Method and system for detecting paper pulp distribution uniformity in toilet paper production

Abstract

The invention provides a paper pulp distribution uniformity detection method and system in toilet paper production, and relates to the technical field of papermaking processes and detections.The method comprises the steps that according to a filtered image, an edge distribution diagram of paper pulp texture is extracted to obtain a paper pulp texture edge distribution diagram; according to the paper pulp texture edge distribution diagram, calculating a width index of a paper pulp surface macrostructure so as to obtain a paper pulp surface preliminary evaluation result of toilet paper pulp distribution uniformity; and according to the paper pulp surface preliminary evaluation result, subdividing an image region and calculating a local index to obtain the local index. The method for detecting the width index of the macroscopic structure of the surface of the paper pulp and the non-uniform coefficient of the internal structure is combined, more comprehensive quality evaluation indexes are provided for toilet paper production, and the product quality is controlled more accurately.

Description

Technical Field

[0001] The invention relates to the field of papermaking technology and detection technology, in particular to a method and system for detecting the uniformity of pulp distribution in toilet paper production. Background Art

[0002] In the production process of toilet paper, the uniform distribution of pulp is a key factor in ensuring product quality. The uniformity of pulp distribution not only affects the core properties of toilet paper such as thickness, strength and water absorption, but is also directly related to the product's appearance and user experience. If the pulp is unevenly distributed, it may cause the toilet paper to have problems such as uneven thickness, easy tearing or poor water absorption, thereby reducing the market competitiveness of the product.

[0003] Traditional pulp distribution uniformity detection methods mainly rely on manual observation and simple physical tests. Manual observation methods usually rely on the operator's experience and subjective judgment, which is not only inefficient, but also severely limited in accuracy and reliability. At the same time, although simple physical tests can provide certain quantitative data, they often cannot fully reflect the true situation of pulp distribution, especially the distribution differences at the microscopic level. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a method and system for detecting the uniformity of pulp distribution in toilet paper production, which combines the width index of the macroscopic structure of the pulp surface and the detection method of the unevenness coefficient of the internal structure to provide more comprehensive quality evaluation indicators for toilet paper production and more accurately control product quality.

[0005] In order to solve the above technical problems, the technical solution of the present invention is as follows:

[0006] In a first aspect, a method for detecting pulp distribution uniformity in toilet paper production comprises:

[0007] Capturing texture images of pulp surfaces during tissue production;

[0008] Separating the background and the pulp texture of the pulp image to obtain a separated pulp texture image;

[0009] performing filtering processing on the separated pulp texture image to obtain a filtered image;

[0010] Extracting the edge distribution map of the pulp texture according to the filtered image to obtain the edge distribution map of the pulp texture;

[0011] According to the pulp texture edge distribution map, the width index of the pulp surface macro structure is calculated to obtain the preliminary evaluation result of the pulp surface for the uniformity of the pulp distribution of toilet paper;

[0012] According to the preliminary evaluation results of the pulp surface, the image area is subdivided and the local index is calculated to obtain the local index;

[0013] According to the subdivided image area and local indexes, the uniform complexity of the pulp surface image is obtained;

[0014] Select pulp samples from the production line and analyze the tomographic images of the samples to obtain the internal structure of the pulp, including the distribution of fibers and fillers;

[0015] Calculate the non-uniformity coefficient of the internal structure of pulp, including horizontal non-uniformity coefficient and vertical non-uniformity coefficient;

[0016] The distribution uniformity of toilet paper pulp is comprehensively evaluated based on the non-uniformity coefficient and uniformity complexity, so as to realize the detection of pulp distribution uniformity in toilet paper production.

[0017] Further, the background and the pulp texture of the pulp image are separated to obtain a separated pulp texture image, including:

[0018] Get each pixel of the image data and set the window size according to the image;

[0019] In each window, calculate the local mean and local standard deviation of each pixel;

[0020] According to the local mean and local standard deviation, Get the adaptive threshold of each pixel, where W is the window radius, x is the horizontal position of the pixel in the image, y is the vertical position of the pixel in the image, i is the horizontal offset in the window, j is the vertical offset in the window, I(x+i,y+j) is the pixel value at position (x+i,y+j) in the image, k is an empirical constant with a value between -0.2 and -0.5; T(x,y) is the adaptive threshold of each pixel;

[0021] According to the adaptive threshold, each pixel is subjected to threshold processing to separate the background of the pulp image from the pulp texture, thereby obtaining a separated pulp texture image.

[0022] Further, according to the filtered image, an edge distribution map of the pulp texture is extracted to obtain an edge distribution map of the pulp texture, including:

[0023] Convolve the filtered image with the horizontal and vertical Sobel matrices to obtain the horizontal and vertical gradients of each pixel;

[0024] The gradient amplitude and direction of each pixel are calculated based on the horizontal gradient and the vertical gradient to obtain the gradient amplitude map of the image;

[0025] According to the gradient amplitude map of the image, a threshold is set for edge judgment to extract the edge distribution map of the pulp texture.

[0026] Furthermore, according to the pulp texture edge distribution map, the width index of the pulp surface macro structure is calculated to obtain the preliminary evaluation results of the pulp surface for the uniformity of the toilet paper pulp distribution, including:

[0027] On the pulp texture edge distribution map, all edges are identified and marked using image processing technology to obtain accurate edge position and shape information;

[0028] For each marked edge, measure its width, average all measured edge widths to get the average edge width;

[0029] By calculating the standard deviation or coefficient of variation of the edge width, the degree of width dispersion can be obtained;

[0030] Determine the width index based on the average edge width and the dispersion of the width;

[0031] Based on the value of the width index, the preliminary evaluation results of the pulp surface for the uniformity of the pulp distribution of toilet paper are obtained.

[0032] Furthermore, according to the preliminary evaluation results of the pulp surface, the image area is subdivided and the local indicators are calculated to obtain the local indicators, including:

[0033] According to the preliminary evaluation results of the pulp surface, the sub-graph division method is determined;

[0034] According to the sub-image division method, the global pulp texture edge distribution map is divided into several sub-images, wherein each sub-image should contain a part of the pulp texture;

[0035] Calculate the width index of the edge distribution of the sub-image texture. The uniformity index includes the identification edge, the measured width, the average width and the discrete degree of the width;

[0036] The width index is normalized to obtain the local index.

