CN113936001B - Textile surface flaw detection method based on image processing technology - Google Patents

Textile surface flaw detection method based on image processing technology Download PDF

Info

Publication number
CN113936001B
CN113936001B CN202111545882.6A CN202111545882A CN113936001B CN 113936001 B CN113936001 B CN 113936001B CN 202111545882 A CN202111545882 A CN 202111545882A CN 113936001 B CN113936001 B CN 113936001B
Authority
CN
China
Prior art keywords
fabric
image
neural network
silk
artificial neural
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202111545882.6A
Other languages
Chinese (zh)
Other versions
CN113936001A (en
Inventor
周兴华
范海宁
姜思炎
秦志磊
黄建强
杨茁筠
徐皓
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hangzhou Youhuasijie Culture And Art Development Co ltd
Original Assignee
Hangzhou Youhuasijie Culture And Art Development Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Hangzhou Youhuasijie Culture And Art Development Co ltd filed Critical Hangzhou Youhuasijie Culture And Art Development Co ltd
Priority to CN202111545882.6A priority Critical patent/CN113936001B/en
Publication of CN113936001A publication Critical patent/CN113936001A/en
Application granted granted Critical
Publication of CN113936001B publication Critical patent/CN113936001B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0004—Industrial image inspection
    • G06T7/001—Industrial image inspection using an image reference approach
    • D—TEXTILES; PAPER
    • D03—WEAVING
    • D03D—WOVEN FABRICS; METHODS OF WEAVING; LOOMS
    • D03D15/00—Woven fabrics characterised by the material, structure or properties of the fibres, filaments, yarns, threads or other warp or weft elements used
    • D03D15/20—Woven fabrics characterised by the material, structure or properties of the fibres, filaments, yarns, threads or other warp or weft elements used characterised by the material of the fibres or filaments constituting the yarns or threads
    • D03D15/233—Woven fabrics characterised by the material, structure or properties of the fibres, filaments, yarns, threads or other warp or weft elements used characterised by the material of the fibres or filaments constituting the yarns or threads protein-based, e.g. wool or silk
    • D03D15/235—Cashmere or silk
    • D—TEXTILES; PAPER
    • D03—WEAVING
    • D03D—WOVEN FABRICS; METHODS OF WEAVING; LOOMS
    • D03D15/00—Woven fabrics characterised by the material, structure or properties of the fibres, filaments, yarns, threads or other warp or weft elements used
    • D03D15/50—Woven fabrics characterised by the material, structure or properties of the fibres, filaments, yarns, threads or other warp or weft elements used characterised by the properties of the yarns or threads
    • D03D15/54—Woven fabrics characterised by the material, structure or properties of the fibres, filaments, yarns, threads or other warp or weft elements used characterised by the properties of the yarns or threads coloured
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0004—Industrial image inspection
    • G06T7/0006—Industrial image inspection using a design-rule based approach
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/13—Edge detection
    • D—TEXTILES; PAPER
    • D10—INDEXING SCHEME ASSOCIATED WITH SUBLASSES OF SECTION D, RELATING TO TEXTILES
    • D10B—INDEXING SCHEME ASSOCIATED WITH SUBLASSES OF SECTION D, RELATING TO TEXTILES
    • D10B2211/00—Protein-based fibres, e.g. animal fibres
    • D10B2211/01—Natural animal fibres, e.g. keratin fibres
    • D10B2211/04—Silk
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20024—Filtering details
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30108—Industrial image inspection
    • G06T2207/30124—Fabrics; Textile; Paper

Abstract

The invention relates to a textile surface flaw point detection method based on an image processing technology, which comprises the following steps: step 1: determining and classifying the types and colors of silk threads used by the fabric by using an artificial neural network, and extracting and selecting color matching scheme data by using an artificial neural network technology; step 2: weaving according to a preset weaving scheme; and step 3: and collecting a fabric image, detecting the fabric image by using a Gabor filter, and marking and segmenting flaw points in the fabric image. The method and the device have the advantages that the odd-symmetry Gabor filter and the even-symmetry Gabor filter are utilized to carry out flaw detection on the fabric image, so that the accuracy of flaw detection is increased, the analysis and the modification of the cause of the fabric flaw are facilitated, and the production efficiency of the fabric is increased.

