This repository is a resource for anyone interested in implementing the basics of image-based visual servoing methods . With the help of the included scripts and files, you'll be able to quickly create an implementation.
The Gazebo simulation environment and the ROS environment and the integration of these two are required to use this method. (Used ROS version: Noetic, Gazebo version:11.12.0)
You can use the following link to install these:https://classic.gazebosim.org/tutorials?tut=install_ubuntu
After the installation you can build the 'pack' package.
Starting roscore to use ROS communication:
roscore
Launch the Gazebo world file:
roslaunch pack launcher.launch
Setup a topic to adjust camera position:
rosrun pack Set_camera_position_from_topic.py
Set the camera position to the hard coded coordinates, this position will be the desired position:
rosrun pack Publish_camera_position_to_topic.py
Create desired input values (image,Z value):
rosrun pack Feature_image.py
Change hard coded camera position in Publish_camera_position_to_topic.py to the starting position, and then rerun that.
rosrun pack Publish_camera_position_to_topic.py
Start image based visual servoing method:
rosrun pack IBVS.py
Running the following commands in separated command windows:
roscore
roslaunch pack launcher.launch
rosrun pack Set_camera_position_from_topic.py
Define the coordinates in Publish_camera_position_to_topic.py
if __name__ == '__main__':
values = np.array([0,0,0.3,180,0,0]) # desired camera position for now
publish_camera_pose(values)
Create desired input values (image,Z value):
rosrun pack Feature_image.py
Define the coordinates in Publish_camera_position_to_topic.py
if __name__ == '__main__':
values = np.array([0.015,-0.01,0.5,179,2,30]) #starting camera position
publish_camera_pose(values)
Start image based visual servoing method:
rosrun pack IBVS.py
The movement of the camera object can be seen on the Gazebo simulation window (in the eaxmple from position 0 to position 6, along the white arrow), and the error vector can be seen on the commander:
The camera position during the method:
