Mixed fiber web uniformity online detection method and device and storage medium

Abstract

The invention discloses a mixed fiber web uniformity online detection method and device and a storage medium. The method comprises the following steps: acquiring and preprocessing an image of a mixed fiber web; converting the preprocessed image into a color space to obtain image data of nine channels; evaluating the response degrees of different fibers in nine channels through sensitivity analysis so as to determine channel weights; combining the channel weight, utilizing a clustering algorithm to obtain the proportion of each type of fiber in the mixed fiber net, and judging the distribution uniformity of different fibers in the mixed fiber net; and extracting a pore region of the preprocessed image through a self-adaptive threshold method, and judging the porosity uniformity in the mixed fiber net through a porosity variation coefficient. According to the method, sensitivity analysis and a clustering algorithm are combined, an advanced image processing algorithm is adopted, accurate segmentation of different types of fibers and quantitative analysis of the mixing proportion are achieved, the method is suitable for a complex fiber web structure with multiple mixed fibers, and the detection efficiency and precision are greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of fiber web quality detection, and in particular to an online detection method and device for uniformity of a mixed fiber web, and a computer-readable storage medium. Background Art

[0002] The production process of nonwoven fabrics generally includes combing, laying and reinforcement. In the production and application of traditional nonwoven fabrics, they are mainly used to meet the needs of civilian or low-end industries (such as geotextiles, etc.), and the quality requirements are relatively loose. The detection method is mostly concentrated on random sampling of the final product, and the quality of the nonwoven fabric is judged by evaluating the uniformity of the gram weight. However, this kind of detection not only cannot fully reflect the overall performance of the product, but also cannot play a role in process guidance in the product production process. The adjustment of the production process is mostly completed by experience.

[0003] As nonwoven fabrics are gradually used in high-end industrial products such as aerospace and energy storage batteries, their quality requirements have increased significantly. It is necessary not only to detect defects in nonwoven fabrics (such as impurities, uneven thickness, fiber orientation, etc.), but also to control quality in real time during the production process. For example, in the production of high-performance fiber composite preforms, nonwoven fabrics can be reinforced to form felt or other forms of products, but traditional end-product weight testing alone can no longer meet the requirements of high-end products for quality uniformity and performance stability. Therefore, there is an urgent need to use advanced technology to conduct online quality testing of intermediate products in the intermediate links of nonwoven product production, such as the combing and web-forming link.

[0004] Online detection of the uniformity of fiber web distribution during the combing and web-forming process can not only predict whether the quality of subsequent products meets the standards, but also provide real-time feedback on the process adjustment of the combing machine. At present, domestic and foreign scholars and enterprises have conducted a lot of research on the technical difficulties in the field of non-woven fabrics and developed some advanced processes and detection methods. Some companies at home and abroad have realized the use of machine vision technology to detect fabric defects and quality online. For example, the density of the wool web is detected by a CCD camera, but the existing technology can only determine the grayscale distribution, and it is difficult to deeply analyze the uniformity of the fiber distribution. It is mainly suitable for a single fiber type and it is difficult to meet the needs of uniformity analysis of mixed fiber webs with complex distribution. Summary of the invention

[0005] The technical problem to be solved by the embodiments of the present application is: how to perform online detection of the uniformity of a mixed fiber web.

[0006] In order to solve the above technical problems, in a first aspect, an embodiment of the present application provides an online detection method for uniformity of a mixed fiber web, the method comprising the following steps:

[0007] Acquire an image of the hybrid fiber web;

[0008] Preprocessing the image;

[0009] The preprocessed images were converted into three color spaces: BGR, HSV, and LAB, respectively, to obtain image data of nine channels. Through sensitivity analysis, the response degree of different fibers in the mixed fiber network in the nine channels was evaluated, and then the channel weights were determined.

[0010] Combined with the channel weights, the K-mean clustering algorithm is used to obtain the proportion of each fiber in the mixed fiber web and to judge the uniformity of distribution of different fibers in the mixed fiber web;

[0011] The pore area of ​​the preprocessed image was extracted by the adaptive threshold method, the porosity of the hybrid fiber web was calculated, and the porosity uniformity in the hybrid fiber web was evaluated by the porosity variation coefficient.

