Variable-thickness curved-surface woven ceramic matrix composite three-dimensional reconstruction method

Abstract

The application discloses a kind of variable-thickness curved surface weaving ceramic matrix composite three-dimensional reconstruction methods, belong to the field of microstructure identification and reconstruction.The application can quickly and accurately calculate the area of each cross section of the composite material by identifying the outer edge of the curved surface weaving composite material combined with the area of a single pixel;Meanwhile, by deleting the surface fibers arranged complexly, the classification accuracy of internal warp and weft can be improved, and it is more effective for subsequent reduction of human participation;By processing the three views of XCT data of the composite material respectively, the pore position can be accurately extracted from the original XCT image, the fiber bundles in different slices are numbered from the overhead view and the side view through the number of fiber columns and the spatial position relationship, and the spatial paths of warp yarns and weft yarns are respectively fitted, the fiber bundle cross section is represented by oblique elliptic function, the three-dimensional model reconstruction of curved surface weaving CMC can be quickly realized, the algorithm is practical, and the whole process is simple to operate.

Description

Technical Field

[0001] This invention belongs to the field of microstructure identification and reconstruction, specifically relating to a method for microstructure identification and reconstruction of three-dimensional woven ceramic matrix composites. Background Technology

[0002] Ceramic matrix composites (CMCs) possess advantages such as low density, high temperature resistance, corrosion resistance, and high specific strength, making them ideal candidate materials for advanced aero-engines and gas turbines, especially for hot-end high-temperature components like turbine blades and nozzle trimmers. These components commonly employ curved surface structures, meaning the fibers within the material are not strictly a single type of weave, but rather a mixture of multiple structures. Compared to regular sheet shapes, curved woven CMCs have a more complex microstructure. Directly applying existing microstructure identification methods to curved woven CMCs can lead to various types of misclassified pixels due to significant structural variations, requiring extensive manual correction. Furthermore, inherent limitations in the manufacturing process result in numerous random defects within the material, such as pores and fiber misalignment, making it difficult for idealized three-dimensional finite element models to accurately represent the microstructure, further reducing the accuracy of mechanical property analysis results.

[0003] Therefore, it is necessary to conduct research on the microstructure identification and model reconstruction methods for woven surface CMC. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a three-dimensional reconstruction method for woven ceramic matrix composites with variable thickness surfaces, which addresses the shortcomings of the prior art.

[0005] To achieve the above-mentioned technical objectives, the technical solution adopted by the present invention is as follows:

[0006] A method for three-dimensional reconstruction of woven ceramic matrix composites with variable thickness curved surfaces, characterized by the following steps:

[0007] Step 1: XCT was used to scan the curved woven CMC to obtain... X Zhang contains a grayscale image of the material's internal microstructure. Original_Image Extract aperture pixels from multiple different regions of the image and determine the maximum value of the aperture pixels. Hole_ Thresh and search Original_Image Less than Hole_Thresh The pixel position is recorded, and its position index is used to distinguish curved woven composite materials. Original_Image fiber bundles, to obtain images ST_Image ;

[0008] Step 2: Export Original_Image and ST_Image Top view Top_View and Top_ST ,according to Top_ST The characteristics of the R, G, and B values ​​of warp and weft fiber bundles in the RGB image are used to distinguish between the two. Top_View and Top_ ST The number is the same, both being Num_Top sheets;

[0009] Step 3: Sharpen Top_View Images and calculate each one Top_View Image edges are analyzed, stray edges are removed, and the complete cross-sectional edge of the woven curved surface (CMC) is obtained. The number of pixels enclosed by the edge is counted, and the area of ​​the cross-section is calculated by combining the area of ​​each individual pixel. Area The image is obtained by fitting a function to the cross-sectional edge and narrowing the edge function inward until the outer fiber bundles, which are difficult to distinguish, are outside the edge function. The pixel regions of the fiber bundles outside the edge are then removed. Delete_Top_View ;

[0010] Step 4: Determine each image Original_Image Number of fiber bundles at the thickest point in the image Max_Num and the minimum number of fiber bundles on both sides Min_Num And determine the total number of fiber bundles in each slice. Total_Col ;

