Method, device and computer program for dimensional control of a real part
Patent Information
- Application Number
- FR2022005695
- Authority / Receiving Office
- FR · FR
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2022-06-13
- Publication Date
- 2025-03-14
- Estimated Expiration
- 2042-06-13
Smart Images
Figure 00000018_0000 
Figure 00000019_0000 
Figure 00000019_0001
Abstract
Description
Title of the invention: Method, device and computer program for dimensional control of a real part
[0001] The invention relates to a method, a device and a computer program for the dimensional control of a real part.
[0002] The field of the invention relates to parts made of a ceramic matrix composite material (CMC for short), particularly in the aeronautical field.
[0003] To access civil applications, CMC materials must have stability of properties over very long periods (30,000 to 100,000 hours), with admissible stresses of an acceptable level (120 to 150 MPa). To meet this specification, we seek to redesign the CMC material so that it operates in an undamaged domain. This principle requires a material with a high elastic limit.
[0004] A known manufacturing process for CMC parts involves successive operations starting with weaving sheets of ceramic wires into a preform, followed by molding the preform into a molded blank in a former, then consolidation to set the molded blank, which is long and expensive and is followed by subsequent manufacturing operations.
[0005] In order to optimize costs during the weaving operation, a short weaving / molding / control loop on a few preform copies is carried out in order to validate by control the weaving of the preform but also the preform / former suitability in order to obtain the correct dimensions of the raw material. The technical field of the invention relates more particularly to the step of controlling this short loop, carried out by tomography in the former.
[0006] A disadvantage is that, in the case of short loop tomography, the conformer is present and is found in the tomographic images produced of the preform molded in the conformer. However, the gray levels between conformer and preform are too close to be separated via simple thresholding of the tomographic images and will disturb the dimensional measurements of the molded preform, which can be carried out by automated processing of the tomographic images. It is therefore sought to remove the conformer from the tomographic images of the molded preform by an automated process.
[0007] An objective of the invention is to obtain a method, a device and a computer program for the dimensional control of a real part, which solve the problem mentioned above.
[0008] To this end, a first object of the invention is a dimensional control method of a real part, located in an interior space of a conformer used for the manufacture of the real part, from a three-dimensional reference representation of the real part, the method comprising the following steps: obtaining by tomography at least a first three-dimensional tomographic representation of the real part in the conformer, the first three-dimensional tomographic representation of the real part in the conformer having brightness levels in the real part and in the conformer, characterized in that the method further comprises the following steps, carried out by at least one computer: obtaining a second three-dimensional tomographic representation of the conformer, which is a negative of the first three-dimensional tomographic representation of the actual part in the conformer, and obtaining a second negative three-dimensional reference representation of the conformer, obtaining a first roto-translation matrix making it possible to move from a first reference frame of the second three-dimensional reference representation of the conformer to a second reference frame of the second three-dimensional tomographic representation of the conformer, registration of the first three-dimensional tomographic representation of the real part in the conformer by the first roto-translation matrix applied to the second reference frame, thresholding of the first three-dimensional tomographic representation of the real part in the conformer, having been registered, according to a first threshold, to obtain a second corrected three-dimensional tomographic representation of the real part, using a uniformization mask putting first voxels thereof, located outside the interior space of the conformer, at a black level, which is uniform and predefined, and keeping the brightness levels in the interior space, processing of the second corrected three-dimensional tomographic representation of the real part and of the target three-dimensional representation of the real part, to provide dimensional deviations of the real part with respect to the target three-dimensional representation.
[0009] By means of the invention, the image of the conformer is removed from the second corrected three-dimensional tomographic representation by automated steps performed by a computer. This avoids a user having to manually manipulate the tomographic representations to remove the image of the conformer each time, which is tedious for a user. Thus, by means of the invention, the correction of the tomographic dimensional representation to remove the image of the conformer does not depend on the operator and is therefore more rigorous than manual removal.
