Method of characterizing a woven fiber structure

Abstract

The invention relates to a method of characterizing a fibrous structure from a volume image, said fibrous structure having a three-dimensional weaving between a plurality of warp yarns extending along a first direction and a plurality of weft yarns extending along a second direction perpendicular to the first direction, said method comprising: a first processing (E10) of the volume image by filtering along a third direction perpendicular to the first and second directions to attenuate a periodic pattern along the third direction; obtaining (E20) a two-dimensional image corresponding to an intermediate plane along the third direction of the filtered volume image; a second processing (E31, E41) of the two-dimensional image by filtering along the first or second direction to attenuate the periodic pattern; obtaining (E32, E33) a one-dimensional profile representative of the position of a warp or weft column and corresponding to an intermediate line along the first or second direction of the filtered two-dimensional image; and comparing (E33, E43) the one-dimensional profile with a reference profile.

Description

Technical Field

[0001] This invention pertains to the design, characterization, and monitoring of industrial components, particularly those subjected to significant mechanical stress, such as aircraft engine components. More specifically, this invention relates to the characterization of components with woven fiber structures or made of composite materials, including fibrous structures as fiber reinforcements densified by a matrix. Background Technology

[0002] Most braided fiber reinforcements used to manufacture parts made of composite materials are first provided in a planar shape at the loom exit. The resulting preform must then be matched to the shape of a die, in which it is densified, for example, by injection and polymerization of resin. The forming of the preform and the injection of the preform (or preform) significantly alter the arrangement of the yarns or plies within the preform. For example, these alterations can include movements corresponding to modifications of the initial 90° angle between the warp and weft yarns, or slippage of warp or weft rows relative to each other. These changes in the reinforcement affect the mechanical strength of the part and its behavior under stress, demonstrating the need for accurate characterization in order to optimize manufacturing methods.

[0003] X-ray computed tomography (CT) is a known technique. This experimental method utilizes the differential absorption of X-rays by different materials to reconstruct a three-dimensional image of the component under study from a series of X-ray images through calculation. The information contained in the tomographic images is valuable because it involves the entire volume of the component, allowing us to see not only its microstructure but also, and potentially, its defects.

[0004] When attempting to study the displacement of yarns or strands in a fiber preform obtained through three-dimensional weaving, a fiber preform obtained by molding such a preform, or a composite component that includes such a preform as a fiber reinforcement, problems may arise as illustrated by the following examples.

[0005] Figure 1 An aero-turbine engine blade 1 is shown, comprising fiber reinforcements densified from a matrix. Such a blade 1 can have a double bend, also known as a "twist," along its longitudinal length L. To obtain such a shape, the initial planar blank must be formed in a suitable tooling or directly in an injection mold.

[0006] Figure 2 This is a simplified cross-sectional view of the blade fiber preform 2, obtained by three-dimensional weaving of longitudinal yarns (extending along the longitudinal direction L) or warp yarns and transverse yarns (extending along the transverse direction T) or weft yarns, which has been shaped to obtain the characteristic curvature of the blade 1. After the forming step, the expected simplified orientation of several rows of warp yarns is identified by line 3.

[0007] Figure 3 This is a cross-sectional X-ray tomographic image of the fiber preform 4, produced by forming a fiber blank through three-dimensional weaving. In this image, the warp yarns are arranged in rows... Figure 2 The deformed line 5 is observed in the image. This deformation (geometric transformation) is caused by the manufacturing process, manifesting as the original warp and weft planes becoming curved surfaces.

[0008] Figure 4 Two examples of continuous simplified weave planes 6 and 7, used to create a woven fiber structure with interlocking, are shown. The warp is denoted by c, and the weft by t. It can be seen that the weave patterns are very close together, typically distinguishable by a simple phase shift between two adjacent planes. The weave patterns are actually periodic or nearly periodic within the woven structure.

