Zipper puller plug visual identification method of full-automatic zipper threading machine

Abstract

The present application relates to the technical field of image recognition, and particularly relates to a zipper puller blockage visual recognition method of a full-automatic zipper puller threading machine, which comprises: acquiring a guide puller image and a puller region thereof; acquiring a matching coefficient of a sliding window in adjacent two puller regions according to the structure and the gray scale distribution of the sliding window in the sliding window; acquiring a matching window of the sliding window according to the matching coefficient and the structural features of the two sliding windows; analyzing the consistency of the positional relationship of all sliding windows in the puller region and the positional relationship of other adjacent puller regions in a constructed structure matching space to acquire a blockage coefficient of the puller region; and identifying the blockage of the zipper puller threading machine by using the blockage coefficient. The present application aims to improve the accuracy of the puller distance calculation by dividing the puller into regions and eliminating the influence of non-fixed structures in the region matching process, and then judging the puller blockage condition by using the distance between the matching structures.

Description

Technical Field

[0001] This invention relates to the field of image recognition technology, and specifically to a visual recognition method for zipper pull blockage in a fully automatic zipper threading machine. Background Technology

[0002] The fully automatic zipper threading machine is a core piece of automated equipment in the zipper processing stage. With its multi-material compatibility and end-to-end integration capabilities, it is widely used in clothing, bags, home textiles, and special equipment industries. Its proper operation is paramount to its production. In the production process of the fully automatic zipper threading machine, zipper head blockage is a major efficiency bottleneck, often occurring in the vibratory feeder track, linear conveyor channel, and gripping station. Therefore, to minimize the impact of zipper head blockage on production efficiency, a precise zipper head blockage detection method is essential.

[0003] For zipper head blockage detection, some existing technologies introduce photoelectric sensors for preliminary monitoring. However, these devices can only identify the presence or absence of material and cannot distinguish between normal arrangement of zipper heads and stacked blockage. Other technologies use visual detection, but traditional visual detection often directly uses the edges between the zipper head areas to calculate the distance between the zipper heads. Since some metal zipper heads are reflective, the reflection can obscure some edges of the zipper heads, resulting in the distance between the edges of the zipper heads not accurately reflecting the actual distance between the zipper heads. Summary of the Invention

[0004] This invention provides a visual recognition method for zipper pull blockage in a fully automatic zipper threading machine, in order to solve the existing problems.

[0005] The present invention provides a visual recognition method for zipper pull blockage in a fully automatic zipper threading machine, which adopts the following technical solution:

[0006] One embodiment of the present invention provides a visual recognition method for zipper pull blockage in a fully automatic zipper threading machine, the method comprising the following steps:

[0007] Acquire images of the sliders on the guide rail and construct several slider regions according to the guide rail direction;

[0008] For the i-th zipper head region on the guide rail and the j-th zipper head region adjacent to the i-th zipper head region, based on the similarity of the structure and grayscale distribution of each sliding window in the i-th zipper head region and each sliding window in the j-th zipper head region, obtain the matching coefficient between each sliding window in the i-th zipper head region and each sliding window in the j-th zipper head region;

[0009] Based on the matching coefficients of two sliding windows belonging to the i-th and j-th zipper regions respectively, and verifying the information content and consistency of the structural features of the two sliding windows, obtain the matching window of each sliding window in the i-th zipper region in the j-th zipper region;

[0010] A structural matching space is constructed based on the positional relationship between each sliding window in the i-th zipper head region and its matching window in the j-th zipper head region in the zipper head image. In the structural matching space, the consistency between the positional relationship of all sliding windows in each zipper head region and the corresponding matching window in the adjacent zipper head region and the positional relationship of other adjacent zipper head regions is analyzed to obtain the blockage coefficient of each zipper head region. The blockage coefficient is used to identify the blockage of the zipper threading machine.

[0011] Preferably, the step of obtaining the matching coefficient between each sliding window in the i-th slider region and each sliding window in the j-th slider region based on the similarity of their structure and grayscale distribution includes:

[0012] Construct a sliding window; obtain the structural similarity index between the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region;

[0013] Obtain the absolute value of the difference in grayscale values ​​between the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region, and record the mean of the absolute values ​​of the grayscale differences as the grayscale similarity between the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region.

[0014] Obtain the matching coefficient between each sliding window in the i-th zipper region and each sliding window in the j-th zipper region. The matching coefficient is positively proportional to the structural similarity index and inversely proportional to the grayscale similarity.

[0015] Preferably, obtaining the matching coefficient between each sliding window in the i-th zipper region and each sliding window in the j-th zipper region includes:

[0016] The matching coefficient between the m-th sliding window in the i-th slider region and the n-th sliding window in the j-th slider region. The calculation method is as follows:

[0017]

[0018] in, For the m-th sliding window in the i-th pull head region, For the j-th slider region, the n-th sliding window, Let be the structural similarity index between the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region; The grayscale similarity between the m-th sliding window in the i-th slider region and the n-th sliding window in the j-th slider region; It is an exponential function with the natural constant as its base.