Description

Textile surface flaw detection method based on image processing technology
Technical Field
The invention relates to the technical field of textile processes, in particular to a textile surface flaw point detection method based on an image processing technology.
Background
With the development of times and the advancement of science and technology, the production quantity of the fabric cannot meet the market demand due to the fact that the fabric produced by the existing brocade process consumes a lot of time in production because of the complex process, one reason of which is that a lot of flaws exist in the finished fabric product due to the machine or manual errors in the weaving process of the fabric, and the fabric design and the fabric entity have more differences due to the different coloring degrees of the fabric in the design process of the fabric.
In the textile industry, the flaw detection of cloth is an important link, and most enterprises at home and abroad still use the traditional manual detection. The manual detection obviously has many limitations, and the workman is unfavorable for healthily for long-time operation, and to the enterprise, long-term input cost is high, detection efficiency is low, the missed measure rate is high.
In the existing technical field of fabric defect point detection, commonly used fabric defect point monitoring algorithms can be classified into three categories according to the surface characteristics of fabrics: statistical feature-based algorithms, neural network-based algorithms, and spectral analysis-based algorithms.
The algorithm based on statistical features includes: the method comprises a gray scale statistical method, a fractal dimension defect point detection method, a morphological algorithm and a defect point detection algorithm based on symbiotic gray moment characteristics, wherein the four algorithms have great limitations in practical application, and the detection algorithm based on the frequency spectrum analysis is expensive and is not suitable for random texture materials.
The fabric flaw detection algorithm based on the neural network has proved to have good performance on detecting complex fabric flaw due to the characteristic of nonparametric influence, and the training and calculation process of the network is relatively simple, but in the actual application process, detection errors are often caused by folds, light rays and the like, so that the detection accuracy is low.
Disclosure of Invention
In order to overcome the technical defects in the prior art, the invention provides a textile surface flaw point detection method based on an image processing technology, which can effectively solve the problems in the background art.
In order to solve the technical problems, the technical scheme provided by the invention is as follows:
the embodiment of the invention discloses a textile surface flaw point detection method based on an image processing technology, which comprises the following steps:
step 1: determining and classifying the types and colors of silk threads used by the fabric by using an artificial neural network, and extracting and selecting color matching scheme data by using an artificial neural network technology;
step 2: weaving according to a preset weaving scheme;
and step 3: and collecting a fabric image, detecting the fabric image by using a Gabor filter, and marking and segmenting flaw points in the fabric image.
The artificial neural network technology is a network system established by simulating a human brain neural network, in the artificial intelligence technology, the scale of the artificial neural network is large, main information can be organized and arranged, and the information is inferred, so that a method for solving problems is found, meanwhile, the artificial neural network has strong adaptability and can comprehensively arrange various information types, so that the type and the color of silk threads required by the fabric can be determined by utilizing the artificial neural network in the brocade process, and the visual and sensory effects required by the fabric are achieved.
In any of the above solutions, it is preferable that, when determining the type and color of the silk thread used in the fabric by using the artificial neural network technology, the method includes the following steps:
step 1.1: inputting parameters of the selectable fabric into an artificial neural network to form a database;
step 1.2, inputting parameters of the fabric to be woven into an artificial neural network, and selecting the most similar raw material scheme, processing steps and required raw materials by using the artificial neural network;
step 1.3: simulating the result of the scheme selected by the artificial neural network by using a genetic algorithm;
step 1.4: and judging the scheme by using a computer vision technology.
In any of the above solutions, preferably, when determining the type and color of the silk used by the fabric by using the artificial neural network technology, the parameters of the fabric include: the overall style of the fabric, the color of the fabric, the size of the fabric, the shape of the fabric, and the use of the fabric.
In any of the above schemes, preferably, when determining the type and color of the silk thread used by the fabric by using the artificial neural network technology, the parameters of the fabric to be woven include: roughness of the silk thread, difficulty of dyeing the silk thread, coloring effect of the silk thread, brightness of the silk thread, use path of the silk thread and thickness of the silk thread.
In any of the above aspects, it is preferable that, when the aspect is judged by using the computer vision technology, the method includes the steps of:
and establishing a model of the knitted fabric by using 3Dmax software, endowing different characteristics on the surface of the knitted fabric, comparing a fabric model image with a fabric design image after the modeling is finished, if the difference between the pre-fabric model image and the fabric design image is large, selecting the type and the color of the selected silk thread again if the type and the color of the selected silk thread are unqualified, and selecting the type and the color of the silk thread again until the pre-knitted fabric image is consistent with the fabric reference image.
In any of the above embodiments, when different properties are imparted to the surface of the knitted fabric, the imparted properties include roughness of the yarn, brightness of the yarn, and dyeing effect of the yarn.
The genetic algorithm mainly carries out related calculation on data of the biological evolution process, mainly converts the biological evolution process into a data form, and establishes a mathematical model according to the obtained data. And then finding the problems, and calculating a solution to the problems according to the problems, wherein in the brocade process, the genetic algorithm can perform data conversion on the products to be prevented, a mathematical model is established, deduction and calculation are performed on the textile products, and the final results of the textile products are obtained according to the input parameters.
In any one of the above aspects, it is preferable that when knitting is performed according to a predetermined knitting scheme, the method includes the steps of:
step 2.1: performing primary processing on raw silk, specifically, reeling the raw silk, twisting the raw silk into a twisted silk thread, refining and enzymatic refining the raw silk, and dyeing the raw silk;
step 2.2: re-processing raw silk, specifically, firstly carrying out silk resistance and silk adjustment on dyed silk threads, then carrying out pattern arrangement and warp drawing on the silk threads, and then carrying out drafting and denting;
step 2.3: and (4) carrying out weft winding operation on the yarns subjected to drafting and reeding, and weaving after finishing the weft winding operation.
In any of the above schemes, preferably, when detecting the woven fabric, the method comprises the following steps:
step 3.1: inspecting the edges of the woven fabric;
step 3.2: sampling the woven fabric, and collecting surface images of the sample by using a high-definition camera to obtain fabric images;
step 3.3: processing the collected fabric image to obtain a processed fabric image;
step 3.4: and identifying and calculating the processed fabric image by using an artificial neural network, and marking the flaw points.
In any of the above schemes, preferably, when processing the acquired image, the method comprises the following steps:
the method comprises the following steps: filtering the image by using a median filtering method to remove irrelevant noise in the image;
step two: carrying out gray processing on the obtained surface image of the fabric to obtain a gray image of the surface texture of the fabric;
step three: performing mathematical morphology binarization processing on the surface image of the fabric subjected to the graying processing;
step four: and performing thresholding operation on the image after the graying processing and the binarization processing of mathematical morphology, and converting the image into a binary image.
In any of the above schemes, preferably, when the artificial neural network is used for identifying and calculating the image, the method includes the following steps:
step 3.4.1: inputting a standard clear defect image and a standard defect image of a fabric with the same texture into an artificial neural network, and processing the images;
step 3.4.2: inputting the processed standard clear defect image into a odd-symmetric Gabor filter, inputting the processed standard clear defect image into an even-symmetric Gabor filter, and performing spatial filtering on the standard clear defect image and the standard clear defect image by using the odd-symmetric Gabor filter and the even-symmetric Gabor filter to obtain a filtering result image of the clear defect image and the defect image;
step 3.4.3: calculating a threshold value according to the standard clear defect image and the filtering result image of the standard defect image, and performing binarization operation on the filtering result image to respectively obtain a odd detection template and an even detection template;
step 3.4.4: and fusing the odd detection template and the even detection template, inputting the fabric image into an artificial neural network, and detecting the fabric image by using the fused odd detection template and even detection template to obtain a flaw point segmentation result.
In any of the above embodiments, it is preferable that a Tophat operation method is used when the binarization processing is performed on the filtering result image; wherein, Tophat operation is defined as the difference between the on operations of image A and image B, and is defined as: b = a- (a ∘ S); where A is the image before processing and S is a structural element.
In any one of the above aspects, it is preferable that the method includes, when dividing the defective dot image, the steps of:
step 3.4.4.1: performing convolution operation on the two groups of 3 multiplied by 3 matrixes Gx and Gy and the fabric image to respectively obtain two groups of gradient value components;
step 3.4.4.2: solving the size of the approximate gradient value by using a gradient value formula;
step 3.4.4.3: and comparing the approximate gradient value with a threshold value to obtain the edge of the flaw.
In any of the above aspects, when detecting a defect in a fabric image, it is preferable that the type of defect in the fabric image indicates that a bright defect exists in the fabric image if the average gray scale value of the fabric image is greater than the gray scale value of the standard image, and indicates that a dark defect exists in the fabric image if the average gray scale value of the fabric image is less than the gray scale value of the standard image.
In any of the above schemes, preferably, when the artificial neural network is used to identify and calculate the image, the selected artificial neural network is a convolutional neural network.
In any of the above solutions, it is preferable that the fabric flaw detection and positioning system further includes:
the image acquisition module is used for acquiring surface images of the fabric and storing the acquired images;
the file conversion module is used for storing and marking the acquired image and recording the position of the image on the fabric;
the image processing module is used for processing the collected fabric surface image and calculating the gray value of each point in the image;
the flaw point detection module is used for carrying out flaw detection on the processed image;
the image storage module is used for storing the image information and the position information of the flaw;
and the image display module is used for displaying the shot image and the processed image.
In any of the above aspects, preferably, the principle of the defect point monitoring module is as follows: reading in a flaw image and a non-flaw image, searching the maximum gray value and the minimum gray value of each row in a standard image, searching the maximum gray value and the minimum gray value of each column in the standard image, determining the probability of abnormal pixels in the flaw image, determining a threshold value by using a gray histogram of the flaw image, thresholding the flaw image, performing median filtering, and outputting a flaw detection result.
Compared with the prior art, the invention has the beneficial effects that:
1. according to the invention, the fabric is subjected to three-dimensional modeling, and is analyzed by using an artificial neural network and a computer vision technology, so that the design drawing of the fabric is closer to a real object, and the fabric is more attractive.
2. According to the invention, the processing raw materials of the fabric are input into the artificial neural network, and the calculation is carried out by matching with the genetic algorithm, so that the finished fabric product is closer to the design drawing, and the attractiveness of the fabric is improved.
3. The method and the device have the advantages that the odd-symmetry Gabor filter and the even-symmetry Gabor filter are utilized to carry out flaw detection on the fabric image, so that the accuracy of flaw detection is increased, the analysis and the modification of the cause of the fabric flaw are facilitated, and the production efficiency of the fabric is increased.
Drawings
The drawings are included to provide a further understanding of the invention, and are incorporated in and constitute a part of this specification.
FIG. 1 is a schematic structural diagram of a method for detecting a textile surface defect point based on an image processing technology according to an embodiment of the present invention;
FIG. 2 is a fabric image in a method for detecting a defect point on a surface of a textile based on an image processing technology according to an embodiment of the present invention;
fig. 3 is a segmented image of a defect point in a method for detecting a defect point on a textile surface based on an image processing technology according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
It will be understood that when an element is referred to as being "secured to" or "disposed on" another element, it can be directly on the other element or be indirectly on the other element. When an element is referred to as being "connected to" another element, it can be directly connected to the other element or be indirectly connected to the other element.
In the description of the present invention, it is to be understood that the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", and the like, indicate orientations or positional relationships based on the orientations or positional relationships illustrated in the drawings, and are used merely for convenience in describing the present invention and for simplicity in description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed in a particular orientation, and be operated, and thus, are not to be construed as limiting the present invention.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of the present invention, "a plurality" means two or more unless specifically defined otherwise.
For better understanding of the above technical solutions, the technical solutions of the present invention will be described in detail below with reference to the drawings and the detailed description of the present invention.
Referring to fig. 1, a method for detecting a textile surface defect point based on an image processing technology includes the following steps:
step 1: determining and classifying the types and colors of silk threads used by the fabric by using an artificial neural network, and extracting and selecting color matching scheme data by using an artificial neural network technology;
step 2: weaving according to a preset weaving scheme;
and step 3: and collecting a fabric image, detecting the fabric image by using a Gabor filter, and marking and segmenting flaw points in the fabric image.
In the tapestry process based on artificial intelligence, the artificial neural network technology is a network system established by simulating a human brain neural network, in the artificial intelligence technology, the scale of the artificial neural network is large, main information can be organized and sorted, and the information can be inferred, so that a method for solving the problem is found, meanwhile, the artificial neural network has strong adaptability and can be used for comprehensively sorting various information types, so that the type and color of silk threads required by the fabric can be determined by using the artificial neural network in the tapestry process, and the visual and sensory effects required by the fabric are achieved.
In the tapestry process based on artificial intelligence, when the artificial neural network technology is used for determining the types and colors of silk threads used by a fabric, the process comprises the following steps:
step 1.1: inputting parameters of the selectable fabric into an artificial neural network to form a database;
step 1.2, inputting parameters of the fabric to be woven into an artificial neural network, and selecting the most similar raw material scheme, processing steps and required raw materials by using the artificial neural network;
step 1.3: simulating the result of the scheme selected by the artificial neural network by using a genetic algorithm;
step 1.4: and judging the scheme by using a computer vision technology.
In the tapestry process based on artificial intelligence in the embodiment of the invention, when the type and color of the silk used by the fabric are determined by using the artificial neural network technology, the parameters of the fabric comprise: the overall style of the fabric, the color of the fabric, the size of the fabric, the shape of the fabric, and the use of the fabric.
In the tapestry process based on artificial intelligence in the embodiment of the invention, when the type and color of silk used by a fabric are determined by using an artificial neural network technology, parameters of the fabric to be woven comprise: roughness of the silk thread, difficulty of dyeing the silk thread, coloring effect of the silk thread, brightness of the silk thread, use path of the silk thread and thickness of the silk thread.
In the tapestry process based on artificial intelligence, when a computer vision technology is used for judging a scheme, the process comprises the following steps:
and establishing a model of the knitted fabric by using 3Dmax software, endowing different characteristics on the surface of the knitted fabric, comparing a fabric model image with a fabric design image after the modeling is finished, if the difference between the pre-fabric model image and the fabric design image is large, selecting the type and the color of the selected silk thread again if the type and the color of the selected silk thread are unqualified, and selecting the type and the color of the silk thread again until the pre-knitted fabric image is consistent with the fabric reference image.
In the brocade process based on artificial intelligence, when different characteristics are endowed on the surface of the knitted fabric, the endowed characteristics comprise the roughness of the silk threads, the brightness of the silk threads and the dyeing effect of the silk threads.
In the tapestry process based on artificial intelligence, when weaving is carried out according to a preset weaving scheme, the process comprises the following steps:
step 2.1: performing primary processing on raw silk, specifically, reeling the raw silk, twisting the raw silk into a twisted silk thread, refining and enzymatic refining the raw silk, and dyeing the raw silk;
step 2.2: re-processing raw silk, specifically, firstly carrying out silk resistance and silk adjustment on dyed silk threads, then carrying out pattern arrangement and warp drawing on the silk threads, and then carrying out drafting and denting;
step 2.3: and (4) carrying out weft winding operation on the yarns subjected to drafting and reeding, and weaving after finishing the weft winding operation.
In the tapestry process based on artificial intelligence, when detecting the woven fabric, the process comprises the following steps:
step 3.1: inspecting the edges of the woven fabric;
step 3.2: sampling the woven fabric, and collecting surface images of the sample by using a high-definition camera to obtain fabric images;
step 3.3: processing the collected fabric image to obtain a processed fabric image;
step 3.4: and identifying and calculating the processed fabric image by using an artificial neural network, and marking the flaw points.
In the tapestry process based on artificial intelligence, when processing the collected image, the process comprises the following steps:
the method comprises the following steps: filtering the image by using a median filtering method to remove irrelevant noise in the image;
step two: carrying out gray processing on the obtained surface image of the fabric to obtain a gray image of the surface texture of the fabric;
step three: performing mathematical morphology binarization processing on the surface image of the fabric subjected to the graying processing;
step four: and performing thresholding operation on the image after the graying processing and the binarization processing of mathematical morphology, and converting the image into a binary image.
In the tapestry process based on artificial intelligence, when the artificial neural network is used for identifying and calculating the image, the process comprises the following steps:
step 3.4.1: inputting a standard clear defect image and a standard defect image of a fabric with the same texture into an artificial neural network, and processing the images;
step 3.4.2: inputting the processed standard clear defect image into a odd-symmetric Gabor filter, inputting the processed standard clear defect image into an even-symmetric Gabor filter, and performing spatial filtering on the standard clear defect image and the standard clear defect image by using the odd-symmetric Gabor filter and the even-symmetric Gabor filter to obtain a filtering result image of the clear defect image and the defect image;
step 3.4.3: calculating a threshold value according to the standard clear defect image and the filtering result image of the standard defect image, and performing binarization operation on the filtering result image to respectively obtain a odd detection template and an even detection template;
step 3.4.4: and fusing the odd detection template and the even detection template, inputting the fabric image into an artificial neural network, and detecting the fabric image by using the fused odd detection template and even detection template to obtain a flaw point segmentation result.
In the tapestry process based on artificial intelligence, a Tophat operation method is adopted when binarization processing is carried out on a filtering result image; wherein, Tophat operation is defined as the difference between the on operations of image A and image B, and is defined as: b = a- (a ∘ S); where A is the image before processing and S is a structural element.
In the tapestry process based on artificial intelligence, when a flaw point image is segmented, the method comprises the following steps:
step 3.4.4.1: performing convolution operation on the two groups of 3 multiplied by 3 matrixes Gx and Gy and the fabric image to respectively obtain two groups of gradient value components;
step 3.4.4.2: solving the size of the approximate gradient value by using a gradient value formula;
step 3.4.4.3: comparing the approximate gradient value with a threshold value to obtain the edge of the flaw point, wherein Gx =
Figure 373745DEST_PATH_IMAGE001
，Gy=
Figure 573783DEST_PATH_IMAGE002
The approximate gradient value is calculated by the following formula: g =
Figure DEST_PATH_IMAGE003
Figure 761223DEST_PATH_IMAGE003
(ii) a Wherein G is an approximate gradient value; the image A is an original image, the edge of the flaw is divided, and the edge of the flaw can be more accurately determined, so that the shape and size of the flaw are determined, and workers can conveniently analyze the occurrence reason of the flaw.
In the tapestry process based on artificial intelligence, when detecting the flaws in the fabric image, if the average gray value of the fabric image is larger than the gray value of the standard image, it is indicated that bright flaws exist in the fabric image, if the average gray value of the fabric image is smaller than the gray value of the standard image, it is indicated that dark flaws exist in the fabric image, and when the artificial neural network is used for identifying and calculating the image, the selected artificial neural network is a convolutional neural network.
In the tapestry process based on artificial intelligence in the embodiment of the invention, the invention further comprises a fabric flaw detection and positioning system, wherein the fabric flaw detection and positioning system comprises:
the image acquisition module is used for acquiring surface images of the fabric and storing the acquired images;
the file conversion module is used for storing and marking the acquired image and recording the position of the image on the fabric;
the image processing module is used for processing the collected fabric surface image and calculating the gray value of each point in the image;
the flaw detection module is used for carrying out flaw detection on the processed image, and the principle of the flaw monitoring module is as follows: reading in a flaw image and a non-flaw image, searching the maximum gray value and the minimum gray value of each row in a standard image, searching the maximum gray value and the minimum gray value of each column in the standard image, determining the probability of abnormal pixels in the flaw image, determining a threshold value by using a gray histogram of the flaw image, thresholding the flaw image, performing median filtering, and outputting a flaw detection result;
the image storage module is used for storing the image information and the position information of the flaw;
and the image display module is used for displaying the shot image and the processed image.
Example 3
A textile surface flaw point detection method based on an image processing technology comprises the following steps:
step 1: according to the designed water painting pattern of the Sichuan brocade, yellow, blue, purple, black and secondary colors in the design picture are extracted by utilizing an artificial neural network, and then the dyeing scheme of silk threads is screened and extracted by utilizing the artificial neural network technology, and the method specifically comprises the following steps:
step 1.1: inputting parameters of selectable fabrics into an artificial neural network to form a database, wherein the glossiness of available silk threads comprises three types of softness, roughness and smoothness, and the thickness of the available silk threads comprises four types of 48NM, 60NM, 80NM and 120 NM;
step 1.2, inputting parameters of a fabric to be woven into an artificial neural network, selecting the most similar raw material scheme and processing steps and required raw materials by using the artificial neural network, wherein 12 selectable dyes are selected, and screening and matching 32 selectable schemes according to a fabric design drawing;
step 1.3: simulating the result of the scheme selected by the artificial neural network by using a genetic algorithm;
step 1.4: judging the scheme by using a computer vision technology, and selecting the optimal collocation from 32 selectable schemes, wherein the silk threads are selected from soft 80NM silk threads, rough 120NM silk threads and soft 48NM silk threads which are smooth in knot;
step 2: weaving according to a predetermined weaving scheme, comprising the steps of:
step 2.1: performing primary processing on raw silk, specifically, reeling the raw silk, twisting the raw silk into a twisted silk thread, refining and enzymatic refining the raw silk, and dyeing the raw silk;
step 2.2: re-processing raw silk, specifically, firstly carrying out silk resistance and silk adjustment on dyed silk threads, then carrying out pattern arrangement and warp drawing on the silk threads, and then carrying out drafting and denting;
step 2.3: performing weft winding operation on the fully-penetrated and reeded silk threads, and weaving after the weft winding operation is completed;
and step 3: collecting a fabric image, detecting the fabric image by using a Gabor filter, and marking and segmenting flaw points in the fabric image; the method comprises the following steps:
step 3.1: inspecting the edges of the woven fabric;
step 3.2: sampling the woven fabric, and collecting surface images of the sample by using a high-definition camera to obtain fabric images;
step 3.3: processing the collected fabric image to obtain a processed fabric image, as shown in fig. 2;
step 3.4: identifying and calculating the processed fabric image by using an artificial neural network, and marking the flaw points; when the artificial neural network is used for identifying and calculating the image, the method comprises the following steps:
step 3.4.1: inputting a standard clear defect image and a standard defect image of a fabric with the same texture into an artificial neural network, and processing the images;
step 3.4.2: inputting the processed standard clear defect image into a odd-symmetric Gabor filter, inputting the processed standard clear defect image into an even-symmetric Gabor filter, and performing spatial filtering on the standard clear defect image and the standard clear defect image by using the odd-symmetric Gabor filter and the even-symmetric Gabor filter to obtain a filtering result image of the clear defect image and the defect image;
step 3.4.3: calculating a threshold value according to the standard clear defect image and the filtering result image of the standard defect image, performing binarization operation on the filtering result image to respectively obtain a odd detection template and an even detection template, and adopting a Tophat operation method when performing binarization processing on the filtering result image; wherein, Tophat operation is defined as the difference between the on operations of image A and image B, and is defined as: b = a- (a ∘ S); wherein A is an image before processing, and S is a structural element;
step 3.4.4: fusing the odd detection template and the even detection template, inputting the fabric image into an artificial neural network, detecting the fabric image by using the fused odd detection template and even detection template to obtain a flaw segmentation result, wherein the segmented flaw image is shown in FIG. 3, and the method comprises the following steps when segmenting the flaw image:
step 3.4.4.1: performing convolution operation on the two groups of 3 multiplied by 3 matrixes Gx and Gy and the fabric image to respectively obtain two groups of gradient value components;
step 3.4.4.2: solving the size of the approximate gradient value by using a gradient value formula;
step 3.4.4.3: comparing the approximate gradient value with a threshold value to obtain the edge of the flaw point, wherein Gx =
Figure 922077DEST_PATH_IMAGE001
，Gy=
Figure 190247DEST_PATH_IMAGE002
The approximate gradient value is calculated by the following formula: g =
Figure 120157DEST_PATH_IMAGE003
Figure 753264DEST_PATH_IMAGE003
(ii) a Wherein G is an approximate gradient value; the image A is an original image, when the image A is used, the edge of the flaw is divided, and the edge of the flaw can be more accurately determined, so that the shape and the size of the flaw are determined, and a worker can conveniently analyze the occurrence reason of the flaw.
Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art will understand that various changes, modifications and substitutions can be made without departing from the spirit and scope of the invention as defined by the appended claims. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (9)