[0009] Due to this deformation and the minimal difference between the two continuous weave planes, it is difficult, if not impossible, to imagine a perpendicular to this. Figure 3 The planar extension of the weave plane controls the weave quality and infers any defects related to yarn or ply displacement. Therefore, it is desirable to be able to identify the transformations occurring during forming so that the obtained image can be straightened. In this context, the term "straightening" corresponds to the re-transformation of the geometry of the woven fabric in a simplified state.

[0010] In this example, the identification of these displacements was done manually through the analysis of tomographic images.

[0011] Therefore, a more robust characterization method is needed when studying fiber structures obtained through three-dimensional weaving. Summary of the Invention

[0012] To this end, the present invention proposes a method for characterizing fiber structures based on volumetric images, wherein the fiber structure has a three-dimensional weave according to a given pattern, the given pattern being located between a plurality of warp yarns or warp plies extending along a first direction and a plurality of weft yarns or weft plies extending along a second direction perpendicular to the first direction, the method comprising:

[0013] The volumetric image undergoes a first processing step by filtering along a third direction perpendicular to the first and second directions to attenuate the periodic pattern along the third direction.

[0014] Obtain a two-dimensional image representing warp yarns or warp strands and weft yarns or weft strands, the image corresponding to a third-direction intermediate plane along the filtered volume image.

[0015] The two-dimensional image is subjected to a second processing by filtering along the first or second direction to attenuate the periodic pattern along the first or second direction.

[0016] Obtain a one-dimensional profile representing the position of warp or weft yarns or ply yarns, the profile corresponding to the midline along a first or second direction of a filtered two-dimensional image, and

[0017] The one-dimensional profile is compared with a reference profile.

[0018] "Three-dimensional weaving," "3D weaving," and "multi-layer weaving" refer here to a weaving pattern in which at least some warp yarns (or longitudinal yarns) are combined with weft yarns (or transverse yarns) on several layers of weft yarns. This weaving can be performed on a jacquard type loom in a manner known per se. In one exemplary embodiment, a given weaving pattern can be interlocked. "Interlocking" refers to a three-dimensional weaving in which each layer of warp yarns is combined with several layers of weft yarns, and all yarns in the same warp row have the same movement within the weaving plane. Document WO2006 / 136755 describes the production of such fabrics.

[0019] A “middle” plane or line refers to a plane or line that is not located at the edge of the volumetric or two-dimensional image under consideration.

[0020] The inventors have developed a characterization method that, instead of relying on a direct comparison between two states of the same woven fiber structure to infer the transition linking these two states, assumes that the transformation between the two states of the linked structure can be separated into two quadratic transformations. These quadratic transformations link each state to a model of the structure, such as an undeformed model of the structure. The combination of these quadratic transformations allows finding the transformation linking the two considered states. The simplified model is reflected in the method according to the invention as a reference contour, which is compared to a one-dimensional contour obtained from an initial volumetric image.

[0021] The method according to the invention is further significant because it is applicable to the characterization of woven fiber structures with a given periodic pattern. The characterization problem is simplified using each image processing operation: first, a filtered three-dimensional image is obtained, then a filtered two-dimensional image, and then at least one one-dimensional profile is extracted from the two-dimensional image, representing the position of the yarn or radial ply columns along the weave direction at the considered volume location. Filtering overcomes the difficulties associated with repeating weave patterns along different directions. Information about the displacement of the yarn or ply columns at a given line level relative to this simplified weave model can then be obtained by comparison with, for example, a reference one-dimensional profile constructed from a simplified weave model. This greatly simplifies the comparison because it is performed in a lower dimension, rather than directly with the volume image. Furthermore, the digital image correlation (DIC) algorithm that can be used for comparison (Sutton, MA, Orteu, JJ, & Schreier, H. (2009), Image Correlation in Shape, Motion and Deformation Measurements: Basic Concepts, Theory and Applications, Springer Science & Business Media) converges more easily with filtering.

