Product quality inspection method and system applied to garment processing production line

Abstract

The invention relates to the technical field of garment visual inspection, in particular to a product quality inspection method and system applied to a garment processing production line, and the method comprises the steps: enabling all pixel points in a correction image to map corresponding pixel points in a garment contour image, and recording the pixel points as contrast points; analyzing the abscissa difference between all the same abscissa pixel points between the image block where each standard point is located and the image block where the contrast point is located, and the abscissa difference between all the same abscissa pixel points, so as to determine a local window; and based on the difference of pixel values of all pixel points in a neighborhood between each contour pixel point in the local window and the corresponding standard point, obtaining a target point, used for size measurement, of each standard point in the clothing contour map, and carrying out quality inspection on the to-be-detected clothing. The objective of the invention is to improve the precision and efficiency of quality inspection of the size of the to-be-detected garment.

Description

Technical Field

[0001] This application relates to the technical field of clothing visual inspection, and specifically relates to a product quality inspection method and system applied to a clothing processing production line. Background Art

[0002] The processing and production process of clothing usually includes production preparation, fabric inspection, cutting, sewing, ironing and shaping, finished product inspection, packaging, storage and transportation, etc. Among them, the quality inspection of clothing finished products is a key step in the clothing processing and production process, such as clothing surface defect detection, clothing size detection, etc.

[0003] However, the existing clothing size detection method based on visual inspection usually uses a corner detection algorithm to extract a large number of key points from the entire image or contour image of the clothing to be detected, and screens the clothing feature points for clothing size measurement from the extracted key points. Finally, the size measurement of the clothing to be detected is realized based on the screened clothing feature points. However, the method of detecting key points in the image by the corner detection algorithm to measure the clothing size not only has errors in the extracted key points, that is, the screening results, due to lighting conditions and environmental noise, but also this method has a high time complexity, which in turn affects the accuracy and efficiency of clothing product quality inspection. Summary of the Invention

[0004] In order to solve the above technical problems, the purpose of this application is to provide a product quality inspection method and system applied to a clothing processing production line, and the specific technical solutions adopted are as follows:

[0005] In the first aspect, an embodiment of this application provides a product quality inspection method applied to a clothing processing production line, and this method includes the following steps:

[0006] Obtain the clothing contour map and its template contour map of the clothing to be detected, as well as each preset standard point in the template contour map;

[0007] Rotate the central axis of the template contour map to be parallel to the central axis of the clothing contour map, and denote the rotated template contour map as the corrected map; select the pixel point with the smallest ordinate on the central axis of the clothing contour map and the pixel point with the smallest ordinate on the central axis of the corrected map as a reference point pair, and based on the coordinates of the reference point pair, map the coordinates of all pixel points in the corrected map to the clothing contour map, and denote the pixel points corresponding to the mapped pixel points of all pixel points in the corrected map in the clothing contour map as the comparison points;

[0008] Divide the corrected map and the clothing contour map into multiple image blocks in the same way, analyze the differences in the abscissas between all pixel points with the same ordinate and the differences in the ordinates between all pixel points with the same abscissa between the image block where each standard point is located and the image block where its comparison point is located, and determine the position deviation values of each standard point and its comparison point;

[0009] Based on the position deviation value, determine the local window of each standard point and its corresponding point. Based on the difference in pixel values of all pixel points in the neighborhood between each contour pixel point in the local window and the corresponding standard point, determine the feature value of each contour pixel point, so as to obtain the target points for dimension measurement of each standard point in the clothing contour map;

[0010] Based on the coordinates of the target points corresponding to all standard points in the clothing contour map, obtain the dimension data of the clothing to be detected, and conduct quality inspection on the quality of the clothing to be detected.

[0011] Preferably, the rotation of the central axis of the template contour map to be parallel to the central axis of the clothing contour map includes:

[0012] Fit the coordinates of all pixel points on the central axis of the template contour map and the coordinates of all pixel points on the central axis of the clothing contour map respectively to obtain two fitted straight lines. Map the two fitted straight lines to the same spatial coordinate system, use the image rotation algorithm to rotate the template contour map, and take the included angle between the two fitted straight lines as the rotation angle in the image rotation algorithm to obtain the rotated template contour map.

