🟠 Point Cloud Stitching

Function Description

Receives multiple sets of corresponding point clouds, 3D ROIs, and hand-eye calibration matrices, which can be used to merge point cloud data collected from multiple different spatial positions or viewpoints in the robot base coordinate system.

  1. For each set of data, first use its corresponding 3D ROI to crop and filter the point cloud;

  2. Use its corresponding hand-eye calibration matrix to transform the cropped point cloud to the unified robot base coordinate system;

  3. Stitch all point clouds processed through the above steps into a single point cloud.

Usage Scenarios

Multi-camera stitching or single camera multi-position stitching.

Input Output

Input

Point cloud: A point cloud list where each element is point cloud data from one acquisition.

3D ROI: A 3D ROI list that corresponds one-to-one in order and quantity with the input point cloud list, used to crop each point cloud.

Hand-eye calibration matrix: A transformation matrix list that also corresponds one-to-one in order and quantity with the input point cloud list, used to transform each point cloud from its respective camera coordinate system to the robot base coordinate system.

Output

Point cloud: A single point cloud that is the final result of merging all input point clouds after their respective ROI cropping and coordinate transformations.

Parameter Description

  • The three input ports—"Point cloud", "3D ROI", "Hand-eye calibration matrix"—must receive lists, and the number of elements in these three lists must be exactly equal and correspond one-to-one. The operator will process the first point cloud, first ROI, and first matrix as one group, and so on. If the quantities do not match, the operator will report an error.

  • This operator has two versions:

    • Point Cloud Stitching: Processes point clouds without normal information.

    • Point Cloud Stitching (with normals): Processes point clouds with normal information.

Both have identical core functionality and parameters, differing only in the type of point cloud data they process.