[0022] Then, by considering the results obtained in a single dimension of the intermediate plane, iterations can be performed on all lines of the same intermediate plane and in both weaving directions, and then on the multi-dry continuous plane of the volumetric image, to obtain information on the displacements occurring in the volume of the image compared to the simplified weaving model. By obtaining the transformation information between the true volumetric image and the simplified weaving model, the displacements involved in the different steps of processing the woven fiber structure to manufacture composite parts can be more easily characterized.

[0023] In one exemplary embodiment, the method may further include determining a displacement field based on the comparison result. In other words, the result of the comparison step may be a displacement field. This displacement field can be obtained using the DIC algorithm.

[0024] Studying the displacement field between different steps in the manufacturing process of composite parts can help understand the effects of loom parameters (weaving step), preform forming kinematics to allow for a digital pathway from one configuration to another (forming step), or to assess three-dimensional residual stress at the end of manufacturing (resin injection step).

[0025] In one exemplary embodiment, the comparison step may correspond to the step of aligning the one-dimensional contour with a reference contour. In particular, this step may be performed using the DIC algorithm suitable for the one-dimensional case.

[0026] In one exemplary embodiment, the filtering of the first and / or second processing may be Gaussian filtering, wherein the width of the filter is determined as a function of the average spacing between yarn or ply layers along the considered filtering direction. Specifically, the width of the Gaussian filter may be contained between T and several T, where T is the average spacing between yarn or ply layers along the considered filtering direction and in the considered sample.

[0027] In one exemplary embodiment, the method may further include the steps of obtaining several one-dimensional contours along a continuous line of the filtered two-dimensional image, and iteratively comparing each obtained one-dimensional contour with a reference contour from the intermediate line, taking into account the results of a comparison performed at the previous line. Specifically, the comparison steps can be performed iteratively by performing two passes from the intermediate line to the edge of the filtered two-dimensional image: towards each relative edge of the image. Therefore, all results obtained for the same plane are consistent, and no discontinuities occur. The results obtained from the intermediate plane can be used to initiate a comparison between a two-dimensional image of a continuous plane of volumetric image and a simplified two-dimensional model.

[0028] In one exemplary embodiment, the method may further include the steps of obtaining a plurality of two-dimensional images at continuous planes of the filtered volumetric image after comparing all one-dimensional contours in the first and second directions with reference contours, and iteratively comparing each obtained two-dimensional image with a reference two-dimensional image from an intermediate plane, taking into account the results of comparisons performed on the previous plane. Specifically, the comparison steps may be performed iteratively by performing two passes from the intermediate plane up to the edges of the filtered volumetric image: toward each relative edge of the image. Therefore, alignment of the entire volumetric image can be achieved by ensuring the convergence of the DIC algorithm, which is initialized with the results obtained in previous iterations and can be started from the results obtained in the intermediate one-dimensional plane.

[0029] In one exemplary embodiment, the volumetric image can be obtained by X-ray tomography.

[0030] The present invention also relates to a system for characterizing fiber structures based on volumetric images, the fiber structures having a three-dimensional weave according to a given pattern, the given pattern being located between a plurality of warp yarns or warp plies extending along a first direction and a plurality of weft yarns or weft plies extending along a second direction perpendicular to the first direction, the system comprising:

[0031] A first module is configured to perform a first processing on the volumetric image by filtering along a third direction perpendicular to the first and second directions to attenuate the periodic pattern along the third direction.

[0032] A module for obtaining a two-dimensional image representing warp yarns or warp strands and weft yarns or weft strands, the image corresponding to a third-direction intermediate plane along a filtered volume image.

[0033] The second module is used to perform a second processing on the two-dimensional image by filtering along the first direction or the second direction to attenuate the periodic pattern along the first direction or the second direction.

[0034] A module for obtaining a one-dimensional profile representing the position of warp or weft yarns or ply yarns, the profile corresponding to the midline along a first or second direction of a filtered two-dimensional image, and

[0035] A module for comparing the one-dimensional profile with a reference profile.

