CN115082482A - Metal surface defect detection method - Google Patents

Metal surface defect detection method Download PDF

Info

Publication number
CN115082482A
CN115082482A CN202211009351.XA CN202211009351A CN115082482A CN 115082482 A CN115082482 A CN 115082482A CN 202211009351 A CN202211009351 A CN 202211009351A CN 115082482 A CN115082482 A CN 115082482A
Authority
CN
China
Prior art keywords
abnormal
suspected
gray value
gray
metal surface
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.)
Granted
Application number
CN202211009351.XA
Other languages
Chinese (zh)
Other versions
CN115082482B (en
Inventor
雷科
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.)
Shandong Youyipang Pump Technology Co ltd
Original Assignee
Shandong Youyipang Pump Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shandong Youyipang Pump Technology Co ltd filed Critical Shandong Youyipang Pump Technology Co ltd
Priority to CN202211009351.XA priority Critical patent/CN115082482B/en
Publication of CN115082482A publication Critical patent/CN115082482A/en
Application granted granted Critical
Publication of CN115082482B publication Critical patent/CN115082482B/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
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/40—Analysis of texture
    • G06T7/41—Analysis of texture based on statistical description of texture
    • G06T7/44—Analysis of texture based on statistical description of texture using image operators, e.g. filters, edge density metrics or local histograms
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/762—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using clustering, e.g. of similar faces in social networks
    • G06V10/763—Non-hierarchical techniques, e.g. based on statistics of modelling distributions
    • 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
    • G06T2207/30136—Metal

Abstract

The invention relates to the field of image data processing, in particular to a metal surface defect detection method, which comprises the steps of obtaining a suspected abnormal gray value and a suspected normal gray value in a metal surface image to be detected, obtaining suspected abnormal pixel points according to the suspected abnormal gray value and the suspected normal gray value, obtaining the number and the gray value of the suspected abnormal pixel points of each suspected abnormal pixel point in a window with each size by using windows with different sizes to obtain the abnormal degree of each suspected abnormal pixel point, obtaining a suspected abnormal area and a suspected normal area in the metal surface image to be detected according to the abnormal degree and the abnormal degree threshold of each suspected abnormal pixel point, obtaining an accurate abnormal gray value by using the suspected abnormal area, performing characteristic enhancement on the abnormal gray value by using a piecewise linear transformation function, and inhibiting the normal gray value to ensure that abnormal defects in the metal surface image to be detected are more obvious, the detection precision is improved.

Description

Metal surface defect detection method
Technical Field
The application relates to the field of image data processing, in particular to a metal surface defect detection method.
Background
Along with the rapid development of the industry, the demand on metals is also increasing, the smelting and processing technology of the metals is also rapidly developed, the quality is emphasized more while the yield is improved, the detection on the surface defects of the metals is very important in the quality control of products, and in the production process of the metals, the common quality problems include slag inclusion, loosening, decarburization and the like.
The existing method for detecting defects of a metal surface comprises the steps of obtaining a gray level histogram of the metal surface, analyzing the histogram, clustering gray levels according to the probability of each gray level value in the gray level histogram, using a class of gray levels with a high probability obtained after clustering as normal gray levels, using a class of gray levels with a low probability as abnormal gray levels, and using the normal gray levels and the abnormal gray levels obtained by the classification method as data bases to perform subsequent analysis.
Disclosure of Invention
The invention provides a metal surface defect detection method, which solves the problem that detection precision is low due to errors in gray value division, and adopts the following technical scheme:
acquiring a suspected abnormal gray value and a suspected normal gray value in a metal surface image to be detected;
acquiring a difference value between each suspected abnormal gray value and the average value of the suspected normal gray values, and selecting a pixel point corresponding to the suspected abnormal gray value with the largest difference value as a suspected abnormal pixel point;
taking each suspected abnormal pixel point as a center, and acquiring the number and the gray value of the suspected abnormal pixel points of the abnormal pixel point in a window area with each size;
obtaining the abnormal degree of the central suspected abnormal pixel point according to the number and the gray value of the suspected abnormal pixel points in the window area with each size;
obtaining a suspected abnormal area in the metal surface image to be detected according to the abnormal degree and the abnormal degree threshold of each suspected abnormal pixel point;
correcting the suspected abnormal gray value in the metal surface image to be detected by using the suspected abnormal area to obtain the accurate abnormal gray value in the metal surface image to be detected;
and performing characteristic enhancement on the accurate abnormal gray value in the metal surface image to be detected.
The method for acquiring the suspected abnormal gray value and the suspected normal gray value in the metal surface image to be detected comprises the following steps:
acquiring a gray level histogram of a metal surface image to be detected;
acquiring the occurrence probability of each gray value in the gray histogram, and performing K-means clustering on the occurrence probability of each gray value in the gray histogram;
and taking the gray value class with the small sum of the probability values obtained after clustering as a suspected abnormal gray value, and taking the gray value class with the large sum of the probability values as a suspected normal gray value.
