CN106846313B - Workpiece surface defect detection method and device - Google Patents

Workpiece surface defect detection method and device Download PDF

Info

Publication number
CN106846313B
CN106846313B CN201710051079.4A CN201710051079A CN106846313B CN 106846313 B CN106846313 B CN 106846313B CN 201710051079 A CN201710051079 A CN 201710051079A CN 106846313 B CN106846313 B CN 106846313B
Authority
CN
China
Prior art keywords
image
workpiece
contour
processing
gabor
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
CN201710051079.4A
Other languages
Chinese (zh)
Other versions
CN106846313A (en
Inventor
陈新度
杨明江
吴磊
关日钊
赖火生
徐焯基
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.)
GUANGZHOU ZSROBOT INTELLIGENT EQUIPMENT Co.,Ltd.
Original Assignee
Guangdong University of Technology
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 Guangdong University of Technology filed Critical Guangdong University of Technology
Priority to CN201710051079.4A priority Critical patent/CN106846313B/en
Publication of CN106846313A publication Critical patent/CN106846313A/en
Application granted granted Critical
Publication of CN106846313B publication Critical patent/CN106846313B/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
    • G06T7/0008—Industrial image inspection checking presence/absence
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
    • 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
    • G06T2207/20036—Morphological image processing
    • 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
    • G06T2207/30164—Workpiece; Machine component

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Quality & Reliability (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Image Analysis (AREA)
  • Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
  • Image Processing (AREA)

Abstract

The invention discloses a method for detecting surface defects of a workpiece, which comprises the steps of converting an input image of the workpiece into a gray image, then carrying out binarization processing, then extracting the outline of the workpiece, extracting the gray image of the workpiece, filtering the gray image to remove strip-shaped textures with an included angle of 5 degrees between the vertical direction and the horizontal direction of the surface of the workpiece, carrying out edge sharpening processing, then carrying out self-adaptive binarization processing to obtain an outline image, carrying out outline detection processing by using a patch with a preset size to obtain a corresponding characteristic vector, bringing a pre-trained SVM model into a discrimination operation, and outputting the result of the discrimination operation, so that the image of the corresponding range of the workpiece can be accurately and quickly extracted, the operation amount of a subsequent processing process is greatly reduced, and meanwhile, interference can be avoided, the accuracy of discrimination is improved, and the robustness is also improved.

