Evaluation method for the degree of uniform dispersion of fiber slurry based on image processing technology

Abstract

The present invention proposes an evaluation method for the dispersion uniformity of fiber slurries based on image processing technology. The fiber slurries are disintegrated using a sheet former, and the fibers are distributed on a flat cardboard and then dried in an oven to obtain a formed fiber sheet. The sheet is placed on a camera stand, and an industrial camera is used to collect the image of the sheet. Edge detection is performed based on the Scharr operator to obtain the edge information of the entire image. The image is converted into the hsv color space, the grayscale image is converted into a pure black-and-white image, and after masking, the black-and-white is transposed. The ROI region is selected, and the image is traversed cyclically with a sliding neighborhood to calculate the proportion of black pixels within the neighborhood range of each pixel. The variance of the proportion of black pixels within the neighborhood range of all pixels is calculated to characterize the dispersion uniformity of the fiber slurries. The present invention can quantitatively and quickly compare the dispersion uniformity of fibers in the slurries.

Description

Technical Field

[0001] The present invention relates to the production technology of filter paper products, and specifically to an evaluation method for the dispersion uniformity of fiber slurries based on image processing technology. Background Art

[0002] In the process of filter paper production, it is a very important technological process to defibrate and disperse fibers with different length and diameter ratios through defibrating equipment. If the glass fibers are not fully dispersed, the fibers that are not fully dispersed will form lumps on the paper surface, reducing the quality of the filter paper. The quality of the defibrating effect, that is, the dispersion degree of the fibers in the slurry, will directly affect the efficiency of the subsequent processes and the quality of the final paper sheet. For the raw materials of glass fiber filter paper, the diameter distribution is relatively wide, and different raw materials have different requirements for the defibrating process. A large number of experiments are required to evaluate the influence of defibrating process parameters on the dispersion degree of the raw materials. The traditional method for evaluating the dispersion degree of slurry fibers is time-consuming, and due to a more qualitative judgment, there will be a large error. Therefore, a method that can quickly, simply and quantitatively evaluate the dispersion degree of fiber slurries is needed. Summary of the Invention

[0003] The object of the present invention is to propose an evaluation method for the dispersion uniformity of fiber slurries based on image processing technology.

[0004] The technical solution to achieve the object of the present invention is: an evaluation method for the dispersion uniformity of fiber slurries based on image processing technology, including the following steps:

[0005] Step 1: Conduct a defibrating experiment to obtain a sheet: Use a sheet former to defibrate the fiber slurry, and through vacuum suction, make the fibers distribute on a flat cardboard, and then put it into an oven to dry to obtain a formed fiber sheet;

[0006] Step 2: Image acquisition of the sheet: Place the sheet on a camera stand, and use an industrial camera to acquire the image of the sheet;

[0007] Step 3: Edge feature extraction of the image: Based on the Scharr operator, perform edge detection to obtain the edge feature information of the entire picture;

[0008] Step 4: Binary processing: Convert the image into the hsv color space, define the black threshold range, convert the grayscale image into a pure black and white image, and transpose the black and white colors after masking;

[0009] Step 5: Image sliding neighborhood operation: Select the ROI area, use the sliding neighborhood to loop through the image, and calculate the proportion of black pixels in the neighborhood range of each pixel;

[0010] Step 6: Calculate the variance: Calculate the variance of the proportion of black pixels in the neighborhood range of all pixels, which characterizes the dispersion uniformity of the fiber slurry.

[0011] Further, in step 1, the fibers are distributed on a flat cardboard and placed in an oven at 150 °C for drying for 1 hour and then taken out to obtain a formed fiber sheet.

[0012] Further, in step 2, the sheets under multiple working conditions need to be subjected to centralized image acquisition to avoid the influence of environmental factors such as light sources, and it is ensured that the focal length and exposure settings of the industrial camera remain unchanged, and the placement position of each sheet is unchanged.

[0013] Further, in step 3, the Scharr operator kernel is:

[0014]

[0015] The edges in the vertical direction are obtained by convolving the image with the G x operator, and the edges in the horizontal direction are obtained by convolving the image with the G y operator;

[0016] Further, in step 3, the steps of edge detection are as follows:

[0017] First, perform Gaussian blur for smoothing and noise reduction; further, convert the image into a grayscale image; then, based on the Scharr operator kernel, calculate the gradients in the X and Y directions; finally, add the absolute values of the gradients in the X and Y directions to obtain the amplitude image of the comprehensive gradient.

