Filter element quality monitoring method based on image processing
By calculating the pixel point characteristic value and histogram of the filter element non-woven fabric, and combining with the segmentation threshold interval, the problem of low accuracy in the quality monitoring of the filter element in the prior art is solved, and higher monitoring accuracy and Flaw detection accuracy.
Patent Information
- Application Number
- CN202411533459.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-31
- Publication Date
- 2024-11-29
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing filter filter element quality monitoring method based on image processing is large in the setting parameter value deviation, resulting in low accuracy of filter element quality monitoring and inability to effectively detect defects in non-woven fabrics.
By calculating the characteristic value of each pixel point in the grayscale image of the filter filter element non-woven fabric, building a sliding window to obtain the grayscale run matrix, calculating the characteristic value of the pixel point, and determining whether the pixel point is a defective point based on the characteristic value. Combining histogram and segmentation threshold intervals improve monitoring accuracy.
It improves the accuracy of filter element quality monitoring, can detect defects in non-woven fabrics more accurately, and enhances the monitoring ability of filter element quality.
Smart Images

Figure CN119048509A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the field of filter element quality monitoring, and in particular to a filter element quality monitoring method based on image processing. Background Art
[0002] Filter elements are usually made of non-woven fabrics that are processed and folded. During the production process, some parts of the non-woven fabrics will be dirty and uneven in thickness. After the non-woven fabrics are folded, these defective parts are not easy to detect, which will make the quality of the filter element poor, and further lead to poor filtering effect.
[0003] A Chinese patent document with announcement number CN114841982B discloses a method for monitoring the quality of air-conditioning filters based on image processing, the method comprising: obtaining a plurality of continuous frames of grayscale images of the surface of activated carbon non-woven fabrics used for processing air-conditioning filters; obtaining the fiber uniformity corresponding to each grayscale image of the surface of activated carbon non-woven fabrics according to the grayscale histogram corresponding to each grayscale image of the surface of activated carbon non-woven fabrics; screening each grayscale image of the surface of activated carbon non-woven fabrics according to the fiber uniformity to obtain each target non-woven fabric surface image corresponding to the processing of air-conditioning filters; obtaining a Gaussian mixture model corresponding to each target non-woven fabric surface image according to the grayscale histogram corresponding to each target non-woven fabric surface image; and obtaining each abnormal pixel point corresponding to each target non-woven fabric surface image according to the Gaussian mixture model.
[0004] When obtaining abnormal pixel points on the surface of non-woven fabrics through the Gaussian mixture model, the results are greatly affected by the specific values of the set parameters. When the deviation of the set parameter values is large, the error of the obtained abnormal pixel points is large, which leads to low accuracy of filter element quality monitoring. Summary of the invention
[0005] In order to solve the problem that the accuracy of filter element quality monitoring is low when the set parameter value deviation is large, the present invention provides a filter element quality monitoring method based on image processing.
[0006] The present invention provides a filter element quality monitoring method based on image processing, which adopts the following technical solution: A filter element quality monitoring method based on image processing comprises the following steps: Obtain a grayscale image of the filter element non-woven fabric, and calculate the characteristic value of each pixel in the grayscale image; In response to the feature value of the pixel point exceeding the preset segmentation threshold interval, the pixel point is marked as a defect point to monitor the quality of the filter element; The calculation method of the characteristic value of each pixel is: A sliding window is constructed with the pixel point to be calculated as the center, and the grayscale run-length matrix of the pixel point in the sliding window is obtained. The eigenvalue of the pixel point is calculated according to the grayscale run-length matrix. The calculation formula is:
[0007] in, Represents the characteristic value of the pixel, Indicates that the grayscale run matrix direction angle is hour, Indicates that the grayscale run matrix direction angle is When the gray value is The run length is The number of times it appears in the sliding window, represents the maximum grayscale value in the sliding window, Represents the pixel value in the sliding window The total number of Indicates the maximum value of the run length in the grayscale run matrix, Represents the pixel value in the sliding window The total number of Represents the number of pixels in the sliding window, Represents the gray value, Indicates the run length.