[0012] Preferably, the acquiring the image of the mixed fiber web specifically comprises: acquiring the image of the mixed fiber web on a conveying device between the carding machine output system and the cross-lapping system.

[0013] Preferably, the preprocessing of the image specifically includes:

[0014] Performing brightness uniformity processing on the image;

[0015] Image enhancement via contrast-limited histogram equalization;

[0016] Convert the image to grayscale.

[0017] Preferably, the grayscale processing of the image is specifically:

[0018] I gray =0.299R+0.587G+0.114B

[0019] Among them, R, G, and B are the red, green, and blue channels of the color image respectively, and I gray is the gray value.

[0020] Preferably, the sensitivity analysis is used to evaluate the response of different fibers in the mixed fiber web in the nine channels, and then determine the channel weights, specifically:

[0021] Weaken one of the nine channels respectively to determine the influence of the channel on the clustering error;

[0022] After weakening each channel, the K-mean clustering algorithm is used to cluster the hybrid fiber network image data, and the clustering error is calculated, and the change in clustering error before and after weakening the channel Δm is recorded. i , calculate the change Δn of the channel data caused by the weakening operationi , the sensitivity of each channel is calculated by the following formula:

[0023]

[0024] Among them, S i is the sensitivity of the channel, Δm i is the clustering error change caused by channel data change, Δn i is the change in channel data;

[0025] The above steps are repeated for all sample data to obtain the average sensitivity of each channel, and a weight is assigned to each channel according to the average sensitivity.

[0026] Preferably, the channel weights are combined to obtain the proportion of each fiber in the mixed fiber web using a K-mean clustering algorithm to judge the uniformity of distribution of different fibers in the mixed fiber web, specifically:

[0027] The converted nine-channel image data is organized into a data set. Assume that there are n data points x in the data set. i =(x i1 ,x i2 ,…,x id ), i = 1, 2, ... n, where d is the dimension of the data point, and d and n are both positive integers;

[0028] Divide the data points into k categories using C j represents the jth cluster, j = 1, 2, ... k, k is a positive integer; select k initial centroids u j =(u j1 ,u j2 ,…,u jd ), as the center of k clusters;

[0029] Using the K-mean clustering algorithm, the goal is to minimize the sum of the weighted distances from all data points to the center of the class to which they belong; the objective function is:

[0030]

[0031] Among them, ||x i -u j || is the data point x i To the Class Center j The Euclidean distance, p i For data point x i The weight of the channel;

[0032] After clustering is completed, the proportion of each type of fiber in the mixed fiber network is determined according to the proportion of the number of data points in each class to the total number of data points, and then the uniformity of distribution of different fibers in the mixed fiber network is judged.

[0033] Preferably, the pore area of ​​the pre-processed image is extracted by an adaptive threshold method, the porosity of the mixed fiber web is calculated, and the porosity uniformity in the mixed fiber web is judged by the porosity variation coefficient, specifically:

[0034] The local threshold T(x,y) of the preprocessed image P(x,y) is calculated using the neighborhood average method:

[0035]

[0036] Where w is the local window, N is the number of pixels in the local window, P(i,j) is the pixel point in the local window of the preprocessed image, and C is the offset constant;

[0037] Based on the comparison between P(x,y) and T(x,y), a binary image is generated to extract the pore area;

[0038] The binary image is divided into several regions, the porosity of the mixed fiber mesh in each region is calculated, and then the porosity variation coefficient cv is calculated:

[0039]

[0040] Among them, a1 is the standard deviation of porosity in each region, and b1 is the mean porosity in each region;

[0041] The porosity uniformity in the mixed fiber web was evaluated based on the porosity variation coefficient.

[0042] In a second aspect, the present application also provides an online detection device for uniformity of a mixed fiber web, the device comprising:

[0043] An image acquisition module, used for acquiring an image of the mixed fiber web;

[0044] An image preprocessing module, used for preprocessing the image;

[0045] The space conversion and weight assignment module is used to convert the preprocessed image into three color spaces: BGR, HSV, and LAB, to obtain image data of nine channels; through sensitivity analysis, the response degree of different fibers in the mixed fiber network in the nine channels is evaluated, and then the channel weight is determined;

[0046] A module for judging the uniformity of distribution of different fibers in a mixed fiber web, which is used to combine the channel weights and use a K-mean clustering algorithm to obtain the proportion of each fiber in the mixed fiber web and judge the uniformity of distribution of different fibers in the mixed fiber web;

[0047] The porosity uniformity evaluation module in the mixed fiber network is used to extract the pore area of ​​the preprocessed image through the adaptive threshold method, calculate the porosity of the mixed fiber network, and evaluate the porosity uniformity in the mixed fiber network through the porosity variation coefficient.

