Point Cloud Clustering and Output

Function Description

Performs clustering operations on point clouds and outputs point clouds that meet the set rules from the clustering results.

Use Cases

Generally used for denoising point clouds of individual objects.

Inputs

  • Point Clouds with Normals (Required): Point cloud to be processed.

  • Labels: Classification labels of the input point cloud.

  • Scores: Scores of the input point cloud.

Outputs

  • Processed Point Clouds with Normals: Point clouds that meet the set rules.

  • Labels: Classification labels corresponding to point clouds that meet the set rules.

  • Scores: Scores corresponding to point clouds that meet the set rules.

  • Out Flags: Output flags corresponding to point clouds that meet the set rules.

Parameters

  • Cluster Method

    Used to select the point cloud clustering method.

    • EuclideanCluster: Euclidean clustering, determines whether point clouds belong to the same class based on distance.

    • RegionGrowingSeg: Region growing segmentation, determines whether point clouds belong to the same class based on adjacent region distance.

    Using the default method is recommended.

  • Maximum Distance Between Adjacent Points

    Used to set the maximum distance between points in the same class.

    When this parameter is increased, points that are farther apart will be classified as the same class; when this parameter is decreased, points that are closer together will be classified as different classes.

  • Minimum Point Number

    Used to filter clustering results. Point clouds with point counts greater than this parameter value after clustering will only be output.

    If you need to filter out more outliers, increase this parameter.

  • Minimum Ratio of Output Class Points to Total Points (0-1.0)

    Used to filter clustering results. Point clouds will only be output when the ratio of their point count to the total input point count exceeds this threshold.

    Increase this parameter when you only need to retain the main point cloud in the scene; decrease this parameter when you only need to retain multiple separated point clouds in the scene.

  • Neighborhood Core Radius

    When “Clustering Method” is set to “RegionGrowingSeg”, this parameter needs to be set, using this pixel radius as the adjacent region range.

    Increase this value when there are large discontinuities in the point cloud but they need to be classified as the same class. Generally, it is recommended to use the default value.