[0013] Preferably, the mapping of the coordinates of all pixel points in the correction map to the clothing contour map includes:

[0014] Assume that the coordinates of the reference point pair correspond to (x a1 , y a1 ) in the correction map and (x c , y c ) in the clothing contour map respectively. Then the coordinates of any pixel point (x, y) in the correction map in the clothing contour map are (x + (x c - x a1 ), y + (y c - y a1 ).

[0015] Preferably, the method for determining the position deviation value of each standard point and its corresponding point is:

[0016] Respectively take the average of the differences in abscissa between all pixel points with the same ordinate and the average of the differences in ordinate between all pixel points with the same abscissa between the image block where each standard point is located and the image block where its corresponding point is located, and denote them as the first average and the second average. Take the maximum value of the first average and the second average as the position deviation value of each standard point and its corresponding point.

[0017] Preferably, the method for determining the local window of each standard point and its corresponding point is:

[0018] The result of adding the ceiling value of the position deviation value between each standard point and its corresponding point to a preset value is used as the local window length of each standard point and its corresponding point. A square window constructed with the corresponding point of each standard point as the center and the local window length as the window side length is used as the local window of each standard point and its corresponding point.

[0019] Preferably, the method for determining the eigenvalue of each contour pixel point is as follows:

[0020] Within the local window of the corresponding point of standard point i, the pixel values of all contour pixel points in the neighborhood of each contour pixel point and in the neighborhood of standard point i are assigned as 1, and the pixel values of the remaining pixel points are assigned as 0. The difference between the re-assigned results of the pixel values of all pixel points between each contour pixel point and the neighborhood of standard point i is used as the eigenvalue of each contour pixel point within the local window of the corresponding point of standard point i.

[0021] Preferably, the target point for dimension measurement of each standard point in the clothing contour map is the contour pixel point corresponding to the minimum eigenvalue within the local window of the corresponding point of each standard point.

[0022] Preferably, the process of obtaining the dimension data of the clothing to be detected is as follows:

[0023] By using the method of calibrating the position coordinates of the target points in the clothing contour map, and calculating the straight-line distances between all target points to calculate the corresponding geometric dimensions, all the dimension data of the clothing to be detected are obtained, where all the dimension data include: sleeve length, shoulder width, garment length, chest width, collar width, collar height, and cuff width.

[0024] Preferably, the quality inspection of the clothing to be detected includes:

[0025] Obtain all the dimension data of the clothing template used for the clothing to be detected, calculate the differences in the corresponding dimensions between the clothing to be detected and the clothing template, and take the average value of all the differences in the corresponding dimensions as the dimension deviation of the clothing to be detected;

[0026] According to the dimension deviation of the clothing to be detected, grade the quality of the clothing to be detected.

[0027] In a second aspect, the embodiment of the present application also provides a product quality inspection system applied to a clothing processing production line, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the steps of the product quality inspection method applied to the clothing processing production line described in any one of the above are implemented.

[0028] The present application has at least the following beneficial effects:

[0029] This application can effectively reduce the influence of factors such as the illumination conditions and environmental noise in the environment on the target points used for measuring the size of the garment to be detected in the garment contour map obtained subsequently by collecting surface images of the garment to be detected from multiple angles and successively performing denoising and image fusion processing on the collected surface images; further, by performing rotation correction on the template contour map of the garment template, the central axis of the template contour map is aligned with the central axis of the garment contour map, improving the accuracy of mapping the pixel points in the template contour map to the garment contour map subsequently, thereby improving the accuracy of screening the target points for garment size measurement in the garment contour;