[0037] Furthermore, pulp samples from the production line were selected and the tomographic images of the samples were analyzed to obtain the internal structure of the pulp, including the distribution of fibers and fillers, including:

[0038] Randomly select pulp samples from the production line;

[0039] Using a high-resolution tomography device, the processed pulp sample is scanned to obtain tomographic image data;

[0040] According to the tomographic image data, the pixels in the image are divided into two or more categories by iteratively calculating the threshold value to obtain the internal structure of the pulp, including the distribution of fibers and fillers.

[0041] Furthermore, the non-uniformity coefficient of the internal structure of the pulp is calculated, including the horizontal non-uniformity coefficient and the vertical non-uniformity coefficient, including:

[0042] According to the tomographic image, the analysis area is determined and divided into regular grids;

[0043] For each mesh, the fibers and fillers in the mesh are quantified to obtain quantitative data of each mesh;

[0044] According to the quantitative data of each grid, for all grids on the same horizontal layer, The fiber non-uniformity coefficient of the horizontal layer is obtained by The packing non-uniformity coefficient of the horizontal layer is obtained, where and are the inhomogeneities of the fiber and filler at the mth level, F m，n is the fiber quantification data of the nth grid in the mth horizontal layer, P m,n is the packing quantization data of the nth grid in the mth horizontal layer, N is the total number of grids in the mth horizontal layer, and m and n are indexes;

[0045] According to the quantitative data of each grid, for grids on different horizontal layers, The fiber vertical non-uniformity coefficient is obtained by The vertical non-uniformity coefficient of the filler is obtained, where and are the heterogeneity of fibers and fillers in multiple horizontal layers, F m，n is the fiber quantification data of the nth grid in the mth horizontal layer, P m，n is the packing quantization data of the nth grid in the mth horizontal layer, N is the total number of grids in the mth horizontal layer, m and n are indices, M is the total number of horizontal layers, and w m is the weight of each horizontal layer.

[0046] In a second aspect, a system for detecting uniformity of pulp distribution in toilet paper production comprises:

[0047] An acquisition module for capturing texture images of the pulp surface during the toilet paper production process;

[0048] A surface processing module is used to separate the background and pulp texture of the pulp image to obtain a separated pulp texture image; filter the separated pulp texture image to obtain a filtered image; extract the edge distribution map of the pulp texture based on the filtered image to obtain a pulp texture edge distribution map; calculate the width index of the macro structure of the pulp surface based on the pulp texture edge distribution map to obtain a preliminary evaluation result of the pulp surface for the uniformity of the pulp distribution of toilet paper; subdivide the image area and calculate the local index based on the preliminary evaluation result of the pulp surface to obtain the local index; obtain the uniform complexity of the pulp surface image based on the subdivided image area and the local index;

[0049] The internal processing module is used to select pulp samples from the production line and analyze the tomographic images of the samples to obtain the internal structure of the pulp, including the distribution of fibers and fillers; calculate the non-uniformity coefficient of the internal structure of the pulp, including the horizontal non-uniformity coefficient and the vertical non-uniformity coefficient;

[0050] The comprehensive processing module comprehensively evaluates the distribution uniformity of toilet paper pulp according to the non-uniformity coefficient and uniformity complexity, so as to realize the detection of pulp distribution uniformity in toilet paper production.

[0051] According to a third aspect, a computing device includes:

[0052] one or more processors;

[0053] The storage device is used to store one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the method described.

[0054] In a fourth aspect, a computer-readable storage medium stores a program, and when the program is executed by a processor, the method described is implemented.

[0055] The above solution of the present invention includes at least the following beneficial effects:

[0056] By capturing and analyzing the texture image of the pulp surface, the uniformity of pulp distribution can be accurately evaluated. By using the background and texture separation technology of pulp images, combined with filtering and edge extraction, the flow and distribution of pulp in the toilet paper production process can be deeply understood. By real-time monitoring of the uniformity of pulp distribution, problems in the production process can be discovered and adjusted in time to avoid unnecessary downtime and material waste. By analyzing the tomographic images of pulp samples from the production line, the distribution of fibers and fillers inside the pulp can be deeply understood. Combined with the width index of the macroscopic structure of the pulp surface and the non-uniformity coefficient of the internal structure, this method provides a comprehensive quality evaluation index for toilet paper production, which helps to control product quality more accurately and meet the diverse needs of the market and consumers. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a schematic flow chart of a method for detecting pulp distribution uniformity in toilet paper production provided by an embodiment of the present invention.

[0058] Figure 2 It is a schematic diagram of a detection system for pulp distribution uniformity in toilet paper production provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0059] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in a form and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0060] like Figure 1 As shown, an embodiment of the present invention provides a method for detecting pulp distribution uniformity in toilet paper production, the method comprising the following steps:

[0061] Step 11, capturing a texture image of the pulp surface during the toilet paper production process;

[0062] Step 12, separating the background and the pulp texture of the pulp image to obtain a separated pulp texture image;

[0063] Step 13, filtering the separated pulp texture image to obtain a filtered image;

[0064] Step 14, extracting the edge distribution map of the pulp texture according to the filtered image to obtain the edge distribution map of the pulp texture;

[0065] Step 15, calculating the width index of the macroscopic structure of the pulp surface according to the pulp texture edge distribution map, so as to obtain a preliminary evaluation result of the pulp surface for the uniformity of the pulp distribution of the toilet paper;

[0066] Step 16, based on the preliminary evaluation result of the pulp surface, subdividing the image area and calculating the local index to obtain the local index;

[0067] Step 17, obtaining uniform complexity of the pulp surface image according to the subdivided image regions and local indicators;

[0068] Step 18, selecting a pulp sample from the production line and analyzing the tomographic image of the sample to obtain the internal structure of the pulp, including the distribution of fibers and fillers;

[0069] Step 19, calculating the non-uniformity coefficient of the internal structure of the pulp, including the horizontal non-uniformity coefficient and the vertical non-uniformity coefficient;

[0070] Step 20, comprehensively evaluating the distribution uniformity of toilet paper pulp based on the non-uniformity coefficient and uniformity complexity, so as to realize the detection of pulp distribution uniformity in toilet paper production.

