RegionGrowingMultiplePlaneSegmentation

jsk_pcl/RegionGrowingMultiplePlaneSegmentation.

jsk_pcl/RegionGrowingMultiplePlaneSegmentation estimates multiple plenes from pointcloud.

It extracts planes based on region growing and evaluation function of connectivity if based on the following equation:

Subscribing Topics

  • ~input (sensor_msgs/PointCloud2):

    input pointcloud.

  • ~input_normal (sensor_msgs/PointCloud2):

    normal pointcloud of ~input

Publishing Topics

  • ~output/clustering_result (jsk_recognition_msgs/ClusterPointIndices):

    Result of region growing as cluster.

  • ~output/inliers (jsk_recognition_msgs/ClusterPointIndices):

    Set of indices of the polygons.

  • ~output/coefficients (jsk_recognition_msgs/ModelCoefficientsArray):

    Array of coefficients of the polygons.

  • ~output/polygons (jsk_recognition_msgs/PolygonArray):

    Polygons

  • ~output/latest_time (std_msgs/Float32)

    latest computation time

  • ~output/average_time (std_msgs/Float32)

    average computation time

Parameters

  • ~angular_threshold (Double, default: 0.04)

    Angular threshold to connect two points in one cluster. See

  • ~distance_threshold (Double, default: 0.1)

    Distance threshold to connect two points in one cluster.

  • ~max_curvature (Double, default: 0.1)

    Before extracting planes, filtering out the points which have higer curvature than this value.

  • ~min_size (Integer, default: 100)

    The minimum number of the points of each plane.

  • ~min_area (Double, default: 0.1)

    The minimum area of the convex areas.

  • ~max_area (Double, default: 100)

    The max area of the convex areas.

  • ~cluster_tolerance (Double, default: 0.1)

    The spatial tolerance for new cluster candidates.

  • ~ransac_refine_outlier_distance_threshold (Double, default: 0.1)

    Outlier threshold for plane estimation using RANSAC.

  • ~ransac_refine_max_iterations (Integer, default: 100)

    The maximum number of the iterations for plane estimation using RANSAC.