[0010] According to one embodiment of the invention, the treatment comprises: a second thresholding of the second corrected three-dimensional tomographic representation of the real part according to a second threshold, to detect edges of the real part, a uniformization of a region of the actual part located between the edges by voxels having a uniform and prescribed brightness level, which is greater than the black level, providing a third three-dimensional tomographic representation of the actual part, a second recalibration of the third three-dimensional tomographic representation of the real part with respect to the reference three-dimensional representation of the real part according to a second prescribed roto-translation matrix, to obtain a fourth three-dimensional tomographic representation of the real part, recalibrated in a third reference frame of the reference three-dimensional representation of the real part, a calculation of the dimensional deviations between points of the fourth three-dimensional tomographic representation and the three-dimensional reference representation.
[0011] According to one embodiment of the invention, the second threshold is less than or equal to the first threshold.
[0012] According to another embodiment of the invention, the second threshold is greater than or equal to the first threshold.
[0013] According to one embodiment of the invention, the second roto-translation matrix is equal to the first roto-translation matrix.
[0014] According to an embodiment of the invention, the second roto-translation matrix is calculated by an inertial pre-registration configured for making a center of gravity, having been calculated, of the third three-dimensional tomographic representation of the real part coincide with a center of gravity, having been calculated, of the three-dimensional reference representation of the real part, making the calculated axes of inertia of the third three-dimensional tomographic representation of the real part coincide with the calculated axes of inertia of the setpoint three-dimensional representation of the real part.
[0015] A second object of the invention is a device for dimensional control of a real part, located in an interior space of a conformer used for the manufacture of the real part, from a three-dimensional reference representation of the real part, the dimensional control device comprising a tomographic imaging device for obtaining by tomography at least a first three-dimensional tomographic representation of the real part in the conformer, the first three-dimensional tomographic representation of the real part in the conformer having brightness levels in the real part and in the conformer, characterized in that the method further comprises at least one calculator configured to carry out the following steps: obtaining a second three-dimensional tomographic representation of the conformer, which is a negative of the first three-dimensional tomographic representation of the actual part in the conformer, and obtaining a second negative three-dimensional reference representation of the conformer, obtaining a first roto-translation matrix making it possible to move from a first reference frame of the second three-dimensional reference representation of the conformer to a second reference frame of the second three-dimensional tomographic representation of the conformer, recalibration of the first three-dimensional tomographic representation of the real part in the conformer by the first roto-translation matrix in the second reference frame, thresholding the first three-dimensional tomographic representation of the real part in the conformer, having been recalibrated, according to a first threshold, to obtain a second corrected three-dimensional tomographic representation of the real part, using a uniformization mask putting first voxels thereof, located outside the interior space of the conformer, at a black level, which is uniform and predefined, and keeping the brightness levels in the interior space, processing the second corrected three-dimensional tomographic representation of the real part and the target three-dimensional representation of the real part, to provide dimensional deviations of the real part compared to the target three-dimensional representation.
[0016] A third subject of the invention is a computer program comprising code instructions for implementing the following steps of the method for dimensional control of a real part, as described above, when executed on at least one computer: obtaining a second three-dimensional tomographic representation of the conformer, which is a negative of the first three-dimensional tomographic representation of the actual part in the conformer, and obtaining a second negative three-dimensional target representation of the conformer, obtaining a first roto-translation matrix making it possible to move from a first reference point of the second three-dimensional setpoint representation of the conformer to a second reference point of the second three-dimensional tomographic representation of the conformer, registration of the first three-dimensional tomographic representation of the real part in the conformer by the first roto-translation matrix in the second marker, thresholding the first three-dimensional tomographic representation of the real part in the conformer, having been recalibrated, according to a first threshold, to obtain a second corrected three-dimensional tomographic representation of the real part, using a uniformization mask putting first voxels thereof, located outside the interior space of the conformer, at a black level, which is uniform and predefined, and keeping the brightness levels in the interior space, processing the second corrected three-dimensional tomographic representation of the real part and the target three-dimensional representation of the real part, to provide dimensional deviations of the real part compared to the target three-dimensional representation.
[0017] The invention will be better understood on reading the description which follows, given solely by way of non-limiting example with reference to the figures below of the attached drawings.