[0071] In the embodiment of the present invention, by capturing and analyzing the texture image of the pulp surface, the surface quality of toilet paper can be monitored in real time, and the pulp texture and background can be automatically separated by image processing technology, thereby improving the detection efficiency and accuracy and reducing the error of manual detection. Filtering can remove noise and interference in the image, making the subsequent edge detection and width index calculation more accurate. By extracting the edge distribution map of the pulp texture, the distribution of pulp on the toilet paper surface can be accurately understood, providing data support for uniformity evaluation. The width index of the macroscopic structure of the pulp surface provides a fast and practical preliminary evaluation result of the pulp surface for quality control in the production process. By analyzing the tomographic image of the pulp sample, the distribution of fibers and fillers inside the pulp can be deeply understood, thereby optimizing the production formula and process. The calculation of the non-uniformity coefficient, including horizontal and vertical, provides a quantitative index for evaluating the uniformity of the internal structure of the pulp. Combining the surface uniformity complexity and the non-uniformity coefficient of the internal structure, the distribution uniformity of toilet paper pulp can be more comprehensively evaluated to ensure product quality. According to the comprehensive evaluation results, the production process parameters can be adjusted in time to improve the uniformity of pulp distribution, thereby improving product quality.

[0072] In a preferred embodiment of the present invention, the above step 11 includes:

[0073] By aiming the camera lens at the pulp surface, during the toilet paper production process, the camera is triggered to shoot in real time or at a scheduled time to capture the texture image of the pulp surface.

[0074] In the embodiment of the present invention, by real-time or timed shooting, the texture image of the pulp surface can be instantly obtained, so as to monitor the production process of toilet paper in real time, which helps to find problems in production in time, such as uneven pulp distribution, fiber aggregation, etc., so as to take corrective measures quickly. The texture image of the pulp surface is an important embodiment of the quality of toilet paper. By shooting and analyzing these images, the surface quality of toilet paper can be accurately controlled to ensure that the product meets the quality standards and reduce the defective rate.

[0075] The specific implementation steps of the present invention include: setting the parameters of the camera and setting the trigger mode of the camera to perform real-time triggering. In the real-time trigger mode, a fixed time interval is set, the camera will continuously photograph the pulp surface, preview the image taken by the camera on the software interface, and then adjust the camera position and parameters according to actual conditions; during the toilet paper production process, start the image capture function, the camera will start to capture the texture image of the pulp surface, and the captured image data will be transmitted to the computer or image processing system in real time and stored.

[0076] In a preferred embodiment of the present invention, the above step 12 includes:

[0077] Step 121, obtaining each pixel of the image data and setting the window size according to the image;

[0078] Step 122, in each window, calculating the local mean and local standard deviation of each pixel;

[0079] Step 123, based on the local mean and local standard deviation, Get the adaptive threshold of each pixel, where W is the window radius, x is the horizontal position of the pixel in the image, y is the vertical position of the pixel in the image, i is the horizontal offset in the window, j is the vertical offset in the window, I(x+i, y+j) is the pixel value at position (x+i, y+j) in the image, k is an empirical constant with a value between -0.2 and -0.5; T(x, y) is the adaptive threshold of each pixel;

[0080] Step 124 , performing threshold processing on each pixel point according to the adaptive threshold value, so as to separate the background and the pulp texture of the pulp image, and obtain a separated pulp texture image.

[0081] In an embodiment of the present invention, by calculating the local mean and standard deviation of each pixel point and setting an adaptive threshold accordingly, the pulp texture and background can be segmented more accurately. This method is more flexible and accurate than using a global threshold, and is particularly suitable for situations where the contrast between the background and the texture varies greatly. The adaptive thresholding method can adjust the threshold according to the local characteristics of different regions, thereby better retaining the subtle features of the pulp texture. By accurately separating the background and texture, the interference of background noise on texture analysis can be eliminated. This method can reduce dependence on manual intervention and achieve a higher level of automated processing. Accurate separation of background and texture is the basis for subsequent complex image processing tasks such as pulp uniformity assessment and quality inspection. The use of the adaptive thresholding method can ensure the consistency of image segmentation under different lighting conditions and pulp color changes.

[0082] The specific implementation steps of the present invention include:

[0083] Step 121, read the captured pulp surface texture image into the computer memory through the image processing library, and determine the window radius W according to the characteristics of the pulp texture and the image resolution to ensure that sufficient local information can be captured for calculating the local statistical information of each pixel point to ensure that sufficient local information can be captured.

[0084] Step 122, use a nested loop structure to traverse each pixel in the image. For each pixel, create a window of size (2W+1)×(2W+1) with it as the center. In the current window, accumulate the grayscale values ​​of all pixels and divide it by the total number of pixels in the window (i.e. (2W+1) 2 ), obtain the local mean of the current pixel, where the local mean reflects the average brightness level of the area around the pixel; calculate the local standard deviation of the current pixel based on the local mean of the pixel. The local standard deviation reflects the degree of fluctuation of the brightness of the area around the pixel.

[0085] Steps 123-124, for each pixel, use the formula to calculate its adaptive threshold, traverse each pixel in the image again, and compare the grayscale value of each pixel with the corresponding adaptive threshold. If the grayscale value is greater than or equal to the threshold, the pixel is set to the foreground color (usually white, indicating pulp texture); otherwise, it is set to the background color (usually black, indicating background). After the above processing, the pulp texture and background in the original pulp image will be effectively separated.

[0086] In a preferred embodiment of the present invention, the above step 13 includes:

[0087] According to the selected Sobel filter and parameters, a filter kernel is created, and a convolution operation is performed on the kernel and each local area of ​​the separated pulp texture image to obtain a new pixel value to obtain a filtered image.

[0088] In the embodiment of the present invention, the Sobel filter is a commonly used edge detection operator, which can effectively detect the edge information in the image. In the pulp texture image, the edge of the pulp fiber can be accurately identified by Sobel filtering. The filtering process can effectively suppress the noise in the image, especially the high-frequency noise. The Sobel filter can smooth the image while enhancing the edge information, and reduce the interference of noise on the subsequent analysis. Through the filtering process, the edge features in the pulp texture image are enhanced, which is helpful for the subsequent feature extraction and analysis. The clear edge information can make the distribution, direction and density of the pulp fiber more obvious. Compared with other complex image processing algorithms, the Sobel filter has the characteristics of simple calculation and high efficiency. Therefore, the use of the Sobel filter for filtering can improve the calculation efficiency of image processing and meet the requirements of real-time or large-scale data processing. The edge information of the image after Sobel filtering is clearer. In the process of toilet paper production, the texture and background of the pulp may change due to factors such as raw materials and processes. The use of the Sobel filter for filtering can improve the overall robustness and stability.