Description

Workpiece surface defect detection method and device
Technical Field
The invention relates to the field of workpiece quality detection, in particular to a method and a device for detecting surface defects of a workpiece.
Background
The continuous progress of the human society is closely connected with the use of new materials to some extent, and the introduction and use of bronze and steel for example greatly improve the productivity and the production efficiency of the human society, thereby having epoch-making significance. Since the 20 th century, the development of human society is being more and more deeply influenced by a polymer material as another innovation in the field of materials. Plastics are well-known polymer materials, have the advantages of small density, light weight, good insulating property, low dielectric loss, high chemical stability, good wear resistance and the like, are widely applied to various fields such as industry, agriculture, building, national defense advanced industry and the like which are closely related to the life of people, and have the growth rate which is leaping over the first of four industrial materials (plastics, steel, wood and cement), namely, steel is replaced by plastics and wood is replaced by plastics, so that the development trend of the material world in the present world is formed. In the field of plastic processing, the most common processing method is injection molding, which is characterized by being capable of processing plastic products with complicated shapes, precise sizes and compact textures. And the injection molding process has good adaptability to the processing of various plastics, has higher production capacity and is easy to realize automation. Today, the plastic industry is rapidly developing, and the injection molding machine applied to the injection molding process is important in both quantity and variety, so that the injection molding machine becomes one of the processing methods which are the fastest in growth and the largest in production quantity in the plastic processing and molding at present. However, in the production process of injection molded articles, defects of appearance, dimensional accuracy or functionality such as spots, pits, scratches, color difference, defects and the like are often generated on the surface of a workpiece due to interference of various factors such as poor injection molding conditions, and the defects are likely to greatly influence further processing of the articles and even the usability of the articles. The existing defect detection method of the injection molding product mainly depends on a manual method combining manual visual inspection and offline sampling analysis for detection. However, the manual detection method can only sort the products with obvious shapes and surface defects, and perform necessary intervention on production equipment when the number of the defective products is obviously increased, and eliminate faults by means of overhauling equipment, adjusting parameters and the like, and recover the production; moreover, because the manual detection is greatly influenced by the experience, psychological and physiological factors of the detection personnel, namely the manual detection is greatly influenced by the subjective factors of the detection personnel, the manual detection method has poor real-time performance and high false detection rate, and cannot perform quantitative description, so that the accuracy and reliability of the evaluation are influenced. In addition, in a dangerous working environment, manual detection cannot be performed.
Disclosure of Invention
The invention mainly aims to provide a workpiece surface defect detection method and a workpiece surface defect detection device, and aims to solve the technical problems that only products with obvious shapes and surface defects can be sorted by a manual detection method, and due to the fact that the manual detection method is greatly influenced by subjective factors of detection personnel, the real-time performance is poor, the false detection rate is high, quantitative description cannot be carried out, and therefore the accuracy and the reliability of evaluation are influenced.
In order to achieve the above object, the present invention provides a method for detecting and processing surface defects of a workpiece, comprising: inputting an image I of a workpiece with a surface having a vertical stripe pattern and a horizontal stripe pattern at an angle of 5 DEGIConversion into a grayscale image IGThen, binarization processing is carried out, thereby obtaining a binarized image I for separating the workpiece from the backgroundBWherein the surface of the workpiece has the stripe-shaped texture and the image I in the vertical directionIAre overlapped in the vertical direction;
from the binarized image IBIn which the contour F of the workpiece is extractediAccording to said profile FiFrom the grey scale image IGIn which a gray-scale image I of the workpiece is extractedGi；
By said gray-scale image I of said workpieceGiFiltering to filter out the bar-shaped textures in the vertical direction of the surface of the workpiece and the bar-shaped textures forming an included angle of 5 degrees with the horizontal direction, and then obtaining a filtered image BGi；
For the image BGiCarrying out edge sharpening processing to obtain an edge sharpened image EGiThen, the self-adaptive binarization processing is carried out to obtain a contour image DGi；
Using preset sizesTo the contour image DGiCarrying out contour detection processing to obtain a feature vector V corresponding to each patchijAnd all the workpiece feature vectors V are usedijNormalized and formed into a feature vector set ViWherein the subscript j is the feature vector in the feature vector set ViThe serial number within;
collecting the feature vectors ViOf each feature vector VijAnd bringing in a pre-trained SVM model to perform discrimination operation, and outputting a result of the discrimination operation, wherein the discrimination result of the SVM model comprises two discrimination result types of normal and defect.
Preferably, the surface to be input is provided with an image I of a workpiece with stripe textures in the vertical direction and stripe textures forming an included angle of 5 degrees with the horizontal directionIConversion into a grayscale image IGThen, binarization processing is carried out, thereby obtaining a binarized image I for separating the workpiece from the backgroundBIn the step (2), the binarization processing is an OTSU algorithm.
Preferably, said deriving from said binarized image IBIn which the contour F of the workpiece is extractediAccording to said profile FiFrom the grey scale image IGIn which a gray-scale image I of the workpiece is extractedGiFrom the binarized image IBIn which the contour F of the workpiece is extractediThe method used is an erosion algorithm. Preferably, said passing is through said grayscale image I of said workpieceGiFiltering to obtain filtered image B after the bar textures on the surface of the workpiece, wherein the bar textures on the vertical direction and the horizontal direction form an included angle of 5 degreesGiComprises the following steps:
gray scale image I of the workpieceGiCarrying out first Gabor filtering processing to obtain the first filtered image AGi＝IGi*g1；
For the image AGiCarrying out second Gabor filtering processing to obtain the graph after the second filteringImage BGi＝AGi*g2，
Wherein, the said x is convolution operation, the Gabor kernel function g of the first Gabor filtering1And the Gabor kernel function g of the second Gabor filtering2The formula of (c) is defined as:
Figure GDA0002079752590000031
the above-mentioned
Figure GDA0002079752590000032
The above-mentioned
Figure GDA0002079752590000033
λ is the wavelength of the Gabor kernel function, δ is the scale of the Gabor kernel function, θ is the suppression angle, and
Figure GDA0002079752590000034