The method for acquiring the abnormal degree of each suspected abnormal pixel point comprises the following steps:
selecting a suspected abnormal pixel point;
taking the suspected abnormal pixel point as the center, and sequentially counting the size as
Figure 456418DEST_PATH_IMAGE001
、
Figure 204931DEST_PATH_IMAGE002
、
Figure 975703DEST_PATH_IMAGE003
The window of (2) contains the number of suspected abnormal pixel points, wherein,
Figure 151470DEST_PATH_IMAGE004
；
the abnormal degree of each suspected abnormal pixel point is as follows:
Figure 404596DEST_PATH_IMAGE005
in the formula (I), the compound is shown in the specification,
Figure 691221DEST_PATH_IMAGE006
for the position in the image of the metal surface
Figure 782412DEST_PATH_IMAGE007
Suspected abnormal pixel point of
Figure 863500DEST_PATH_IMAGE008
The degree of abnormality of (a) is,
Figure 603923DEST_PATH_IMAGE009
is as follows
Figure 523600DEST_PATH_IMAGE009
The rows of the image data are, in turn,
Figure 767499DEST_PATH_IMAGE010
is as follows
Figure 550648DEST_PATH_IMAGE010
The columns of the image data are,
Figure 981629DEST_PATH_IMAGE011
is of size
Figure 397608DEST_PATH_IMAGE001
The number of suspected abnormal pixel points in the window of (1),
Figure 761594DEST_PATH_IMAGE012
is of the size of
Figure 450064DEST_PATH_IMAGE002
The number of suspected abnormal pixel points in the window of (1),
Figure 633920DEST_PATH_IMAGE013
is of size
Figure 567504DEST_PATH_IMAGE003
The number of suspected abnormal pixel points in the window of (1),
Figure 785995DEST_PATH_IMAGE014
is of size
Figure 848629DEST_PATH_IMAGE001
In the window of
Figure 519782DEST_PATH_IMAGE015
The number of the suspected abnormal pixel points is,
Figure 254127DEST_PATH_IMAGE016
the total number of suspected abnormal pixel points in the window,
Figure 327125DEST_PATH_IMAGE017
is of size
Figure 685294DEST_PATH_IMAGE002
In the window of
Figure 345208DEST_PATH_IMAGE018
Each suspected abnormal pixel point is selected from the group of abnormal pixel points,
Figure 384708DEST_PATH_IMAGE019
the total number of suspected abnormal pixel points in the window,
Figure 312212DEST_PATH_IMAGE020
is of size
Figure 513387DEST_PATH_IMAGE003
In the window of
Figure 657667DEST_PATH_IMAGE021
The number of the suspected abnormal pixel points is,
Figure 438541DEST_PATH_IMAGE022
the total number of suspected abnormal pixel points in the window,
Figure 220552DEST_PATH_IMAGE023
is the average of the values of the suspected normal gray values,
Figure 858207DEST_PATH_IMAGE024
is composed of
Figure 492713DEST_PATH_IMAGE001
Window openingThe weight of (a) is calculated,
Figure 936332DEST_PATH_IMAGE025
is composed of
Figure 838429DEST_PATH_IMAGE002
The weight of the window is calculated based on the weight of the window,
Figure 879941DEST_PATH_IMAGE026
is composed of
Figure 765858DEST_PATH_IMAGE003
The weight of the window.
The method for obtaining the suspected abnormal area in the metal surface image to be detected according to the abnormal degree and the abnormal degree threshold value of each suspected abnormal pixel point comprises the following steps:
if the abnormal degree of the suspected abnormal pixel point is larger than the abnormal degree threshold value, the suspected abnormal pixel point is taken as the center, and the size is
Figure 419693DEST_PATH_IMAGE003
The window area is a suspected abnormal area in the metal surface image to be detected; otherwise it is not
Figure 379559DEST_PATH_IMAGE003
The window area of (a) is a suspected normal area in the metal surface image.
In the suspected abnormal areas in the metal surface image to be detected, if the distance between two adjacent suspected abnormal areas is smaller than a distance threshold, the two adjacent suspected abnormal areas with the distance smaller than the distance threshold are merged.
The distance threshold value obtaining method comprises the following steps:
if the size of two adjacent suspected abnormal areas is
Figure 860481DEST_PATH_IMAGE027
If the distance threshold of two adjacent suspected abnormal areas is
Figure 968114DEST_PATH_IMAGE028
。
The method for correcting the suspected abnormal gray value in the metal surface image to be detected by using the suspected abnormal area to obtain the accurate abnormal gray value in the metal surface image to be detected comprises the following steps:
acquiring a suspected abnormal gray value set P in a metal surface image to be detected;
acquiring a gray value set Q in a suspected abnormal area in a metal surface image to be detected;
and taking the gray value belonging to the gray value set Q in the gray value P as an accurate abnormal gray value in the metal surface image to be detected, and taking other gray values as accurate normal gray values.