[0036] The present invention also proposes a computer program comprising instructions for performing the steps of the above-described method when the program is executed by a computer.

[0037] It should be noted that the computer programs mentioned in this disclosure can be in any programming language and can be in the form of source code, object code, or intermediate code between source code and object code, such as in a partially compiled form or any other desired form.

[0038] The present invention also proposes a computer-readable recording medium having a computer program recorded thereon, including instructions for performing the steps of the above-described method.

[0039] The recording (or information) medium mentioned in this disclosure can be any entity or device capable of storing programs. For example, the medium can include storage devices such as ROMs, such as CD ROMs or microelectronic circuit ROMs, or magnetic recording devices such as floppy disks or hard disks.

[0040] On the other hand, the recording medium can correspond to a transmissible medium such as an electrical signal or an optical signal, which can be transmitted via cable or optical fiber, radio, or other means. The program according to the invention can be downloaded specifically from an Internet-type network.

[0041] Alternatively, the recording medium may correspond to an integrated circuit containing a program, the circuit being adapted to perform or be used to perform the method.

[0042] The present invention also relates to a method for manufacturing a component made of a composite material comprising a matrix-dense fiber reinforcement, the method comprising manufacturing the composite component from a fiber structure obtained by three-dimensional weaving, and characterizing the component by the characterization methods described above. The manufacturing of the composite component may include the steps of weaving the fiber structure, shaping the fiber structure, and injecting a matrix (e.g., resin) into the pores of the shaped fiber structure.

[0043] In one exemplary embodiment, the method for weaving a fiber structure may include weaving the fiber structure by three-dimensional weaving and characterizing the fiber structure by the characterization method described above.

[0044] In one exemplary embodiment, the method for forming a fiber structure obtained by three-dimensional weaving may include forming the fiber structure and characterizing the woven fiber structure by the characterization method described above.

[0045] In one exemplary embodiment, a method for densifying a fiber structure obtained by three-dimensional weaving may include injecting resin into the pores of the fiber structure to form a component made of a composite material, and characterizing the component thus formed by the characterization method described above. Attached Figure Description

[0046] Figure 1 This is a schematic diagram of the fan blades of an aircraft turbine engine.

[0047] Figure 2 It is used to manufacture such as Figure 1 The simplified cross-sectional view of the woven fiber structure of the blade shown illustrates the simplified alignment of several warp yarns.

[0048] Figure 3 From the manufacture of such Figure 1 The cross-sectional view obtained from the X-ray tomography scan of the woven fiber structure of the blade shown reveals the misalignment (tilt, bend, twist, or more complex shape) of several warp yarns.

[0049] Figure 4 Two simplified braided planes in a braided fiber structure with interlocking braids are shown.

[0050] Figure 5 This illustrates the decomposition of the transformation between two deformable states of a structure linked by a structural model.

[0051] Figure 6 A simplified weaving model of interlocking weaving is explained.

[0052] Figure 7This is a volumetric image of a component, including a woven fiber structure with interlocking weave, obtained through X-ray tomography. In this visualization, the X, Y, and Z scales are not equal.

[0053] Figure 8 This is a flowchart illustrating the main steps of a characterization method according to an embodiment of the present invention.

[0054] Figure 9 A volumetric image is displayed to illustrate the method and results after the first image processing. In this visualization, the X, Y, and Z scales are not equal.

[0055] Figure 10 The visualization shows the extracted 2D image and the result obtained after processing the second image. In this visualization, the X and Y scales are not equal.

[0056] Figure 11 The extracted one-dimensional contour and the comparison between the contour and the reference contour are shown.

[0057] Figure 12 This visualization displays a reference 2D image used for aligning 2D images and an aligned or corrected 2D image of the intermediate plane. In this visualization, the X and Y scales are not equal.

