A method and system for automatically detecting defects in artificial leather

Abstract

The invention relates to the technical field of machine vision detection, and in particular to an automated detection method and system for artificial leather defects, comprising: a collection layer, an analysis layer and a determination layer; the artificial leather is fixed in the collection layer and enters a state to be detected, the appearance parameters of the artificial leather in the state to be detected are collected by the collection layer, the appearance parameters of the artificial leather are preprocessed synchronously, the analysis layer further receives the preprocessed appearance parameters of the artificial leather, and analyzes the artificial leather surface defect position and the defect degree of the defect position based on the artificial leather appearance parameters, the determination layer synchronously receives the artificial leather surface defect position and the defect degree of the defect position analyzed in the analysis layer, the invention fixes the artificial leather by specifying a detection tool, collects the artificial leather appearance parameters in a fixed state, the artificial leather appearance parameters include the artificial leather thickness parameter and the artificial leather surface image data, the overall quality of the artificial leather is detected based on the comprehensive analysis of the two, and the quality qualification determination of the artificial leather is realized.

Description

Technical Field

[0001] The invention relates to the technical field of machine vision detection, and in particular to an automated detection method and system for artificial leather defects. Background Art

[0002] Artificial leather is a plastic product that looks and feels like leather and can be used instead of it. It is made by artificial synthesis, usually with fabric as the base and coated with synthetic resin. Artificial leather is relatively cheap, durable and easy to process, and has a certain degree of waterproofness. It is widely used in clothing, shoes, furniture, luggage and other fields, but its breathability is generally not as good as natural leather.

[0003] The invention patent with application number 202310116293.9 discloses an automated synthetic leather defect detection method, which is characterized in that the method includes: S1. Obtaining the product identification number of the synthetic leather to be detected; S2. Inputting the detection formula according to the product identification number; S3. Controlling the acquisition of synthetic leather images that meet the defect detection requirements of the synthetic leather to be detected; S4. Inputting the synthetic leather image into the defect defect detection model to perform defect type detection and defect location on the synthetic leather. The detection formula includes the detection standard parameters of the synthetic leather and the light source control parameters of the image acquisition. The surface image of a certain type of synthetic leather product is obtained in advance under the lighting conditions of a standard light intensity light source through a color camera; by calculating the grayscale histogram of the surface image, the grayscale characteristics of the image under standard lighting conditions are obtained, and the light source control parameters are inversely calculated, and the light source control parameters include light intensity value parameters and light source color parameters: the light source control parameters corresponding to the synthetic leather product type are recorded.

[0004] The application aims to solve the problem that "existing detection methods have a series of problems such as low detection efficiency, high false detection / missed detection rate, difficulty in standardization, poor automation and high cost."

[0005] However, the current artificial leather detection technology has greatly improved the detection efficiency, but compared with the accuracy, there is still room for optimization and improvement.

[0006] To this end, we proposed an automated detection method and system for artificial leather defects. Summary of the invention

[0007] In view of the above-mentioned shortcomings of the prior art, the present invention provides an automated detection method and system for artificial leather defects, which solves the technical problems raised in the above-mentioned background technology.

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0009] In a first aspect, an automated detection system for artificial leather defects includes: a collection layer, an analysis layer, and a determination layer;

[0010] The artificial leather is fixed in the collection layer and enters a state to be tested. The appearance parameters of the artificial leather in the state to be tested are collected by the collection layer, and the appearance parameters of the artificial leather are preprocessed synchronously. The analysis layer further receives the preprocessed appearance parameters of the artificial leather, and analyzes the position of the artificial leather surface defect and the degree of the defect position based on the artificial leather appearance parameters. The determination layer synchronously receives the position of the artificial leather surface defect and the degree of the defect position analyzed in the analysis layer, and determines whether the artificial leather is qualified based on the position of the artificial leather surface defect and the degree of the defect position;

[0011] The analysis layer includes a segmentation module, a conversion module and an analysis module. The segmentation module is used to receive the artificial leather appearance parameters and perform segmentation processing on the artificial leather surface image data in the artificial leather appearance parameters. The conversion module is used to receive the sub-artificial leather surface image data obtained by segmentation in the segmentation module and convert the sub-artificial leather surface image data into a contour image. The analysis module is used to obtain the artificial leather appearance parameters received in the segmentation module and the contour image converted from the sub-artificial leather surface image data, and combine the two to analyze the artificial leather surface defect position and the defect degree of the defect position.

[0012] The analysis logic of the position and degree of defects on the artificial leather surface is expressed as follows:

[0013] ;

[0014] Where: The defect levels shown by two groups of outer surface artificial leather image data and two groups of inner surface artificial leather image data based on the same segmentation position; is the similarity of the two groups of inner-face artificial leather image data; is the similarity of the contour images corresponding to the two groups of outer sub-leather image data; , is the weight;

[0015] Among them, the weight , The sum of is 1, and < , based on the above formula, the degree of defect of each segmented position in the artificial leather surface image data is calculated, and the analysis result of the defect position on the artificial leather surface in the analysis module is, the segmented positions in the artificial leather surface image data arranged in descending order based on the defect degree calculation result;

[0016] Said The calculation formula is:

[0017] ;

[0018] Where: , for distance and direction; is the image length and width; is grayscale; The two sets of inner surface artificial leather image data are at a distance d and in a direction Gray-level co-occurrence matrix under and The absolute value of the difference between corresponding elements; is the correction value;

[0019] Among them, the distance For the inner face leatherette image data center based on direction The linear distance to the image data boundary, direction Choose 0°, 45°, 90°, 135°, 180°, 225°, 270°, and 315°.

[0020] Furthermore, the acquisition layer includes a detection tool, a camera module and a pre-processing module, the detection tool is used to grab the artificial leather, fix the artificial leather, and collect the artificial leather structure parameters at the same time, the camera module is used to collect the surface image data of the artificial leather in a fixed state, and the pre-processing module is used to receive the artificial leather structure parameters collected by the detection tool and the artificial leather surface image data collected by the camera module, and perform noise reduction processing on the artificial leather surface image data;

[0021] Among them, the collection of artificial leather structure parameters and artificial leather surface image data is recorded as artificial leather appearance parameters. The camera module is integrated by two groups of high-definition cameras, which respectively collect the inner and outer surface image data of artificial leather. When the high-definition camera collects the inner and outer surface image data of artificial leather, the image acquisition direction of the high-definition camera lens end is perpendicular to the artificial leather surface, and the boundary of the artificial leather surface image data falls on the artificial leather surface.

[0022] Furthermore, the detection tooling includes a support member and a support plate, one side of the support member is symmetrically mounted with grab arms through an electric slide rail, the bottom surfaces of the two groups of grab arms are equidistantly mounted with electric suction cups, the surfaces of the two groups of grab arms are both mounted with grab plates through electric rotating shafts, the bottom surfaces of the two groups of grab arms are equidistantly mounted with thickness sensors, the support plate is located directly below the two groups of grab arms, and the top surface of the support plate is placed with artificial leather;

[0023] Among them, the two groups of grabbing plates are evenly equipped with electric rollers on the side that contacts the artificial leather in the moving state, the electric rollers installed on each group of grabbing plates rotate in the same direction, and the electric rollers installed on the two groups of grabbing plates rotate in opposite directions.

[0024] Furthermore, the noise reduction processing logic for the artificial leather surface image data in the preprocessing module is expressed as:

[0025] ;

[0026] Where: After noise reduction The pixel value of is the weighted median; is the pixel value at position (x, y) in the original image; is the adaptive threshold; The index that meets the conditions in the sorted sequence; is the weight of the kth pixel in the sorted sequence; is the total number of pixels in the window;

[0027] Where find represents the query, based on the above formula (1) for the original image Each pixel in the image is processed to output a denoised image, and the adaptive threshold Defined by the system end user, or subject to: :

[0028] Where: is a constant; is a positive integer; is the total number of pixels in the local window; 2n+1 is the number of pixels on the edge of the local window; is the pixel value at the coordinate (i, j) position in the original image; is the weighted median The average value of the pixel values ​​in the corresponding window;

[0029] Among them, the constant > 0, in the original image denoising stage, a window is selected in the image as the denoising target area, and the window size is ;

[0030] When the inspection tool grabs the artificial leather through the electric suction cup, the camera module collects image data of the inner and outer surfaces of the artificial leather once. When the inspection tool grabs the artificial leather through the grab plate and the electric roller, the camera module collects image data of the inner and outer surfaces of the artificial leather again.

[0031] Furthermore, after selecting a window in the image, weighted calculation is further performed on each pixel in the window, and the weighted calculation results of each pixel are further arranged in ascending order;

[0032] The weighted calculation logic of the pixel is:

[0033] ;

[0034] Where: is the weighted calculation result of pixel (i, j); is the pixel value of pixel (i, j); is the weight function;

[0035] in, , is a parameter that controls the speed of weight decay. > 0, and the closer the pixel (i, j) is to the center of the window The smaller it is, the larger the setting logic is.

[0036] Furthermore, the artificial leather surface image data is four groups of artificial leather inner and outer surface image data, and the artificial leather inner and outer surface image data is the outer surface and the inner surface of the artificial leather. When the artificial leather surface image data is segmented, the number of segments of each group of artificial leather surface image data is q×q, and the length-width ratios of the sub-artificial leather surface image data obtained by segmenting each group of artificial leather surface image data are the same;

[0037] The number of segments of the artificial leather surface image data is subject to the following: the higher the artificial leather defect detection accuracy requirement is, the more the segmentation number is; conversely, the fewer the segmentation number is;

[0038] The contour image conversion targets are all sub-artificial leather surface image data representing the outside;

[0039] Said The calculation formula is:

[0040] ;

[0041] Where: is the total amount of closed contours in the contour images corresponding to the two groups of outer sub-leather image data; , is the perimeter of the vth group of closed contours in the contour images corresponding to the two groups of outer sub-leather image data; , is the area of ​​the vth group of closed contours in the contour image corresponding to the two groups of outer sub-leather image data; , is the average curvature of the inner ring of the vth group of closed contours in the contour images corresponding to the two groups of outer sub-leather image data; is the coordination factor;

[0042] Among them, the coordination factor The value is 1 or -1, coordination factor Make the value of the fraction it contains always less than or equal to 1;

[0043] Correction value The value of , is the total amount of thickness measurement results; is the value of the thickness measurement result of group p; To get the maximum value in brackets, Express The mean value of Express The mean value of .

[0044] Furthermore, the determination layer includes a receiving module, a determination module and an output module. The receiving module is used to receive the defect degree calculation results of each segmented position in the analysis layer. The determination module is used to set the determination logic, obtain the defect degree calculation results of each segmented position received by the receiving module, and determine whether the artificial leather is qualified based on the determination logic and the calculation results. The output module is used to receive the determination result of whether the artificial leather is qualified in the determination module, and output the determination result when the determination result is yes, and output the determination result and the defect position on the artificial leather when the determination result is no;

[0045] Among them, the decision logic set in the decision module is expressed as:

[0046] ;

[0047] Where: It is the judgment value of the overall defect degree of artificial leather; is the total amount of split positions; is the degree of defects displayed by the two groups of outer surface artificial leather image data and the two groups of inner surface artificial leather image data corresponding to the h-th segmentation position; is the reference value; is the unqualified judgment value;

[0048] Among them, the reference value and unqualified judgment value Customized by the system user. , represents the decision function, If established, =1, otherwise, =0, when formula (2) holds true, it means the artificial leather is qualified, otherwise, it means the artificial leather is unqualified.

[0049] Furthermore, in the determination logic application stage in the determination module, each segmentation position in the artificial leather surface image data arranged in descending order is synchronously obtained, and based on each segmentation position in the artificial leather surface image data arranged in descending order, the determination function is used in sequence to perform calculations, and when the determination result of any one determination function is 1, the corresponding determination function values ​​of the remaining segmentation positions in the artificial leather surface image data arranged in descending order that have not been calculated are all 1;

[0050] The defective position on the artificial leather is: When it is not established, the degree of defect applied corresponds to the corresponding area on the artificial leather where the segmentation position is located.

[0051] Furthermore, the segmentation module is interactively connected to the preprocessing module via a wireless network, the preprocessing module is interactively connected to the camera module and the detection tooling via a wireless network, the segmentation module is interactively connected to the conversion module and the analysis module via a wireless network, the analysis module is interactively connected to the receiving module via a wireless network, and the receiving module is interactively connected to the determination module and the output module via a wireless network.

[0052] In a second aspect, an automated method for detecting defects in artificial leather comprises:

[0053] The artificial leather is grasped by the detection tooling, and during the process of grasping the artificial leather by the detection tooling, the surface image data of the artificial leather is collected at the first grasping stage and the tightening grasping stage;

[0054] Acquire the collected artificial leather surface image data, and perform noise reduction processing on the artificial leather surface image data;

[0055] Segmenting the artificial leather surface image data after noise reduction processing to obtain sub-artificial leather surface image data, and converting the inner and outer sub-artificial leather surface image data into contour images;

[0056] Analyze the position of defects on the surface of artificial leather and the degree of defects at the defect position according to the surface image data of artificial leather;

[0057] Setting the artificial leather qualified judgment logic, obtaining the artificial leather surface defect position and the defect degree of the defect position, and determining whether the artificial leather is qualified based on the artificial leather surface defect position and the defect degree of the defect position and the artificial leather qualified judgment logic;

[0058] If the determination result is yes, the process ends, and if the determination result is no, the position of the defect on the surface of the artificial leather is output.

[0059] Compared with the known public technology, the technical solution provided by the present invention has the following beneficial effects:

[0060] The present invention provides an automated detection method and system for artificial leather defects. During operation, the system collects artificial leather appearance parameters in a fixed state by fixing a designated detection tool. The artificial leather appearance parameters include artificial leather thickness parameters and artificial leather surface image data. Based on a comprehensive analysis of the two, the overall quality of the artificial leather is detected to achieve qualified quality judgment of the artificial leather.

[0061] In the stage of collecting artificial leather appearance parameters, the specified collection logic is set, and the artificial leather surface image data in the artificial leather appearance parameters is subjected to specific noise reduction processing to improve data accuracy. At the same time, based on image segmentation, the accuracy of artificial leather qualification judgment is effectively controlled. In the process of judging the qualification of artificial leather, the position of artificial leather surface defects can also be located, thereby achieving high-precision detection of artificial leather defects and improving the quality of artificial leather leaving the factory;

[0062] In addition, based on the execution of the above method, further operation logic support is provided for the above system, ensuring that the technical solution composed of the system and method can complete the artificial leather defect detection work more stably and effectively. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0064] Figure 1 It is a structural schematic diagram of an automated detection system for artificial leather defects;

[0065] Figure 2 A schematic diagram of a process for automated detection of artificial leather defects;

[0066] Figure 3 It is a schematic diagram of the three-dimensional structure of the detection tooling in the present invention;

[0067] Figure 4 This is a schematic diagram of the upward viewing angle of the detection tooling in the present invention;

[0068] The numbers in the figure represent: 1. support member; 2. grabbing arm; 3. electric suction cup; 4. grabbing plate; 5. thickness sensor; 6. supporting plate; 7. artificial leather. DETAILED DESCRIPTION

[0069] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0070] The present invention will be further described below in conjunction with the embodiments.

[0071] Embodiment 1:

[0072] An automatic detection system for artificial leather defects in this embodiment is as follows: Figure 1 As shown, it includes: a collection layer, an analysis layer and a decision layer;

[0073] The artificial leather is fixed in the collection layer and enters a state to be tested. The appearance parameters of the artificial leather in the state to be tested are collected by the collection layer, and the appearance parameters of the artificial leather are preprocessed synchronously. The analysis layer further receives the preprocessed appearance parameters of the artificial leather, and analyzes the position of the artificial leather surface defect and the degree of the defect position based on the artificial leather appearance parameters. The determination layer synchronously receives the position of the artificial leather surface defect and the degree of the defect position analyzed in the analysis layer, and determines whether the artificial leather is qualified based on the position of the artificial leather surface defect and the degree of the defect position;

[0074] The acquisition layer includes a detection tool, a camera module and a pre-processing module. The detection tool is used to grab the artificial leather, fix the artificial leather, and collect the structural parameters of the artificial leather at the same time. The camera module is used to collect the surface image data of the artificial leather in a fixed state. The pre-processing module is used to receive the artificial leather structural parameters collected by the detection tool and the artificial leather surface image data collected by the camera module, and perform noise reduction processing on the surface image data of the artificial leather.

[0075] The set of artificial leather structural parameters and artificial leather surface image data is recorded as artificial leather appearance parameters. The camera module is integrated by two sets of high-definition cameras, which respectively collect the inner and outer surface image data of artificial leather. When the high-definition camera collects the inner and outer surface image data of artificial leather, the image collection direction of the high-definition camera lens end is perpendicular to the surface of artificial leather, and the boundary of the artificial leather surface image data falls on the surface of artificial leather.

[0076] The noise reduction logic for the artificial leather surface image data in the preprocessing module is expressed as:

[0077] When the inspection tool grabs the artificial leather through the electric suction cup, the camera module collects the image data of the inner and outer surfaces of the artificial leather once. When the inspection tool grabs the artificial leather through the grab plate and the electric roller, the camera module collects the image data of the inner and outer surfaces of the artificial leather again.

[0078] ;

[0079] Where: After noise reduction The pixel value of is the weighted median; is the pixel value at position (x, y) in the original image; is the adaptive threshold; The index that meets the conditions in the sorted sequence; is the weight of the kth pixel in the sorted sequence; is the total number of pixels in the window;

[0080] Where find represents the query, based on the above formula (1) for the original image Each pixel in the image is processed to output a denoised image, and the adaptive threshold Defined by the system end user, or subject to: :

[0081] Where: is a constant; is a positive integer; is the total number of pixels in the local window; 2n+1 is the number of pixels on the edge of the local window; is the pixel value at the coordinate (i, j) position in the original image; is the weighted median The average value of the pixel values ​​in the corresponding window;

[0082] Among them, the constant > 0, in the original image denoising stage, a window is selected in the image as the denoising target area, and the window size is ;

[0083] After selecting a window in the image, further weighted calculation is performed on each pixel in the window, and the weighted calculation results of each pixel are further arranged in ascending order;

[0084] The pixel weighted calculation logic is:

[0085] ;

[0086] Where: is the weighted calculation result of pixel (i, j); is the pixel value of pixel (i, j); is the weight function;

[0087] in, , is a parameter that controls the speed of weight decay. > 0, and the closer the pixel (i, j) is to the center of the window The smaller the value, the larger the setting logic;

[0088] The analysis layer includes a segmentation module, a conversion module and an analysis module. The segmentation module is used to receive the artificial leather appearance parameters and perform segmentation processing on the artificial leather surface image data in the artificial leather appearance parameters. The conversion module is used to receive the sub-artificial leather surface image data obtained by segmentation in the segmentation module and convert the sub-artificial leather surface image data into a contour image. The analysis module is used to obtain the artificial leather appearance parameters received in the segmentation module and the contour image converted from the sub-artificial leather surface image data, and combine the two to analyze the artificial leather surface defect position and the defect degree of the defect position.

[0089] The analysis logic of the defect position and degree of artificial leather surface defects is expressed as:

[0090] ;

[0091] Where: The defect levels shown by two groups of outer surface artificial leather image data and two groups of inner surface artificial leather image data based on the same segmentation position; is the similarity of the two groups of inner-face artificial leather image data; is the similarity of the contour images corresponding to the two groups of outer sub-leather image data; , is the weight;

[0092] Among them, the weight , The sum of is 1, and < , based on the above formula, the degree of defect of each segmented position in the artificial leather surface image data is calculated, and the analysis result of the defect position on the artificial leather surface in the analysis module is, the segmented positions in the artificial leather surface image data arranged in descending order based on the defect degree calculation result;

[0093] It should be noted that artificial leather, as a kind of fabric, is usually divided into inner and outer surfaces. The terms inner and outer surfaces recorded in this technical solution are used to indicate the artificial inner and outer surfaces;

[0094] The calculation formula is:

[0095] ;

[0096] Where: , for distance and direction; is the image length and width; is grayscale; The two sets of inner surface artificial leather image data are at a distance d and in a direction Gray-level co-occurrence matrix under and The absolute value of the difference between corresponding elements; is the correction value;

[0097] Among them, the distance For the inner face leatherette image data center based on direction The linear distance to the image data boundary, direction Select 0°, 45°, 90°, 135°, 180°, 225°, 270°, 315°;

[0098] The artificial leather surface image data are four groups of artificial leather inner and outer surface image data, the artificial leather inner and outer surface image data are the outer surface and the inner surface of the artificial leather. When the artificial leather surface image data are segmented, the number of segments of each group of artificial leather surface image data is q×q, and the length-width ratios of the sub-artificial leather surface image data obtained by segmenting each group of artificial leather surface image data are the same;

[0099] The number of segments of the artificial leather surface image data is subject to the following: the higher the artificial leather defect detection accuracy requirement is, the more the segmentation number is, and vice versa;

[0100] The targets of the contour image conversion are all the sub-leather surface image data representing the outside;

[0101] The calculation formula is:

[0102] ;

[0103] Where: is the total amount of closed contours in the contour images corresponding to the two groups of outer sub-leather image data; , is the perimeter of the vth group of closed contours in the contour images corresponding to the two groups of outer sub-leather image data; , is the area of ​​the vth group of closed contours in the contour image corresponding to the two groups of outer sub-leather image data; , is the average curvature of the inner ring of the vth group of closed contours in the contour images corresponding to the two groups of outer sub-leather image data; is the coordination factor;

[0104] Among them, the coordination factor The value is 1 or -1, coordination factor Make the value of the fraction it contains always less than or equal to 1;

[0105] Correction value The value of , is the total amount of thickness measurement results; is the value of the thickness measurement result of group p; To get the maximum value in brackets, Express The mean value of Express The mean of ;

[0106] The determination layer includes a receiving module, a determination module and an output module. The receiving module is used to receive the defect degree calculation results of each segmented position in the analysis layer. The determination module is used to set the determination logic, obtain the defect degree calculation results of each segmented position received by the receiving module, and determine whether the artificial leather is qualified based on the determination logic and the calculation results. The output module is used to receive the determination result of whether the artificial leather is qualified in the determination module, and output the determination result when the determination result is yes, and output the determination result and the defect position on the artificial leather when the determination result is no;

[0107] Among them, the decision logic set in the decision module is expressed as:

[0108] ;

[0109] Where: It is the judgment value of the overall defect degree of artificial leather; is the total amount of split positions; is the degree of defects displayed by the two groups of outer surface artificial leather image data and the two groups of inner surface artificial leather image data corresponding to the h-th segmentation position; is the reference value; is the unqualified judgment value;

[0110] Among them, the reference value and unqualified judgment value Customized by the system user. , represents the decision function, If established, =1, otherwise, =0, when formula (2) holds true, it means the artificial leather is qualified, otherwise, it means the artificial leather is unqualified;

[0111] The segmentation module is interactively connected to the preprocessing module through a wireless network, the preprocessing module is interactively connected to the camera module and the detection tooling through a wireless network, the segmentation module is interactively connected to the conversion module and the analysis module through a wireless network, the analysis module is interactively connected to the receiving module through a wireless network, and the receiving module is interactively connected to the determination module and the output module through a wireless network.

[0112] In this embodiment, the detection tooling is operated to grab the artificial leather, fix the artificial leather, and collect the structural parameters of the artificial leather at the same time, the camera module synchronously collects the surface image data of the artificial leather in the fixed state, the pre-processing module further receives the artificial leather structural parameters collected by the detection tooling and the artificial leather surface image data collected by the camera module, performs noise reduction processing on the artificial leather surface image data, the segmentation module is post-operated to receive the artificial leather appearance parameters, and performs segmentation processing on the artificial leather surface image data in the artificial leather appearance parameters, the conversion module receives the sub-artificial leather surface image data obtained by the segmentation in the segmentation module in real time, converts the sub-artificial leather surface image data into a contour image, and then performs The analysis module obtains the appearance parameters of the artificial leather received in the segmentation module and the contour image converted from the sub-artificial leather surface image data, combines the two to analyze the defect position on the artificial leather surface and the degree of the defect position, and finally receives the defect degree calculation result of each segmentation position in the analysis layer through the receiving module. The judgment module transports and sets the judgment logic, obtains the defect degree calculation result of each segmentation position received by the receiving module, and judges whether the artificial leather is qualified based on the judgment logic and the calculation result. The output module receives the judgment result of whether the artificial leather is qualified in the judgment module, and outputs the judgment result when the judgment result is yes, and outputs the judgment result and the defect position on the artificial leather when the judgment result is no.

[0113] Through the operation of the system in the above embodiment, comprehensive defect detection is provided for the surface of artificial leather, which effectively improves the quality of finished artificial leather products.

[0114] During the operation of the system, based on the linkage operation of each module in the operation layer, as well as the image noise reduction processing, the analysis logic setting of the degree of defects, and the setting of the artificial leather qualification judgment logic, it is ensured that the system can stably judge the qualification of artificial leather and locate surface defects.

[0115] See also Figure 3 , Figure 4 As shown, the figure further shows the specific structure of the detection tooling in the acquisition layer of the above system.

[0116] Embodiment 2:

[0117] In terms of specific implementation, based on Example 1, this example refers to Figure 1 The following is a further detailed description of an automated artificial leather defect detection system in Example 1:

[0118] The detection tooling includes a support and a support plate. A grabbing arm is symmetrically installed on one side of the support through an electric slide rail. Electric suction cups are equidistantly installed on the bottom surfaces of the two groups of grabbing arms. Grabbing plates are installed on the surfaces of the two groups of grabbing arms through electric rotating shafts. Thickness sensors are equidistantly installed on the bottom surfaces of the two groups of grabbing arms. The support plate is located directly below the two groups of grabbing arms, and artificial leather is placed on the top surface of the support plate.

[0119] Among them, the two groups of grabbing plates are evenly equipped with electric rollers on the side in contact with the artificial leather in the moving state. The electric rollers installed on each group of grabbing plates rotate in the same direction, and the electric rollers installed on the two groups of grabbing plates rotate in opposite directions.

[0120] Based on the operation stage of the monitoring tooling, the artificial leather 7 is laid flat on the top surface of the supporting plate 6, and the two groups of grabbing arms 2 slide downward synchronously on the side of the support member 1 through the electric slide rail until each group of electric suction cups 3 is in contact with the surface of the artificial leather 7, and the electric suction cups 3 are started, and the artificial leather 7 is grabbed by the suction force of the electric suction cups 3, and the two groups of grabbing arms 2 are controlled to reset synchronously, and the artificial leather 7 grabbed by the electric suction cups 3 is suspended in the air. At this time, the first collection of artificial leather surface image data is performed, that is, two groups of inner and outer surface image data;

[0121] It should be noted that, during the stage when each set of electric suction cups 3 is in contact with the surface of the artificial leather 7, each set of thickness sensors 5 is used to detect the thickness of the artificial leather 7;

[0122] When the artificial leather 7 is suspended in the air, the grabbing plate 4 on the grabbing arm 2 is driven to flip through the electric shaft, and cooperates with the bottom surface of the grabbing arm 2 to clamp the artificial leather 7. The surface of the grabbing plate 4 and the artificial leather 7 rotate synchronously due to the electric rollers that clamp against each other, driving the two sides of the artificial leather 7 to stretch in opposite directions to present a taut state. At this time, the collection of artificial leather surface image data is performed again, that is, two sets of inner and outer surface image data.

[0123] It should be noted that the number of rotations of each set of electric rollers is fixed and equal, and the rotation speed remains consistent during the rotation process.

[0124] like Figure 1 As shown, in the determination logic application stage in the determination module, each segmentation position in the artificial leather surface image data arranged in descending order is synchronously obtained, and based on each segmentation position in the artificial leather surface image data arranged in descending order, the determination function is used in sequence to perform calculations, and when the determination result of any determination function is 1, the corresponding determination function values ​​of the remaining segmentation positions in the artificial leather surface image data arranged in descending order that have not been calculated are all 1;

[0125] The defective position on the artificial leather is: When it is not established, the degree of defect applied corresponds to the corresponding area on the artificial leather where the segmentation position is located.

[0126] Through the above settings, the positioning logic of the area with defects on the surface of the unqualified artificial leather is further limited when the artificial leather is judged to be unqualified, so as to provide users with more data references and provide support for the recycling of defective artificial leather.

[0127] Embodiment 3:

[0128] In terms of specific implementation, based on Example 1, this example refers to Figure 2 The following is a further detailed description of an automated artificial leather defect detection system in Example 1:

[0129] An automated method for detecting artificial leather defects, comprising:

[0130] The artificial leather is grasped by the detection tooling, and during the process of grasping the artificial leather by the detection tooling, the surface image data of the artificial leather is collected at the first grasping stage and the tightening grasping stage;

[0131] Acquire the collected artificial leather surface image data, and perform noise reduction processing on the artificial leather surface image data;

[0132] Segmenting the artificial leather surface image data after noise reduction processing to obtain sub-artificial leather surface image data, and converting the inner and outer sub-artificial leather surface image data into contour images;

[0133] Analyze the position of defects on the surface of artificial leather and the degree of defects at the defect position according to the surface image data of artificial leather;

[0134] Setting the artificial leather qualified judgment logic, obtaining the artificial leather surface defect position and the defect degree of the defect position, and determining whether the artificial leather is qualified based on the artificial leather surface defect position and the defect degree of the defect position and the artificial leather qualified judgment logic;

[0135] If the determination result is yes, the process ends, and if the determination result is no, the position of the defect on the surface of the artificial leather is output.

[0136] In summary, in the above-mentioned embodiment, during operation, the system fixes the specified detection tooling and collects the appearance parameters of artificial leather in a fixed state. The artificial leather appearance parameters include artificial leather thickness parameters and artificial leather surface image data. Based on a comprehensive analysis of the two, the overall quality of the artificial leather is inspected to achieve qualified quality judgment of the artificial leather. In the stage of collecting the appearance parameters of artificial leather, the specified acquisition logic is set, and the artificial leather surface image data in the artificial leather appearance parameters is subjected to specific noise reduction processing to improve data accuracy. At the same time, based on image segmentation, the accuracy of the qualified judgment of artificial leather is effectively controlled. In the process of qualified judgment of artificial leather, the position of defects on the surface of artificial leather can also be located, thereby achieving high-precision monitoring of artificial leather defects and improving the quality of artificial leather leaving the factory.

[0137] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An automated detection system for artificial leather defects, characterized in that: include: Collection layer, analysis layer and decision layer; The artificial leather is fixed in the collection layer and enters a state to be tested. The appearance parameters of the artificial leather in the state to be tested are collected by the collection layer, and the appearance parameters of the artificial leather are preprocessed synchronously. The analysis layer further receives the preprocessed appearance parameters of the artificial leather, and analyzes the position of the artificial leather surface defect and the degree of the defect position based on the artificial leather appearance parameters. The determination layer synchronously receives the position of the artificial leather surface defect and the degree of the defect position analyzed in the analysis layer, and determines whether the artificial leather is qualified based on the position of the artificial leather surface defect and the degree of the defect position; The analysis layer includes a segmentation module, a conversion module and an analysis module. The segmentation module is used to receive the artificial leather appearance parameters and perform segmentation processing on the artificial leather surface image data in the artificial leather appearance parameters. The conversion module is used to receive the sub-artificial leather surface image data obtained by segmentation in the segmentation module and convert the sub-artificial leather surface image data into a contour image. The analysis module is used to obtain the artificial leather appearance parameters received in the segmentation module and the contour image converted from the sub-artificial leather surface image data, and combine the two to analyze the artificial leather surface defect position and the defect degree of the defect position. The position and degree of the defect on the surface of the artificial leather are determined based on the similarity between the two groups of inner sub-artificial leather image data and the similarity between the contour images corresponding to the two groups of outer sub-artificial leather image data; In the determination layer application stage, each segmentation position in the artificial leather surface image data arranged in descending order based on the defect degree calculation result is obtained, and the determination function is used in sequence to perform calculations to determine the defect position on the artificial leather surface.

2. The automatic detection system for artificial leather defects according to claim 1, characterized in that: The acquisition layer includes a detection tool, a camera module and a pre-processing module. The detection tool is used to grab the artificial leather, fix the artificial leather, and collect the artificial leather structure parameters at the same time. The camera module is used to collect the surface image data of the artificial leather in a fixed state. The pre-processing module is used to receive the artificial leather structure parameters collected by the detection tool and the artificial leather surface image data collected by the camera module, and perform noise reduction processing on the artificial leather surface image data. Among them, the collection of artificial leather structure parameters and artificial leather surface image data is recorded as artificial leather appearance parameters. The camera module is integrated by two groups of high-definition cameras, which respectively collect the inside and outside image data of artificial leather. When the high-definition camera collects the inside and outside image data of artificial leather, the image acquisition direction of the high-definition camera lens end is perpendicular to the artificial leather surface, and the boundary of the artificial leather surface image data falls on the artificial leather surface.

3. The automatic detection system for artificial leather defects according to claim 2, characterized in that: The detection tooling includes a support member and a support plate, one side of the support member is symmetrically mounted with grab arms through an electric slide rail, the bottom surfaces of the two groups of grab arms are equidistantly mounted with electric suction cups, the surfaces of the two groups of grab arms are both mounted with grab plates through electric rotating shafts, the bottom surfaces of the two groups of grab arms are equidistantly mounted with thickness sensors, the support plate is located directly below the two groups of grab arms, and the top surface of the support plate is placed with artificial leather; The two groups of grab plates are evenly equipped with electric rollers on the sides in contact with the artificial leather in the moving state, and the electric rollers installed on each group of grab plates rotate in the same direction, and the electric rollers installed on the two groups of grab plates rotate in opposite directions; When the inspection tool grabs the artificial leather through the electric suction cup, the camera module collects image data of the inner and outer surfaces of the artificial leather once. When the inspection tool grabs the artificial leather through the grab plate and the electric roller, the camera module collects image data of the inner and outer surfaces of the artificial leather again.

4. The automatic detection system for artificial leather defects according to claim 2, characterized in that: The noise reduction processing logic for the artificial leather surface image data in the preprocessing module is expressed as follows: Where: I d (x, y) is the pixel value of I(x, y) after noise reduction; M w (x, y) is the weighted median; I(x, y) is the pixel value at position (x, y) in the original image; T(x, y) is the adaptive threshold; m is the index that meets the conditions in the sorted sequence; w k is the weight of the kth pixel in the sorted sequence; N is the total number of pixels in the window; Wherein, find represents a query. Based on the above formula (1), each pixel in the original image I(x, y) is processed to output a denoised image. The adaptive threshold T(x, y) is defined by the system end user or is subject to: Where: k is a constant; n is a positive integer; (2n+1) 2 is the total number of pixels in the local window; 2n+1 is the number of pixels on the edge of the local window; I(i,j) is the pixel value at the coordinate (i,j) position in the original image; is the weighted median M w (x,y) corresponds to the average value of the pixel value in the window; Among them, the constant k>0, in the original image denoising stage, a window is selected in the image as the denoising target area, and the window size is (2n+1) 2 .

5. The automatic detection system for artificial leather defects according to claim 4, characterized in that: After selecting a window in the image, further weighted calculation is performed on each pixel in the window, and the weighted calculation results of each pixel are further arranged in ascending order; The weighted calculation logic of the pixel is: w(i,j)=p i,j ×f(i,j); Where: w(i,j) is the weighted calculation result of pixel (i,j); p i,j is the pixel value of pixel (i, j); f(i,j) is the weight function; in, σ is a parameter that controls the weight decay speed, σ>0, and follows the setting logic that the closer the pixel (i, j) is to the window center, the smaller σ is, and vice versa.

6. The automatic detection system for artificial leather defects according to claim 1, characterized in that: The analysis logic of the position and degree of defects on the artificial leather surface is expressed as follows: K(A,B)=1-[sim(A in ,B in )×ω1+sim(A out ,B out )×ω2]; Where: K(A, B) is the degree of defects displayed by two sets of outer surface artificial leather image data and two sets of inner surface artificial leather image data based on the same segmentation position; sim(A in ,B in ) is the similarity of the two groups of inner surface artificial leather image data; sim(A out ,B out ) is the similarity of the contour images corresponding to the two groups of outer sub-leather image data; ω1, ω2 are weights; Wherein, the sum of weights ω1 and ω2 is 1, and ω1<ω2. Based on the above formula, the degree of defect of each segmented position in the artificial leather surface image data is calculated, and the analysis result of the artificial leather surface defect position in the analysis module is each segmented position in the artificial leather surface image data arranged in descending order based on the defect degree calculation result; The sim(A in ,B in ) is calculated as: Where: d, θ are distance and direction; M, N are image length and width; L is grayscale; is the gray-level co-occurrence matrix of two sets of inner-face artificial leather image data at a distance of d and a direction of θ and The absolute value of the difference between the corresponding elements; γ is the correction value; Wherein, the distance d is the straight-line distance from the inner surface artificial leather image data center to the image data boundary based on the direction θ, and the direction θ is 0°, 45°, 90°, 135°, 180°, 225°, 270°, and 315°; The artificial leather surface image data is four groups of artificial leather inner and outer surface image data, the artificial leather inner and outer surface image data is the outer surface and the inner surface of the artificial leather, when the artificial leather surface image data is segmented, the number of segments of each group of artificial leather surface image data is q×q, and the length-width ratios of the sub-artificial leather surface image data obtained by segmenting each group of artificial leather surface image data are the same; The number of segments of the artificial leather surface image data is subject to the following: the higher the artificial leather defect detection accuracy requirement is, the more the segmentation number is; conversely, the fewer the segmentation number is; The contour image conversion targets are all sub-artificial leather surface image data representing the outside; The sim(A out ,B out ) is calculated as: Where: u is the total amount of closed contours in the contour images corresponding to the two sets of outer surface artificial leather image data; C(A out ) v , C(B out ) v is the perimeter of the vth group of closed contours in the contour image corresponding to the two groups of outer sub-leather image data; s(A out ) v 、s(B out ) v is the area of ​​the vth group of closed contours in the contour image corresponding to the two groups of outer sub-leather image data; κ(A out ) v ,κ(B out ) v is the average curvature of the inner ring of the vth group of closed contours in the contour image corresponding to the two groups of outer sub-leather image data; ε is the coordination factor; Among them, the coordination factor ε takes the value of 1 or -1, and the coordination factor ε makes the value of the fraction in which it is located always less than or equal to 1; The correction value γ is Q is the total amount of thickness measurement results; T p is the value of the thickness measurement result of the pth group; max(T p , T p+1 , T p+2 , ...) is to take the maximum value in the brackets, Express The mean value of Express The mean value of .

7. The automatic detection system for artificial leather defects according to claim 2, characterized in that: The determination layer includes a receiving module, a determination module and an output module. The receiving module is used to receive the defect degree calculation results of each segmented position in the analysis layer. The determination module is used to set the determination logic, obtain the defect degree calculation results of each segmented position received by the receiving module, and determine whether the artificial leather is qualified based on the determination logic and the calculation results. The output module is used to receive the determination result of whether the artificial leather is qualified in the determination module, and output the determination result when the determination result is yes, and output the determination result and the defect position on the artificial leather when the determination result is no; Among them, the decision logic set in the decision module is expressed as: Where: G is the overall defect level of artificial leather; F is the total number of segmentation positions; K(A,B) h is the degree of defects displayed by the two groups of outer surface artificial leather image data and the two groups of inner surface artificial leather image data corresponding to the hth segmentation position; G0 is the reference value; G norr is the unqualified judgment value; Among them, the reference value G0 and the unqualified judgment value G norr Defined by the system user, f(K(A,B) h ≤G0), represents the decision function, K(A,B) h ≤G0 holds, then f(K(A,B) h ≤G0)=1, otherwise, f(K(A,B) h ≤G0)=0, when formula (2) holds, it means that the artificial leather is qualified, otherwise, it means that the artificial leather is unqualified.

8. The automatic detection system for artificial leather defects according to claim 7, characterized in that: In the determination logic application stage in the determination module, each segmentation position in the artificial leather surface image data arranged in descending order is synchronously obtained, and based on each segmentation position in the artificial leather surface image data arranged in descending order, the determination function is used in sequence to perform calculations, and when the determination result of any one determination function is 1, the remaining segmentation positions in the artificial leather surface image data arranged in descending order that have not been calculated and corresponding to the determination function value are all 1; The defect position on the artificial leather is: f(K(A,B) h ≤G0) is not true, the degree of defect applied corresponds to the corresponding area on the artificial leather where the segmentation position is located.

9. The automatic detection system for artificial leather defects according to claim 8, characterized in that: The segmentation module is interactively connected to the preprocessing module through a wireless network, the preprocessing module is interactively connected to the camera module and the detection tooling through a wireless network, the segmentation module is interactively connected to the conversion module and the analysis module through a wireless network, the analysis module is interactively connected to the receiving module through a wireless network, and the receiving module is interactively connected to the determination module and the output module through a wireless network.

10. An automated detection method for artificial leather defects, the method being an implementation method of the automated detection system for artificial leather defects as claimed in claim 9, characterized in that: include: The artificial leather is grasped by the detection tooling, and during the process of grasping the artificial leather by the detection tooling, the surface image data of the artificial leather is collected at the first grasping stage and the tightening grasping stage; Acquire the collected artificial leather surface image data, and perform noise reduction processing on the artificial leather surface image data; Segmenting the artificial leather surface image data after noise reduction processing to obtain sub-artificial leather surface image data, and converting the inner and outer sub-artificial leather surface image data into contour images; Analyze the position of defects on the surface of artificial leather and the degree of defects at the defect position according to the surface image data of artificial leather; Setting the artificial leather qualified judgment logic, obtaining the artificial leather surface defect position and the defect degree of the defect position, and determining whether the artificial leather is qualified based on the artificial leather surface defect position and the defect degree of the defect position and the artificial leather qualified judgment logic; If the determination result is yes, the process ends, and if the determination result is no, the position of the defect on the surface of the artificial leather is output.

Images