[0018] Further, in step 5, an operation is performed using a 2×3 neighborhood.

[0019] Further, in step 6, calculate the variance of the proportion of black pixels within the neighborhood range of all pixels. The formula is:

[0020]

[0021] where x i represents the ratio of the number of black pixels within the neighborhood range of the i-th pixel to the total number of pixels in the neighborhood, x represents the mean value of the proportion of black pixels within the neighborhood range of all pixels, and s 2 represents the variance of the proportion of black pixels within the neighborhood range of all pixels.

[0022] An evaluation system for the dispersion uniformity of fiber slurries based on image processing technology, based on the evaluation method for the dispersion uniformity of fiber slurries, realizes the evaluation of the dispersion uniformity of fiber slurries.

[0023] A computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, based on the evaluation method for the dispersion uniformity of fiber slurries, the evaluation of the dispersion uniformity of fiber slurries is realized.

[0024] A computer-readable storage medium has a computer program stored thereon. When the computer program is executed by a processor, an evaluation of the degree of uniform dispersion of fiber slurry is achieved based on the evaluation method of the degree of uniform dispersion of the fiber slurry.

[0025] Compared with the prior art, the significant advantages of the present invention are as follows: 1) low detection cost and convenient use; 2) by performing image processing on the sheet made from the pulping experiment, it is more stable compared to directly evaluating the fiber dispersion in the slurry; 3) during the research process of pulping process parameters, feedback on the degree of uniform fiber dispersion can be quickly obtained and quantitatively compared, which can provide a basis for judging and optimizing the research of pulping process parameters. Description of the Drawings

[0026] Figure 1 It is a flowchart of the method of the present invention.

[0027] Figure 2 It is a schematic diagram of a visual recognition system.

[0028] Figure 3 It is the original image of the sheet obtained by an industrial camera.

[0029] Figure 4 It is the comprehensive gradient image after edge detection.

[0030] Figure 5 It is the image after further binaryzation processing.

[0031] Figure 6 It is the original images of sheets with three different raw material concentrations.

[0032] Figure 7 It is the images of sheets with three different raw material concentrations after edge detection.

[0033] Figure 8 It is the variance curve graph of the sheet images under three different raw material concentrations. Detailed Embodiments

[0034] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0035] The present invention proposes an evaluation method for the degree of uniform dispersion of fiber slurry based on image processing technology. The specific process is as Figure 1 shown, and the steps are as follows:

[0036] Step 1: Conduct defibration experiments to obtain sheets: Conduct defibration experiments, defibrate the fiber slurry using a small sheet former, and through vacuum suction, make the fibers distribute on a flat cardboard, then place it in an oven at 150 °C for drying. Take it out after 1 hour to obtain a formed fiber sheet.

[0037] Step 2: Collect sheet images: As Figure 2 shown, use a visual recognition system to collect sheet images. Place the sheet on a camera stand and use a professional industrial camera to collect sheet images. Multiple sheets under different working conditions need to be centrally image-collected to avoid the influence of environmental factors such as light sources, and at the same time ensure that the placement position of each sheet remains fixed, and ensure that the settings such as the focal length and exposure of the industrial camera are appropriate and unchanged.

[0038] Step 3: Extract image edge features: Import relevant modules in the Python integrated development environment and open the sheet image. Since the pixel values of the edge part are significantly different from the adjacent pixels, taking the local extreme values of the picture can obtain the edge information of the entire picture. Import the CV2 module and perform gradient calculation by convolving the edge detection operator with the image. Through comparison of the algorithm and the effect after sheet processing, it is obtained that the Scharr operator is more accurate than the Sobel operator, Prewitt operator, and Roberts operator, is not afraid of interference, and has a faster operation speed. The Scharr operator kernel is:

[0039]

[0040] Use the G x operator to convolve with the original image to detect the vertical edge, and use the G y operator to convolve with the original image to detect the horizontal edge. The steps for extracting the edge are as follows:

[0041] 1. First, perform Gaussian blur for smoothing and noise reduction:

[0042] GaussianBlur(src,dst,Size(3,3),0,0,BORDER_DEFAULT);

[0043] 2. Further convert the image to a grayscale image:

[0044] cvtColor(src,gray,COLOR_RGB2GRAY);

[0045] 3. Calculate the gradients in the X and Y directions (derivation):

[0046] Scharr(gray_src,xgrad,CV_16S,1,0)、

[0047] Scharr(gray_src, ygrad, CV_16S, 0, 1);

[0048] 4. Pixel absolute value: Calculate the absolute value of the pixels of image A and output it to image B:

[0049] convertScaleAbs(A, B);

[0050] 5. Add X and Y to image B to obtain the amplitude image of the comprehensive gradient: Use a loop to obtain pixels, and directly add the gradients in the X and Y directions at each point. By comparison, the present invention has a better effect than adding with a mixed weight, and the result is as Figure 4 shown.

[0051] Step 4, Binarization processing: Convert the image to the hsv color space, define the black range, and use the cv2.inRange() function and mask similar to threshold processing to convert the grayscale image into a pure black-and-white image. After passing through the mask, transpose the black and white, and the result is as Figure 5 shown.

[0052] lower_black = np.array([0, 0, 0]);

[0053] upper_black = np.array([180, 255, 46]);

[0054] mask = cv2.inRange(HSV, lower_red, upper_red)cv2.imshow('mask', mask).

[0055] Step 5, Define the ROI (region of interest) area: Use the circle function to specify the center coordinates and radius of the circle to define a circular area as the ROI area.

[0056] circle(mask, circleCenter, radius, Scalar(255), -1);

[0057] Step 6, Image sliding neighborhood operation: There are two types of image block operations: non-overlapping block operation and sliding neighborhood operation. In the present invention, variance is used to evaluate the sheet forming uniformity. Compared with the non-overlapping block operation, more sample data can be obtained by adopting the sliding neighborhood operation, and the variance value can more reflect the non-uniformity of the feature distribution in the image. It is also found through the comparison of sheet forming images that the variance change range of the sliding neighborhood operation is larger than that of the non-overlapping block operation, so it can be compared more accurately. Define a sliding neighborhood of m×n, and the central pixel position is: floor([(m + 1) / 2, (n + 1) / 2]). By comparing sliding neighborhoods of various sizes in the present invention, including neighborhoods of 2×3, 3×5, 5×5, and 5×10, it is found that the smaller the neighborhood, the larger the variance difference between two different sheet formings, that is, the more obvious the difference in uniformity, and at the same time, the relatively longer the calculation time. After comprehensive comparison, the present invention adopts a 2×3 neighborhood for operation. The specific steps for the sliding neighborhood operation are as follows:

[0058] 1. Select the pixels in the ROI area;

[0059] 2. Define a sliding pixel neighborhood of 2×3;

[0060] 3. Perform a cyclic traversal of the pixel neighborhood throughout the image;

[0061] 4. Nest a traversal loop of the pixel points within the neighborhood in the neighborhood traversal loop within the image, and call a function to calculate the ratio of the number of black pixels in each neighborhood to the total number of pixels in the neighborhood;

[0062] Function definition: Nested row and column traversal loops within the neighborhood. When img[i, j] == 0, the number of black pixels is incremented by 1, otherwise the number of white pixels is incremented by 1, and calculate the ratio of the number of black pixels to the total number of pixels.

[0063]

[0064] 5. Add the black pixel ratios of each sliding neighborhood to the specified list in sequence.

[0065] Step 7, Calculate the variance of the list: According to the variance calculation formula: Solve for the average value x of all elements in the list, count the number of list elements, and then calculate the variance. Use the variance value of the sheet forming image to characterize the dispersion uniformity of the fiber slurry. The smaller the variance, the better the dispersion uniformity.