The method for performing feature enhancement on the accurate abnormal gray value in the metal surface image to be detected comprises the following steps:
constructing a piecewise linear transformation function according to the minimum gray value in the accurate normal gray values, the maximum gray value in the accurate abnormal gray values, the accurate normal gray value mean value and the accurate abnormal gray value mean value:
the piecewise linear transformation function for constructing the abnormal gray value is as follows:
Figure 425640DEST_PATH_IMAGE029
in the formula (I), the compound is shown in the specification,
Figure 302329DEST_PATH_IMAGE030
for the exact transformed gray values of the abnormal gray values,
Figure 13540DEST_PATH_IMAGE031
as a function of the number of the coefficients,
Figure 811732DEST_PATH_IMAGE032
for an accurate average of the abnormal gray values,
Figure 72949DEST_PATH_IMAGE033
the gray scale value is the original abnormal gray scale value,
Figure 305609DEST_PATH_IMAGE031
has a value of
Figure 954765DEST_PATH_IMAGE034
，
Figure 771412DEST_PATH_IMAGE035
Is the smallest gray value among the exact normal gray values,
Figure 69276DEST_PATH_IMAGE036
the maximum gray value in the accurate abnormal gray values;
the piecewise linear transformation function for constructing the normal gray value is:
Figure 920557DEST_PATH_IMAGE037
in the formula (I), the compound is shown in the specification,
Figure 615981DEST_PATH_IMAGE038
for an accurate normal gray value transformed gray value,
Figure 919923DEST_PATH_IMAGE039
as a function of the number of the coefficients,
Figure 24407DEST_PATH_IMAGE040
is an accurate average of the normal gray values,
Figure 58091DEST_PATH_IMAGE041
the gray level value is the original normal gray level value,
Figure 455575DEST_PATH_IMAGE039
has a value of
Figure 745348DEST_PATH_IMAGE042
；
And enhancing the accurate abnormal gray value in the metal surface image to be detected by using the segmented linear transformation function of the abnormal gray value, and inhibiting the accurate normal gray value in the metal surface image to be detected by using the segmented linear transformation function of the normal gray value.
The invention has the beneficial effects that:
(1) the method realizes the preliminary division of the gray value by analyzing the gray distribution probability in the gray histogram of the metal surface image and clustering the distribution probability to divide the gray value into a suspected normal gray value and a suspected abnormal gray value;
(2) the method comprises the steps of obtaining the abnormal degree of each suspected abnormal pixel point position through the number and the gray level of the suspected abnormal pixel points around the suspected abnormal gray level pixel points, obtaining abnormal areas and final abnormal gray levels and normal gray levels according to the abnormal degree, and further carrying out secondary analysis on the gray level results of primary division based on the pixel points of each suspected abnormal gray level in consideration of errors possibly caused by illumination or clustering in the primary division to obtain accurate gray level classification, so that the judgment precision of the normal gray level values and the abnormal gray levels is improved;
(3) the accurate defect region is obtained according to the accurate abnormal gray value and the accurate normal gray value, and the characteristic enhancement is carried out on the accurate defect region.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments or the description of the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flow chart of a metal surface defect detection method of the present invention;
FIG. 2 is a schematic diagram of two adjacent suspected abnormal areas with a distance less than a distance threshold in a metal surface defect detection method according to the present invention;
FIG. 3 is a schematic diagram of two adjacent suspected abnormal areas with a distance greater than a distance threshold in a metal surface defect detection method according to the present invention;
fig. 4 is a schematic diagram of suspected abnormal areas obtained by merging two adjacent suspected abnormal areas in the metal surface defect detection method of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
An embodiment of a method for detecting defects on a metal surface according to the present invention is shown in fig. 1:
the method comprises the following steps: acquiring a suspected abnormal gray value and a suspected normal gray value in an image of the metal surface to be detected;
the method comprises the steps of collecting an image of a metal surface to be detected, carrying out gray level processing on the image to obtain a gray level histogram of the metal surface to be detected, obtaining the probability of each gray level according to the gray level value and the number of pixel points in the gray level histogram, and dividing the gray level value according to the probability to obtain a suspected abnormal gray level value and a suspected normal gray level value.
In the embodiment, a camera is firstly arranged to collect the image of the metal surface to be detected, then the image of the metal surface is grayed to obtain the gray level image of the metal surface to be detected, and finally the gray level histogram corresponding to the gray level image of the metal surface to be detected is obtained.
The method for acquiring the suspected abnormal gray value and the suspected normal gray value in the metal surface image to be detected comprises the following steps:
(1) acquiring a gray level histogram of a metal surface image;
the gray histogram is the probability of occurrence of each gray value in the statistical gray map, and due to the particularity of the metal image, the gray values in the image are very close and the number of pixels under each gray value, namely the probability, is similar, so that the spatial quotient presents an aggregation state on data. If the image has defects, the image is represented as a gray value with a low probability of occurrence on the gray histogram, because the range of the defects is generally small, and the corresponding probability value is small.
(2) Acquiring the occurrence probability of each gray value in the gray histogram, and performing K-means clustering on the occurrence probability of each gray value in the gray histogram;
according to the analysis, the probability value corresponding to the normal gray level of the metal surface image is very large, and the probability value corresponding to the gray level of the defect part is very small, so that the final histogram data can be segmented through a K-means clustering algorithm to obtain two groups of data;
(3) taking a class of gray values with small sum of the probability values obtained after clustering as suspected abnormal gray values, and taking a class of gray values with large sum of the probability values as suspected normal gray values;
the K-means clustering algorithm can only segment data but cannot distinguish data types, because the probability value of normal data is far greater than that of abnormal data, a gray value with a high probability value sum is used as a suspected normal pixel point, a gray value with a low probability value sum is used as a suspected abnormal gray value in two groups of clustered data, and because the probability value of the normal data is greater than that of the abnormal data, the types of the gray values can be distinguished through comparison of the two groups of data.