[0030] This application constructs a position deviation value by analyzing the difference in pixel values of the pixels in the local regions where the standard points and their corresponding points are located between the corrected template contour map, that is, the corrected map and the garment contour map. For each standard point in the corrected map of the garment template, it is the local window where the target point is located in the garment contour map of the garment to be detected, which can avoid the situation that the local window is too small and misses the garment standard points, or the wrong pixel points in the local window are misidentified as target points. At the same time, it shortens the screening time of the target points, can effectively reduce the processing process of redundant pixel points in the garment contour map of the garment to be detected, and further improves the efficiency of obtaining all the target points for size measurement in the garment contour map of the garment to be detected;

[0031] This application obtains the target points for size measurement in the garment contour map by comparing the differences in pixel values of all the pixels in the neighborhood between each contour pixel point in the local window and the corresponding standard point. Compared with the existing method of screening multiple garment feature points from all the key points extracted by the corner detection algorithm, it can more accurately obtain the target points for size measurement in the garment contour map of the garment to be detected, thereby improving the accuracy and efficiency of quality inspection for the size of the garment to be detected. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following-described drawings are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0033] Figure 1 It is a flowchart of the steps of a product quality inspection method applied to a garment processing production line provided by an embodiment of the present application;

[0034] Figure 2 It is a schematic diagram of the acquisition process of the target points for size measurement in the garment contour map provided by an embodiment of the present application. Detailed Implementation Manner

[0035] In order to further elaborate on the technical means and effects adopted by this application to achieve the predetermined invention purpose, the following, in combination with the accompanying drawings and preferred embodiments, details the specific implementation manner, structure, features, and effects of the product quality inspection method and system applied to the clothing processing production line proposed according to this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0036] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which this application belongs.

[0037] The following specifically describes the specific solutions of the product quality inspection method and system applied to the clothing processing production line provided by this application in combination with the accompanying drawings.

[0038] Please refer to Figure 1 , which shows a flowchart of the steps of the product quality inspection method applied to the clothing processing production line provided by an embodiment of this application. The method includes the following steps:

[0039] Step S1: Obtain the clothing contour map and its template contour map of the clothing to be detected, as well as each preset standard point in the template contour map.

[0040] In this embodiment, the product quality inspection system applied to the clothing processing production line includes a data acquisition module, an image processing module, and a clothing size detection module.

[0041] S1.1 Obtain the clothing contour map and related data of the clothing to be detected.

[0042] In the data acquisition module, when performing size detection on a batch of clothing, in this embodiment, the grayscale image of the surface image of the clothing template used to produce this batch of clothing is obtained. Taking the grayscale image as an example, the contour extraction algorithm is used to extract the contour in the grayscale image, and the obtained contour image is recorded as the template contour map. Multiple pixel points for measuring the clothing template size are manually marked in the template contour map. Usually, all the corner points on the clothing template are selected and recorded as the standard points of the template contour map A. Among them, the standard points in the template contour map can be selected by technicians according to the type of clothing, and all the size data of the clothing template are obtained. The sizes include sleeve length, shoulder width, clothing length, chest width, collar width, collar height, and cuff width, and the size data of the clothing template are transmitted to the clothing size detection module.

[0043] Lay the clothing to be detected in the above batch flat on the detection platform, and place multiple industrial CCD cameras at different positions above the position to be detected to collect the surface images of the clothing to be detected, and transmit all the surface images of the clothing to be detected collected to the data processing module. Among them, the background of the detection platform needs to use a solid color background with a large color difference from the clothing to be detected, such as: white or black background. The number and placement positions of the industrial CCD cameras can be set by the implementer himself. In this embodiment, an industrial CCD camera is placed on each of the left and right sides directly above the detection platform. The installation schematic diagram of the industrial CCD camera is as Figure 2 shown.

[0044] Among them, the contour extraction algorithm is a well-known technology, and its specific principle will not be elaborated here.

[0045] S1.2 Preprocess the obtained surface images.