[0011] Step 5: Cutting Delete_Top_View The image is divided into fiber bundle regions with a defined number of columns, and morphological erosion is performed to ensure that no fiber bundles in each column have overlapping edges. The centroid of each fiber bundle is calculated, and each fiber bundle is numbered sequentially from top to bottom and from left to right according to its spatial relationship. The number of fiber bundles in each column must not be less than the minimum number. Min_Num The quantity must not exceed the maximum. Max_Num ;

[0012] Step 6: Represent the cross-section of each fiber bundle using the equation of an oblique ellipse, sharpen the cross-sectional image of each fiber bundle, calculate the edge of the cross-section of each fiber bundle, calculate the two points that are furthest apart on the edge, and connect them to obtain the cross-section. Line_1 ,calculate Line_1 length Long_Axis Find the perpendicularity of the line connecting any two points on the edge. Line_1 And record the length of the line segment at the maximum distance. Minor_Axis Connecting the top-left vertex of the image to the centroid of the ellipse yields... Line_2 ,calculate Line_2 The angle relative to the horizontal axis of the image, and calculation. Line_1 and Line_2 The angle between the two points can be subtracted to obtain the angle of inclination of the ellipse, Angle, and the major axis of the ellipse. Long_Axis With short axis Minor_Axis ;

[0013] Step 7: Divide the image into left and right sides using the central axis of the first image as the boundary, and place the corresponding lines on the central axis as... Middle_Col Column, calculate the first Y Calculate the angle between the left edge of the image and the top left corner vertex of the image, and the line connecting the center of the first fiber bundle in each column of fiber bundles on the left. Y+Num_Top The angle between the right edge of the cross-sectional image of the fiber bundle and the line connecting the top right corner of the image to the center of the first fiber bundle in each column is calculated to be a total of [missing information]. Total_ Col-1 One; because the curved surface structure causes the fibers corresponding to the central axis to deflect continuously in a certain fixed direction, therefore, the vertical center line of the image and the center of the upper edge are recorded simultaneously. Middle_Col The angle between the lines connecting the centroids of the first fiber bundles;

[0014] Step 8: Based on the number of cross-sectional images of the fiber bundle Num_Top The average angle change of the fiber bundles in each image is used to match each fiber bundle in each column of fiber bundles in the previous and next slices. The angle change is added to each slice to obtain the approximate position of the centroid of the first fiber of the fiber bundle in the next slice corresponding to a certain column of fiber bundle in the current image, thus providing approximate information for the positioning of each column of fiber bundle in each image.

[0015] Step 9: Pair the front and back images of the fiber bundle columns according to the fiber column number. When locating a column, count the columns from left to right. Left_Col At the same time, count the columns from right to left. Right_Col And the sum of the two is the total number of columns. Total_Col When pairing previous and subsequent images, the column numbers of the same fiber bundles are the same; in addition, because the yarn will be added or removed on the surface layer of the curved CMC, the next slice... Next_Col If the number of fiber bundles in a column of fibers is greater than the number of fiber bundles in the same column of the current slice, then the first fiber bundle counting from top to bottom in the next slice is marked as U+1, where U is the number of fiber bundles in the corresponding column of the current slice. If it is still greater, then the bottommost fiber bundle is marked as U+2. If it is still greater, then the second fiber bundle is marked as U+3, and so on, until the number of fiber bundles in the corresponding column of the current slice and the next slice are the same. If the number of fiber bundles in the G-th column of the next slice is less than the number of fiber bundles in the same column of the current slice, then U-1, U-2, ... are marked.

[0016] Step 10: Connect the centroids of the fiber bundles with the same number on the front and back slices, calculated in Step 5, and fit the curve using the least squares method to obtain the three-dimensional spatial orientation function of the warp fiber bundles.

[0017] Step 11: Export the left view of the grayscale image from Step 3. Left_Image Then, assign the warp fiber bundle pixel value to 0; at this time, only the weft fiber bundle cross section is in the current view. Repeat step 6 to calculate the oblique ellipse geometric parameters of each weft fiber bundle cross section in each image; since the change of weft yarn is similar to the change of warp yarn, but the number of yarn columns is relatively more obvious and easier to determine, but yarn addition and subtraction will also occur. Therefore, repeat steps 7, 8 and 9 to calculate the centroid of each weft fiber bundle cross section and obtain the three-dimensional spatial orientation function of the weft fiber bundle.

