U.S. patent application number 12/827670 was filed with the patent office on 2010-12-30 for adaptive 3d scanning.
Invention is credited to Tais Clausen, Nikolaj Deichmann, Rune Fisker, Henrik Ojelund.
Application Number | 20100332196 12/827670 |
Document ID | / |
Family ID | 35107074 |
Filed Date | 2010-12-30 |
United States Patent
Application |
20100332196 |
Kind Code |
A1 |
Fisker; Rune ; et
al. |
December 30, 2010 |
ADAPTIVE 3D SCANNING
Abstract
The present invention relates to adaptive 3D scanning wherein a
scan sequence for obtaining full geometrical coverage of a physical
object are created automatically and specifically for the physical
object, by using a method and a system for producing a 3D computer
model of a physical object, wherein the method comprises the
following steps providing a scanner system, said scanner system
comprising a scanner, and a computer connectable to and/or
integrated in said scanner, said computer comprising a virtual
model of said scanner, entering shape information of the physical
object into the computer, creating in said computer a visibility
function based on said virtual model and the shape information,
said visibility function being capable of evaluating the coverage
of areas of interest of the physical object by at least one
predetermined scan sequence, establishing at least one scan
sequence based on the evaluation of the visibility function,
performing a scan of the physical object using said at least one
scan sequence, and obtaining a 3D computer model of the physical
object.
Inventors: |
Fisker; Rune; (Virum,
DK) ; Clausen; Tais; (Klagshamn, SE) ;
Deichmann; Nikolaj; (Klagshamn, SE) ; Ojelund;
Henrik; (Lyngby, DK) |
Correspondence
Address: |
GIFFORD, KRASS, SPRINKLE,ANDERSON & CITKOWSKI, P.C
PO BOX 7021
TROY
MI
48007-7021
US
|
Family ID: |
35107074 |
Appl. No.: |
12/827670 |
Filed: |
June 30, 2010 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
11572424 |
Aug 6, 2008 |
|
|
|
PCT/DK2005/000507 |
Jul 22, 2005 |
|
|
|
12827670 |
|
|
|
|
Current U.S.
Class: |
703/1 |
Current CPC
Class: |
Y10S 359/90 20130101;
G06T 17/00 20130101 |
Class at
Publication: |
703/1 |
International
Class: |
G06F 17/50 20060101
G06F017/50 |
Foreign Application Data
Date |
Code |
Application Number |
Jul 23, 2004 |
DK |
PA 2004 01143 |
Claims
1-38. (canceled)
39. A system for producing a 3D computer model of a physical object
comprising, a computer software product stored on a computer
readable medium in which computer instructions are stored, which
instructions, when executed by a computer system including a
scanner and a computer, and the computer storing a virtual model of
the scanner, cause the system to, receive into the computer shape
information of a physical object, create in the computer a
visibility function based on the virtual model and the shape
information, the visibility function operating to evaluate the
coverage of areas of interest of the physical object by at least
one predetermined scan sequence, establish at least one scan
sequence based on the evaluation of the visibility function, scan
the physical object using the at least one scan sequence; and
obtain a 3D computer model of the physical object.
40. The system of claim 39, wherein the scanner comprises a light
source, a camera, and a support for supporting the physical
object.
41. The system of claim 39, wherein the computer comprises an input
device, an central processing unit, a memory and a display, and
wherein said computer readable medium includes a data processing
system of the computer.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application is a continuation of U.S. patent
application Ser. No. 11/572,424 filed Aug. 6, 2008, which is a 371
of PCT/DK05/00507 filed Jul. 22, 2005, which claims priority of
Denmark Patent Application PA 2004 01143 filed Jul. 23, 2004.
FIELD OF THE INVENTION
[0002] The present invention relates to production of 3D computer
models of a physical object.