for phase difference, the (x, y) is the image IGiAnd said image AGiWherein, the suppression angle θ of the corresponding pixel point is 90 ° when the first Gabor filtering process is performed, and the suppression angle θ of the corresponding pixel point is 5 ° when the second Gabor filtering process is performed.
Preferably, the contour image D is subjected to a patch with a preset sizeGiCarrying out contour detection processing to obtain a feature vector V corresponding to each patchijAnd all the workpiece feature vectors V are usedijNormalized and formed into a feature vector set ViIn the step (2), the feature vector VijIncluding the length of the contour, the width of the contour, the position coordinates (x, y) of the contour, and the gray level mean value of the contour, which are obtained through the contour detection processing, are 5 feature values.
The present invention further provides a workpiece surface defect detecting apparatus, comprising:
an image input preprocessing module for making the input surface have vertical stripesImage I of a workpiece with stripe texture at an angle of 5 degrees to the horizontalIConversion into a grayscale image IGThen, binarization processing is carried out, thereby obtaining a binarized image I for separating the workpiece from the backgroundBWherein the surface of the workpiece has the stripe-shaped texture and the image I in the vertical directionIAre overlapped in the vertical direction;
a workpiece image acquisition module for acquiring the binary image IBIn which the contour F of the workpiece is extractediAccording to said profile FiFrom the grey scale image IGIn which a gray-scale image I of the workpiece is extractedGi；
A filter processing module for processing the gray image I of the workpieceGiFiltering to filter out the bar-shaped textures in the vertical direction of the surface of the workpiece and the bar-shaped textures forming an included angle of 5 degrees with the horizontal direction, and then obtaining a filtered image BGi；
A contour extraction module for extracting a contour from the image BGiCarrying out edge sharpening processing to obtain an edge sharpened image EGiThen, the self-adaptive binarization processing is carried out to obtain a contour image DGi；
A feature extraction module for matching the contour image D with a patch of a preset sizeGiCarrying out contour detection processing to obtain a feature vector V corresponding to each patchijAnd all the workpiece feature vectors V are usedijNormalized and formed into a feature vector set ViWherein the subscript j is the feature vector in the feature vector set ViThe serial number within;
a discrimination output module for collecting the feature vectors ViOf each feature vector VijAnd bringing in a pre-trained SVM model to perform discrimination operation, and outputting a result of the discrimination operation, wherein the discrimination result of the SVM model comprises two discrimination result types of normal and defect.
Preferably, in the image input preprocessing module, the binarization processing is OTSAnd (4) a U algorithm. Preferably, in the workpiece image acquisition module, the binarized image I is obtained from the image data of the workpieceBIn which the contour F of the workpiece is extractediThe method used is an erosion algorithm.
Preferably, the filtering processing module includes:
a first-time Gabor filtering processing unit for processing gray image I of the workpieceGiCarrying out first Gabor filtering processing to obtain the first filtered image AGi＝IGi*g1；
A second Gabor filter processing unit for processing the image AGiCarrying out secondary Gabor filtering processing to obtain an image B after the secondary filteringGi＝AGi*g2，
Wherein, the said x is convolution operation, the Gabor kernel function g of the first Gabor filtering1And the Gabor kernel function g of the second Gabor filtering2The formula of (c) is defined as:
Figure GDA0002079752590000041
the above-mentioned
Figure GDA0002079752590000042
The above-mentioned
Figure GDA0002079752590000043
λ is the wavelength of the Gabor kernel function, δ is the scale of the Gabor kernel function, θ is the suppression angle, and
Figure GDA0002079752590000044
for phase difference, the (x, y) is the image IGiAnd said image AGiWherein, the suppression angle θ of the corresponding pixel point is 90 ° when the first Gabor filtering process is performed, and the suppression angle θ of the corresponding pixel point is 5 ° when the second Gabor filtering process is performed.
Preferably, among the feature extraction modules, the feature directionQuantity VijIncluding the length of the contour, the width of the contour, the position coordinates (x, y) of the contour, and the gray level mean value of the contour, which are obtained through the contour detection processing, are 5 feature values.
According to the invention, through the preprocessing, the image of the corresponding range of the workpiece can be accurately and quickly extracted, so that the calculation amount in the subsequent processing process can be greatly reduced, meanwhile, the interference of the image content outside the corresponding range of the workpiece on the judgment processing can be avoided, the judgment accuracy is improved, and meanwhile, the robustness is also improved. And through the filtering processing, the interference and the influence of the strip texture with the vertical direction and the strip texture with the 5-degree included angle with the horizontal direction on the subsequent characteristic extraction of the profile and the surface of the workpiece are filtered out. In addition, the image after the filtering processing is performed with edge sharpening and then is subjected to contour extraction, so that the reliability and robustness of contour feature extraction are greatly improved, and meanwhile, the contour feature is adopted as the feature parameter of the SVM discrimination model, and the method has the characteristics of simplicity in operation, high operation speed and good reliability.
And, different from the ordinary binarization processing method, the OTSU algorithm is called the maximum between-class variance algorithm, and the optimal threshold value capable of separating the two classes needs to be calculated firstly in the main processing process, so that the intra-class variance of the two classes is minimum, therefore, the OTSU algorithm has the characteristic of automatically acquiring the optimal classification threshold value according to the content of the image. Moreover, because the input workpiece image has simple content and single color or gray distribution, the adoption of the algorithm can avoid the interference of adverse factors such as very sensitive noise and target size, incapability of effectively dealing with the situation of complicated image content and the like, and has the characteristics of simple operation, high speed and capability of realizing real-time processing. Furthermore, in some circumstances, if a form of vectorization is used, the loop can be operated faster, and a method of multi-thread parallel processing can be easily adopted.
Secondly, since the Gabor kernel function is sensitive to the edge of the image, good direction selection and scale selection characteristics can be provided, and the Gabor kernel function is insensitive to illumination variation and can provide good adaptability to illumination variation. And, since the Gabor filtering method is very similar to the visual stimulus response of simple cells in the human visual system, it has good characteristics in extracting local spatial and frequency domain information of the target. In addition, by adopting a Gabor filtering mode, the method has the advantages of simple operation, easy understanding, easy parameter adjustment, reduced calculation complexity, reduced calculation amount and improved response speed.
Drawings
Fig. 1 is a hardware configuration diagram of an image input section that implements various embodiments of the present invention;
FIG. 2 is a schematic diagram of an external structure of a workpiece for implementing various embodiments of the present invention;
FIG. 3 is a schematic flow chart illustrating a method for detecting surface defects of a workpiece according to a first embodiment of the present invention;
FIG. 4 is a schematic flow chart illustrating a method for detecting surface defects of a workpiece according to a second embodiment of the present invention;
FIG. 5 is a functional block diagram of a first embodiment of an apparatus for detecting surface defects of a workpiece according to the present invention;
FIG. 6 is a functional block diagram of an apparatus for detecting surface defects of a workpiece according to a second embodiment of the present invention.
The implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
An apparatus arrangement implementing various embodiments of the present invention will now be described with reference to the accompanying drawings. In the following description, suffixes such as "module", "component", or "unit" used to denote elements are used only for facilitating the explanation of the present invention, and have no specific meaning in themselves. Thus, "module" and "component" may be used in a mixture.
The workpiece surface defect detection method and apparatus may be implemented in various forms. For example, the processing method and apparatus described in the present invention may include a mobile device apparatus and a fixed device apparatus. In the following, it is assumed that the terminal is a fixed equipment device. However, it will be understood by those skilled in the art that the configuration according to the embodiment of the present invention can be applied to a mobile type equipment device in addition to elements particularly used for fixing purposes.
Fig. 1 is a hardware configuration diagram of an image input section that implements various embodiments of the present invention. As shown in the figure, a camera 1 for image input is arranged above a Z-axis 2, and the height of the camera 1 can move up and down along the vertical direction of the Z-axis 2, so that the focal length and the size of the field range of the camera 1 can be adjusted; one end of the Z shaft 2 is connected with a Y shaft 3, two ends of the Y shaft 3 are respectively connected with two X shafts 4 which are arranged on two sides of the image input component frame in parallel, and the Y shaft 3 can move along the X shaft 4 direction, so that the view field range of the camera 1 can be adjusted. A light source 6 is arranged on the base of the image input part, and a plurality of ground glass rollers 5 are arranged in parallel on the upper part of the light source 6, so that a workpiece placing platform consisting of the ground glass rollers 5 is formed. The ground glass roller 5 is driven by a motor to synchronously rotate. When the driving motor of the ground glass roller 5 rotates for 45 degrees, the ground glass roller 5 correspondingly rotates for 45 degrees, and meanwhile, the camera 1 collects a picture and inputs the picture into the workpiece surface defect detection method and device for processing. Each workpiece needs to rotate 8 degrees and 45 degrees, and the pictures shot corresponding to each angle are sequentially input into the workpiece surface defect detection method and device for processing, so as to judge whether defects exist.
Fig. 2 is a schematic structural diagram of an appearance of a workpiece for implementing various embodiments of the present invention. As shown in the figure. The shape of the workpiece is similar to a cylinder shape and can roll for 360 degrees; the surface is light-transmitting and has stripe patterns in the vertical direction and stripe patterns forming an angle of 5 degrees with the horizontal direction.
Referring to fig. 3, fig. 3 is a schematic flow chart of a method for detecting surface defects of a workpiece according to a first embodiment of the present invention. In the embodiment shown in fig. 3, the method for detecting the surface defects of the workpiece comprises the following steps:
step S10, image input preprocessing.
An image I of a workpiece having a surface with stripe patterns in the vertical direction and stripe patterns at an angle of 5 degrees to the horizontal directionIConversion into a grayscale image IGThen, binarization processing is carried out, thereby obtaining a binarized image I for separating the workpiece from the backgroundBWherein the surface of the workpiece has the stripe-shaped texture and the image I in the vertical directionICoincide.
And step S20, acquiring a workpiece image.
I.e. from the binarized image IBIn which the contour F of the workpiece is extractediAccording to said profile FiFrom the grey scale image IGIn which a gray-scale image I of the workpiece is extractedGi。
And step S30, filtering processing.
I.e. by the grey-scale image I of the workpieceGiFiltering to filter out the bar-shaped textures in the vertical direction of the surface of the workpiece and the bar-shaped textures forming an included angle of 5 degrees with the horizontal direction, and then obtaining a filtered image BGi。
And step S40, extracting the contour.
Namely for the image BGiCarrying out edge sharpening processing to obtain an edge sharpened image EGiThen, the self-adaptive binarization processing is carried out to obtain a contour image DGi。
And step S50, feature extraction.
I.e. using a patch of a predetermined size to pair the contour image DGiCarrying out contour detection processing to obtain a feature vector V corresponding to each patchijAnd all the workpiece feature vectors V are usedijNormalized and formed into a feature vector set ViWherein the subscript j is the feature vector in the feature vector set ViThe serial number therein.
And step S60, judging and outputting.
I.e. the set of feature vectors ViEach of themFeature vector VijAnd bringing in a pre-trained SVM model to perform discrimination operation, and outputting a result of the discrimination operation, wherein the discrimination result of the SVM model comprises two discrimination result types of normal and defect.
Through the steps of S10 and S20, the image corresponding to the workpiece can be extracted accurately and quickly, so that the calculation amount in the subsequent processing process can be greatly reduced, the interference of the image content outside the corresponding range of the workpiece on the judgment processing can be avoided, the judgment accuracy is improved, and the robustness is also improved. And filtering out interference and influence of the contour of the workpiece and the strip texture with the surface having the vertical direction and the strip texture with the horizontal direction forming an included angle of 5 degrees on subsequent feature extraction through the filtering processing of the step S30. In addition, the image after the filtering processing is performed with edge sharpening and then is subjected to contour extraction, so that the reliability and robustness of contour feature extraction are greatly improved, and meanwhile, the contour feature is adopted as the feature parameter of the SVM discrimination model, and the method has the characteristics of simplicity in operation, high operation speed and good reliability.
Further, based on the above-described embodiment of fig. 3, in step S10, an image input preprocessing process is performed, namely, the image I of the workpiece whose surface to be input has the bar-shaped texture in the vertical direction and the bar-shaped texture at an angle of 5 degrees to the horizontal directionIConversion into a grayscale image IGThen, binarization processing is carried out, thereby obtaining a binarized image I for separating the workpiece from the backgroundBIn the step (2), the binarization processing is an OTSU algorithm. Maximum between-class variance