[0089] In a preferred embodiment of the present invention, the above step 14 includes:

[0090] Step 141, convolving the filtered image with horizontal and vertical Sobel matrices to obtain the horizontal gradient and vertical gradient of each pixel;

[0091] Step 142, calculating the gradient magnitude and direction of each pixel point according to the horizontal gradient and the vertical gradient to obtain a gradient magnitude map of the image;

[0092] Step 143, according to the gradient amplitude map of the image, a threshold is set to perform edge judgment to extract the edge distribution map of the pulp texture.

[0093] In the embodiment of the present invention, the horizontal gradient and vertical gradient of each pixel can be accurately calculated by convolving the image with the horizontal and vertical Sobel matrices; the gradient amplitude and direction of each pixel can be further calculated by combining the horizontal gradient and the vertical gradient to form a gradient amplitude map of the image, which is helpful to describe the structure and characteristics of the pulp texture in more detail. By setting the threshold for edge judgment, the sensitivity of edge extraction can be flexibly adjusted according to actual needs, and the edge distribution map of the extracted pulp texture can be customized according to different product standards or production environments. The Sobel operator has a certain inhibitory effect on noise. Through gradient calculation and threshold judgment, the interference of non-edge areas can be further reduced, the real pulp texture edge can be highlighted, and the signal-to-noise ratio of the analysis can be improved. The accurately extracted pulp texture edge distribution map is an important basis for subsequent complex image processing tasks such as pulp uniformity evaluation, quality inspection, and fault identification. Through the accurate extraction and analysis of the pulp texture edge, problems in the production process, such as uneven fiber distribution and impurity mixing, can be discovered in time, so as to adjust the production parameters in time, optimize the production process, and ensure product quality.

[0094] The specific implementation steps of the present invention include:

[0095] Step 141, first, use standard horizontal and vertical Sobel matrices. The horizontal Sobel matrix is ​​used to detect horizontal edges in the image, while the vertical Sobel matrix is ​​used to detect vertical edges. These two matrices are convolved with the filtered image respectively. During the convolution process, each element of the Sobel matrix is ​​multiplied by the pixel value at the corresponding position in the image, and all the products are added to obtain the horizontal gradient and vertical gradient of each pixel.

[0096] Step 142, using the horizontal gradient and vertical gradient obtained in step 141, calculate the gradient amplitude and direction of each pixel, wherein the gradient amplitude is obtained by calculating the square root of the sum of the squares of the horizontal gradient and the vertical gradient, and the gradient amplitude reflects the strength of the edge. At the same time, the gradient direction can be obtained by calculating the inverse tangent function of the ratio of the vertical gradient to the horizontal gradient, and the gradient direction indicates the direction of the edge. The gradient amplitude value of each pixel is used as the grayscale value or color value of the pixel in the gradient amplitude map. In this way, the gradient amplitude map intuitively shows the intensity of changes in various regions in the image, and the edge region usually has a higher gradient amplitude value. In the gradient amplitude map, the brighter area indicates a larger gradient amplitude, that is, the edge or the area with obvious changes; while the darker area indicates a smaller gradient amplitude, that is, the area with gentle or uniform changes in the image.

[0097] Step 143, determine the edge based on the gradient amplitude map of the image. First, set a suitable threshold. If the threshold is set too high, it may cause the loss of edge information, while if it is set too low, it may introduce too much noise. By comparing the gradient amplitude of each pixel with this threshold, it can be determined which pixels belong to the edge. Specifically, if the gradient amplitude of a pixel is greater than or equal to the threshold, it is marked as an edge pixel; otherwise, it is regarded as a non-edge pixel. Finally, all the pixels marked as edges together constitute the edge distribution map of the pulp texture.

[0098] In a preferred embodiment of the present invention, the above step 15 includes:

[0099] Step 151, using image processing technology to identify and mark all edges on the pulp texture edge distribution map to obtain accurate position and shape information of the edges;

[0100] Step 152, for each marked edge, measure its width, and average all measured edge widths to obtain an average edge width;

[0101] Step 153, calculating the standard deviation or coefficient of variation of the edge width to obtain the degree of width dispersion;

[0102] Step 154, determining a width index based on the average edge width and the degree of width dispersion;

[0103] Step 155, obtaining a preliminary evaluation result of the pulp surface of the toilet paper pulp distribution uniformity according to the value of the width index.

[0104] In the embodiment of the present invention, by identifying and marking all edges of the pulp texture through image processing technology, the accurate position and shape information of the edge can be obtained, which is helpful for accurately analyzing the uniformity of pulp distribution and avoiding the subjectivity and error of manual identification. By measuring and averaging the width of all marked edges, a specific average edge width value is obtained. This quantitative index provides an objective basis for evaluating the uniformity of pulp distribution, so that the product quality can be understood more accurately. By calculating the standard deviation or coefficient of variation of the edge width, the discrete degree of the width can be understood, which is helpful for identifying the uneven areas in the pulp distribution and providing guidance for improving the production process. Determining the width index based on the average edge width and the discrete degree of the width provides a comprehensive evaluation standard for the uniformity of toilet paper pulp distribution. According to the value of the width index, the preliminary evaluation result of the pulp surface of the uniformity of toilet paper pulp distribution can be obtained, which allows the products that do not meet the uniformity requirements to be quickly screened out during the production process and adjusted in time, thereby improving product quality and qualified rate. Through the accurate evaluation of the uniformity of pulp distribution, the production process can be optimized more effectively, the waste of raw materials can be reduced, and the production cost can be reduced. At the same time, improving product quality also helps to enhance brand image and market competitiveness.

[0105] The specific implementation steps of the present invention include:

[0106] Step 151, based on the edge distribution map obtained by the Sobel operator, refinement is performed, and some morphological operations (such as dilation, erosion, opening and closing operations) are used to optimize the continuity and accuracy of the edge, and image processing technology is used to identify and mark all edges on the pulp texture edge distribution map. The purpose of identification and marking is to obtain the accurate position and shape information of the edge.

[0107] Step 152, after the edge is accurately marked, scan the image row by row or column by column, record the width value of each edge, determine the starting and ending positions of the edge, and calculate the width. After completing the width measurement of all edges, these width values ​​are averaged to obtain the average edge width, where the average value represents the average size of the pulp texture edge and is one of the important parameters for evaluating the uniformity of pulp distribution.

