Defect detection method for miniature bearing manufacturing based on image enhancement

Abstract

The invention discloses a defect detection method for micro bearing manufacturing based on image enhancement, which relates to the technical field of image enhancement. A data acquisition terminal is set to collect manufacturing picture data and standard pattern data of a micro bearing; an input standard is set to perform standard proofreading on the manufacturing picture data and the standard pattern data to obtain a standard manufacturing picture and a standard standard pattern; a filter matrix is ​​set to perform pixel optimization and update on the standard manufacturing picture to obtain replacement pixels, and the filter matrix is ​​smoothly traversed in the standard manufacturing picture according to the replacement pixels to obtain an optimized pixel picture; pixel optimization and update and smooth traversal are performed on the standard standard pattern to obtain a standard optimized pattern, a matching comparison value is obtained by performing pixel verification with the optimized pixel picture, a defect threshold is set to perform threshold comparison with the matching comparison value to obtain an abnormal defect picture; the defect detection accuracy and detection speed are improved, the probability of defect misjudgment is reduced, and the image quality is improved.

Description

Technical Field

[0001] The invention relates to the technical field of image enhancement, and in particular to a defect detection method for miniature bearing manufacturing based on image enhancement. Background Art

[0002] Image enhancement refers to the use of various algorithms and techniques to improve the quality of images, enhance the features and details of images, and make images clearer, more vivid, and easier to analyze. Common image enhancement techniques include contrast enhancement, brightness adjustment, color balance, sharpening, and denoising. Image enhancement techniques can make images more visually appealing, accurate, and informative, and improve image readability and comprehensibility.

[0003] Miniature bearings are usually used in small equipment, precision instruments, electronic products and other fields because they are small, light, high speed, low noise and other characteristics. Miniature bearing manufacturing refers to the process of producing and manufacturing these tiny bearings. Because miniature bearings are usually used in some equipment with high requirements for precision and detail, the manufacturing process of miniature bearings also needs to be very precise and rigorous; therefore, intelligent defect detection of miniature bearings used for manufacturing has important theoretical and practical significance.

[0004] How to use image enhancement technology to perform standard proofreading on the collected manufacturing image data and standard drawing data to obtain standard manufacturing images and standard standard drawings; set the filter matrix to perform pixel optimization update and smooth traversal on the standard manufacturing images and standard standard drawings respectively to obtain optimized pixel images and standard optimized drawings, perform pixel verification on the standard optimized drawings and optimized pixel images to obtain matching comparison values, set defect thresholds and compare thresholds with matching comparison values ​​to obtain abnormal defect images, which are the problems we need to solve; for this purpose, a defect detection method for miniature bearing manufacturing based on image enhancement is now provided. Summary of the invention

[0005] The purpose of the present invention can be achieved through the following technical solutions:

[0006] A defect detection method for miniature bearing manufacturing based on image enhancement comprises the following steps:

[0007] Step S1: Setting a data acquisition terminal, collecting comprehensive data of the miniature bearing through the set data acquisition terminal, and obtaining manufacturing picture data and standard drawing data;

[0008] Step S2: setting input standards to perform standard proofreading on manufacturing picture data and standard pattern data, obtaining standard manufacturing pictures and standard pattern data, setting a filter matrix, uploading the filter matrix to the standard manufacturing picture, obtaining a filter interval, and setting optimized pixel points and traversal order within the filter interval;

[0009] Step S3: Optimize and update the optimized pixel points in the filter interval according to the obtained filter matrix to obtain replacement pixels, perform statistics on the obtained replacement pixels to obtain the optimized pixel interval, smoothly traverse the obtained filter matrix in the standard manufacturing picture, and perform comprehensive statistics on the obtained optimized pixel interval to obtain the optimized pixel picture;

[0010] Step S4: Perform pixel optimization update and smooth traversal on the obtained standard pattern to obtain a standard optimized pattern, perform standard matching on the obtained standard optimized pattern and the optimized pixel image to obtain a matching optimized pixel image and a matching standard pattern, perform pixel verification on the obtained matching optimized pixel image and the matching standard pattern to obtain a matching comparison value, set a defect threshold and perform threshold comparison with the matching comparison value to obtain an abnormal defect image.