[0008] By calculating the characteristic value of each pixel, comprehensively considering the grayscale non-uniformity and grayscale value of the pixel when calculating the characteristic value, judging whether the pixel area is abnormal according to the characteristic value of the pixel, the monitoring accuracy of the defect area is improved.
[0009] Preferably, the method further includes: calculating the eigenvalues of the pixels in the grayscale image, and constructing a histogram of the eigenvalues and quantities.
[0010] By establishing a histogram, we can intuitively understand the types and quantity of characteristic values of pixels in the grayscale image, which is convenient for monitoring the grayscale image.
[0011] Preferably, the method for determining the segmentation threshold interval is: Obtain the grayscale image of the non-woven fabric in real time, construct a histogram based on the grayscale image, and determine the segmentation threshold interval based on the histogram; The segmentation threshold interval is: ; in, is the largest eigenvalue in the histogram, is the distance value from the left end of the segmentation threshold interval, is the distance value to the right end of the segmentation threshold interval.
[0012] By obtaining the segmentation threshold interval, it is possible to accurately determine whether a pixel point is a defect point, thereby improving the accuracy of the monitoring process.
[0013] Preferably, the method for obtaining the left end distance value and the right end distance value is: Obtain multiple historical grayscale images of non-woven fabrics, mark the normal area of each historical grayscale image, and calculate the characteristic values of the pixels in the marked normal area; Construct a histogram according to the characteristic value of each historical grayscale image; Calculate the first difference between the largest eigenvalue and the smallest eigenvalue in the histogram, and use the smallest first difference among the multiple histograms as the left end distance value; The second difference between the maximum eigenvalue and the eigenvalue with the largest number in the histogram is calculated, and the smallest second difference among the multiple histograms is used as the right end distance value.
[0014] Preferably, the method for obtaining the left end distance value and the right end distance value is: Obtain multiple historical grayscale images of non-woven fabrics, mark the normal area of each historical grayscale image, and calculate the characteristic values of the pixels in the marked normal area; Construct a histogram according to the characteristic value of each historical grayscale image; Calculate the first difference between the largest eigenvalue and the smallest eigenvalue in the histogram, and take the average of the first differences in multiple histograms as the left end distance value; The second difference between the maximum eigenvalue and the eigenvalue with the largest number in the histogram is calculated, and the average of the second differences in multiple histograms is used as the right end distance value.
[0015] The left end distance value and the right end distance value are obtained through multiple historical images, so that the segmentation threshold interval can be further obtained, thereby improving the accuracy of the threshold segmentation interval.
[0016] Preferably, outlier analysis is used to remove abnormal defect points, and the remaining defect points are clustered to obtain defect areas.
[0017] By removing abnormal defect points and eliminating the interference of noise on the clustering results, the defect area can be accurately obtained, which provides a theoretical basis for cutting the defect area and improves the utilization rate of non-woven fabrics.
[0018] Preferably, the method further comprises: cutting the defective area of the non-woven fabric to obtain a normal area.
[0019] Preferably, the remaining defect points are clustered using a K-Means algorithm.
[0020] The present invention has the following technical effects: By calculating the eigenvalue of each pixel, comprehensively considering the grayscale non-uniformity and grayscale value of the pixel when calculating the eigenvalue, and determining the segmentation threshold interval, it is determined whether the pixel area is abnormal based on whether the eigenvalue of the pixel is within the segmentation threshold interval, thereby improving the monitoring accuracy of the defect area. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts.
[0022] Figure 1 The present invention is a method flow chart of a filter element quality monitoring method based on image processing according to an embodiment of the present invention. DETAILED DESCRIPTION
[0023] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.
[0024] It should be understood that when the terms "first", "second", etc. are used in the claims, descriptions, and drawings of the present invention, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the description and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their collections.
[0025] The embodiment of the present invention discloses a filter element quality monitoring method based on image processing, referring to Figure 1 , including steps S1 to S5, which are specifically as follows: S1: Obtain a grayscale image of the non-woven fabric of the filter element.
