What’s the Best Way to Capture Shape Data?
Scanning technology, hardware, and software have advanced such that more and more firms can make use of digital representations of physical objects in production workflows.
Digital shape-sampling and
processing (DSSP) adds value to all stages of design, from conception to
mass customization.
There are different data representations at different stages of the design
cycle, from physical object to CAD model. The ultimate goal is a manufactured
product (not all products require each kind of data representation).
A physical object, of course, has properties such as shape. A coordinate-
measuring machine or an articulated arm takes digital measurements
from the shape’s
outer surfaces to generate
a digital point cloud.
Point clouds are most
useful in metrology
for example, a sparse
point selection can be
replaced by richer color
error maps to show deviation
trends.
Cloud data from 3D
scanners can be converted
into surface data.
Programs including
Raindrop GeoMagic,
Innovmetric, RapidForm, and AliasStudio among others, convert scanner
point clouds into watertight polygonal meshes. When imported into
industrial-design software or 3D CAD, these meshes can, for instance,
provide close representations of the physical object for packaging development.
Meshes also allow interference checking in assemblies and support
aerodynamic and mold-flow studies so users can pinpoint needed
design changes earlier in the cycle. Meshes also generate photorealistic
images for marketing studies as well as let dispersed teams share 3D
models. Finally, meshes can build forms as Class A Nurbs surfaces.
Tools for data capture range in cost from a few thousand dollars to
hundreds of thousands. A complete device will include a registration
method, a scanning or capturing capability, and a way to extract points
into a cohesive cloud. Software that converts the cloud to a mesh eliminates
typical scan problems such as removing registration marks and
filling in holes. Depending on the scanning method, smoothing and
reduction can help make polygonal meshes easier to work with. Finally,
software that converts meshes to Nurbs might be necessary for downstream
users who lack a way to work with meshes.
Paul Deyo
Paul Deyo has over 28 years experience in transportation and industrial
design, including reverse engineering, scan processing, visualization, and
Class A surfacing. He is currently a product designer on the Alias team at
Autodesk, focusing on modeling.