A carding web three-dimensional modeling method based on single fiber original distribution characteristics

Abstract

The present application relates to a kind of carding fiber web three-dimensional modeling method based on single fiber original distribution characteristics, comprising: preparation and fiber color difference significantly tracer fiber, after blending, get blended fiber web;Obtain the two-dimensional image of blended fiber web on different microscope focusing plane, identify the tracer fiber in each image, obtain the two-dimensional coordinate point on corresponding plane;Each two-dimensional coordinate point and its corresponding plane height are combined into spatial coordinate point set, obtain the three-dimensional point cloud data of tracer fiber;According to three-dimensional point cloud data, extract the main stem feature point of tracer fiber, and then fit out continuous fiber center line;Based on fiber center line, construct single fiber model, make single fiber model with set orientation with random order in turn free fall motion, construct fiber web three-dimensional model.The present application can efficiently construct the carding fiber web model that accurately exhibits single fiber original distribution characteristics, while reducing modeling cost.

Description

Technical Field

[0001] This invention relates to the fields of image processing and fiber web 3D modeling technology, and in particular to a method for 3D modeling of combed fiber webs based on the original distribution characteristics of single fibers. Background Technology

[0002] Carding is a crucial step in the spinning process, loosening and separating the fiber clumps treated in the previous process into individual fibers, forming a uniform fiber web. The integrity of this web structure and the arrangement of the fibers directly affect the continuity of subsequent production processes (drafting, twisting), ultimately influencing key indicators such as yarn evenness, strength, and hairiness. Therefore, a deep understanding and effective control of the carded fiber web structure is a core scientific issue and key technology for achieving efficient and high-quality spinning.

[0003] For a long time, the optimization of spinning processes has relied heavily on empirical adjustments from numerous spinning experiments, and the evaluation of fiber web quality has been largely limited to macroscopic indicators, lacking visualization and quantitative analysis methods for the true three-dimensional microstructure of the fiber web. Although there are some digital methods aimed at constructing fiber web models, most focus on the macroscopic stacking and random distribution of fibers, failing to fully incorporate the typical characteristics of fibers arranged with a certain degree of orientation in the combed fiber web, and especially lacking the ability to reasonably reconstruct the arrangement structure of individual fibers in three-dimensional space while preserving the true morphological details of individual fibers (such as bends and hooks).

[0004] CN120182473A discloses a fiber network generation method. This method generates multiple lines in a square space, performs polynomial interpolation to generate fiber segments with smooth trajectories, and then uses random cyclic settings to form a fiber distribution in space, ultimately constructing a fiber network model with a stacked distribution of fiber segments. While this fiber network model, constructed using a random generation algorithm, allows for parameterized control, it fails to accurately reflect the true morphology of the fibers and their arrangement characteristics within the network. CN112164134A discloses a random curve modeling method based on image processing. This method uses SEM to capture images of the microstructure of nanocellulose fibers, performs image processing to calculate the fiber diameter, and assigns the obtained diameter to a spline curve drawn at the same scale as the SEM image. The wireframe diagrams corresponding to the contours of each layer's curves are imported into modeling software to ultimately construct a three-dimensional fiber network. Although this approach increases the fiber diameter attribute and improves the geometric accuracy of the model, the size of the established three-dimensional model is significantly limited by the SEM magnification, thus affecting the applicability of this modeling scheme. CN112380741A discloses a finite element modeling method for random fiber meshes, which constructs beam element models through image processing. While this improves modeling efficiency, it still relies on random generation rather than the actual fiber distribution. CN112329300A discloses a micro-finite element modeling method for fiber-reinforced composite materials based on CT scan images. This method reconstructs the microstructure based on CT scans, achieving high modeling accuracy, but the equipment is expensive and the data processing is complex, making it difficult to widely apply to fiber web modeling and analysis in conventional spinning processes. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a three-dimensional modeling method for combed fiber web based on the original distribution characteristics of single fibers, which can efficiently construct a combed fiber web model that accurately represents the original distribution characteristics of single fibers, while reducing the modeling cost.

[0006] The technical solution adopted by this invention to solve its technical problem is: to provide a three-dimensional modeling method for combed fiber webs based on the original distribution characteristics of single fibers, including:

[0007] Tracer fibers with significantly different colors from the original fibers were prepared, and the two were blended to obtain a blended fiber web.

