Point Cloud

3D Scanners typically create these aptly named point clouds.  With the continual innovation in computer capacity and capability we now have measurement solutions of a variety of configurations and design which can capture millions of discrete measurements in minutes, thus point cloud, point-cloud, or even pointcloud.  Apparently they are all used but 'point cloud' is the most used and correct designation.

Aligned and merged point clouds of a Ferrari 550 using Geomagic Studio

While a typical CMM or portable metrology solution such as the ROMER or FARO ARM capture tens or

hundreds of points, measurement solutions with 3D laser scanner technology justify the term point cloud with thousands, and even millions of measurements on the surface of an object.  The output created is a point cloud and can be found in a variety of data formats the simplest is basic XYZ Ascii data in a CSV (comma separated value) format.  In this way the surface of an object and its actual surface conditions are accurately modeled thanks to the point cloud created.

Color point cloud of a house

Post processing of point cloud data

A point cloud can be a useful tool by itself, however generally its important to use software to interpret the data and intelligently convert it into a triangulated mesh model, NURBS surface model, or CAD model.  This process of reverse engineering is very useful and, prior to the advent of the 3D laser scanner, was time consuming.

Post processing the point cloud data serves two major purposes; to simplify the data/information and to give the objects, surfaces, or environment intelligence.  Multiple point cloud data sets can take up hundreds of megabytes of computer storage and processing power.  Thus it is encouraged to either mesh the point cloud or simply intelligently compress the data so that maximum resolution is retained where it is needed and in planar regions (which could technically be defined by just 3 points) redundant data eliminated.  Usually compression of the point cloud is lossy, but of course it is also voluntary.

Techniques to convert a point cloud to a polygon mesh
Delaunay triangulation
Marching triangles
the Ball-Pivoting algorithm

Applications using point cloud data

Corvette racing team by Pratt & Miller tuned and optimized using point clouds

Among a plethora of applications, point cloud data is used to create 3D CAD models of manufactured parts, Metrology/quality inspection, reverse engineering, mass customization, visualization, animation, and rendering.

Industrial Metrology or inspection are applications which directly use the point cloud as an essential data tool. The point cloud of a manufactured part is aligned to a CAD model (or even another point cloud), compared to design or other products from the production run, and then a 3D inspection report is generated.  Polyworks, Geomagic, and Rapidform are three highly capable software packages which interface with the 3D Laser Scanner and generate this information.  Any differences can be displayed as color maps or deviation report which provides a visual indicator of the difference between the manufactured part and the CAD model. Geometric dimensions and tolerances can also be extracted from the point cloud data.

Medical imaging equipment also utilize point clouds to represent volumetric data.


ADD IMAGES!!!

MORE TO COME