[0046] Since the lighting conditions in the environment usually change over time, which may cause shadows or reflections in the collected surface images of the clothing, affecting the quality of the images. Therefore, in this embodiment, images of the surface of the clothing to be detected are collected at different angles, and the collected images are fused to reduce the impact of the lighting conditions in the environment on the quality of the subsequent collected surface images of the clothing to be detected. This is because when the lighting at a certain angle is weak and the quality of the collected surface image of the clothing is poor, images with better lighting at other angles can be used to balance the lighting intensity in the collected surface image of the clothing, thereby improving the quality of the collected surface image of the clothing.

[0047] Based on the above analysis, in the data processing module of this embodiment, each surface image of the clothing to be detected collected is respectively converted into a grayscale image, and the obtained grayscale image of each is respectively denoised by using an image denoising algorithm to reduce the impact of the noise received during the collection and transmission of the surface image on the subsequent image processing, and all the surface images of the clothing to be detected after denoising processing are obtained.

[0048] It should be noted that there are many commonly used image denoising algorithms. In this embodiment, the Gaussian filtering algorithm is used to denoise the image. In the actual application process, as other implementation manners, the implementer can also use the median filtering algorithm or the wavelet threshold denoising algorithm to denoise the image. Regarding the selection of the image denoising algorithm, no special restrictions are made in this embodiment.

[0049] Among them, both the Gaussian filtering algorithm and the process of converting the color image collected by the industrial camera into a grayscale image are well-known technologies, and the specific denoising process of the Gaussian filtering algorithm and the specific conversion principle of the grayscale image will not be elaborated here.

[0050] Use the correction algorithm based on contour extraction to correct each denoised surface image respectively, so as to align the objects in all surface images spatially, facilitating subsequent fusion processing of the image information at the same position in all surface images using the image fusion method. Then, use the image fusion method to fuse all the corrected surface images. Thus, the preprocessing of the obtained surface images is completed, and the surface images of the to-be-detected clothing after preprocessing are obtained. The contour extraction algorithm is used to obtain the contour in this surface image, and the obtained contour image is denoted as the clothing contour map C of the to-be-detected clothing.

[0051] It should be added that there are many common image fusion methods. In this embodiment, the Laplacian pyramid fusion method is used to fuse the surface images taken from different angles of the to-be-detected clothing. In the actual application process, as other implementation manners, the implementer can also use image fusion methods such as the linear weighted average fusion method or the Poisson fusion method. Regarding the selection of the image fusion method, no special limitation is made in this embodiment.

[0052] Among them, both the Laplacian pyramid fusion method and the correction algorithm based on contour extraction are well-known technologies, and their specific principles will not be elaborated here.

[0053] Step S2: Perform a contrast mapping on the template contour map and the clothing contour map to obtain the target points for dimension measurement in the clothing contour map corresponding to the standard points in the template contour map.

[0054] During the process of clothing processing and production, usually batch production of clothing is carried out according to a certain clothing template, and the shape and size of the produced clothing usually do not have a large deviation from the shape and size of the clothing template. As a result, there is no large difference between the positions of the clothing standard points for measuring its size in the surface image of the to-be-detected clothing and the positions of the clothing standard points in the surface image of the clothing template. Moreover, the clothing standard points for measuring its size in the surface image of the clothing are usually only distributed on the edge contour of the clothing in its surface image. Therefore, in this embodiment, by obtaining the local areas where each standard point in the template contour map of the clothing template is located in the clothing contour map of the to-be-detected clothing, the screening of the clothing standard points in the clothing contour map of the to-be-detected clothing is accelerated subsequently, and all the clothing standard points in the clothing contour map of the to-be-detected clothing are accurately screened based on the local contour features of all the contour pixel points in the local area.

[0055] S2.1 Rotate the central axis of the template contour diagram to be parallel to the central axis of the clothing contour diagram, and denote the rotated template contour diagram as the corrected diagram; select the pixel point with the smallest ordinate on the central axis of the clothing contour diagram and the pixel point with the smallest ordinate on the central axis of the corrected diagram as a reference point pair. Based on the coordinates of the reference point pair, map the coordinates of all pixel points in the corrected diagram to the clothing contour diagram, and denote the pixel points corresponding to all pixel points in the corrected diagram in the clothing contour diagram as comparison points.