BACKGROUND
[0003] Structured light 3D scanner systems basically functioning
the same way have been described in the prior art. They basically
function as described in FIG. 1, wherein a monochromatic or multi
spectral light pattern 101, such as laser dots, laser lines, white
or colored strips, is projected from a light source 102 onto the
object 103. The projected light is then reflected 104 and one or
more cameras 105 acquire(s) images of the projection. The light
pattern is detected in the image and well established projection
geometry such as triangulation or stereo is used to derive the 3D
coordinates, e.g. a line laser is projected onto the object forming
a line. The 3D coordinates are then reconstructed along that
particular line. The scanner may contain one or more light
sources/patterns and one or more cameras.
[0004] The next step is then to move the object and scanner
relative to each other e.g. by rotation 106 or linear motion 107 of
the object 103. This way the scanner can reconstruct the surface on
a new part of the object, e.g. a new line on the surface in the
line laser example. The scanners in the prior art have the motion
manually programmed in a predefined scan sequence or the
object/scanner is simply manually moved around.
[0005] An inherited problem with structured light 3D scanning is
that both camera and light pattern need to "see" each surface point
at the same time to be able to make a 3D reconstruction of that
particular point. This leads to "occluded" or uncovered areas which
appear as surface holes in the final scan, i.e. areas without
surface measurement information. Holes in the scan are in most
cases undesirable or unacceptable both from a visual and
application point of view.
[0006] The problem is illustrated in FIG. 2, where the point cloud
2a of the initial scan of a toy bear is shown. The initial scan is
performed by a predefined scan sequence of two rotation scans. When
the surface model 2b is created the uncovered areas appear as holes
e.g. 204. Adaptive scanning is then used to make a scan sequence
that scans the holes in an additional scan. In fact two holes 205
have already been adaptively scanned and covered by new points.
After the first adaptive scan a single hole 206 is still present in
the surface model 2c of the merged result of the initial and
adaptive scan. A second adaptive scan is then performed and full
coverage is obtained 2d.
[0007] In the prior art the occlusion problem is attempted to be
solved by manually definition of complex scan sequences and
constraints on how the object is positioned in the scanner. However
long and time consuming scan sequences is required to cover just
simple shapes or objects with moderate shape variation. In the case
of objects with varying shapes this does still not guarantee full
coverage. Another problem is that the creation of the scan
sequences can be very cumbersome and requires expert knowledge.
[0008] To fix the problem with uncovered areas some commercial
scanners artificially close the holes in the scan using the surface
information around the hole. The artificial hole closing might be
performed by fitting parametric surface such as spline surface or
second order surfaces. Artificial hole closing might give visually
pleasant results, but the accuracy is very low, which is
unacceptable for most application.
SUMMARY OF THE INVENTION
[0009] The present invention relates to adaptive 3D scanning
wherein a scan sequence to obtain full geometrical coverage are
created automatically and specifically for the physical object.
[0010] Accordingly, the present invention relates to a method and a
system for producing a 3D computer model of a physical object,
wherein the method comprises the following steps [0011] a)
providing a seamier system, said scanner system comprising [0012]
i. a scanner, and [0013] ii. a computer connectable to and/or
integrated in said scanner, said computer comprising a virtual
model of said scanner, [0014] b) entering shape information of the
physical object into the computer, [0015] c) creating in said
computer a visibility function based on said virtual model and the
shape information, said visibility function being capable of
evaluating the coverage of areas of interest of the physical object
by at least one predetermined scan sequence, [0016] d) establishing
at least one scan sequence based on the evaluation of the
visibility function, [0017] e) performing a scan of the physical
object using said at least one scan sequence, [0018] f) optionally
repeating steps d) and e) at least once, [0019] g) obtaining a 3D
computer model of the physical object.
[0020] In another aspect the invention relates to a data processing
system for producing a 3D computer model of a physical object,
including an input device, a central processing unit, a memory, and
a display, wherein said data processing system has stored therein
data representing sequences of instructions which when executed
cause the method as defined above be performed.
