A three-dimensional modeling method based on yarn hairiness characteristics

Abstract

The present application relates to a kind of three-dimensional modeling methods based on yarn hair feature, comprising the following steps: intercepting yarn sample, the yarn hair image of yarn sample is obtained by camera equipment;The yarn hair image is preprocessed, and the preprocessed image is obtained, and the hair path matching tracking algorithm is used to extract the hair skeleton in preprocessed image, obtain the feature library of yarn hair on macroscopic;Combining the fiber curve path in feature library, fiber radius, hair length, hair random distribution coefficient, hair length coefficient, hair distortion coefficient, hair lodging coefficient, hair section radius, hair point ordinal maximum and hair number constructs hair three-dimensional model.The present application utilizes computer vision to detect the hair skeleton of yarn, so as to obtain the feature of yarn hair on macroscopic, and establish four coefficients and hair shape formation relationship, make two-dimensional yarn hair modeling to three-dimensional parameterization modeling.

Description

Technical Field

[0001] This invention relates to the field of yarn 3D modeling technology, specifically a 3D modeling method based on yarn hairiness characteristics. Background Technology

[0002] Hair refers to the part of some fiber ends that protrude from the yarn body or arch up in a loop from the yarn surface. Hair is caused by a combination of factors that accompany fiber movement in the production process. The causes of hair are broken fibers and excessively twisted hooked fibers in the raw cotton. It is formed due to the irregular movement of fibers caused by irregular fiber arrangement and the interaction forces during the stretching process.

[0003] Hair is mainly generated in the twisting triangle area during yarn spinning. Some hair is generated by friction during subsequent processing and use. The quantity and length of hair determine the quality of yarn and fabric. Too much hair can easily cause itching and friction in dry environments.

[0004] With the development of science and the advancement of computer technology and image processing and analysis technology, yarn images can be directly acquired using machine vision and digital image processing. The appearance image information of the yarn can then be directly obtained, and a 3D model of the yarn can be created based on the appearance image information. However, current 3D modeling of yarn is limited to the yarn part and basically does not involve the content of hairiness. The mainstream 3D modeling software defines hairiness as a fluffy effect, and hairiness modeling is mostly still at the level of 2D images. This means that the fidelity and accuracy of 3D modeling of yarn and fabrics need to be improved. Summary of the Invention

[0005] The technical problem to be solved by this invention is to provide a three-dimensional modeling method based on yarn hairiness characteristics to address the above-mentioned shortcomings. This invention uses computer vision to detect the yarn hairiness skeleton, thereby obtaining the macroscopic characteristics of yarn hairiness, and establishing the relationship between four coefficients and hairiness morphology. It moves from traditional two-dimensional yarn hairiness modeling to three-dimensional parametric modeling, and combines graphic fractal theory to draw hairiness, thereby improving the fidelity and accuracy of yarn hairiness modeling.

[0006] To solve the above technical problems, the present invention adopts the following technical solution:

[0007] A 3D modeling method based on yarn hairiness features includes the following steps:

[0008] Step 1: Cut a yarn sample and use a camera to obtain an image of the yarn hairiness of the yarn sample;

[0009] Step 2: Preprocess the yarn hair image to obtain a preprocessed image, and use the hair path matching and tracking algorithm to extract the hair skeleton in the preprocessed image to obtain a macroscopic feature library of yarn hair.

[0010] Step 3: Construct a three-dimensional model of the hair by combining the fiber curve path, fiber radius, hair length, hair random distribution coefficient, hair length coefficient, hair twist coefficient, hair lodging coefficient, hair cross-sectional radius, maximum hair point number, and hair number in the feature library.

[0011] Furthermore, the camera device is an industrial camera.

[0012] Further preprocessing of the yarn fuzz image includes image grayscale conversion, filtering, OTSU image segmentation, and morphological operations.