1. A textile surface flaw point detection method based on an image processing technology is characterized by comprising the following steps: the method comprises the following steps:
step 1: determining and classifying the types and colors of silk threads used by the fabric by using an artificial neural network, and extracting and selecting color matching scheme data by using an artificial neural network technology;
step 2: weaving according to a preset weaving scheme;
and step 3: collecting a fabric image, detecting the fabric image by using a Gabor filter, and marking and segmenting flaw points in the fabric image; when the artificial neural network technology is used for determining the types and the colors of the silk threads used by the fabric, the method comprises the following steps:
step 1.1: inputting parameters of the selectable fabric into an artificial neural network to form a database;
step 1.2: inputting parameters of the fabric to be woven into an artificial neural network, and selecting the most similar raw material scheme, processing steps and required raw materials by using the artificial neural network;
step 1.3: simulating the result of the scheme selected by the artificial neural network by using a genetic algorithm;
step 1.4: and judging the scheme by using a computer vision technology.
2. The method for detecting textile surface flaws based on image processing technology as claimed in claim 1, wherein: when the artificial neural network technology is used for determining the types and colors of silk threads used by the fabric, the parameters of the fabric and the parameters of the fabric to be woven can be selected to comprise: the overall style of the fabric, the color of the fabric, the size of the fabric, the shape of the fabric, and the use of the fabric.
3. The method for detecting textile surface flaws based on image processing technology as claimed in claim 2, wherein: when the artificial neural network technology is used for determining the types and colors of silk threads used by the fabric, the parameters of the fabric to be woven comprise: roughness of the silk thread, difficulty of dyeing the silk thread, coloring effect of the silk thread, brightness of the silk thread, use path of the silk thread and thickness of the silk thread.
4. The method for detecting textile surface flaws based on image processing technology as claimed in claim 3, wherein: when the computer vision technology is used for judging the scheme, the method comprises the following steps: and establishing a model of the knitted fabric by using 3Dmax software, endowing different characteristics on the surface of the knitted fabric, comparing a fabric model image with a fabric design image after the modeling is finished, if the difference between the pre-fabric model image and the fabric design image is large, selecting the type and the color of the selected silk thread again if the type and the color of the selected silk thread are unqualified, and selecting the type and the color of the silk thread again until the pre-knitted fabric image is consistent with the fabric reference image.
5. The method for detecting textile surface flaws based on image processing technology as claimed in claim 4, wherein: when different properties are imparted to the surface of the knitted fabric, the imparted properties include the roughness of the yarn, the lightness of the yarn, and the dyeing effect of the yarn.
6. The method for detecting textile surface flaws based on image processing technology as claimed in claim 5, wherein: when weaving is carried out according to a predetermined weaving scheme, the method comprises the following steps:
step 2.1: performing primary processing on raw silk, reeling the raw silk, twisting the raw silk into a hank-shaped silk thread, refining and enzymatic refining the raw silk, and dyeing the raw silk;
step 2.2: re-processing raw silk, performing silk resistance and silk adjustment on the dyed silk threads, then performing pattern arrangement and warp drawing on the silk threads, and then performing drafting and denting;
step 2.3: and (4) carrying out weft winding operation on the yarns subjected to drafting and reeding, and weaving after finishing the weft winding operation.
7. The method for detecting textile surface flaws based on image processing technology as claimed in claim 6, wherein: when the fabric image is detected, the method comprises the following steps:
step 3.1: inspecting the edges of the woven fabric;
step 3.2: sampling the woven fabric, and collecting surface images of the sample by using a high-definition camera to obtain fabric images;
step 3.3: processing the collected fabric image to obtain a processed fabric image;
step 3.4: and detecting the fabric image by using a Gabor filter, identifying and calculating the processed fabric image by using an artificial neural network, and marking and segmenting the flaw points.
8. The method for detecting textile surface flaws based on image processing technology as claimed in claim 7, wherein: when the artificial neural network is used for identifying and calculating the image, the method comprises the following steps:
step 3.4.1: inputting a standard clear defect image and a standard defect image of a fabric with the same texture into an artificial neural network, and processing the images;
step 3.4.2: inputting the processed standard clear defect image into an odd-symmetric Gabor filter, inputting the processed standard clear defect image into an even-symmetric Gabor filter, calculating the optimal parameters of the Gabor filter by using a genetic algorithm, performing spatial filtering on the standard clear defect image and the standard clear defect image by using the odd-symmetric Gabor filter and the even-symmetric Gabor filter, and obtaining filtering result images of the clear defect image and the defect image;
step 3.4.3: calculating a threshold value according to the standard clear defect image and the filtering result image of the standard defect image, and performing binarization operation on the filtering result image to respectively obtain a odd detection template and an even detection template;
step 3.4.4: and fusing the odd detection template and the even detection template, inputting the fabric image into an artificial neural network, and detecting the fabric image by using the fused odd detection template and even detection template to obtain a flaw point segmentation result.
9. The method for detecting textile surface flaws based on image processing technology as claimed in claim 8, wherein: when the flaw point image is divided, the method comprises the following steps:
step 3.4.4.1: performing convolution operation on the two groups of 3 multiplied by 3 matrixes Gx and Gy and the fabric image to respectively obtain two groups of gradient value components;
step 3.4.4.2: solving the size of the approximate gradient value by using a gradient value formula;
step 3.4.4.3: and comparing the approximate gradient value with a threshold value to obtain the edge of the flaw.
CN202111545882.6A 2021-12-17 2021-12-17 Textile surface flaw detection method based on image processing technology Active CN113936001B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111545882.6A CN113936001B (en) 2021-12-17 2021-12-17 Textile surface flaw detection method based on image processing technology

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111545882.6A CN113936001B (en) 2021-12-17 2021-12-17 Textile surface flaw detection method based on image processing technology

Publications (2)

Publication Number Publication Date
CN113936001A CN113936001A (en) 2022-01-14
CN113936001B true CN113936001B (en) 2022-03-04

Family

ID=79289214

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111545882.6A Active CN113936001B (en) 2021-12-17 2021-12-17 Textile surface flaw detection method based on image processing technology

Country Status (1)

Country Link
CN (1) CN113936001B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115018826B (en) * 2022-08-02 2022-12-13 南通市爱诺家用纺织品有限公司 Fabric flaw detection method and system based on image recognition
CN117132593B (en) * 2023-10-25 2024-03-26 济宁华晟服装股份有限公司 Cloth roughness detection method for resisting periodic texture influence

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1475944A (en) * 2003-06-29 2004-02-18 浙江大学银河自动化有限公司 Fabric fleuron designing method and data storage device
CN101866427A (en) * 2010-07-06 2010-10-20 西安电子科技大学 Method for detecting and classifying fabric defects
US11761128B2 (en) * 2018-03-17 2023-09-19 Drexel University Topology optimization for modeling and prediction of complex fabric structures and properties
CN110349146B (en) * 2019-07-11 2020-06-02 中原工学院 Method for constructing fabric defect identification system based on lightweight convolutional neural network

Also Published As

Publication number Publication date
CN113936001A (en) 2022-01-14

Similar Documents

Publication Publication Date Title
Rasheed et al. Fabric defect detection using computer vision techniques: a comprehensive review
CN114549522B (en) Textile quality detection method based on target detection
CN113936001B (en) Textile surface flaw detection method based on image processing technology
Zhu et al. Yarn-dyed fabric defect detection based on autocorrelation function and GLCM
CN107437243B (en) Tire impurity detection method and device based on X-ray image
CN108364291A (en) Grey cloth rapid detection method based on computer vision technique
CN114723704B (en) Textile quality evaluation method based on image processing
CN110889838A (en) Fabric defect detection method and device
CN109509171A (en) A kind of Fabric Defects Inspection detection method based on GMM and image pyramid
CN105631481B (en) Iron based on genetic programming composes abrasive grain compound characteristics building method
CN114998321A (en) Textile material surface hairiness degree identification method based on optical means
CN110458809B (en) Yarn evenness detection method based on sub-pixel edge detection
Pereira et al. Computer vision techniques for detecting yarn defects
CN117541557A (en) Textile jacquard defect detection method and system based on deep learning
CN115266732A (en) Carbon fiber tow defect detection method based on machine vision
CN111402225A (en) Cloth folding false detection defect discrimination method
CN114782426B (en) Knitted fabric broken yarn defect detection method based on artificial intelligence system
Jeffrey Kuo et al. Self-organizing map network for automatically recognizing color texture fabric nature
CN113610843A (en) Real-time defect identification system and method for optical fiber braided layer
Niles et al. A system for analysis, categorisation and grading of fabric defects using computer vision
Kopaczka et al. Fully Automatic Faulty Weft Thread Detection using a Camera System and Feature-based Pattern Recognition.
Ramakrishnan et al. A Novel Fabric Defect Detection Network in textile fabrics based on DLT
CN115100481B (en) Textile qualitative classification method based on artificial intelligence
Rauf et al. Fabric weave pattern recognition and classification by machine learning
CN117745730B (en) Polyester filament yarn detection method and system based on image processing technology

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
PE01 Entry into force of the registration of the contract for pledge of patent right

Denomination of invention: A Method for Detecting Surface Flaw Points on Textiles Based on Image Processing Technology

Effective date of registration: 20230330

Granted publication date: 20220304

Pledgee: Hangzhou joint rural commercial bank Limited by Share Ltd. three pier sub branch

Pledgor: Hangzhou youhuasijie culture and Art Development Co.,Ltd.

Registration number: Y2023980036481

PE01 Entry into force of the registration of the contract for pledge of patent right