[0056] Since the clothing contour generally exhibits symmetry, the central axes of the template contour diagram and the clothing contour diagram are respectively obtained, and the coordinates of all pixel points on the central axis of the template contour diagram and the coordinates of all pixel points on the central axis of the clothing contour diagram are respectively fitted to obtain two fitted straight lines. Map the two fitted straight lines to the same spatial coordinate system, and use the image rotation algorithm to rotate the template contour diagram. Take the included angle between the two fitted straight lines as the rotation angle in the image rotation algorithm to obtain the rotated template contour diagram, and denote the rotated template contour diagram as the corrected diagram; make the central axis of the template contour diagram parallel to the central axis of the clothing contour diagram to facilitate obtaining the pixel points at the corresponding positions of the standard points in the template contour diagram in the clothing contour diagram.

[0057] It should be noted that there are many commonly used fitting methods. In this embodiment, the least squares fitting method is used to fit the coordinates of pixel points. In the actual application process, the implementer can also use other fitting methods such as linear regression. Regarding the selection of fitting methods, this embodiment does not make special restrictions.

[0058] Among them, the least squares fitting method and the image rotation algorithm are both well-known technologies, and their specific principles will not be elaborated here.

[0059] Furthermore, in this embodiment, the pixel point with the smallest ordinate on the central axis of the clothing contour diagram and the pixel point with the smallest ordinate on the central axis of the template contour diagram are selected to form a reference point pair. Assume the coordinates of the reference point pair are (x a1 , y a1 ), (x c , y c ), corresponding to the corrected diagram and the clothing contour diagram respectively. Furthermore, map the coordinates of all pixel points in the corrected diagram to the clothing contour diagram. Specifically: the coordinate (x, y) of any pixel point in the corrected diagram in the clothing contour diagram is (x + (x c - x a1 ), y + (y c - y a1 ))). Traverse all pixel points in the corrected diagram, map the coordinates of all pixel points in the corrected diagram to the clothing contour diagram, and denote the pixel points corresponding to all pixel points in the corrected diagram in the clothing contour diagram as comparison points.

[0060] So far, by performing calibration mapping on the template contour map, the corresponding points of all pixel points in the calibration map in the clothing contour map have been obtained.

[0061] S2.2 Divide the calibration map and the clothing contour map into multiple image blocks in the same way, analyze the differences in the abscissas between all pixel points with the same ordinate and the differences in the ordinates between all pixel points with the same abscissa between the image block where each standard point is located and the image block where its corresponding point is located, and determine the position deviation values of each standard point and its corresponding point.

[0062] During the processing and production of clothing, due to the precision of machine cutting and sewing, there will be a certain error between the size of the produced clothing and the size of the template clothing, which will further lead to a certain position deviation between the positions of the clothing standard points used to measure its size in the surface image of the clothing to be detected and the clothing standard points at the corresponding positions in the surface image of the clothing template. Moreover, this position deviation generated at different positions of different produced clothing is usually different, which makes the size of the local area where the clothing standard points in the surface contour image of the clothing template are located in the surface contour image of the clothing to be detected affected by the position deviation. This is because if the local area is too small, it may cause the local area to miss clothing standard points and misidentify wrong pixel points as clothing standard points; if the local area is too large, it will increase the time for subsequently screening out clothing standard points from the local area. Therefore, it is necessary to select a suitable size for the local area.

[0063] First, divide the calibration map and the clothing contour map into W image blocks in the same way. In this embodiment, the value of W is 16. Implementers can also set the division quantity according to specific situations by themselves, and this embodiment does not make special restrictions.

[0064] Furthermore, respectively take the average value of the differences in the abscissas between all pixel points with the same ordinate and the average value of the differences in the ordinates between all pixel points with the same abscissa between the image block where each standard point is located and the image block where its corresponding point is located, and denote them as the first average value and the second average value. Take the maximum value of the first average value and the second average value as the position deviation value of each standard point and its corresponding point, which is used to evaluate the offset distance between the positions of the standard points and the corresponding target points in the clothing contour map.