[0066] The present invention is implemented in the above manner by taking fiber slurries of different concentrations as examples, Figure 6 For the sheet forming images corresponding to three concentrations (2.93‰, 3.52‰, 4.40‰) of the slurry, it can be qualitatively seen from the images that as the fiber concentration increases, fiber flocculation increases, the dispersion situation deteriorates, and the uniformity decreases. After edge detection of the sheet forming images of the three concentrations, asFigure 7 As shown, by comparing with the sheet-forming images, it can be seen that the detection results can better reflect the image features. The variance curves of sheet-forming at three concentrations are distributed as Figure 8 shown. The results of this embodiment are consistent with the qualitative understanding and can quantitatively compare the different dispersion uniformity degrees of fibers, which is helpful for the experimental study on optimizing the technological parameters of the filter paper disintegration process.

[0067] The present invention also provides an evaluation system for the dispersion uniformity degree of fiber slurries based on image processing technology, which realizes the evaluation of the dispersion uniformity degree of fiber slurries based on the above-mentioned evaluation method for the dispersion uniformity degree of fiber slurries.

[0068] A computer device includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, it realizes the evaluation of the dispersion uniformity degree of fiber slurries based on the above-mentioned evaluation method for the dispersion uniformity degree of fiber slurries.

[0069] A computer-readable storage medium stores a computer program thereon. When the computer program is executed by a processor, it realizes the evaluation of the dispersion uniformity degree of fiber slurries based on the above-mentioned evaluation method for the dispersion uniformity degree of fiber slurries.

[0070] The technical features of the above embodiments can be combined arbitrarily. For the sake of brief description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0071] The above-described embodiments only represent several implementation manners of the present application. Their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

Claims

1. An evaluation method for the degree of uniformity of fiber slurry dispersion based on image processing technology, characterized in that, It includes the following steps: Step 1, conduct a beating experiment to obtain a sheet: Use a sheet former to beat the fiber slurry. Through vacuum suction, the fibers are distributed on a flat cardboard and then placed in an oven to dry to obtain a formed fiber sheet; Step 2, image acquisition of the sheet: Place the sheet on a camera stand and use an industrial camera to acquire the image of the sheet; Step 3, edge feature extraction of the image: Based on the Scharr operator, perform edge detection to obtain the edge information of the entire image; Step 4, binary processing: Convert the image to the hsv color space, define the black range, convert the grayscale image to a pure black-and-white image, and perform black-and-white transposition after masking; Step 5, image sliding neighborhood operation: Select the ROI area, use a sliding neighborhood to traverse the image cyclically, and calculate the proportion of black pixels within the neighborhood range of each pixel; Step 6, calculate the variance: Calculate the variance of the proportion of black pixels within the neighborhood range of all pixels, which characterizes the degree of uniform dispersion of the fiber slurry; In Step 1, the fibers are distributed on a flat cardboard and placed in an oven at 150 °C for 1 hour and then taken out to obtain a formed fiber sheet; In Step 2, the sheets under multiple working conditions need to be centrally imaged, and it is ensured that the focal length and exposure settings of the industrial camera remain unchanged, and the placement position of each sheet remains unchanged; In Step 3, the Scharr operator kernel is: ； The vertical edges are obtained by convolving the image with the G x operator, and the horizontal edges are obtained by convolving the image with the G y operator; The steps of edge detection are as follows: First, perform Gaussian blur for smoothing and noise reduction; further convert the image to a grayscale image; then calculate the gradients in the X and Y directions based on the Scharr operator kernel; finally, add the absolute values of the gradients in the X and Y directions to obtain the amplitude image of the comprehensive gradient; In Step 5, a 2×3 neighborhood is used for the operation; In Step 6, calculate the variance of the proportion of black pixels within the neighborhood range of all pixels, and the formula is: ； Among them, x i represents the ratio of the number of black pixels within the i -th pixel neighborhood to the total number of pixels in the neighborhood, x represents the mean value of the proportion of black pixels within all pixel neighborhoods, s 2 represents the variance of the proportion of black pixels within all pixel neighborhoods.

2. A computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, based on the evaluation method for the degree of uniform dispersion of the fiber slurry described in claim 1, the evaluation of the degree of uniform dispersion of the fiber slurry is realized.

3. A computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, based on the evaluation method for the degree of uniform dispersion of the fiber slurry described in claim 1, the evaluation of the degree of uniform dispersion of the fiber slurry is realized.

Images