[0108] Step 153, in order to more comprehensively evaluate the uniformity of pulp distribution, the degree of dispersion of edge width is calculated by calculating the standard deviation or coefficient of variation of edge width. Standard deviation is a statistic that measures the degree of dispersion of data distribution and reflects the fluctuation of data points relative to the average value. The coefficient of variation is the ratio of standard deviation to average value, which eliminates the influence of average value on the degree of dispersion measurement and enables comparison between data sets with different average values. The standard deviation or coefficient of variation is selected to quantify the range of variation of edge width, thereby evaluating the uniformity of pulp distribution.

[0109] Step 154, based on the average edge width and the degree of width dispersion (standard deviation or coefficient of variation) of step 153, determine a comprehensive width index that can comprehensively reflect the average size and size consistency of the pulp texture edge. For example, the average edge width and the standard deviation (or coefficient of variation) can be weighted combined to obtain a single width index value. The larger the value, the worse the uniformity of the pulp distribution; conversely, the better the uniformity.

[0110] Step 155, based on the calculated width index value, a preliminary evaluation result of the pulp surface of the toilet paper pulp distribution uniformity is obtained. The preliminary evaluation result is a series of thresholds, which are used to map the width index value to a specific uniformity level. For example, several thresholds can be set to divide the width index value into several levels such as "excellent", "good", "average" and "poor". When the width index value is lower than a certain threshold, the pulp distribution uniformity is considered to be excellent; as the index value increases, the uniformity level gradually decreases.

[0111] In a preferred embodiment of the present invention, the above step 16 includes:

[0112] Step 161, determining a sub-graph division method according to the preliminary evaluation result of the pulp surface;

[0113] Step 162, dividing the global pulp texture edge distribution map into a plurality of sub-maps according to the sub-map division method, wherein each sub-map should contain a portion of the pulp texture;

[0114] Step 163, calculating the width index of the sub-image texture edge distribution, the uniformity index includes the identification edge, the measurement width, the average width and the discrete degree of the width;

[0115] Step 164 , normalizing the width index to obtain a local index.

[0116] In an embodiment of the present invention, by sub-graph division, the global pulp texture edge distribution map can be subdivided into smaller areas, so that each area can be evaluated more finely, which is helpful to find local differences and unevenness in pulp distribution. When determining the sub-graph division method, the size and number of sub-graphs can be flexibly adjusted according to the preliminary evaluation results of the pulp surface, so that the evaluation process can adapt to pulp texture images of different sizes and resolutions, thereby improving the versatility and practicality of the evaluation method. By calculating the width index of the sub-graph texture edge distribution, the characteristics of each local area can be highlighted. By comparing the index values ​​of different sub-graphs, it is easier to identify abnormal areas or potential problem points in the pulp distribution, providing strong support for quality control in the production process. Normalizing the width index can eliminate the dimensional differences caused by factors such as size and resolution between different sub-graphs, making the local index comparable, and helping to make a unified quantitative evaluation of the uniformity of pulp distribution on a global scale. Through a detailed analysis of the local index, the uniformity of toilet paper pulp distribution can be more accurately evaluated. The sub-graph division and local index calculation process can be automated through programming to reduce the influence of manual intervention and subjective judgment. This helps to improve evaluation efficiency, reduce evaluation costs, and lay the foundation for the intelligent management of toilet paper production.

[0117] The specific implementation steps of the present invention include:

[0118] Step 161, analyze the preliminary evaluation results of the pulp surface, determine which areas may show high non-uniformity based on the indicators in the preliminary evaluation results of the pulp surface (such as the average edge width and the degree of width dispersion), and determine the appropriate sub-image size considering the characteristics of the pulp texture and the image size. The sub-image should be large enough to contain sufficient texture information, and small enough to facilitate the analysis of local features. Simple grid division can be used to evenly divide the global image into several sub-images. If the preliminary evaluation results of the pulp surface indicate that some specific areas need more detailed analysis, denser division can be used in these areas, while sparser division can be used in other areas.

[0119] Step 162, according to the sub-image division method determined in step 161, use the image processing software or the function in the programming library to actually segment the global pulp texture edge distribution map, cut the global image into several sub-images according to the division method, each sub-image should contain a part of the pulp texture, and the size should meet the determination in step 161, ensure that the boundaries between the sub-images are clear, and save the generated sub-images in an appropriate format for subsequent analysis and processing.

[0120] Step 163, for each sub-image, repeat the process of step 151 to step 154, that is, identify the edge, measure the width, calculate the average width and the discrete degree of the width. The calculation results (average edge width, standard deviation or coefficient of variation, etc.) of each sub-image are recorded to form the width index data at the sub-image level. By comparing the width indexes of different sub-images, the distribution uniformity of pulp texture in different local areas can be analyzed.

[0121] Step 164, in order to facilitate the comparison between different sub-images, it is necessary to normalize the width index, convert the width index of each sub-image into a uniform value range (such as between 0 and 1), and the normalized width index is the local index, which reflects the relative uniformity of the pulp texture distribution in each sub-image. By comparing the local indexes of different sub-images, it is easier to identify areas with more uniform or less uniform pulp distribution.

[0122] In a preferred embodiment of the present invention, the above step 17 includes:

[0123] The calculated local indicators (i.e., normalized width indicators) of each sub-image are collected, the indicator data are cleaned and formatted, and the processed local indicator data are statistically analyzed to obtain the distribution in different sub-images. According to the distribution of local indicators, those areas that are obviously deviated from the normal range are identified. These areas correspond to the parts where the pulp distribution is extremely uneven or defective. The uniform complexity is calculated based on the variation range and distribution of local indicators. The uniform complexity is a comprehensive indicator used to quantify the uniformity of texture distribution in the pulp surface image. The variation range of local indicators in different sub-images is analyzed by calculating statistics such as the difference between the maximum and minimum values, the standard deviation, or the interquartile range. The uniform complexity is calculated using weighted average based on the average value, variation range, and distribution of local indicators.

[0124] The specific implementation steps of the present invention include:

[0125] By collecting local indicators of each sub-graph and performing data cleaning and formatting, the accuracy and consistency of the data can be ensured. Statistical analysis of the processed local indicator data can accurately identify areas that are significantly deviated from the normal range, which helps to quickly locate parts where the pulp distribution is extremely uneven or defective, so that timely adjustments and improvements can be made. By calculating the comprehensive indicator of uniform complexity, the uniformity of texture distribution in the pulp surface image can be quantified, providing an objective and comparable metric that helps to evaluate product quality and consistency. Analyzing the distribution of local indicators in different sub-graphs and gaining a comprehensive understanding of the overall status of pulp distribution can help to discover potential systemic problems or trends. Through the above analysis, production parameters and processes can be adjusted more accurately to improve production efficiency and product quality.

[0126] In a preferred embodiment of the present invention, the above step 18 includes:

[0127] Step 181, randomly selecting pulp samples from the production line;

[0128] Step 182, using a high-resolution tomography device to scan the processed pulp sample to obtain tomographic image data;

[0129] Step 183, based on the tomographic image data, pixels in the image are divided into two or more categories by iteratively calculating the threshold value, so as to obtain the internal structure of the pulp, including the distribution of fibers and fillers.

[0130] In the embodiment of the present invention, pulp samples are randomly selected from the production line to ensure that the selected samples are representative and can reflect the actual situation in the production process, which is helpful to obtain accurate and reliable data; by scanning the pulp samples with a high-resolution tomography device, high-precision tomographic image data can be obtained, which can show the microstructure inside the pulp in detail, including the distribution of fibers and fillers, and provide intuitive and comprehensive visual information for in-depth understanding of the properties of pulp. By iteratively calculating the threshold, the pixels in the image are divided into two or more categories, which can achieve quantitative analysis and classification of the internal structure of pulp. This classification capability helps to accurately identify different types of fibers and fillers and quantify their proportions and distribution states in pulp. Understand the detailed information of the internal structure of pulp and better control product quality. By monitoring the distribution of fibers and fillers, problems in the production process, such as uneven fiber distribution and filler aggregation, can be discovered in time, and the production process parameters can be adjusted accordingly to optimize the production process. In-depth analysis of the internal structure of pulp can provide strong support for the research and development of new products. By understanding the influence of different combinations of fibers and fillers on pulp properties, new products with specific properties can be developed to meet market demand. Optimizing the production process and formulation of pulp can not only improve product quality, but also help achieve energy conservation, emission reduction and efficient use of resources. By reducing waste and unreasonable use in the production process, production costs can be reduced while reducing the impact on the environment. With a deep understanding of the internal structure of pulp and the ability to optimize it, higher quality and more innovative products can be provided, thereby enhancing market competitiveness. This helps to expand market share and improve customer satisfaction and loyalty.

[0131] The specific implementation steps of the present invention include:

[0132] Step 181, according to the characteristics of the production line and the uniformity of the pulp flow, set sampling points at different stages of the production line (such as mixing, stirring, conveying, etc.) to ensure that the sampling points can represent the pulp quality of the entire production line. At each sampling point, use appropriate tools (such as sampling spoons, sampling tubes, etc.) to randomly grab a certain amount of pulp as a sample, and properly process the sampled pulp.

[0133] Step 182, select a high-resolution tomography device, such as an X-ray tomography scanner or a nuclear magnetic resonance imager; set appropriate scanning parameters, such as scanning speed, layer thickness, resolution, etc., according to the characteristics of the pulp sample and the analysis requirements; place the processed pulp sample at an appropriate position of the scanning device, start the scanning device, and perform a high-resolution tomography scan on the pulp sample. During the scanning process, the device will penetrate the sample layer by layer and record the image data of each layer.

[0134] Step 183, pre-processing the scanned tomographic image data, using the Otsu algorithm to determine the optimal segmentation threshold of different tissues (such as fibers and fillers) in the image, and finding the threshold that can maximize the inter-class variance through multiple iterative calculations. Based on the calculated threshold, using the segmentation tool in the image processing software, the pixels in the image are divided into two or more categories, corresponding to different components in the pulp (such as fibers, fillers, etc.), and the classified images are visualized to more intuitively observe the internal structure of the pulp. The classification results are statistically analyzed, such as calculating the volume fraction and distribution of each component, so as to quantitatively evaluate the uniformity and quality of the pulp.

[0135] In a preferred embodiment of the present invention, the above step 19 includes:

[0136] Step 191, determining an analysis region according to the tomographic image, and dividing the analysis region into regular grids;

[0137] Step 192, for each grid, quantify the fibers and fillers in the grid to obtain quantified data of each grid;

[0138] Step 193, based on the quantized data of each grid, for all grids on the same horizontal layer, The fiber non-uniformity coefficient of the horizontal layer is obtained by The packing non-uniformity coefficient of the horizontal layer is obtained, where and are the inhomogeneities of the fiber and filler at the mth level, F m，n is the fiber quantification data of the nth grid in the mth horizontal layer, P m，n is the packing quantization data of the nth grid in the mth horizontal layer, N is the total number of grids in the mth horizontal layer, and m and n are indexes;

[0139] Step 194, based on the quantized data of each grid, for grids on different horizontal layers, The fiber vertical non-uniformity coefficient is obtained by The vertical non-uniformity coefficient of the filler is obtained, where and are the heterogeneity of fibers and fillers in multiple horizontal layers, F m，n is the fiber quantification data of the nth grid in the mth horizontal layer, P m，n is the packing quantization data of the nth grid in the mth horizontal layer, N is the total number of grids in the mth horizontal layer, m and n are indices, M is the total number of horizontal layers, and w m is the weight of each horizontal layer.

[0140] In an embodiment of the present invention, by dividing the analysis area into regular grids, a more refined analysis of the internal structure of the pulp can be performed, which helps to accurately capture the distribution of fibers and fillers in different positions; by quantifying the fibers and fillers in each grid, comprehensive quantitative data can be obtained, which can objectively and accurately reflect the content and distribution of each component inside the pulp, avoid errors caused by subjective judgment, and improve the accuracy and reliability of the analysis. By calculating the fiber and filler unevenness coefficient of the horizontal layer, the distribution uniformity of the fibers and fillers on the same horizontal layer can be evaluated, which helps to identify the distribution differences and problem areas in the horizontal direction, and provide a strong basis for optimizing the pulp formula and production process. By calculating the vertical unevenness coefficient of the fibers and fillers, the distribution uniformity of the fibers and fillers between different horizontal layers can be evaluated, which helps to understand the structural changes of the pulp in the thickness direction, discover potential inter-layer differences, and provide guidance for improving the overall performance of the product. Combining the horizontal and vertical unevenness evaluation results, a multi-level comprehensive evaluation of the distribution of fibers and fillers inside the pulp can be conducted, which can more comprehensively reveal the characteristics and existing problems of the pulp structure; based on the quantitative analysis results, the production process can be optimized in a targeted manner, and the ratio and distribution of fibers and fillers can be adjusted to improve the uniformity and quality stability of the product. At the same time, this also helps to achieve quality control and cost optimization in the production process. By improving the uniformity of the distribution of fibers and fillers inside the pulp, the overall performance and user experience of the product can be improved.