[0013] Furthermore, the method for constructing the three-dimensional model of the feathers includes the following steps:

[0014] Step a1: Using the fiber curve path and fiber radius R y Establish fiber space, through hair length l min Establish a hair space outside the fiber space;

[0015] Step a2: Based on the random distribution coefficient of feathers Determine the initial positions P of nh randomly distributed hairs on the outer surface of the yarn space. 0，m m = [1, nh];

[0016] Step a3, with the initial position P of the feathers 0，m Let P be the initial point, and let P be the projection of the initial position of the feather onto the outer surface of the feather space. i，m Construct the initial vector of the feathers for the endpoint And based on the feather length coefficient and feather initial vector Calculate the reference normal vector L of the feathers h ;

[0017] Step a4: Let the initial vector of the m-th feather be given by the i-th point P. j,m Composition, j = [0, i], m = [1, nh], j ∈ N, m ∈ N; based on the random feather distortion coefficient Calculate the random twist amount of the hairs at each point on all the hairs. Based on the random feather lodging coefficient Calculate the amount of random lodging of feathers at each point on all feathers. Based on the reference normal vector L h , distortion amount and lodging amount Calculate the position of each point on all the feathers to obtain the point set matrix P of the feathers:

[0018]

[0019] Connect all points in the same column of the point set matrix P to obtain nh feather curves. Based on the feather curves and the feather cross-sectional radius r... h Construct a feather model.

[0020] Furthermore, in step 1, the hair length l min It is less than the tilt height of the twisting triangle zone during spinning.

[0021] Furthermore, the reference normal vector L of the feathers is calculated. h The formula is:

[0022]

[0023] in, Let l be the initial vector of the yarn surface hairiness. min Let l be the vector length. h for Random numbers within a range This is the feather length coefficient.

[0024] Furthermore, when calculating the amount of random fall of feathers, 95% of the feathers are randomly selected as terminal feathers, and all points in the terminal feathers except the initial point are randomly twisted; 5% of the feathers are randomly selected as floating feathers, and all points in the floating feathers are randomly twisted.

[0025] Furthermore, in step a4, the random twisting amount of the feathers is calculated. The formula is:

[0026]

[0027] Where, θ a and θ b It is a random array, ranging from [0°, 360°]; r ran It is a random array, with a range of This is the lodging coefficient.

[0028] Furthermore, in step a4, the amount of random lodging of feathers is calculated. The formula is:

[0029]

[0030] Where τ is 1 or -1, indicating whether the feathers are tilted forward or backward. The lodging coefficient is... The basic values ​​for the variable representing the lodging of feathers, A radius of 2 times R is formed from the initial position P0 of the corresponding feather. y It is obtained by connecting the points where the spherical space and the yarn curve intersect.

[0031] Furthermore, in step a4, point P j,m With reference normal vector L h , distortion amount and lodging amount The relationship formula is:

[0032]

[0033] Among them, P 0，m Let m be the initial position of the feathers, where m = [1, nh] and i is the total number of points.

[0034] Compared with the prior art, the present invention, by adopting the above technical solution, has the following advantages:

[0035] Compared to current methods that only focus on two-dimensional modeling of fabric hair and rarely involve hair in yarn modeling, this invention significantly improves the fidelity and accuracy of yarn fabric modeling. Traditional two-dimensional hair modeling methods, which are closer to visual perception, cannot be directly applied to three-dimensional modeling. However, this invention utilizes computer vision to extract four hair coefficients, establishes the relationship between hair morphology and the four coefficients, and thus achieves three-dimensional parametric modeling of hair. Hair modeling using this method has high fidelity and accuracy.

[0036] This invention extracts the yarn hair skeleton from physical samples and summarizes the corresponding hair feature library. Based on the feature library and four hair coefficients, and combined with graphic fractal theory, a high-fidelity and high-precision yarn hair model is finally obtained. On the basis of traditional yarn modeling, hair modeling has moved from two-dimensional to three-dimensional, making the details of the three-dimensional yarn model richer. The parametric relationship between the yarn hair features of physical samples and model construction is established, and the influence of hair features on the appearance and quality of yarn fabrics can be compared more intuitively with the help of finite element analysis.

[0037] Furthermore, the yarn hair model obtained by this invention has high precision and rich details, and can be applied to scenarios such as finite element analysis, advertising and promotion, and scientific research drawing.

[0038] The present invention will now be described in detail with reference to the accompanying drawings and embodiments. Attached Figure Description

[0039] Figure 1 This is a schematic diagram illustrating the construction of the feather space in an embodiment of the present invention;

[0040] Figure 2 This is a schematic diagram illustrating the setting of random initial positions for feathers in an embodiment of the present invention;

[0041] Figure 3 (a) is a schematic diagram of the normal vector of the feathers in an embodiment of the present invention;

[0042] Figure 3 (b) is a schematic diagram of the feather twisting variable in an embodiment of the present invention;

[0043] Figure 4 This is a schematic diagram of the feathers lying flat in an embodiment of the present invention, wherein: (a) the figure shows Figure 1 shows the representation of different values; Figure 2b shows the representation when τ is 1 or -1.

[0044] Figure 5 The diagram shows the length of the feathers, where (a) represents short feathers, (b) represents medium feathers, and (c) represents long feathers.

[0045] Figure 6 (d) is a picture of a physical sample of cotton yarn; (e) is a top view of a yarn fuzz model; (f) is a perspective view of a yarn fuzz model.

[0046] Figure 7 The following are fabric effect diagrams based on mechanical simulation, where: (a) represents the simulation model, (b) represents the fabric sample, (c) represents the perspective model, and (d) represents the fabric structure diagram;

[0047] Figure 8 This is how the yarn hairiness model is represented in UE5. Detailed Implementation

[0048] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0049] A 3D modeling method based on yarn hairiness features includes the following steps:

[0050] Step 1: Cut a yarn sample and use a camera to obtain an image of the yarn hairiness of the yarn sample;

[0051] Step 2: Preprocess the yarn hair image to obtain a preprocessed image, and use the hair path matching and tracking algorithm to extract the hair skeleton in the preprocessed image to obtain a macroscopic feature library of yarn hair.

[0052] Step 3: Construct a three-dimensional model of the hair by combining the fiber curve path, fiber radius, hair length, hair random distribution coefficient, hair length coefficient, hair twist coefficient, hair lodging coefficient, hair cross-sectional radius, maximum hair point number, and hair number in the feature library.

[0053] Example 1:

[0054] 1. Take yarn samples, use an industrial camera to acquire images of the yarn and its fuzz, repeat 10 times, and enter the data into the computer system.

[0055] 2. By converting the yarn hairy image to grayscale, filtering, segmenting the image using OTSU, and performing morphological operations, a preprocessed image is obtained. The hairy skeleton is extracted using the hairy path matching and tracking algorithm. The obtained data can be used to obtain a macroscopic feature library of yarn hairy.

[0056] 3. For example Figure 1 As shown, we establish the hairline space by using the fiber curve path, radius, and hairline setting length. Generally, l min It is less than the tilt height of the twisting triangle zone during spinning.

[0057] 4. Determine the initial position of the feathers. 0，m The position m = [1, nh] represents the initial position of the fuzz, which is attached to the yarn surface and randomly distributed. The position of the fuzz can be determined by setting the initial point P. 0，m The hairs are randomly distributed on the surface of the object to determine the total number of hairs, nh, and the number of seeds. Let be the random distribution coefficient of the feathers. The relationship of the random array involved is (m = nh). The initial point P is obtained. 0，m ,See Figure 2 Initial point P 0，m Constructing a matrix (P) 0,m =[P 0,1 P 0,2 ... P 0,nh ]).

[0058] 5. Determine the feather length vector. For example... Figure 3 As shown in Figure (a), P on the yarn 0，m Point P is generated by projecting a point onto the surface of the feather space. i，m Then with P 0，m Let P be the initial point. i，m Construct an initial vector for the feathers at the endpoint. L h The normal vector for the feathers is given by the following formula:

[0059]

[0060] In the formula, Let l be the initial vector based on the hairiness of the yarn surface. min , Feather length coefficient, l h for A random number within a given range.

[0061] 4. Determine the feather twist coefficient. For example... Figure 3 As shown in Figure (b), the amount of distortion has a random coefficient. Feathers are composed of i points P j,m Composition: j = [0, i], m = [1, nh], j ∈ N, m ∈ N. P j,mLet P represent the j-th point on the m-th hair. When terminal hairs are formed, j∈[1,i], the point is randomly displaced in the spherical space, causing the yarn to twist randomly. When wild hairs are formed (floating hairs), P 0，m Random displacement is also performed. It is the random distortion of the feathers, and the formula is:

[0062]

[0063] In the formula, θ a and θ b It is a random array, ranging from [0°, 360°], r ran It is a random array, with a range of

[0064] 5. Determine the feather lodging coefficient. For example... Figure 4 (a) and Figure 4 As shown in (b), there is a vector along the feather axis. As a basic value of the variable of feather lodging, the amount of feather lodging

[0065]

[0066] In the formula, the random coefficient τ is the lodging coefficient, which is 1 or -1, indicating whether the feathers are facing forward or backward. A radius of 2 times R is formed from the initial point P0 of the feathers. y The points where the spherical space and the yarn curve intersect are connected to obtain the shape; each wool has its own...

[0067] 6. Construct a 3D model of yarn hairiness using the sample yarn hairiness feature library. The final formula for the relationship between hairiness points is:

[0068]

[0069] Finally, the point set matrix P of the feathers is obtained. j,m :

[0070]

[0071] By P j,m Connecting points with the same value m in the set matrix (j=[0,i],m=[1,nh],j,m∈N) yields the feather curve, and a feather model is constructed based on the curve.

[0072] Based on the requirements of feather shape and characteristic indicators, the following parameter needs to be set: feather random distribution coefficient. Feather length coefficient Deformation coefficient Lodging coefficient Feather cross-sectional radius r h , Feather setting length l min , maximum value of feather point ordinal number i, number of feather roots nh.

[0073] The main equipment required for this invention includes: an industrial camera, a standard light source box, and a computer;

[0074] Working principle: Yarn samples are cut and processed, and multiple sets of yarn data are captured using an industrial camera to expand the sample data. Image processing technology is used to extract the yarn hairiness skeleton, and a hairiness feature library is summarized and compiled based on this (see Table 1). The specific values ​​of the four hairiness coefficients are determined by the feature library, and a corresponding three-dimensional yarn hairiness model is constructed based on Rhino and Grasshopper (see Table 1). Figure 5 (include Figure 5 Figures (a), (b), and (c) in the middle are shown. Figure 6 (include Figure 6 Figures (d), (e), and (f) show the yarn fuzz model. This model can be applied to finite element analysis, advertising, scientific research drawing, and other scenarios (see Figures (d), (e), and (f)). Figure 7 (include Figure 7 Figures (a), (b), (c), and (d) in the middle are... Figure 8 .

[0075]

[0076] Table 1. Yarn Hairiness Feature Database

[0077] Currently, modeling of yarn fibers remains at a two-dimensional level, while three-dimensional modeling of yarns rarely involves yarn fibers. There is significant room for improvement in the fidelity and accuracy of yarn fabric modeling. Furthermore, two-dimensional yarn fiber modeling is generally a more visual and traditional method, and cannot be directly applied to three-dimensional modeling.

[0078] This patent aims to extract the yarn hair skeleton from physical samples and summarize the hair feature library of such samples. Based on the feature library and four hair coefficients, combined with graphic fractal theory, a high-fidelity and high-precision yarn hair model is finally obtained. On the basis of traditional yarn modeling, hair has moved from two-dimensional to three-dimensional. The details of the three-dimensional yarn model are also richer, making the parameterized relationship between the yarn hair features of the physical sample and the model construction more intuitive. The influence of hair features on the appearance and quality of yarn fabric can be compared more intuitively with the help of finite element analysis.

[0079] The above description provides examples of the preferred embodiments of the present invention. Parts not detailed herein are common knowledge to those skilled in the art. The scope of protection of the present invention is determined by the claims. Any equivalent modifications based on the technical teachings of the present invention are also within the scope of protection of the present invention.

