U.S. patent application number 16/635841 was filed with the patent office on 2021-06-17 for method for the non-destructive inspection of an aeronautical part and system thereof.
This patent application is currently assigned to SAFRAN. The applicant listed for this patent is SAFRAN. Invention is credited to Yann LE GUILLOUX, Sylvaine PICARD.
Application Number | 20210183084 16/635841 |
Document ID | / |
Family ID | 1000005623314 |
Filed Date | 2021-06-17 |
United States Patent
Application |
20210183084 |
Kind Code |
A1 |
PICARD; Sylvaine ; et
al. |
June 17, 2021 |
METHOD FOR THE NON-DESTRUCTIVE INSPECTION OF AN AERONAUTICAL PART
AND SYSTEM THEREOF
Abstract
The invention relates to a method for the non-destructive
inspection of an aeronautical part, by means of acquiring
stereoscopic images and determining a three-dimensional model of
the part, characterised in that it is used to extinguish one or
more portions of the lighting of the part, and subsequently acquire
a stereoscopic image of the surface by each of the sensors, these
steps being performed by projecting a light on the surface by means
of at least two projectors positioned in different locations.
Inventors: |
PICARD; Sylvaine;
(MOISSY-CRAMAYEL, FR) ; LE GUILLOUX; Yann;
(MOISSY-CRAMAYEL, FR) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
SAFRAN |
Paris |
|
FR |
|
|
Assignee: |
SAFRAN
Paris
FR
|
Family ID: |
1000005623314 |
Appl. No.: |
16/635841 |
Filed: |
August 3, 2018 |
PCT Filed: |
August 3, 2018 |
PCT NO: |
PCT/FR2018/052017 |
371 Date: |
January 31, 2020 |
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G06T 7/586 20170101;
G06T 7/521 20170101; H04N 13/254 20180501; G06T 2207/10152
20130101; G06T 2207/20221 20130101; G06T 2207/10012 20130101 |
International
Class: |
G06T 7/521 20060101
G06T007/521; H04N 13/254 20060101 H04N013/254; G06T 7/586 20060101
G06T007/586 |
Foreign Application Data
Date |
Code |
Application Number |
Aug 3, 2017 |
FR |
1757490 |
Claims
1. A method (P1, P2) for non-destructive inspection of an
aeronautical part (5), by acquisition of stereoscopic images and
determination of a three-dimensional model of the part (5), the
part (5) being delimited by a surface (11), said method
implementing: a) a projection of a lighting (20) onto the surface
(11) by a first projector (3); b) an acquisition of a stereoscopic
image of the surface by a first sensor (1) and by a second sensor
(2) that are arranged in two different locations; c) a detection of
one or several specularity/specularities (10) on each of the images
of the sensors; the method being characterized in that it
implements: d) an extinction of one or several portion(s) of the
lighting (20) causing the specularity/specularities in the
direction of the sensor(s); then e) an acquisition of a
stereoscopic image of the surface by each of the sensors (1,2); the
operations a) to e) also being carried out by projecting a lighting
(20) onto the surface by a second projector (4), the second
projector being arranged at a location different from the first
projector; the method implementing a determination of the
three-dimensional model of the part (5) from the stereoscopic
images obtained during the acquisition e) under lighting of the
first projector (3) and stereoscopic images obtained during the
acquisition e) under lighting of the second projector (4), a first
three dimensional model of the part (5) being determined from the
images obtained during an acquisition e) under lighting of the
first projector (3), a second three-dimensional model of the part
(5) being determined from the images obtained during an acquisition
e) under lighting of the second projector (4), and a third
three-dimensional model of the part (5) being determined by fusion
of the first model and the second model.
2. The method (P1, P2) according to claim 1, wherein there is,
during the acquisition e), projection of a lighting (20) onto the
surface by the second projector (4) without simultaneous projection
of a lighting by the first projector (3).
3. The method (P1) according to claim 1, wherein there are, during
the determination of the three-dimensional model, fusion of the
images of the first sensor (1) that are acquired under different
lightings after extinction d), fusion of the images of the second
sensor (2) that are acquired under different lightings after
extinction d), and determination of the three-dimensional model of
the part (5) from the images thus obtained by fusion.
