Adaptive Landmark Color for AUV Docking in Visually Dynamic Environments

Dependencies

Ubuntu 20.04
ROS Noetic
opencv-python (tested with version 4.9.0.80)

Python 3.8
cv-bridge (tested with version 1.16.2)
numpy (tested with version 1.24.4)

A Docker image with all necessary software packages installed is available on DockerHub here and the run-container.sh script is provided to run the example code. It is not necessary to install ROS 1 or the specific Python packages to run the Docker container.

On a system with a valid Docker configuration, simply run the following command to be dropped into a shell with all necessary software installed

./run-container.sh

Usage

Use in your existing catkin_ws

1. Clone this repository

git clone https://github.com/IRVLab/adaptive_landmarks.git

2. Move the adaptive_landmarks package from this repo into your catkin_ws/src directory

mv adaptive_landmarks/catkin_ws/src/adaptive_landmarks catkin_ws/src/

3. (Optional) Add the following lines to your existing launch file to launch the nodes. Replace CAMERA_TOPIC with your desired camera topic.

<node pkg="adaptive_landmarks" type="adaptive_color_node.py" name="adaptive_color_node" output="screen">
    <param name="camera_topic" value="CAMERA_TOPIC"/>
</node>
<node pkg="adaptive_landmarks" type="blob_detector_node.py" name="blob_detector_node" output="screen">
    <param name="camera_topic" value="CAMERA_TOPIC"/>
</node>

By default, the adaptive_color node calculates the ternary marker color. To change this to the complementary marker color, edit the catkin_ws/src/adaptive_landmarks/src/adaptive_color_node.py, line 39, and change the shift to shift=180.

Run on the example bag

It is recommended to run commands in a Docker container to ensure compatibility.

1. Clone this repository

git clone https://github.com/IRVLab/adaptive_landmarks.git

2. Build the catkin_ws

cd adaptive_landmarks
./run-container.sh
cd catkin_ws
catkin_make

3. In the same terminal, launch the robot_meta detect_only.launch file

source devel/setup.bash
roslaunch robot_meta detect_only.launch

4. In terminal 2, play the example bag file

cd adaptive_landmarks
./run-container.sh
rosbag play -l example.bag

5. In terminal 3, open rqt_image_view and view the /blob_detector/image topic for the detected docking station markers.

cd adaptive_landmarks
./run-container.sh
rqt_image_view