[0018] Step 12: Based on the yarn three-dimensional spatial orientation function in Steps 10 and 11, and sweeping according to the oblique ellipse parameters of each fiber bundle, perform Boolean operations on the overlapping parts of the warp and weft fiber bundles to establish a three-dimensional model of the variable thickness curved surface weaving CMC.

[0019] To optimize the above technical solution, the specific measures also include:

[0020] In step 1, the two-dimensional structural tensor method is used to distinguish woven composite materials with curved surfaces. Original_Image fiber bundles, to obtain images ST_Image .

[0021] In step 3, sharpening is performed using the Laplacian operator. Top_View Images and each image is computed using the Canny edge operator. Top_View Image edges are analyzed, stray edges are removed, the complete cross-sectional edge of the surface weave (CMC) is obtained, and the number of pixels enclosed by the edge is counted.

[0022] In step 6, the cross-sectional image of each fiber bundle is sharpened using the Laplacian operator, and the Canny edge operator is used to calculate the cross-sectional edge of each fiber bundle. The two points that are furthest apart on the edge are calculated and connected to obtain the result. Line_1 .

[0023] After step 9 and before step 10, a manual judgment step is set up. The manual judgment step is as follows: after the pairing of the front and back images of the fiber bundle column is completed, the manual judgment is made on whether each fiber bundle column is correctly paired.

[0024] The beneficial effects of this invention are:

[0025] 1. By identifying the outer edge of the woven composite material, the area of ​​each cross-section of the composite material can be calculated quickly and accurately by combining the area of ​​a single pixel; at the same time, by deleting the complexly arranged surface fibers, the accuracy of internal warp and weft yarn classification can be improved, and it is also more effective in reducing manual intervention in the future.

[0026] 2. By processing the three views of the XCT data of the composite material, the pore locations can be accurately extracted from the original XCT image. From the top and side views, the fiber bundles in different slices are numbered according to the number of fiber columns and spatial position relationships. The spatial paths of warp and weft yarns are fitted respectively, and the cross-section of the fiber bundle is represented by the oblique elliptical function. The three-dimensional model reconstruction of the curved surface braided CMC can be realized quickly. The algorithm is highly practical and the whole process is simple to operate. Attached Figure Description

[0027] Figure 1 It is an XCT image of a curved surface woven CMC;

[0028] Figure 2 This is a diagram showing the results of structural tensor calculations on XCT images;

[0029] Figure 3 This is a top view of the curved surface weaving CMC;

[0030] Figure 4 This is a diagram showing the structural tensor calculation results from a top view.

[0031] Figure 5 This is a comparison result chart of the three channels;

[0032] Figure 6 It is the top view edge of the XCT image of the curved surface woven CMC;

[0033] Figure 7 This is a top view of a surface-woven XCT image filled with a single pixel;

[0034] Figure 8 This is a schematic diagram of a cut-out fiber bundle in a certain area;

[0035] Figure 9 The cut fiber bundle images are numbered and marked on the centroid position diagram;

[0036] Figure 10 This is a diagram showing the reordering of the numbering results;

[0037] Figure 11 This is a diagram illustrating angle markings;

[0038] Figure 12 It is the exported Left_Image;

[0039] Figure 13 It is a cross-sectional image of a certain fiber bundle;

[0040] Figure 14 It is an edge view of a cross-sectional image of a certain fiber bundle;

[0041] Figure 15 This is a diagram showing the determination of the geometric parameters of a slanted ellipse for a certain fiber bundle. Detailed Implementation

[0042] The embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0043] The present invention provides a three-dimensional reconstruction method for woven ceramic matrix composites with variable thickness surfaces, comprising the following steps:

[0044] Step 1: XCT was used to scan the curved woven CMC to obtain... X Zhang contains a grayscale image of the material's internal microstructure. Original_Image ,like Figure 1 Extract aperture pixels from multiple different regions of the image and determine the maximum value of the aperture pixels. Hole_Thresh and search Original_Image Less than Hole_Thresh The pixel position is recorded, and its position index is used to distinguish woven composite materials with curved surfaces using the two-dimensional structural tensor method. Original_Image fiber bundles, to obtain images ST_Image, like Figure 2 ;

[0045] Step 2: Export Original_Image and ST_Image Top view Top_View and Top_ST ,like Figure 3 and 4 ,according to Top_ST The characteristics of the R, G, and B values ​​of warp and weft fiber bundles in RGB images are used to distinguish between them, such as... Figure 5 ;