[0008] Two-dimensional images of the blended fiber web on different microscope focusing planes are obtained, the tracer fibers in each image are identified, and the two-dimensional coordinate points on the corresponding plane are obtained;

[0009] The three-dimensional point cloud data of the tracer fiber is obtained by combining each two-dimensional coordinate point with the height of its corresponding plane.

[0010] The main feature points of the tracer fiber are extracted from the 3D point cloud data, and then a continuous fiber centerline is fitted.

[0011] A single fiber model is constructed based on the fiber centerline. The single fiber model is then subjected to free fall motion in a random order according to a set orientation to obtain a three-dimensional model of the fiber web.

[0012] Furthermore, the step of extracting the backbone feature points of the tracer fiber based on the three-dimensional point cloud data includes:

[0013] The 3D point cloud data is divided into multiple clusters;

[0014] Extract the two-dimensional contour points of each cluster projected onto the horizontal plane;

[0015] Based on the analysis of the two-dimensional contour points, the main extension direction of the corresponding cluster is analyzed. The cluster is evenly segmented along its main extension direction and the weighted centroid of each segment is calculated to obtain the main feature points of the tracer fiber.

[0016] Furthermore, fitting the continuous fiber centerline includes:

[0017] Construct the distance matrix of the main feature points and generate the minimum spanning tree;

[0018] Search for the longest path in the minimum spanning tree;

[0019] By fitting the centerline of the main feature points on the longest path, the fiber centerline is obtained.

[0020] Furthermore, centerline fitting is performed along the main extension direction.

[0021] Furthermore, the step of identifying tracer fibers in each image to obtain two-dimensional coordinate points on the corresponding plane includes:

[0022] The acquired 2D image is converted to the HSV color space and denoised to obtain the processed image;

[0023] Based on the color threshold, the tracer fibers in each processed image are identified, and the pixel coordinates on the corresponding plane are obtained;

[0024] The coordinates of pixels with gradient magnitudes greater than a set threshold are retained and mapped to real physical coordinates to obtain two-dimensional coordinate points on the corresponding plane.

[0025] Furthermore, the mapping to real physical coordinates is achieved by multiplying the pixel coordinates by the physical length corresponding to each unit pixel.

[0026] Furthermore, the construction of the single-fiber model based on the fiber centerline includes:

[0027] Import the fiber centerline into the simulation software;

[0028] Using the three-dimensional data points on the fiber centerline as the center coordinates of the particle entity, and the average diameter of the fiber as the diameter of the particle entity, the particle entities are bonded together into a single fiber model according to the set strength.

[0029] Furthermore, the set strength is infinity.

[0030] Furthermore, the different microscope focusing planes are arranged at equal intervals in the height direction.

[0031] Furthermore, the tracer fiber accounts for 0.5% to 2% of the mass of the blended fiber web.

[0032] Beneficial effects

[0033] By adopting the above-mentioned technical solution, the present invention has the following advantages and positive effects compared with the prior art:

[0034] This invention uses dyeing tracing and micro-chromatographic techniques to directly obtain the distribution data of real fibers in three-dimensional space, overcoming the shortcomings of traditional random generation models that cannot reflect the actual fiber morphology and arrangement characteristics.

[0035] This invention uses a common optical microscope and image processing methods, avoiding the use of high-cost equipment such as CT and SEM, and significantly reducing the economic and technical threshold for modeling;

[0036] This invention analyzes and identifies the main extension direction of the fiber, ensuring that subsequent processing proceeds along the natural direction of the fiber and avoiding morphological distortion caused by cross-cutting the fiber. In addition, by using segmented weighted centroids with Gaussian weights to highlight the core points of the fiber edge and trunk, redundant noise inside the fiber can be effectively filtered out, while retaining key features describing the morphology. By combining minimum spanning tree with the longest path algorithm, the longest connection is extracted from discrete centroid points, accurately restoring the trunk structure of the fiber.

[0037] The model, constructed through particle bonding, simulates the mechanical properties of real fibers and conforms to the deformation laws of fibers during carding and drafting. The constructed fiber model possesses both flexibility and mechanical realism, and can be further used for fiber behavior simulation under multi-physics fields (such as electric fields and flow fields), providing a reliable model foundation for process optimization and performance prediction. At the same time, the constructed fiber web model can realistically reflect the orientation characteristics and spatial overlap structure of fibers after carding, providing theoretical basis and model support for optimizing carding process parameters, evaluating fiber web uniformity, and digital design of subsequent spinning processes.