Step two: acquiring a difference value between each suspected abnormal gray value and the average value of the suspected normal gray values, and selecting a pixel point corresponding to the suspected abnormal gray value with the largest difference value as a suspected abnormal pixel point; taking each suspected abnormal pixel point as a center, and acquiring the number and gray value of the suspected abnormal pixel points of the abnormal pixel points in a window area with each size; obtaining the abnormal degree of the central suspected abnormal pixel point according to the number and the gray value of the suspected abnormal pixel points in the window area with each size;
the step aims to analyze a gray value which possibly has abnormality, namely a suspected abnormal gray value, select the suspected abnormal gray value with the maximum difference according to the difference between the abnormal gray value and the average value of the normal gray values, and calculate the abnormal degree of each suspected abnormal pixel point by using the number and the gray values of the suspected abnormal pixel points in different windows;
the method for obtaining the difference value between each suspected abnormal gray value and the average value of the suspected normal gray values and selecting the pixel point corresponding to the suspected abnormal gray value with the largest difference value as the suspected abnormal pixel point comprises the following steps:
(1) calculating the mean value of all suspected normal gray values
Figure 355321DEST_PATH_IMAGE023
；
(2) Set the abnormal gray value as
Figure 915615DEST_PATH_IMAGE043
Denotes the first
Figure 749579DEST_PATH_IMAGE044
Each suspected abnormal gray value;
(3) calculating the sum of each abnormal gray value
Figure 264000DEST_PATH_IMAGE023
Difference of (2)
Figure 208822DEST_PATH_IMAGE045
；
Figure 623623DEST_PATH_IMAGE046
Is shown as
Figure 628488DEST_PATH_IMAGE044
The difference between the average value of the suspected abnormal gray value and the average value of the suspected normal gray value is larger, the larger the difference is, the larger the abnormal degree of the gray value is, the more likely the corresponding gray value is to be a defect gray value, and the more likely the pixel point corresponding to the gray value is to be a defect pixel point;
(4) selecting and
Figure 627274DEST_PATH_IMAGE023
the suspected abnormal gray value with the maximum difference is compared with
Figure 110208DEST_PATH_IMAGE023
And taking the pixel point corresponding to the suspected abnormal gray value with the largest difference value as a suspected abnormal pixel point, and performing subsequent abnormal judgment.
The meaning of the method is to determine the degree of abnormality of each gray value in the combination of abnormal gray values by calculating the difference between the abnormal gray value and the normal gray value.
It should be noted that, in the step one, the suspected normal pixel points and the suspected abnormal pixel points in the image are distinguished according to the clustering characteristics, and in the clustering result, the determination of the abnormal pixel points cannot determine that the image is abnormal, which may be caused by illumination or clustering errors, and if the image has defects, the clustered abnormal pixel points present an aggregation state in space, and the suspected abnormal gray value obtained by clustering is further analyzed and determined based on the above analysis, so as to determine whether the suspected abnormal gray value is an abnormal gray value.
The method for obtaining the abnormal degree of the central suspected abnormal pixel point according to the number and the gray value of the suspected abnormal pixel points in the window area with each size comprises the following steps:
(1) selecting each suspected abnormal pixel point; in the embodiment, the gray value of the suspected abnormality
Figure 645095DEST_PATH_IMAGE043
Selecting abnormal pixel points from all corresponding suspected abnormal pixel points
Figure 820861DEST_PATH_IMAGE008
I.e. the position in the image of the metal surface
Figure 309874DEST_PATH_IMAGE007
The suspected abnormal pixel point of (c) is detected,
Figure 596498DEST_PATH_IMAGE009
is a line of the image that is,
Figure 720312DEST_PATH_IMAGE010
is a column of the image;
(2) using the suspected abnormal pixel pointAs a center, sequentially counting the sizes of
Figure 863718DEST_PATH_IMAGE001
、
Figure 860534DEST_PATH_IMAGE002
、
Figure 950850DEST_PATH_IMAGE003
The window of (2) contains the number of suspected abnormal pixel points, wherein,
Figure 194749DEST_PATH_IMAGE004
；
for the selection of three windows, a small area such as a defect window may be formed due to the wide variety of metal defects
Figure 181160DEST_PATH_IMAGE047
Is normal in the window of (A), and
Figure 644764DEST_PATH_IMAGE048
or
Figure 273192DEST_PATH_IMAGE049
Is abnormal, and therefore only judges
Figure 637177DEST_PATH_IMAGE047
The abnormal degree of the window size is not reasonable, but the window size is too large, so that certain noise is counted and judgment is wrong, and therefore the embodiment selects the windows with three sizes for counting, namely
Figure 60068DEST_PATH_IMAGE050
、
Figure 181608DEST_PATH_IMAGE051
、
Figure 440158DEST_PATH_IMAGE052