[0141] The specific implementation steps of the present invention include:

[0142] Step 191, according to the characteristics of the tomographic image, select a representative area as the analysis object, this area can reflect the overall situation of the internal structure of the pulp, divide the selected analysis area into a series of regular grids, the shape of the grid can be square, rectangular; determine the size of the grid to ensure that each grid contains enough pixels for subsequent quantitative analysis.

[0143] Step 192, select a quantification method, including pixel counting, gray value statistics, etc. For fibers, quantification can be performed based on their shape, size, gray value and other characteristics; for fillers, quantification can be performed based on their gray value, distribution density and other characteristics; the quantification data of the fibers and fillers in each grid are recorded to form a data set.

[0144] Step 193, extract the fiber and filler quantitative data of all grids on the same horizontal layer from the quantitative data set. For the fiber, use the provided formula to calculate the fiber non-uniformity coefficient of the horizontal layer, and the coefficient reflects the non-uniformity of the fiber distribution on the same horizontal layer. For the filler, also use the provided formula to calculate the filler non-uniformity coefficient of the horizontal layer, and the coefficient reflects the non-uniformity of the filler distribution on the same horizontal layer.

[0145] Step 194, extract the quantitative data of fibers and fillers of the grids on different horizontal layers from the quantitative data set, assign a weight to each horizontal layer, the weight can be determined according to the importance, thickness or other relevant factors of the horizontal layer, and use the provided formula to combine the weight of each horizontal layer to calculate the vertical non-uniformity coefficient of the fiber, which reflects the degree of non-uniform distribution of the fiber between different horizontal layers. Similarly, calculate the vertical non-uniformity coefficient of the filler.

[0146] In a preferred embodiment of the present invention, the above step 20 includes:

[0147] In toilet paper production, uniform complexity, as a surface evaluation index, reflects the smoothness of the toilet paper surface and the distribution consistency of fibers and fillers; while the non-uniformity coefficient, as an internal structure evaluation index, reveals the uniformity of the dispersion of fibers and fillers inside the pulp. Combining these two indicators, the distribution uniformity of toilet paper pulp can be comprehensively evaluated. The following is the specific implementation process:

[0148] According to the calculation uniformity complexity index, the index can quantify the uniformity of the toilet paper surface. Based on the quantitative data, the horizontal layer non-uniformity coefficient and the vertical non-uniformity coefficient are calculated respectively to evaluate the uniformity of fiber and filler distribution inside the pulp.

[0149] The collected uniformity complexity and non-uniformity coefficient data are standardized to eliminate the dimensional differences between different indicators. A comprehensive evaluation index is constructed by weighted average based on the uniformity complexity and non-uniformity coefficient. When constructing the comprehensive evaluation index, appropriate weights are assigned to the uniformity complexity and non-uniformity coefficient based on historical data to reflect their importance in evaluating the uniformity of pulp distribution. Based on historical data, thresholds and evaluation criteria are set for the comprehensive evaluation index. These thresholds and criteria will be used to judge the degree of uniformity of toilet paper pulp distribution. For example, evaluation grades such as "excellent", "good", "average" and "poor" can be set.

[0150] The calculated comprehensive evaluation index is compared with the set evaluation standard to determine the evaluation grade of toilet paper pulp distribution uniformity to obtain the uniformity of pulp distribution.

[0151] In an embodiment of the present invention, by simultaneously considering the surface uniformity and internal structural uniformity of toilet paper, a more comprehensive and accurate evaluation of product quality can be obtained. Both uniform complexity and non-uniformity coefficient are quantitative evaluation indicators that can provide objective and accurate data support, eliminate the uncertainty caused by subjective evaluation, and make the evaluation results more comparable and credible. By monitoring the changes in uniform complexity and non-uniformity coefficient, problems in the production process, such as uneven fiber distribution and rough surface, can be discovered in a timely manner, which helps to quickly take corresponding corrective measures to avoid the expansion of problems and affect product quality. Based on the feedback of comprehensive evaluation indicators, the production process can be optimized in a targeted manner, the raw material ratio can be adjusted, the process parameters can be improved, etc., which helps to improve production efficiency while reducing production costs and scrap rates. By continuously improving the uniformity of toilet paper pulp distribution, higher quality products can be provided to consumers, which will help to enhance the brand image, enhance the market competitiveness of products, and thus expand market share. The comprehensive evaluation indicators and their thresholds and evaluation criteria provide a basis for continuous improvement.

[0152] like Figure 2 As shown, an embodiment of the present invention further provides a detection system 20 for pulp distribution uniformity in toilet paper production, comprising:

[0153] An acquisition module 21 is used to capture a texture image of a pulp surface during a toilet paper production process;

[0154] The surface processing module 22 is used to separate the background and pulp texture of the pulp image to obtain a separated pulp texture image; filter the separated pulp texture image to obtain a filtered image; extract the edge distribution map of the pulp texture based on the filtered image to obtain a pulp texture edge distribution map; calculate the width index of the macro structure of the pulp surface based on the pulp texture edge distribution map to obtain a preliminary evaluation result of the pulp surface for the uniformity of the pulp distribution of toilet paper; subdivide the image area and calculate the local index based on the preliminary evaluation result of the pulp surface to obtain the local index; obtain the uniform complexity of the pulp surface image based on the subdivided image area and the local index;

[0155] The internal processing module 23 is used to select pulp samples from the production line and analyze the tomographic images of the samples to obtain the internal structure of the pulp, including the distribution of fibers and fillers; calculate the non-uniformity coefficient of the internal structure of the pulp, including the horizontal non-uniformity coefficient and the vertical non-uniformity coefficient;

[0156] The comprehensive processing module 24 comprehensively evaluates the distribution uniformity of the toilet paper pulp according to the non-uniformity coefficient and the uniformity complexity, so as to realize the detection of the pulp distribution uniformity in the toilet paper production.