4. The method (P1, P2) according to claim 1, wherein there is
determination of the portion(s) of the lighting (20) to be turned
off, during the extinction d), by: projecting a light pattern (7)
onto the surface (11) by means of the first projector (3) and/or
the second projector (4); associating the image of a light pattern
on the surface (11) and the projected light pattern (7); turning
off one or several portion(s) of a projector (3, 4) that are
associated with one or several portion(s) of the image of the light
pattern (7) corresponding to one or several
specularity/specularities (10).
5. The method (P1, P2) according to claim 1, wherein there is
determination of the portion(s) of the lighting (20) to be turned
off by: projecting a sequence (15) of light patterns (7), each
light pattern of the sequence (15) comprising several portions of
binary light intensities, the sequence of the intensities of each
portion of light pattern making it possible to identify said
portion of light pattern (7); filming the surface (11) with a
sensor (1, 2) during the projection of the sequence of light
patterns (7), detecting a specularity and identifying one said
portion of the lighting to be turned off by the sequence of one
portion of the image of the sensor comprising the specularity.
6. The method (P1, P2) according to claim 1, wherein there is
inspection of the part (5) in a working space and wherein the first
projector (3) and the second projector (4) are arranged so that,
for each sensor (1, 2) and throughout the working space, the angle
.alpha. of intersection of the ellipse (13) having as foci the
first projector (3) and the sensor (1, 2) and of the ellipse (13)
having as foci the second projector (4) and the sensor (1,2), is
greater than 10.degree..
7. A system for non-destructive inspection of an aeronautical part
(5), by determination of a three-dimensional model of said part
(5), said part (5) being delimited by a surface (11) comprising a
specular portion (12), the system comprising at least a first
projector (3), a first sensor (1) and a second sensor (2) that are
arranged at two different locations, and a control unit (17),
characterized in that the system also comprises a second projector
(4) arranged at a location different from the first projector, and
in that said control unit (17) is configured to: control a lighting
(20) of the surface (11) by the first projector (3) and/or by the
second projector (4); control the acquisition of a stereoscopic
image of the surface by the first sensor (1) and/or by the second
sensor (2); detect one or several specularity/specularities (10) on
the images of the sensors; control the extinction of at least one
portion of the lighting (20) causing specularities in the direction
of the sensor(s) by the first projector and/or by the second
projector.
Description
FIELD OF THE INVENTION
[0001] The invention relates to a method for the non-destructive
inspection of an aeronautical part, by acquisition of stereoscopic
images and determination of a three-dimensional model of the part,
as well as to a system for acquiring this type of images and
inspecting such an aeronautical part.
STATE OF THE ART
[0002] The three-dimensional measurement of a surface is typically
carried out by contact. The surface of a part, fixed on a measuring
table, is traveled by a measuring head, making it possible to
acquire the spatial coordinates of the surface of the part. This
method is particularly invasive and the speed of acquisition of the
spatial data is limited by the travel time of the measuring head on
the surface.
[0003] For this purpose, it is known to acquire a three-dimensional
image of a part without contact through stereoscopy. During a
stereoscopic measurement, two images of the surface are produced by
two optical sensors at two different locations of the space. It is
thus possible to reconstruct the three-dimensional structure of the
surface by comparing the two images.
[0004] In order to facilitate the reconstruction of the surface in
three dimensions, it is also known to use a structured light
projection method. This method consists of projecting a known light
pattern onto the surface of the part to be measured, then imaging
the surface with one or several optical sensor(s). The structure of
the surface is then calculated by comparing the original pattern
with the pattern diffused by the surface, then imaged by each of
the sensors, or by comparing the imaged patterns with each other.
This method can be implemented only if the reflection of the
pattern on the surface of the measured part is a diffusive (or
Lambertian) reflection: the reflected luminance is the same in all
the directions of the half-space delimited by the surface. Thus,
light rays emitted by one point on the surface can reach all the
sensors, which makes it possible to associate a pixel of each of
the sensors with the same point on the surface.
[0005] This measurement method is not suitable for measuring
surfaces causing specular reflections, that is to say when a ray
incident on the surface is reflected along a single direction, or
more generally along a preferred direction. In this case, the
images acquired by the two sensors are not reliably and accurately
matchable. Moreover, the matchings between specular reflections
lead to an erroneous reconstruction because these reflections do
not generally correspond to the same point on the surface.
[0006] For this purpose, it is known to matify the surface before
carrying out a stereoscopic measurement thereof. The matifying of a
surface consists of depositing a powder on the surface to be
measured, the powder causing diffusive or Lambertian reflection
properties on the surface.