[0065] It should be noted that there are many methods to measure the differences between data. In this embodiment, the method of taking the absolute value of the difference is used as the calculation method of the difference. In the actual application process, implementers can also use other methods to measure the differences between data, such as the square or ratio of the difference. Regarding the selection of the method to measure the differences between data, this embodiment does not make special restrictions.

[0066] It should be noted that in this embodiment, for all content related to measuring the difference between data, the method of taking the absolute value of the difference is adopted.

[0067] S2.3: Based on the position deviation value, determine the local window of each reference point's corresponding point. Based on the difference between the pixel values of all pixel points in the neighborhood between each contour pixel point in the local window and the corresponding reference point, determine the feature value of each contour pixel point, so as to obtain the target points for size measurement of each reference point in the clothing contour map.

[0068] For the local feature values of general pixel points, such as Scale-Invariant Feature Transform (SIFT) feature descriptors and Speeded-Up Robust Features (SURF) feature descriptors, although they can well evaluate the local shape of pixel points, these local feature values usually not only have relatively complex calculations, but also the obtained local feature values are usually a high-dimensional vector, making it easy to increase the time for subsequently screening out target points from the local window when directly using the existing local feature values to evaluate the local shape of contour pixel points in the local window of pixel points. Therefore, to shorten the screening time, the following processing is carried out, specifically:

[0069] (1) Based on the position deviation value, determine the local window of each reference point's corresponding point, specifically: the result of adding the rounded-up value of the position deviation value of each reference point's corresponding point to a preset value is used as the length of the local window of each reference point's corresponding point, and a square window constructed with the corresponding point of each reference point as the center and the local window length as the window side length is used as the local window of each reference point's corresponding point.

[0070] It should be noted that the value of the preset value is set manually. In this embodiment, the value of the preset value is 1, and the implementer can also set it according to the specific situation. This embodiment does not make special restrictions.

[0071] (2) Further, based on the difference between the pixel values of all pixel points in the neighborhood between each contour pixel point in the local window and the corresponding reference point, determine the feature value of each contour pixel point, specifically:

[0072] Since the feature points for measuring the size of clothing are generally located at the corners of the clothing contour, such as the corners of the clothes and the cuffs, there is a large distribution difference between the feature points located at the corners of the clothing contour and the remaining contour pixel points in the corner area. Therefore, based on the above analysis, determine the feature value of the contour pixel point, specifically:

[0073] Within the local window of the corresponding point of the standard point i, the pixel values of all the contour pixel points in the neighborhood of each contour pixel point and in the neighborhood of the standard point i are assigned the value 1, and the pixel values of the remaining pixel points are assigned the value 0. The difference between the re-assignment results of the pixel values of all the pixel points between each contour pixel point and the neighborhood of the standard point i is used as the feature value of each contour pixel point within the local window of the corresponding point of the standard point i, which is used to evaluate the contour difference between the local area where the contour pixel point is located and the local area where the standard point is located. The larger the feature value, the greater the contour difference, and the less likely the contour pixel point is to be the pixel point corresponding to the standard point in the clothing contour map.

[0074] It should be noted that there are many methods to measure the difference between the re-assignment results of all the pixel points between the neighborhoods. In this embodiment, the Euclidean distance between the re-assignment results of the pixel values of all the pixel points between each contour pixel point and the neighborhood of the standard point i is used as the difference between the re-assignment results of the pixel values of all the pixel points between each contour pixel point and the neighborhood of the standard point i. In the actual application process, as other implementation manners, the implementer can also adopt other methods such as the DTW distance, and this embodiment does not make special restrictions.

[0075] Among them, the calculation method of the Euclidean distance is a well-known technology, and its specific calculation process will not be elaborated here.

[0076] It should be added that in this embodiment, the method for dividing the neighborhood of each contour pixel point within the local window of the corresponding point of each standard point is: a window of n×n is divided with each contour pixel point as the center. In this embodiment, the value of n is 3, and the implementer can also set it by combining the specific situation, and this embodiment does not make special restrictions.

