liveBook · Manning

3d point cloud

As explained in Robotics for Programmers

A point cloud is a collection of three-dimensional points that may include attributes such as color in addition to position.

A 3D point cloud is a set of three-dimensional points, often carrying attributes such as color. Unlike a mesh or voxel model, it has no explicit size, surface, volume, or connectivity; these properties remain implicit. section 6.1.5

In robotics, point clouds are useful because they:

  • Capture sensor data for representing observed shapes. section 6.1.5
  • Support object detection and visualization. section 6.1.5
  • Can be converted into other representations when operations require explicit geometry. section 6.1.5

RGB-D camera view

An RGB-D camera can produce a point cloud showing the visible surface of an object. Points from the side facing away from the camera are absent, so the cloud does not describe the complete object volume.

How Robotics for Programmers covers it

How deep it goes
Defines point clouds and contrasts their implicit geometry with meshes and voxels, but defers detailed processing methods.
Examples
Illustrates point clouds with RGB-D camera observations that omit surfaces hidden from the camera.
What you can do after
Apply point clouds to sensor data, object detection, and visualization, converting them when explicit geometry is needed.
How it connects
Connects point clouds with sensors, object detection, visualization, meshes, and voxel models.

Also explained in

Compare how 2 books explain 3d point cloud →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Robotics for Programmers. The section links point to the parts of the book it draws on.