The OTSU algorithm is different from a common binarization processing method and is called an inter-class variance maximization algorithm, and the main processing process of the OTSU algorithm needs to calculate an optimal threshold value capable of separating two classes first so as to minimize the intra-class variance of the two classes, so that the OTSU algorithm has the characteristic of automatically acquiring the optimal classification threshold value according to the content of an image. Moreover, because the input workpiece image has simple content and single color or gray distribution, the adoption of the algorithm can avoid the interference of adverse factors such as very sensitive noise and target size, incapability of effectively dealing with the situation of complicated image content and the like, and has the characteristics of simple operation, high speed and capability of realizing real-time processing. Furthermore, in some circumstances, if a form of vectorization is used, the loop can be operated faster, and a method of multi-thread parallel processing can be easily adopted.
Further, based on the above-mentioned embodiment of fig. 3, in step S20, a workpiece image is obtained, i.e. the binary image I is obtainedBIn which the contour F of the workpiece is extractediAccording to said profile FiFrom the grey scale image IGIn which a gray-scale image I of the workpiece is extractedGiFrom the binarized image IBIn which the contour F of the workpiece is extractediThe method used is an erosion algorithm.
The erosion algorithm is a process of eliminating boundary points and shrinking boundaries inward, and can be used to eliminate small and meaningless objects, thereby effectively avoiding noise interference.
Referring to fig. 4, fig. 4 is a schematic flowchart of the step S30 of the method for detecting surface defects of a workpiece according to the second embodiment of the present invention. As shown in fig. 4, based on the above-mentioned embodiment of fig. 3, the step S30 of filtering includes:
and step S310, carrying out first Gabor filtering processing.
I.e. a grey scale image I of the workpieceGiCarrying out first Gabor filtering processing to obtain the first filtered image AGi＝IGi*g1。
And step S320, second Gabor filtering processing.
Namely to the image AGiCarrying out secondary Gabor filtering processing to obtain an image B after the secondary filteringGi＝AGi*g2。
Wherein, the said x is convolution operation, the Gabor kernel function g of the first Gabor filtering1And the Gabor kernel function g of the second Gabor filtering2The formula of (c) is defined as:
Figure GDA0002079752590000091
the above-mentioned
Figure GDA0002079752590000092
The above-mentioned
Figure GDA0002079752590000093
λ is the wavelength of the Gabor kernel function, δ is the scale of the Gabor kernel function, θ is the suppression angle, and
Figure GDA0002079752590000094
for phase difference, the (x, y) is the image IGiAnd said image AGi, the suppression angle θ of the corresponding pixel point is 90 ° when the first Gabor filtering process is performed, and the suppression angle θ of the corresponding pixel point is 5 ° when the second Gabor filtering process is performed.
The Gabor kernel function is sensitive to the edge of the image, so that good direction selection and scale selection characteristics can be provided, and the Gabor kernel function is insensitive to illumination change and can provide good adaptability to the illumination change. And, since the Gabor filtering method is very similar to the visual stimulus response of simple cells in the human visual system, it has good characteristics in extracting local spatial and frequency domain information of the target. In addition, by adopting a Gabor filtering mode, the method has the advantages of simple operation, easy understanding, easy parameter adjustment, reduced calculation complexity, reduced calculation amount and improved response speed.
Further, based on the above-mentioned embodiment of fig. 4, the S50, feature extraction process, that is, the process of using a preset size of patch to match the contour image DGiCarrying out contour detection processing to obtain a feature vector V corresponding to each patchijAnd all the workpiece feature vectors V are usedijNormalized and formed into a feature vector set ViIn the step (2), the feature vector VijIncluding the length of the contour obtained by the contour detection processing, the contourThe position coordinates (x, y) of the contour and the mean value of the gray levels of the contour are 5 eigenvalues.
The feature vector V is increased by using 5 feature values of the length of the contour, the width of the contour, the position coordinates (x, y) of the contour, and the gray level average of the contour obtained by the contour detection processingijThe latitude of the method enhances the robustness of the method, enables the SVM discrimination to be more mature and reliable, has high response speed and is easy to realize in engineering.
The workpiece surface defect detecting method in the first embodiment of the workpiece surface defect detecting method of the present invention described above can be realized by the workpiece surface defect detecting apparatus provided in the first embodiment of the workpiece surface defect detecting apparatus of the present invention.
Referring to fig. 5, fig. 5 is a diagram illustrating a workpiece surface defect detecting apparatus 100 according to a first embodiment of the present invention, wherein the workpiece surface defect detecting apparatus 100 includes:
an image input preprocessing module 10 for inputting an image I of a workpiece having a surface with stripe textures in a vertical direction and stripe textures at an angle of 5 degrees with respect to a horizontal directionIConversion into a grayscale image IGThen, binarization processing is carried out, thereby obtaining a binarized image I for separating the workpiece from the backgroundBWherein the surface of the workpiece has the stripe-shaped texture and the image I in the vertical directionICoincide.
A workpiece image acquisition module 20 for acquiring the binarized image I from the binarized imageBIn which the contour F of the workpiece is extractediAccording to said profile FiFrom the grey scale image IGIn which a gray-scale image I of the workpiece is extractedGi(ii) a A filter processing module for processing the gray image I of the workpieceGiFiltering to obtain filtered image B after the bar textures on the surface of the workpiece, wherein the bar textures on the vertical direction and the horizontal direction form an included angle of 5 degreesGi。
A filter processing module 30 for passing the ash of the workpieceDegree image IGiFiltering to filter out the bar-shaped textures in the vertical direction of the surface of the workpiece and the bar-shaped textures forming an included angle of 5 degrees with the horizontal direction, and then obtaining a filtered image BGi。
A contour extraction module 40 for extracting the contour of the image BGiCarrying out edge sharpening processing to obtain an edge sharpened image EGiThen, the self-adaptive binarization processing is carried out to obtain a contour image DGi。
A feature extraction module 50, configured to apply a preset size of patch to the contour image DGiCarrying out contour detection processing to obtain a feature vector V corresponding to each patchijAnd all the workpiece feature vectors V are usedijNormalized and formed into a feature vector set ViWherein the subscript j is the feature vector in the feature vector set ViThe serial number therein.
A judgment output module 60 for collecting the feature vectors ViOf each feature vector VijAnd bringing in a pre-trained SVM model to perform discrimination operation, and outputting a result of the discrimination operation, wherein the discrimination result of the SVM model comprises two discrimination result types of normal and defect.
Through the image input preprocessing module 10 and the workpiece image acquisition module 20, the image of the corresponding range of the workpiece can be accurately and quickly extracted, so that the calculation amount of the subsequent processing process can be greatly reduced, meanwhile, the interference of the image content outside the corresponding range of the workpiece on the judgment processing can be avoided, the judgment accuracy is improved, and meanwhile, the robustness is also improved. And the interference and influence of the contour of the workpiece and the strip texture with the surface having the vertical direction and the strip texture forming the 5-degree included angle with the horizontal direction on the subsequent feature extraction are filtered out through the filtering processing of the filtering processing module 30. In addition, the image after the filtering processing is performed with edge sharpening and then is subjected to contour extraction, so that the reliability and robustness of contour feature extraction are greatly improved, and meanwhile, the contour feature is adopted as the feature parameter of the SVM discrimination model, and the method has the characteristics of simplicity in operation, high operation speed and good reliability.
Further, based on the above-mentioned embodiment of fig. 5, the image I of the workpiece whose surface to be input has the bar-shaped texture in the vertical direction and the bar-shaped texture forming an angle of 5 degrees with the horizontal direction in the image input preprocessing module 10IConversion into a grayscale image IGThen, binarization processing is carried out, thereby obtaining a binarized image I for separating the workpiece from the backgroundBIn the processing procedure of (3), the binarization processing is an OTSU algorithm.
The OTSU algorithm is different from a common binarization processing method and is called an inter-class variance maximization algorithm, and the main processing process of the OTSU algorithm needs to calculate an optimal threshold value capable of separating two classes first so as to minimize the intra-class variance of the two classes, so that the OTSU algorithm has the characteristic of automatically acquiring the optimal classification threshold value according to the content of an image. Moreover, because the input workpiece image has simple content and single color or gray distribution, the adoption of the algorithm can avoid the interference of adverse factors such as very sensitive noise and target size, incapability of effectively dealing with the situation of complicated image content and the like, and has the characteristics of simple operation, high speed and capability of realizing real-time processing. Furthermore, in some circumstances, if a form of vectorization is used, the loop can be operated faster, and a method of multi-thread parallel processing can be easily adopted.
Further, based on the embodiment of fig. 5, the binarized image I is obtained from the workpiece image obtaining module 20BIn which the contour F of the workpiece is extractediAccording to said profile FiFrom the grey scale image IGIn which a gray-scale image I of the workpiece is extractedGiIn the processing of (1), the secondary binarized image IBIn which the contour F of the workpiece is extractediThe method used is an erosion algorithm.
The erosion algorithm is a process of eliminating boundary points and shrinking boundaries inward, and can be used to eliminate small and meaningless objects, thereby effectively avoiding noise interference.
The workpiece surface defect detecting method in the second embodiment of the workpiece surface defect detecting method of the present invention described above can be realized by the workpiece surface defect detecting apparatus provided in the second embodiment of the workpiece surface defect detecting apparatus of the present invention.
Referring to fig. 6, a second embodiment of the workpiece surface defect detecting apparatus according to the present invention provides a workpiece surface defect detecting apparatus, based on the embodiment shown in fig. 5, the filtering processing module 30 includes:
a first-time Gabor filter processing unit 31 for processing the gray image I of the workpieceGiCarrying out first Gabor filtering processing to obtain the first filtered image AGi＝IGi*g1；
A second Gabor filter processing unit 32 for processing the image AGiCarrying out secondary Gabor filtering processing to obtain an image B after the secondary filteringGi＝AGi*g2，
Wherein, the said x is convolution operation, the Gabor kernel function g of the first Gabor filtering1And the Gabor kernel function g of the second Gabor filtering2The formula of (c) is defined as:
Figure GDA0002079752590000121
the above-mentioned
Figure GDA0002079752590000122
The above-mentioned
Figure GDA0002079752590000123
λ is the wavelength of the Gabor kernel function, δ is the scale of the Gabor kernel function, θ is the suppression angle, and
Figure GDA0002079752590000124
for phase difference, the (x, y) is the image IGiAnd said image AGiWherein the coordinates of the corresponding pixels are the suppression angle during the first Gabor filtering processθ is 90 °, and the suppression angle θ at the time of the second Gabor filter processing is 5 °.
The Gabor kernel function is sensitive to the edge of the image, so that good direction selection and scale selection characteristics can be provided, and the Gabor kernel function is insensitive to illumination change and can provide good adaptability to the illumination change. And, since the Gabor filtering method is very similar to the visual stimulus response of simple cells in the human visual system, it has good characteristics in extracting local spatial and frequency domain information of the target. In addition, by adopting a Gabor filtering mode, the method has the advantages of simple operation, easy understanding, easy parameter adjustment, reduced calculation complexity, reduced calculation amount and improved response speed.
Further, based on the embodiment of fig. 6, the contour image D is subjected to the patch with the preset size in the feature extraction module 50GiCarrying out contour detection processing to obtain a feature vector V corresponding to each patchijAnd all the workpiece feature vectors V are usedijNormalized and formed into a feature vector set ViIn the processing of (2) the feature vector VijIncluding the length of the contour, the width of the contour, the position coordinates (x, y) of the contour, and the gray level mean value of the contour, which are obtained through the contour detection processing, are 5 feature values.
The feature vector V is increased by using 5 feature values of the length of the contour, the width of the contour, the position coordinates (x, y) of the contour, and the gray level average of the contour obtained by the contour detection processingijThe latitude of the method enhances the robustness of the method, enables the SVM discrimination to be more mature and reliable, has high response speed and is easy to realize in engineering.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.
It will be apparent to those skilled in the art that the above-described block units or steps of the present invention may be implemented by a general purpose computing device, or alternatively, they may be implemented by program code executable by a computing device, so that they may be stored in a storage device and executed by a computing device, and in some cases, the steps shown or described may be performed in a different order than here, or they may be separately fabricated into individual integrated circuit modules, or multiple ones of them may be fabricated into a single integrated circuit module. Thus, the present invention is not limited to any specific combination of hardware and software.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which is stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk) and includes instructions for enabling a terminal device (such as a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the method according to the embodiments of the present invention.
The above description is only a preferred embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes, which are made by using the contents of the present specification and the accompanying drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.

Claims (10)

1. A method for detecting surface defects of a workpiece, comprising:
inputting an image I of a workpiece with a surface having a vertical stripe pattern and a horizontal stripe pattern at an angle of 5 DEGIConversion into a grayscale image IGThen, binarization processing is carried out, thereby obtaining a binarized image I for separating the workpiece from the backgroundBWherein the surface of the workpiece has the stripe-shaped texture and the image I in the vertical directionIAre overlapped in the vertical direction;
from the binarized image IBIn which the contour F of the workpiece is extractediAccording to said profile FiFrom the grey scale image IGIn which a gray-scale image I of the workpiece is extractedGi；
By said gray-scale image I of said workpieceGiFiltering to filter out the bar-shaped textures in the vertical direction of the surface of the workpiece and the bar-shaped textures forming an included angle of 5 degrees with the horizontal direction, and then obtaining a filtered image BGi；
For the image BGiCarrying out edge sharpening processing to obtain an edge sharpened image EGiThen, the self-adaptive binarization processing is carried out to obtain a contour image DGi；
Using a patch with a preset size to pair the contour image DGiCarrying out contour detection processing to obtain a feature vector V corresponding to each patchijAnd all the workpiece feature vectors V are usedijNormalizing and forming a feature vector set ViWherein the subscript j is the feature vector in the feature vector set ViThe serial number within;
collecting the feature vectors ViOf each feature vector VijAnd bringing in a pre-trained SVM model to perform discrimination operation, and outputting a result of the discrimination operation, wherein the discrimination result of the SVM model comprises two discrimination result types of normal and defect.
2. As in claimThe method for detecting surface defects of a workpiece according to claim 1, wherein the image I of the workpiece having the strip-shaped texture in the vertical direction and the strip-shaped texture forming an angle of 5 degrees with the horizontal direction on the surface to be input isIConversion into a grayscale image IGThen, binarization processing is carried out, thereby obtaining a binarized image I for separating the workpiece from the backgroundBIn the step (2), the binarization processing is an OTSU algorithm.
3. The method for detecting surface defects of a workpiece as set forth in claim 1, wherein said binarized image I is obtained from said imageBIn which the contour F of the workpiece is extractediAccording to said profile FiFrom the grey scale image IGIn which a gray-scale image I of the workpiece is extractedGiFrom the binarized image IBIn which the contour F of the workpiece is extractediThe method used is an erosion algorithm.
4. The method of claim 1, wherein said passing of said grayscale image I of said workpieceGiFiltering to obtain filtered image B after the bar textures on the surface of the workpiece, wherein the bar textures on the vertical direction and the horizontal direction form an included angle of 5 degreesGiComprises the following steps:
gray scale image I of the workpieceGiCarrying out first Gabor filtering processing to obtain the first filtered image AGi＝IGi*g1；
For the image AGiCarrying out secondary Gabor filtering processing to obtain an image B after the secondary filteringGi＝AGi*g2，
Wherein, the said x is convolution operation, the Gabor kernel function g of the first Gabor filtering1And the Gabor kernel function g of the second Gabor filtering2The formula of (c) is defined as:
Figure FDA0002079752580000021
the above-mentioned
Figure FDA0002079752580000022
The above-mentioned
Figure FDA0002079752580000023
λ is the wavelength of the Gabor kernel function, δ is the scale of the Gabor kernel function, θ is the suppression angle, and
Figure FDA0002079752580000024
for phase difference, the (x, y) is the image IGiAnd said image AGiWherein, the suppression angle θ of the corresponding pixel point is 90 ° when the first Gabor filtering process is performed, and the suppression angle θ of the corresponding pixel point is 5 ° when the second Gabor filtering process is performed.
5. The method of claim 1, wherein the contour image D is processed by using a preset size of patchGiCarrying out contour detection processing to obtain a feature vector V corresponding to each patchijAnd all the workpiece feature vectors V are usedijNormalized and formed into a feature vector set ViIn the step (2), the feature vector VijIncluding the length of the contour, the width of the contour, the position coordinates (x, y) of the contour, and the gray level mean value of the contour, which are obtained through the contour detection processing, are 5 feature values.
6. An apparatus for detecting surface defects of a workpiece, comprising:
an image input preprocessing module for inputting an image I of a workpiece with a surface having a vertical bar-shaped texture and a 5-degree included angle with the horizontal bar-shaped textureIConversion into a grayscale image IGThen performing binarizationSo as to obtain a binarized image I separating the workpiece from the backgroundBWherein the surface of the workpiece has the stripe-shaped texture and the image I in the vertical directionIAre overlapped in the vertical direction;
a workpiece image acquisition module for acquiring the binary image IBIn which the contour F of the workpiece is extractediAccording to said profile FiFrom the grey scale image IGIn which a gray-scale image I of the workpiece is extractedGi；
A filter processing module for processing the gray image I of the workpieceGiFiltering to filter out the bar-shaped textures in the vertical direction of the surface of the workpiece and the bar-shaped textures forming an included angle of 5 degrees with the horizontal direction, and then obtaining a filtered image BGi；
A contour extraction module for extracting a contour from the image BGiCarrying out edge sharpening processing to obtain an edge sharpened image EGiThen, the self-adaptive binarization processing is carried out to obtain a contour image DGi；
A feature extraction module for matching the contour image D with a patch of a preset sizeGiCarrying out contour detection processing to obtain a feature vector V corresponding to each patchijAnd all the workpiece feature vectors V are usedijNormalizing and forming a feature vector set ViWherein the subscript j is the feature vector in the feature vector set ViThe serial number within;
a discrimination output module for collecting the feature vectors ViOf each feature vector VijAnd bringing in a pre-trained SVM model to perform discrimination operation, and outputting a result of the discrimination operation, wherein the discrimination result of the SVM model comprises two discrimination result types of normal and defect.
7. The apparatus for detecting surface defects of workpieces as set forth in claim 6, wherein said image input preprocessing module and said binarization processing are OTSU algorithm.
8. The apparatus for detecting surface defects of a workpiece as set forth in claim 7, wherein said binarized image I is obtained from said workpiece image obtaining moduleBIn which the contour F of the workpiece is extractediThe method used is an erosion algorithm.
9. The apparatus for detecting surface defects of a workpiece according to claim 6, wherein the filter processing module comprises:
a first-time Gabor filtering processing unit for processing gray image I of the workpieceGiCarrying out first Gabor filtering processing to obtain the first filtered image AGi＝IGi*g1；
A second Gabor filter processing unit for processing the image AGiCarrying out secondary Gabor filtering processing to obtain an image B after the secondary filteringGi＝AGi*g2，
Wherein, the said x is convolution operation, the Gabor kernel function g of the first Gabor filtering1And the Gabor kernel function g of the second Gabor filtering2The formula of (c) is defined as:
Figure FDA0002079752580000041
the above-mentioned
Figure FDA0002079752580000042
The above-mentioned
Figure FDA0002079752580000043
λ is the wavelength of the Gabor kernel function, δ is the scale of the Gabor kernel function, θ is the suppression angle, and
Figure FDA0002079752580000044
for phase difference, the (x, y) is the image IGiAnd said image AGiWherein, the suppression angle θ of the corresponding pixel point is 90 ° when the first Gabor filtering process is performed, and the suppression angle θ of the corresponding pixel point is 5 ° when the second Gabor filtering process is performed.
10. The apparatus of claim 6, wherein the feature extraction module extracts the feature vector VijIncluding the length of the contour, the width of the contour, the position coordinates (x, y) of the contour, and the gray level mean value of the contour, which are obtained through the contour detection processing, are 5 feature values.
CN201710051079.4A 2017-01-23 2017-01-23 Workpiece surface defect detection method and device Active CN106846313B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710051079.4A CN106846313B (en) 2017-01-23 2017-01-23 Workpiece surface defect detection method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710051079.4A CN106846313B (en) 2017-01-23 2017-01-23 Workpiece surface defect detection method and device