[0046] Step 3: Sharpen the image using the Laplacian operator and calculate each edge using the Canny edge operator. Top_View Image edges, remove stray edges, obtain the complete cross-sectional edge of the curved weave CMC (Curved Surface Weave CMC). Figure 6 It also counts the number of pixels enclosed by the edge, and combines this with the area of ​​each individual pixel to calculate the area of ​​the cross section. Area like Figure 7 The image is obtained by fitting a function to the cross-sectional edge and narrowing the edge function inward until the outer fiber bundles, which are difficult to distinguish, are outside the edge function. The pixel regions of the fiber bundles outside the edge are then removed. Delete_Top_View ;

[0047] Step 4: Determine each image Original_Image Number of fiber bundles at the thickest point in the image Max_Num and the minimum number of fiber bundles on both sides Min_Num And determine the total number of fiber bundles in each slice. Total_Col ;

[0048] Step 5: Cutting Delete_Top_ViewThe image is divided into columns of fiber bundle regions, and morphological erosion is performed to ensure that no fiber bundles in each column have overlapping edges. The centroid of each fiber bundle is calculated, and the fiber bundles are numbered using a connected component labeling method. At this point, the numbering is randomized, such as... Figure 9 Then, each fiber bundle is numbered according to its spatial relationship, from top to bottom and from left to right, such as... Figure 10 The number of fiber bundles in each column must not be less than the minimum number. Min_Num The quantity must not exceed the maximum. Max_Num ;

[0049] Step 6: Taking a specific fiber bundle as an example, such as... Figure 11 The cross-section of each fiber bundle is represented using the equation of an oblique ellipse. The image is sharpened using the Laplacian operator, and the cross-sectional edges of each fiber bundle are calculated using the Canny edge operator, as shown below. Figure 12 Calculate the two points on the edge that are furthest apart and connect them to get Line_1 ,calculate Line_1 length Long_Axis Find the perpendicularity of the line connecting any two points on the edge. Line_1 And record the length of the line segment at the maximum distance. Minor_Axis Connecting the top-left vertex of the image to the centroid of the ellipse yields... Line_2 ,calculate Line_2 The angle relative to the horizontal axis of the image, and calculation. Line_1 and Line_2 The angle between the two points can be subtracted to obtain the angle of inclination of the ellipse, Angle, and the major axis of the ellipse. Long_Axis With short axis Minor_Axis ,like Figure 13 ;

[0050] Step 7: Divide the image into left and right sides using the central axis of the first image as the boundary, and place the corresponding lines on the central axis as... Middle_Col Column, calculate the first Y Calculate the angle between the left edge of the image and the top left corner vertex of the image, and the line connecting the center of the first fiber bundle in each column of fiber bundles on the left. Y+R( ) The angle between the right edge of the image and the upper right corner vertex of the image and the center line of the first fiber bundle in each column of fiber bundles, the total angle is ( Total_Col- 1 ) each, such as Figure 14 Because the curved surface structure causes the fibers corresponding to the central axis to deflect continuously in a fixed direction, it is necessary to simultaneously record the vertical center line of the image and the center of the upper edge. Middle_Col The angle between the lines connecting the centroids of the first fiber bundles;

[0051] Step 8: Based on the number of images RThe average angle change of the fiber bundles in each image is used to match each fiber bundle in each column of fiber bundles in the previous and next slices. The angle change is added to each slice to obtain the approximate position of the centroid of the first fiber of the fiber bundle in the next slice corresponding to a certain column of fiber bundle in the current image, thus providing approximate information for the positioning of each column of fiber bundle in each image.

[0052] Step 9: Pair the front and back images of the fiber bundle columns according to the fiber column number. When locating a column, count the columns from left to right. Left_Col At the same time, count the columns from right to left. Right_Col And the sum of the two is the total number of columns. Total_Col When pairing images, the columns containing the same fiber bundles are numbered the same. Furthermore, since yarn increases or decreases on the surface layer of the curved CMC weave, if the number of fiber bundles in column G of the next slice is greater than the number of fiber bundles in the same column of the current slice, the first fiber bundle of the next slice is marked as U+1, where U is the number of fiber bundles in the corresponding column of the current slice. If it is still greater, the bottommost fiber bundle is marked as U+2. If it is still greater, the second fiber bundle is marked as U+3, and so on, until the number of fiber bundles in the corresponding columns of the current and next slices is the same. If the number of fiber bundles in column G of the next slice is less than the number of fiber bundles in the same column of the current slice, then U-1, U-2, ... are marked. After all images are paired, manual verification is required to ensure that each column of fiber bundles is correctly paired.