[0038] This invention can be widely applied to the field of textile carding web modeling and multiphysics simulation. It is not only suitable for the structural analysis of short fiber webs such as cotton and wool, but can also be extended to the modeling and research of fiber network materials such as nonwoven fabrics and composite preforms. The constructed web model can be placed in external field environments such as electric field and flow field to study the kinematic behavior of fibers under external influence, mutual entanglement mechanism and network formation law, thereby providing a simulation platform and theoretical support for the optimization design and intelligent control of advanced process parameters such as electric field assisted carding and airflow web formation. Attached Figure Description

[0039] Figure 1 This is a flowchart of an embodiment of the present invention;

[0040] Figure 2 This is a flowchart of a preferred embodiment 1 of the present invention;

[0041] Figure 3 This is a schematic diagram of image recognition and three-dimensional point set according to a preferred embodiment of the present invention;

[0042] Figure 4 This is a schematic diagram of the fitting centroid and curve of the preferred embodiment of the present invention;

[0043] Figure 5 This is a single-fiber model diagram of preferred embodiment 1 of the present invention;

[0044] Figure 6 This is a three-dimensional model diagram of the fiber web of the preferred embodiment 1 of the present invention. Detailed Implementation

[0045] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be understood that after reading the teachings of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent forms also fall within the scope defined by the appended claims.

[0046] The embodiments of the present invention relate to a method for three-dimensional modeling of combed fiber webs based on the original distribution characteristics of single fibers, such as... Figure 1 As shown, it includes:

[0047] Tracer fibers with significantly different colors from the original fibers were prepared, and the two were blended to obtain a blended fiber web.

[0048] Two-dimensional images of the blended fiber web on different microscope focusing planes are obtained, the tracer fibers in each image are identified, and the two-dimensional coordinate points on the corresponding plane are obtained;

[0049] The three-dimensional point cloud data of the tracer fiber is obtained by combining each two-dimensional coordinate point with the height of its corresponding plane.

[0050] The main feature points of the tracer fiber are extracted from the 3D point cloud data, and then a continuous fiber centerline is fitted.

[0051] A single fiber model is constructed based on the fiber centerline. The single fiber model is then subjected to free fall motion in a random order according to a set orientation to construct a three-dimensional fiber web model.

[0052] The tracer fiber can be prepared by immersing and dyeing the fiber in a mixed solution containing dye. The fiber can be cotton fiber, wool staple fiber, etc., and the dye can be rhodamine, indigo pigment, etc. The preferred mass fraction of the dye is 1%~2% (omf). After dyeing, the fiber is taken out and washed with water, and then dried after repeating 3-5 times.

[0053] Then, 0.5-2% of the dried dyed fiber is thoroughly mixed with the undyed fiber and fed into a carding machine for 2-3 carding processes to obtain a blended fiber web.

[0054] The blended fiber web was cut into rectangular samples for observation and placed horizontally under a microscope. The focusing plane was changed at equal intervals by adjusting the fine-tuning knob, and the height of each focusing plane was recorded as... A series of two-dimensional images were obtained. In these images, the side length of the sample was greater than the average length of a single fiber, and the sample was placed in the microscope observation area without applying any pressure along the thickness direction.

[0055] Locate the tracer fibers and adjust the magnification so that the tracer fibers are fully visible in the field of view. Adjust the microscope's initial focusing plane (height denoted as...) by rotating the fine-tuning knob. =0) is located on the upper surface of the tracer fiber. Then, the fine-tuning knob is rotated continuously at equal angles to change the focusing plane height (denoted as ). ), and a series of two-dimensional images are obtained by recording them sequentially.

[0056] and , The relationship between them is as follows: .

[0057] in, The knob rotation angle, The height that rises by 1° when the fine adjustment knob is turned can be obtained by testing an object of known thickness and rotating the fine adjustment knob by the angle at which the focusing plane is lowered from the upper surface to the lower surface.