[0011] The process of obtaining manufacturing image data and standard drawing data includes:

[0012] Obtain a manufacturing workshop of a miniature bearing, and set a data collection terminal according to the obtained manufacturing workshop;

[0013] The comprehensive data of the miniature bearing is collected through the set data collection terminal to obtain manufacturing picture data and standard drawing data;

[0014] The obtained manufacturing picture data is associated with the corresponding standard pattern data.

[0015] The process of setting up the filter matrix includes:

[0016] Setting input standards, and performing standard proofreading on the obtained manufacturing picture data according to the obtained input standards to obtain standard manufacturing pictures;

[0017] Perform standard proofreading on the obtained standard drawing data according to the obtained input standard to obtain the standard drawing;

[0018] Marking pixels of the obtained standard manufacturing pictures to obtain standard pixel points;

[0019] A filter matrix is ​​set according to the obtained standard manufacturing picture, and a matrix center point is obtained according to the filter matrix.

[0020] The process of obtaining the filtering interval includes:

[0021] A traversal starting point is set in the obtained standard manufacturing picture, the obtained filter matrix is ​​uploaded to the traversal starting point of the standard manufacturing picture, and the corresponding area of ​​the filter matrix in the standard manufacturing picture is marked as a filter interval;

[0022] Set the optimized pixels and traversal order within the filtering interval.

[0023] The process of obtaining the replacement pixel includes:

[0024] The matrix center point of the obtained filter matrix is ​​matched with the optimized pixel point in the filter interval to obtain the pixel dot product, and the obtained pixel dot products are summed to obtain the pixel product sum;

[0025] The obtained filter matrix is ​​counted to obtain the total number of elements, and the value-added pixel is obtained according to the total number of elements and the pixel product number;

[0026] The optimized pixel points are updated based on the obtained value-added pixels to obtain replacement pixels.

[0027] The process of obtaining pixel-optimized images includes:

[0028] The next optimized pixel point in the filtering interval is obtained according to the obtained traversal order, and the obtained standard pixel points are optimized and updated to obtain replacement pixels, until all standard pixel points in the filtering interval have completed pixel optimization and update, and the obtained replacement pixels are counted according to the obtained filtering interval to obtain an optimized pixel interval;

[0029] The obtained filter matrix is ​​smoothly traversed in the standard manufacturing picture until the filter matrix covers all filter intervals of the standard manufacturing picture, and the obtained optimized pixel intervals are comprehensively counted to obtain the optimized pixel picture.

[0030] The process of obtaining a matching comparison value includes:

[0031] Marking pixels of the obtained standard pattern to obtain standard pixel points;

[0032] Upload the obtained filter matrix to the standard drawing to obtain the standard filter interval;

[0033] Setting standard optimized pixels according to standard pixels within the standard filtering interval;

[0034] Perform pixel optimization update on the standard optimized pixel points in the standard filtering interval according to the obtained filtering matrix to obtain a standard optimized pattern;

[0035] Performing standard matching on the optimized pixel image according to the obtained standard optimized pattern to obtain a matching optimized pixel image and a matching standard pattern;

[0036] The obtained matching specification pattern is equally divided to obtain a local inspection frame, and the obtained local inspection frame is uploaded to the matching optimization pixel map to obtain a matching verification area;

[0037] The obtained matching verification area is pixel-verified according to the obtained local inspection frame to obtain a matching comparison value.

[0038] Set a defect threshold, compare the obtained matching comparison value with the defect threshold, obtain an abnormal matching value, and mark the matching verification area corresponding to the abnormal matching value as an abnormal matching area;

[0039] The abnormal matching areas in the matching optimization pixel map are counted to obtain an abnormal defect map.

[0040] Compared with the prior art, the beneficial effects of the present invention are as follows: standard proofreading is performed on the collected manufacturing picture data and standard pattern data, a standard manufacturing picture and a standard pattern are obtained, a filter matrix is ​​set, the filter matrix is ​​uploaded to the standard manufacturing picture, a filter interval is obtained, and an optimized pixel point and a traversal order are set within the filter interval; the filter matrix is ​​set to update the pixels of the optimized pixel points, thereby improving the visual effect of the picture and improving the readability and recognizability of the picture;

[0041] According to the obtained filter matrix, the optimized pixel points in the filter interval are optimized and updated and smoothly traversed to obtain an optimized pixel picture; providing clear and high-quality images for defect detection, which is conducive to pixel comparison between the optimized pixel picture and the standard optimized pattern, obtaining an abnormal defect map, and improving the defect detection speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0043] Figure 1 It is a schematic diagram of the present invention. DETAILED DESCRIPTION