[0021] In yet another, aspect the invention relates to a computer
software product containing sequences of instructions which when
executed cause the method as defined above to be performed.
[0022] In a fourth aspect the invention relates to an integrated
circuit product containing sequences of instructions which when
executed cause the method as defined above to be performed.
DRAWINGS
[0023] FIG. 1 shows a schematic drawing of a 3D scanner system
[0024] FIGS. 2a-d shows an example of a physical object to be
scanned
[0025] FIG. 3 shows a flow chart for a scanning of uncovered
areas
[0026] FIG. 4 shows a flow chart for automated creation of initial
scan
[0027] FIG. 5 shows a 3D computer model of a dental impression
DETAILED DESCRIPTION OF THE INVENTION
[0028] The invention is related to the field of 3D scanning of
physical objects, i.e. the creation of 3D computer models that in
details resemble the physical objects geometry and optional
textural, surface and material features. The invention will be
described for 3D optical structured light scanners, but can be
applied to most other types of 3D scanners, such as touch probe
scanners, laser range scanners, MR, MRI, CT, x-rays, ultra sound,
range cameras, time-of-flight sensors or optical scanners based on
silhouettes, structure and motion, shape from shading, shape from
texture or color keying.
[0029] A scan sequence defines the actual scan including the
motions, the active cameras, the active lights sources, active
light patterns and other relevant settings. A scan sequence can be
anything from a single scan without motion to scanning with long
and complex motion sequences.
[0030] For simplicity motion of the object is only mentioned in the
following text. However it should be interpreted as relative motion
where at least one component of the system and/or the object is
moved, such as wherein either object or scanner is moved. This also
covers movement of light pattern and the cameras field of view
preferably using mirrors or prisms even though the scanner is not
physically moved.
[0031] The structured light 3D scanner systems comprising a scanner
and a computer connected and/or integrated into the scanner.
Preferably the scanner contains one or more light sources 102, one
or more cameras 105 and a motion system 106-107, which is able to
move the object relative to the camera and light source. Note that
the motion system is not necessarily limited to a single rotation
106 and a single linear motion 107 as shown in FIG. 1, but might
preferably be a multi axis system with many degrees of freedom such
as a CNC machine or a robot. Camera, light source and motion system
are all connected 108 to a computer 109, which communicate with and
controls each component. The computer might be a separate unit or
integrated into the scanner.
[0032] The invention is adaptive 3D scanning where scan sequences
for obtaining full geometrical coverage are created automatically.
In other embodiments of the invention sequence files are
automatically created for optimal scanning of texture, surface
material parameters, shinny surfaces and other relevant features.
Note that coverage should be widely interpreted with respect to
scanning these features.
[0033] Shape information of step b) may be any information of the
shape of the physical object. It may be a direct shape information,
such as shape information which may e.g. originate from one or more
initial scans,
[0034] It may alternatively or additionally be indirect shape
information, such as information in the form of template shapes,
simplified or approximate shapes such as boxes and cylinders or
combinations hereof, average shapes, CAD models or be derived from
one or more 2D images. Note that the shape information does not
need to be 3D information, but may also be 1D, 2D, 4D or 5D. The
shape information may originate from other scanners than structured
light scanners preferably touch probe scanners, MR, MRI, CT, x-rays
or ultra sound. Different shape information may also be fused e.g.
by registration of the result of an initial scan to other shape
information preferably a CAD model, template or average shape. In
one embodiment shape information is a combination of direct shape
information and indirect shape information.
[0035] In one embodiment the shape information of the physical
object is a point cloud, such as wherein the shape information of
the physical object is obtained by fitting a surface to a point
cloud. The fitting may be conducted by any suitable method, such as
conducted as triangulation.
[0036] The shape information is preferably aligned to the physical
object, such as wherein the shape information and the object is
aligned by a fixation of the object in the scanner. In another
embodiment the shape information and the object is aligned by an
initial scan which is aligned to the shape information.