[0019] Preferably, the specific steps for obtaining the matching window of each sliding window in the i-th slider region in the j-th slider region include:

[0020] Perform edge detection on the m-th sliding window in the i-th zipper region to obtain the number of edge pixels and the number of edges in the m-th sliding window in the i-th zipper region;

[0021] The structural feature value of the m-th sliding window in the i-th pull head region is obtained based on the number of edge pixels and the number of edges.

[0022] Using the structural feature values ​​of the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region, the matching coefficients of the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region are corrected to obtain the matching correction coefficients of the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region.

[0023] The matching correction coefficient is used to obtain the matching window of each sliding window in the i-th slider region in the j-th slider region.

[0024] Preferably, the step of using the structural feature value of the m-th sliding window in the i-th sliding head region and the structural feature value of the n-th sliding window in the j-th sliding head region to correct the matching coefficient between the m-th sliding window in the i-th sliding head region and the n-th sliding window in the j-th sliding head region, and obtaining the matching correction coefficient between the m-th sliding window in the i-th sliding head region and the n-th sliding window in the j-th sliding head region includes:

[0025] The difference between the structural feature value of the m-th sliding window in the i-th zipper region and the structural feature values ​​of the sliding windows of all pixels in the i-th zipper region is denoted as the standard structural feature value of the m-th sliding window in the i-th zipper region. ;

[0026] The difference between the structural feature value of the nth sliding window in the j-th zipper region and the structural feature values ​​of the sliding windows of all pixels in the j-th zipper region is denoted as the standard value of the structural feature of the nth sliding window in the j-th zipper region. ;

[0027] Matching correction coefficient between the m-th sliding window in the i-th slider region and the n-th sliding window in the j-th slider region The calculation method is as follows:

[0028]

[0029] in, Let be the matching coefficient between the m-th sliding window in the i-th slider region and the n-th sliding window in the j-th slider region.

[0030] Preferably, obtaining the matching window of each sliding window in the i-th slider region in the j-th slider region using the matching correction coefficient includes:

[0031] In the sliding window of all pixels in the j-th zipper region, the sliding window of the pixel with the largest matching correction coefficient with the m-th sliding window in the i-th zipper region is denoted as the matching window of each sliding window in the i-th zipper region in the j-th zipper region.

[0032] Preferably, the construction of the structural matching space based on the positional relationship between each sliding window in the i-th zipper region and its matching window in the j-th zipper region in the zipper image includes:

[0033] Construct a Cartesian coordinate system with any pixel in the zipper image as the origin, where the horizontal axis represents the row direction of the pixels in the zipper image and the vertical axis represents the column direction of the pixels in the zipper image.

[0034] For the i-th slider region on the guide rail and the j-th slider region adjacent to the i-th slider region, connect the center pixel of each sliding window in the i-th slider region and its matching window in the j-th slider region in the slider image. Record the connection obtained as the connection line of each sliding window in the i-th slider region, and obtain the slope of the connection line in the Cartesian coordinate system.

[0035] Construct a structural matching space, where the horizontal axis of the structural matching space is the length of the connecting line, and the vertical axis is the slope of the connecting line in the Cartesian coordinate system.

[0036] Preferably, the step of analyzing the positional relationship between all sliding windows in each zipper head region and the corresponding matching windows in adjacent zipper head regions in the structural matching space, and the consistency of the positional relationship between these windows and other adjacent zipper head regions, to obtain the blockage coefficient of each zipper head region, and using the blockage coefficient to identify blockages in the zipper threading machine includes:

[0037] Project the line connecting all pixels in the i-th zipper region and its slope onto the structure matching space to obtain several discrete points. Use the K-means clustering algorithm with K=1 to cluster all discrete points, and denote the obtained cluster as the matchable window cluster of the i-th zipper region.

[0038] Based on the distance between the discrete point of each sliding window in the structural matching space in the i-th zipper region and the cluster of matchable windows, and combined with the matching correction coefficient of each sliding window in the i-th zipper region and its matching window in the j-th zipper region, the elimination coefficient of each sliding window in the i-th zipper region is obtained.

[0039] Based on the elimination coefficient, all sliding windows of the i-th zipper area are filtered to obtain the distance judgment window of the i-th zipper area;

[0040] Based on the difference in length between the i-th zipper head area and the lengths of the lines connecting all distance judgment windows in other zipper head areas, the blockage coefficient of each zipper head area is obtained, and the blockage coefficient is used to identify the blockage of the zipper threading machine.

[0041] Preferably, the specific steps for obtaining the elimination coefficient include:

[0042] The Euclidean distance between the discrete point of each sliding window in the i-th zipper region and the cluster center of the matchable window cluster in the i-th zipper region is denoted as the distance deviation of each sliding window in the i-th zipper region.

[0043] The elimination coefficient of the m-th sliding window in the i-th slider region. The calculation method is as follows:

[0044]

[0045] in, Let m be the distance deviation of the m-th sliding window in the i-th slider region. Let be the matching correction coefficient between the m-th sliding window in the i-th slider region and its matching window in the j-th slider region; It is an exponential function with the natural constant as its base.

