Floor Detection From Point Cloud

This ROS2 package detects holes in ground point clouds. The package uses OpenCV to detect the holes. The package should be used with the linefit_ground_segmentation_ros2 package.

Description

This package provides a node, floor_detection_from_pointcloud_node, which takes in point cloud data (presumably from a LiDAR or similar sensor), and applies algorithms to detect the presence of holes in the ground.

The list of ROS parameters that the node uses are:

• camera.range_near: Range start for the camera (default: 0.5),
• camera.range_far: Range end for the camera (default: 6.0),
• camera.fov_degrees: Field of view for the camera in degrees (default: 50.0),
• camera.gravity_aligned_frame: The frame aligned with gravity (default: camera_gravity_link),
• resolution: Resolution of the images (default: 0.1),
• img_size: Size of the images (default: 20.0),
• show_debug_window: Indicates if a debug window is shown or not (default: false),
• use_sim_time: If true, the system will use simulated time (default: true).

The package also includes a debug viewer which can show the results in a user-friendly manner.

Topics

The node subscribes to the following topics for the input ground and obstacles pointcloud data:

• /segmentation/ground: This topic should publish messages of type sensor_msgs/msg/PointCloud2. The node uses this as the input for the ground pointcloud.

• /segmentation/obstacle: This topic should publish messages of type sensor_msgs/msg/PointCloud2. The node uses this as the input for the obstacles pointcloud.

Additionally, the node publishes the detected holes to the following topics:

• detected_holes: This topic publishes messages of type sensor_msgs/msg/LaserScan. It publishes the range of detected holes.

• detected_holes_cloud: This topic publishes messages of type sensor_msgs/msg/PointCloud2. It publishes the pointcloud of the detected holes.