[0077] It should be understood that in this embodiment, since both the collected clothing contour map and the template contour map are grayscale images, therefore, the pixel value of the pixel point is the grayscale value of the pixel point.

[0078] (3) Further, the contour pixel point corresponding to the minimum feature value within the local window of the corresponding point of each standard point is used as the target point for size measurement of each standard point in the clothing contour map.

[0079] Preferably, the schematic diagram of the process for obtaining the target point for size measurement in the clothing contour map provided by this embodiment is as Figure 2 shown.

[0080] Step S3: Based on the coordinates of the target points corresponding to all the standard points in the clothing contour map, the size data of the clothing to be detected is obtained, and the quality of the clothing to be detected is inspected.

[0081] In the clothing size detection module, a method of calibrating the position coordinates of target points in the clothing contour map is adopted. By calculating the straight-line distances between all target points, the corresponding geometric sizes are calculated to obtain all the size data of the clothing to be detected. Among them, all the size data include: sleeve length, shoulder width, clothing length, chest width, collar width, collar height, and cuff width.

[0082] Based on all the size data of the clothing template used for the clothing to be detected obtained in step S1, calculate the differences in the corresponding sizes between the clothing to be detected and the clothing template, and take the average value of all the differences in the corresponding sizes as the size deviation of the clothing to be detected.

[0083] In this embodiment, the specific method for dividing the quality grade according to the size deviation is as follows: If the size deviation is less than 1 cm, the quality grade of the clothing to be detected is set to level 1. If the size deviation is between 1 cm and 2 cm, the quality grade of the clothing to be detected is set to level 2. If the size deviation is greater than 2 cm, the quality grade of the clothing to be detected is set to level 3. The larger the level, the worse the quality.

[0084] Implementers can also set the quality grade judgment criteria and judgment methods according to specific circumstances by themselves, and this embodiment does not make special restrictions.

[0085] Among them, calculating the corresponding geometric sizes by calculating the straight-line distances between all corner points is a well-known technology, and its specific principle process will not be elaborated here.

[0086] Based on the same inventive concept as the above method, the embodiment of the present application also provides a product quality inspection system applied to a clothing processing production line, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above product quality inspection methods applied to the clothing processing production line.

[0087] It should be noted that: The above sequence of the embodiments of the present application is only for description and does not represent the superiority or inferiority of the embodiments. And the above has described specific embodiments of this specification. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0088] Each embodiment in this specification is described in a progressive manner. The same or similar parts between each embodiment can be referred to each other, and the key points of each embodiment are the differences from other embodiments.

[0089] The foregoing are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present application shall be included within the protection scope of the present application.

Claims

1. A product quality inspection method applied to a clothing processing production line, characterized in that, The method includes the following steps: Obtain the clothing contour map and its template contour map of the clothing to be detected, as well as each preset standard point in the template contour map; Rotate the central axis of the template contour map to be parallel to the central axis of the clothing contour map, and denote the rotated template contour map as the corrected map; select the pixel point with the minimum ordinate on the central axis of the clothing contour map and the pixel point with the minimum ordinate on the central axis of the corrected map as a reference point pair, and based on the coordinates of the reference point pair, map the coordinates of all pixel points in the corrected map to the clothing contour map, and denote the pixel points corresponding to the mapped pixel points of all pixel points in the corrected map in the clothing contour map as comparison points; Divide the corrected map and the clothing contour map into multiple image blocks in the same way, analyze the differences in the abscissas between all pixel points with the same ordinate and the differences in the ordinates between all pixel points with the same abscissa between the image block where each standard point is located and the image block where its comparison point is located, and determine the position deviation values of each standard point and its comparison point; Based on the position deviation values, determine the local windows of each standard point and its comparison point, and based on the differences in the pixel values of all pixel points in the neighborhood between each contour pixel point in the local window and the corresponding standard point, determine the feature value of each contour pixel point, so as to obtain the target points for size measurement of each standard point in the clothing contour map; Based on the coordinates of the target points corresponding to all standard points in the clothing contour map, obtain the size data of the clothing to be detected and conduct quality inspection on the quality of the clothing to be detected.