[0044] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] like Figure 1 As shown, the defect detection method for miniature bearing manufacturing based on image enhancement includes the following steps:

[0046] Step S1: Setting a data acquisition terminal, collecting comprehensive data of the miniature bearing through the set data acquisition terminal, and obtaining manufacturing picture data and standard drawing data;

[0047] Step S2: setting input standards to perform standard proofreading on manufacturing picture data and standard pattern data, obtaining standard manufacturing pictures and standard pattern data, setting a filter matrix, uploading the filter matrix to the standard manufacturing picture, obtaining a filter interval, and setting optimized pixel points and traversal order within the filter interval;

[0048] Step S3: Optimize and update the optimized pixel points in the filter interval according to the obtained filter matrix to obtain replacement pixels, perform statistics on the obtained replacement pixels to obtain the optimized pixel interval, smoothly traverse the obtained filter matrix in the standard manufacturing picture, and perform comprehensive statistics on the obtained optimized pixel interval to obtain the optimized pixel picture;

[0049] Step S4: Perform pixel optimization update and smooth traversal on the obtained standard pattern to obtain a standard optimized pattern, perform standard matching on the obtained standard optimized pattern and the optimized pixel image to obtain a matching optimized pixel image and a matching standard pattern, perform pixel verification on the obtained matching optimized pixel image and the matching standard pattern to obtain a matching comparison value, set a defect threshold and perform threshold comparison with the matching comparison value to obtain an abnormal defect image.

[0050] It should be further explained that, in the specific implementation process, the process of collecting manufacturing image data and standard pattern data includes:

[0051] Obtain a manufacturing workshop of a miniature bearing, and set a data collection terminal according to the obtained manufacturing workshop;

[0052] Furthermore, the manufacturing workshop is a manufacturing plant for manufacturing miniature bearings, and has a complete production line for manufacturing miniature bearings;

[0053] According to the obtained manufacturing workshop, the data acquisition terminal is set at the corresponding position of the manufacturing workshop, and the comprehensive data of the miniature bearing is collected through the set data acquisition terminal to obtain the manufacturing picture data and the standard drawing data;

[0054] It should be further explained that, in the specific implementation process, the data acquisition end is set at the corresponding production node on the production line of the miniature bearing manufacturing workshop to collect the manufacturing image data at the production node;

[0055] Associating the obtained manufacturing picture data with corresponding standard pattern data;

[0056] Furthermore, the standard drawing data is a standard manufacturing drawing required for reference in manufacturing miniature bearings, including the bearing name, bearing drawing, and standard bearing size.

[0057] Setting input criteria, including pixel limits and channel criteria;

[0058] It should be further explained that, in the specific implementation process, the pixel limit represents the scale of the manufacturing image data in the spatial dimension, including the length limit and the width limit, the length limit represents the number of pixels in the horizontal direction, and the width limit represents the number of pixels in the vertical direction, for example, 224×224 pixels, 128×128 pixels, 512×512 pixels; the channel standard represents the number of identifiable color channels, for example, if the channel standard is 1, it means there is only 1 color channel, and if the channel standard is 3, it means there are 3 color channels;

[0059] Performing standard proofreading on the manufacturing picture data according to the input standard obtained to obtain a standard manufacturing picture;

[0060] Perform standard proofreading on standard drawing data according to the input standards obtained to obtain standard drawings;

[0061] Marking pixels of the obtained standard manufacturing pictures to obtain standard pixel points;

[0062] The standard pixel point is the position occupied by each pixel in the standard manufacturing image, recorded as the standard pixel point, and the value of the standard pixel point is marked as the standard pixel value;

[0063] A filter matrix is ​​set according to the obtained standard manufacturing picture, wherein the filter matrix includes a matrix length and a matrix width, and the matrix length and the matrix width are equal, the matrix width represents the number of rows, and the matrix length represents the number of columns, for example, a 3×3 filter matrix represents a matrix of 3 rows and 3 columns, with a total of 9 elements, and a 5×5 filter matrix represents a matrix of 5 rows and 5 columns, with a total of 25 elements;

[0064] Setting a traversal starting point in the obtained standard manufacturing picture;

[0065] Further, in this embodiment, the traversal starting point is the upper left corner of the standard manufacturing picture;

