CN117788887A - An automatic rating device and method for fabric pleat appearance retention grade - Google Patents

Abstract

The invention discloses an automatic grading method for a fabric pleat appearance retention grade, which comprises the following steps: (1) image acquisition: the method comprises the steps that an image acquisition device is used for acquiring images of projection of a grating on a test fabric by using a camera to obtain a grating image; (2) image preprocessing: firstly, performing primary image segmentation on the grating image acquired in the step (1), selecting a target area, and intercepting a required target area; noise reduction is carried out after the primary image segmentation, and a clear grating image is obtained; (3) pleat feature extraction: extracting morphological characteristics of the segmented grating; (4) rank classification and assessment: calculating the height distribution of the pleats by analyzing the grating form and the angle of incidence points, analyzing to obtain a regression equation, comparing and verifying with a standard template grating image, and evaluating the grade value of the pleat retention of the fabric; three measurements were repeated and the average was taken as the final grade.

Description

An automatic rating device and method for fabric pleat appearance retention grade

Technical field

The invention belongs to the field of fabric pleat testing, and specifically relates to an automatic rating device and method for the appearance retention level of fabric pleats.

Background technique

In clothing styling design, pleats have both practical and decorative functions, and their shape retention is closely related to fiber properties, yarn properties, fabric structure and care conditions. Currently, GB/T 13770-2009 "Test method for textiles to evaluate the pleated appearance of fabrics after washing", ISO 7769:2009 "Textile fabrics - Test method for evaluation of pleated appearance after home washing and drying", AATCC 88C-2018t "Fabrics Standards such as "Pleat Retention after Home Laundering" use eye ratings to subjectively evaluate the pleat retention of fabrics after washing. Manual assessment has low accuracy and weak reproducibility, and requires multiple people to participate in the assessment, which is time-consuming and labor-intensive. There is a large gap between the current development needs for accurate and rapid detection.

The application of image processing technology provides a feasible alternative to this type of gaze assessment method. Existing research mainly evaluates the appearance of fabric pleats through two-dimensional images or three-dimensional reconstruction methods. The surface pleat characteristics of two-dimensional images, such as surface grayscale information, shadow area, etc., are greatly affected by ambient lighting, fabric surface patterns and textures. Three-dimensional reconstruction can overcome the influence of illumination, fabric pattern, texture and other factors on the collected images to a certain extent. However, as the accuracy improves, the equipment cost increases and the calculation and rating time are too long.

The detection and evaluation of fabric pleats using two-dimensional images mainly extracts surface features such as gray area, shadow area, and contour of the image, and compares and converts them to achieve detection and evaluation of pleats. In 1996, XU ^[1] et al. based on fabric wrinkle image analysis and rated fabric wrinkles by measuring the wrinkle gray surface area and shadow area. This method can quantitatively describe the degree and quality of fabric wrinkles. In 1999, NA ^[2] et al. proposed a method for characterizing the wrinkle degree of fabrics, which includes parameters such as wrinkle intensity, contour, power spectral density, sharpness, degree of random distribution, and overall appearance. This method provides a more comprehensive and accurate assessment of wrinkling and helps further understand the mechanisms and properties of fabric wrinkling. In 2010, HESARIAN ^[3] provided a technology to objectively evaluate the flatness grade of fabrics. By observing the side projection of the wrinkled fabric under light illumination, the degree of wrinkles and smoothness of the fabric surface can be visually evaluated. This method is relatively simple and can be used in practical applications for fabric quality control and evaluation. The algorithm flow of using three-dimensional images to detect fabric pleats is: 1), three-dimensional image acquisition; 2), three-dimensional data processing; 3), feature extraction. Three-dimensional image collection can be done through three-dimensional laser scanners, binocular stereo vision cameras, etc. ^{[4, 5]} . Three-dimensional data processing can be done through threshold segmentation, Fourier transform, wavelet transform, etc. Appearance feature extraction for 3D reconstruction includes interquartile range, mean deviation, roughness, standard deviation, distortion, kurtosis, etc. In 2006, Xu Jianming ^[6] et al. developed an objective evaluation system for fabric flatness levels based on the photometric stereovision method. They used multi-view images and photometric information to perform three-dimensional reconstruction of the fabric surface, calculated and obtained the three-dimensional depth information of the fabric surface, and combined 4 Characteristic values can provide more comprehensive and accurate fabric wrinkle assessment results. Take multiple images of the fabric under different lighting conditions, use photometric stereo vision technology to perform 3D reconstruction of the fabric surface, and apply this algorithm to the 3D reconstruction of the AATCC fabric flatness template image to obtain 3D depth information and combine 4 eigenvalues Indicates the degree of wrinkling of the fabric. In 2013, Liu Ruixin ^[7] et al. captured images under illumination in 8 different directions to obtain two-dimensional grayscale images, combined with three-dimensional reconstruction to obtain fabric surface morphology information, and characterized the features with height values and sharpness to objectively evaluate the fabric. Pleat level. In 2018, Fang Su ^[8] and others used a three-dimensional laser scanner to obtain pleat template point cloud data, extract three-dimensional features, and establish a model integrating feature indicators and pleat levels to provide a certain basis for objective evaluation of fabric pleat levels. Their application Related patent 1: CN 107945279 B ^[9] , a way to evaluate the pleat level of clothing. The assessment of the pleat level of the garment to be tested can be completed through three-dimensional scanning. However, the three-dimensional scanning test takes a long time and is computationally expensive. Large and costly.

Image processing for pleat appearance retention rating is affected by factors such as ambient lighting, fabric surface pattern and texture, and feature information extraction is greatly interfered with, ultimately affecting the accuracy of pleat rating assessment. In addition, in order to enhance the clarity of pleats, preprocessing operations such as brightness and image contrast are required. For multiple types of fabrics, parameter settings are not universal.

In addition, existing three-dimensional scanning equipment has shortcomings such as long scanning time, expensive equipment, and high algorithm cost, and is not suitable for the development of fabric pleat rating instruments with large quantities and wide areas.

Summary of the invention

The purpose of the present invention is to provide a device and method for automatically grading the appearance retention of fabric pleats, wherein a parallel grating emitted by a laser module forms a number of distorted light rays when irradiating fabric pleats; the grating image is enhanced and segmented by using image processing and analysis technology; the height information of the pleats is obtained after shape feature extraction; the information is analyzed to obtain a regression equation, which is compared and verified with a standard sample to evaluate the pleat grade of the fabric sample.

In order to solve the above technical problems, the following technical solutions are adopted:

A method for automatically grading the appearance retention of fabric pleats, characterized by comprising the following steps:

(1) Image acquisition: With the help of the image acquisition device, the camera is used to collect the image of the grating projected on the test fabric to obtain the grating image;

(2) Image preprocessing: performing image segmentation on the grating image acquired in step (1), selecting the target area, and cutting out the desired target area; performing noise reduction after the image segmentation to obtain a clear grating image;

(3) Pleat feature extraction: Extract morphological features from the segmented gratings;

(4) Grade classification and evaluation: By analyzing the grating shape and incident point angle, calculate the pleat height distribution, analyze and obtain the regression equation, and compare and verify it with the standard sample grating image to evaluate the grade value of the fabric pleat retention; Repeat the measurement three times and take the average as the final grade.

After optimization, the step (1): place the test fabric on the image acquisition device, irradiate more than 30 laser beams in parallel on the test fabric, adjust the illumination, screen brightness and camera parameters, and photograph the grating brightness of the test fabric.

After optimization, step (2): Use bilateral filtering to reduce noise, the calculation formula is:

Where: I ^B (p) is a bilateral filter, N (p) is the neighborhood of pixel p, I _p is the pixel value, W _p is the normalization factor, and the calculation formula is (2):

in: is a spatial Gaussian kernel, which depends on the positions of pixel points p and q, /> is the scale Gaussian kernel, which depends on I _p and I _q , σ _s is the spatial parameter, and σ _r is the scale parameter.

After optimization, step (3): Binarize the raster image obtained after processing in step (2), then use an edge detection operator to detect the laser curve, obtain the skeleton of the curve, and extract the pleat irradiation point area lines. Relationship to the distance from the vertex of the curve.

After optimization, step (4) grade classification: establish the relationship between fabric pleats and grades, perform statistical analysis on the characteristics of fabric pleats, and convert the extracted laser curve point and line feature parameters into intuitive pleat height information, Finally, the pleat grade is compared and evaluated; the point-line distance formula is as follows:

Among them: A _i , B _i , C _i represent the parameters of the fitted straight line equation at the i-th laser irradiation point, x _i , y _i are the coordinates of the curve vertex;

Convert the obtained distance to pleat height using the formula:

h _i =d _i ×tanα _i (4)

Among them: α _i represents the i-th laser irradiation angle.

After optimization, the angle of each irradiation point of the laser on the fabric surface will be different, and the height of the same level of pleats in different areas will also be different. The maximum height, minimum height, average height, variance, and average height of pleats are obtained through the height values of each area as follows:

in: represents the average height of pleats, n represents the number of lasers;

Where: h _σ ² represents the variance.

After optimization, step (4) grade evaluation: according to the pleat height data obtained in step (3), after converting it into actual unit distance, the data is fitted to obtain the relationship between the average height of pleats and the grade of pleats. , and calculate and obtain the regression equation. The equation form is as follows:

Among them: Grade represents the sample grade, a ₁ and c ₁ are regression parameters respectively;

The average height range of AATCC standard samples: CR-1: 0~1mm, CR-2: 1~3mm, CR-3: 3~4mm, CR-4: 4~5mm, CR-5: 7mm and above;

The relationship between the highest pleat height and pleat grade is calculated and the regression equation is obtained. The equation form is as follows:

Grade = a ₂ × h _max + c ₂ (8)

Among them: Grade represents the sample grade, a ₂ and c ₂ are regression parameters respectively, h _max represents the highest pleat height. Spend;

Finally, the measured height data of the test sample are brought into the regression equation to evaluate the pleat level of the sample.

An image acquisition device for an automatic rating method of the appearance retention level of fabric pleats comprises a closed box and a stage for placing the fabric to be tested, and is characterized in that a shooting module, a laser module and an illumination module are installed in the closed box, the laser module is used to emit a laser beam, the shooting module is used to collect a grating image of the fabric after laser irradiation, and the illumination module is used for illumination.

After optimization, the laser module emits more than 30 parallel beams, adjusts the beam distance, and evenly covers the pleats of the test fabric.

After optimization, the photographing module is connected to the computer through a network interface, and the photographing module is controlled by the computer.

Due to the adoption of the above technical solution, the following beneficial effects are achieved:

This invention uses parallel gratings emitted by a laser module to form a number of distorted light rays when irradiating fabric pleats; image processing and analysis technology is used to enhance and segment the grating image; after shape feature extraction, the height information of the pleats is obtained ; Analyze the information to obtain the regression equation, compare and verify the standard sample, and evaluate the pleat grade of the fabric sample.

The present invention reduces the time and cost of pleating grade assessment. When subjectively evaluating pleat grades, in order to ensure accuracy and credibility, specialized testers are usually required to evaluate. The subjective impact is large and the labor cost is high. The present invention realizes automatic objective rating through laser, and can switch the laser grating. The module changes the number of lasers, the method is simple, low-cost, repeatable, and the reproducibility is 100%.

The present invention is not affected by the pattern, color, surface texture and ambient light of the fabric itself, and the detection is objective and stable. The existing method collects the appearance image of the fabric, and the image is a comprehensive mapping of the color, pattern, texture and pleat morphology, which is affected by the surface properties and ambient light during the analysis process; the present invention uses laser irradiation, and the influence of ambient light is small. Even if the light is uneven, it can still effectively obtain the curve information on the pleats on the surface of the fabric and extract the pleat features, so as to perform stable detection.

The present invention calculates extracted features to achieve objective grade evaluation of fabric surface pleats. Extract two-dimensional image features through image processing, calculate and obtain the three-dimensional height information of pleats, and realize fabric pleat grade evaluation with controllable accuracy. It does not require a large amount of calculation cost, is easy to use, and reduces the complexity of pleat grade evaluation to a certain extent. Difficulty, there are certain application prospects.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention will be further described below in conjunction with the accompanying drawings:

Figure 1 is a flow chart of the automatic grade rating method of the present invention;

Figure 2 is a schematic structural diagram of the automatic grade rating device;

Figure 3 is a schematic diagram for calculating the height of the laser irradiation point;

Figure 4 shows the relationship between the average pleat height and pleat grade;

Figure 5 shows the relationship between the highest pleat height and pleat grade.

Detailed ways

The present invention aims to provide an automatic rating device and method for the appearance maintenance level of fabric pleats. The parallel grating emitted by the laser module forms a number of distorted light rays when the fabric pleats are irradiated. Image processing and analysis technology is used to evaluate the grating. The image is enhanced and segmented; after shape feature extraction, the pleat height information is obtained; the information is analyzed to obtain the regression equation, and the standard sample is compared and verified to evaluate the pleat grade of the fabric sample.

The technical solution of the present invention will be described in detail below with reference to specific examples:

An automatic rating method for fabric pleat appearance retention grade, which is characterized by including the following steps:

(1) Image acquisition: With the help of the image acquisition device, the camera is used to collect the image of the grating projected on the test fabric to obtain the grating image;

The image acquisition device includes a closed box and a stage on which the fabric to be tested is placed. The lower end of the closed box is provided with an opening, and the stage is set at the opening. The stage is drawer-type and can be pushed out or pushed in. , the stage is equipped with a loading area for placing test fabrics. The camera module, laser module and lighting module are installed in the closed box. The closed box is equipped with a bracket. The camera module, laser module and lighting module are installed through the bracket. The camera module can be industrial In a camera or other digital shooting equipment, the laser module is a laser pointer and the lighting module is a lighting lamp. There are two sets.

During the collection process, the test fabric or standard sample (CR-1~5 standard sample used in the AATCC 88C-2018t standard) is placed in the loading area of the stage at the bottom of the box, and the laser module emits more than 30 lines The parallel beam is adjusted to allow the beam spacing to evenly cover the 38cm wide fabric pleats. Due to the height difference of the pleats, the parallel beam is distorted and deformed. Use an industrial camera or other digital device to collect the distorted raster image, plug it into the computer host through the network port, then turn on the power of the instrument, open the software, select the corresponding industrial camera, and enter the shooting preview window. After entering the shooting interface, open the laser test, turn on the laser pointer, and use the laser module to irradiate multiple laser beams in parallel on the test fabric. You can observe the shooting effect through the software interface, select the screen brightness, and adjust the lighting through the lighting lamp to make it suitable. parameters, and then click Shoot to save the image.

(2) Image preprocessing: Preprocessing includes one-time fabric segmentation and noise reduction processing to obtain a clear raster image:

a. One-time segmentation: The segmentation can obtain the effective area image, speed up the image processing efficiency, select the target area through one-time segmentation selection, and intercept the required target area;

b. Noise reduction: Noise reduction eliminates the influence of noise and other factors that exist during the acquisition process. In order to maintain the edges of the extracted lines and ensure accuracy, bilateral filtering is used to reduce noise. The calculation formula is:

Among them: I ^B (p) is a bilateral filter, N (p) is the neighborhood of pixel point p, I _p is the pixel value, W _p is the normalization factor, and the calculation formula is Equation (2):

in: is a spatial Gaussian kernel, depending on the position of pixels p and q,/> is the scaled Gaussian kernel, which depends on I _p and I _q , σ _s is the spatial parameter, and σ _r is the scale parameter.

(3) Pleat feature extraction: Extract morphological features from the segmented grating; perform binarization processing on the grating image obtained after processing in step (2), and then use edge detection operators to detect the laser curve to obtain the skeleton of the curve , extract the distance relationship between the pleat illumination point area lines and the curve vertices.

(4) Grade classification and evaluation: By analyzing the grating shape and incident point angle, calculate the pleat height distribution, analyze and obtain the regression equation, and compare and verify it with the standard sample grating image to evaluate the grade value of the fabric pleat retention. specific:

a. Grade classification: establish the relationship between fabric pleats and grades, perform statistical analysis on the characteristics of fabric pleats, convert the extracted laser curve point and line feature parameters into intuitive pleat height information, and finally compare and evaluate pleat grades; The point-line distance formula is as follows:

Among them: A _i , B _i , C _i represent the parameters of the fitted straight line equation at the i-th laser irradiation point, x _i , y _i are the coordinates of the curve vertex;

Convert the obtained distance to pleat height using the formula:

h _i = d _i × tan α _i (4)

Among them: α _i represents the i-th laser irradiation angle.

The angle of each irradiation point of the laser irradiated on the surface of the fabric will be different, and the height of pleats of the same grade will also be different in different areas. Generally speaking, the lower the grade, the smaller the height, and the higher the grade, the greater the height. . In order to better quantitatively analyze the data, the maximum height, minimum height, average height, and variance are obtained through the height values of each area. The algorithm for average pleat height is as follows:

in: represents the average height of pleats, n represents the number of lasers;

Among them: h _σ ² represents the variance.

b. Grade evaluation: Based on the pleat height data of multiple areas obtained in step (3), use the ruler marked on the stage to convert it into actual unit distance, and then fit the data to obtain the average pleat height. It is related to the pleat level, and the regression equation is calculated and obtained. The equation form is as follows:

Among them: Grade represents the sample grade, a ₁ and c ₁ are regression parameters respectively;

The average height range of AATCC standard samples: CR-1: 0~1mm, CR-2: 1~3mm, CR-3: 3~4mm, CR-4: 4~5mm, CR-5: 7mm and above;

The relationship between the highest pleat height and pleat grade is calculated and the regression equation is obtained. The equation form is as follows:

Grade = a ₂ × h _max + c ₂ (8)

Among them: Grade represents the sample grade, a ₂ and c ₂ are regression parameters respectively, h _max represents the highest pleat height. Spend;

Finally, the measured height data of the test sample are brought into the regression equation to evaluate the pleat level of the sample. Repeat the measurement three times and take the average as the final grade.

The above are only specific embodiments of the present invention, but the technical features of the present invention are not limited thereto. Any simple changes, equivalent substitutions or modifications made based on the present invention to solve basically the same technical problems and achieve basically the same technical effects are all covered by the protection scope of the present invention.

Claims (10)

1. An automatic grading method for the appearance retention grade of a fabric pleat is characterized by comprising the following steps:

(1) And (3) image acquisition: the method comprises the steps that an image acquisition device is used for acquiring images of projection of a grating on a test fabric by using a camera to obtain a grating image;

(2) Image preprocessing: firstly, performing primary image segmentation on the grating image acquired in the step (1), selecting a target area, and intercepting a required target area; noise reduction is carried out after the primary image segmentation, and a clear grating image is obtained;

(3) Pleat feature extraction: extracting morphological characteristics of the segmented grating;

(4) Classification and assessment of grades: calculating the height distribution of the pleats by analyzing the grating form and the angle of incidence points, analyzing to obtain a regression equation, comparing and verifying with a standard template grating image, and evaluating the grade value of the pleat retention of the fabric; three measurements were repeated and the average was taken as the final grade.

2. An automatic fabric pleat appearance retention rating method according to claim 1, characterized in that: the step (1): and (3) placing the test fabric in an image acquisition device, radiating more than 30 lasers on the test fabric in parallel, adjusting illumination, picture brightness and camera parameters, and shooting the grating brightness of the test fabric.

3. An automatic fabric pleat appearance retention rating method according to claim 1, characterized in that: the step (2): the double-sided filtering noise reduction is adopted, and the calculation formula is as follows:

wherein: i ^B (p) is a bilateral filter, N (p) is the neighborhood of the pixel point p, I _p Is the pixel value, W _p For normalization factor, the calculation formula is formula (2):

wherein: g _σs (||p-q|) is a spatial Gaussian kernel, G depending on the location of the pixel points p and q _σr (|I _p -I _q I) is a scaled gaussian kernel, depending on I _p And I _q ，σ _s Is a space parameter, sigma _r Is a scale parameter.

4. An automatic fabric pleat appearance retention rating method according to claim 1, characterized in that: the step (3): and (3) performing binarization processing on the grating image obtained after the processing in the step (2), detecting a laser curve by using an edge detection operator to obtain a framework of the curve, and extracting the distance relation between the line of the pleat irradiation point area and the vertex of the curve.

5. An automatic fabric pleat appearance retention rating method according to claim 1, characterized in that: the step (4) is classified according to the grades: establishing a relation between the fabric pleats and the grades, carrying out statistical analysis on the characteristics of the fabric pleats, converting the extracted laser curve dotted line characteristic parameters into visual pleat height information, and finally comparing and evaluating the pleat grades; the dotted line distance formula is as follows:

wherein: a is that _i ，B _i ，C _i Represents the parameters of the fitting linear equation at the ith laser irradiation part, x _i ，y _i Is the curve vertex coordinates;

the distance obtained is converted into pleat height, the formula is:

h _i ＝d _i ×tanα _i (4)

wherein: alpha _i And expressing the irradiation angle of the ith laser.

6. An automatic fabric pleat appearance retention rating method according to claim 5, characterized in that: the angle of each irradiation point of the surface of the fabric irradiated by the laser is different, the height of the pleats of the same grade is also different in different areas, the maximum height, the minimum height, the average height, the variance and the average height of the pleats are obtained by the height values of the areas, and the algorithm is as follows:

wherein:represents the average height of the pleats, n represents the number of lasers;

wherein: h is a _σ ² Representing the variance.

7. An automatic fabric pleat appearance retention rating method according to claim 6, characterized in that: and (3) grading in the step (4): and (3) converting the data into actual unit distance according to the data of the height of the pleats obtained in the step (3), fitting the data to obtain the relation between the average height of the pleats and the pleat grade, and calculating and obtaining a regression equation, wherein the equation is as follows:

wherein: grade represents sample Grade, a ₁ ，c ₁ Regression parameters respectively;

AATCC standard average height range: CR-1: 0-1 mm, CR-2: 1-3 mm, CR-3:3 to 4mm of the diameter of the steel plate,

CR-4: 4-5 mm, CR-5:7mm and above;

the relation between the highest pleat height and the pleat grade is calculated and a regression equation is obtained, and the equation is as follows:

Grade＝a ₂ ×h _max +c ₂ (8)

wherein: grade represents sample Grade, a ₂ ，c ₂ Regression parameters, h _max Indicating that the pleats are highest. A degree;

finally, the sample pleat rating is assessed based on the measured height data of the test sample taken into a regression equation.

8. The apparatus for automatic grading of fabric pleat appearance retention level as in claim 1, comprising a closed housing and a stage for holding fabric to be tested, wherein: the novel laser device comprises a closed box body, wherein a shooting module, a laser module and an illumination module are arranged in the closed box body, the laser module is used for emitting laser beams, the shooting module is used for collecting grating images of fabrics after laser irradiation, and the illumination module is used for illumination.

9. An apparatus for an automatic fabric pleat appearance retention rating method according to claim 8, wherein: the laser module emits more than 30 parallel light beams, the distance of the light beams is adjusted, and the light beams are uniformly covered on the test fabric pleats.

10. An image acquisition device for an automatic fabric pleat appearance retention rating method according to claim 9, characterized in that: the shooting module is connected with a computer through a network interface and is controlled by the computer.