[0026] A white light source is deployed below the non-woven fabric used to process the filter element, and a device for collecting images of the non-woven fabric is deployed above the non-woven fabric. According to the width of the non-woven fabric and the length of the non-woven fabric actually used to process a filter element, the shooting angle and parameters of the image acquisition device are adjusted to ensure that the image acquisition device can completely capture the area of the non-woven fabric actually used to process a filter element in one image acquisition. After the image of the non-woven fabric surface is collected, it is converted into a grayscale image and the image is denoised to reduce the impact of noise.
[0027] S2: Calculate the eigenvalue of each pixel in the grayscale image.
[0028] The calculation method of the characteristic value of a pixel is: S21: Construct a sliding window with the pixel point whose eigenvalue is to be calculated as the center.
[0029] In this embodiment, a sliding window with a size of n×m and a moving step of 1 pixel is set, and the sliding window is used to slide on the grayscale image of the non-woven fabric surface. When sliding, it starts from the upper left corner of the image, takes the first boundary pixel in the upper left corner as the center of the sliding window, and slides horizontally on the image with a step of 1 pixel. When sliding to the edge pixel in the horizontal direction, it moves down by one pixel and then slides from left to right again in sequence until all the pixels in the grayscale image are traversed. When the sliding window slides at the boundary of the image, the pixels of the grayscale image cannot completely fill the sliding window. The sliding window is filled by copying and filling the boundary pixels, so that the sliding operation of the window can be performed normally, so that the sliding window corresponds to the pixels in the grayscale image.
[0030] S22: Obtain the grayscale run-length matrix of the pixel points in the sliding window, and calculate the eigenvalues of the pixel points according to the grayscale run-length matrix.
[0031] The calculation formula of the eigenvalue is:
[0032] in, Represents the characteristic value of the pixel, Indicates that the grayscale run matrix direction angle is hour, Indicates that the grayscale run matrix direction angle is When the gray value is The run length is The number of occurrences in the sliding window can also be understood as the grayscale run matrix Line The element value of the column, represents the maximum grayscale value in the sliding window, Represents the pixel value in the sliding window The total number of Indicates the maximum value of the run length in the grayscale run matrix, Represents the pixel value in the sliding window The total number of Indicates the number of pixels in the sliding window. In this embodiment, the grayscale run matrix direction angle is 0 degrees, Represents the gray value, Indicates the run length.
[0033] For example, in the grayscale run length matrix, the first row represents the pixel value , the second line indicates the pixel value , ..., the first column indicates the run length , the second column indicates the run length , ..., in the grayscale run matrix middle, It means that the grayscale run matrix direction angle is 0, the grayscale value is 1, and the run length is 2, which appears 2 times. It can also be understood as: Represents the value 2 in the first row and second column.
[0034] In the formula, Indicates the grayscale non-uniformity of the pixels in the sliding window. The larger the value, the more uneven the grayscale distribution of the pixels in the sliding window. Indicates the distribution of run lengths of different grayscale values in the matrix. The larger the value, the more run lengths are relatively large in the image, and the more uneven the grayscale distribution is. A normalization factor is provided to facilitate comparison of grayscale distributions of different images.
[0035] Represents the influence of different gray values on the eigenvalue. For example, when When the value is small, it means that there are more pixels with smaller grayscale values. When the value of is large, it means that there are more pixels with larger grayscale values; the eigenvalue represents the abnormality of the pixel. When the eigenvalue of the pixel is too large or too small, it means that the pixel is a defective point. For example, when the eigenvalue of the pixel is too large, it means that the pixel area is thin or the pixel area is a hole defect. When the eigenvalue of the pixel is too small, it means that the pixel area is thick or there are debris attached to the pixel area. Obtaining the grayscale run matrix of the pixel in the sliding window is a prior art, and the specific process will not be repeated here.