[0046] Preferably, the step of obtaining the blockage coefficient of each zipper pull area based on the difference in the length of the line connecting all distance judgment windows in the i-th zipper pull area and other zipper pull areas, and using the blockage coefficient to identify the blockage of the zipper threading machine includes:

[0047] The average length of the line connecting all distance judgment windows in the i-th zipper area is denoted as the zipper distance between the i-th zipper area and its adjacent zipper areas. ;

[0048] Obtain the distance between each slider region and its adjacent slider regions in the slider head image;

[0049] The congestion coefficient of the i-th pull-head area The calculation method is as follows:

[0050]

[0051] in, This is the average distance between all slider regions and their adjacent slider regions in the slider image. It is an absolute value function;

[0052] A preset congestion judgment coefficient is set; if the congestion coefficient of the i-th pull-head area is... When the value is greater than or equal to 0.8, the i-th zipper head area is marked as a blocked area. The fully automatic zipper threading machine will issue a blockage alarm and notify the operator to stop the machine for inspection, guiding the staff to carry out inspection and maintenance.

[0053] The beneficial effects of the technical solution of this invention are as follows: This application acquires the image of the zipper pull on the guide rail and constructs several zipper pull regions according to the direction of the guide rail; based on the similarity of the structure and grayscale distribution of each sliding window in the i-th zipper pull region and each sliding window in the j-th zipper pull region, the matching coefficient of each sliding window in the i-th zipper pull region and each sliding window in the j-th zipper pull region is obtained; through structural and grayscale analysis of different sliding windows in adjacent zipper pull regions, the similarity of the small structures belonging to the zipper pull between the two windows is reflected; based on the matching coefficient of two sliding windows belonging to the i-th zipper pull region and the j-th zipper pull region respectively, and verifying the information content and consistency of the structural features of the two sliding windows, the matching window of each sliding window in the i-th zipper pull region in the j-th zipper pull region is obtained; by analyzing the information content of the structural features of the two sliding windows, the influence of non-zipper structure sliding windows on matching is screened, thereby balancing the non-zipper structure sliding windows during sliding window matching. The application constructs a structural matching space based on the positional relationship between each sliding window in the i-th zipper head region and its matching window in the j-th zipper head region in the zipper head image. This application constructs a structural matching space by representing the positional relationship between two matching sliding windows in different zipper head regions. The structural matching space analyzes the consistency between the positional relationship of all sliding windows in each zipper head region and the corresponding matching window in adjacent zipper head regions, as well as the positional relationship with other adjacent zipper head regions, to obtain the blockage coefficient of each zipper head region. The blockage coefficient is then used to identify the blockage of the zipper head machine. By filtering out two matching sliding windows with inconsistent positional relationships in the structural matching space, and retaining only the sliding windows with consistent positional relationships, the distance between two zipper head regions is calculated using the consistent sliding windows. The blockage of the zipper head machine is then identified based on the distance between the zipper head regions, thus improving the accuracy of the zipper head distance calculation. Attached Figure Description

[0054] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0055] Figure 1 This is a flowchart illustrating the steps of a visual recognition method for zipper pull blockage in a fully automatic zipper threading machine according to the present invention. Detailed Implementation

[0056] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a visual recognition method for zipper pull blockage in a fully automatic zipper threading machine according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0057] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0058] The following description, in conjunction with the accompanying drawings, details a specific scheme for a visual recognition method for zipper pull blockage in a fully automatic zipper threading machine provided by the present invention.

[0059] Please see Figure 1 The diagram illustrates a flowchart of a visual recognition method for zipper pull blockage in a fully automatic zipper threading machine according to an embodiment of the present invention. The method includes the following steps:

[0060] Step S001: Obtain the image of the slider on the guide rail and construct several slider areas according to the direction of the guide rail.

[0061] It should be noted that the fully automatic zipper threading machine can thread the zipper head onto the zipper tape at fixed intervals, and then use it to cut the zipper to produce zippers of specific specifications. It fixes the two zipper tapes with two guide rails respectively. The zipper head is clamped by the zipper head fixing seat, ensuring that the two entrances of the zipper head are aligned with the two zipper tapes. When threading the zipper head, the insert of the threading machine first inserts into the rear opening of the zipper head and passes through to the front opening to open the zipper head. After the guide rails pass the fixed zipper tape through the two entrances of the zipper head, the insert is pulled out. At this time, the internal channel of the opened zipper head will return to its original shape due to the elasticity of the metal, thus tightly engaging the zipper tape and completing the threading of the zipper head.

[0062] However, in practice, the teeth of the two zipper tapes may not mesh perfectly, causing the zipper pull to jam, or the zipper pull may get stuck with the zipper tape when it regains its elasticity, resulting in the zipper pull getting stuck in the zipper guide rail and causing blockage. Therefore, it is necessary to identify blockages in the fully automatic zipper pull machine. Therefore, it is first necessary to acquire an image of the zipper pull on the guide rail and construct the zipper pull area according to the guide rail direction, so as to identify the position of the zipper pull for blockage analysis.

[0063] Preferably, the specific steps for acquiring the slider image on the guide rail and constructing several slider regions according to the guide rail direction are as follows:

[0064] An image sensor is installed above the plane where the two guide rails of the zipper head are located to capture images of the zipper head on the guide rails;

[0065] The image of the zipper pull on the guide rail is acquired using an image sensor; the zipper pull image is a grayscale image.

[0066] The Otsu thresholding algorithm is used to perform threshold segmentation on the zipper image, and each highlighted area after threshold segmentation is used as the initial area of ​​the zipper.

[0067] Based on the direction of the guide rail, construct the minimum outer rectangle of the initial region of each slider in the slider image to obtain several slider regions.

[0068] It should be noted that the image sensor described in this embodiment is installed directly above the guide rail and 60cm away from the guide rail. The parameters of the installed image sensor directly affect the shooting quality. In this embodiment, the accuracy needs to be above 0.05mm. Therefore, the resolution of the image sensor is 1600*1200 pixels. The shooting interval between two adjacent images of the pull head is 0.5 seconds in this embodiment.

[0069] Step S002: For the i-th zipper head region on the guide rail and the j-th zipper head region adjacent to the i-th zipper head region, based on the similarity of the structure and grayscale distribution of each sliding window in the i-th zipper head region and each sliding window in the j-th zipper head region, obtain the matching coefficient between each sliding window in the i-th zipper head region and each sliding window in the j-th zipper head region.

[0070] It should be noted that in an automatic zipper threading machine, the zipper pull needs to be conveyed to the threading position via a linear guide rail running at a certain speed to achieve zipper threading. The running speed of the guide rail is consistent with the threading speed, so the distance between the zipper pulls in the guide rail remains consistent. However, when zipper pull blockage occurs, the zipper pull at the threading position is not removed from the guide rail, causing the distance between the zipper pulls in the subsequent guide rails to decrease. The captured image of the zipper pulls includes all the zipper pulls in the guide rail. At this time, the distance between the zipper pull areas in the image can reflect whether zipper pull blockage has occurred. However, because the metal zipper pulls and the guide rails have similar colors, and the metal zipper pulls are reflective, the edges of the zipper pulls are obscured, making it impossible to directly determine the distance between the edges of the zipper pulls, and therefore impossible to determine the distance between the zipper pulls.

[0071] It should be further explained that the zipper pull has some small structures, such as the zipper body, the pull tab, and the pull tab retaining ring. These small structures form obvious feature structures in the zipper pull area, and all zipper pulls in the same guide rail have the same structure. Therefore, this embodiment calculates the distance between zipper pulls by measuring the distance between the feature structures in the image. Therefore, in each zipper pull area of ​​the zipper pull image, in order to extract feature structures while facilitating distance calculation, this embodiment sets a sliding window in each zipper pull area, using the sliding window as a separate structure extraction unit for extracting feature structures.

[0072] Based on the above, this embodiment constructs several sliding windows in two adjacent slider regions on the guide rail, and then analyzes the similarity of the sliding windows in the two adjacent slider regions, thereby analyzing the matching of the sliding window in each slider region with the sliding window in the adjacent slider region, and obtaining the matching coefficient of each sliding window in each slider region with each sliding window in the adjacent slider region.

[0073] It should be noted that in the fully automatic zipper threading machine, the zipper heads are arranged in a single row. In this embodiment, the opposite direction of the guide rail's forward movement is used as the selection direction for the adjacent zipper head area of ​​each zipper head area. That is, if the guide rail moves to the left, the nearest zipper head area to the right of each zipper head area is taken as the adjacent zipper head area of ​​each zipper head area.

[0074] Preferably, for the i-th zipper pull area on the guide rail and the j-th zipper pull area adjacent to the i-th zipper pull area, the specific steps for obtaining the matching coefficient between each sliding window in the i-th zipper pull area and each sliding window in the j-th zipper pull area based on the similarity in structure and grayscale distribution between each sliding window in the i-th zipper pull area and each sliding window in the j-th zipper pull area are as follows:

[0075] A sliding window is constructed. In this embodiment, the sliding window is a 7*7 directional window centered on each pixel in the pull head area. This embodiment does not limit the shape and side length of the sliding window. Other embodiments may choose other methods.

[0076] Obtain the structural similarity index between the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region;

[0077] Obtain the absolute value of the difference in grayscale values ​​between the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region, and record the mean of the absolute values ​​of the grayscale differences as the grayscale similarity between the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region.

[0078] Obtain the matching coefficient between each sliding window in the i-th zipper region and each sliding window in the j-th zipper region. The matching coefficient is positively proportional to the structural similarity index and inversely proportional to the grayscale similarity.

[0079] As an example, the matching coefficient between the m-th sliding window in the i-th slider region and the n-th sliding window in the j-th slider region. The calculation method is as follows:

[0080]

[0081] in, For the m-th sliding window in the i-th pull head region, For the j-th slider region, the n-th sliding window, Let be the structural similarity index between the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region; The grayscale similarity between the m-th sliding window in the i-th slider region and the n-th sliding window in the j-th slider region; It is an exponential function with the natural constant as its base, used to inversely normalize the gray-level similarity.

[0082] It should be noted that the structural similarity index This is a known existing technique used to evaluate the similarity of two sliding windows based on the mean, standard deviation, and covariance of grayscale values. In this embodiment, a structural similarity index is used to evaluate the similarity of the zipper pull structures in the two sliding windows. The larger the value, the higher the structural similarity. This embodiment uses the difference in grayscale values ​​of corresponding pixels in different zipper pull regions as grayscale similarity, thereby avoiding the problem of excessively large matching coefficients caused by slight deviations in the two zipper pull structures but relatively similar overall structures, and balancing the difference in matching coefficients when matching in sliding windows.