[0018] [Fig. 1] represents a flowchart of the dimensional control method of a real part according to an embodiment of the invention.
[0019] [Fig.2] represents a tomographic image of a real part and its conformer.
[0020] [Fig.3] represents a schematic view of a device for dimensional control of a real part according to an embodiment of the invention.
[0021] [Fig.4] represents a negative reference image of a vacuum conformer without the actual part.
[0022] [Fig.5] schematically represents a roto-translation matrix between a second three-dimensional reference representation of the conformer and the second three-dimensional tomographic representation of the conformer according to an embodiment of the invention.
[0023] [Fig.6] schematically represents the superposition of a negative three-dimensional tomographic representation of the conformer and a negative three-dimensional reference representation of the conformer, which can be used to calculate a roto-translation matrix used according to an embodiment of the invention.
[0024] [Fig.7] schematically represents a second corrected three-dimensional tomographic representation of the real part according to an embodiment of the invention.
[0025] [Fig.8] schematically represents an image of a reconstructed exterior surface of the actual part according to an embodiment of the invention.
[0026] [Fig.9] schematically represents a fourth three-dimensional tomographic representation of the real part P superimposed on the three-dimensional reference representation of the real part according to an embodiment of the invention.
[0027] Different steps of the method, of the device 100 and of the computer program, for the dimensional control of a real part P according to embodiments of the invention are described below, with reference to [Fig.l].
[0028] The actual part P was produced by an operation of weaving sheets of ceramic wires into a preform, then molding this preform in a CONF former or CONF mold (operations called short loop, following which dimensional control is carried out on the actual part P, before carrying out the subsequent manufacturing operations). The actual part P mentioned below is therefore formed in the CONF former and is therefore located in an interior space INTCONF of the CONF former used for the manufacturing of the actual part P. This actual part P is then subjected to other manufacturing operations to produce the finished actual part P.
[0029] The finished manufacturing actual part P is made of a ceramic matrix composite. The finished manufacturing part P may be a part forming part of an aircraft turbomachine, in particular an airplane. The ceramic matrix composite part may be, for example, a high-pressure turbine ring sector of the turbomachine, a blade of the turbomachine, in particular a low-pressure turbine blade, an air distributor of the turbomachine, a combustion chamber of the turbomachine, or others. By way of non-limiting example, in FIGS. 2, 4, 6, 7, 8 and 9, the actual part P is a high-pressure turbine ring sector of a turbomachine.
[0030] The method according to the invention carries out a dimensional check of the real part P from a three-dimensional XCao representation of the real part P. The dimensional check is a non-destructive check of the real part P.
[0031] During a first step E1, one (or more) first three-dimensional tomographic representations Xtomo of the real part P are obtained by tomography in the presence of the conformer CONF. The first three-dimensional tomographic representation Xtomo of the real part P in the presence of the conformer has different brightness levels in the real part P and in the conformer CONF. The first three-dimensional tomographic representation Xtomo is also called short-loop tomography. Generally speaking, the brightness levels can be gray levels and can include a color or not. Thus, the brightness levels can range from black to white, via gray levels or brightness levels of a color other than gray.
[0032] During the first step E1, the first three-dimensional tomographic representation Xtomo of the real part P is acquired by a tomographic imaging device 10. For example, the first three-dimensional tomographic representation Xtomo of [Fig. 2] is obtained. This three-dimensional tomographic representation Xtomo can be, for example, recorded in a permanent memory 17 of a computer 16. Each Xtomo tomographic three-dimensional representation can be taken in one or more image planes passing through the ceramic matrix composite part and distinct from each other (parallel to each other or not parallel to each other). Each Xtomo tomographic three-dimensional representation is made up of voxels and can be made up of several two-dimensional images taken in image planes passing through the ceramic matrix composite part, which are distinct and neighboring each other (parallel to each other or not parallel to each other).
[0033] The first step E1 can be carried out for example in the following manner. The conformer CONF, in the interior space INTCONF of which the real part P is located, is introduced into the tomographic imaging device 10, shown by way of example in [Fig. 3]. The measuring device 11 can comprise an X-ray cabin 14 into which the conformer CONF and the real part P are introduced. The tomographic imaging device 10 captures by one (or more) sensors 12 in the cabin 14 during the step E1 one (or more) images of the real part P. The measuring device 11 can be a tomographic or tomograph imaging device, which can for example be a transmission X-ray device, and makes it possible to obtain images of the representation Xtomo of the real part P by the X-ray sensor 12, an X-ray source 13 being provided in the cabin 14.The tomographic imaging device 10 can acquire a large number of radiographs during a rotation, most of the time complete, of the real part P and the conformer CONF in the cabin 14 (the real part P and the conformer CONF being arranged on a rotating support 15 in the cabin 14), with the aim of calculating by the computer 16 connected to the sensor 12 the complete three-dimensional tomographic representation Xtomo of the real part P.
[0034] During a step E01 subsequent to step E1, a second three-dimensional tomographic representation CONFNEGtomo of the conformer CONF is calculated by a calculator CAL, which is a negative of the first three-dimensional tomographic representation Xtomo. This second three-dimensional tomographic representation CONFNEGtomo makes it possible to make the real part P disappear from the first three-dimensional tomographic representation Xtomo and to visualize the holes in the conformer CONF. Thus, the brightness levels of the voxels of the second three-dimensional tomographic representation CONFNEGtomo of the conformer CONF are complementary to the brightness levels of the voxels of the first three-dimensional tomographic representation Xtomo of the real part P in the conformer CONF.During step E01, the computer calculates a second three-dimensional representation CONFNEGcao of the negative setpoint of the empty CONF conformer without the actual part P. An example of the second three-dimensional representation CONFNEGcao of the negative setpoint of the conformer at . empty without the actual part P is shown in [Fig.4].
[0035] During a step E02 subsequent to step E01, the computer CAL calculates a first roto-translation matrix M making it possible to move from a first reference frame RCao of the second three-dimensional representation CONFNEGcao of the CONF conformator to a second reference frame RConf of the second negative three-dimensional tomographic representation CONFNEGtomo of the CONF conformator, as shown diagrammatically in [Fig. 5]. As an example, [Fig. 6] shows that to calculate the first roto-translation matrix M, the computer CAL actually realigns the captured drillings of the CONF conformator of the second negative three-dimensional tomographic representation CONFNEGtomo with drillings of the CONF conformator of the second negative three-dimensional representation CONFNEGcao of the CONF conformator.This strategy has the advantage of requiring registration with clouds having a lower number of points than taking all the points of the tomography and of being able to threshold the drillings in the conformer more easily than the reverse.
[0036] During a step E03 subsequent to step E02, the first roto-translation matrix M is applied by the computer CAL to the first three-dimensional tomographic representation Xtomo of the real part P in the conformer CONF to re-align this first three-dimensional tomographic representation Xtomo in the second reference frame Rconf of the conformer CONF. This first three-dimensional tomographic representation Xtomo of the real part P having been re-aligned in the second reference frame Rconf of the conformer CONF is designated by Xtomo rec in [Fig.2].
[0037] During a step E04 subsequent to step E03, the computer CAL performs a first automatic thresholding of the first three-dimensional tomographic representation Xtomo rec of the real part P in the conformer CONF, having been realigned by the first roto-translation matrix M in the second reference frame Rconf of the conformer CONF, according to a first threshold SI, this first thresholding providing the second corrected three-dimensional tomographic representation X2tomo of the real part P. This first thresholding is configured to obtain, from the first realigned three-dimensional tomographic representation Xtomo rec, contours of the conformer CONF in the second corrected three-dimensional tomographic representation X2tomo, these contours delimiting the boundary of the interior space INTCONF of the conformer CONF.This thresholding is calculated by the CAL calculator using a uniformization mask MS, which places the first VI voxels of the second corrected three-dimensional tomographic representation X2tomo, located outside the interior space INTCONF of the CONF conformer, at the same predefined and uniform black level NN. This calculation of the uniformization mask MS can be carried out by filling the VI voxels located outside the contours of the with the black level NN. CONF conformer. The MS uniformization mask has the same resolution as the first three-dimensional tomographic representation Xtomo. The first threshold S1 can be a 1 inside / 0 outside thresholding according to a given position in the image (thanks to registration).