And to ensure accuracy, when the final result is obtained, it will be right
Figure 393070DEST_PATH_IMAGE049
The window is subjected to abnormity verification;
after the window size is selected, the step is firstly counted
Figure 252442DEST_PATH_IMAGE047
Counting the number of abnormal pixel points in the window
Figure 595698DEST_PATH_IMAGE048
Counting the number of abnormal pixel points in the window, and finally counting
Figure 332973DEST_PATH_IMAGE049
The number of suspected abnormal pixel points in the window;
(3) the abnormal degree of each suspected abnormal pixel point is as follows:
Figure 405971DEST_PATH_IMAGE005
in the formula (I), the compound is shown in the specification,
Figure 170664DEST_PATH_IMAGE006
for the position in the image of the metal surface
Figure 329113DEST_PATH_IMAGE007
Suspected abnormal pixel point of
Figure 601569DEST_PATH_IMAGE008
The degree of abnormality of (a) is,
Figure 325812DEST_PATH_IMAGE009
is as follows
Figure 526986DEST_PATH_IMAGE009
The rows of the image data are, in turn,
Figure 172731DEST_PATH_IMAGE010
is as follows
Figure 720649DEST_PATH_IMAGE010
The columns of the image data are,
Figure 768240DEST_PATH_IMAGE011
is of the size of
Figure 140315DEST_PATH_IMAGE001
The number of suspected abnormal pixel points in the window of (1),
Figure 945460DEST_PATH_IMAGE012
is of size
Figure 61184DEST_PATH_IMAGE002
The number of suspected abnormal pixel points in the window of (1),
Figure 196237DEST_PATH_IMAGE013
is of size
Figure 739213DEST_PATH_IMAGE003
The number of suspected abnormal pixel points in the window of (1),
Figure 890709DEST_PATH_IMAGE014
is of size
Figure 810123DEST_PATH_IMAGE001
In the window of
Figure 68192DEST_PATH_IMAGE015
The number of the suspected abnormal pixel points is,
Figure 516491DEST_PATH_IMAGE016
the total number of suspected abnormal pixel points in the window,
Figure 889703DEST_PATH_IMAGE017
is of size
Figure 81650DEST_PATH_IMAGE002
In the window of
Figure 191295DEST_PATH_IMAGE018
The number of the suspected abnormal pixel points is,
Figure 341654DEST_PATH_IMAGE019
the total number of suspected abnormal pixel points in the window,
Figure 671004DEST_PATH_IMAGE020
is of size
Figure 604325DEST_PATH_IMAGE003
In the window of
Figure 69941DEST_PATH_IMAGE021
The number of the suspected abnormal pixel points is,
Figure 627087DEST_PATH_IMAGE022
the total number of suspected abnormal pixel points in the window,
Figure 443733DEST_PATH_IMAGE023
is the average of the normal gray-scale values,
Figure 243062DEST_PATH_IMAGE024
to represent
Figure 828764DEST_PATH_IMAGE001
The weight of (a) is calculated,
Figure 616198DEST_PATH_IMAGE025
to represent
Figure 654561DEST_PATH_IMAGE002
The weight of (a) is calculated,
Figure 257581DEST_PATH_IMAGE026
is composed of
Figure 963369DEST_PATH_IMAGE003
The weight of the window;
in the present embodiment, the first and second electrodes are,
Figure 862317DEST_PATH_IMAGE024
the value of (a) is 1/2,
Figure 387976DEST_PATH_IMAGE025
the value of (a) is 1/5,
Figure 529107DEST_PATH_IMAGE026
the value of (a) is 3/10, the total weight is 1, the three windows in the calculation formula of the abnormal degree of each suspected abnormal pixel point are different in size, if the window is smaller, the number of the abnormal pixel points is more, the abnormal degree is larger, and a larger weight is given, and the formula has the meaning that the current pixel point is used as the center
Figure 89402DEST_PATH_IMAGE047
、
Figure 156321DEST_PATH_IMAGE048
、
Figure 434856DEST_PATH_IMAGE049
Judging the abnormal condition of the position of the current suspected abnormal pixel point according to the number of the abnormal pixel points in the window and the difference of the gray values, if only one suspected abnormal pixel point is present, namely the selected central suspected abnormal pixel point, judging that the abnormal condition is present
Figure 379678DEST_PATH_IMAGE053
0, the abnormal degree of the suspected abnormal pixel
Figure 794479DEST_PATH_IMAGE054
。
Step three: obtaining a suspected abnormal area in the metal surface image to be detected according to the abnormal degree and the abnormal degree threshold of each suspected abnormal pixel point;
the method comprises the steps of determining a suspected abnormal area and a suspected normal area in a gray level image of the metal surface according to the abnormal degree of suspected abnormal pixel points, and combining adjacent suspected abnormal areas according to the distance between the adjacent suspected abnormal areas to obtain a combined suspected abnormal area.
The method for obtaining the suspected abnormal area and the suspected normal area in the metal surface image to be detected according to the abnormal degree and the abnormal degree threshold of each suspected abnormal pixel point comprises the following steps:
if the abnormal degree of the suspected abnormal pixel point is larger than the abnormal degree threshold value, the suspected abnormal pixel point is taken as the center, and the size is
Figure 300809DEST_PATH_IMAGE003
The window area of the image is a suspected abnormal area in the metal surface image; otherwise it is not
Figure 473164DEST_PATH_IMAGE003
The window area of (1) is a suspected normal area in the metal surface image, and the abnormal degree threshold value in the embodiment
Figure 956098DEST_PATH_IMAGE055
Is 10.