[0036] S3: In response to the feature value of the pixel point exceeding the preset segmentation threshold interval, the pixel point is marked as a defect point to monitor the quality of the filter element.
[0037] In one embodiment, the segmentation threshold interval is determined by: S31: Calculate the eigenvalues of the pixels in the grayscale image and construct a histogram of the eigenvalues and quantities.
[0038] In the histogram, the horizontal axis is the eigenvalue and the vertical axis is the quantity. The histogram reflects the distribution state of the number of eigenvalues of pixels in the grayscale image. One grayscale image corresponds to one histogram.
[0039] S32: acquiring a grayscale image of the non-woven fabric in real time, constructing a histogram according to the acquired grayscale image, and determining the largest number of eigenvalues according to the histogram.
[0040] When a grayscale image is obtained, the eigenvalues of the pixels in the grayscale image are calculated, and a histogram is constructed, and the eigenvalue m with the largest number is determined through the histogram.
[0041] S33: Obtain multiple historical grayscale images of non-woven fabrics, mark the normal area of each historical grayscale image, and calculate the characteristic values of the pixels in the marked normal area; construct a histogram according to the characteristic values of each historical grayscale image; calculate the first difference between the largest eigenvalue and the smallest eigenvalue in the histogram, and use the smallest first difference among multiple histograms as the left end distance value L; calculate the second difference between the largest eigenvalue and the largest eigenvalue in the histogram, and use the smallest second difference among multiple histograms as the right end distance value R.
[0042] S34: The segmentation threshold interval is: ; in, is the largest eigenvalue in the histogram, is the distance value from the left end of the segmentation threshold interval, is the distance value to the right end of the segmentation threshold interval.
[0043] According to the obtained segmentation threshold, defect points can be identified. Specifically, if the feature value is within the segmentation threshold interval, the pixel point is a normal pixel point. If it is outside the segmentation threshold interval, it means that the pixel point is a defect point. If it is located on the left side of the segmentation threshold interval, it indicates that the non-woven fabric area corresponding to the pixel point has defects such as thick thickness, stains on the surface, or scratches on the surface. If it is located on the right side of the segmentation threshold interval, it indicates that the non-woven fabric area corresponding to the pixel point has defects such as thin thickness, holes on the surface, or scratches on the surface. It should be noted that the feature value of the pixel point exceeds the preset segmentation threshold interval, including: the feature value is less than and eigenvalues greater than Two situations.
[0044] In other embodiments, the method for obtaining the left end distance value and the right end distance value is: obtain multiple historical grayscale images of non-woven fabrics, mark the normal area of each historical grayscale image, and calculate the characteristic values of the pixels in the marked normal area; construct a histogram according to the characteristic values of each historical grayscale image; calculate the first difference between the largest eigenvalue and the smallest eigenvalue in the histogram, and use the average of the first difference in multiple histograms as the left end distance value; calculate the second difference between the largest eigenvalue and the largest eigenvalue in the histogram, and use the average of the second difference in multiple histograms as the right end distance value.
[0045] S4: Use outlier analysis to remove abnormal defect points, and cluster the remaining defect points to obtain the defect area.
[0046] By removing abnormal defect points and using the K-Means algorithm to cluster the remaining defect points to obtain the defect area, the accuracy of the clustering result is improved, and the defect area of the grayscale image can be accurately obtained.
[0047] S5: Cut the defective area to obtain a normal area of the nonwoven fabric.
[0048] The defective area of the grayscale image of the non-woven fabric surface currently used for processing the filter element and the position of the defective area on the grayscale image are obtained, the cutting position of the current non-woven fabric is calculated, the defective area of the non-woven fabric currently used for processing the filter element is cut off, the normal area of the non-woven fabric used for processing the filter element is retained, and then the non-woven fabric is processed into a filter element.
[0049] By calculating the eigenvalue of each pixel, comprehensively considering the grayscale non-uniformity and grayscale value of the pixel when calculating the eigenvalue, and determining the segmentation threshold interval, it is determined whether the pixel area is abnormal based on whether the eigenvalue of the pixel is within the segmentation threshold interval, thereby improving the monitoring accuracy of the defect area.