[0058] Figure 13 A characterization system according to an embodiment of the present invention is shown. Detailed Implementation

[0059] The method according to the invention relates to a fiber structure woven by three-dimensional braiding, and also to a component made of a composite material that integrates such a braided fiber structure as a fiber reinforcement.

[0060] Figure 5 The basic principle of this invention is explained. According to this principle, in order to characterize the fiber structure, the transformation of two states of the linked structure (according to the definition of deformation) is decomposed into several quadratic transformations, and each state is linked to a simplified or reference model of the structure. Each of these quadratic transformations is determined by a digital image correlation (DIC) algorithm.

[0061] Consider the fiber structure in deformation state 1, which then undergoes a transformation [T] during method P. Method P can be, for example, a step of weaving the fiber structure, forming, or densifying the matrix by injection. The transformation [T] can be decomposed by using the states of the structure corresponding to a simplified model of the structure. The transformation [TR1] that links state 1 of the structure to the model, and the transformation [TR2] that links state 2 of the structure to the model, are then defined. The transformation [T] can then be viewed as transformations [TR1] and [TR2]. -1The combination of [TR1] and [TR2] allows for the acquisition of the transformations [T] of the two states of the linked fiber structure. This decomposition further enables the correction of the volumetric image of the structure in a given state, allowing for the study of its deformation for quality control purposes.

[0062] In this text, the term "yarn" is used interchangeably to refer to yarn or ply.

[0063] Figure 6 Several views of a simplified model of a woven fiber structure obtained through three-dimensional weaving with interlocking weaving are shown. The weft direction corresponds to the X-axis, the warp direction corresponds to the Y-axis, and the thickness direction of the structure corresponds to the Z-axis. The X, Y, and Z axes are perpendicular. The weft yarn t typically extends along the weft direction given by the X-axis, while the warp yarn c typically extends along the warp direction given by the Y-axis. Figure 6 It also shows a side view of the fiber structure in the XZ plane, with the weft yarns t undulating, and a top view in the XY plane, with the warp yarns c and weft yarns t forming a grid pattern that intersects at 90°.

[0064] Figure 7 An example of a volumetric image obtained by X-ray tomography of a woven fiber structure within a component made of composite material is shown. Figure 6 The model with interlocking weaving is used to weave the fiber structure of this example. The X, Y, and Z axes are the same as those defined in the fiber structure model. Yarns are identified in light gray, and the matrix is ​​identified in dark gray.

[0065] Now we will combine Figure 8 The flowchart in the diagram describes an example of a method for characterizing the structure of woven fibers.

[0066] The first step, E10, includes processing from... Figure 7 Volumetric image extracted from the image 10 ( Figure 9 A first filtering process is applied. This first process involves applying a Gaussian filter (by convolution) along the Z-direction to the volume image 10, that is, over the thickness of the sample, in order to minimize the effect of modulation of the warp and weft yarns in the Z-direction. The radius of the Gaussian filter is advantageously defined within an interval between T and several T, where T is the average value of the intervals between yarn layers along the Z-direction. This average value can be ideal, for example, an average value defined in a loom to obtain the woven fiber structure, or it can be calculated based on the sample under consideration. This yields the filtered volume image 12. Figure 9 ).

[0067] Then, by extracting at the level of the intermediate plane 16 or intermediate layer of the filtered volumetric image 12, a two-dimensional image 14 is obtained (step E20). Figure 10Image 16 represents the intersecting warp (along the X direction) and weft (along the Y direction), and filtering is used to gather information about adjacent planes. The intermediate plane 16 is preferably chosen at a central location, for example, z0 = D / 2, where D is the depth of the volumetric image 10 along the Z-axis. The choice of the intermediate plane 16 allows for layers of fiber structure, where deformation is generally weaker than at the edges, making the initialization of the DIC algorithm more robust.

[0068] Next, in step E31, a second processing of the two-dimensional image 14 is performed by filtering in the Y direction. In addition to applying a Gaussian filter in the Y direction and considering the defined filter width and the average spacing between the warp planes, a Gaussian filter with characteristics similar to the previously used filter is applied. This filtering operation allows for the attenuation of high frequencies due to yarn crossovers. This results in the filtered two-dimensional image 18. Figure 10 ).