[0058] The obtained two-dimensional image is processed using a Python program. Color thresholding is used to identify the tracer fiber and obtain its coordinates within the corresponding two-dimensional image. , The three-dimensional coordinates of this point are determined by combining the height of the focal plane at that point with the coordinates of the focal plane. , , This operation is performed on all the resulting series of two-dimensional images to obtain the spatial coordinate point set representing the tracer fiber.

[0059] In some preferred embodiments, to obtain more accurate modeling results, the key morphological coordinate points describing the fiber can be selected by using gradient magnitude. The specific method is as follows:

[0060] 1) Edit a Python program to process a series of two-dimensional images, convert them to the HSV color space, remove noise using morphological and Gaussian blurring, and identify and extract the set of pixel coordinates containing the dye color {( , )};

[0061] 2) Calculate the coordinate set {( , The gradient magnitude of the coordinate point within the original image region is selected, and the gradient magnitude greater than a certain set value is retained. The coordinates of the point {( , This method preserves the edge contour pixels that describe the core features of the fiber shape.

[0062] 3) Then convert the coordinate point to physical coordinates consistent with the fiber length {( , The three-dimensional coordinates of the point are then determined by combining the height of the focal plane at that point with the coordinates of the focal plane. , , ), thereby obtaining the three-dimensional spatial coordinate point set representing the tracer fiber {( , , )}.

[0063] , and , The relationship between them is as follows:

[0064]

[0065]

[0066] in, To determine the actual object length corresponding to a unit pixel in an image, one can take a picture of an object of known length and use a program to calculate the number of pixels corresponding to that length.

[0067] The modeling process can be carried out in simulation software. The obtained three-dimensional spatial coordinate point set is drawn into a point cloud. After determining the neighborhood radius, the point cloud is divided into multiple clusters, and the XY plane projection of each cluster is contoured. PCA analysis is performed on the XY coordinates within the cluster to obtain its main extension direction. Then, the cluster is segmented along its main direction, and Gaussian weighted centroid is calculated for each segment.

[0068] When performing PCA analysis, the eps parameter of DBSCAN can be automatically calculated using the quantiles of the k-nearest neighbor distance. DBSCAN is then run to identify dense regions and divide the point cloud into multiple clusters. Each cluster is then projected onto the XY plane, and the projected area contour is extracted. The mean Z-coordinate of the cluster is then added to the 2D contour points to restore them to 3D points. PCA analysis is performed on the XY coordinates of the clusters to obtain the main extension direction. Based on the projection distance, the clusters are divided into multiple segments. Gaussian weights (with higher weight for the center point) are used to define the weights of each contour point within a segment, and the weighted sum is calculated as the centroid of each segment, serving as the backbone feature point.

[0069] Construct the distance matrix of the centroids and generate the minimum spanning tree (MST). Find the longest path in the MST as the backbone and fit the centerline of the ordered centroids to obtain the fiber centerline.

[0070] The fiber centerline is imported into the modeling software, and a single-fiber model is constructed by combining continuous particle entities. The 3D data points serve as the center coordinates of the particle entities, and the diameter of each particle entity matches the average diameter of the fiber. During data acquisition, the fiber diameter is used as the step size. The bonding force between the particle entities is also set to ensure the flexibility and continuity of the fiber entities, resulting in a single-fiber 3D model.

[0071] A three-dimensional space is constructed, and a series of single fiber models are oriented along a certain side. Since the fibers are relatively randomly distributed during the combing process, the single fiber models are subjected to free fall in random order to achieve contact and overlap between fibers, and finally the three-dimensional model of combed fiber web (RVE model) with fiber orientation arrangement is constructed.

[0072] A preferred embodiment 1 of the present invention is a method for three-dimensional modeling of combed fiber webs based on the original distribution characteristics of single fibers, such as... Figure 2 As shown, it includes the following steps:

[0073] S1: Prepare dyed tracer fibers and blend them with the original fibers in a certain proportion to obtain a blended fiber web;

[0074] S2: Place the blended fiber web obtained in S1 horizontally under a microscope, and change the focusing plane at equal intervals by adjusting the fine adjustment knob. Its height gauge is Zn, and a series of two-dimensional images are obtained.

[0075] S3: The two-dimensional image obtained in S2 is processed using a Python program. The color threshold is used to identify the tracer fiber and obtain its coordinates in the corresponding two-dimensional image. , The three-dimensional coordinates of this point are determined by combining the height of the focal plane at that point with the coordinates of the focal plane. , , This operation is performed on all the resulting series of two-dimensional images to obtain the spatial coordinate point set representing the tracer fiber.

