🟠 Point Cloud Align to Image Shape

Function Description

This operator is used to reorganize an ordered point cloud data structure from a one-dimensional list (or any shape) into a two-dimensional grid shape that is completely consistent with the input image dimensions (height x width).

It does not change the coordinates or normal information of each point in the point cloud, only changes their arrangement in the data array.

Usage Scenarios

As a data preprocessing step, adapted to input formats with image shapes required by certain specific operators, making point cloud data have dimensions and structure consistent with the corresponding image.

Input Output

Input

Image: Input image, the height and width of this image will be used as target dimensions to reshape the point cloud.

Point cloud: Ordered point cloud to be processed, the number of point cloud points is the same as the number of image pixels.

Output

Output point cloud: Point cloud data structure is "height x width x number of channels", where height and width are the same as the input image.

Parameter Description

  • The input point cloud must be an ordered point cloud, where the storage order of points in the data is consistent with their pixel positions in the original depth image;

  • The total number of points in the input point cloud must strictly equal the total number of pixels in the input image (i.e., image height × image width), otherwise correspondence will not be possible.

  • This operator has two versions:

    • Point Cloud Align to Image Shape : Processes point clouds without normal information.

    • Point Cloud Align to Image Shape (with normals) : Processes point clouds with normal information.

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