CN111027577B - Fabric abnormal texture type identification method and device - Google Patents
Fabric abnormal texture type identification method and device Download PDFInfo
- Publication number
- CN111027577B CN111027577B CN201911108615.5A CN201911108615A CN111027577B CN 111027577 B CN111027577 B CN 111027577B CN 201911108615 A CN201911108615 A CN 201911108615A CN 111027577 B CN111027577 B CN 111027577B
- Authority
- CN
- China
- Prior art keywords
- image
- abnormal texture
- texture
- abnormal
- level
- 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
Links
- 230000002159 abnormal effect Effects 0.000 title claims abstract description 178
- 239000004744 fabric Substances 0.000 title claims abstract description 67
- 238000000034 method Methods 0.000 title claims abstract description 25
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 61
- 230000004927 fusion Effects 0.000 claims abstract description 55
- 230000009466 transformation Effects 0.000 claims abstract description 25
- 230000006870 function Effects 0.000 claims description 37
- 238000004590 computer program Methods 0.000 claims description 12
- 238000013145 classification model Methods 0.000 claims description 9
- 238000010801 machine learning Methods 0.000 claims description 9
- 238000000844 transformation Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 238000012706 support-vector machine Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 5
- 238000005259 measurement Methods 0.000 claims description 5
- 230000001737 promoting effect Effects 0.000 abstract 1
- 238000004422 calculation algorithm Methods 0.000 description 7
- 238000004891 communication Methods 0.000 description 6
- 230000007547 defect Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 230000004807 localization Effects 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 230000001131 transforming effect Effects 0.000 description 2
- 241001388119 Anisotremus surinamensis Species 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000000877 morphologic effect Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
- 230000037303 wrinkles Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/42—Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation
- G06V10/431—Frequency domain transformation; Autocorrelation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/40—Analysis of texture
-
- 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
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of automatic classification of fabrics, and provides a method and a device for identifying abnormal texture types of fabrics, wherein the method comprises the following steps: performing quality enhancement on the original abnormal texture image through a preset image enhancement model to generate an enhanced abnormal texture image; performing multi-level and multi-directional decomposition on the enhanced abnormal texture image through a preset dual-tree complex wavelet decomposition model to obtain a frequency domain texture feature set; extracting 6 multiplied by N high-frequency abnormal texture sub-images from the frequency domain texture feature set, fusing the 6 high-frequency abnormal texture sub-images of each level into high-frequency abnormal texture fused images of corresponding levels through a preset dual-tree complex wavelet fusion model, and identifying the abnormal texture types of the fabric according to all the high-frequency abnormal texture fused images, thereby promoting the self-adaption and stable transformation in different levels and different directions to obtain the frequency domain texture feature set, and well supporting the accurate and efficient identification of the abnormal texture types of the fabric according to the frequency domain texture feature set.
Description
Technical Field
The invention relates to the technical field of automatic classification of fabrics, in particular to a method and a device for identifying abnormal texture types of fabrics.
Background
With the development and wide application of image identification technology, the fabric abnormal texture type is automatically identified according to the image, and compared with a manual mode, the automatic identification degree and identification efficiency of the fabric abnormal texture type are greatly improved, for example: the fabric has abnormal texture types such as flaw, pompon, fold, structural damage and the like.
The existing method for automatically identifying the abnormal texture types of the fabrics according to images mainly comprises two types: firstly, an airspace feature set is extracted from an original abnormal texture image to which a fabric belongs through an airspace-based image recognition model, and the abnormal texture type of the fabric is recognized according to the airspace feature set, wherein the airspace-based image recognition model can adopt a gray level co-occurrence matrix algorithm, a digital morphology algorithm, a Markov random field algorithm and the like, and the airspace feature set can comprise airspace texture features, morphological features, pixel gray level features and the like; secondly, performing space-domain to frequency-domain transformation on an original abnormal texture image to which the fabric belongs through an image identification model based on space-frequency domain transformation to obtain a transformed abnormal texture image, and extracting a frequency domain texture feature set from the transformed abnormal texture image.
However, frequency domain texture features are distributed in different levels and different directions of an abnormal texture image to which the fabric belongs, and the conventional image identification model is difficult to adaptively and stably extract the frequency domain texture features in different levels and different directions, so that the efficiency of stably identifying the abnormal texture type of the fabric according to the frequency domain texture feature set is limited.
Disclosure of Invention
The invention provides a method and a device for identifying fabric abnormal texture types, intelligent equipment and a computer readable storage medium, aiming at the problems that the existing image identification model is difficult to adaptively and stably extract frequency domain texture features in different levels and different directions and the efficiency of stably identifying the fabric abnormal texture types according to a frequency domain texture feature set is restricted.
The invention provides a method for identifying the type of an abnormal texture of a fabric, which comprises the following steps:
obtaining an original abnormal texture image to which a fabric belongs;
performing quality enhancement processing on the original abnormal texture image through a preset image enhancement model to obtain an enhanced abnormal texture image;
respectively carrying out 6 directional decomposition transformations on the enhanced abnormal texture image at each level of N levels through a preset dual-tree complex wavelet decomposition model, and obtaining a frequency domain texture feature set after the N levels of decomposition transformations are completed;
extracting 6 multiplied by N high-frequency abnormal texture subgraphs from the frequency domain texture feature set;
respectively carrying out fusion transformation on the 6 high-frequency abnormal texture subgraphs of each level through a preset dual-tree complex wavelet fusion model to obtain a high-frequency abnormal texture fusion image of the corresponding level;
and identifying the abnormal texture type of the fabric according to all the high-frequency abnormal texture fusion images of the N levels.
The invention provides a fabric abnormal texture type identification device in a second aspect, which comprises:
the acquiring module is used for acquiring an original abnormal texture image to which the fabric belongs;
the enhancement module is used for carrying out quality enhancement processing on the original abnormal texture image through a preset image enhancement model to obtain an enhanced abnormal texture image;
the decomposition module is used for respectively carrying out 6 directional decomposition transformations on the enhanced abnormal texture image at each level of N levels through a preset dual-tree complex wavelet decomposition model, and obtaining a frequency domain texture feature set after the N levels of decomposition transformations are completed;
the extraction module is used for extracting 6 multiplied by N high-frequency abnormal texture subgraphs from the frequency domain texture feature set;
the fusion module is used for respectively carrying out fusion transformation on the 6 high-frequency abnormal texture subgraphs of each level through a preset dual-tree complex wavelet fusion model to obtain a high-frequency abnormal texture fusion image of the corresponding level;
and the identification module is used for identifying the abnormal texture type of the fabric according to all the high-frequency abnormal texture fusion images of the N levels.
A third aspect of the present invention provides an intelligent device, comprising: a memory and a processor coupled with the memory, the memory being configured to store a computer program, the processor being configured to load and execute the computer program, the computer program being executed by the processor to implement the operational steps performed by the fabric abnormal texture type identification method according to the first aspect.
A fourth aspect of the present invention provides a computer-readable storage medium configured to be couplable to a smart device and to store at least one instruction or at least one program or set of codes or set of instructions, which is loaded and executed by the smart device to implement the operation steps performed by the fabric abnormal texture type identification method according to the first aspect.
The fabric abnormal texture type identification method, the fabric abnormal texture type identification device, the intelligent equipment and the computer readable storage medium have the advantages that: compared with the original abnormal texture image, the definition of the enhanced abnormal texture image is enhanced through the preset image enhancement model, the abnormal texture image is enhanced through the preset dual-tree complex wavelet decomposition model multi-level multi-directional decomposition and transformation, the self-adaption rapid, accurate and stable extraction of the frequency domain texture feature set is promoted by means of the maintenance of the frequency domain localization characteristic and the invariance of extracting the frequency domain texture feature in multi-level multi-direction of the dual-tree complex wavelet decomposition model, the 6 high-frequency abnormal texture sub-images under each level are fused through the preset dual-tree complex wavelet fusion model, the preset dual-tree complex wavelet fusion model fusion transformation image belongs to the inverse transformation associated with the process of decomposing and transforming the image through the preset dual-tree complex wavelet decomposition model, the self-adaption rapid, accurate and stable generation of the multi-level high-frequency abnormal texture fusion image is promoted, the defect that the existing image identification model is difficult to adaptively and stably extract the frequency domain texture feature in different levels and different directions is overcome, the accurate and efficient identification of the abnormal texture type of the fabric according to the frequency domain texture feature set is well supported, and the identification difficulty and the identification efficiency of the abnormal texture type of the fabric are well balanced.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a schematic flow chart of a method for identifying the type of an abnormal texture in a fabric according to the present invention;
FIG. 2 is a schematic diagram of a fabric abnormal texture type identification system according to the present invention;
fig. 3 is a schematic structural diagram of a fabric abnormal texture type identification device provided by the invention.
Detailed Description
The principles and features of this invention are described below in conjunction with the following drawings, which are set forth by way of illustration only and are not intended to limit the scope of the invention.
Example one
As shown in fig. 1, a method for identifying the type of an abnormal texture of a fabric includes:
step 10, obtaining an original abnormal texture image of a fabric;
step 20, performing quality enhancement processing on the original abnormal texture image through a preset image enhancement model to obtain an enhanced abnormal texture image;
step 30, respectively carrying out 6 directional decomposition transformations on the enhanced abnormal texture image at each level of N levels through a preset dual-tree complex wavelet decomposition model, and obtaining a frequency domain texture feature set after the N levels of decomposition transformations are completed, wherein the frequency domain texture feature set comprises 2 low-frequency basic texture sub-images and 6 multiplied by N high-frequency abnormal texture sub-images;
step 40, extracting 6 multiplied by N high-frequency abnormal texture subgraphs from the frequency domain texture feature set;
step 50, respectively carrying out fusion transformation on the 6 high-frequency abnormal texture subgraphs of each level through a preset dual-tree complex wavelet fusion model to obtain a high-frequency abnormal texture fusion image of the corresponding level;
and step 60, identifying the abnormal texture type according to all the high-frequency abnormal texture fusion images.
As shown in fig. 2, the fabric abnormal texture type identification system comprises a lamp box 1, a communication cable 7 and an intelligent device 8, a light source 2 is arranged in the lamp box 1, a CCD camera 3, a support 4, a fabric 5 to be detected and an object placing plate 6 are arranged on a bottom plate of the lamp box 1, the object placing plate 6 is arranged in a shooting range of the CCD camera 3, the fabric 5 to be detected is arranged on a top surface of the object placing plate 6, a bottom end of the support 4 is fixed on the bottom plate of the lamp box 1, the CCD camera 3 is fixed on a top end of the support 4 and is positioned above the fabric 5 to be detected, the CCD camera 3 is used for shooting the fabric 5 to be detected to obtain an original abnormal texture image to which the fabric 5 to be detected belongs, the light source 2 is fixed on a top plate of the lamp box 1, the light source 2 is used for supplementing light to the CCD camera 3, so that the original abnormal texture image has good definition, two ends of the communication cable 7 are respectively in communication connection with the CCD camera 3 and the intelligent device 8, so that the CCD camera 3 inputs the original abnormal texture image to the communication cable 7, the intelligent device 8 positioned outside the lamp box 1, and stores the original abnormal texture image.
Taking the intelligent device 8 as a desktop computer as an example, the desktop computer may be pre-installed with a computer program with a fabric texture anomaly category identification function, the computer program may be provided with a preset image enhancement model, a preset dual-tree complex wavelet decomposition model and a preset dual-tree complex wavelet fusion model, the preset image enhancement model may be pre-constructed by using an image gray scale measurement algorithm, an image denoising algorithm, an image definition enhancement algorithm and the like, and the preset dual-tree complex wavelet decomposition model and the preset dual-tree complex wavelet fusion model may be pre-constructed by using a dual-tree complex wavelet algorithm, respectively.