[0048] In a third aspect, an embodiment of the present application further provides an online detection device for the uniformity of a mixed fiber web, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned online detection method for the uniformity of a mixed fiber web when executing the program.

[0049] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the steps of the above-mentioned online detection method for uniformity of a mixed fiber web.

[0050] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0051] 1. The method provided by the present invention, based on artificial intelligence technology, applies the weight of sensitivity analysis to the K-means clustering algorithm, realizes the precise segmentation of different types of fibers and the quantitative analysis of the mixing ratio, can deeply analyze the uniformity of the distribution of different fibers, is suitable for complex fiber network structures mixed with multiple fibers, and overcomes the limitation of the prior art to a single fiber type;

[0052] 2. The method provided by the present invention adopts advanced image processing algorithms and porosity analysis technology. It is no longer limited to the simple judgment of grayscale distribution and gets rid of the traditional method of relying on random sampling, cutting and weighing. It can realize real-time and comprehensive evaluation of the global quality of non-woven fabrics, greatly improve the detection efficiency and accuracy, meet the needs of industrial online detection, and is particularly suitable for the quality requirements in high-end application scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0054] Figure 1 This is a flow chart of the online detection method for uniformity of a mixed fiber web provided in Example 1 of the present application;

[0055] Figure 2 This is a comparison diagram of the preprocessed image and the binarized image in Example 1 of the present application; (a) preprocessed image, (b) binarized image;

[0056] Figure 3 This is a comparison chart of the pore CV values ​​and weight CV values ​​obtained for 26 samples in Example 1 of the present application;

[0057] Figure 4 This is a schematic diagram of the structure of the online detection device for uniformity of a mixed fiber web provided in Example 2 of the present application.

[0058] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0059] In order to better understand the above technical solution, exemplary embodiments will be described in detail here, and examples thereof are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are only examples of devices consistent with some aspects of the present application as detailed in the attached claims.

[0060] Embodiment 1

[0061] Figure 1 This is a flow chart of the online detection method for uniformity of a mixed fiber web provided in Example 1 of the present application. The online detection method for uniformity of a mixed fiber web comprises the following steps:

[0062] Step S1: acquiring an image of a mixed fiber web;

[0063] Specifically, an image of the mixed fiber web on a conveyor between the carding machine output system and the cross-lapping system is acquired.

[0064] In a specific implementation, 23 CCD cameras are arranged on a mixed fiber web with a width of 2.5 m. The specific layout of the cameras is as follows: the cameras are evenly distributed above the detection area, and the interval between every two cameras is 100 mm to ensure that the entire detection width is covered. The acquisition resolution of each CCD camera is 2704×2704 pixels, and continuous acquisition is achieved with the assistance of light sources. Among them, the sampling frequency is adjusted according to the running speed of the mixed fiber web, exposure time, etc.

[0065] The full inspection method was used to continuously collect images of the mixed fiber web area after removing the 10 cm boundary on both sides: the coverage area of ​​each collected image was 100 mm × 100 mm, ensuring that the image collection of the mixed fiber web was completely covered without omission. The collected image data was used for subsequent processing and analysis.

[0066] Step S2: preprocessing the image;

[0067] To ensure the accuracy of subsequent analysis, the collected original images were preprocessed.

[0068] The specific steps are as follows:

[0069] Step S21: performing brightness uniformity processing on the image;

[0070] In order to eliminate the brightness inconsistency of the image caused by uneven light source or reflection on the surface of the mixed fiber web, the following formula is used:

[0071]

[0072] Among them, I(x,y) is the original image, G(x,y) is the illumination distribution model, and I′(x,y) is the corrected image.

