Hailo RPi5

This repository provides a detection-tracker algorithm for the Raspberry Pi 5, utilizing the AI capabilities of the Hailo-8L chip. The algorithm is implemented within a GStreamer pipeline, specifically designed to track the first object detected by the YOLO model.

Overview

This project offers a detection and tracking solution integrated with the GStreamer pipeline. It aims to track objects identified by the YOLO model, enhancing the AI capabilities of your Raspberry Pi 5 setup.

Prerequisites

Before running this code, ensure you have the following hardware and software set up:

• Raspberry Pi 5
• AI Kit, based on Hailo-8L
• Display connected to the Raspberry Pi

Find how to set them correctly here.

How to run?

Initialize and Update Submodules

First, clone the repository:

git clone --recursive https://github.com/dataroot/hailo-rpi5.git

Configure the environment

To set up the environment, run the following commands:

cd hailo-rpi5/hailo-rpi5-examples
source setup_env.sh
pip install -r requirements.txt
./download_resources.sh
cd ..

If you already have a configured environment, simply activate it:

cd hailo-rpi5/hailo-rpi5-examples
source setup_env.sh
cd ..

Export display

Export display to visualize the application's output:

export DISPLAY=:0

Run the application

Finally, execute the application:

python custom_pipeline.py

If you want to save the processed video, add the --save argument with the directory and filename with the .mkv extension:

python custom_pipeline.py --save "directory/to/file.mkv"

Example of running:

The following are examples of the application running, illustrating its tracking capabilities exclusively: