Clothing fabric inspection, analysis and processing system based on artificial intelligence

Abstract

The present invention discloses an inspection, analysis and processing system for clothing fabrics based on artificial intelligence, comprising a data acquisition unit, a preprocessing unit, an inspection and analysis unit and an optimization and control unit. The data acquisition unit acquires image data of clothing fabrics, and extracts fabric feature information through image analysis. The inspection and analysis unit then analyzes the printing and dyeing effects and the physical properties of stretching and wear resistance from the two perspectives of the appearance features and physical features of the fabrics, and then comprehensively analyzes and evaluates the clothing fabric effects. In particular, machine processing is used to perform intelligent inspection operations on clothing fabrics, and image processing technology is used to extract and analyze feature information, thereby performing quantitative analysis on image data, and artificial intelligence is used to perform data analysis and processing, thereby improving the accuracy and comprehensiveness of clothing inspection and making the operation efficient and convenient. The optimization and control unit self-grows, optimizes and improves the clothing inspection model, thereby improving the stability and reliability of data processing.

Description

Technical Field

[0001] The present invention relates to the technical field of clothing inspection, and in particular to a clothing fabric inspection, analysis and processing system based on artificial intelligence. Background Art

[0002] Artificial intelligence is a discipline that studies the use of computers to simulate certain human thought processes and intelligent behaviors (such as learning, reasoning, thinking, planning, etc.). It mainly includes the principles of computer intelligence and the manufacture of computers similar to human brain intelligence, so that computers can achieve higher-level applications. Artificial intelligence must also borrow mathematical tools. Mathematics is often considered to be the basic science of many disciplines. When mathematics enters the artificial intelligence discipline, they will promote each other and develop faster. Data analysis and processing based on artificial intelligence can make clothing inspection more efficient and convenient, thereby expanding production scale and rapid industrial development.

[0003] Existing clothing fabric inspection methods mainly rely on manual inspection, which has the defects of low inspection efficiency and insufficient refinement. It is difficult to quickly determine the printing and dyeing deviation for fabrics with complex printing and dyeing patterns, especially the comprehensive differences in printing and dyeing gloss, chroma and color difference. Therefore, it is difficult to ensure the comprehensiveness of clothing inspection, and it is difficult to quantify the physical properties of fabrics such as stretch and wear resistance. There is a problem that the clothing fabric inspection analysis and processing model is not intelligent enough, so it is difficult to ensure stability and reliability.

[0004] In view of the above technical defects, a solution is now proposed. Summary of the invention

[0005] The purpose of the present invention is to solve the defects of low efficiency, insufficient refinement and comprehensiveness of manual inspection, and difficulty in ensuring stability and reliability in existing clothing inspection, to perform intelligent inspection operations on clothing fabrics through machine processing, to extract and analyze feature information through image processing technology, thereby performing quantitative analysis on image data, and to use artificial intelligence for data analysis and processing, thereby improving the accuracy and comprehensiveness of clothing inspection and making the operation efficient and convenient, and to improve the stability and reliability of data processing by optimizing and improving the clothing inspection model through self-growth optimization and improvement of the optimization and control unit.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] The clothing fabric inspection, analysis and processing system based on artificial intelligence includes a data acquisition unit, a preprocessing unit, an inspection and analysis unit and an optimization and control unit, wherein the data acquisition unit, the preprocessing unit, the inspection and analysis unit and the optimization and control unit are signal-connected;

[0008] The data acquisition unit is used to collect image data of clothing fabrics: by inspecting clothing fabrics and combining image acquisition technology to obtain corresponding image data;

[0009] The pre-processing unit is used to extract fabric feature information: by performing preliminary analysis and processing on the image data of clothing fabrics, the corresponding fabric feature information is extracted;

[0010] The inspection and analysis unit is used to construct an image comparison model, and comprehensively evaluate the fabric effect of clothing by analyzing fabric feature information, wherein the fabric feature information includes appearance feature parameters and physical feature parameters, and the printing and dyeing effect index of the fabric is obtained by analyzing the appearance feature parameters, and the physical performance index of the fabric is obtained by the physical feature parameters, and then the clothing fabric effect index is comprehensively obtained, so as to evaluate the clothing fabric effect;

[0011] The optimization and control unit is used to automatically optimize the model: first, by comparing the manual inspection results with the system inspection results, the abnormal deviations in the inspection and analysis are identified, and then the image comparison model is verified and corrected to achieve automatic optimization and upgrading of the model.

[0012] Furthermore, the specific process of collecting image data of clothing fabrics is as follows:

[0013] The image data includes fabric printing and dyeing related images, fabric stretching related images and fabric polishing related images;

[0014] Inspection operations for clothing fabrics include printing and dyeing operations, stretching operations and polishing operations, among which polishing operations include flat polishing and folding polishing;

[0015] The fabric printing and dyeing related images include the fabric printing and dyeing preset image Pa1 and the fabric printing and dyeing actual image Pa2; the fabric stretching related images include the fabric stretching operation image Pb1, the fabric stretching operation image Pb2 and the fabric stretching operation image Pb3; the fabric polishing related images include the fabric polishing operation image Pc1, the fabric polishing operation image Pc2 and the fabric folding polishing image Pc3.

[0016] Furthermore, the specific process of preliminary analysis and processing of image data of clothing fabrics is as follows:

[0017] A1: Fabric feature information includes appearance feature parameters and physical feature parameters;

[0018] A1-1: Appearance characteristic parameters include chromaticity value, brightness value and color value of printed and dyed fabrics. Appearance characteristic parameters are extracted by preliminary analysis of fabric printing and dyeing related images;

[0019] A1-101: Synchronously extract Na0 feature regions from the image Pa1 and the image Pa2, respectively, mark any feature region of the image Pa1 as Ia1, and mark the feature region of the image Pa2 that is synchronously corresponding to the feature region Ia1 as Ia2;

[0020] A1-102: Synchronously extract Na1 feature points of the feature area Ia1 and the feature area Ia2, mark any feature point of the feature area Ia1 as Da1, and mark the feature point in the feature area Ia2 that is synchronously corresponding to the feature point Da1 as Da2;

[0021] A1-103: The chromaticity value, brightness value and color value of the marked feature point Da1 are SDa1, LDa1 and YSa1 respectively; the chromaticity value, brightness value and color value of the marked feature point Da2 are SDa2, LDa2 and YSa2 respectively;

[0022] A1-2: Physical characteristic parameters include the chromaticity value of fabric stretching and the chromaticity value of fabric polishing;

[0023] A1-201: Extract the chromaticity value of fabric stretch by performing preliminary analysis on fabric stretch correlation images;

[0024] A1-201-1: First, Nb0 feature regions are extracted from the image Pb1, any feature region of the image Pb1 is marked as Ib1, and the coordinates of the boundary points of the feature region Ib1 are marked, and then the fabric stretching operation is performed, and the feature regions corresponding to the feature region Ib1 are extracted from the images Pb2 and Pb3 according to the coordinates of the marked feature region boundary points, and are marked as feature regions Ib2 and Ib3 respectively;

[0025] A1-201-2: extract Nb1 feature points of feature area Ib1, feature area Ib2 and feature area Ib3 respectively and mark them, and mark the corresponding points in feature area Ib1, feature area Ib2 and feature area Ib3 as Db1, Db2 and Db3 respectively;

[0026] A1-201-3: The chromaticity values ​​of the marking feature points Db1, Db2 and Db3 are SDb1, SDb2 and SDb3 respectively;

[0027] A1-202: Extract the chromaticity value of fabric polishing by performing preliminary analysis on the fabric polishing related images;

[0028] A1-202-1: Synchronously extract Nc0 feature regions from the images Pc1, Pc2, and Pc3, respectively, mark any feature region of the image Pc1 as Ic1, and mark the feature regions in the images Pc2 and Pc3 that are synchronously corresponding to the feature region Ic1 as Ic2 and Ic3, respectively;

[0029] A1-202-2: Synchronously extract Nc1 feature points of feature regions Ic1, Ic2, and Ic3, mark any feature point of feature region Ic1 as Dc1, and mark the feature points in feature regions Ic2 and Ic3 that are synchronously corresponding to feature point Dc1 as Dc2 and Dc3, respectively;

[0030] A1-202-3: The chromaticity values ​​of the marked feature points Dc1, Dc2 and Dc3 are SDc1, SDc2 and SDc3 respectively.

[0031] Furthermore, the specific process of constructing the image contrast model is as follows:

[0032] The image contrast model includes a colorimetric analysis model and a verification analysis model:

[0033] B1: First build a chromaticity analysis model and obtain the chromaticity contrast evaluation coefficient PG under the current operation:

[0034] B1-1: Input the chromaticity data of n0 image groups for comparison, set any image group to include two images, mark the two images as image A and image B, preset image A and image B have N0 feature regions, mark any feature region of image A as ia, and mark the feature region of image B that is synchronously corresponding to feature region ia as ib;

[0035] B1-2: It is assumed that both the feature region ia and the feature region ib have N1 feature points, and the corresponding feature points in the feature region ia and the feature region ib are marked as da and db, respectively, and the chromaticity values ​​of the marked feature point da and the feature point db are SDda and SDdb, respectively;

[0036] B1-3: Setting the difference coefficient analysis model: Obtain the chromaticity difference coefficient CY of the current image group through the chromaticity values ​​of the feature points of the two images;

[0037] B1-4: Substitute the chromaticity values ​​of the feature points of the two images into the difference coefficient analysis model to obtain the chromaticity difference coefficient CY of the current image group;

[0038] B1-5: Then, by combining the chromaticity difference coefficients CY of the n0 image groups, an evaluation coefficient PG of the chromaticity contrast under the current operation is generated and output;

[0039] B2: Substitute the chromaticity value of the printed and dyed fabric, the chromaticity value of the fabric stretching, and the chromaticity value of the fabric polishing into the chromaticity analysis model in sequence to obtain the printing and dyeing chromaticity evaluation coefficient, the stretching evaluation coefficient, and the polishing evaluation coefficient respectively;

[0040] B2-1: For the chromaticity value of printed and dyed fabrics: by inputting an image group, namely image Pa1 and image Pa2, the printing and dyeing evaluation coefficient PGyr is output;

[0041] B2-2: For the color value of fabric stretch: by inputting two image groups, namely, image Pb1 and image Pb2, image Pb1 and image Pb3, the stretch evaluation coefficient PGls is output;

[0042] B2-3: For the chromaticity value of fabric polishing: by inputting two image groups, namely, image Pc1 and image Pc2, image Pc1 and image Pc3, the polishing evaluation coefficient PGdm is output;

[0043] B3: Reconstruct the inspection and analysis model, and obtain the printing and dyeing gloss evaluation coefficient and printing and dyeing color evaluation coefficient by analyzing the brightness value and color value of the printed and dyed fabrics, and combine them with the printing and dyeing degree evaluation coefficient to integrate and generate the printing and dyeing effect index of the fabrics.

[0044] Furthermore, the construction process of the coefficient of variation analysis model is as follows:

[0045] Input the parameter values ​​of N1 groups of feature points da and db to obtain the parameter mean of the feature area;

[0046] Then, the average value is calculated by the parameter means of N0 feature regions to obtain the parameter amplitude of the feature region of the current image group;

[0047] Then, the standard deviation of the parameter means of N0 feature regions is calculated to obtain the fluctuation coefficient of the feature region parameters;

[0048] Finally, by assigning corresponding weight factor coefficients to the characteristic region parameter amplitude and fluctuation coefficient of the current image group, the parameter difference coefficient is obtained and output;

[0049] Among them, the parameter values ​​in the difference coefficient analysis model include chromaticity value, brightness value and color value.

[0050] Furthermore, the specific process of generating the printing and dyeing effect index of the fabric is as follows:

[0051] B3-1: Substitute the brightness values ​​of image Pa1 and image Pa2 into the difference coefficient analysis model to generate the corresponding parameter difference coefficient and mark it as the printing and dyeing gloss evaluation coefficient PGgz;

[0052] B3-2: Substitute the color values ​​of image Pa1 and image Pa2 into the difference coefficient analysis model to generate the corresponding parameter difference coefficient and mark it as the printing and dyeing color evaluation coefficient PGys;

[0053] B3-3: The printing and dyeing effect index Zyr of the fabric is generated by combining the printing and dyeing gloss evaluation coefficient PGgz, the printing and dyeing color evaluation coefficient PGys and the printing and dyeing degree evaluation coefficient PGyr.

[0054] Furthermore, the specific process of obtaining the clothing fabric effect index and evaluating the clothing fabric effect is as follows:

[0055] The physical performance index Zwl of the fabric is generated by combining the tensile evaluation coefficient PGls and the polishing evaluation coefficient PGdm;

[0056] By combining the fabric printing and dyeing effect index Zyr and the physical performance index Zwl, the clothing fabric effect index XG is comprehensively obtained;

[0057] Then set the standard range of the clothing fabric effect index XG, and evaluate the clothing fabric effect through range comparison.

[0058] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0059] The present invention acquires image data of clothing fabrics through a data acquisition unit, extracts fabric feature information through image analysis, and then analyzes the printing and dyeing effect and the physical properties of stretching and wear resistance of the fabric from the two perspectives of appearance features and physical features in the fabric feature information through an inspection and analysis unit, thereby comprehensively analyzing and evaluating the clothing fabric effect, wherein intelligent inspection operation of clothing fabrics is performed through machine processing, and feature information extraction and analysis is performed through image processing technology, thereby quantitatively analyzing the image data, and using artificial intelligence to perform data analysis and processing, thereby improving the accuracy and comprehensiveness of clothing inspection and making the operation efficient and convenient, and optimizing and improving the clothing inspection model through the self-growth optimization and improvement of the optimization and control unit, thereby improving the stability and reliability of data processing;

[0060] The present invention analyzes the chromaticity, brightness and color information in the fabric image data, thereby comparing and analyzing the difference in fabric states before and after printing and dyeing operations, stretching operations and polishing operations, and then comprehensively analyzes the printing and dyeing effects of the fabric through the differences in printing and dyeing gloss, color and chromaticity, and then comprehensively evaluates the physical properties of the fabric through the stretching difference and the polishing difference, and then integrates the printing and dyeing effects and physical properties of the fabric to analyze the fabric effects of clothing, ensure the comprehensiveness and precision of data analysis, combine artificial intelligence algorithms with manual inspection results, and establish a detection system in which artificial intelligence and human work complement and promote each other, making the system more complete and efficient. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 Shows a schematic diagram of the modules of the present invention;

[0062] Figure 2 A schematic diagram of the process of the present invention is shown. DETAILED DESCRIPTION

[0063] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0064] Embodiment 1:

[0065] like Figure 1-2 As shown, the clothing fabric inspection, analysis and processing system based on artificial intelligence includes a data acquisition unit, a preprocessing unit, an inspection and analysis unit and an optimization and control unit, wherein the data acquisition unit, the preprocessing unit, the inspection and analysis unit and the optimization and control unit are signal connected;

[0066] Among them, artificial intelligence is an important driving force for the new round of scientific and technological revolution and industrial transformation. It is a new technical science that studies and develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence. Artificial intelligence is a very broad science, including robots, language recognition, image recognition, natural language processing, expert systems, machine learning, computer vision, etc. The present invention is based on artificial intelligence to design inspection models and analyze and process data, and optimize the model through self-growth through deep learning, so as to realize the transformation of the clothing industry in the new era;

[0067] The working steps are as follows:

[0068] Sa: The data acquisition unit collects image data of clothing fabrics: by inspecting clothing fabrics and combining image acquisition technology to obtain corresponding image data, the specific process is as follows:

[0069] The image data includes fabric printing and dyeing related images, fabric stretching related images and fabric polishing related images;

[0070] Inspection operations for clothing fabrics include printing and dyeing operations, stretching operations and polishing operations, among which polishing operations include flat polishing and folding polishing;

[0071] Sa-1: Fabric printing and dyeing related images include the fabric printing and dyeing preset image Pa1 and the fabric printing and dyeing actual image Pa2; the original fabric to be printed and dyed is marked as a0, the printing and dyeing effect image of the original fabric a0 is Pa1, the marked fabric a0 after the printing and dyeing operation is completed is fabric a2, and the image Pa2 is obtained by scanning and collecting the image of fabric a2;

[0072] Sa-2: fabric stretching associated images include image Pb1 before fabric stretching operation, image Pb2 during fabric stretching operation, and image Pb3 after fabric stretching operation;

[0073] The original fabric to be stretched is marked as b1, and then a stretching force is applied to the fabric b1 to perform a stretching operation. When the fabric is stretched to the maximum extension value, an image is scanned and acquired to obtain an image Pb2. Then, the stretching force is withdrawn to allow the fabric to rebound. The fabric after rebound is marked as b3. By scanning and acquiring images of the original fabric b1 and the fabric b3 respectively, images Pb1 and image Pb3 are acquired in sequence.

[0074] Sa-3: Fabric grinding related images include the image Pc1 before fabric grinding operation, the image Pc2 after fabric flat grinding, and the image Pc3 after fabric folding grinding;

[0075] The original fabric before the grinding operation is marked as c1, fabric c2 ​​is obtained by performing a flat grinding operation on fabric c1, fabric c3 is obtained by performing a folding grinding operation on fabric c1, and images of fabric c1, fabric c2 ​​and fabric c3 are scanned and collected in sequence to obtain images Pc1, image Pc2 and image Pc3 respectively;

[0076] Sb: The pre-processing unit extracts fabric feature information: By performing preliminary analysis and processing on the image data of clothing fabrics, the corresponding fabric feature information is extracted. The specific process is as follows:

[0077] Fabric feature information includes appearance feature parameters and physical feature parameters;

[0078] Sb-1: Appearance characteristic parameters include chromaticity value, brightness value and color value of printed and dyed fabrics, among which the chromaticity value of the image is extracted by a colorimeter, the brightness value of the image is extracted by a brightness meter, and the color value of the image is extracted by image processing tools such as color picker software;

[0079] By making a preliminary analysis of the fabric printing and dyeing related images, the appearance feature parameters are extracted;

[0080] Sb-101: Synchronously extract Na0 feature regions from the image Pa1 and the image Pa2 respectively, mark any feature region of the image Pa1 as Ia1, and mark the feature region of the image Pa2 that is synchronously corresponding to the feature region Ia1 as Ia2;

[0081] Sb-102: Synchronously extract Na1 feature points of the feature area Ia1 and the feature area Ia2, mark any feature point of the feature area Ia1 as Da1, and mark the feature point in the feature area Ia2 that is synchronously corresponding to the feature point Da1 as Da2;

[0082] Sb-103: The chromaticity value, brightness value and color value of the marked feature point Da1 are SDa1, LDa1 and YSa1 respectively; the chromaticity value, brightness value and color value of the marked feature point Da2 are SDa2, LDa2 and YSa2 respectively;

[0083] Sb-2: Physical characteristic parameters include the chromaticity value of fabric stretching and the chromaticity value of fabric polishing;

[0084] Sb-2-1: Extract the chromaticity value of fabric stretch by preliminary analysis of fabric stretch correlation image;

[0085] Sb-2-101: First, Nb0 feature regions are extracted from the image Pb1, any feature region of the image Pb1 is marked as Ib1, and the coordinates of the boundary points of the feature region Ib1 are marked, and then the fabric stretching operation is performed, and the feature regions corresponding to the feature region Ib1 are extracted from the image Pb2 and the image Pb3 respectively according to the coordinates of the marked feature region boundary points, and are marked as the feature region Ib2 and the feature region Ib3 respectively;

[0086] Sb-2-102: extract Nb1 feature points of the feature area Ib1, the feature area Ib2 and the feature area Ib3 respectively and mark them, and mark the corresponding points in the feature area Ib1, the feature area Ib2 and the feature area Ib3 as Db1, Db2 and Db3 respectively;

[0087] Sb-2-103: the chromaticity values ​​of the marking feature points Db1, Db2 and Db3 are SDb1, SDb2 and SDb3 respectively;

[0088] Sb-2-2: Extract the chromaticity value of fabric polishing by preliminary analysis of fabric polishing related images;

[0089] Sb-2-201: Synchronously extract Nc0 feature regions from the images Pc1, Pc2, and Pc3, respectively, mark any feature region of the image Pc1 as Ic1, and mark the feature regions in the images Pc2 and Pc3 that are synchronously corresponding to the feature region Ic1 as Ic2 and Ic3, respectively;

[0090] Sb-2-202: Synchronously extract Nc1 feature points of feature regions Ic1, Ic2, and Ic3, mark any feature point of feature region Ic1 as Dc1, and mark the feature points in feature regions Ic2 and Ic3 that are synchronously corresponding to feature point Dc1 as Dc2 and Dc3, respectively;

[0091] Sb-2-203: the chromaticity values ​​of the marking feature points Dc1, Dc2 and Dc3 are SDc1, SDc2 and SDc3 respectively;

[0092] Sc: The inspection and analysis unit constructs an image comparison model and comprehensively evaluates the fabric effect of clothing by analyzing fabric feature information, wherein the fabric feature information includes appearance feature parameters and physical feature parameters. The printing and dyeing effect index of the fabric is obtained by analyzing the appearance feature parameters, and the physical performance index of the fabric is obtained by the physical feature parameters, and then the clothing fabric effect index is comprehensively obtained to evaluate the clothing fabric effect;

[0093] The specific process of building an image contrast model is as follows:

[0094] The image contrast model includes a colorimetric analysis model and a verification analysis model:

[0095] Sc-1: First build a colorimetric analysis model. The specific process is as follows:

[0096] Sc-1-1: Input the chromaticity data of n0 image groups for comparison, set any image group to include two images, mark the two images as image A and image B, preset image A and image B have N0 feature regions, mark any feature region of image A as ia, and mark the feature region of image B that is synchronously corresponding to feature region ia as ib;

[0097] Sc-1-2: It is assumed that both the feature region ia and the feature region ib have N1 feature points, and the corresponding feature points in the feature region ia and the feature region ib are marked as da and db, respectively, and the chromaticity values ​​of the marked feature point da and the feature point db are SDda and SDdb, respectively;

[0098] Sc-1-3: Set the difference coefficient analysis model: Get the chromaticity difference coefficient CY of the current image group through the chromaticity values ​​of the feature points of the two images:

[0099] The construction process of the coefficient of variation analysis model is as follows:

[0100] Input the parameter values ​​of the N1 groups of feature points da and db to obtain the parameter mean of the feature area; then calculate the average value through the parameter means of the N0 feature areas to obtain the parameter amplitude of the feature area of ​​the current image group; then calculate the standard deviation of the parameter means of the N0 feature areas to obtain the fluctuation coefficient of the feature area parameters; finally, assign corresponding weight factor coefficients to the parameter amplitude and fluctuation coefficient of the feature area of ​​the current image group, thereby obtaining the parameter difference coefficient; wherein the parameter values ​​in the difference coefficient analysis model refer to chromaticity value, brightness value and color value;

[0101] Substitute the chromaticity values ​​of the feature points of the two images into the difference coefficient analysis model to obtain the chromaticity difference coefficient CY of the current image group. The specific process is as follows:

[0102] Sc-1-301: First, obtain the color difference mean jz of the feature area through the chromaticity values ​​of N1 groups of feature points da and feature points db:

[0103] Sc-1-302: Then, the color difference amplitude sc of the feature area of ​​the current image group is obtained through the color difference mean jz of N0 feature areas:

[0104] Sc-1-303: Then obtain the color difference fluctuation coefficient σc of the characteristic area of ​​the current image group:

[0105] Sc-1-304: Output the chromaticity difference coefficient CY by combining the color difference amplitude sc and the color difference fluctuation coefficient σc of the characteristic area of ​​the current image group: CY=α1*sc+α2*σ C ;

[0106] Among them, α1 and α2 are weight factor coefficients of the characteristic region color difference amplitude sc and the color difference fluctuation coefficient σc of the current image group, respectively, and α1 and α2 are both greater than 0; when the characteristic region color difference amplitude sc is higher, it means that the overall average amplitude of the chromaticity difference between the characteristic regions of several fabrics under the current specific operation is higher, and when the color difference fluctuation coefficient σc is higher, it means that the fluctuation amplitude of the chromaticity difference between the characteristic regions of several fabrics under the current specific operation is higher, and the difference amplitude is unbalanced, then the chromaticity difference coefficient CY is higher, which comprehensively indicates that the fabric effect evaluation under the current specific operation is worse;

[0107] Sc-1-305: Then, by combining the chromaticity difference coefficients CY of the n0 image groups, the chromaticity contrast evaluation coefficient PG under the current operation is generated and output:

[0108] Among them, φi is the weight coefficient of the chromaticity difference coefficient CY, and φi is greater than 0; when the chromaticity difference coefficient CY of any pair of image groups is higher, it means that the difference between the fabrics under the current specific operation is higher, which means that the comprehensive evaluation of the fabric effect is worse;

[0109] Sc-2: Substitute the chromaticity value of the printed and dyed fabric, the chromaticity value of the fabric stretching, and the chromaticity value of the fabric polishing into the chromaticity analysis model in turn to obtain the printing and dyeing evaluation coefficient, the stretching evaluation coefficient, and the polishing evaluation coefficient respectively;

[0110] For the chromaticity value of printed and dyed fabrics: by inputting an image group, namely image Pa1 and image Pa2, the printing and dyeing evaluation coefficient PGyr is output;

[0111] For the chromaticity value of fabric stretch: by inputting two image groups, namely, image Pb1 and image Pb2, image Pb1 and image Pb3, the stretch evaluation coefficient PGls is output;

[0112] For the chromaticity value of fabric polishing: by inputting two image groups, namely, image Pc1 and image Pc2, image Pc1 and image Pc3, the polishing evaluation coefficient PGdm is output;

[0113] Sc-3: Reconstruct the inspection and analysis model, and obtain the printing and dyeing gloss evaluation coefficient and printing and dyeing color evaluation coefficient by analyzing the brightness value and color value of the printed and dyed fabrics, and combine them with the printing and dyeing degree evaluation coefficient to integrate and generate the printing and dyeing effect index of the fabrics. The specific process is as follows:

[0114] Sc-3-1: Substitute the brightness values ​​of images Pa1 and Pa2 into the difference coefficient analysis model to generate the corresponding parameter difference coefficient and mark it as the printing and dyeing gloss evaluation coefficient PGgz; Substitute the color values ​​of images Pa1 and Pa2 into the difference coefficient analysis model to generate the corresponding parameter difference coefficient and mark it as the printing and dyeing color evaluation coefficient PGys;

[0115] Sc-3-2: The printing and dyeing effect index Zyr of the fabric is generated by combining the printing and dyeing gloss evaluation coefficient PGgz, the printing and dyeing color evaluation coefficient PGys and the printing and dyeing degree evaluation coefficient PGyr:

[0116]

[0117] Among them, ω1, ω2 and ω3 are the weight coefficients of the printing and dyeing gloss evaluation coefficient PGgz, the printing and dyeing color evaluation coefficient PGys and the printing and dyeing degree evaluation coefficient PGyr, respectively, and ω1, ω2 and ω3 are all greater than 0; when the printing and dyeing gloss evaluation coefficient PGgz, the printing and dyeing color evaluation coefficient PGys and the printing and dyeing degree evaluation coefficient PGyr are higher, the printing and dyeing effect index Zyr of the fabric is lower, indicating that the greater the difference in printing and dyeing gloss, the greater the difference in printing and dyeing color and the greater the difference in printing and dyeing degree, the worse the comprehensive printing and dyeing effect of the fabric is;

[0118] Sc-3-3: The physical performance index Zwl of the fabric is generated by combining the tensile evaluation coefficient PGls and the polishing evaluation coefficient PGdm:

[0119] Among them, υ1 and υ2 are weight coefficients of the stretching evaluation coefficient PGls and the grinding evaluation coefficient PGdm, respectively, and υ1 and υ2 are both greater than 0; when the stretching evaluation coefficient PGls and the grinding evaluation coefficient PGdm are higher, the physical performance index Zwl of the fabric is lower, indicating that the greater the color difference before and after stretching and the greater the color difference before and after grinding, it means that the fabric may be damaged and deformed due to stretching, and the fabric may be damaged by burrs and breakage due to grinding, which means that the comprehensive physical properties of the fabric are worse;

[0120] Sc-3-4: Then, by combining the printing and dyeing effect index Zyr and the physical performance index Zwl of the fabric, the clothing fabric effect index XG is comprehensively obtained: XG = β1*Zyr+β2*Zwl;

[0121] Among them, β1 and β2 are the weight coefficients of the printing and dyeing effect index Zyr and the physical performance index Zwl, respectively, and both β1 and β2 are greater than 0; when the printing and dyeing effect index Zyr and the physical performance index Zwl are higher, the clothing fabric effect index XG is higher, indicating that the printing and dyeing effect and physical performance of the fabric are better, which in turn comprehensively indicates that the evaluation effect of the clothing fabric is better;

[0122] Sc-3-5: Set the standard range of clothing fabric effect index XG, and evaluate the clothing fabric effect by comparing the ranges:

[0123] There are M0 standard intervals preset, and any standard interval is marked as Hm (where m is the serial number of the standard interval, 0<m≤M0). When the clothing fabric effect index XG is in the standard interval Hm, an m-level clothing fabric effect signal is generated, and the clothing fabric effect level is determined to be m-level, and it is output to the background as a system inspection result for visualization; the higher the level of the clothing fabric effect signal, the better the clothing fabric effect;

[0124] Sd: The optimization control unit automatically optimizes the model: First, by comparing the manual inspection results with the system inspection results, the abnormal deviation of the inspection analysis can be identified, and then the image comparison model can be verified and corrected to achieve automatic optimization and upgrading of the model. The specific process is as follows:

[0125] Professional clothing inspectors conduct manual inspections to determine the level of fabric effects, thereby obtaining manual inspection results, and comparing the manual inspection results with the system inspection results to perform verification and correction operations on the image comparison model. The verification and correction operations include adaptive adjustments to the weight coefficients and preset standard intervals in the model, thereby continuously optimizing the image comparison model to make it closer to the actual inspection results, and ultimately achieving automatic optimization and upgrading of the model.

[0126] In summary, the present invention acquires image data of clothing fabrics through a data acquisition unit, extracts fabric feature information through image analysis, and then analyzes the printing and dyeing effect and the physical properties of stretching and wear resistance of the fabric from the two perspectives of appearance features and physical features in the fabric feature information through an inspection and analysis unit, thereby comprehensively analyzing and evaluating the effect of the clothing fabric, wherein intelligent inspection operation of clothing fabrics is performed through machine processing, and feature information extraction and analysis is performed through image processing technology, thereby quantitatively analyzing the image data, and using artificial intelligence to perform data analysis and processing, thereby improving the accuracy and comprehensiveness of clothing inspection and making the operation efficient and convenient, and optimizing and improving the clothing inspection model through the self-growth optimization and improvement of the optimization and control unit, thereby improving the stability and reliability of data processing;

[0127] The present invention analyzes the chromaticity, brightness and color information in the fabric image data, thereby comparing and analyzing the difference in fabric states before and after printing and dyeing operations, stretching operations and polishing operations, and then comprehensively analyzes the printing and dyeing effects of the fabric through the differences in printing and dyeing gloss, color and chromaticity, and then comprehensively evaluates the physical properties of the fabric through the stretching difference and the polishing difference, and then integrates the printing and dyeing effects and physical properties of the fabric to analyze the fabric effects of clothing, ensure the comprehensiveness and precision of data analysis, combine artificial intelligence algorithms with manual inspection results, and establish a detection system in which artificial intelligence and human work complement and promote each other, making the system more complete and efficient.

[0128] The size of the interval and threshold is set to facilitate comparison. The size of the threshold depends on the amount of sample data and the number of bases set by technical personnel in this field for each group of sample data; as long as it does not affect the proportional relationship between the parameter and the quantized value.

[0129] The above formulas are all dimensionless and numerical calculations. The formula is a formula obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0130] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. The clothing fabric inspection, analysis and processing system based on artificial intelligence is characterized by: It includes a data acquisition unit, a preprocessing unit, a test and analysis unit and an optimization and control unit, wherein the data acquisition unit, the preprocessing unit, the test and analysis unit and the optimization and control unit are signal connected; The data acquisition unit is used to collect image data of clothing fabrics: by inspecting clothing fabrics and combining image acquisition technology to obtain corresponding image data; The pre-processing unit is used to extract fabric feature information: by performing preliminary analysis and processing on the image data of clothing fabrics, the corresponding fabric feature information is extracted; The inspection and analysis unit is used to construct an image comparison model, and comprehensively evaluate the fabric effect of clothing by analyzing fabric feature information, wherein the fabric feature information includes appearance feature parameters and physical feature parameters, and the printing and dyeing effect index of the fabric is obtained by analyzing the appearance feature parameters, and the physical performance index of the fabric is obtained by the physical feature parameters, and then the clothing fabric effect index is comprehensively obtained, so as to evaluate the clothing fabric effect; The optimization control unit is used to automatically optimize the model: first, by comparing the manual inspection results with the system inspection results, the abnormal deviation of the inspection analysis can be identified, and then the image comparison model can be verified and corrected to achieve automatic optimization and upgrading of the model; The specific process of collecting image data of clothing fabrics is as follows: The image data includes fabric printing and dyeing related images, fabric stretching related images and fabric polishing related images; Inspection operations for clothing fabrics include printing and dyeing operations, stretching operations and polishing operations, among which polishing operations include flat polishing and folding polishing; The fabric printing and dyeing related images include the fabric printing and dyeing preset image Pa1 and the fabric printing and dyeing actual image Pa2; the fabric stretching related images include the fabric stretching operation image Pb1, the fabric stretching operation image Pb2 and the fabric stretching operation image Pb3; the fabric grinding related images include the fabric grinding operation image Pc1, the fabric flat grinding image Pc2 and the fabric folding grinding image Pc3; The specific process of preliminary analysis and processing of image data of clothing fabrics is as follows: A1: Fabric feature information includes appearance feature parameters and physical feature parameters; A1-1: Appearance characteristic parameters include chromaticity value, brightness value and color value of printed and dyed fabrics. Appearance characteristic parameters are extracted by preliminary analysis of fabric printing and dyeing related images; A1-101: Synchronously extract Na0 feature regions from the image Pa1 and the image Pa2, respectively, mark any feature region of the image Pa1 as Ia1, and mark the feature region of the image Pa2 that is synchronously corresponding to the feature region Ia1 as Ia2; A1-102: Synchronously extract Na1 feature points of the feature area Ia1 and the feature area Ia2, mark any feature point of the feature area Ia1 as Da1, and mark the feature point in the feature area Ia2 that is synchronously corresponding to the feature point Da1 as Da2; A1-103: The chromaticity value, brightness value and color value of the marked feature point Da1 are SDa1, LDa1 and YSa1 respectively; the chromaticity value, brightness value and color value of the marked feature point Da2 are SDa2, LDa2 and YSa2 respectively; A1-2: Physical characteristic parameters include the chromaticity value of fabric stretching and the chromaticity value of fabric polishing; A1-201: Extract the chromaticity value of fabric stretch by performing preliminary analysis on fabric stretch correlation images; A1-201-1: First, Nb0 feature regions are extracted from the image Pb1, any feature region of the image Pb1 is marked as Ib1, and the coordinates of the boundary points of the feature region Ib1 are marked, and then the fabric stretching operation is performed, and the feature regions corresponding to the feature region Ib1 are extracted from the images Pb2 and Pb3 according to the coordinates of the marked feature region boundary points, and are marked as feature regions Ib2 and Ib3 respectively; A1-201-2: extract Nb1 feature points of feature area Ib1, feature area Ib2 and feature area Ib3 respectively and mark them, and mark the corresponding points in feature area Ib1, feature area Ib2 and feature area Ib3 as Db1, Db2 and Db3 respectively; A1-201-3: The chromaticity values ​​of the marking feature points Db1, Db2 and Db3 are SDb1, SDb2 and SDb3 respectively; A1-202: Extract the chromaticity value of fabric polishing by performing preliminary analysis on the fabric polishing related images; A1-202-1: Synchronously extract Nc0 feature regions from the images Pc1, Pc2, and Pc3, respectively, mark any feature region of the image Pc1 as Ic1, and mark the feature regions in the images Pc2 and Pc3 that are synchronously corresponding to the feature region Ic1 as Ic2 and Ic3, respectively; A1-202-2: Synchronously extract Nc1 feature points of feature regions Ic1, Ic2, and Ic3, mark any feature point of feature region Ic1 as Dc1, and mark the feature points in feature regions Ic2 and Ic3 that are synchronously corresponding to feature point Dc1 as Dc2 and Dc3, respectively; A1-202-3: The chromaticity values ​​of the marked feature points Dc1, Dc2 and Dc3 are SDc1, SDc2 and SDc3 respectively.

2. The clothing fabric inspection, analysis and processing system based on artificial intelligence according to claim 1 is characterized by: The specific process of building an image contrast model is as follows: The image contrast model includes a colorimetric analysis model and a verification analysis model: B1: First build a chromaticity analysis model and obtain the chromaticity contrast evaluation coefficient PG under the current operation: B1-1: Input the chromaticity data of n0 image groups for comparison, set any image group to include two images, mark the two images as image A and image B, preset image A and image B have N0 feature regions, mark any feature region of image A as ia, and mark the feature region of image B that is synchronously corresponding to feature region ia as ib; B1-2: It is assumed that both the feature region ia and the feature region ib have N1 feature points, and the corresponding feature points in the feature region ia and the feature region ib are marked as da and db, respectively, and the chromaticity values ​​of the marked feature point da and the feature point db are SDda and SDdb, respectively; B1-3: Setting the difference coefficient analysis model: Obtain the chromaticity difference coefficient CY of the current image group through the chromaticity values ​​of the feature points of the two images; B1-4: Substitute the chromaticity values ​​of the feature points of the two images into the difference coefficient analysis model to obtain the chromaticity difference coefficient CY of the current image group; B1-5: Then, by combining the chromaticity difference coefficients CY of the n0 image groups, an evaluation coefficient PG of the chromaticity contrast under the current operation is generated and output; B2: Substitute the chromaticity value of the printed and dyed fabric, the chromaticity value of the fabric stretching, and the chromaticity value of the fabric polishing into the chromaticity analysis model in sequence to obtain the printing and dyeing chromaticity evaluation coefficient, the stretching evaluation coefficient, and the polishing evaluation coefficient respectively; B2-1: For the chromaticity value of printed and dyed fabrics: by inputting an image group, namely image Pa1 and image Pa2, the printing and dyeing evaluation coefficient PGyr is output; B2-2: For the color value of fabric stretch: by inputting two image groups, namely, image Pb1 and image Pb2, image Pb1 and image Pb3, the stretch evaluation coefficient PGls is output; B2-3: For the chromaticity value of fabric polishing: by inputting two image groups, namely, image Pc1 and image Pc2, image Pc1 and image Pc3, the polishing evaluation coefficient PGdm is output; B3: Reconstruct the inspection and analysis model, and obtain the printing and dyeing gloss evaluation coefficient and printing and dyeing color evaluation coefficient by analyzing the brightness value and color value of the printed and dyed fabrics, and combine them with the printing and dyeing degree evaluation coefficient to integrate and generate the printing and dyeing effect index of the fabrics.

3. The clothing fabric inspection, analysis and processing system based on artificial intelligence according to claim 2 is characterized in that: The construction process of the coefficient of variation analysis model is as follows: Input the parameter values ​​of N1 groups of feature points da and db to obtain the parameter mean of the feature area; Then, the average value is calculated by the parameter means of N0 feature regions to obtain the parameter amplitude of the feature region of the current image group; Then, the standard deviation of the parameter means of N0 feature regions is calculated to obtain the fluctuation coefficient of the feature region parameters; Finally, by assigning corresponding weight factor coefficients to the characteristic region parameter amplitude and fluctuation coefficient of the current image group, the parameter difference coefficient is obtained and output; Among them, the parameter values ​​in the difference coefficient analysis model include chromaticity value, brightness value and color value.

4. The clothing fabric inspection, analysis and processing system based on artificial intelligence according to claim 3 is characterized by: The specific process of generating the printing and dyeing effect index of fabrics is: B3-1: Substitute the brightness values ​​of image Pa1 and image Pa2 into the difference coefficient analysis model to generate the corresponding parameter difference coefficient and mark it as the printing and dyeing gloss evaluation coefficient PGgz; B3-2: Substitute the color values ​​of image Pa1 and image Pa2 into the difference coefficient analysis model to generate the corresponding parameter difference coefficient and mark it as the printing and dyeing color evaluation coefficient PGys; B3-3: The printing and dyeing effect index Zyr of the fabric is generated by combining the printing and dyeing gloss evaluation coefficient PGgz, the printing and dyeing color evaluation coefficient PGys and the printing and dyeing degree evaluation coefficient PGyr.

5. The clothing fabric inspection, analysis and processing system based on artificial intelligence according to claim 4 is characterized in that: The specific process of obtaining the clothing fabric effect index and evaluating the clothing fabric effect is as follows: The physical performance index Zwl of the fabric is generated by combining the tensile evaluation coefficient PGls and the polishing evaluation coefficient PGdm; By combining the fabric printing and dyeing effect index Zyr and the physical performance index Zwl, the clothing fabric effect index XG is comprehensively obtained; Then set the standard range of the clothing fabric effect index XG, and evaluate the clothing fabric effect through range comparison.

Images