👷🏽‍♀️🚀 PPE Detection Demo with Axelera

Detect safety compliance with AI

This PPE (Personal Protective Equipment) detection demo, powered by Axelera AI and developed by RidgeRun.ai, helps you prevent workplace hazards.

Upload your own video and watch as the AI—running on Axelera’s high-performance hardware identifies persons, hardhats, and safety vests. A person is marked safe only if they are wearing both a hardhat and a safety vest. Unsafe persons are clearly highlighted in red.

✨ Features

• PPE detection powered by Axelera Metis hardware.
• Custom video input – bring your own footage.
• Clear compliance feedback – unsafe visual marking.
• Simple Docker setup – minimal configuration required.

🗒️ Requirements

• An Axelera Metis device supporting Voyager SDK 1.3.1.
• Installed Metis PCIe driver.

1. Download the Metis PCIe driver:

wget https://software.axelera.ai/artifactory/axelera-apt-source/metis-dkms/metis-dkms_1.0.2_all.deb

2. Install the Metis PCIe driver:

sudo dpkg -i metis-dkms_1.0.2_all.deb
sudo modprobe metis

👷🏽‍♀️ Running the PPE Detection Demo

1. Set your Axelera AI credentials:

export AXELERA_EMAIL=johndoe@example.com
export AXELERA_TOKEN=xxxxxx

2. Start Docker container up:

docker compose up

3. Open the web UI at http://127.0.0.1:7860 in your browser, upload your video, and watch PPE detection in action.

To build the image before starting the container:

docker compose up --build

Installing the Project for Development

This is a Python project and, as such, standard Python tooling may be used. However, we recommend using UV to handle the project.

Note: the demo needs Python 3.10 or higher.

Using UV (Recommended)

UV is a fast Python package and project manager. It's proven to be a more robust dependency manager than similar alternatives.

Note: Install UV, if you haven't, following the official instructions.

To install the project and its dependencies:

uv sync

Then, install the pre-commit hooks.

uv run pre-commit install

• The sync portion will create a venv, install the dependencies and finally install the project in development mode.
• The pre-commit line will install hooks to automatically perform checks before a commit or a merge.

After installing the project, you can activate the virtual environment and continue to work without UV:

. .venv/bin/activate
python 

Warning: Avoid installing the project by running uv pip install -e .[dev] as it won't use the uv.lock to match dependencies.