Claims

1. A three-dimensional modeling method based on yarn hairiness characteristics, characterized in that, Includes the following steps: Step 1: Cut a yarn sample and use a camera to obtain an image of the yarn hairiness of the yarn sample; Step 2: Preprocess the yarn hair image to obtain a preprocessed image, and use the hair path matching and tracking algorithm to extract the hair skeleton in the preprocessed image to obtain a macroscopic feature library of yarn hair. Step 3: Construct a three-dimensional model of the hair by combining the fiber curve path, fiber radius, hair length, hair random distribution coefficient, hair length coefficient, hair twist coefficient, hair lodging coefficient, hair cross-sectional radius, maximum hair point number, and hair number in the feature library. The method for constructing the three-dimensional model of the feathers includes the following steps: Step a1: Using the fiber curve path and fiber radius R y Establish fiber space, through hair length l min Establish a hair space outside the fiber space; Step a2: Based on the random distribution coefficient of feathers Determine the initial positions P of nh randomly distributed hairs on the outer surface of the yarn space. 0，m m = [1, nh]; Step a3, with the initial position P of the feathers 0，m Let P be the initial point, and let P be the projection of the initial position of the feather onto the outer surface of the feather space. i，m Construct the initial vector of the feathers for the endpoint And based on the feather length coefficient and feather initial vector Calculate the reference normal vector L of the feathers h ; Step a4: Let the initial vector of the m-th feather be given by the (i+1)-th point P. j,m Composition, j = [0, i], m = [1, nh], j ∈ N, m ∈ N; based on the random feather distortion coefficient Calculate the random twist amount of the hairs at each point on all the hairs. Based on the random feather lodging coefficient Calculate the amount of random lodging of feathers at each point on all feathers. Based on the reference normal vector L h , distortion amount and lodging amount Calculate the position of each point on all the feathers to obtain the point set matrix P of the feathers: Connect all points in the same column of the point set matrix P to obtain nh feather curves. Based on the feather curves and the feather cross-sectional radius r... h Construct a feather model.

2. The three-dimensional modeling method based on yarn hairiness features according to claim 1, characterized in that, The camera device is an industrial camera.

3. The three-dimensional modeling method based on yarn hairiness features according to claim 1, characterized in that, The preprocessing of the yarn fuzz image includes image grayscale conversion, filtering, OTSU image segmentation, and morphological operations.

4. The three-dimensional modeling method based on yarn hairiness features according to claim 1, characterized in that, In step 1, the hair length l min It is less than the tilt height of the twisting triangle zone during spinning.

5. The three-dimensional modeling method based on yarn hairiness features according to claim 1, characterized in that, Calculate the reference normal vector L of the feathers h The formula is: in, Let l be the initial vector of the yarn surface hairiness. h for Random numbers within a range This is the feather length coefficient.

6. The three-dimensional modeling method based on yarn hairiness features according to claim 1, characterized in that, When calculating the amount of random fall of feathers, 95% of the feathers are randomly selected as terminal feathers, and all points in the terminal feathers except the initial point are randomly twisted; 5% of the feathers are randomly selected as floating feathers, and all points in the floating feathers are randomly twisted.

7. The three-dimensional modeling method based on yarn hairiness characteristics according to claim 1, characterized in that, In step a4, the random twisting amount of the hair is calculated. The formula is: Where, θ a and θ b It is a random array, ranging from [0°, 360°]; r ran It is a random array, with a range of This is the lodging coefficient.

8. The three-dimensional modeling method based on yarn hairiness features according to claim 1, characterized in that, In step a4, the amount of random lodging of feathers is calculated. The formula is: Where τ is 1 or -1, indicating whether the feathers are tilted forward or backward. The lodging coefficient is... The basic values ​​for the feather-feather-collapse variable. From the initial position P of the corresponding feather 0,m Formation radius is 2 times R y It is obtained by connecting the points where the spherical space and the yarn curve intersect.

9. The three-dimensional modeling method based on yarn hairiness features according to claim 1, characterized in that, In step a4, point P j,m With reference normal vector L h , distortion amount and lodging amount The relationship formula is: Among them, P 0，m Let m be the initial position of the feather, where m = [1, nh] and i is the maximum ordinal number of the point.

Images