advantage of AI on the STM32MP1 Series.
The meta layer contains frameworks, tools and applications to run AI examples. Different image flavors are to be made available targeting different use cases such as computer vision (CV).

STM32MPU Embedded Software Distribution Package with the OpenSTLinux AI Expansion Package

1. Prerequisite[edit source]

Info white.png Information
The artificial intelligence expansion package meta-st-stm32mpu-ai has been validated against the ecosystem release v1.1.0 More info.png

Install the STM32MP1 Distribution Package v1.1.0, but do not initialize the OpenEmbedded environment (sourcing the envsetup.sh) before having installed the meta-st-stm32mpu-ai meta layer (see next chapter).

2. Installation of the meta layer[edit source]

2.1. STM32MP1 DK2 or EV1 boards[edit source]

The software package is provided AS IS, and by downloading it, you agree to be bound to the terms of the software license agreement (SLA). The detailed content licenses can be found here.

Info white.png Information
The meta-st-stm32mpu-ai layer depends on the meta-clang layer.
  • Clone following git repositories into <Distribution Package installation directory>/layers/meta-st
 cd <Distribution Package installation directory>/layers
 git clone https://github.com/kraj/meta-clang -b thud
 cd <Distribution Package installation directory>/layers/meta-st
 git clone https://github.com/STMicroelectronics/meta-st-stm32mpu-ai.git -b v1.1.0
  • Set up the build environment
 cd ../..
 DISTRO=openstlinux-weston MACHINE=stm32mp1 source layers/meta-st/scripts/envsetup.sh
  • Add the new layers in that order
 bitbake-layers add-layer ../layers/meta-clang
 bitbake-layers add-layer ../layers/meta-st/meta-st-stm32mpu-ai

2.2. STM32MP1 Avenger96 board[edit source]

The software package is provided AS IS, and by downloading it, you agree to be bound to the terms of the software license agreement (SLA). The detailed content licenses can be found here.

Info white.png Information
The meta-st-stm32mpu-ai layer depends on the meta-clang layer.
  • Clone following git repositories into <Distribution Package installation directory>/layers/meta-st
 cd <Distribution Package installation directory>/layers
 git clone https://github.com/kraj/meta-clang -b thud
 git clone https://github.com/vinceab/meta-av96.git -b thud
 cd <Distribution Package installation directory>/layers/meta-st
 git clone https://github.com/STMicroelectronics/meta-st-stm32mpu-ai.git -b v1.1.0
  • Set up the build environment
 cd ../..
 META_LAYER_ROOT=layers DISTRO=openstlinux-weston MACHINE=stm32mp1-av96 source layers/meta-st/scripts/envsetup.sh
  • Add the new layers in that order
 bitbake-layers add-layer ../layers/meta-av96
 bitbake-layers add-layer ../layers/meta-clang
 bitbake-layers add-layer ../layers/meta-st/meta-st-stm32mpu-ai

3. Build the software image[edit source]

Different OpenSTLinux expansion packages are available to target different use cases. The following commands need to be executed in the build environment.

  • X-LINUX-AI-CV for computer vision use case
For further information about this OpenSTLinux expansion package, read this article X-LINUX-AI-CV OpenSTLinux expansion package
To build it, execute the following command:
 bitbake st-image-ai-cv
Info white.png Information
Note that building the image could take more than 2 hours depending on the host computer performance.

4. Flash the built image[edit source]

Follow this link to know how to flash the built image.

5. Launch the AI software[edit source]

5.1. The AI demo launcher[edit source]

The AI demo launcher is a derivative of the GTK demo launcher application.

It is written in python3[1] and uses GTK[2] to display the user interface. It allows easy launching of the AI application examples.
A "single tap" on the touch screen or a "single click" with a mouse connected to the board is sufficient to launch the AI application.

Example of AI demo launcher for Computer Vision

The AI demo launcher is updated with the available examples according the built image.
If a computer vision package is built, the AI demo launcher then contains the supported computer vision application examples.

Info white.png Information
Source code of the AI demo launcher is located here:
  • in the meta layer:
meta-st-stm32mpu-ai/recipes-samples/demo/ai-demo-launcher/ai-demo-launcher.py
  • on the target:
/usr/local/demo-ai/ai-demo-launcher.py

6. References[edit source]