When the computer program is run by a processor in a desktop computer, the steps 10-60 are executed step by step according to a time sequence, the step 10 executes the reading of an original abnormal texture image from a storage of the desktop computer, the step 20 executes the preprocessing processes of graying, denoising, sharpening and the like on the original abnormal texture image, compared with the original abnormal texture image, the sharpening of the abnormal texture image is enhanced through a preset image enhancement model, the abnormal texture image is enhanced through the preset dual-tree complex wavelet decomposition model multi-level multi-directional decomposition and transformation, the frequency domain localization characteristic and the invariance of extracting the frequency domain texture characteristic in multi-level multi-direction are kept by means of the dual-tree complex wavelet, the self-adaptive fast, accurate and stable extraction of the frequency domain texture characteristic set are promoted, the 6-amplitude high-frequency abnormal subgraph texture under each level is fused through the preset dual-tree wavelet fusion model, the preset dual-tree complex wavelet fusion model fusion transformation image belongs to the inverse transformation associated with the process of decomposing and transforming the preset dual-tree complex wavelet decomposition model, the self-adaptive fast, the accurate and stable generation of the multi-level high-frequency abnormal texture characteristic fusion image is promoted, the defects that the self-adaptive fast, the high-frequency abnormal texture characteristic fusion image identification model is difficult to adaptively and the identification of multiple levels is more accurate and the identification of the frequency domain texture identification of the abnormal texture characteristic sets in different directions are overcome, and the defects of the existing dual-frequency identification of the high-frequency fabric identification in different levels are more accurate and the identification of the high-frequency identification efficiency identification of the abnormal texture types are more good identification.
The preset dual-tree complex wavelet decomposition model is provided with a decomposition function, the decomposition function is expressed by a first formula, and the first formula specifically comprises the following steps:
wherein f represents a decomposition function, M represents the total number of pixel points to which the enhanced abnormal texture image belongs, and A N A first decomposition coefficient, phi (x), corresponding to the total layer number N m ,y m ) Representing the scale function corresponding to the mth pixel point to which the enhanced abnormal texture image belongs, wherein M is more than or equal to 1 and less than or equal to M, x m Line coordinate, y, of the mth pixel point on the enhanced abnormal texture image m To representAnd the m-th pixel point is used for enhancing the column coordinate of the abnormal texture image.
Wherein the content of the first and second substances,a second decomposition coefficient representing the kth direction of the nth level to which the mth pixel belongs, N is more than or equal to 1 and less than or equal to N, the kth direction is epsilon (+ 15 degrees, -45 degrees, -75 degrees), k is less than or equal to 6 and/or is greater than or equal to N>Represents a second decomposition coefficient->The corresponding wavelet function.
The abnormal texture image is decomposed and enhanced through the decomposition function presented by the first formula, and the accuracy, reliability and efficiency of the decomposition image are ensured.
Step 60 specifically includes:
step 61, respectively carrying out gray value summation calculation on the high-frequency abnormal texture fusion image of each level through a preset image energy measuring and calculating model to obtain an image energy value of a corresponding level;
and step 62, performing classification prediction on all image energy values of the N levels through a pre-trained machine learning classification model to obtain a classification prediction result for pointing to an abnormal texture type.
A preset image energy estimation model and a machine learning classification model trained in advance, which is a support vector machine model such as: the machine learning classification model is provided with a linear support vector machine function, and the classification prediction result can comprise classification probability values, such as: the classification probability is the probability value output by the linear support vector machine function.
Step 60 specifically includes: and step 63, inquiring the abnormal texture type matched with the classification probability value from the preset texture abnormal classification table.
The preset texture abnormal classification table can be obtained by associating multiple abnormal texture types and multiple classification probability ranges one by one in advance and storing the abnormal texture types in the desktop computer, wherein the classification probability values are in the classification probability range associated with the abnormal texture types, the multiple abnormal texture types can comprise a text information type for describing that a flaw point appears on a fabric texture, a text information type for describing that a ball appears on the fabric texture, a text information type for describing that a wrinkle appears on the fabric texture and a text information type for describing that the fabric texture is structurally damaged, and the method supports the self-adaption, high speed and stable identification of the abnormal texture types according to multiple high-frequency abnormal texture fusion images.
The preset image energy measurement and calculation model is provided with a pixel energy level summation function, the pixel energy level summation function is expressed by a second formula, and the second formula specifically comprises the following steps:
wherein E represents a pixel energy level summation function, N represents the total level number, P represents the total number of pixel points of the high-frequency abnormal texture fused image of the nth level,representing the gray value of the p-th pixel point in 6 directions of the high-frequency abnormal texture fused image of the nth level, wherein N is more than or equal to 1 and less than or equal to N, x P Expressing the line coordinate, y, of the p-th pixel point on the n-th high-frequency abnormal texture fusion image P And expressing the column coordinates of the p-th pixel point on the high-frequency abnormal texture fusion image of the n-th level.
The accuracy, reliability and efficiency of measuring the energy value are ensured by measuring the energy value of each high-frequency abnormal texture fusion image through the pixel energy level summation function presented by the second formula.
Example two
As shown in fig. 2, a device for identifying the type of an abnormal texture of a fabric comprises an obtaining module, a judging module and a judging module, wherein the obtaining module is used for obtaining an original abnormal texture image to which the fabric belongs; the enhancement module is used for carrying out quality enhancement processing on the original abnormal texture image through a preset image enhancement model to obtain an enhanced abnormal texture image; the decomposition module is used for respectively carrying out 6-direction decomposition transformation on the enhanced abnormal texture image at each level of the N levels through a preset dual-tree complex wavelet decomposition model, and obtaining a frequency domain texture feature set after the N levels are decomposed and transformed; the extraction module is used for extracting 6 multiplied by N high-frequency abnormal texture subgraphs from the frequency domain texture feature set; the fusion module is used for respectively carrying out fusion transformation on the 6 high-frequency abnormal texture subgraphs of each level through a preset dual-tree complex wavelet fusion model to obtain a high-frequency abnormal texture fusion image of the corresponding level; and the identification module is used for identifying the abnormal texture type of the fabric according to all the high-frequency abnormal texture fusion images of the N levels.
The preset dual-tree complex wavelet decomposition model is provided with a decomposition function, the decomposition function is expressed by a first formula, and the first formula specifically comprises the following steps:
wherein f represents a decomposition function, M represents the total number of pixel points to which the enhanced abnormal texture image belongs, and A N A first decomposition coefficient, phi (x), corresponding to the total layer number N m ,y m ) Representing the scale function corresponding to the mth pixel point to which the enhanced abnormal texture image belongs, wherein M is more than or equal to 1 and less than or equal to M, x m Line coordinate, y, of the mth pixel point on the enhanced abnormal texture image m And the column coordinates of the mth pixel point on the enhanced abnormal texture image are represented.
Wherein the content of the first and second substances,a second decomposition coefficient representing the kth direction of the nth level to which the mth pixel belongs, N is more than or equal to 1 and less than or equal to N, the kth direction is epsilon (+ 15 degrees, -45 degrees, -75 degrees), k is less than or equal to 6 and/or is greater than or equal to N>Represents a second decomposition coefficient->The corresponding wavelet function.
The identification module is specifically configured to: respectively carrying out gray value summation calculation on the nth high-frequency abnormal texture fusion image through a preset image energy measuring and calculating model to obtain an image energy value of a corresponding level; and carrying out classification prediction on all image energy values of the N levels through a machine learning classification model to obtain a classification prediction result pointing to the abnormal texture type, wherein the machine learning classification model is a support vector machine model.
The preset image energy measurement and calculation model is provided with a pixel energy level summation function, the pixel energy level summation function is expressed by a second formula, and the second formula specifically comprises the following steps:
wherein E represents a pixel energy level summation function, N represents the total level number, P represents the total number of pixel points of the high-frequency abnormal texture fused image of the nth level,expressing the gray value of the p pixel point in 6 directions of the N-level high-frequency abnormal texture fusion image, wherein N is more than or equal to 1 and less than or equal to N, x P Expressing the line coordinate, y, of the p-th pixel point on the n-th high-frequency abnormal texture fusion image P And expressing the column coordinates of the p-th pixel point on the high-frequency abnormal texture fusion image of the n-th level.
EXAMPLE III
A smart device, comprising: the fabric abnormal texture type identification method comprises a memory and a processor coupled with the memory, wherein the memory is configured to store a computer program, the processor is configured to load and execute the computer program, and the computer program is executed by the processor to realize any operation steps executed by the fabric abnormal texture type identification method in the embodiment one.
The memory and the processor can be respectively and electrically connected on a communication control bus, so that the memory is coupled with the processor through the communication control bus, and the intelligent device is a mobile device such as a mobile phone or an ipad or a notebook computer or a wearable device.
Example four
A computer-readable storage medium configured to be couplable to a smart device and to store at least one instruction or at least one program or set of codes or set of instructions, which is loaded and executed by the smart device to implement any one of the operation steps performed by the fabric abnormal texture type identification method according to an embodiment.
The computer readable storage medium such as flash memory or a removable hard disk or a flash disk or an optical disk may be driven by an intelligent device such as an industrial personal computer or a server.
The reader should understand that in the description of this specification, reference to the description of the terms "aspect," "embodiment," and "exemplary" etc., means that a particular feature, step, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention, and the terms "first" and "second," etc., 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, whereby the features defined as "first" and "second," etc., may explicitly or implicitly include at least one such feature.
In this specification, where the terminology above is used for the purpose of describing particular features, steps or characteristics in general, it is not necessary for the terminology above to be restricted to the same embodiments or examples, and the described particular features, steps or characteristics may be combined in any suitable manner in one or more particular examples or examples, or those skilled in the art may combine or/and combine features of different embodiments or examples described in this specification and other features of different embodiments or examples without conflict with one another.
Claims (8)
1. A fabric abnormal texture type identification method is characterized by comprising the following steps:
acquiring an original abnormal texture image to which a fabric belongs;
performing quality enhancement processing on the original abnormal texture image through a preset image enhancement model to obtain an enhanced abnormal texture image;
respectively carrying out 6 directional decomposition transformations on the enhanced abnormal texture image at each level of N levels through a preset dual-tree complex wavelet decomposition model, and obtaining a frequency domain texture feature set after the N levels of decomposition transformations are completed;
extracting 6 multiplied by N high-frequency abnormal texture subgraphs from the frequency domain texture feature set;
respectively carrying out fusion transformation on the 6 high-frequency abnormal texture subgraphs of each level through a preset dual-tree complex wavelet fusion model to obtain a high-frequency abnormal texture fusion image of the corresponding level;
identifying the abnormal texture type of the fabric according to all the high-frequency abnormal texture fusion images of the N levels;
the identifying of the abnormal texture type to which the fabric belongs according to all the high-frequency abnormal texture fusion images of the N levels specifically includes:
carrying out gray value summation calculation on the high-frequency abnormal texture fusion image of each level through a preset image energy measuring and calculating model to obtain an image energy value of the corresponding level;
classifying and predicting all image energy values of N levels through a machine learning classification model trained in advance to obtain a classification prediction result for pointing to the abnormal texture type;
the preset image energy measurement model is provided with a pixel energy level summation function, and the pixel energy level summation function is expressed as:
wherein E represents the pixel energy level summation function, N represents the total level number, P represents the total number of pixel points of the high-frequency abnormal texture fused image of the nth level,representing the gray value of the p-th pixel point in 6 directions of the high-frequency abnormal texture fused image of the nth level, wherein N is more than or equal to 1 and less than or equal to N, x P Expressing the line coordinate, y, of the p-th pixel point on the n-th high-frequency abnormal texture fusion image P And expressing the column coordinates of the p-th pixel point on the high-frequency abnormal texture fusion image of the n-th level.
2. The method for identifying the fabric abnormal texture type according to claim 1, wherein the preset dual-tree complex wavelet decomposition model is provided with a decomposition function, and the decomposition function is expressed as:
wherein f represents the decomposition function, M represents the total number of pixel points to which the enhanced abnormal texture image belongs, and A N A first decomposition coefficient, phi (x), corresponding to the total layer number N m ,y m ) Representing the scale function corresponding to the mth pixel point to which the enhanced abnormal texture image belongs, wherein M is more than or equal to 1 and less than or equal to M, and x m The line coordinate, y, of the mth pixel point on the enhanced abnormal texture image m The column coordinates of the mth pixel point on the enhanced abnormal texture image are represented,a second decomposition coefficient which represents the kth direction of the nth level to which the mth pixel belongs, N is more than or equal to 1 and less than or equal to N, the kth direction is epsilon and is (+ 15 degrees, -45 degrees, -75 degrees), k is less than or equal to 6 and is greater than or equal to>Represents said second decomposition coefficient>The corresponding wavelet function.
3. The method for identifying fabric abnormal texture types according to claim 1, wherein the machine learning classification model is a support vector machine model.
4. An abnormal texture type identification device for fabric, comprising:
the acquiring module is used for acquiring an original abnormal texture image to which the fabric belongs;
the enhancement module is used for carrying out quality enhancement processing on the original abnormal texture image through a preset image enhancement model to obtain an enhanced abnormal texture image;
the decomposition module is used for respectively carrying out 6-direction decomposition transformation on the enhanced abnormal texture image at each level of N levels through a preset dual-tree complex wavelet decomposition model, and obtaining a frequency domain texture feature set after the N levels are decomposed and transformed;
the extraction module is used for extracting 6 multiplied by N high-frequency abnormal texture subgraphs from the frequency domain texture feature set;
the fusion module is used for respectively carrying out fusion transformation on the 6 high-frequency abnormal texture subgraphs of each level through a preset dual-tree complex wavelet fusion model to obtain a high-frequency abnormal texture fusion image of the corresponding level;
the identification module is used for identifying the abnormal texture type of the fabric according to all the high-frequency abnormal texture fusion images of the N levels;
the identification module is specifically configured to:
carrying out gray value summation calculation on the high-frequency abnormal texture fusion image of each level through a preset image energy measuring and calculating model to obtain an image energy value of the corresponding level;
classifying and predicting all energy values of N levels through a machine learning classification model trained in advance to obtain a classification prediction result for pointing to the abnormal texture type;
the identification module is further configured to:
the preset image energy measurement model is provided with a pixel energy level summation function, and the pixel energy level summation function is expressed as:
wherein E represents the pixel energy level summation function, N represents the total level number, P represents the total number of pixel points of the high-frequency abnormal texture fused image of the nth level,representing the gray value of the p-th pixel point in 6 directions of the high-frequency abnormal texture fused image of the nth level, wherein N is more than or equal to 1 and less than or equal to N, x P The line coordinate, y, of the p pixel point on the n-level high-frequency abnormal texture fusion image is represented P And the column coordinates of the p pixel point on the n-th level high-frequency abnormal texture fusion image are represented.
5. The fabric abnormal texture type recognition device according to claim 4, wherein the decomposition module is further configured to:
the preset dual-tree complex wavelet decomposition model is provided with a decomposition function, and the decomposition function is expressed as:
wherein f represents the decomposition function, M represents the total number of pixel points to which the enhanced abnormal texture image belongs, and A N A first decomposition coefficient, phi (x), corresponding to the total layer number N m ,y m ) Representing the scale function corresponding to the mth pixel point to which the enhanced abnormal texture image belongs, wherein M is more than or equal to 1 and less than or equal to M, and x m The line coordinate, y, of the mth pixel point on the enhanced abnormal texture image m The column coordinate of the mth pixel point on the enhanced abnormal texture image is represented,a second decomposition coefficient representing the kth direction of the nth level to which the mth pixel belongs, N is more than or equal to 1 and less than or equal to N, the kth direction is epsilon (+ 15 degrees, -45 degrees, -75 degrees), k is less than or equal to 6 and/or is greater than or equal to N>Represents the second decomposition coefficient->The corresponding wavelet function.
6. The fabric abnormal texture type recognition device according to claim 4, wherein the machine learning classification model is a support vector machine model.
7. A smart device, comprising: a memory and a processor coupled with the memory, the memory configured to store a computer program, the processor configured to load and execute the computer program, the computer program being executed by the processor to implement the operational steps performed by the fabric abnormal texture type identification method according to any one of claims 1 to 3.
8. A computer-readable storage medium, configured to be couplable with a smart device and to store at least one instruction or at least one program or set of codes or set of instructions, which is loaded and executed by said smart device to implement the operating steps performed by the fabric abnormal texture type identification method according to any one of claims 1 to 3.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911108615.5A CN111027577B (en) | 2019-11-13 | 2019-11-13 | Fabric abnormal texture type identification method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911108615.5A CN111027577B (en) | 2019-11-13 | 2019-11-13 | Fabric abnormal texture type identification method and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111027577A CN111027577A (en) | 2020-04-17 |
CN111027577B true CN111027577B (en) | 2023-03-31 |
Family
ID=70201416
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201911108615.5A Active CN111027577B (en) | 2019-11-13 | 2019-11-13 | Fabric abnormal texture type identification method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111027577B (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113409288B (en) * | 2021-06-29 | 2023-06-27 | 北京百度网讯科技有限公司 | Image definition detection method, device, equipment and storage medium |
CN114693677B (en) * | 2022-05-31 | 2022-08-23 | 南通倍拓工贸有限公司 | Knitted fabric pad dyeing process abnormity detection method |
CN116740053B (en) * | 2023-08-08 | 2023-11-07 | 山东顺发重工有限公司 | Management system of intelligent forging processing production line |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105445022A (en) * | 2015-11-17 | 2016-03-30 | 中国矿业大学 | Planetary gear fault diagnosis method based on dual-tree complex wavelet transform-entropy feature fusion |
Family Cites Families (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
GB0427737D0 (en) * | 2004-12-17 | 2005-01-19 | Univ Cambridge Tech | Method of identifying features within a dataset |
CN103458772B (en) * | 2011-04-07 | 2017-10-31 | 香港中文大学 | Retinal images analysis method and device |
CN104949990B (en) * | 2013-11-01 | 2017-06-23 | 南通大学 | A kind of flaw online test method suitable for Woven textiles |
CN104268543B (en) * | 2014-10-04 | 2017-07-04 | 上海工程技术大学 | A kind of acquisition method of complex texture fabric Pilling Image |
CN105893971A (en) * | 2016-04-01 | 2016-08-24 | 上海理工大学 | Traffic signal lamp recognition method based on Gabor and sparse representation |
CN108629757A (en) * | 2018-05-08 | 2018-10-09 | 山东理工大学 | Image interfusion method based on complex shear wave conversion Yu depth convolutional neural networks |
CN109255748B (en) * | 2018-06-07 | 2023-04-28 | 上海出版印刷高等专科学校 | Digital watermark processing method and system based on double-tree complex wavelet |
-
2019
- 2019-11-13 CN CN201911108615.5A patent/CN111027577B/en active Active
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105445022A (en) * | 2015-11-17 | 2016-03-30 | 中国矿业大学 | Planetary gear fault diagnosis method based on dual-tree complex wavelet transform-entropy feature fusion |
Non-Patent Citations (1)
Title |
---|
融合频谱变换的板材纹理缺陷分类;苏耀文等;《东北林业大学学报》(第02期);全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN111027577A (en) | 2020-04-17 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111027577B (en) | Fabric abnormal texture type identification method and device | |
US11062453B2 (en) | Method and system for scene parsing and storage medium | |
CN106557778B (en) | General object detection method and device, data processing device and terminal equipment | |
CN110148130B (en) | Method and device for detecting part defects | |
US8108324B2 (en) | Forward feature selection for support vector machines | |
US20150030239A1 (en) | Training classifiers for deblurring images | |
US20200134382A1 (en) | Neural network training utilizing specialized loss functions | |
CN111753863A (en) | Image classification method and device, electronic equipment and storage medium | |
CN111400528B (en) | Image compression method, device, server and storage medium | |
JP7393472B2 (en) | Display scene recognition method, device, electronic device, storage medium and computer program | |
CN113850838A (en) | Ship voyage intention acquisition method and device, computer equipment and storage medium | |
CN113920296B (en) | Text recognition method and system based on comparative learning | |
CN114120349A (en) | Test paper identification method and system based on deep learning | |
CN115311469A (en) | Image labeling method, training method, image processing method and electronic equipment | |
CN111401309A (en) | CNN training and remote sensing image target identification method based on wavelet transformation | |
CN106611148B (en) | Image-based offline formula identification method and device | |
Han et al. | Circular array targets detection from remote sensing images based on saliency detection | |
CN116797533B (en) | Appearance defect detection method and system for power adapter | |
CN114255381B (en) | Training method of image recognition model, image recognition method, device and medium | |
CN114595741B (en) | High-dimensional data rapid dimension reduction method and system based on neighborhood relation | |
US11816909B2 (en) | Document clusterization using neural networks | |
US11715288B2 (en) | Optical character recognition using specialized confidence functions | |
CN114511862A (en) | Form identification method and device and electronic equipment | |
CN112418321A (en) | Identification method and device of sign image | |
CN108596167B (en) | Variable-length optical character fast recognition system, method and device |
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 |