[0007] The deposition of the powder is long and expensive. In
addition, the thickness of the powder layer on the surface to be
measured introduces a bias in the measurement.
[0008] It is also known for this purpose to light the part with
several projectors. Sun et al. (Sun, J., Smith, M., Smith, L.,
Midha, S., & Bamber, J. (2007), Object surface recovery using a
multi-light photometric stereo technique for non-Lambertian
surfaces subject to shadows and specularities, Image and Vision
Computing, 25(7), 1050-1057) describes the determination of a
three-dimensional model of a part through stereoscopy, in which
stereoscopic images are acquired with different lightings of the
part. Thus, it is possible to acquire stereoscopic images
presenting different locations of the specularities. The
specularities can be digitally erased. However, the accuracy of
this method may be limited, and the implementation of this method
requires six lighting sources, the installation and inspection of
which can be complex.
SUMMARY OF THE INVENTION
[0009] An object of the invention is to propose a solution to be
able to acquire stereoscopic images, making it possible to
determine a three-dimensional model of an aeronautical part,
without direct mechanical contact with the surface and without
matifying step.
[0010] Particularly, one object of the invention is a method for
non-destructive inspection of an aeronautical part, by acquisition
of stereoscopic images and determination of a three-dimensional
model of the part, the part being delimited by a surface, said
method implementing:
[0011] a) a projection of a lighting onto the surface by a first
projector;
[0012] b) an acquisition of a stereoscopic image of the surface by
a first sensor and by a second sensor that are arranged in two
different locations;
[0013] c) a detection of one or several specularity/specularities
on each of the images of the sensors;
[0014] the method being characterized in that it implements:
[0015] d) an extinction of one or several portion(s) of the
lighting causing the specularity/specularities in the direction of
the sensor(s); then
[0016] e) an acquisition of a stereoscopic image of the surface by
each of the sensors;
[0017] the operations a) to e) also being carried out by projecting
a lighting onto the surface by a second projector, the second
projector being arranged at a location different from the first
projector; the method implementing a determination of the
three-dimensional model of the part from the stereoscopic images
obtained during the acquisition e) under lighting of the first
projector and stereoscopic images obtained during the acquisition
under lighting of the second projector, a first three-dimensional
model of the part being determined from the images obtained during
an acquisition e) under lighting of the first projector, a second
three-dimensional model of the part being determined from the
images obtained during an acquisition e) under lighting of the
second projector, and a third three-dimensional model of the part
being determined by fusion of the first model and the second
model.
[0018] It is understood that with such a method, it is possible to
acquire stereoscopic images and to determine a model of a part
whose surface comprises specularities, in a non-invasive manner,
without interaction with the part (differently from known
measurement methods, with a probe or a powder).
[0019] The invention is advantageously completed by the following
characteristics, taken individually or in any of their technically
possible combinations: [0020] during the acquisition e), projection
of a lighting onto the surface by the second projector without
simultaneous projection of a lighting by the first projector;
[0021] during the determination of the three-dimensional model,
fusion of the images of the first sensor that are acquired under
different lightings after extinction d), fusion of the images of
the second sensor that are acquired under different lightings after
extinction d), and determination of the three-dimensional model of
the part from the images thus obtained by fusion; [0022]
determination of the portion(s) of the lighting to be turned off,
during the extinction d), by: [0023] projecting a light pattern
onto the surface by means of the first projector and/or the second
projector; [0024] associating the image of a light pattern on the
surface and the projected light pattern; [0025] turning off one or
several portion(s) of a projector that are associated with one or
several portion(s) of the image of the light pattern corresponding
to one or several specularity/specularities; [0026] determination
of the portion(s) of the lighting to be turned off by: [0027]
projecting a sequence of light patterns, each light pattern of the
sequence comprising several portions of binary light intensities,
the sequence of the intensities of each portion of light pattern
making it possible to identify said portion of light pattern;
[0028] filming the surface with a sensor during the projection of
the sequence of light patterns, detecting a specularity and
identifying one said portion of the lighting to be turned off by
the sequence of one portion of the image of the sensor comprising
the specularity; [0029] determination of a straight line normal to
the surface at one point of a specular portion, by implementing:
[0030] inspection of the part in a working space and wherein the
first projector and the second projector are arranged so that, for
each sensor and throughout the working space, the angle .alpha. of
intersection of the ellipse having as foci the first projector and
the sensor and of the ellipse having as foci the second projector
and the sensor, is greater than 10.degree..