[0038] This second corrected three-dimensional tomographic representation X2tomo of the real part P has the predefined black level NN in its first voxels VI, located outside the interior space INTCONF, and keeps the brightness levels of the first three-dimensional tomographic representation Xtomo in the interior space INTCONF. We thus obtain the second corrected three-dimensional tomographic representation X2tomo, where only the real part P is present and where the conformer CONF has been removed.The voxel uniformization mask MS can be applied by the fact that the CAL calculator applies a Boolean operation "AND" between the voxel uniformization mask MS and the initial tomographic representation X tomo, the black level NN mentioned above corresponding to a zero brightness level value, the voxel uniformization mask MS can have for example a brightness value equal to one (or equal to the maximum non-zero brightness) in the interior space INTCONF of the conformer CONF. An example of this second corrected three-dimensional tomographic representation X2tomo of the real part P is shown in [Fig.7]. In order not to obscure [Fig.7], the black level NN outside the interior space INTCONF is shown in white.
[0039] After step E05, the second corrected three-dimensional tomographic representation X2tomo of the real part P and the target three-dimensional X CAD representation of the real part P are processed by the CAL computer, for example according to steps E2, E3, E4, E5 described below. The processing carried out by the CAL computer provides dimensional deviations of the real part P with respect to the target three-dimensional representation XCAo.
[0040] According to an embodiment of the invention, the processing carried out by the CAL computer comprises a step E2, during which the CAL computer carries out a second automatic thresholding of the second corrected three-dimensional tomographic representation X2tomo of the real part P according to a second threshold S2, to detect edges PEXT of the real part P in this second corrected three-dimensional tomographic representation X2tomo. These edges P|X1 form the outer surface PEXT (or skin) of the real part P. The outer surface PEXT of the real part P is formed by a cloud of points in the second corrected three-dimensional tomographic representation X2tomo. Several types of automatic thresholding are possible, for example of the Otsu type, or of the minimum type, or of the Shanbhag type, or others. The CAL computer constructs the outer surface PEXT of the real part P from the second thresholding.The CAL calculator generates for example as surface . exterior surface PEXT a skin model of the real part P, by a computational algorithm according to Lorensen, William E.; Cline, Harvey E., August 1, 1987, "Marching cubes: A high resolution 3D surface construction algorithm", in ACM SIGGRAPH Computer Graphics 21 (4): 163-169. [Fig. 8] shows an example of the exterior surface PEXT of the real part P. The second threshold is an automatic grayscale detection to detect part edges.
[0041] During step E3, subsequent to step E2, the computer CAL performs, in the second corrected three-dimensional tomographic representation X2tomo, having been thresholded, a uniformization of one or more regions of the real part P located between the edges P|-XT by voxels having the same uniform and prescribed brightness level NGINT, which is greater than the black level NN. This uniformization provides a third three-dimensional tomographic representation X3tomo of the real part P. This third three-dimensional tomographic representation X3tomo of the real part P has the same brightness level NGINT in its second voxels V2, located in the interior space INTCONF.
[0042] During step E4, subsequent to step E3, the computer CAL performs a second recalibration of the third three-dimensional tomographic representation X3 tomo of the real part P to superimpose it on the three-dimensional representation XCao of the real part P according to a second prescribed roto-translation matrix M2, to obtain a fourth three-dimensional tomographic representation X4tomo of the real part P, recalibrated in a third reference frame R3Cao of the three-dimensional representation XCao of the real part P. An example of a fourth three-dimensional tomographic representation X4tomo of the real part P superimposed on the three-dimensional representation XCao of the real part P is shown in [Fig.9].The third reference point R3Cao of the three-dimensional representation XCao of the actual part P may be equal to the first reference point RCao of the three-dimensional representation CONFCao of the conformator CONF in a first embodiment of the invention, or be different from the latter in other embodiments of the invention.
[0043] During step E5, subsequent to step E4, the CAL calculator calculates the dimensional deviations Ex4_CAo between points of the fourth three-dimensional tomographic representation X4tomo and the reference three-dimensional representation XCao. For example, during step E5, the CAL calculator creates a file (for example a point cloud of type .stl) of the fourth three-dimensional tomographic representation X4tomo to provide on an input lines of computer code (for example of the python type), analyzing the deviations. The lines of computer code make it possible to return, from the file, figures representing the dimensional deviations Ex4.cao. The dimensional deviations Ex4_Cao allow a user to decide on the conformity of the actual part P according to criteria.
[0044] The processing by steps E2, E3, E4, E5 makes it possible to avoid two pitfalls mentioned below in the state of the art. A first pitfall of the state of the art is that there is no way to compare the short loop measurements with production measurements, while the parts all come from a conformer setting and this would make it possible to validate the short loops on the basis of the production data. A second pitfall of the state of the art is having to manually manipulate the tomography to obtain information: this makes the study variable, operator dependent and tedious. The invention makes it possible to generate source data for carrying out quantitative dimensional checks of the open loop process on the parts P in CMC and to be able to compare it with the production data.The processing by steps E2, E3, E4, E5 makes it possible to set up an automatic process for processing the tomography to find the same information as that from the production cycle. The invention makes it possible to move away from a qualitative, manual study of validation of the open loops of manufacturing of parts P in CMC. The processing makes it possible to extract from the tomography a point cloud representing the surface of part P in the reference frame of the three-dimensional Xcao representation of the part P. The processing of the tomographies of the short loop is thus automated and provides input data compatible with a quantitative analysis of the conformity of this manufacturing phase.
[0045] According to an embodiment of the invention, the second threshold S2 may be less than or equal to the first threshold SL.
[0046] According to another embodiment of the invention, the second threshold S2 may be greater than or equal to the first threshold SL.
[0047] According to the first embodiment of the invention, the second roto-translation matrix M2 is equal to the first roto-translation matrix M.
[0048] According to a second embodiment of the invention, the second roto-translation matrix M2 is calculated by the calculator CAL during step E4 by an inertial pre-registration configured for make coincide a center of gravity, having been calculated, of the third three-dimensional tomographic representation X3tomo of the real part P and a center of gravity, having been calculated, of the three-dimensional representation XCao of the target of the real part P, to make coincide the calculated axes of inertia of the third three-dimensional tomographic representation X3tomo of the real part P and the calculated axes of inertia of the three-dimensional representation XCao of the target of the real part P. According to a third embodiment of the invention, the second matrix M2 of roto-translation is calculated by the CAL computer during step E4 by an inertial pre-registration or optimized adjustment (in English: best fit), configured to perform at least one minimization of a mean square distance, for example by gradient descent, between the third three-dimensional tomographic representation X3tomo of the real part P and the three-dimensional representation XCao of reference of the real part P. The CAL computer performs for example two successive minimizations of the mean square distance, for example by gradient descent, with different filtering parameters between the two minimizations (with for example 20% use of the point cloud in the first minimization, then 70% use of the point cloud in the second minimization).
[0049] Of course, the third embodiment of the invention can be carried out without the second embodiment or in addition to the second embodiment.
[0050] Of course, in other embodiments, the processing by steps E2, E3, E4, E5 could be provided without the prior steps E01, E02, E03 and E04.
[0051] The invention also relates to a device 100 for dimensional control of the real part P, comprising the tomographic imaging device 10 and the CAL computer(s), which are configured to implement some or all of the steps mentioned above.
[0052] The invention also relates to a computer program comprising code instructions for implementing the steps E02, E03, E04 mentioned above and the above-mentioned processing of the dimensional control method of the actual part P, when the method is executed on one (or more) CAL computers. The computer program is recorded in a permanent memory of the dimensional control device 100, namely in the memory MEM of the CAL computer.
[0053] One or more computers 12 and / or one or more computers and / or one or more processors and / or one or more computer programs may be provided as a CAL calculator.
[0054] Of course, the embodiments, features, possibilities and examples described above can be combined with each other or selected independently of each other.