[0037] Adaptive 3D scanning facilitates the possibility for truly
one button scan of arbitrary objects, i.e. the user just insert an
arbitrary object in the scanner and press a single button. The
entire scanning is then performed automatically in an optimal way
and outputs a full coverage scan. No need for expert knowledge for
creation of complex scan sequences, tuning of scanning parameters
or manual stitching/merging of different scans. No scanner in the
prior art is close to achieve this goal. FIG. 3 shows one
embodiment of the invention where the shape acquisition is
performed by an initial scan 300 preferably using combination of
linear scans, rotation scans or simultaneous motion of several
axes. By a linear scan is understood a relative linear motion of
the object or a linear sweep of the light pattern over the object.
The resulting scan is then analyzed for uncovered areas and these
areas of interest are automatically rescanned using the optimal
combination of motions, light patterns, light sources and cameras.
The new adaptive scan might then be merged with the result of the
initial scan to form a full coverage scan.
[0038] By the term "full coverage" is meant the degree of coverage
defined with respect to the specific 3D computer model. Thus, it
may be accepted that not all occluded holes are covered if the
holes are of a size or localization of minor interest. It may also
be relevant to predetermine that the time for scanning is the
limiting factor, so that "full coverage" is obtained when the
scanning has been running for a predetermined period of time.
Accordingly, full coverage stop criteria may be different than
absolute coverage of all holes, the stop criteria may be holes
under a certain threshold are ignored, only a certain number of
iterations are allowed, maximal scan time or a certain overall
coverage is only needed.
[0039] The first step in the adaptive scanning might be to
determine which areas that are not properly covered in the initial
scan. Preferably this is done by creating the surface model 301
e.g. by triangulation of the point cloud or fitting of parametric
surfaces such as spline surfaces. If surface models are supposed to
be created in the post processing of the scan it is advantageous to
apply the same surface creation algorithm. When the surface model
is present full coverage can be directly evaluated 302, since
uncovered areas correspond to the holes in the surface. Uncovered
areas can also be determined directly from the point cloud, voxel
data or other raw data. Dependent on the application some holes
such as the bottom of the object might be ignored. The full
coverage stop criteria 302 might be modified to express other
priorities as described above.
[0040] Adaptively determining the combination of motions, light
patterns, light sources and cameras ensuring optimal coverage of
each area of interest is the crucial step of the adaptive scanning.
Recall that the area of interest in this embodiment corresponds to
the uncovered areas. The first step is to construct the visibility
function, i.e. given shape information and a specific configuration
of camera, light source/pattern and object motion the current
visibility of the uncovered area can be evaluated. To be able to
evaluate the visibility a virtual scanner model that replicates the
physical scanner is created preferably including camera model and
parameters, light sources/pattern model and parameters, motion
models and parameters, motion values and other relevant parameters.
The camera model might for example be a standard projective model
with parameters such as camera position and orientation. Many
parameters in the virtual scanner are specific to the scanner and
are obtained through a calibration procedure. Preferably these
parameters should be included in the virtual scanner model.
[0041] The virtual model of the scanner, ie. the virtual scanner
model, forms part of the basis for creating the visibility
function. The virtual scanner function may be created in relation
to each production of a 3D computer model, or it may be created
once in the scanner system and only re-created if changes are made
to the scanner system. The virtual scanner model may be created
before, after or simultaneous with the entrance of the shape
information of the physical object into the computer, however it is
preferred that the virtual scanner model is created before entrance
of the shape information.
[0042] The virtual scanner model is then used to evaluate the
visibility of an uncovered area given position and orientation
information of the object, camera settings, light source/pattern
settings and other relevant features. When evaluating the
visibility it might be advantageous to evaluate the visibility in a
number of sample points/cells/patches sampled over the uncovered
area. In the case of missing shape information sampling points
might be approximated e.g. by interpolation or surface fitting. The
visibility of each sample point might be calculated from the
point's degree of visibility for both light source/pattern and
camera simultaneous. For many scanner configurations the object
need to be moved while scanning an uncovered area. This motion need
to be taken into account during the visibility evaluation.