[0031] Another object of the invention is a system for
non-destructive inspection of an aeronautical part, by
determination of a three-dimensional model of said part, said part
being delimited by a surface comprising a specular portion, the
system comprising at least a first projector, a first sensor and a
second sensor that are arranged at two different locations, and a
control unit, characterized in that the system also comprises a
second projector arranged at a location different from the first
projector, and in that said control unit is configured to: [0032]
control a lighting of the surface by the first projector and/or by
the second projector; [0033] control the imaging of the surface by
the first sensor and/or by the second sensor; [0034] detect one or
several specularity/specularities on the images of the sensors;
[0035] control the extinction of at least one portion of the
lighting causing specularities in the direction of the sensor(s) by
the first projector and/or by the second projector.
PRESENTATION OF THE FIGURES
[0036] Other characteristics and advantages will also emerge from
the following description, which is purely illustrative and
non-limiting, and should be read in relation to the appended
figures, among which:
[0037] FIG. 1 illustrates a method for acquiring stereoscopic
images and determining a three-dimensional model of a part;
[0038] FIG. 2 illustrates a method for acquiring stereoscopic
images and determining a three-dimensional model of a part;
[0039] FIGS. 3 to 7 illustrate steps of a method for acquiring a
stereoscopic image of a part;
[0040] FIG. 8 schematically illustrates a light pattern in the form
of a sequence;
[0041] FIG. 9 schematically illustrates the surface of a part, a
sensor and a projector;
[0042] FIG. 10 schematically illustrates the surface of a part, a
sensor and two projectors.
DEFINITIONS
[0043] The term "specular" refers to the ability of a surface to
reflect an incident light ray along a preferred direction and more
specifically a substantially unique direction, along the half-space
delimited by the surface. In other words, the light of an incident
ray is not or hardly diffused by a surface: a specular reflection
is different from a diffuse or Lambertian reflection.
[0044] The term "specularity" refers to the specular reflection of
a surface, at one point. It is directed in the preferred direction
of reflection of the incident light ray.
[0045] The term "image fusion" refers to an image processing,
taking into account the information coming from several input
images, and producing one data or a set of data comprising more
information than the input images considered individually.
[0046] The term "angle of intersection of two conics", particularly
"of two ellipses", at one point common to the two conics, refers to
the minimum angle formed by the straight lines tangent to the
conics at this point.
DESCRIPTION OF THE INVENTION
[0047] FIG. 1 illustrates a method P1 for non-destructive
inspection of an aeronautical part 5 and for determination of a
three-dimensional model of the part 5. In one embodiment of the
invention, the surface 11 of the considered part 5 comprises a
specular portion 12. However, the acquisition of stereoscopic
images, and the determination of a three-dimensional model of a
part 5 is possible in the absence of specular reflection on the
surface 11 of the part 5.
[0048] During a step 101 (illustrated in FIG. 3), the user can
project a lighting 20 onto the surface 11 of a part 5, with a first
projector 3. The lighting can advantageously implement the
projection of a light pattern 7, or be a light pattern 7.
[0049] During step 102 (illustrated in FIG. 3), the acquisition of
a stereoscopic image of the surface 11 of the part 5 is carried out
by a first sensor 1 and by a second sensor 2. The sensors 1, 2 are
arranged at different locations in the space. The sensors 1, 2 used
can be standard photographic sensors, for example of the CCD, CMOS
type, industrial cameras, or any other device forming a resolved
image of the surface 11 observed.
[0050] An image 8 of a first sensor 1 is illustrated on the left of
FIG. 3 and an image 9 of a second sensor 2 is illustrated on the
right of FIG. 3. On each of these images, a specularity 10 is
imaged. The two specularities 10 are illustrated by gray points.
The specularity 10 of each of the images is different: in the image
8 of the first sensor 1, the specularity 10 observed corresponds to
a specular portion 12 comprised in the left portion of the part 5
observed, and in the image 9 of the second sensor 2, the
specularity 10 observed corresponds to a specular portion 12
comprised in the right portion of the part 5 observed.
[0051] During a step 103, the user and/or a control unit 17 detect
the specularity/specularities 10 in each of the images of the
sensors 1, 2 obtained during step 102 or 202. The specularities 10
can be detected in the image by the presence of local saturations
of one or several neighboring pixels in the image. In general, the
specularities 10 can be detected in post-processing, for example by
segmentation of the images, by thresholding of the gray or color
levels. The location of the specularities 10 in the image 8, 9 is
directly dependent on the relative positions of the projector, of
the surface 11 of the part 5 and of the sensor 1, 2.
