Apply Masks to Point Cloud

Function Description

This step applies masks to point clouds, retaining points inside the mask and removing points outside it.

Use Cases

This step extracts the 3D point cloud defined by a mask. It is typically used with the “Project 3D Point Cloud to 2D Image” step: first generate the mask image, then use that mask to extract the corresponding region from the original point cloud.

Inputs

• Mask Images (Required): 2D mask image list.

• Original Point Cloud (Required): Original point cloud.

• Labels (Optional): Labels of target objects, generally the labels of target objects output by deep learning related steps.

• Confidences (Optional): Confidence of target objects, generally the confidence of target objects output by deep learning related steps.

Outputs

• Point Cloud After Intersection: Point clouds in each mask region.

• Classification Labels: Labels of target objects (related to deep learning steps).

• Confidence: Confidence of target objects (related to deep learning steps).