[0076] S4: Draw the 3D point set obtained in S3 into a point cloud, determine the neighborhood radius, run DBSCAN, divide the point cloud into multiple clusters, perform contour recognition on the XY plane projection of each cluster, perform PCA on the XY coordinates within the cluster, and segment along its main direction, and calculate the weighted centroid of each segment using Gaussian weighting.

[0077] S5: Construct the distance matrix of the centroids in S4 and generate the MST. Find the longest path in the MST as the backbone and fit the centerline of the ordered centroids.

[0078] S6: Import the fiber centerline obtained based on S1-S5 into EDEM, set up a continuous particle entity combination to construct a single fiber model, where the diameter of the particle entity is consistent with the average diameter of the fiber, and set the bonding force between the particle entities to ensure the flexibility and continuity of the fiber entity, and finally obtain the fiber three-dimensional model. The obtained single fiber model is subjected to free fall motion in random order to construct the oriented fiber web three-dimensional model.

[0079] Step S1 is detailed below:

[0080] S11: The fiber is immersed in a mixed solution containing dye for dyeing. The fiber can be cotton fiber, wool staple fiber, etc., and the dye can be rhodamine, indigo pigment, etc. The mass fraction of the dye is 1%~2% (omf). After that, it is taken out and washed with water. This process is repeated 3-5 times and then dried.

[0081] S12: 0.5-2% of dry dyed fibers are thoroughly mixed with undyed fibers and fed into a carding machine for 2-3 carding processes to obtain a blended fiber web.

[0082] Step S2 is detailed below:

[0083] S21: Cut the blended fiber web into rectangular specimens for observation. The side length of each specimen is greater than the average length of a single fiber. The length is 5cm and the width is 3cm. Place the specimen in the microscope observation area without applying pressure along the thickness direction.

[0084] S22: Locate the tracer fiber and adjust the magnification so that the tracer fiber appears completely in the field of view. Adjust the initial focus plane (height denoted as S22) by rotating the fine-tuning knob. =0) is located on the upper surface of the tracer fiber, and then continuously at equal angles (denoted as ). Rotate the fine-tuning knob to change the focus plane height (denoted as...). ), and sequentially record to obtain a series of two-dimensional images;

[0085] in and , The relationship between them is shown in the following formula:

[0086]

[0087] in The height that rises by 1° when the fine adjustment knob is turned can be obtained by testing an object of known thickness and rotating the fine adjustment knob by the angle at which the focusing plane is lowered from the upper surface to the lower surface.

[0088] Step S3 is as follows: Figure 3 As shown:

[0089] S31: Edit a Python program to process a series of two-dimensional images, convert them to the HSV color space, remove noise using morphological and Gaussian blurring, and identify and extract the set of pixel coordinates containing the dye color {( , )};

[0090] S32: Calculate the coordinate set {( , The gradient magnitude of the coordinate point within the original image region is selected, and the gradient magnitude greater than a certain set value is retained. The coordinates of the point {( , Then convert the coordinate point to the actual coordinates that match the fiber length. , The three-dimensional coordinates of the point are then determined by combining the height of the focal plane at that point with the coordinates of the focal plane. , , ), thereby obtaining the three-dimensional coordinate point set representing the tracking fiber {( , , )}.

[0091] , and , The relationship between them is as follows:

[0092]

[0093]

[0094] in To determine the actual length of an object corresponding to one pixel in an image, you can take a picture of an object of known length and use a program to calculate the number of pixels corresponding to that length.

[0095] Step S4 is as follows: Figure 4 As shown:

[0096] S41: Edit the Python program to convert the 3D point set {( , , The point cloud is plotted as a point cloud. The eps parameter of DBSCAN is automatically calculated using the quantile of the k-nearest neighbor distance. DBSCAN is then run to identify dense regions and divide the point cloud into multiple clusters.

