Vineyard Row Middle Detection

ROS package for detecting the middle of the passage between two rows of grapevines.

Point Cloud Library (PCL) is used to detect the border rows from a 3D LiDAR scan.

❓ What does it do?

1. Removes pointcloud points inside the Remove Box
2. Removes pointcloud points outside the Keep Box
3. Flattens the remaining pointcloud (sets the z component to zero)
4. Detects all the lines subject to max_angle and min_points parameters
5. Selects right and left borders from detected lines
6. Calculates the middle line and publishes it to the middle_line topic

📄 Dependences

This project depends on ROS, PCL and Eigen

🛠️ Usage

⚙️ Build the project

Clone into a catkin workspace and build with:

catkin build vineyard_midrow_detection

🖥️ Launching the project

To launch the node along with Rviz visualization:

roslaunch vineyard_midrow_detection mid_row_detection.launch  

Parameters

Reconfigurable parameters

• Line Detection - PCL line detection using RANSAC

  • line_detection_distance_threshold - minimum distance for a point to be considered a line inlier
  • line_detection_max_angle - maximum angle of a line to be considered a candidate for border selection
  • line_detection_min_points - minimum number of points to perform line detection on

• Keep Box - points inside the Keep Box cuboid are considered for line detection

  • keep_box_lower_bound_x
  • keep_box_upper_bound_x
  • keep_box_lower_bound_y
  • keep_box_upper_bound_y
  • keep_box_lower_bound_z
  • keep_box_upper_bound_z

• Remove Box - points inside the Remove Box cuboid are removed. Used to remove irrelevant points such as vehicle parts

  • remove_box_lower_bound_x
  • remove_box_upper_bound_x
  • remove_box_lower_bound_y
  • remove_box_upper_bound_y
  • remove_box_lower_bound_z
  • remove_box_upper_bound_z

Launch file parameters

• rviz - start RVIZ (default true)
• rqt_reconfigure - start rqt_reconfigure (default true)
• ros_rate - ros rate (default 50Hz)
• input_pointcloud_topic - Pointcloud2 topic (default /rslidar_points)