[0043] The optimal sequence of motions, light patterns, light
sources and cameras for scanning an uncovered area is then found as
the sequence that maximizes the visibility. In one embodiment the
at least one scan sequence of step d) is established by simulating
the coverage of the physical object of at least two scan sequences,
and selecting the scan sequence having an optimal coverage.
[0044] The scan sequence established in step d) is preferably
established for obtaining coverage of occluded holes in the shape
information of the physical object.
[0045] Preferably the optimal sequence is found by creation of one
or more scan sequences 303 which are then optimized. The
optimization of the visibility function 304 might be performed with
respect to any free parameter in the virtual scanner model, though
limiting the optimization to motion parameters and different
combinations of camera and light source might be advantageous. The
actual optimization might be performed by a standard optimization
algorithm such as steepest descent, conjugated gradient,
Levenberg-Marquardt, Newton or Quasi-Newton methods, BFGS,
simulated annealing or generic algorithms. Constraints on
parameters such as physical axis limits might be added as hard
constraints or soft constraints.
[0046] Additional factors might also be integrated into the
visibility function such that the function value reflects other
priorities than just visibility. Even though the visibility
function contains more factors than pure visibility the function
will still be named visibility. Additional factors might be scan
time, visibility for several cameras, reflection suppression,
surface normal to camera/light source angle or textural visibility.
It might also be advantageous to split uncovered areas in high
curvature surface to improve visibility or combined the scanning of
neighboring areas to achieve lower scan times.
[0047] Thus, in one embodiment the scan sequence is established by
optimising the scanner for best angle surface to camera and/or
laser angle. In another embodiment the scan sequence established by
optimising for speed of scan sequence. In a third embodiment the
scan sequence established is optimised for minimal reflections from
physical object, and in a fourth embodiment the scan sequence
established is optimised for visibility for more than one camera.
In further embodiments the scan sequence established is optimised
for two or more of the factors mentioned above. Thereby, a scan
sequence may be created by moving at least one of the components of
the scanner relative to at least one of the other components.
[0048] The optimal scan sequence is then selected as the one which
have the maximal visibility. Additional factors can also be used to
select the optimal configuration if several sequences achieve full
visibility. A scan is performed using the optimal scan sequence 305
and the result is merged with previously performed scans 305. The
surface model might then be created for the merged scans 301 and
the resulting scan might be checked for full coverage 302. If full
coverage is not obtained in the combined scan then a second
adaptive scan might be performed. For complex shapes several
adaptive scan sequences might be required to obtain full coverage,
because uncovered areas can contain shapes that occlude other parts
of the uncovered area.
[0049] The coverage of areas of the physical object may be
evaluated by any suitable method. In one embodiment the coverage of
areas of the physical object is evaluated by evaluating percentage
of area of holes as compared to estimated area of the physical
object. In another embodiment the coverage of areas of the physical
object is evaluated by evaluating the size of holes. In yet another
embodiment steps d) and e) are repeated until the coverage of the
physical object is above a predetermined value. Combination of the
different methods of determining coverage may also be envisaged by
the invention, for example by first evaluating the size of holes
and second evaluate the percentage of area of holes.
[0050] The scanner system according to the invention may be any
suitable scanner system, examples of these are mentioned below. In
principle the scanner system according to the invention comprises
the following components: at least one light source, at least one
camera, and at least one support for supporting the physical
object.
[0051] It is preferred that at least one of the following
components: the light source, the camera and the support is movable
relative to one of the other components, and the present invention
encompasses systems wherein two or more the components are movable
in relation to at least one of the other components. Accordingly,
examples of the invention include systems wherein the support may
be movable in relation to the light source, and/or the support may
be movable in relation to the camera, preferably the support for
supporting the physical object is capable of conducting at least
one of the following movements: a linear movement, a rotational
movement, or a combination thereof.
