The PointCloud object

The PointCloud class is essential for all point cloud operations and processing within this Toolbox. The class encapsulates a NumPy array that contains the point coordinates and optionally colors as a 3D array respectively. An PointCloud instance has many methods that perform useful operations on a point cloud and wrap low-level operations performed using NumPy or Open3d.

class PointCloud(arg, depth_scale=1.0, **kwargs)

__init__(arg, depth_scale=1.0, **kwargs)

Create new point cloud object

Parameters:

• arg (o3d.geometry.PointCloud, PointCloud, ndarray(3,N)) – point cloud data as an Open3D or MVTB PointCloud object, 3xN array of points.

• depth_scale (float, optional) – depth scale factor, defaults to 1.0

Raises:

• RuntimeError – PointCloud class requires Open3D to be installed: pip install open3d

• ValueError – bad arguments

This object wraps an Open3D open3d.geometry.PointCloud object. It can be created from:

• an Open3D point cloud object

• a 3xN array of point coordinates

• a depth image as a 1-plane Image instance

• an RGBD image as a 4-plane Image instance

• a tuple of two images (RGB, D) as 3-plane and 1-plane Image instances

Warning

Open3D must be installed.

The wrapped Open3D open3d.geometry.PointCloud object has many methods and properties but rather than writing individual wrappers the __getattr__ picks up these references and invokes them on the underlying Open3D object. Any returned Open3D point cloud objects are converted to a PointCloud instance.

Seealso:

__getattr__ open3d.geometry.PointCloud Image CentralCamera

Informational

__str__	Concise string representation of point cloud parameters
__repr__	Return repr(self).
__len__	Number of points

Access to point cloud data

pcd
points	Get points as array
colors	Get point color data as array

Point cloud i/o

Read	Create point cloud from file
write	Write point cloud to file
disp	Display point cloud using Open3D

Point cloud operations

copy	Copy point cloud
transform	Transform point cloud
downsample_voxel	Downsample point cloud by voxelization
downsample_random	Downsample point cloud by random selection
voxel_grid	Voxelize point cloud
normals	Get point normal data as array
remove_outlier	Remove point cloud outliers
segment_plane
select	Select points by index
paint	Colorize point cloud
ICP	Register point cloud using ICP
__getattr__	Access Open3D PointCloud methods and attributes

Overloaded operators

__rmul__	Overloaded * operator to transform points
__imul__	Overloaded *= operator to transform points
__add__	Overloaded the + operator to concatenate point clouds