In the present embodiment, according to each
Figure 490985DEST_PATH_IMAGE008
Obtained with pixel points as centres
Figure 135593DEST_PATH_IMAGE049
And (4) calculating the abnormity of the window to judge whether the area is abnormal, and if all the areas are normal, indicating that the image is not abnormal. If some area has abnormity, it indicates that metal has defect, and then the subsequent treatment is carried out;
further, calculating the distance between the center points of any two adjacent suspected abnormal areas, and if the distance is smaller than a distance threshold, merging the two adjacent suspected abnormal areas to obtain a merged suspected abnormal area;
the distance threshold value obtaining method comprises the following steps: if the size of two adjacent suspected abnormal areas is
Figure 621676DEST_PATH_IMAGE027
If the distance threshold of two adjacent suspected abnormal areas is
Figure 908300DEST_PATH_IMAGE028
。
The specific combination method is as follows:
since the plurality of suspected abnormal areas are close in spatial position, the plurality of suspected abnormal areas need to be merged to determine the position of the defect area, and the suspected abnormal area determined by the suspected abnormal pixel point with the maximum abnormal degree is taken as a reference area, which is assumed to be
Figure 94431DEST_PATH_IMAGE008
And the suspected abnormal area determined by the residual suspected abnormal pixel points is assumed to be
Figure 441099DEST_PATH_IMAGE056
I.e. the position in the image of the metal surface
Figure 417407DEST_PATH_IMAGE057
Suspected abnormal pixel points are located;
computing
Figure 242144DEST_PATH_IMAGE056
And
Figure 486043DEST_PATH_IMAGE008
spatial positional relationship between the regions, i.e., distance therebetween:
Figure 3612DEST_PATH_IMAGE058
in the above formula, the first and second carbon atoms are,
Figure 690987DEST_PATH_IMAGE059
representing the spatial distance between two suspected abnormality areas,
Figure 319415DEST_PATH_IMAGE060
and
Figure 417821DEST_PATH_IMAGE061
respectively represent
Figure 106291DEST_PATH_IMAGE008
And
Figure 57192DEST_PATH_IMAGE056
the spatial positions of the two suspected abnormal areas in the image show that the two suspected abnormal areas jointly form an abnormal area when the two suspected abnormal areas are close to each other, the distance between every two suspected abnormal areas is sequentially and circularly calculated, and the distance between every two suspected abnormal areas is set as
Figure 223731DEST_PATH_IMAGE062
It indicates that the two suspected abnormal areas are close,
Figure 442223DEST_PATH_IMAGE063
is a distance threshold;
the distance threshold value obtaining method comprises the following steps:
if the size of two adjacent suspected abnormal areas is
Figure 301594DEST_PATH_IMAGE027
If the distance threshold of two adjacent suspected abnormal areas is
Figure 471282DEST_PATH_IMAGE028
Since the size of each suspected abnormal area in the present embodiment is the size
Figure 707091DEST_PATH_IMAGE049
Window size of (i.e.
Figure 780090DEST_PATH_IMAGE064
So that the size of each suspected abnormal area is
Figure 75942DEST_PATH_IMAGE049
If the distance threshold of two adjacent suspected abnormal areas is
Figure 735855DEST_PATH_IMAGE065
；
Fig. 2 shows two adjacent suspected abnormal areas with a distance smaller than the distance threshold, fig. 3 shows two adjacent suspected abnormal areas with a distance larger than the distance threshold, and fig. 4 shows a merged suspected abnormal area of the two adjacent suspected abnormal areas with a distance smaller than the distance threshold in fig. 2;
when an overlapping area exists between two suspected abnormal areas, the two suspected abnormal areas are considered as the same suspected abnormal area, and the maximum distance (distance threshold) between the two windows when the two suspected abnormal areas are overlapped is as follows:
Figure 244197DEST_PATH_IMAGE066
in the above formula, the first and second carbon atoms are,
Figure 437281DEST_PATH_IMAGE063
in order to judge the space distance threshold between two suspected abnormal areas, the maximum distance of the overlapping area is obtained according to the size of the window
Figure 372876DEST_PATH_IMAGE065
Then when
Figure 782735DEST_PATH_IMAGE067
The two are considered to be the same suspected abnormal area, and the fused suspected abnormal area in fig. 4 is obtained.
Step four: correcting the suspected abnormal gray value in the metal surface image to be detected by using the suspected abnormal area to obtain the accurate abnormal gray value in the metal surface image to be detected; and performing characteristic enhancement on the accurate abnormal gray value in the metal surface image to be detected.
The method comprises the steps of correcting a suspected abnormal gray value in a metal surface image to be detected to obtain an accurate abnormal gray value in the metal surface image to be detected, utilizing a piecewise linear transformation function to restrain a normal gray value in the metal surface image, and performing characteristic enhancement on the abnormal gray value in the metal surface image to enable defects to be more obvious.
The method for correcting the suspected abnormal gray value in the metal surface image to be detected by using the suspected abnormal area to obtain the accurate abnormal gray value in the metal surface image to be detected comprises the following steps:
acquiring a suspected abnormal gray value set P in a metal surface image to be detected;
acquiring a gray value set Q in a suspected abnormal area in a metal surface image to be detected;
and taking the gray value belonging to the gray value set Q in the gray value P as an accurate abnormal gray value in the metal surface image to be detected, and taking other gray values as accurate normal gray values.
In this embodiment, the suspected abnormal gray value contained in the merged suspected abnormal area in the metal surface image to be detected is used as an accurate abnormal gray value, other gray values are used as accurate normal gray values, connected domain analysis can be performed according to the pixel points corresponding to the accurate abnormal gray value and the pixel points corresponding to the accurate normal gray value in the metal surface image to be detected to obtain the defect area and the normal area, however, since many defects are not obvious in the defect types of the metal, the abnormal gray value is very close to the gray value of the normal image, therefore, the method is not friendly to defect detection under the condition, and is easy to generate false alarm and false alarm, the gray value of the defect part is determined by analyzing the probability, therefore, the defect part is more obvious through enhancing the characteristics of the defect part, and the subsequent detection is convenient.
