Jeff's MCAD Blogging
Jeffrey Rowe has more than 40 years of experience in all aspects of industrial design, mechanical engineering, and manufacturing. On the publishing side, he has written well over 1,000 articles for CAD, CAM, CAE, and other technical publications, as well as consulting in many capacities in the … More »
Point Clouds Moving Beyond Surface Representation To 3D Modeling
January 25th, 2018 by Jeff Rowe
For quite some time, I’ve been saying to my peers and detractors that 2018 might finally be the year the “cloud” takes off, whether we’re talking storage in the cloud, CAD in the cloud, simulation in the cloud – whatever in the cloud. I also think that another “cloud” innovation that will get its just attention is the point cloud that will grow far beyond its traditional role for representing surfaces to becoming an integral component of 3D modeling and maybe even virtual reality.
So, exactly what is a point cloud? Technically, a point cloud is a data base containing points in a 3D coordinate system. A point cloud is a very accurate digital record of an object or space. It is saved as a (very) large number of points that cover surfaces of a scanned/sensed object. The points in a point clouds are always located on the external surfaces of visible objects, because they are the points reflected from scanned objects.
In a three-dimensional coordinate system, these points are usually defined by X, Y, and Z coordinates, and intended to represent the external surface of an object.
Point clouds can be created by several methods, including 3D scanners and photographs. These devices measure a large number of points on an object’s surface, and often output a point cloud as a data file. The point cloud represents the set of points that the device has measured.
Reality Capture – Converting Photos Into 3D Models
The key factor in acquiring point cloud data is the access/visibility to scanned surfaces. It is important to remember, that point clouds are created with visible access to real objects. Regardless of the method of acquisition (scanner or photos). It is impossible to obtain points on the surfaces that are not visible from the position from which data is collected. This means that to cover entire objects, many scanning positions must be combined.
Point cloud density describes resolution on the collected dataset this usually means the distance from point to point. Less dense point clouds are obviously much quicker to capture.
As the output of 3D scanning processes, point clouds are used for many purposes, including to create 3D CAD models for manufactured parts, metrology/quality inspection, and a multitude of visualization, animation, rendering, and mass customization applications.
While point clouds can be directly rendered and inspected, point clouds themselves are generally not directly usable in most 3D applications, and therefore are usually converted to polygon mesh or triangle mesh models, NURBS surface models, or CAD models through a process known as surface reconstruction.
There are many techniques for converting a point cloud to a 3D surface. Some approaches, like Delaunay triangulation, alpha shapes, and ball pivoting, build a network of triangles over the existing vertices of the point cloud, while other approaches convert the point cloud into a volumetric distance field and reconstruct the defined implicit surface defined by an algorithm known as marching cubes.
A Point Cloud (a) Is Triangulated (b), Meshed (c), and Rendered (d) As A 3D Model
One application in which point clouds are directly usable is industrial metrology or inspection using industrial computed tomography (CT). The point cloud of a manufactured part can be aligned to a CAD model (or even another point cloud) and compared to check for differences. These differences can be displayed as color maps that give a visual indicator of the deviation between the manufactured part and the CAD model. Geometric dimensions and tolerances (GD&T) can also be extracted directly from the point cloud.
It is important to understand that the point cloud is a set of individual, unrelated points with defined position and color. This makes point clouds relatively easy to edit, filter, and display.
The Future of Point Clouds
The future of point clouds is not so much about hardware (scanners, cameras, and computers). Rather, the bigger issue and emphasis is on point cloud software.
Like just about every other technical software, point cloud processing software has been getting a face lift with modern User Interface (UI) and User Experience (UX) design so that it can be used by mere mortals, and not just specialists – analogous with what has been happening with simulation/analysis software.
Paul Tice, 3D scanning consultant and specialist with 20 years of experience in the field, has predicted that the evolution of 3D scanning could one day make 3D modeling obsolete. Tice is the CEO of ToPa 3D visualization and design company and his article published on LinkedIn, is quite interesting overall, but a statement referring to a point-cloud “baking” process used by Australian 3D imagers Euclideon is especially interesting.
In registration software, which is used to combine individual puzzle pieces (scans) together into a complete picture, developments in this software include noise cleaning, which sounds basic but because of this, point cloud data firstly gets reduced in size with un-necessary data through so-called point skipping where data has been removed enabling far less false positives within automatic point cloud recognition software.
Newly developed software also has greater functionality enabling re-registering data against itself, what this means is the ability to fine tune the point cloud overlap by finding similarities in the point cloud and closing the gap with its corresponding scan, giving far greater accuracy enabling a better picture of what is happening with a scanned object.
The future in point cloud data won’t be solely the collection of data, but with the processing of that data, software enabling feature-based extraction, as well as software enabling the use of 3D models in asset management and documentation.
So, if 2018 is really the year of the cloud, and although it’s different than other traditionally regarded aspects, I think we need to include point clouds as part of the “cloud” conversation.
Meet Us At SOLIDWORKS World 2018!
We’ll be at SOLIDWORKS World 2018 in Los Angeles, February 4-7, 2018 conducting video interviews. If your company is interested and you haven’t signed up yet, click on this link to schedule a video interview. If you have any questions, contact me at 719-221-1867 or firstname.lastname@example.org. Stop by and say “hello” during the conference in exhibit booth #302. Hope to see you there!