PointCloudNode

The PointCloudNode class in Geomapi represents the data and metadata of point cloud data. The data itself and methods build upon Open3D PointCloud concepts while the metadata builds upon the RDFlib framework:

http://www.open3d.org/docs/release/python_api/open3d.geometry.PointCloud.html

https://rdflib.readthedocs.io/

The code below shows how to create a PointCloudNode from various inputs.

First the geomapi and external packages are imported

#IMPORT PACKAGES
from rdflib import Graph
import os
import numpy as np

#IMPORT MODULES
from context import geomapi #context import for documentation only
from geomapi.nodes import PointCloudNode

Jupyter environment detected. Enabling Open3D WebVisualizer.
[Open3D INFO] WebRTC GUI backend enabled.
[Open3D INFO] WebRTCWindowSystem: HTTP handshake server disabled.

PointCloudNode Creation

A PointCloudNode is constructed using the same parameters as the base Node. Please refer to Node Tutorial For more info about Node Creation

PointCloudNode( subject = None,             # (URIRef, optional) : A subject to use as identifier for the Node.
                graph = None,               # (Graph, optional) : An RDF Graph to parse.
                graphPath = None,           # (Path, optional) : The path of an RDF Graph to parse.
                name = None,                # (str, optional) : A name of the Node.
                path = None,                # (Path, optional) : A filepath to a resource.
                timestamp = None,           # (str, optional) : Timestamp for the node.
                resource = None,            # (optional) : Resource associated with the node.
                cartesianTransform = None,  # (np.ndarray, optional) : The (4x4) transformation matrix.
                orientedBoundingBox = None, # (o3d.geometry.OrientedBoundingBox, optional) : The oriented bounding box of the node.
                convexHull = None,          # (o3d.geometry.TrianglePointCloud, optional) : The convex hull of the node.
                loadResource = False,       # Load the resource at initialization?
                e57Index = None             # (int, optional) : index of the scan you want to import from an e57 file. Defaults to 0.
              )

Creation From E57 file

Since E57 files contain a large amount of elements, they cannot be loaded directly into a PointCloudNode. Use the geomapi.tools.e57_to_pointcloud_nodes function to load all elements into a list of PointCloudNodes.

import geomapi.tools as tl

PointCloudnodes = tl.e57_to_pointcloud_nodes(e57Path = r"../../..\tests\testfiles\pcd\lidar.e57", loadResource = True)
print(len(PointCloudnodes))

PointCount cannot be set directly when a resource is present
PointCount cannot be set directly when a resource is present
2

Creation From E57 xml file

Since xml files contain a large amount of elements, they cannot be loaded directly into a PointCloudNode. Use the geomapi.tools.e57xml_to_pointcloud_nodes function to load all elements into a list of PointCloudNodes.

import geomapi.tools as tl

PointCloudnodes = tl.e57xml_to_pointcloud_nodes(xmlPath = r"../../..\tests\testfiles\pcd\lidar.xml", loadResource = True)
print(len(PointCloudnodes))

PointCount cannot be set directly when a resource is present
PointCount cannot be set directly when a resource is present
2

PointCloudNode Resource

When creating a PointCloudNode with a resource, it can be done either directly with the resource, or with the path to the resource. The resource can be a o3d.geometry.TriangleMesh, trimesh.base.Trimesh or an ifcopenshell.entity_instance. The resource is always converted to a o3d.geometry.TriangleMesh

A resource can be a big piece of data, this is why it is not always wanted to load the whole resource at initialization. This is why the loadResource parameter is default to False

Loading The Resource

node = PointCloudNode(path=r"../../..\tests\testfiles\pcd\parking.pcd", loadResource=False)
print("resource before loading:",node.resource)
node.load_resource() # Use specialized node fo each type of resource.
print("resource after loading:",node.resource)

Resource not loaded, but path is defined, call `load_resource()` to access it.
Resource not loaded, but path is defined, call `load_resource()` to access it.
resource before loading: None
resource after loading: PointCloud with 556485 points.

Saving The Resource

A PointCloud resource can be saved to disk using the save_resource() function. Currently supports: .ply, .obj

node = PointCloudNode(path=r"../../..\tests\testfiles\pcd\parking.pcd", loadResource=True)
node.save_resource(directory=r"../../../tests/testfiles/resources", extension=".pcd") # Save the resource to the resourcePath

True

PointCloudNode Transformation

Since every nod has a cartesian transform, it can be transformed using the node.transform() function.

The transformation also updates the convexHull and orientedBoundingBox.

Furthermore, if the PointCloudNode has a resource, that resource is also transformed.

node = PointCloudNode()
print(node.cartesianTransform)
transformation = np.array([[0,0,1,0],[0,1,0,0],[1,0,0,0],[0,0,0,1]])
node.transform(transformation=transformation)
print("applying transformation: (-1)")
print(node.cartesianTransform,"\n")

node = PointCloudNode()
rotation = np.array([90,0,0]) #eulers in degrees
node.transform(rotation=rotation)
print("applying rotation: (90,0,0)")
print(node.cartesianTransform,"\n")

node = PointCloudNode()
translation = np.array([1,2,3])
node.transform(translation=translation)
print("applying translation: (1,2,3)")
print(node.cartesianTransform)

[[1. 0. 0. 0.]
 [0. 1. 0. 0.]
 [0. 0. 1. 0.]
 [0. 0. 0. 1.]]
applying transformation: (-1)
[[0. 0. 1. 0.]
 [0. 1. 0. 0.]
 [1. 0. 0. 0.]
 [0. 0. 0. 1.]] 

applying rotation: (90,0,0)
[[ 1.000000e+00  0.000000e+00  0.000000e+00  0.000000e+00]
 [ 0.000000e+00  6.123234e-17 -1.000000e+00  0.000000e+00]
 [ 0.000000e+00  1.000000e+00  6.123234e-17  0.000000e+00]
 [ 0.000000e+00  0.000000e+00  0.000000e+00  1.000000e+00]] 

applying translation: (1,2,3)
[[1. 0. 0. 1.]
 [0. 1. 0. 2.]
 [0. 0. 1. 3.]
 [0. 0. 0. 1.]]

PointCloudNode Visualisation

When a PointCloudNode has a resource, the show() function displays the PointCloud using either open3d or Trimesh, depending on the workspace.

Use the inline = True parameter to display the Mesh using the Trimesh viewer in your jupyter notebook file. Otherwise the function opens a new python window to display the open3d viewer

node = PointCloudNode(path=r"../../..\tests\testfiles\pcd\parking.pcd", loadResource=True)
node.show(inline=True) # The standard node has no resource to display

Further reading

Please refer to the full API documentation of the PointCloudNode class for more details about the functionality