Lane Detection AI

This repository contains the ros2 jazzy package for the ai lane detection based on the ultra fast lane detection v2.

Lane Detection AI

Installation

1. Clone the package into the smarty_workspace/src directory:

   git clone <repository-url> smarty_workspace/src

2. Install the required Python dependencies using the requirements.txt file:

   pip install -r requirements.txt

3. Download the trained model and its corresponding configuration file. (See the Usage section for more details.)
4. Please set the following ENV variable in your .bashrc/.zshrc if not already done:

   PYTHON_EXECUTABLE="/home/$USER/.pyenv/versions/default/bin/python3" # Change this to the python3 executable path of your pyenv

Usage

To run this package, you will need the pre-trained model weights. As of September 2024, two versions are available: a dense model and a sparse model. You can download them here. Place the downloaded model in the models/ folder, and the corresponding configuration file in the config/ folder.

Important: Ensure that you update the model_config_path in the config/ros_params.yaml file, and set the correct path for the model in the configuration file's test_model field.

Running the Node

To launch the node, use the following command:

ros2 launch lane_detection_ai lane_detection_ai.launch.py

For running with the debug image enabled, use:

ros2 launch lane_detection_ai lane_detection_ai.launch.py debug:=true

Structure

• config/: All configurations for ROS and the model
• launch/: Contains the launch files
• models/: Contains the models
• resource/: Contains the package name (required to build with colcon)
• lane_detection_ai/: Contains all nodes and sources for the ros package
• lane_detection_ai/model/: Contains all sources of the ufldv2 model
• test/: Contains the tests
• package.xml: Contains metadata about the package
• setup.py: Used for Python package configuration
• setup.cfg: Additional configuration for the package
• requirements.txt: Python dependencies

License

This repository is licensed under the MIT license. See LICENSE for details.