PointCloudData<T>

Models & Types · AiDotNet.PointCloud.Data

Represents a point cloud data structure with coordinates and optional features.

How It Works

For Beginners: This class stores 3D point cloud data in an organized way.

A point cloud consists of:

  • Points: XYZ coordinates representing positions in 3D space
  • Features: Optional additional information per point (color, intensity, normals, etc.)
  • Labels: Optional category or class information for each point

Think of it as a spreadsheet where:

  • Each row is a point
  • First 3 columns are X, Y, Z coordinates
  • Additional columns can store colors, surface properties, etc.
  • Another column can store what category each point belongs to

This structure makes it easy to:

  • Load point cloud data from sensors or files
  • Pass data to neural networks for processing
  • Store results from classification or segmentation

Constructors

Constructor Summary
PointCloudData(Tensor<T> points, Vector<T> labels) Initializes a new instance of the PointCloudData class.

Properties

Property Summary
Tensor<T> Points { get; set; } Gets or sets the tensor containing point coordinates and features.
Vector<T> Labels { get; set; } Gets or sets optional labels for classification or segmentation tasks.
int NumFeatures { get; set; } Gets or sets the number of features per point (including XYZ coordinates).
int NumPoints { get; set; } Gets or sets the number of points in the cloud.

Methods

Method Summary
PointCloudData<T> FromCoordinates(Matrix<T> coordinates, Vector<T> labels) Creates a point cloud from coordinates only (no additional features).
Tensor<T> GetCoordinates() Extracts only the XYZ coordinates from the point cloud.
Tensor<T> GetFeatures() Extracts additional features (excluding XYZ coordinates).