[0073] Step S22: performing image enhancement by contrast-limited histogram equalization;

[0074] The specific process is as follows:

[0075] (1) Image division: The image is divided into multiple non-overlapping small blocks, so that the local area of ​​the image can be processed and the contrast changes of different areas in the image can be better adapted;

[0076] (2) Histogram equalization: Perform histogram equalization on each small block. By counting the number of pixels at each gray level in the small block, the distribution probability of the gray value is calculated, and then the gray value in the small block is remapped according to the cumulative distribution function to enhance the local contrast.

[0077] (3) Cropping and redistribution: The histogram of each small block is cropped. When the number of pixels at a certain gray level exceeds the preset "limited contrast threshold", the excess pixels will be cropped and evenly distributed to other gray levels. This can limit the increase in contrast and avoid excessive amplification of noise;

[0078] (4) Interpolation merging: Use the interpolation method to merge the results of each small block to form the final enhanced image, so that the image maintains continuity and consistency as a whole.

[0079] Step S23: grayscale processing of the image;

[0080] The human eye has different sensitivities to different colors of light, with the highest sensitivity to green light, followed by red light, and the lowest sensitivity to blue light. Each pixel in a color image consists of three color channels: red (R), green (G), and blue (B). By taking the weighted average of the pixel values ​​of the three channels according to a certain weight, we can get a value that represents the brightness or grayscale value of the pixel.

[0081] According to the different perception of red, green and blue by human eyes, the image is grayed out using the following formula:

[0082] I gray =0.299R+0.587G+0.114B

[0083] Among them, R, G, and B are the red, green, and blue channels of the color image respectively, and I gray is the grayscale value. 0.299, 0.587 and 0.114 are the weights of the red, green and blue channels respectively.

[0084] The grayscale value obtained by the above processing is more consistent with the visual perception of image brightness.

[0085] Step S3: Convert the preprocessed image into three color spaces, namely BGR, HSV, and LAB, to obtain image data of nine channels; evaluate the response degree of different fibers in the mixed fiber network in the nine channels through sensitivity analysis, and then determine the channel weights.

[0086] The specific process is as follows:

[0087] First, convert the preprocessed image into three color spaces: BGR, HSV, and LAB, and obtain nine-channel image data; this can be achieved through Python or MATLAB;

[0088] Then, weaken one of the nine channels respectively to determine the influence of the channel on the clustering error;

[0089] After weakening each channel, the hybrid fiber web image data is clustered using the K-mean clustering algorithm and the clustering error is calculated. The clustering error can be determined by comparing the clustering results with known fiber categories or other reference standards. The change in clustering error before and after weakening the channel, Δm, is recorded. i ;

[0090] Calculate the change Δn of the channel data caused by the weakening operation i It can be measured by calculating the changes in statistics such as the mean and variance of the channel data.

[0091] The sensitivity of each channel is calculated by the following formula:

[0092]

[0093] Among them, S i is the sensitivity of the channel, Δm i is the clustering error change caused by channel data change, Δn i is the change in channel data.

[0094] Repeat the above steps for all sample data to obtain the average sensitivity of each channel. Then assign a weight to each channel based on the average sensitivity. The more sensitive the channel, the greater its weight, indicating that the channel is more important to the response of different fibers in the mixed fiber network.

[0095] Step S4: combining the channel weights and using the K-mean clustering algorithm to obtain the proportion of each fiber in the mixed fiber web, and to judge the uniformity of distribution of different fibers in the mixed fiber web.

[0096] The converted nine-channel image data is organized into a data set. Assume that there are n data points x in the data set. i =(x i1 ,x i2 ,…,x id ),i=1,2,…n, where d is the dimension of the data point, and d and n are both positive integers.

[0097] According to actual needs and understanding of the mixed fiber network, it is determined that the data points should be divided into k categories, using C j represents the j-th cluster, j = 1, 2, … k, and k is a positive integer.

[0098] Select k initial centroids u randomly or according to a certain strategy j =(u j1 ,u j2 ,…,u jd ), as the center of k classes, each class has a centroid;

[0099] Using the K-mean clustering algorithm, the goal is to minimize the sum of the weighted distances from all data points to the center of the cluster to which they belong. The objective function is:

[0100]

[0101] Among them, ||x i -u j || is the data point x i To the Class Center j The Euclidean distance, p i For data point x i The weight of the channel.