[0097] S42: Project each cluster onto the XY plane, extract the projected area contour, and then add the average Z coordinate of the cluster to the two-dimensional contour points to restore them to 3D points;

[0098] S43: Perform PCA on the XY coordinates of the cluster to obtain the main extension direction, then divide the cluster into multiple segments according to the projection distance, and use Gaussian weights (with higher weight for the center point) to obtain the centroid of each segment;

[0099] In step S5, for example Figure 5 As shown, the specific steps are as follows: construct the distance matrix of the centroid points and generate the MST, find the global connectivity structure, use Dijkstra's algorithm to find the longest path in the MST as the trunk of the center line, and fit the discrete weighted centroids.

[0100] In step S6, the specific details are as follows: Figure 6 As shown:

[0101] S61: Import the 3D fiber centerline data obtained from S1-S5 into an EDEM, and construct a single-fiber model by combining continuous particle entities. The 3D data points serve as the center coordinates of the particle entities, and the diameter of each particle entity matches the average diameter of the fiber. During data acquisition, the fiber diameter is used as the step size. The particle entities are connected by bonds with infinite force to ensure the flexibility and continuity of the fiber entities. Based on the above operations, a series of 3D models of single fibers are obtained.

[0102] S62: Construct a three-dimensional space, arrange the obtained series of single fiber models along a certain side length, and let them fall freely in random order to achieve contact and overlap between fibers, and finally realize the construction of a three-dimensional model of the combed fiber web (RVE model) with fiber orientation arrangement.

Claims

1. A method for three-dimensional modeling of combed fiber webs based on the original distribution characteristics of single fibers, characterized in that, include: Tracer fibers with significantly different colors from the original fibers were prepared, and the two were blended to obtain a blended fiber web. Two-dimensional images of the blended fiber web on different microscope focusing planes are obtained, the tracer fibers in each image are identified, and the two-dimensional coordinate points on the corresponding plane are obtained; The three-dimensional point cloud data of the tracer fiber is obtained by combining each two-dimensional coordinate point with the height of its corresponding plane. The main feature points of the tracer fiber are extracted from the 3D point cloud data, and then a continuous fiber centerline is fitted. A single fiber model is constructed based on the fiber centerline. The single fiber model is then subjected to free fall motion in a random order according to a set orientation to obtain a three-dimensional model of the fiber web.

2. The method according to claim 1, characterized in that, The step of extracting the backbone feature points of the tracer fiber based on the 3D point cloud data includes: The 3D point cloud data is divided into multiple clusters; Extract the two-dimensional contour points of each cluster projected onto the horizontal plane; Based on the analysis of the two-dimensional contour points, the main extension direction of the corresponding cluster is analyzed. The cluster is evenly segmented along its main extension direction and the weighted centroid of each segment is calculated to obtain the main feature points of the tracer fiber.

3. The method according to claim 2, characterized in that, The process of fitting a continuous fiber centerline includes: Construct the distance matrix of the main feature points and generate the minimum spanning tree; Search for the longest path in the minimum spanning tree; By fitting the centerline of the main feature points on the longest path, the fiber centerline is obtained.

4. The method according to claim 3, characterized in that, When performing centerline fitting, it is done along the main extension direction.

5. The method according to claim 1, characterized in that, The process of identifying tracer fibers in each image and obtaining corresponding two-dimensional coordinate points on the plane includes: The acquired 2D image is converted to the HSV color space and denoised to obtain the processed image; Based on the color threshold, the tracer fibers in each processed image are identified, and the pixel coordinates on the corresponding plane are obtained; The coordinates of pixels with gradient magnitudes greater than a set threshold are retained and mapped to real physical coordinates to obtain two-dimensional coordinate points on the corresponding plane.

6. The method according to claim 5, characterized in that, The mapping to real physical coordinates is achieved by multiplying the pixel coordinates by the physical length corresponding to each unit pixel.

7. The method according to claim 1, characterized in that, The construction of a single-fiber model based on the fiber centerline includes: Import the fiber centerline into the simulation software; Using the three-dimensional data points on the fiber centerline as the center coordinates of the particle entity, and the average diameter of the fiber as the diameter of the particle entity, the particle entities are bonded together into a single fiber model according to the set strength.

8. The method according to claim 7, characterized in that, The set strength is infinity.

9. The method according to claim 1, characterized in that, The different microscope focusing planes are arranged at equal intervals in the height direction.

10. The method according to claim 1, characterized in that, The tracer fiber accounts for 0.5% to 2% of the mass of the blended fiber web.

Images