[0050] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.
[0051] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. A filter element quality monitoring method based on image processing, It is characterized in that Includes steps: Obtain a grayscale image of the filter element non-woven fabric, and calculate the characteristic value of each pixel in the grayscale image; In response to the feature value of the pixel point exceeding a preset segmentation threshold interval, marking the pixel point as a defect point; The calculation method of the characteristic value of each pixel is: A sliding window is constructed with the pixel point to be calculated as the center, and the grayscale run-length matrix of the pixel point in the sliding window is obtained. The eigenvalue of the pixel point is calculated according to the grayscale run-length matrix. The calculation formula is: in, Represents the characteristic value of the pixel, Indicates that the grayscale run matrix direction angle is When the gray value is The run length is The number of times it appears in the sliding window, represents the maximum grayscale value in the sliding window, Represents the pixel value in the sliding window The total number of Indicates the maximum value of the run length in the grayscale run matrix, Represents the number of pixels in the sliding window, Represents the gray value, Indicates the run length.
2. The filter element quality monitoring method based on image processing according to claim 1, It is characterized in that Also includes: Calculate the eigenvalue of each pixel in the grayscale image and construct a histogram of the eigenvalue and quantity.
3. The filter element quality monitoring method based on image processing according to claim 2, It is characterized in that The method for determining the segmentation threshold interval is: Obtain the grayscale image of the non-woven fabric in real time, construct a histogram based on the grayscale image, and determine the segmentation threshold interval based on the histogram; The segmentation threshold interval is: ,in, is the largest eigenvalue in the histogram, is the distance value from the left end of the segmentation threshold interval, is the distance value to the right end of the segmentation threshold interval.
4. The filter element quality monitoring method based on image processing according to claim 3, It is characterized in that The method for obtaining the left end distance value and the right end distance value is: Obtain multiple historical grayscale images of non-woven fabrics, mark the normal area of each historical grayscale image, and calculate the characteristic values of the pixels in the marked normal area; Construct a histogram according to the characteristic value of each historical grayscale image; Calculate the first difference between the largest eigenvalue and the smallest eigenvalue in the histogram, and use the smallest first difference among the multiple histograms as the left end distance value; The second difference between the maximum eigenvalue and the eigenvalue with the largest number in the histogram is calculated, and the smallest second difference among the multiple histograms is used as the right end distance value.
5. The filter element quality monitoring method based on image processing according to claim 3, It is characterized in that The method for obtaining the left end distance value and the right end distance value is: Obtain multiple historical grayscale images of non-woven fabrics, mark the normal area of each historical grayscale image, and calculate the characteristic values of the pixels in the marked normal area; Construct a histogram according to the characteristic value of each historical grayscale image; Calculate the first difference between the largest eigenvalue and the smallest eigenvalue in the histogram, and take the average of the first differences in multiple histograms as the left end distance value; The second difference between the maximum eigenvalue and the eigenvalue with the largest number in the histogram is calculated, and the average of the second differences in multiple histograms is used as the right end distance value.
6. The filter element quality monitoring method based on image processing according to claim 1, It is characterized in that Outlier analysis method is used to remove abnormal defect points, and the remaining defect points are clustered to obtain the defect area.
7. The filter element quality monitoring method based on image processing according to claim 6, It is characterized in that Also includes: The defective area of the nonwoven fabric is cut to obtain a normal area.
8. The filter element quality monitoring method based on image processing according to claim 6, It is characterized in that The K-Means algorithm is used to cluster the remaining defect points.
Citation Information
Patent Citations
Textile defect detection and recognition device
CN114627111A
Fabric fuzzing and pilling detection method based on gray run matrix
CN115170570A
Textile defect rapid detection method based on image features
CN116934749A
Laminated yarn defect detection method for cloth surface
CN117237346A
Cited By
Self-adaptive cleaning control method of cleaning robot based on image texture analysis
CN121767759A