The piecewise linear transformation can perform different gray enhancement or suppression on different segments according to needs, and the embodiment suppresses a normal gray value and enhances an abnormal gray value according to the two gray data obtained in the above steps, so that the gray values of the two are greatly different, and the defect characteristics are more obvious.
The method for performing feature enhancement on the accurate abnormal gray value in the metal surface image to be detected comprises the following steps:
it should be noted that the abnormal gray value refers to an accurate abnormal gray value obtained, and the normal gray value refers to an accurate normal gray value obtained;
(1) the piecewise linear transformation function for constructing the abnormal gray value is as follows:
Figure 94768DEST_PATH_IMAGE029
in the formula (I), the compound is shown in the specification,
Figure 876779DEST_PATH_IMAGE030
is the gray value after the abnormal gray value is transformed,
Figure 248855DEST_PATH_IMAGE031
as a function of the number of the coefficients,
Figure 883361DEST_PATH_IMAGE032
is the average value of the abnormal gray-scale values,
Figure 999084DEST_PATH_IMAGE033
the gray scale value is the original abnormal gray scale value,
Figure 901181DEST_PATH_IMAGE031
has a value of
Figure 444158DEST_PATH_IMAGE034
，
Figure 297451DEST_PATH_IMAGE035
Is the minimum gray value among the normal gray values,
Figure 216866DEST_PATH_IMAGE036
is the maximum gray value among the abnormal gray values;
(2) the piecewise linear transformation function for constructing the normal gray value is:
Figure 239048DEST_PATH_IMAGE037
in the formula (I), the compound is shown in the specification,
Figure 952926DEST_PATH_IMAGE038
is a gray value after the normal gray value is transformed,
Figure 562024DEST_PATH_IMAGE039
as a function of the number of the coefficients,
Figure 613026DEST_PATH_IMAGE040
is the average value of the normal gray-scale values,
Figure 486785DEST_PATH_IMAGE041
the gray level value is the original normal gray level value,
Figure 824095DEST_PATH_IMAGE039
has a value of
Figure 843127DEST_PATH_IMAGE042
；
The final purpose of the linear gray-scale transformation is to make the gray-scale values of two sets of data have large difference and highlight the characteristics of the defect part, the coefficients a and m of the linear transformation determine the magnitude of the linear transformation range, and the expressions of a and m are taken to adjust the variation range of the linear transformation
Figure 884770DEST_PATH_IMAGE068
Is to satisfy
Figure 242064DEST_PATH_IMAGE069
Get it
Figure 156799DEST_PATH_IMAGE070
Is to satisfy
Figure 206402DEST_PATH_IMAGE071
And the intercept of the linear gray scale transformation is represented by the gray scale mean value of the two types of gray scales,
Figure 271310DEST_PATH_IMAGE072
represents the difference between the normal gradation value and the abnormal gradation value, and is therefore calculated by
Figure 857012DEST_PATH_IMAGE072
Adjusting the amplitude of the gray level enhancement;
(3) and enhancing the abnormal gray value in the metal surface image to be detected by using the segmented linear transformation function of the abnormal gray value, and inhibiting the normal gray value in the metal surface image to be detected by using the segmented linear transformation function of the normal gray value to realize the characteristic enhancement of the defect region.
Furthermore, the enhanced metal surface image can be compressed and transmitted in different modes, and different coding and compression modes are adopted for different areas in the defect enhanced metal surface image for compression and transmission, so that the defect characteristics are ensured, and the compression efficiency is improved.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.

Claims (8)

1. A method for detecting defects on a metal surface, the method comprising:
acquiring a suspected abnormal gray value and a suspected normal gray value in an image of the metal surface to be detected;
acquiring a difference value between each suspected abnormal gray value and the average value of the suspected normal gray values, and selecting a pixel point corresponding to the suspected abnormal gray value with the largest difference value as a suspected abnormal pixel point;
taking each suspected abnormal pixel point as a center, and acquiring the number and the gray value of the suspected abnormal pixel points of the abnormal pixel point in a window area with each size;
obtaining the abnormal degree of the central suspected abnormal pixel point according to the number and the gray value of the suspected abnormal pixel points in the window area with each size;
obtaining a suspected abnormal area in the metal surface image to be detected according to the abnormal degree and the abnormal degree threshold of each suspected abnormal pixel point;
correcting the suspected abnormal gray value in the metal surface image to be detected by using the suspected abnormal area to obtain the accurate abnormal gray value in the metal surface image to be detected;
and performing characteristic enhancement on the accurate abnormal gray value in the metal surface image to be detected.
2. The method for detecting the defects of the metal surface according to claim 1, wherein the method for acquiring the suspected abnormal gray value and the suspected normal gray value in the image of the metal surface to be detected comprises the following steps:
acquiring a gray level histogram of a metal surface image to be detected;
acquiring the occurrence probability of each gray value in the gray histogram, and performing K-means clustering on the occurrence probability of each gray value in the gray histogram;
and taking the gray value class with the small sum of the probability values obtained after clustering as a suspected abnormal gray value, and taking the gray value class with the large sum of the probability values as a suspected normal gray value.