[0069] Then, in step E32, a one-dimensional profile 20 representing the position of the weft yarn row is obtained. Figure 11 This corresponds to the extraction at the median line 22 along the Y direction of the filtered two-dimensional image 18. The median line 22 is preferably selected at the middle position, for example, at y0 = H / 2, where H is the height of the image.

[0070] Next, in step E33, the one-dimensional profile 20 obtained in the previous step is compared with the reference profile 24. Here, the reference profile 24 is obtained from the previously proposed simplified weaving model. The model equation used as the reference profile 24 is shown below, where μ and σ are the mean and variance of the considered one-dimensional profile 20, and ω x This corresponds to the spatial frequency. The comparison step here implements the DIC algorithm, which determines the displacement field UX and the brightness correction of the image, thus relating the one-dimensional profile 20 to its simplified model.

[0071]

Mathematical Formula 1

[0072] M 1D (x,y,z)=μ+σ·sin(2π·ω x ·x)

[0073] To obtain the displacement field UX of the weft yarn rows in the entire intermediate plane 16, it is necessary to perform two passes through the filtered two-dimensional image 18, starting from the intermediate line 22 and continuing to the edge. Each pass uses the displacement field results obtained from the previous line to initialize the DIC algorithm. The first pass is performed in the direction of reference 26, and the second pass is performed in the direction of reference 28. Figure 10 This is used to cover the entire filtered two-dimensional image.

[0074] The same operation can then be performed in steps E41, E42, and E43 for the other direction of weaving (here, the X direction). After passing through the entire filtered 2D image along the X direction, the displacement field UY of the warp yarns passing through the intermediate plane 16 is obtained.

[0075] At the end of the iteration of the entire intermediate plane 16, the displacement fields UX and UY of the warp and weft yarns of the entire intermediate plane (as well as image brightness correction) are provided.

[0076] Finally, in step E50, the images of all planes extracted from the filtered volumetric image 12 using the DIC algorithm are aligned with the simplified two-dimensional model, and iteration is performed from the intermediate plane 16. The model equation used as the simplified two-dimensional model is shown below, where μ and σ are the mean and variance of the two-dimensional images extracted on the considered planes, and ω x and ω y It is the spatial frequency of the identification.

[0077]

Mathematical Formula 2

[0078] M 2D (x,y,z)=μ+σ·[sin(2π·ω x ·x)+sin(2π·ω y ·y)]

[0079] Two-dimensional image 14 is first extracted at the level of the intermediate plane 16 that provides the displacement fields UX and UY. This two-dimensional image 14 is then compared with the simplified model defined above, which is represented by the reference two-dimensional image 30. Figure 12 The DIC algorithm is initialized using displacement fields UX and UY (steps E31 to E43), which are found using calculations performed in one dimension. The two-dimensional image 14 can then be corrected to align it, resulting in an aligned two-dimensional image 32 and the corresponding displacement fields.

[0080] Alignment of the entire volume can now be achieved through two iterations, starting from the intermediate plane 16 and moving towards the edge in two opposite directions. The first iteration is performed in reference direction 34, and the second iteration in reference direction 36... Figure 9 The second iteration is performed on the previous plane to cover the entire filtered volumetric image. In each iteration, the DIC algorithm is initialized with the results obtained on the previous plane.

[0081] Following step E50, displacement fields UX and UY are provided for all planes. All displacement fields UX and UY in the volumetric image allow access to the transformation that links the deformed woven fiber structure to its simplified (non-deformed) model, corrects its volumetric image 10, and characterizes the deformation it has undergone.

[0082] Figure 13 An example of a characterization system 40 for implementing a method according to an embodiment of the present invention is illustrated schematically. System 40 includes:

[0083] The first module 42 is used to process the volumetric image to perform step E10.

