Skip to content

Latest commit

 

History

History
76 lines (44 loc) · 2.65 KB

agribot_visualservoing.md

File metadata and controls

76 lines (44 loc) · 2.65 KB

AgriBot Project (VisualServoing)


About VisualServoing Method

The visualServoing is a technique which uses feedback information extracted from a vision sensor to control the motion of the robot. Its usage in case of AgriBot is to navigate the robot through the crop rows in the field. the main graph of this node is shown.

visualservoing

In AgriBot navigation strategy, input image will receive from the camera mounted on the front of the robot and as the robot is supposed to work on field expected image is the scene of soil and some crop rows. Hence, features can efficiently be define based on crops and the background will always contain soil and some other plants out of the row, which based on experiments they should be weeds. In figure below the robot and its camera consolation are shown with respect to crop rows in the field.

robotinrow croprow vsout

About the agribot_visualservoing package

In Agribot_visualservoing package, a vision sensor as input is used to recognize crop rows, which in the launch file under /launch directory in the package user change the input source, too. it has access to agribot_local_planner and navigation (including map_server) to set new goals in /map frame.

To launch the agribot_visualservoing simply run:

$ roslaunch agribot_visualservoing visualservoing.launch

NOTE to have full functionality, you will need a localization service running to provide position of the robot. In case which you use the robot in the lab, use mocap system as localization server.

for more information about mocap system refer to this page Motion Capture

Node: [/agribot_vs]

Publications:

  • /cmd_vel [geometry_msgs/Twist]
  • /rosout [rosgraph_msgs/Log]

Subscriptions:

  • /amcl_pose [unknown type]
  • /odom [unknown type]
  • /usb_cam/image_raw [unknown type]
  • /zed/camera/left/image_raw [sensor_msgs/Image]

Services:

  • None

Parameters

  • None


by: Alireza Ahmadi
University of Bonn- Robotics & Geodetic Engineering

Alireza.Ahmadi@uni-bonn.de