[0053] Step 10: Connect the centroids of the fiber bundles with the same number on the front and back slices, calculated in Step 5, and fit the curve using the least squares method to obtain the three-dimensional spatial orientation function of the warp fiber bundles.

[0054] Step 11: Export the left view of the grayscale image from Step 3. Left_Image ,like Figure 15 Then, assign the warp fiber bundle pixel value to 0; at this time, only the weft fiber bundle cross section is in the current view. Repeat step 6 to calculate the oblique ellipse geometric parameters of each weft fiber bundle cross section in each image; since the change of weft yarn is similar to the change of warp yarn, but the number of yarn columns is relatively more obvious and easier to determine, but yarn addition and subtraction will also occur. Therefore, repeat steps 8 and 9 to calculate the centroid of each weft fiber bundle cross section and obtain the three-dimensional spatial orientation function of the weft fiber bundle.

[0055] Step 12: Based on the yarn three-dimensional spatial orientation function in Steps 10 and 11, and sweeping according to the oblique ellipse parameters of each fiber bundle, perform Boolean operations on the overlapping parts of the warp and weft fiber bundles to establish a three-dimensional model of the curved surface weaving CMC.

[0056] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should be considered within the scope of protection of the present invention.

Claims

1. A method for three-dimensional reconstruction of woven ceramic matrix composites with variable thickness surfaces, characterized by: Includes the following steps: Step 1: XCT was used to scan the curved woven CMC to obtain... X Zhang contains a grayscale image of the material's internal microstructure. Original_Image Extract aperture pixels from multiple different regions of the image and determine the maximum value of the aperture pixels. Hole_ Thresh and search Original_Image Less than Hole_Thresh The pixel position is recorded, and its position index is used to distinguish curved woven composite materials. Original_Image fiber bundles, to obtain images ST_Image ; Step 2: Export Original_Image and ST_Image Top view Top_View and Top_ST ,according to Top_ST The characteristics of the R, G, and B values ​​of warp and weft fiber bundles in the RGB image are used to distinguish between the two. Top_View and Top_ST The number is the same, both being Num_Top sheets; Step 3: Sharpen Top_View Images and calculate each one Top_View Image edges are analyzed, stray edges are removed, and the complete cross-sectional edge of the woven curved surface (CMC) is obtained. The number of pixels enclosed by the edge is counted, and the area of ​​the cross-section is calculated by combining the area of ​​each individual pixel. Area The image is obtained by fitting a function to the cross-sectional edge and narrowing the edge function inward until the outer fiber bundles, which are difficult to distinguish, are outside the edge function. The pixel regions of the fiber bundles outside the edge are then removed. Delete_Top_View ; Step 4: Determine each image Original_Image Number of fiber bundles at the thickest point in the image Max_Num and the minimum number of fiber bundles on both sides Min_Num ; And determine the total number of fiber bundles in each slice. Total_Col ; Step 5: Cutting Delete_Top_View The image is divided into fiber bundle regions with a defined number of columns, and morphological erosion is performed to ensure that no fiber bundles in each column have overlapping edges. The centroid of each fiber bundle is calculated, and each fiber bundle is numbered sequentially from top to bottom and from left to right according to its spatial relationship. The number of fiber bundles in each column must not be less than the minimum number. Min_ Num The quantity must not exceed the maximum. Max_Num ; Step 6: Represent the cross-section of each fiber bundle using the equation of an oblique ellipse, sharpen the cross-sectional image of each fiber bundle, calculate the edge of the cross-section of each fiber bundle, calculate the two points that are furthest apart on the edge, and connect them to obtain the cross-section. Line_1 ,calculate Line_1 length Long_Axis Find the perpendicularity of the line connecting any two points on the edge. Line_1 And record the length of the line segment at the maximum distance. Minor_Axis Connecting the top-left vertex of the image to the centroid of the ellipse yields... Line_2 ,calculate Line_2 The angle relative to the horizontal axis of the image, and calculation. Line_1 and Line_2 The angle between the two values ​​can be subtracted to obtain the angle of inclination of the ellipse, Angle. Step 7: Divide the image into left and right sides using the central axis of the first image as the boundary, and place the corresponding lines on the central axis as... Middle_Col Column, calculate the first Y Calculate the angle between the left edge of the image and the top left corner vertex of the image, and the line connecting the center of the first fiber bundle in