[0084] Module 44 is used to obtain a two-dimensional image to perform step E20.

[0085] The second module 46 is used to process the two-dimensional image to perform step E31 and / or step E41.

[0086] Module 48 is used to obtain a one-dimensional profile to perform steps E32 and / or E42.

[0087] Module 50 for comparing a one-dimensional profile with a reference profile to perform steps E33 and / or E43, and

[0088] Module 52 is used to compare two-dimensional images of different planes of a volumetric image to compare the two-dimensional image with a reference two-dimensional image to perform step E50.

Claims

1. A method for characterizing a fiber structure based on a volumetric image, the fiber structure having a three-dimensional weave according to a given pattern, the given pattern being located between a plurality of warp yarns or warp plies extending along a first direction and a plurality of weft yarns or weft plies extending along a second direction perpendicular to the first direction, the method comprising: The volumetric image undergoes a first processing step by filtering along a third direction perpendicular to the first and second directions to attenuate the periodic pattern along the third direction. Obtain a two-dimensional image representing warp yarns or warp strands and weft yarns or weft strands, the two-dimensional image corresponding to the middle plane of a third direction along the filtered volume image. The two-dimensional image is subjected to a second processing by filtering along the first or second direction to attenuate the periodic pattern along the first or second direction. Obtain a one-dimensional profile representing the position of the warp or weft yarn or ply, the one-dimensional profile corresponding to the midline along a first or second direction of the filtered two-dimensional image, and The one-dimensional profile is compared with a reference profile.

2. The method according to claim 1, further comprising determining a displacement field based on the result of the comparison.

3. The method according to claim 1, wherein, The comparison step is performed using a digital image correlation algorithm.

4. The method according to claim 1, wherein, The filtering in the first and / or second processing is Gaussian filtering, wherein the width of the filter is determined as a function of the average spacing between yarn or ply layers along the filtering direction under consideration.

5. The method of claim 1 further includes the steps of obtaining a plurality of one-dimensional contours on a continuous line of the filtered two-dimensional image, and the step of iteratively comparing each obtained one-dimensional contour with a reference contour from the intermediate line by taking into account the result of a comparison made at the previous line.

6. The method of claim 5 further comprises the steps of obtaining a plurality of two-dimensional images at a continuous plane of the filtered volume image after comparing all one-dimensional contours in the first and second directions with a reference contour, and the steps of iteratively comparing each obtained two-dimensional image with a reference two-dimensional image from an intermediate plane by taking into account the results of the comparisons made on the previous plane.

7. The method according to any one of claims 1 to 6, wherein, The volumetric image was obtained by X-ray tomography.

8. A system for characterizing a fiber structure based on a volumetric image, the fiber structure having a three-dimensional weave according to a given pattern, the given pattern being located between a plurality of warp yarns or warp plies extending along a first direction and a plurality of weft yarns or weft plies extending along a second direction perpendicular to the first direction, the system comprising: A first module is configured to perform a first processing on the volumetric image by filtering along a third direction perpendicular to the first and second directions to attenuate the periodic pattern along the third direction. A module for obtaining a two-dimensional image representing warp yarns or warp strands and weft yarns or weft strands, the two-dimensional image corresponding to a third-direction intermediate plane along a filtered volume image. The second module is used to perform a second processing on the two-dimensional image by filtering along the first direction or the second direction to attenuate the periodic pattern along the first direction or the second direction. A module for obtaining a one-dimensional profile representing the position of warp or weft yarns or ply yarns, the one-dimensional profile corresponding to the center line along a first or second direction of a filtered two-dimensional image, and A module for comparing the one-dimensional profile with a reference profile.

9. A computer program comprising, when executed by a computer, instructions for performing the steps of the method according to any one of claims 1 to 7.

10. A computer-readable recording medium having a computer program recorded thereon including instructions for performing the steps of the method according to any one of claims 1 to 7.

Images