[0052] The scanner system according to the invention may comprise
more components, such as a the scanner system comprising at least
two light sources, and/or a scanner system comprising at least two
cameras.
[0053] For the scanner showed in FIG. 1 the simplest virtual
scanner model will contain a camera model, a light model and models
for the rotation and linear motion. To complete the virtual scanner
model calibrated values for all scanner model parameters such as
camera position for a specific scanner is inserted. Given shape
information the visibility function can now be evaluated. The only
free parameters in the scanner are the rotation and linear motion.
Any combination of rotation and linear motion can be used to scan
holes in the initial scan. However using rotation for turning the
hole for maximal visibility and linear motion to actually scan the
area is a simple and powerful configuration. In practice this
limits the free parameter to rotation since the linear motion can
be derived from the rotation angle and the size of the hole. Given
a scan sequence 303 with an initial guess of the rotation angle the
optimal scan sequence can be found by optimization the visibility
function f( . . . ) 304 with respect to the rotation angle
.theta.:
Max.sub.0f(.theta.|shape information, virtual scanner model)
[0054] The output of the optimization, .theta..sub.max, and the
corresponding sequence file ensures maximal visibility. The actual
optimization might be performed using one of the previously
mentioned optimization algorithms e.g. a steepest descent
preferably using numerical gradient evaluation with a steep size,
.DELTA..theta., on 1 degree. The next step is then to scan using
the optimal sequence file 305. The resulting scan is then merged
with previous scans 306 and the surface is created 301 and checked
for full coverage 302. If full coverage is not obtained yet a new
iteration can be performed. As discussed elsewhere herein, the term
"full coverage" means full coverage in relation to the specific 3D
computer model, and does not necessarily mean that every part of
the physical object is covered.
[0055] FIG. 4 shows another embodiment of the invention where the
virtual scanner model and the visibility function is used to create
the first initial scan 300 based on indirect shape information not
scanned directly from the object. Preferably the indirect shape
information is obtained from a CAD model 400, an average or a
template model. In this case the area of interest corresponds to
the full model.
[0056] In the same virtual scanner framework as in the previous
embodiment one or more scan sequences 401 are created and the free
parameters are optimized 402 with respect to the visibility
function. Dependent on the actual application additional factors
might be added to the visibility function such as the number of
sub-scans, visibility for mutual sub-scans and scan time. The
optimal scan sequence is then used for the initial scan of the
object 403.
[0057] The absolute positions of the object need to be known to
perform a successful scan. Preferably the position is obtained from
a fixation of the object in the scanner or more flexible by running
a fast predefined scan, which is then registered/aligned to the
shape model. If required this might then be followed by a second
scan 301-305 to close possible holes.
[0058] A simple way to speed up the scanning is to let the user
select the area of interest on some representation of the object.
If shape information is known prior to scanning the representation
is preferably a CAD model, template model or average model
otherwise the representation might be the result of a fast scan or
even faster just a 2D image captured by the cameras. Note that a
selection on a 2D image created by the scanner directly can be
transformed to a 3D selection by the use of the virtual scanner
model. This user selected area of interest is then used directly
for determining the optimal scan sequence.
[0059] Another embodiment of the invention can be applied to
scanners that capture the texture of the object where modifications
of the visibility function can ensure full coverage of texture all
over the object. The object geometry can either originate from a
scan or an indirect obtained model. The main difference to the
plain geometry scanning is the formulation of the visibility
function which now might incorporate texture features such as the
visibility of the texture and surface normal to camera angle. The
visibility function can either be formulated to incorporate
simultaneous or separate geometry and texture capture depending on
the scan strategy.
[0060] In a similar way the invention can also be applied for
capturing lightning, shading and material parameters mainly used
for rendering images of realistic visual appearance of objects.