[0102] Specifically, the following steps are repeated until the clustering results converge or the maximum number of iterations is reached: For each data point x i , calculate its weighted Euclidean distance p to the k centroids i ||x i -u j ||, assign it to the class C to which the nearest centroid belongs j For each class C j, calculate the weighted mean of all data points in the class and update the centroid u j .

[0103] After clustering is completed, the proportion of each type of fiber in the mixed fiber web is determined according to the proportion of the number of data points in each class to the total number of data points, and then the uniformity of the distribution of different fibers in the mixed fiber web is judged. For example, if the proportion of each type of fiber in the mixed fiber web in each picture tends to be consistent, it means that the uniformity of the distribution of different fibers in the mixed fiber web is good. If the proportion of one or more fibers in the mixed fiber web in one or more pictures is significantly different from that in most other pictures, it means that the uniformity of the distribution of different fibers in the mixed fiber web in this area is poor.

[0104] In the above method, by applying the weight of sensitivity analysis to the K-means clustering algorithm, the clustering accuracy can be significantly improved.

[0105] In a specific embodiment, in a fiber web formed by mixing three types of fibers in a mixing ratio of 3:1:1, the ratio of each fiber determined by the above method is as shown in Table 1.

[0106]

[0107]

[0108] In Table 1, the floating range of the ratio is about 7%, while the floating range of the three fiber mixing ratios obtained by directly using the clustering method is about 11%. It is proved that applying the weight of the sensitivity analysis to the K-means clustering algorithm in this embodiment can significantly improve the clustering accuracy.

[0109] Step S5: extract the pore area of ​​the preprocessed image by the adaptive threshold method, calculate the porosity of the mixed fiber web, and judge the porosity uniformity in the mixed fiber web by the porosity variation coefficient (cv value).

[0110] The specific method is:

[0111] Step S51: Calculate the local threshold T(x, y) of the preprocessed image P(x, y) using the neighborhood average method:

[0112]

[0113] Where w is the local window, N is the number of pixels in the local window, P(i,j) is the pixel point in the local window of the preprocessed image, and C is the offset constant. The offset constant C is used to adjust the threshold value, which can be set according to the characteristics of the image and actual needs. If the image is bright as a whole, C can be appropriately increased; if the image is dark as a whole, C can be appropriately reduced.

[0114] Step S52: Generate a binary image based on the comparison between P(x, y) and T(x, y), thereby realizing the extraction of the pore region.

[0115] Specifically:

[0116] By calculating the average value of the gray values of all pixels within the 50*50 neighborhood of a certain pixel in the image, and using this average value as the new gray value of this pixel.

[0117] Compare the gray value P'(x, y) of each pixel in the updated image with the corresponding local threshold T(x, y). If P'(x, y) ≥ T(x, y), then set the gray value of this pixel to 255 (white, representing pores); if P'(x, y) < T(x, y), then set the gray value of this pixel to 0 (black, representing non-pores). In this way, the entire image is converted into a binary image with only black and white colors, thereby realizing the extraction of the pore region. In a specific embodiment, Figure 2 is a comparison chart of the preprocessed image and the binary image.

[0118] Step S53: Divide the binary image into several regions on average, calculate the porosity of the mixed fiber web in each region, and then calculate the coefficient of variation of porosity (cv value):

[0119]

[0120] where a1 is the standard deviation of the porosity of each region, and b1 is the average value of the porosity of each region. Evaluate the uniformity of the porosity in the mixed fiber web according to the coefficient of variation of porosity (cv value).

[0121] According to industry standards, if the cv value is less than 7%, the fiber web is considered very uniform; while in enterprise applications, if the cv value is less than 14%, it is considered applicable to production.

[0122] In a specific embodiment, each binary image is divided into 25 regions of 5*5. Figure 3 Shows the cv values of the porosity of 26 samples obtained by the above method, and their floating values are between 5% and 16%, indicating that the porosity uniformity of most samples meets the enterprise requirements.

[0123] Furthermore, compare the cv values of the porosity of 26 samples with the cv values of the weight. As Figure 3 shown, it is found that the trends between the two are the same, indicating that the method provided in this embodiment for evaluating the porosity uniformity in the mixed fiber web by the coefficient of variation can replace the traditional weighing method.

[0124] The traditional weighing method requires dividing the mixed fiber web into multiple areas and weighing them separately, which is cumbersome and requires destroying the fiber web. In comparison, the method of this embodiment is convenient and fast, with a high degree of intelligence, greatly improving the detection efficiency and accuracy, and can be detected online in real time.

[0125] Embodiment 2

[0126] Based on the same inventive concept as the online detection method for uniformity of a mixed fiber web in the aforementioned embodiment 1, this embodiment further provides an online detection device for uniformity of a mixed fiber web, such as Figure 4 As shown, the device comprises:

[0127] An image acquisition module 10, used to acquire an image of the mixed fiber web;

[0128] An image preprocessing module 20, used for preprocessing the image;

[0129] The space conversion and weight assignment module 30 is used to convert the preprocessed image into three color spaces of BGR, HSV and LAB respectively to obtain image data of nine channels; through sensitivity analysis, the response degree of different fibers in the mixed fiber network in the nine channels is evaluated, and then the channel weight is determined;

[0130] A module 40 for judging the uniformity of distribution of different fibers in a mixed fiber web, which is used to combine the channel weights and use a K-mean clustering algorithm to obtain the proportion of each fiber in the mixed fiber web and judge the uniformity of distribution of different fibers in the mixed fiber web;

[0131] The porosity uniformity evaluation module 50 in the mixed fiber web is used to extract the pore area of ​​the preprocessed image by an adaptive threshold method, calculate the porosity of the mixed fiber web, and evaluate the porosity uniformity in the mixed fiber web by the porosity variation coefficient.

[0132] The various specific processes and specific examples of the online detection method for the uniformity of a mixed fiber web in the aforementioned embodiment 1 are also applicable to the online detection device for the uniformity of a mixed fiber web in this embodiment. Through the detailed description of embodiment 1, those skilled in the art can clearly know the implementation method of the online detection device for the uniformity of a mixed fiber web in this embodiment, so for the sake of brevity of the specification, it will not be described in detail here.

[0133] Embodiment 3

[0134] Based on the same inventive concept as the online detection method for uniformity of a mixed fiber web in the aforementioned embodiment 1, this embodiment also provides an online detection device for uniformity of a mixed fiber web, on which a computer program is stored, and when the program is executed by a processor, the steps of the online detection method for uniformity of a mixed fiber web described in embodiment 1 are implemented.

[0135] Embodiment 4

[0136] Based on the same inventive concept as the online detection method for uniformity of a mixed fiber web in the aforementioned embodiment 1, this embodiment also provides a computer-readable storage medium on which a computer program is stored, which, when executed by a processor, implements the steps of the online detection method for uniformity of a mixed fiber web described in embodiment 1.

[0137] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0138] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0139] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0140] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0141] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0142] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for online detection of uniformity of a mixed fiber web, characterized in that: The method comprises the following steps: Acquire an image of the hybrid fiber web; Preprocessing the image; The preprocessed images are converted into three color spaces: BGR, HSV, and LAB, respectively, to obtain image data of nine channels; Through sensitivity analysis, the response degree of different fibers in the hybrid fiber network in nine channels is evaluated, and the channel weights are determined; Combined with the channel weights, the K-mean clustering algorithm is used to obtain the proportion of each fiber in the mixed fiber web and to judge the uniformity of distribution of different fibers in the mixed fiber web; The pore area of ​​the preprocessed image was extracted by the adaptive threshold method, the porosity of the hybrid fiber web was calculated, and the porosity uniformity in the hybrid fiber web was evaluated by the porosity variation coefficient.

2. The method for online uniformity detection of a mixed fiber web according to claim 1, characterized in that: The acquiring of the image of the mixed fiber web specifically includes acquiring the image of the mixed fiber web on a conveying device between the carding machine output system and the cross-lapping system.

3. The method for online uniformity detection of a mixed fiber web according to claim 1, characterized in that: The preprocessing of the image specifically includes: Performing brightness uniformity processing on the image; Image enhancement via contrast-limited histogram equalization; Convert the image to grayscale.

4. The method for online uniformity detection of a mixed fiber web according to claim 3, characterized in that: The grayscale processing of the image is specifically as follows: I gray =0.299R+0.587G+0.114B Among them, R, G, and B are the red, green, and blue channels of the color image respectively, and I gray is the gray value.