2. The product quality inspection method applied to a clothing processing production line according to claim 1, wherein, The rotation of the central axis of the template contour map to be parallel to the central axis of the clothing contour map includes: Perform fitting on the coordinates of all pixel points on the central axis of the template contour map and the coordinates of all pixel points on the central axis of the clothing contour map respectively to obtain two fitted straight lines, map the two fitted straight lines to the same spatial coordinate system, use an image rotation algorithm to rotate the template contour map, and use the included angle between the two fitted straight lines as the rotation angle in the image rotation algorithm to obtain the rotated template contour map.

3. The product quality inspection method applied to the clothing processing production line according to claim 1, characterized in that, The mapping of the coordinates of all pixel points in the corrected map to the clothing contour map includes: Suppose the coordinates of the reference point pairs correspond to (x a1 , y a1 ) in the calibration image and (x c , y c ) in the clothing contour image respectively. Then the coordinates of any pixel point (x, y) in the calibration image in the clothing contour image are (x + (x c - x a1 ), y + (y c - y a1 ))。 4. The product quality inspection method applied to a clothing processing production line as described in claim 1, wherein, The method for determining the position deviation value of each standard point and its comparison point is: Respectively take the mean of the differences in the abscissas between all pixel points with the same ordinate and the mean of the differences in the ordinates between all pixel points with the same abscissa between the image block where each standard point is located and the image block where its comparison point is located, and denote them as the first mean and the second mean, and take the maximum value of the first mean and the second mean as the position deviation value of each standard point and its comparison point.

5. The product quality inspection method applied to a clothing processing production line according to claim 1, characterized in that, The method for determining the local window of each standard point and its comparison point is: Take the result of rounding up the position deviation value of each standard point and its comparison point and adding a preset value as the length of the local window of each standard point and its comparison point, and construct a square window with the length of the local window as the window side length centered on each standard point and its comparison point as the local window of each standard point and its comparison point.

6. The product quality inspection method applied to the clothing processing production line according to claim 1, characterized in that, The method for determining the feature value of each contour pixel point is: Within the local window of the reference point at the standard point i, the pixel values of all the contour pixel points in the neighborhood of each contour pixel point and in the neighborhood of the standard point i are assigned the value 1, and the pixel values of the remaining pixel points are assigned the value 0. The difference between the re-assigned results of the pixel values of all the pixel points between each contour pixel point and the neighborhood of the standard point i is used as the feature value of each contour pixel point within the local window of the reference point of the standard point i.

7. The product quality inspection method applied to a clothing processing production line according to claim 1, characterized in that, The target points for dimension measurement in the clothing contour map of the respective standard points are the contour pixel points corresponding to the minimum feature values within the local windows of the reference points of the respective standard points.

8. The product quality inspection method applied to a clothing processing production line according to claim 1, characterized in that, The process of obtaining the dimension data of the clothing to be detected is as follows: By using the method of calibrating the position coordinates of the target points in the clothing contour map, the corresponding geometric dimensions are calculated by computing the straight-line distances between all the target points, and all the dimension data of the clothing to be detected are obtained. Among them, all the dimension data include: sleeve length, shoulder width, body length, chest width, collar width, collar height, and cuff width.

9. The product quality inspection method applied to a clothing processing production line according to claim 1, wherein, The quality inspection of the clothing to be detected includes: Obtaining all the dimension data of the clothing template used for the clothing to be detected, calculating the differences in the corresponding dimensions between the clothing to be detected and the clothing template, and taking the mean of the differences in all the corresponding dimensions as the dimension deviation of the clothing to be detected. According to the dimension deviation of the clothing to be detected, the quality of the clothing to be detected is classified into grades.

10. A product quality inspection system applied to a clothing processing production line, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the product quality inspection method applied to the clothing processing production line as described in any one of claims 1-9.

Images