Publications (2)

Publication Number Publication Date
CN106846313A CN106846313A (en) 2017-06-13
CN106846313B true CN106846313B (en) 2020-03-10

Family

ID=59120802

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710051079.4A Active CN106846313B (en) 2017-01-23 2017-01-23 Workpiece surface defect detection method and device

Country Status (1)

Country Link
CN (1) CN106846313B (en)

Families Citing this family (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107680086B (en) * 2017-09-27 2020-10-23 电子科技大学 Method for detecting material contour defects with arc-shaped edges and linear edges
CN107909573B (en) * 2017-12-04 2021-09-28 广东嘉铭智能科技有限公司 Method and device for detecting knife lines on annular surface of metal workpiece
CN108154510A (en) * 2018-01-17 2018-06-12 深圳市亿图视觉自动化技术有限公司 Method for detecting surface defects of products, device and computer readable storage medium
CN108782797B (en) * 2018-06-15 2021-08-27 广东工业大学 Control method for stir-frying tea leaves by arm type tea frying machine and arm type tea frying machine
CN108960325A (en) * 2018-07-03 2018-12-07 柳州市木子科技有限公司 A kind of automobile metal plate work detection Identification of Cracks system based on SVM and Hog
CN109444169B (en) * 2019-01-17 2021-06-15 南昌航空大学 Bearing defect detection method and system
CN111504608B (en) * 2019-01-31 2022-10-04 中强光电股份有限公司 Brightness uniformity detection system and brightness uniformity detection method
CN110287990A (en) * 2019-05-21 2019-09-27 山东大学 Microalgae image classification method, system, equipment and storage medium
CN111390578B (en) * 2020-06-08 2020-09-08 佛山市南海富大精密机械有限公司 Five-axis linkage numerical control machine tool
CN111985497B (en) * 2020-07-21 2023-09-26 国网山东省电力公司电力科学研究院 Crane operation identification method and system under overhead transmission line

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103175844A (en) * 2012-03-16 2013-06-26 沈阳理工大学 Detection method for scratches and defects on surfaces of metal components
CN203908956U (en) * 2013-10-25 2014-10-29 重庆锦和鸿科技有限公司 System for detecting surface defects of workpiece

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103175844A (en) * 2012-03-16 2013-06-26 沈阳理工大学 Detection method for scratches and defects on surfaces of metal components
CN203908956U (en) * 2013-10-25 2014-10-29 重庆锦和鸿科技有限公司 System for detecting surface defects of workpiece

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
A Review of Recent Advances in Surface Defect Detection using Texture analysis Techniques;Xianghua Xie;《Electronic Letters on Computer Vision and Image Analysis》;20081231;第7卷(第3期);第1-22页 *
An Adaptive Support Vector Machine-Based Workpiece Surface Classification System Using High-Definition Metrology;Shi-Chang Du等;《IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT》;20151031;第64卷(第10期);第2590-2604页 *
一种基于机器视觉的轮胎模具表面字符检测方法;陈裕潮等;《锻压技术》;20161231;第41卷(第12期);第127-130页 *
基于机器视觉的曲面工件表面缺陷检测方法研究;唐玮;《中国优秀硕士学位论文全文数据库-信息科技辑》;20160615(第06期);第I138-1057页 *

Also Published As

Publication number Publication date
CN106846313A (en) 2017-06-13

Similar Documents

Publication Publication Date Title
CN106846313B (en) Workpiece surface defect detection method and device
CN111179225B (en) Test paper surface texture defect detection method based on gray gradient clustering
CN108355981B (en) Battery connector quality detection method based on machine vision
Hu Automated defect detection in textured surfaces using optimal elliptical Gabor filters
CN102305798B (en) Method for detecting and classifying glass defects based on machine vision
CN105447512B (en) A kind of detection method and device for the beauty defect that essence slightly combines
CN103175844A (en) Detection method for scratches and defects on surfaces of metal components
CN107437243B (en) Tire impurity detection method and device based on X-ray image
Liong et al. Automatic defect segmentation on leather with deep learning
CN108090494B (en) Textile defect identification method based on Gabor filter and support vector machine
CN108765402B (en) Non-woven fabric defect detection and classification method
KR101477665B1 (en) Defect detection method in heterogeneously textured surface
CN110473201A (en) A kind of automatic testing method and device of disc surface defect
CN111122590B (en) Ceramic surface defect detection device and detection method
CN110781913B (en) Zipper cloth belt defect detection method
Chen et al. Evaluating fabric pilling with light-projected image analysis
CN109101976B (en) Method for detecting surface defects of arc-extinguishing grid plate
Fekri-Ershad et al. A robust approach for surface defect detection based on one dimensional local binary patterns
CN113221881B (en) Multi-level smart phone screen defect detection method
CN116664565A (en) Hidden crack detection method and system for photovoltaic solar cell
Ghosh et al. Counterfeit IC detection by image texture analysis
CN111353992A (en) Agricultural product defect detection method and system based on textural features
CN113822836B (en) Method for marking an image
CN110458231B (en) Ceramic product detection method, device and equipment
CN116740449A (en) Shaving form detection method and system based on AI (advanced technology attachment) computer vision 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
TR01 Transfer of patent right
TR01 Transfer of patent right

Effective date of registration: 20201203

Address after: Fonda YUNPU Industrial Zone Whampoa District Guangzhou road 510000, Guangdong province No. 6 Room 101

Patentee after: GUANGZHOU ZSROBOT INTELLIGENT EQUIPMENT Co.,Ltd.

Address before: 510006 Panyu District, Guangzhou, Guangzhou University,, West Ring Road, No. 100

Patentee before: GUANGDONG University OF TECHNOLOGY