[0157] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for detecting pulp distribution uniformity in toilet paper production, characterized in that: The method comprises: Capturing texture images of pulp surfaces during tissue production; Separating the background and the pulp texture of the pulp image to obtain a separated pulp texture image; performing filtering processing on the separated pulp texture image to obtain a filtered image; Extracting the edge distribution map of the pulp texture according to the filtered image to obtain the edge distribution map of the pulp texture; According to the pulp texture edge distribution map, the width index of the pulp surface macro structure is calculated to obtain the preliminary evaluation result of the pulp surface for the uniformity of the pulp distribution of toilet paper; According to the preliminary evaluation results of the pulp surface, the image area is subdivided and the local index is calculated to obtain the local index; According to the subdivided image area and local indexes, the uniform complexity of the pulp surface image is obtained; Select pulp samples from the production line and analyze the tomographic images of the samples to obtain the internal structure of the pulp, including the distribution of fibers and fillers; Calculate the non-uniformity coefficient of the internal structure of pulp, including horizontal non-uniformity coefficient and vertical non-uniformity coefficient; The distribution uniformity of toilet paper pulp is comprehensively evaluated based on the non-uniformity coefficient and uniformity complexity, so as to realize the detection of pulp distribution uniformity in toilet paper production.

2. The method for detecting pulp distribution uniformity in toilet paper production according to claim 1, characterized in that: The background and the pulp texture of the pulp image are separated to obtain a separated pulp texture image, including: Get each pixel of the image data and set the window size according to the image; In each window, calculate the local mean and local standard deviation of each pixel; Calculate the adaptive threshold of each pixel based on the local mean and local standard deviation; According to the adaptive threshold, each pixel is subjected to threshold processing to separate the background of the pulp image from the pulp texture, thereby obtaining a separated pulp texture image.

3. The method for detecting pulp distribution uniformity in toilet paper production according to claim 2, characterized in that: According to the filtered image, an edge distribution map of the pulp texture is extracted to obtain an edge distribution map of the pulp texture, including: Convolve the filtered image with the horizontal and vertical Sobel matrices to obtain the horizontal and vertical gradients of each pixel; The gradient amplitude and direction of each pixel are calculated based on the horizontal gradient and the vertical gradient to obtain the gradient amplitude map of the image; According to the gradient amplitude map of the image, a threshold is set for edge judgment to extract the edge distribution map of the pulp texture.

4. The method for detecting pulp distribution uniformity in toilet paper production according to claim 3, characterized in that: According to the pulp texture edge distribution map, the width index of the pulp surface macro structure is calculated to obtain the preliminary evaluation results of the pulp surface for the uniformity of the pulp distribution of toilet paper, including: On the pulp texture edge distribution map, all edges are identified and marked using image processing technology to obtain accurate edge position and shape information; For each marked edge, measure its width, average all measured edge widths to get the average edge width; By calculating the standard deviation or coefficient of variation of the edge width, the degree of width dispersion can be obtained; Determine the width index based on the average edge width and the dispersion of the width; Based on the value of the width index, the preliminary evaluation results of the pulp surface for the uniformity of the pulp distribution of toilet paper are obtained.

5. The method for detecting pulp distribution uniformity in toilet paper production according to claim 4, characterized in that: According to the preliminary evaluation results of the pulp surface, the image area is subdivided and the local indicators are calculated to obtain the local indicators, including: According to the preliminary evaluation results of the pulp surface, the sub-graph division method is determined; According to the sub-image division method, the global pulp texture edge distribution map is divided into several sub-images, wherein each sub-image should contain a part of the pulp texture; Calculate the width index of the edge distribution of the sub-image texture. The uniformity index includes the identification edge, the measured width, the average width and the discrete degree of the width; The width index is normalized to obtain the local index.

6. The method for detecting pulp distribution uniformity in toilet paper production according to claim 5, characterized in that: Select pulp samples from the production line and analyze the tomographic images of the samples to obtain the internal structure of the pulp, including the distribution of fibers and fillers, including: Randomly select pulp samples from the production line; Using a high-resolution tomography device, the processed pulp sample is scanned to obtain tomographic image data; According to the tomographic image data, the pixels in the image are divided into two or more categories by iteratively calculating the threshold value to obtain the internal structure of the pulp, including the distribution of fibers and fillers.

7. The method for detecting pulp distribution uniformity in toilet paper production according to claim 6, characterized in that: Calculate the non-uniformity coefficient of the internal structure of the pulp, including the horizontal non-uniformity coefficient and the vertical non-uniformity coefficient, including: According to the tomographic image, the analysis area is determined and divided into regular grids; For each mesh, the fibers and fillers in the mesh are quantified to obtain quantitative data of each mesh; According to the quantitative data of each grid, for all grids on the same horizontal layer, the fiber non-uniformity coefficient of the horizontal layer and the filler non-uniformity coefficient of the horizontal layer are calculated; According to the quantitative data of each grid, the vertical non-uniformity coefficient of fiber and the vertical non-uniformity coefficient of filler are calculated for grids on different horizontal layers.

8. A system for detecting the uniformity of pulp distribution in toilet paper production, characterized in that: The system implements the method according to any one of claims 1 to 7, including: An acquisition module for capturing texture images of the pulp surface during the toilet paper production process; A surface processing module is used to separate the background and pulp texture of the pulp image to obtain a separated pulp texture image; filter the separated pulp texture image to obtain a filtered image; extract the edge distribution map of the pulp texture based on the filtered image to obtain a pulp texture edge distribution map; calculate the width index of the macro structure of the pulp surface based on the pulp texture edge distribution map to obtain a preliminary evaluation result of the pulp surface for the uniformity of the pulp distribution of toilet paper; subdivide the image area and calculate the local index based on the preliminary evaluation result of the pulp surface to obtain the local index; obtain the uniform complexity of the pulp surface image based on the subdivided image area and the local index; The internal processing module is used to select pulp samples from the production line and analyze the tomographic images of the samples to obtain the internal structure of the pulp, including the distribution of fibers and fillers; calculate the non-uniformity coefficient of the internal structure of the pulp, including the horizontal non-uniformity coefficient and the vertical non-uniformity coefficient; The comprehensive processing module comprehensively evaluates the distribution uniformity of toilet paper pulp according to the non-uniformity coefficient and uniformity complexity, so as to realize the detection of pulp distribution uniformity in toilet paper production.

9. A computing device, characterized in that include: one or more processors; A storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the method as claimed in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program, which, when executed by a processor, implements the method according to any one of claims 1 to 7.

Images