[0052] During a step 104, the user or a lighting control unit 17
can hide and/or turn off the portion(s) of the lighting 20 causing
one or several specularity/specularities 10 on the surface 11 of
the part 5 in the direction of the sensors 1, 2. FIG. 4 illustrates
a first projector 3 projecting a lighting 20 onto the surface 11 of
the part 5, in which the incident light rays, represented by arrows
from the first projector 3 to the surface 11, are selectively
turned off. A control unit 17 can be electrically connected to the
first sensor 1 and to the second sensor 2 so as to load the data
from the sensors 1, 2 towards the control unit 17. The portion(s)
of the lighting projected by the first projector 3 causing the
specularity/specularities 10 is/are calculated based on the images
of the surface 11 by the sensors 1, 2. The control unit 17,
electrically connected to the first projector 3, can thus inspect
the operating state of one portion of the projected lighting, for
example, by hiding or turning off the pixels corresponding to the
specularities 10.
[0053] Thus, the specularities 10 of the surface 11 can be turned
off. The localized extinction of the projector causes one or
several local shadow areas 18 on the surface 11 in place of one or
several portion(s) of the lighting 20. In this way, the surface can
be lighted, at least partly, without causing specularity 10.
[0054] During a step 105 of the method (illustrated in FIG. 4), the
acquisition of a stereoscopic image of the surface 11 of the part 5
is carried out by a first sensor 1 and by a second sensor 2. FIG. 4
illustrates an image 8 of a first sensor 1 and an image 9 of a
second sensor 2, each comprising two shadow areas 18 on the surface
11. In the image 8 of the first sensor 1, on the left of FIG. 4,
the shadow area 18 on the left corresponds to a turned off
specularity 10, illustrated by the arrows on the right starting
from the first projector 3, passing through the surface 11 up to
the second sensor 2. The shadow area on the right corresponds to a
turned off specularity 10 illustrated by the arrows on the left
starting from the first projector 3, passing through the surface 11
up to the first sensor 1.
[0055] During steps 111, 112, 113, 114 and 115, steps 101, 102,
103, 104 and 105 are carried out by projecting in step 111 a
lighting onto the surface 11 by a second projector 4, possibly
without the first projector 3. The second projector 4 is arranged
at a location different from the first projector 3, and can thus
cause specularities 10 at other locations on the surface 11, or not
cause specularity.
[0056] FIG. 5 illustrates a system comprising two projectors. The
control unit 17, electrically connected to the second projector 4,
can inspect the operating state of one portion of the projected
lighting, for example by hiding or turning off the pixels
corresponding to the specifics 10.
[0057] At the top left of FIG. 5, an image 8 of a first sensor 1,
produced during a step 101, is illustrated. At the bottom left of
FIG. 5, an image 8 of a first sensor 1, produced during a step 111,
is illustrated. At the top right of FIG. 5, an image 9 of a second
sensor 2, produced during a step 101, is illustrated. At the bottom
right of the image 4, an image 9 of a second sensor 2, produced
during a step 111, is illustrated.
[0058] On each of the four images 8, 9 of sensors 1, 2, illustrated
in FIG. 5, different specularities 10 are imaged. Two specularities
10, caused by the first projector 3, are identical to the
specularities 10 illustrated in FIG. 1. Two other specularities 10,
caused by the second projector 4, are illustrated in the images 8,
9 of the first sensor 1 and of the second sensor 2, respectively at
the bottom left and at the bottom right of FIG. 5.
[0059] During steps 104 and 114, the specularities 10 are turned
off by inspecting the first projector 3 and/or the second projector
4 with the control unit 17. Two images 8 of a first sensor 1 are
illustrated in FIG. 6. The image at the top left corresponds to the
image 8 of a first sensor 1 when the surface 11 is lighted by a
first projector 3 and when the specularities 10 caused by the first
projector 3 are turned off during a step 104. The image at the
bottom left corresponds to the image 8 of a first sensor 1 when the
surface 11 is lighted by a second projector 4 and when the
specularities 10 caused by the second projector 4 are turned off
during a step 114. The image on the top right corresponds to the
image 9 of a second sensor 2 when the surface 11 is lighted by a
first projector 3 and when the specularities 10 caused by the first
projector 3 are turned off during a step 104. The image at the
bottom right corresponds to the image 9 of a second sensor 2 when
the surface 11 is lighted by a second projector 4 and when the
specularities 10 caused by the first projector 3 are turned off
during a step 114.
[0060] During a step 106 of the method (illustrated in FIG. 7), the
images 8 of the first sensor 1 are fused into one image 19, and the
images 9 of the second sensor 2 are also fused into one image 19.
The fusion of several images of the same sensor 1, 2 that are
produced during the projection of a lighting 20 by different
projectors, makes it possible to eliminate the shadow areas 18 in a
fused image. The fusion may comprise a step of selecting the pixel
having the highest value between two same pixels of two images, for
each of the pixels of the images.
[0061] Step 106 is illustrated on the left of FIG. 7 by the fusion
of an image 8 of the first sensor 1 during the projection of a
lighting 20 by the first projector 3 and of an image 8 of the first
sensor 1 during the projection of a lighting 20 by the second
projector 4. The image resulting from the fusion does not include a
shadow area. Step 106 is also illustrated on the right of FIG. 7 by
the fusion of an image 9 of the second sensor 2 during the
projection of a lighting 20 by the first projector 3 and of an
image 9 of the second sensor 2 during the projection of a lighting
20 by the second projector 4. The image resulting from the fusion
does not include a shadow area.
[0062] Two images are then obtained after the image fusion step
106: an image without shadow area or specularity, whose information
comes from the first sensor 1 and an image without shadow area, or
specularity, whose information comes from the second sensor 2.
These two images form a stereoscopic pair without shadows or
specularities.
[0063] During step 107 of the method, a three-dimensional model of
the surface 11 is deduced from the stereoscopic pair. It is
possible to use a known stereoscopy method by using the two images
19 obtained during the fusion step 106. These images are
particularly suitable for use in stereoscopy because they do not
include any shadow areas or specularities.
[0064] FIG. 2 illustrates a method P2 for non-destructive
inspection of an aeronautical part 5 and for determination of a
three-dimensional model of the part 5. Steps 201 to 205 and 211 to
215 of the method P2 are respectively identical to steps 101 to 105
and 111 to 115 of the method P1.
[0065] During step 206 and 216, a first three-dimensional model of
the part 5 is determined from the images acquired during a step 205
(from the images obtained during an acquisition under lighting of
the first projector), and during step 216, a second
three-dimensional model of the part 5 is determined from the images
acquired during a step 215 (from the images obtained during an
acquisition under lighting of the first projector).
[0066] Some areas of the surface 11 not being lighted, the images
acquired by each of the sensors 1, 2 represent only partially the
surface 11. During step 206, a first three-dimensional model of the
part is determined from a pair of stereoscopic images acquired
during step 205. This model can be determined by the control unit
17. In this model, information corresponding to the portions of the
surface 11 that are not lighted by the first projector during step
205 is missing. During step 216, a second three-dimensional model
of the part 5 is determined from a pair of stereoscopic images
acquired during step 215. This model can be determined by the
control unit 17. In this model, information corresponding to the
portions of the surface 11 that are not lighted by the second
projector in step 215, is missing.
[0067] During step 207, the first model and the second model
obtained during steps 206 and 216, are fused so as to determine a
third three-dimensional model of the part 5, more complete than the
first and second models, comprising the information relating to the
three-dimensional structure of the part 5 in the portions of the
surface 11 that are not lighted during steps 205 and/or 215.
[0068] The specularities are turned off during steps 104, 114, 204
and/or 214. The portion(s) of the lighting 20 to be turned off
during these steps are determined from the acquired images
comprising specularities. It is possible to implement the
projection of a light pattern 7 onto the part 5 during steps 102,
112, 202 and/or 212. It is for example possible to project a light
pattern 7 onto the surface 11 by means of the first projector 3
and/or of the second projector 4 during the acquisition of an image
by one of the sensors 1, 2. It is then possible to associate the
image of the light pattern 7 on the surface 11 and the projected
light pattern 7. Thus, it is possible to associate a pixel of the
sensor(s) 1, 2 with a pixel of the projector(s) 3, 4. It is then
possible to turn off one or several portion(s) of a projector 3, 4,
for example a set of pixels of a projector 3, 4, associated with
one or several portion(s) of the image of the light pattern 7
corresponding to one or several specularity/specularities 10.