Claims
Claims
1. Method for dimensional control of a real part (P), located in an interior space (INTCONF) of a conformer (CONF) used for manufacturing the real part (P), from a three-dimensional representation (XCao) of the real part (P), the method comprising the following steps: obtaining (El) by tomography at least one first three-dimensional tomographic representation (Xtomo) of the real part (P) in the conformer (CONF), the first three-dimensional tomographic representation (Xtomo) of the real part (P) in the conformer (CONF) having brightness levels in the real part (P) and in the conformer (CONF), characterized in that the method further comprises the following steps, carried out by at least one computer (CAL): obtaining (E01) a second three-dimensional tomographic representation (CONFNEGtomo) of the conformer (CONF), which is a negative of the first three-dimensional tomographic representation (Xtomo) of the real part (P) in the conformer (CONF), and obtaining (E01) a second three-dimensional representation (CONFNEGcao) of the negative reference of the conformer (CONF), obtaining (E02) a first roto-translation matrix (M) making it possible to move from a first reference frame (RCao) of the second three-dimensional representation (CONFNEGcao) of the reference of the conformer (CONF) to a second reference frame (Rconf) of the second three-dimensional tomographic representation (CONFNEGtomo) of the conformer (CONF), resetting (E03) the first three-dimensional tomographic representation (Xtomo) of the real part (P) in the conformer (CONF) by the first roto-translation matrix (M) applied to the second reference frame (Rconf),thresholding (E04) of the first three-dimensional tomographic representation (Xtomo rec) of the real part (P) in the conformer (CONF), having been readjusted, according to a first threshold (SI), to obtain a second three-dimensional tomographic representation (X2tomo) corrected of the real part (P), using a uniformization mask (MS) placing the first voxels (VI) of the latter, located outside the interior space (INTCONF) of the conformer (CONF), at a black level (NN), which is uniform and predefined, and keeping the brightness levels in the interior space (INTCONF), processing (E2, E3, E4, E5) of the second corrected three-dimensional tomographic representation (X2tomo) of the real part (P) and of the three-dimensional representation (XCao) of the target of the real part (P), to provide dimensional deviations of the real part (P) with respect to the three-dimensional representation (XCao) of the target.
2. Method according to claim 1, characterized in that the processing (E2, E3, E4, E5) comprises: a second thresholding (E2) of the second corrected three-dimensional tomographic representation (X2tomo) of the real part (P) according to a second threshold (S2), to detect there edges (PEXt) of the real part (P), a standardization (E3) of a region of the real part (P) located between the edges (PEXt) by voxels having a uniform and prescribed brightness level (NGINT), which is greater than the black level (NN), providing a third three-dimensional tomographic representation (X3tomo) of the real part (P), a second registration (E4) of the third three-dimensional tomographic representation (X3tomo) of the real part (P) with respect to the set three-dimensional representation (XCao) of the real part (P) according to a second prescribed roto-translation matrix (M2),to obtain a fourth three-dimensional tomographic representation (X4tomo) of the real part (P), recalibrated in a third reference frame of the three-dimensional representation (XCao) of the real part (P), a calculation of the dimensional deviations (Ex4.Cao) between points of the fourth three-dimensional tomographic representation (X4tomo) and the three-dimensional representation (XCao).,
3. Method according to claim 2, characterized in that the second threshold (S2) is less than or equal to the first threshold (SI).
4. Method according to claim 2, characterized in that the second threshold (S2) is greater than or equal to the first threshold (SI).
5. Method according to any one of claims 2 to 4, characterized in that the second roto-translation matrix (M2) is equal to the first roto-translation matrix (M).