[0083] Step S003: Based on the matching coefficients of the two sliding windows belonging to the i-th and j-th zipper regions respectively, and verifying the information content and consistency of the structural features of the two sliding windows, obtain the matching window of each sliding window in the i-th zipper region in the j-th zipper region.

[0084] It should be noted that after obtaining the matching coefficients of the two windows, since the information content contained in different areas of the zipper pull region is different, the information content represents the small structures in the zipper pull, such as the zipper pull body, the zipper tab, and the zipper tab fixing ring. The planar structure in the zipper pull contains low information content and is prone to mismatch. Therefore, this embodiment analyzes the edge situation of each sliding window in each zipper pull region, uses the edge situation to represent the information content of the small structures in the zipper pull, and then uses it as the weight of the information content to correct the matching coefficient, thereby obtaining the matching window of each sliding window in each zipper pull region in the adjacent zipper pull region.

[0085] Preferably, the specific steps for obtaining the matching window of each sliding window in the i-th zipper region in the j-th zipper region, based on the matching coefficients of two sliding windows belonging to the i-th and j-th zipper regions respectively, and verifying the information content and consistency of the structural features of the two sliding windows, are as follows:

[0086] Perform edge detection on the m-th sliding window in the i-th zipper region to obtain the number of edge pixels and the number of edges in the m-th sliding window in the i-th zipper region;

[0087] It should be noted that this embodiment uses the Canny edge detection algorithm for edge detection. The Canny edge detection algorithm is a well-known existing technology, and will not be described in detail in this embodiment.

[0088] The structural feature value of the m-th sliding window in the i-th pull head region is obtained based on the number of edge pixels and the number of edges.

[0089] Using the structural feature values ​​of the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region, the matching coefficients of the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region are corrected to obtain the matching correction coefficients of the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region.

[0090] The matching correction coefficient is used to obtain the matching window of each sliding window in the i-th slider region in the j-th slider region.

[0091] Specifically, the method for obtaining the structural feature value of the m-th sliding window in the i-th zipper region based on the number of edge pixels and the number of edges is as follows:

[0092] Let the number of edge pixels of the m-th sliding window in the i-th zipper area be denoted as . ;

[0093] Let the number of edges of the m-th sliding window in the i-th zipper area be denoted as . ;

[0094] The structural feature value of the m-th sliding window in the i-th pull head region The calculation method is as follows:

[0095]

[0096] in, As a linear normalization function, this embodiment uses the maximum-minimum normalization algorithm to normalize the structural feature values.

[0097] Furthermore, using the structural feature values ​​of the m-th sliding window in the i-th swivel region and the n-th sliding window in the j-th swivel region, the matching coefficients of the m-th sliding window in the i-th swivel region and the n-th sliding window in the j-th swivel region are corrected. The specific method for obtaining the matching correction coefficients of the m-th sliding window in the i-th swivel region and the n-th sliding window in the j-th swivel region is as follows:

[0098] The difference between the structural feature value of the m-th sliding window in the i-th zipper region and the structural feature values ​​of the sliding windows of all pixels in the i-th zipper region is denoted as the standard structural feature value of the m-th sliding window in the i-th zipper region. ;

[0099] The difference between the structural feature value of the nth sliding window in the j-th zipper region and the structural feature values ​​of the sliding windows of all pixels in the j-th zipper region is denoted as the standard value of the structural feature of the nth sliding window in the j-th zipper region. ;

[0100] Matching correction coefficient between the m-th sliding window in the i-th slider region and the n-th sliding window in the j-th slider region The calculation method is as follows:

[0101]

[0102] in, Let be the matching coefficient between the m-th sliding window in the i-th slider region and the n-th sliding window in the j-th slider region.

[0103] It should be noted that the standard value of the structural feature can be positive or negative. When the value is negative, it indicates that the m-th sliding window contains less information than other sliding windows in its corresponding slider area. Therefore, the correction coefficient is reduced to obtain the matching correction coefficient, thereby avoiding interference from sliding windows without structure in the matching results. When the standard value of the structural feature is positive, it indicates that the m-th sliding window contains more information than other sliding windows in its corresponding slider area, better reflects the structure of the slider, thereby improving the accuracy of matching and making it easier to use as a basis for calculating the distance between two sliders. Therefore, the larger the value of the matching correction coefficient, the higher the value of the matching correction coefficient.

[0104] Furthermore, the specific method for obtaining the matching window of each sliding window in the i-th slider region in the j-th slider region using the matching correction coefficient is as follows:

[0105] In the sliding window of all pixels in the j-th zipper region, the sliding window of the pixel with the largest matching correction coefficient with the m-th sliding window in the i-th zipper region is denoted as the matching window of each sliding window in the i-th zipper region in the j-th zipper region.

[0106] Step S004: Construct a structural matching space based on the positional relationship of each sliding window in the i-th zipper head region and its matching window in the j-th zipper head region in the zipper head image; analyze the consistency between the positional relationship of all sliding windows in each zipper head region and the corresponding matching window in the adjacent zipper head region and the positional relationship of other adjacent zipper head regions in the structural matching space, obtain the blockage coefficient of each zipper head region, and use the blockage coefficient to identify the blockage of the zipper threading machine.

[0107] It should be noted that the above matching of sliding windows with the same structure between the zipper pull areas means that the distance between two matched sliding windows represents the distance of the same structure between the two zipper pulls. However, since the zipper pulls themselves have some similar edges, different sliding windows within the same zipper pull area may have the same structure, which causes deviations in the accuracy of the sliding window matching. Furthermore, since the zipper pull tabs wobble during the movement of the guide rail, the wobble of the pull tabs cannot be kept in the same position as the zipper pulls in the image. In other words, the unstable structure of the zipper pulls themselves means that not all the distances between matched sliding windows can accurately represent the distances between the zipper pulls.

[0108] Based on the above, when calculating the distance between two zipper heads using the distance between the matching sliding windows of each zipper head, it is necessary to avoid the influence of matching deviations and sliding windows that are inconsistent with the zipper head position. In this embodiment, the positional relationship between the sliding windows of all pixels in each zipper head region and the matching windows in adjacent zipper head regions is used. Specifically, by constructing a line connecting each sliding window and the matching window, the length and angle of the line connecting the reliable sliding window and the matching window are similar, thereby eliminating the sliding window. Then, the distance between each zipper head region and its adjacent zipper head regions is calculated, and the blockage coefficient of each zipper head region is obtained by analyzing the consistency of the distance, thereby realizing the identification of blockage in the zipper threading machine.

[0109] Preferably, a structural matching space is constructed based on the positional relationship between each sliding window in the i-th zipper head region and its matching window in the j-th zipper head region in the zipper head image; the consistency between the positional relationship of all sliding windows in each zipper head region and the corresponding matching window in adjacent zipper head regions and the positional relationship of other adjacent zipper head regions is analyzed in the structural matching space to obtain the blockage coefficient of each zipper head region; the specific steps for identifying the blockage of the zipper threading machine using the blockage coefficient are as follows:

[0110] Construct a Cartesian coordinate system with any pixel in the zipper image as the origin, where the horizontal axis represents the row direction of the pixels in the zipper image and the vertical axis represents the column direction of the pixels in the zipper image.

[0111] For the i-th slider region on the guide rail and the j-th slider region adjacent to the i-th slider region, connect the center pixel of each sliding window in the i-th slider region and its matching window in the j-th slider region in the slider image. Record the connection obtained as the connection line of each sliding window in the i-th slider region, and obtain the slope of the connection line in the Cartesian coordinate system.

[0112] Construct a structural matching space, where the horizontal axis of the structural matching space is the length of the connecting line, and the vertical axis is the slope of the connecting line in the Cartesian coordinate system.

[0113] Project the line connecting all pixels in the i-th zipper region and its slope onto the structure matching space to obtain several discrete points. Use the K-means clustering algorithm with K=1 to cluster all discrete points, and denote the obtained cluster as the matchable window cluster of the i-th zipper region.

[0114] Based on the distance between the discrete point of each sliding window in the structural matching space in the i-th zipper region and the cluster of matchable windows, and combined with the matching correction coefficient of each sliding window in the i-th zipper region and its matching window in the j-th zipper region, the elimination coefficient of each sliding window in the i-th zipper region is obtained.

[0115] Based on the elimination coefficient, all sliding windows of the i-th zipper area are filtered to obtain the distance judgment window of the i-th zipper area;

[0116] Based on the difference in length between the i-th zipper head area and the lengths of the lines connecting all distance judgment windows in other zipper head areas, the blockage coefficient of each zipper head area is obtained, and the blockage coefficient is used to identify the blockage of the zipper threading machine.

[0117] Specifically, the method for obtaining the elimination coefficient of each sliding window in the i-th zipper region is as follows: based on the distance between the discrete point of each sliding window in the structural matching space and the matchable window cluster in the i-th zipper region, combined with the matching correction coefficient of each sliding window in the i-th zipper region and its matching window in the j-th zipper region.

[0118] The Euclidean distance between the discrete point of each sliding window in the i-th zipper region and the cluster center of the matchable window cluster in the i-th zipper region is denoted as the distance deviation of each sliding window in the i-th zipper region.

[0119] Obtain the rejection coefficient of each sliding window in the i-th zipper area. The rejection coefficient is inversely proportional to the distance deviation and directly proportional to the matching correction coefficient of each sliding window in the i-th zipper area and its matching window in the j-th zipper area.

[0120] As an example, the elimination factor of the m-th sliding window in the i-th pull head region. The calculation method is as follows:

[0121]

[0122] in, Let m be the distance deviation of the m-th sliding window in the i-th slider region. Let be the matching correction coefficient between the m-th sliding window in the i-th slider region and its matching window in the j-th slider region; It is an exponential function with the natural constant as its base.

[0123] Furthermore, the specific method for obtaining the distance judgment window for the i-th zipper area by filtering all sliding windows of the i-th zipper area according to the elimination coefficient is as follows:

[0124] A preset filtering threshold is set, and in this embodiment, 0.75 is used as an example. The rejection coefficient of the m-th sliding window in the i-th zipper area is normalized by the maximum and minimum values. The sliding window whose normalized rejection coefficient in the i-th zipper area is greater than or equal to the filtering threshold is denoted as the distance judgment window of the i-th zipper area.

[0125] Furthermore, based on the difference in length between the distance judgment windows of the i-th zipper head area and all other zipper head areas, a blockage coefficient for each zipper head area is obtained. The specific method for identifying blockages in the zipper threading machine using the blockage coefficient is as follows:

[0126] The average length of the line connecting all distance judgment windows in the i-th zipper area is denoted as the zipper distance between the i-th zipper area and its adjacent zipper areas. ;

[0127] Obtain the distance between each slider region and its adjacent slider regions in the slider head image;

[0128] The congestion coefficient of the i-th pull-head area The calculation method is as follows:

[0129]

[0130] in, This is the average distance between all slider regions and their adjacent slider regions in the slider image. It is an absolute value function;

[0131] The preset blockage judgment coefficient is 0.8. If the blockage coefficient of the i-th pull head area is... When the value is greater than or equal to 0.8, the i-th zipper head area is marked as a blocked area. The fully automatic zipper threading machine will issue a blockage alarm and notify the operator to stop the machine for inspection, guiding the staff to carry out inspection and maintenance.

[0132] It should be noted that the embodiments used in this example The model only represents negative correlations and constraints. The model output results are in... Within the interval, This is the input to this model; in specific implementations, it can be replaced with other models that have the same purpose. This embodiment is merely an example. The model is used as an example for description, without making specific limitations.

[0133] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A visual recognition method for zipper pull blockage in a fully automatic zipper threading machine, characterized in that, The method includes the following steps: Acquire images of the sliders on the guide rail and construct several slider regions according to the guide rail direction; For the i-th zipper head region on the guide rail and the j-th zipper head region adjacent to the i-th zipper head region, based on the similarity of the structure and grayscale distribution of each sliding window in the i-th zipper head region and each sliding window in the j-th zipper head region, obtain the matching coefficient between each sliding window in the i-th zipper head region and each sliding window in the j-th zipper head region; Based on the matching coefficients of two sliding windows belonging to the i-th and j-th zipper regions respectively, and verifying the information content and consistency of the structural features of the two sliding windows, obtain the matching window of each sliding window in the i-th zipper region in the j-th zipper region; A structural matching space is constructed based on the positional relationship between each sliding window in the i-th zipper head region and its matching window in the j-th zipper head region in the zipper head image. In the structural matching space, the consistency between the positional relationship of all sliding windows in each zipper head region and the corresponding matching window in the adjacent zipper head region and the positional relationship of other adjacent zipper head regions is analyzed to obtain the blockage coefficient of each zipper head region. The blockage coefficient is used to identify the blockage of the zipper threading machine. The construction of the structural matching space based on the positional relationship between each sliding window in the i-th zipper region and its matching window in the j-th zipper region in the zipper image includes: Construct a Cartesian coordinate system with any pixel in the zipper image as the origin, where the horizontal axis represents the row direction of the pixels in the zipper image and the vertical axis represents the column direction of the pixels in the zipper image. For the i-th slider region on the guide rail and the j-th slider region adjacent to the i-th slider region, connect the center pixel of each sliding window in the i-th slider region and its matching window in the j-th slider region in the slider image. Record the connection obtained as the connection line of each sliding window in the i-th slider region, and obtain the slope of the connection line in the Cartesian coordinate system. Construct a structural matching space, where the horizontal axis of the structural matching space is the length of the connecting line, and the vertical axis is the slope of the connecting line in the Cartesian coordinate system.

2. The visual recognition method for zipper pull blockage in a fully automatic zipper threading machine according to claim 1, characterized in that, The step of obtaining the matching coefficient between each sliding window in the i-th slider region and each sliding window in the j-th slider region based on the similarity of their structure and grayscale distribution includes: Construct a sliding window; obtain the structural similarity index between the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region; Obtain the absolute value of the difference in grayscale values ​​between the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region, and record the mean of the absolute values ​​of the grayscale differences as the grayscale similarity between the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region. Obtain the matching coefficient between each sliding window in the i-th zipper region and each sliding window in the j-th zipper region. The matching coefficient is positively proportional to the structural similarity index and inversely proportional to the grayscale similarity.

3. The visual recognition method for zipper pull blockage in a fully automatic zipper threading machine according to claim 2, characterized in that, The process of obtaining the matching coefficient between each sliding window in the i-th slider region and each sliding window in the j-th slider region includes: The matching coefficient between the m-th sliding window in the i-th slider region and the n-th sliding window in the j-th slider region. The calculation method is as follows: in, For the m-th sliding window in the i-th pull head region, For the j-th slider region, the n-th sliding window, Let be the structural similarity index between the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region; The grayscale similarity between the m-th sliding window in the i-th slider region and the n-th sliding window in the j-th slider region; It is an exponential function with the natural constant as its base.

4. The visual recognition method for zipper pull blockage in a fully automatic zipper threading machine according to claim 1, characterized in that, The specific steps for obtaining the matching window of each sliding window in the i-th slider region in the j-th slider region include: Perform edge detection on the m-th sliding window in the i-th zipper region to obtain the number of edge pixels and the number of edges in the m-th sliding window in the i-th zipper region; The structural feature value of the m-th sliding window in the i-th pull head region is obtained based on the number of edge pixels and the number of edges. Using the structural feature values ​​of the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region, the matching coefficients of the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region are corrected to obtain the matching correction coefficients of the m-th sliding window in the i-th zipper region and the n-th sliding window in the j-th zipper region. The matching correction coefficient is used to obtain the matching window of each sliding window in the i-th slider region in the j-th slider region.

5. The visual recognition method for zipper pull blockage in a fully automatic zipper threading machine according to claim 4, characterized in that, The process of using the structural feature values ​​of the m-th sliding window in the i-th slider region and the n-th sliding window in the j-th slider region to correct the matching coefficient between the m-th sliding window in the i-th slider region and the n-th sliding window in the j-th slider region, and obtaining the matching correction coefficient between the m-th sliding window in the i-th slider region and the n-th sliding window in the j-th slider region includes: The difference between the structural feature value of the m-th sliding window in the i-th zipper region and the structural feature values ​​of the sliding windows of all pixels in the i-th zipper region is denoted as the standard structural feature value of the m-th sliding window in the i-th zipper region. ; The difference between the structural feature value of the nth sliding window in the j-th zipper region and the structural feature values ​​of the sliding windows of all pixels in the j-th zipper region is denoted as the standard value of the structural feature of the nth sliding window in the j-th zipper region. ; Matching correction coefficient between the m-th sliding window in the i-th slider region and the n-th sliding window in the j-th slider region The calculation method is as follows: in, Let be the matching coefficient between the m-th sliding window in the i-th slider region and the n-th sliding window in the j-th slider region.

6. The visual recognition method for zipper pull blockage in a fully automatic zipper threading machine according to claim 4, characterized in that, The step of using the matching correction coefficient to obtain the matching window of each sliding window in the i-th slider region in the j-th slider region includes: In the sliding window of all pixels in the j-th zipper region, the sliding window of the pixel with the largest matching correction coefficient with the m-th sliding window in the i-th zipper region is denoted as the matching window of each sliding window in the i-th zipper region in the j-th zipper region.

7. The visual recognition method for zipper pull blockage in a fully automatic zipper threading machine according to claim 1, characterized in that, The step of analyzing the positional relationship between all sliding windows in each zipper pull area and the corresponding matching windows in adjacent zipper pull areas, and the consistency of the positional relationship between these windows and other adjacent zipper pull areas in the structural matching space, to obtain the blockage coefficient of each zipper pull area, and using the blockage coefficient to identify blockages in the zipper threading machine includes: Project the line connecting all pixels in the i-th zipper region and its slope onto the structure matching space to obtain several discrete points. Use the K-means clustering algorithm with K=1 to cluster all discrete points, and denote the obtained cluster as the matchable window cluster of the i-th zipper region. Based on the distance between the discrete point of each sliding window in the structural matching space in the i-th zipper region and the cluster of matchable windows, and combined with the matching correction coefficient of each sliding window in the i-th zipper region and its matching window in the j-th zipper region, the elimination coefficient of each sliding window in the i-th zipper region is obtained. Based on the elimination coefficient, all sliding windows of the i-th zipper area are filtered to obtain the distance judgment window of the i-th zipper area; Based on the difference in length between the i-th zipper head area and the lengths of the lines connecting all distance judgment windows in other zipper head areas, the blockage coefficient of each zipper head area is obtained, and the blockage coefficient is used to identify the blockage of the zipper threading machine.

8. The visual recognition method for zipper pull blockage in a fully automatic zipper threading machine according to claim 7, characterized in that, The specific steps for obtaining the elimination coefficient include: The Euclidean distance between the discrete point of each sliding window in the i-th zipper region and the cluster center of the matchable window cluster in the i-th zipper region is denoted as the distance deviation of each sliding window in the i-th zipper region. The elimination coefficient of the m-th sliding window in the i-th slider region. The calculation method is as follows: in, Let m be the distance deviation of the m-th sliding window in the i-th slider region. Let be the matching correction coefficient between the m-th sliding window in the i-th slider region and its matching window in the j-th slider region; It is an exponential function with the natural constant as its base.

9. The visual recognition method for zipper pull blockage in a fully automatic zipper threading machine according to claim 7, characterized in that, The step of obtaining the blockage coefficient of each zipper pull area based on the difference in the length of the line connecting all distance judgment windows in the i-th zipper pull area and other zipper pull areas, and using the blockage coefficient to identify the blockage of the zipper threading machine includes: The average length of the line connecting all distance judgment windows in the i-th zipper area is denoted as the zipper distance between the i-th zipper area and its adjacent zipper areas. ; Obtain the distance between each slider region and its adjacent slider regions in the slider head image; The congestion coefficient of the i-th pull-head area The calculation method is as follows: in, This is the average distance between all slider regions and their adjacent slider regions in the slider image. It is an absolute value function; A preset congestion judgment coefficient is set; if the congestion coefficient of the i-th pull-head area is... When the value is greater than or equal to 0.8, the i-th zipper head area is marked as a blocked area. The fully automatic zipper threading machine will issue a blockage alarm and notify the operator to stop the machine for inspection, guiding the staff to carry out inspection and maintenance.

Images