[0069] FIG. 8 schematically illustrates a sequence 22 of light
patterns 7. During step 104 and 114 of the method, the portion(s)
of the lighting 20 causing one or several specularity/specularities
10 is/are hidden and/or turned off. The operating state (for
example, turned off or turned on state) of one or several
portion(s) of the projected lighting is calculated as a function of
the images of the surface 11, produced by the sensors 1, 2. A user
and/or the control unit 17 can determine the portion(s) of the
lighting to be turned off and/or to be hidden by projecting a
sequence 22 of light patterns 7. FIG. 8 illustrates a sequence of
three light patterns 7, projected successively. The different
portions 16 of a light pattern 7 correspond to the black or white
columns of the pattern 7 at the bottom of the figure: eight
portions 16 are represented. For each pattern 7 of the sequence,
the light intensity of a projected portion 16 is binary,
schematically represented in FIG. 8 by a black or white filling of
each portion 16. In this example, each portion 16 of the pattern is
binary coded along the sequence from 000 (portion completely on the
left of the patterns 7 of the sequence) to 111 (portion completely
on the right of the image of the sequence). In general, the
projected sequence is different for each portion 16, which makes it
possible to identify each of the portions 16. The binary coding
presented here is given by way of example. There are various
codings with interesting properties depending on the goal.
[0070] The surface 11 can be filmed by one or several sensor(s) 1,
2 during the projection of a sequence of light patterns 7. A user
or a control unit 17 can determine a portion of the lighting to be
turned off and/or to be hidden by detecting a specularity, then by
identifying the sequence emitted by a portion of the filmed image
comprising the specularity. The sequence of this portion of the
image can be read and translated by the control unit 17 so as to
identify the portion of the lighting to be turned off and/or
hidden. The portions 16 can for example be pixels or pixel groups
of a projector.
[0071] FIG. 9 schematically illustrates an installation comprising
the surface 11 of a part 5, a sensor 1, 2 and a projector 3, 4.
[0072] During a specular reflection directed towards a sensor 1, 2,
the straight line normal to the surface 11 at the point causing the
specular reflection is aligned with the bisector of the angle
formed by the direction of the incident ray and that of the
reflected ray (in application of the Snell-Descartes law).
[0073] Furthermore, the normal straight line at one point of an
ellipse 13 coincides with the bisector of the angle formed by a
focus of the ellipse 13, said point and the second focus of the
ellipse 13.
[0074] Thanks to this property, it is knows how to establish a
quite simple criterion determining whether a surface element 11 can
produce a specular reflection for a given projector-sensor
configuration: P.sub.1 and P2 being the respective optical centers
of the projector and of the sensor, M being a point on surface 11,
a specular reflection is possible if the ellipse having as foci
P.sub.1 and P2 and passing through M is tangent to the surface 11
at M. FIG. 9 illustrates, in a plane comprising the optical centers
P.sub.1 and P2, the only points on the surface suitable for
presenting a specular reflection. These are the points for which
the local ellipse is tangent to the surface.
[0075] FIG. 10 illustrates a system comprising a sensor 1 and
several projectors 3, 4. In general, the second projector 4 is
arranged in the space so that no point on the surface 11 has a
specular reflection caused by both the first projector 3 and the
second projector 4.
[0076] Thus, if the lighting 20 of the first projector 3 causes on
the sensor 1 a specular reflection at one point on the surface 11,
the portion of the surface 11 at this point is tangent to the
ellipse 13 having as foci the center of the projector 3 and the
center of the sensor 1. This ellipse therefore forms a sufficient
angle, at this point, with the ellipse having as foci the projector
4 and the center of the sensor 1, and the lighting 20 of the second
projector 4 therefore cannot cause specular reflection on the
sensor 1.
[0077] When two sensors 1 and 2 are implemented, the projectors 3,
4 and the sensors 1, 2 can be arranged so that, for each sensor,
the angle .alpha. of intersection of the ellipse 13 having as foci
the first projector 3 and the sensor and of the ellipse 13 having
as foci the second projector 4 and the sensor, is sufficient, that
is to say greater than 10.degree., throughout the working space in
which the part 5 is inspected. Thus, no area of the working space
is likely to cause, on a sensor 1, 2, two specular reflections when
it is successively lighted by the projector 3 and by the projector
4. During the inspection of a part 5, any portion of the part is
therefore lighted by at least one of the projectors, despite the
required extinctions.
* * * * *