6. Method according to any one of claims 2 to 4, characterized in that the second roto-translation matrix (M2) is calculated by an inertial pre-registration configured to
7. making a calculated center of gravity of the third three-dimensional tomographic representation (X3tomo) of the real part (P) coincide with a calculated center of gravity of the setpoint three-dimensional representation (XCAo) of the real part (P), making axes of inertia of the third three-dimensional tomographic representation (X3tomo) of the real part (P) coincide with axes of inertia of the setpoint three-dimensional representation (XCAo) of the real part (P). Device for dimensional control of a real part (P), located in an interior space (INTCONF) of a conformer (CONF) used for the manufacture of the real part (P), from a three-dimensional representation (XCao) of the real part (P), the dimensional control device comprising a tomographic imaging device (10) making it possible to obtain (El) by tomography at least a first three-dimensional tomographic representation (Xtomo) of the real part (P) in the conformer, the first three-dimensional tomographic representation (Xtomo) of the real part (P) in the conformer (CONF) having brightness levels in the real part (P) and in the conformer (CONF), characterized in that the method further comprises at least one computer (CAL) configured to carry out the following steps: obtaining (E01) a second three-dimensional tomographic representation (CONFNEGtomo) of the conformer (CONF), which is a negative of the first three-dimensional tomographic representation (Xtomo) of the real part (P) in the conformer (CONF), and obtaining (E01) a second three-dimensional representation (CONFNEGcao) of the negative setpoint of the conformer (CONF),obtaining (E02) a first roto-translation matrix (M) making it possible to move from a first reference point (RCao) of the second three-dimensional representation (CONFNEGcao) of the setpoint of the conformer (CONF) to a second reference point (RConf) of the second three-dimensional tomographic representation (CONFNEGtomo) of the conformer (CONF), recalibration (E03) of the first three-dimensional tomographic representation (Xtomo) of the real part (P) in the conformer (CONF) by the first roto-translation matrix (M) in the second reference frame (R confX
8. thresholding (E04) of the first three-dimensional tomographic representation (Xtomo rec) of the real part (P) in the conformer (CONF), having been readjusted, according to a first threshold (SI), to obtain a second three-dimensional tomographic representation (X2tomo) corrected of the real part (P), using a uniformization mask (MS) placing first voxels (VI) thereof, located outside the interior space (INTCONF) of the conformer (CONF), at a level (NN) of black, which is uniform and predefined, and keeping the brightness levels in the interior space (INTCONF), processing (E2, E3, E4, E5) of the second three-dimensional tomographic representation (X2tomo) corrected of the real part (P) and of the three-dimensional representation (XCao) of the real part (P), to provide dimensional deviations of the real part (P) compared to the three-dimensional representation (XCao) of instruction.Computer program comprising code instructions for implementing the following steps of the method for dimensional control of a real part (P) according to any one of claims 1 to 6, when executed on at least one computer (CAL): obtaining (E01) a second three-dimensional tomographic representation (CONFNEGtomo) of the conformer (CONF), which is a negative of a first three-dimensional tomographic representation (Xtomo) of the real part (P) in a conformer (CONF), and obtaining (E01) a second three-dimensional representation (CONFNEGcao) of the negative setpoint of the conformer (CONF). obtaining (E02) a first roto-translation matrix (M) making it possible to move from a first reference point (RCao) of the second three-dimensional representation (CONFNEGcao) of the setpoint of the conformer (CONF) to a second reference point (RConf) of the second three-dimensional tomographic representation (CONFNEGtomo) of the conformer (CONF), recalibration (E03) of the first three-dimensional tomographic representation (Xtomo) of the real part (P) in the conformer (CONF) by the first roto-translation matrix (M) in the second reference frame (R confX thresholding (E04) of the first three-dimensional tomographic representation (Xtomo rec) of the real part (P) in the conformer (CONF), having been readjusted, according to a first threshold (SI), to obtain a second three-dimensional tomographic representation (X2tomo) corrected of the real part (P), using a uniformization mask (MS) putting the first voxels (VI) of the latter, located outside the interior space (INTCONF) of the conformer (CONF), at a level (NN) of black, which is uniform and predefined, and keeping the brightness levels in the interior space (INTCONF), processing (E2, E3, E4, E5) of the second three-dimensional tomographic representation (X2tomo) corrected of the real part (P) and of the three-dimensional representation (XCao) of the real part (P), to provide dimensional deviations of the real part (P) compared to the three-dimensional representation (XCao) of the reference.