[0066] Upload the obtained filter matrix to the traversal starting point of the standard manufacturing image, and mark the corresponding area of ​​the filter matrix in the standard manufacturing image as a filter interval;

[0067] Obtaining a matrix center point according to a matrix length and a matrix width included in the filter matrix, wherein the matrix center point is an element at an intersection of diagonal lines of the filter matrix;

[0068] Set the optimized pixel points and traversal order within the filtering interval;

[0069] The obtained optimized pixel point is marked as p, where p=1, 2, 3, ..., v1, v1 is a positive integer;

[0070] Furthermore, the optimized pixel point is to select any standard pixel point in the filtering interval as the optimized pixel point for subsequent processing, and the traversal order is the selection order of the optimized pixel points. For example, the traversal order is from left to right, then the first row of standard pixels is traversed, and then the second row is traversed from left to right until all standard pixels in the filtering interval are traversed; in particular, after the optimized pixel point processing is completed, the pixel point optimization is continued according to the obtained traversal order until all standard pixels in the filtering interval are selected as optimized pixels.

[0071] The matrix center point of the obtained filter matrix is ​​matched with the optimized pixel point in the filter interval to obtain the pixel dot product;

[0072] The element correspondence indicates that after the center point of the matrix corresponds to the optimized pixel point at the same position, the element in the filter matrix is ​​multiplied by the standard pixel value at the corresponding position of the filter interval to obtain the pixel dot product, and the obtained pixel dot product is marked as ,in, , and 2 represents the weight factor, Represents the element value in the filter matrix, represents the standard pixel value in the filtering interval, j represents the element number in the filtering matrix, j=1, 2, 3, ..., v2, v2 is a positive integer, k represents the standard pixel point number in the filtering interval, k=1, 2, 3, ..., v3, v3 is a positive integer;

[0073] The obtained pixel dot products are summed to obtain the pixel product sum, and the obtained pixel product sum is marked as ,in, = , i represents the number of the filtering interval, i=1, 2, 3, ..., v4, v4 is a positive integer;

[0074] Performing statistics on the obtained filter matrix to obtain a total number of elements, where the total number of elements represents the total number of elements in the filter matrix;

[0075] The total amount of elements obtained is marked as ;

[0076] According to the total number of elements and the number of pixel products, the obtained value-added pixels are marked as ,in, , Indicates taking The minimum value in ;

[0077] Perform pixel optimization and update on the optimized pixel points according to the obtained value-added pixels to obtain replacement pixels;

[0078] Furthermore, the pixel optimization update is to replace the pixel value of the optimized pixel point with the value-added pixel, and the pixel value of the optimized pixel point is the replacement pixel;

[0079] The next optimized pixel point in the filtering interval is obtained according to the obtained traversal order, and the obtained standard pixel points are optimized and updated to obtain replacement pixels, until all standard pixel points in the filtering interval have completed pixel optimization and update, and the obtained replacement pixels are counted according to the obtained filtering interval to obtain an optimized pixel interval;

[0080] The range of the optimized pixel interval is the corresponding area of ​​the filtering interval;

[0081] The obtained filter matrix is ​​smoothly traversed in the standard manufacturing picture until the filter matrix covers all filter intervals of the standard manufacturing picture, and the obtained optimized pixel intervals are comprehensively counted to obtain the optimized pixel picture.

[0082] Marking pixels of the obtained standard pattern to obtain standard pixel points, and marking the values ​​of the standard pixel points as standard pixel values;

[0083] Upload the obtained filter matrix to the standard drawing to obtain the standard filter interval;

[0084] Setting standard optimized pixels according to standard pixels within the standard filtering interval;

[0085] Perform pixel optimization update on the standard optimized pixel points in the standard filtering interval according to the obtained filtering matrix to obtain a standard optimized pattern;

[0086] Performing standard matching on the optimized pixel image according to the obtained standard optimized pattern to obtain a matching optimized pixel image and a matching standard pattern;

[0087] Furthermore, the specification matching is to match the manufacturing image data corresponding to the optimized pixel image with the associated standard pattern data, and obtain the specification optimization pattern of the successfully matched standard pattern data, and mark the obtained specification optimization pattern as the matching specification pattern, and mark the optimized pixel image that successfully matches the specification optimization pattern as the matching optimized pixel image;

[0088] The obtained matching standard pattern is equally divided to obtain a local standard image;

