pcalign

Align array of point clouds

Since R2020b

Syntax

ptCloudOut = pcalign(ptClouds,tforms)

ptCloudOut = pcalign(ptClouds,tforms,gridStep)

ptCloudOut = pcalign(___,GridFilter=gridFilterMethod)

Description

ptCloudOut = pcalign(ptClouds,tforms) aligns an array of point clouds, ptClouds, into one point cloud by using the specified transformations tforms.

ptCloudOut = pcalign(ptClouds,tforms,gridStep) applies a box grid filter with 3-D boxes of size specified as gridStep, to the aligned point cloud. The grid filter returns a single point per 3-D box.

ptCloudOut = pcalign(___,GridFilter=gridFilterMethod) sets the method to downsample the aligned point cloud.

Examples

Align Point Clouds from View Set

Load a point cloud view set into the workspace.

data = load('vSetPointClouds.mat');
vSet = data.vSet;

Extract the point clouds and their absolute poses from the view set.

ptClouds = vSet.Views.PointCloud;
tforms   = vSet.Views.AbsolutePose;

Align the point clouds using their absolute poses, and apply a 3-D box filter of size 1 to the aligned point cloud.

gridStep = 1;
ptCloudMap = pcalign(ptClouds,tforms,gridStep);

Display the aligned point cloud map.

pcshow(ptCloudMap)
hold on
plot(vSet)

Input Arguments

ptClouds — Point clouds

pointCloud array

Point clouds, specified as a pointCloud array.

tforms — 3-D rigid or affine geometric transformations

M-by-1 rigidtform3d array | M-by-1 affinetform3d array

3-D rigid or affine geometric transformations, specified as an M-by-1 array of rigidtform3d objects or an M-by-1 array of affinetform3d objects. The tforms argument must contain the same number of elements as the ptClouds argument.

gridStep — Size of 3-D box for grid filter

positive scalar

Size of the 3-D box for the grid filter, specified as a positive scalar. Increasing the value of gridStep can reduce memory usage by merging more points, but it can also reduce the amount of detail captured. The grid filter returns a single point per 3-D box.

Data Types: single | double

gridFilterMethod — Method to downsample aligned point cloud

"average" (default) | "nearest"

Method to downsample the aligned point cloud, specified as one of:

• "average" (default) — Selects a point in the voxel by averaging the location, color, intensity, and normal of all points within the voxel.

• "nearest" — Selects the voxel's location, color, intensity, and normal based on the point nearest to the centroid of the voxel.

Output Arguments

ptCloudOut — Aligned point cloud

pointCloud object

Aligned point cloud, returned as a pointCloud object.

Tips

• If the input point clouds do not all have an assigned value for a property, the function does not assign a value for that property in the returned point cloud. For example, if some of the input point clouds have values for the Color property but another one does not, then the function does not return a value for the Color property of ptCloudOut.

• The function filters out points with NaN or Inf values.

Extended Capabilities

C/C++ Code Generation
Generate C and C++ code using MATLAB® Coder™.

Version History

Introduced in R2020b

R2022b: Supports new geometric transformation objects

Starting in R2022b, most Computer Vision Toolbox™ functions create and perform geometric transformations using the premultiply convention. Accordingly, you can now specify tforms as an array of rigidtform3d or affinetform3d objects, which use the premultiply convention.

Although you can still specify tforms as an array of rigid3d or affine3d objects, these objects are not recommended because they use the postmultiply convention. You can streamline your geometric transformation workflows by switching to the new premultiply geometric transformation objects. For more information, see Migrate Geometric Transformations to Premultiply Convention.