This is implemented by extending the formulation of the visibility
function to incorporate illumination and material parameter
estimation factors. Many different illumination and material models
exist such as Phong shading, Gouraud shading or BRDF models.
However the visibility function should be general applicable to all
different models and estimation algorithms.
[0061] The 3D computer model of the physical object may be based on
one of the results obtained by the scan in step e), or the 3D
computer model of the physical object may obtained by combining
information from the any other information of the physical object
and at least one of the results of the scan performed in step e),
such as a combination with the shape information and scan
results.
[0062] The 3D computer model of the physical object may also be
obtained by combining information from at least two scan sequences
performed in step e), and by optionally combining information from
at least two scans with any other information, such as the shape
information.
[0063] Even though the invention has been described with respect to
structured light scanners it should be clear for the skilled reader
that the invention can be applied to perform adaptive scans for
other types of scanners, such as surface scanners. Thus, the
invention may also be carried out using e.g. touch probe scanners,
laser range scanners, MR, MRI, CT, x-rays, ultra sound, range
cameras, time-of-flight sensors or optical scanners based on
silhouettes, structure and motion, shape from shading, shape from
texture or color keying. The main difference is the different
formulation of the virtual scanner models and the visibility
function.
[0064] The present invention may be used for production of 3D
computer models of any physical object, and thus the adaptive
scanning is relevant in most 3D scanning applications. The physical
objects may have a regular or irregular surface, and the invention
is particular relevant for production of 3D computer models of
physical objects having an irregular surface. Examples of
applications are scanning of dental impressions, dental casts and
dies, lasts, jewelries, art, cultural and historical artifacts,
manufactured parts for quality analysis and mould making, ear
impressions, metrology, reverse engineering, easy creation of
realistic 3D models for home pages, computer games and animation of
movies, cartoons and commercials.
[0065] Optical scanning of arbitrary anatomical or dental
impressions is an application which is impossible to perform
without adaptive scanning. Dental impressions are negative
impression of the teeth and are usually made by some kind of
silicone material. Due to the shape of teeth and their biological
variation achievement of full coverage is very challenging, since
scanning down into an impression of an arbitrary tooth requires
very accurate viewing positions and motion. When the scan is used
for dental restorations very high accuracy is required to obtain a
proper fit, which rules out artificial hole closing. Thus, the
present invention is particular suitable for application in the
dental field, such as for scanning of teeth, a prosthesis, or
impressions of one or more teeth or a prosthesis.
[0066] FIG. 5 shows an adaptive scan of the upper 400 and lower
side 401 of a double sided dental impression, which simultaneous
captures both the negative teeth impressions and the bite. To form
the full scan 402 of a double sided impression the upper and lower
impression scan might be registered preferably automatically but
optionally supported by user marking of one or more corresponding
points 403 on the two scans.
[0067] In another aspect the invention relates to a scanner system
capable of performing the method as described above. Accordingly,
the invention relates to a scanner system for producing a 3D
computer model of a physical object, comprising at least one light
source, at least one camera, and at least one support for
supporting the physical object, and a data processing system
including an input device, a central processing unit, a memory, and
a display, wherein said data processing system has stored therein
data representing sequences of instructions which when executed
cause the method described above to be performed. The components of
the scanner system is as described above in relation to the
method.
[0068] In a third aspect the invention relates to a data processing
system as described above for producing a 3D computer model of a
physical object, including an input device, a central processing
unit, a memory, and a display, wherein said data processing system
has stored therein data representing sequences of instructions
which when executed cause the method of the invention to be
performed.
[0069] In a fourth embodiment the invention relates to a computer
software product containing sequences of instructions which when
executed cause the method of the invention to be performed.
[0070] In a fifth embodiment the invention relates to an integrated
circuit product containing sequences of instructions which when
executed cause the method of the invention to be performed.
* * * * *