[0089] It should be further explained that, in the specific implementation process, the equal segmentation means that the matching standard pattern is evenly segmented into partial images of the same size and shape;

[0090] Mark the obtained local standard image as a local inspection frame;

[0091] Uploading the obtained local inspection frame to the corresponding position of the matching optimized pixel map, and marking the local inspection frame in the corresponding area of ​​the matching optimized pixel map as a matching verification area;

[0092] Perform pixel verification on the obtained matching verification area according to the obtained local inspection frame to obtain a matching comparison value;

[0093] It should be further explained that, in the specific implementation process, the pixel verification is to perform difference calculation between the standard pixel value of the standard pixel point in the local inspection frame and the standard pixel value of the standard pixel point at the corresponding position of the matching verification area to obtain the matching comparison value, and mark the obtained matching comparison value as ,in, = - , Represents the standard pixel value of the standard pixel point, Indicates the standard pixel value of the standard pixel point;

[0094] Setting a defect threshold, marking the obtained defect threshold as Y0, in particular, in this embodiment, Y0=0;

[0095] The obtained matching comparison value is threshold-compared with the defect threshold to obtain an abnormal matching value and a qualified matching value;

[0096] Furthermore, the threshold comparison indicates that =Y0, indicating that the matching value is a qualified matching value, otherwise, it is an abnormal matching value;

[0097] Mark the matching verification area corresponding to the abnormal matching value as the abnormal matching area, and mark the matching verification area corresponding to the qualified matching value as the qualified matching area;

[0098] Count the abnormal matching areas in the matching optimization pixel map to obtain an abnormal defect map;

[0099] Counting qualified matching areas in the matching optimization pixel map to obtain a qualified standard map;

[0100] It should be further explained that, in the specific implementation process, the abnormal matching areas in the matching optimization pixel map are counted, which means that the obtained matching comparison values ​​are resolved. If the matching comparison value is an abnormal matching value, it means that there is an abnormal matching area in the matching optimization pixel map, and the matching optimization pixel map is an abnormal defect map, that is, the manufactured micro-bearing is not exactly the same as the standard pattern data, indicating that the micro-bearing is defective and is an abnormal product; if the obtained matching comparison values ​​are all qualified matching values, it means that the matching optimization pixel map is all qualified matching areas, and the matching optimization pixel map is a qualified specification map, indicating that the manufactured micro-bearing is the same as the standard pattern data, that is, the micro-bearing has no defects and is a qualified product.

[0101] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A defect detection method for miniature bearing manufacturing based on image enhancement, characterized in that: The following steps are involved: Step S1: Setting a data acquisition terminal, collecting comprehensive data of the miniature bearing through the set data acquisition terminal, and obtaining manufacturing picture data and standard drawing data; Step S2: setting input standards to perform standard proofreading on manufacturing picture data and standard pattern data, obtaining standard manufacturing pictures and standard pattern, setting filter matrix, uploading filter matrix to standard manufacturing pictures, obtaining filter interval, setting optimized pixel points and traversal order within the filter interval; the input standards include pixel limit and channel standard; the pixel limit indicates the scale of manufacturing picture data in spatial dimension, including length limit and width limit, the length limit indicates the number of pixels in the horizontal direction, and the width limit indicates the number of pixels in the vertical direction; the channel standard indicates the number of identifiable color channels; Step S3: Optimize and update the optimized pixel points in the filter interval according to the obtained filter matrix to obtain replacement pixels, perform statistics on the obtained replacement pixels to obtain the optimized pixel interval, smoothly traverse the obtained filter matrix in the standard manufacturing picture, and perform comprehensive statistics on the obtained optimized pixel interval to obtain the optimized pixel picture; Step S4: Perform pixel optimization update and smooth traversal on the obtained standard pattern to obtain a standard optimized pattern, perform standard matching on the obtained standard optimized pattern and the optimized pixel image to obtain a matching optimized pixel image and a matching standard pattern, perform pixel verification on the obtained matching optimized pixel image and the matching standard pattern to obtain a matching comparison value, set a defect threshold and perform threshold comparison with the matching comparison value to obtain an abnormal defect image.