each column of fiber bundles on the left. Y The angle between the right edge of the cross-sectional image of the fiber bundle and the line connecting the top right corner of the image to the center of the first fiber bundle in each column is calculated to be a total of [missing information]. Total_Col-1 One; because the curved surface structure causes the fibers corresponding to the central axis to deflect continuously in a certain fixed direction, therefore, the vertical center line of the image and the center of the upper edge are recorded simultaneously. Middle_Col The angle between the lines connecting the centroids of the first fiber bundles; Step 8: Based on the number of cross-sectional images of the fiber bundle Num_Top The average angle change of the fiber bundles in each image is used to match each fiber bundle in each column of fiber bundles in the previous and next slices. The angle change is added to each slice to obtain the position of the centroid of the first fiber of the fiber bundle in the next slice corresponding to a certain column of fiber bundle in the current image, thus providing information for the positioning of each column of fiber bundles in each image. Step 9: Pair the front and back images of the fiber bundle columns according to the fiber column number. When locating a column, count the columns from left to right. Left_Col At the same time, count the columns from right to left. Right_Col And the sum of the two is the total number of columns. Total_ Col When pairing previous and subsequent images, the column numbers of the same fiber bundles are the same; in addition, because the yarn will be added or removed on the surface layer of the curved CMC, the next slice... Next_Col If the number of fiber bundles in a column is greater than the number of fiber bundles in the same column of the current slice, then the first fiber bundle counted from top to bottom in the next slice is marked as U+1, where U is the number of fiber bundles in the corresponding column of the current slice. If it is still more, then the bottommost fiber bundle is marked as U+2. If it is still more, then the second fiber bundle is marked as U+3, until the number of fiber bundles in the corresponding column of the current slice and the next slice are the same. Step 10: Connect the centroids of the fiber bundles with the same number on the front and back slices, calculated in Step 5, and fit the curve using the least squares method to obtain the three-dimensional spatial orientation function of the warp fiber bundles. Step 11: Export the left view of the grayscale image from Step 3. Left_Image Then, assign the warp fiber bundle pixel value to 0; at this time, only the weft fiber bundle cross section is in the current view. Repeat step 6 to calculate the oblique ellipse geometric parameters of each weft fiber bundle cross section in each image. Repeat steps 7, 8 and 9 to calculate the centroid of each weft fiber bundle cross section and obtain the three-dimensional spatial orientation function of the weft fiber bundle. Step 12: Based on the yarn three-dimensional spatial orientation function in Steps 10 and 11, and sweeping according to the oblique ellipse parameters of each fiber bundle, perform Boolean operations on the overlapping parts of the warp and weft fiber bundles to establish a three-dimensional model of the variable thickness curved surface weaving CMC.

2. The method for three-dimensional reconstruction of a variable-thickness woven ceramic matrix composite material according to claim 1, characterized in that: In step 1, the two-dimensional structural tensor method is used to distinguish curved woven composite materials. Original_Image fiber bundles, to obtain images ST_Image .

3. The method for three-dimensional reconstruction of a variable-thickness woven ceramic matrix composite material according to claim 1, characterized in that: In step 3, sharpening is performed using the Laplacian operator. Top_View Images and each image is computed using the Canny edge operator. Top_ View Image edges are analyzed, stray edges are removed, the complete cross-sectional edge of the surface weave (CMC) is obtained, and the number of pixels enclosed by the edge is counted.

4. The method for three-dimensional reconstruction of a variable-thickness woven ceramic matrix composite material according to claim 1, characterized in that: In step 6, the cross-sectional image of each fiber bundle is sharpened using the Laplacian operator, and the Canny edge operator is used to calculate the cross-sectional edge of each fiber bundle. The two points farthest apart on the edge are calculated and connected to obtain the result. Line_1 .

5. The method for three-dimensional reconstruction of a variable-thickness woven ceramic matrix composite material according to claim 1, characterized in that: After step 9 and before step 10, a manual judgment step is set up. The manual judgment step is as follows: after the pairing of the front and back images of the fiber bundle column is completed, the manual judgment is made on whether each fiber bundle column is correctly paired.

Images