5. The method for online uniformity detection of a mixed fiber web according to claim 1, characterized in that: The sensitivity analysis is used to evaluate the response of different fibers in the mixed fiber network in the nine channels, and then determine the channel weights, specifically: Weaken one of the nine channels respectively to determine the influence of the channel on the clustering error; After weakening each channel, the K-mean clustering algorithm is used to cluster the hybrid fiber network image data, and the clustering error is calculated, and the change in clustering error before and after weakening the channel Δm is recorded. i , calculate the change Δn of the channel data caused by the weakening operation i , the sensitivity of each channel is calculated by the following formula: Among them, S i is the sensitivity of the channel, Δm i is the clustering error change caused by channel data change, Δn i is the change in channel data; The above steps are repeated for all sample data to obtain the average sensitivity of each channel, and a weight is assigned to each channel according to the average sensitivity.

6. The method for online uniformity detection of a mixed fiber web according to claim 1, characterized in that: The K-mean clustering algorithm is used to obtain the proportion of each fiber in the mixed fiber network in combination with the channel weights, and to judge the uniformity of distribution of different fibers in the mixed fiber network, specifically: The converted nine-channel image data is organized into a data set. Assume that there are n data points x in the data set. i =(x i1 ,x i2 ,…,x id ), i = 1, 2, ... n, where d is the dimension of the data point, and d and n are both positive integers; Divide the data points into k categories using C j represents the jth cluster, j = 1, 2, ... k, k is a positive integer; select k initial centroids u j =(u j1 ,u j2 ,…,u jd ), as the centers of k clusters; Using the K-mean clustering algorithm, the goal is to minimize the sum of the weighted distances from all data points to the center of the class to which they belong; the objective function is: Among them, ||x i -u j || is the data point x i To the Class Center j The Euclidean distance, p i For data point x i The weight of the channel; After clustering is completed, the proportion of each type of fiber in the mixed fiber network is determined according to the proportion of the number of data points in each class to the total number of data points, and then the uniformity of distribution of different fibers in the mixed fiber network is judged.

7. The method for online detection of uniformity of a mixed fiber web according to claim 1, characterized in that: The method of extracting the pore area of ​​the preprocessed image by the adaptive threshold method, calculating the porosity of the mixed fiber web, and judging the porosity uniformity in the mixed fiber web by the porosity variation coefficient is as follows: The local threshold T(x,y) of the preprocessed image P(x,y) is calculated using the neighborhood average method: Where w is the local window, N is the number of pixels in the local window, P(i,j) is the pixel point in the local window of the preprocessed image, and C is the offset constant; Based on the comparison between P(x,y) and T(x,y), a binary image is generated to extract the pore area; The binary image is divided into several regions, the porosity of the mixed fiber mesh in each region is calculated, and then the porosity variation coefficient cv is calculated: Among them, a1 is the standard deviation of porosity in each region, and b1 is the mean porosity in each region; The porosity uniformity in the mixed fiber web was evaluated based on the porosity variation coefficient.

8. An online detection device for uniformity of a mixed fiber web, characterized in that: The device comprises: An image acquisition module, used for acquiring an image of the mixed fiber web; An image preprocessing module, used for preprocessing the image; The space conversion and weight assignment module is used to convert the preprocessed image into three color spaces: BGR, HSV, and LAB, to obtain image data of nine channels; through sensitivity analysis, the response degree of different fibers in the mixed fiber network in the nine channels is evaluated, and then the channel weight is determined; A module for judging the uniformity of distribution of different fibers in a mixed fiber web, which is used to combine the channel weights and use a K-mean clustering algorithm to obtain the proportion of each fiber in the mixed fiber web and judge the uniformity of distribution of different fibers in the mixed fiber web; The porosity uniformity evaluation module in the mixed fiber network is used to extract the pore area of ​​the preprocessed image through the adaptive threshold method, calculate the porosity of the mixed fiber network, and evaluate the porosity uniformity in the mixed fiber network through the porosity variation coefficient.

9. An online detection device for uniformity of a mixed fiber web, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the online detection method for uniformity of a mixed fiber web as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the online detection method for uniformity of a mixed fiber web as described in any one of claims 1 to 7 are implemented.

Images