2. The method for defect detection in micro bearing manufacturing based on image enhancement according to claim 1, characterized in that: The process of obtaining manufacturing image data and standard drawing data includes: Obtain a manufacturing workshop of a miniature bearing, and set a data collection terminal according to the obtained manufacturing workshop; The comprehensive data of the miniature bearing is collected through the set data collection terminal to obtain manufacturing picture data and standard drawing data; The obtained manufacturing picture data is associated with the corresponding standard pattern data.

3. The method for defect detection in micro bearing manufacturing based on image enhancement according to claim 2, characterized in that: The process of setting up the filter matrix includes: Setting input standards, and performing standard proofreading on the obtained manufacturing picture data according to the obtained input standards to obtain standard manufacturing pictures; Perform standard proofreading on the obtained standard drawing data according to the obtained input standard to obtain the standard drawing; Marking pixels of the obtained standard manufacturing pictures to obtain standard pixel points; A filter matrix is ​​set according to the obtained standard manufacturing picture, and a matrix center point is obtained according to the filter matrix.

4. The method for defect detection in micro bearing manufacturing based on image enhancement according to claim 3, characterized in that: The process of obtaining the filtering interval includes: A traversal starting point is set in the obtained standard manufacturing picture, the obtained filter matrix is ​​uploaded to the traversal starting point of the standard manufacturing picture, and the corresponding area of ​​the filter matrix in the standard manufacturing picture is marked as a filter interval; Set the optimized pixels and traversal order within the filtering interval.

5. The method for defect detection in micro bearing manufacturing based on image enhancement according to claim 4, characterized in that: The process of obtaining the replacement pixel includes: The matrix center point of the obtained filter matrix is ​​matched with the optimized pixel point in the filter interval to obtain the pixel dot product, and the obtained pixel dot product is marked as XY k ,in, β1 and β2 represent weight factors, LY j Indicates the element value in the filter matrix, QX k Indicates the standard pixel value within the filtering interval; The obtained pixel dot products are summed to obtain the pixel product sum, and the obtained pixel product sum is marked as XH i ,in, i represents the number of the filtering interval, i=1, 2, 3, ..., v4, v4 is a positive integer; The obtained filter matrix is ​​counted to obtain the total number of elements, and the value-added pixels are obtained according to the total number of elements and the number of pixel products, and the obtained value-added pixels are marked as X. p ,in, Y 总 Indicates the total amount of elements, XH i represents the sum of pixel products; The optimized pixel points are updated based on the obtained value-added pixels to obtain replacement pixels.

6. The method for defect detection in micro bearing manufacturing based on image enhancement according to claim 5, characterized in that: The process of obtaining pixel-optimized images includes: The next optimized pixel point in the filtering interval is obtained according to the obtained traversal order, and the obtained standard pixel points are optimized and updated to obtain replacement pixels, until all standard pixel points in the filtering interval have completed pixel optimization and update, and the obtained replacement pixels are counted according to the obtained filtering interval to obtain an optimized pixel interval; The obtained filter matrix is ​​smoothly traversed in the standard manufacturing picture until the filter matrix covers all filter intervals of the standard manufacturing picture, and the obtained optimized pixel intervals are comprehensively counted to obtain the optimized pixel picture.

7. The method for defect detection in micro bearing manufacturing based on image enhancement according to claim 6, characterized in that: The process of obtaining a matching comparison value includes: Marking pixels of the obtained standard pattern to obtain standard pixel points; Upload the obtained filter matrix to the standard drawing to obtain the standard filter interval; Setting standard optimized pixels according to standard pixels within the standard filtering interval; Perform pixel optimization update on the standard optimized pixel points in the standard filtering interval according to the obtained filtering matrix to obtain a standard optimized pattern; Performing standard matching on the optimized pixel image according to the obtained standard optimized pattern to obtain a matching optimized pixel image and a matching standard pattern; The obtained matching specification pattern is equally divided to obtain a local inspection frame, and the obtained local inspection frame is uploaded to the matching optimization pixel map to obtain a matching verification area; The obtained matching verification area is pixel-verified according to the obtained local inspection frame to obtain a matching comparison value.

8. The method for defect detection in miniature bearing manufacturing based on image enhancement according to claim 7, characterized in that: The process of obtaining an abnormal defect map includes: Set a defect threshold, compare the obtained matching comparison value with the defect threshold, obtain an abnormal matching value, and mark the matching verification area corresponding to the abnormal matching value as an abnormal matching area; The abnormal matching areas in the matching optimization pixel map are counted to obtain an abnormal defect map.

Images