Official code for the OpenNav paper. ACVR Workshop at ECCV'24 The official implementation of the Twelfth International Workshop on Assistive Computer Vision and Robotics paper OpenNav: Efficient Open Vocabulary 3D Object Detection for Smart Wheelchair Navigation.
The method is tested for ROS2 Humble on Ubuntu 22.04 with IntelRealsense D455 camera
Create a new conda environment
conda create -n opennav python=3.10
conda activate opennav
Install dependencies
pip install -r requirements.txt
To run this project, you first need to install ROS 2 Humble. Follow the instructions below to set up ROS 2 on your system.
Follow the official ROS 2 installation guide for your operating system. You can find the instructions at ROS 2 Humble Installation.
sudo apt update
sudo apt install python3-colcon-common-extensions
sudo apt install ros-humble-desktop
sudo apt install ros-humble-vision-msgs
sudo apt install ros-humble-cv-bridge
After installing, you need to source the ROS 2 setup script to add the ROS 2 commands to your environment:
source /opt/ros/humble/setup.bash
To make this change permanent, add the line above to your ~/.bashrc file:
echo "source /opt/ros/humble/setup.bash" >> ~/.bashrc
source ~/.bashrc
After installing ROS 2 and the required packages, clone the repository of project:
git clone https://github.com/EasyWalk-PRIN/OpenNav.git
Build the project
cd OpenNav
colcon build
source install/setup.bash
First, open a terminal and run node for realsense camera
ros2 launch realsense2_camera rs_launch.py pointcloud.enable:=true depth_module.profile:=640x480x30 rgb_camera.profile:=640x480x30
Launch 3D detection pipeline
ros2 launch rl_detect_bringup rl_detect_launch.py
For visualization, Open a new terminal with ROS2 sourced and run
rviz2
Different Yoloworld models and classes of interest can be specified in
src/rl_detect_bringup/launch/rl_detect_launch.py
@InProceedings{10.1007/978-3-031-92591-7_23,
author="Rahman, Muhammad Rameez ur
and Simonetto, Piero
and Polato, Anna
and Pasti, Francesco
and Tonin, Luca
and Vascon, Sebastiano",
title="OpenNav: Efficient Open Vocabulary 3D Object Detection for Smart Wheelchair Navigation",
booktitle="Computer Vision -- ECCV 2024 Workshops",
year="2025",
publisher="Springer Nature Switzerland",
address="Cham",
pages="372--387",
isbn="978-3-031-92591-7"
}