3D Scanning - PowerPoint PPT Presentation

1 / 35
About This Presentation
Title:

3D Scanning

Description:

3D Scanning How can we get the form? Capturing design intents Generators Fillet radius change Knuckles Scanning 3D Scanning is the process of getting a set of surface ... – PowerPoint PPT presentation

Number of Views:64
Avg rating:3.0/5.0
Slides: 36
Provided by: tdsIcPol
Category:

less

Transcript and Presenter's Notes

Title: 3D Scanning


1
3D Scanning
2
How can we get the form?
  • Projection

3
Capturing design intents
  • Generators
  • Fillet radius change
  • Knuckles

4
Scanning
  • 3D Scanning is the process of getting a set of
    surface points from a physical object
  • Touch probe, optical, laser ranging, etc.
  • Usually the points are structured

5
3D Scanning equipment
  • Contact type
  • Touch probe scanner
  • CMM
  • Non-contact type
  • Laser scanner
  • Optical scanner

6
Touch probe scanning
7
Co-ordinate measurement
8
Laser range scanning
9
Optical scanning
10
Application range
  • From very small (tip of a drill bit) to very
    large (the Lee Ka Shing tower)
  • Rigid or soft
  • Coloured, glossy, transparent
  • Restrictions Smoke, hair, cloud, very fast
    moving objects, etc.

11
Data captured
  • X Y Z location of points
  • Freshly scanned data sets are usually structured,
    as a array in XY plane with different Z values
  • May become un-structured, un-even spaced in all
    directions after re-sampling
  • Usually stored as an ASCII text file

12
(No Transcript)
13
Point cloud
  • The collection of points is called point clouds
  • A point cloud define the SURFACE of an object
  • Some scanner (like CT and MRI) can measure
    physical properties of a objects entire solid
    mass, the collection of points is not point
    cloud, it is called VOXEL

14
(No Transcript)
15
Level of details (LOD)
  • LOD is determined by point density, which is
    determined during scanning
  • LOD is proportional to file size
  • Scanning itself is not selective, the result file
    size of a flat surface is the same as a carved
    board

16
Shape consideration
  • Optical scanner capture a 2D array of points each
    time
  • Steep surface will have less points cast on it,
    thus fewer details

17
3D Optical scanning
  • Take pictures of the object
  • Output a set of 3D points

18
Shape VS position
  • Point clouds gives only position information
  • The shape of the object, if there is any, is not
    defined in the point cloud data
  • Human can infer the shape from the clouds easily,
    but it is very difficult for machines

19
Scan registration and merge
20
Scanning and stitching
  • Scan the object, coat the surface if necessary
  • If the scan result cannot cover entire surface,
    scan from other angles
  • Use point cloud manipulation software to align
    and merge scan results

21
Re-sampling
  • Re-sampling is the process to reduce the number
    of points or polygons so that the file size
    become manageable
  • Adaptive re-sampling reduce more data points on
    smooth surface and keep more in regions that have
    abrupt changes

22
Alignment
  • Points are only x,y,z values

23
Positioning and scaling
  • Data from optical scanner is not necessary
    correct in size
  • Positioning of point cloud is not necessary
    suitable for surface patching
  • Alignment may need to be adjusted

24
Tessellation of Point cloud
25
Triangulation
  • The process of converting point clouds to polygon
    models
  • Triangulation is computation intensive
  • If two points are close together, they are most
    probably on one surface
  • If there are more that one possible shape, the
    shape that form a CONVEX HULL are more likely the
    correct choice

26
Multiple solution
27
Create curve network
28
Feature curves
  • Feature curves are curves that define distinctive
    visual features
  • Sharp edges
  • Fillets
  • Holes
  • 3D scanning cannot capture feature curves
    directly
  • Irregularities may be corrected

29
Repair model
30
Surface modelling
  • Generate surface patches using feature curves as
    guide lines
  • Lofting is easier but may generate surfaces with
    unwanted undulation
  • Sweeping and ruling generate fair surfaces but
    the feature curves are more difficult to draw
  • Irregularities may be corrected

31
NURBS surface creation
32
Surface recreation
  • Cut Sections by user defined plane, and then
    create surface according to these sections
  • Draw 3D feature curves directly on the point
    cloud
  • Full auto generation

33
Steps to retrieve CAD from foam model
  • Scanning and stitching
  • Re-sampling
  • Automatic surfacing or manual surfacing
  • For manual surfacing
  • Positioning and scaling
  • Draw feature curves
  • Draw Surface patches
  • Check deviation

34
Application
35
Related applications
  • Specialised applications
  • Surfacer
  • Rapid Form
  • Raindrop Geomagic
  • Scanning modules
  • Pro/E, UG, CATIA, ALIAS
Write a Comment
User Comments (0)
About PowerShow.com