3. The method of claim 1, wherein the method of obtaining the abnormal degree of each suspected abnormal pixel point comprises:
selecting a suspected abnormal pixel point;
taking the suspected abnormal pixel point as the center, and sequentially counting the size as
Figure 821478DEST_PATH_IMAGE001
、
Figure 125421DEST_PATH_IMAGE002
、
Figure 495484DEST_PATH_IMAGE003
The window of (2) contains the number of suspected abnormal pixel points, wherein,
Figure 935693DEST_PATH_IMAGE004
；
the abnormal degree of each suspected abnormal pixel point is as follows:
Figure DEST_PATH_IMAGE005
in the formula (I), the compound is shown in the specification,
Figure 395493DEST_PATH_IMAGE006
for the position in the image of the metal surface
Figure 593256DEST_PATH_IMAGE007
Suspected abnormal pixel point of
Figure 826398DEST_PATH_IMAGE008
The degree of abnormality of (a) is,
Figure 386692DEST_PATH_IMAGE009
is as follows
Figure 627181DEST_PATH_IMAGE009
The rows of the image data are, in turn,
Figure 905716DEST_PATH_IMAGE010
is as follows
Figure 352003DEST_PATH_IMAGE010
The columns of the image data are,
Figure 766803DEST_PATH_IMAGE011
is of size
Figure 833985DEST_PATH_IMAGE001
The number of suspected abnormal pixel points in the window of (1),
Figure 6341DEST_PATH_IMAGE012
is of size
Figure 276827DEST_PATH_IMAGE002
The number of suspected abnormal pixel points in the window of (1),
Figure 546134DEST_PATH_IMAGE013
is of size
Figure 721900DEST_PATH_IMAGE003
The number of suspected abnormal pixel points in the window of (1),
Figure 709448DEST_PATH_IMAGE014
is of size
Figure 231958DEST_PATH_IMAGE001
In the window of
Figure 621351DEST_PATH_IMAGE015
The number of the suspected abnormal pixel points is,
Figure 764757DEST_PATH_IMAGE016
the total number of suspected abnormal pixel points in the window,
Figure 974021DEST_PATH_IMAGE017
is of size
Figure 766135DEST_PATH_IMAGE002
In the window of
Figure 275613DEST_PATH_IMAGE018
The number of the suspected abnormal pixel points is,
Figure 121078DEST_PATH_IMAGE019
the total number of suspected abnormal pixel points in the window,
Figure 584683DEST_PATH_IMAGE020
is of size
Figure 885214DEST_PATH_IMAGE003
In the window of
Figure 983620DEST_PATH_IMAGE021
The number of the suspected abnormal pixel points is,
Figure 672091DEST_PATH_IMAGE022
the total number of suspected abnormal pixel points in the window,
Figure 590368DEST_PATH_IMAGE023
is the average of the suspected normal gray-scale values,
Figure 786601DEST_PATH_IMAGE024
is composed of
Figure 739513DEST_PATH_IMAGE001
The weight of the window is calculated based on the weight,
Figure 598885DEST_PATH_IMAGE025
is composed of
Figure 535617DEST_PATH_IMAGE002
The weight of the window is calculated based on the weight,
Figure 741733DEST_PATH_IMAGE026
is composed of
Figure 80310DEST_PATH_IMAGE003
The weight of the window.
4. The method for detecting the metal surface defect according to claim 1, wherein the method for obtaining the suspected abnormal area in the metal surface image to be detected according to the abnormal degree and the abnormal degree threshold of each suspected abnormal pixel point comprises the following steps:
if the abnormal degree of the suspected abnormal pixel point is larger than the abnormal degree threshold value, the suspected abnormal pixel point is taken as the center, and the size is
Figure 110583DEST_PATH_IMAGE003
The window area is a suspected abnormal area in the metal surface image to be detected; otherwise it is not
Figure 3453DEST_PATH_IMAGE003
The window area of (a) is a suspected normal area in the metal surface image.
5. The method according to claim 4, wherein in the suspected abnormal areas in the metal surface image to be detected, if the distance between two adjacent suspected abnormal areas is smaller than a distance threshold, the two adjacent suspected abnormal areas with the distance smaller than the distance threshold are merged.
6. The method for detecting the defects on the metal surface as claimed in claim 5, wherein the distance threshold is obtained by:
if the size of two adjacent suspected abnormal areas is
Figure 275909DEST_PATH_IMAGE027
If the distance threshold of two adjacent suspected abnormal areas is
Figure 468993DEST_PATH_IMAGE028
。
7. The method according to claim 1, wherein the method for obtaining the accurate abnormal gray value in the metal surface image to be detected by correcting the suspected abnormal gray value in the metal surface image to be detected by using the suspected abnormal area comprises:
acquiring a suspected abnormal gray value set P in a metal surface image to be detected;
acquiring a gray value set Q in a suspected abnormal area in a metal surface image to be detected;
and taking the gray value belonging to the gray value set Q in the gray value P as an accurate abnormal gray value in the metal surface image to be detected, and taking the other gray values as accurate normal gray values.
8. The method for detecting the defects of the metal surface according to claim 1, wherein the method for performing the feature enhancement on the accurate abnormal gray value in the image of the metal surface to be detected comprises the following steps:
constructing a piecewise linear transformation function according to the minimum gray value in the accurate normal gray values, the maximum gray value in the accurate abnormal gray values, the accurate normal gray value mean value and the accurate abnormal gray value mean value:
the piecewise linear transformation function for constructing the abnormal gray value is as follows:
Figure 670167DEST_PATH_IMAGE029
in the formula (I), the compound is shown in the specification,
Figure 50333DEST_PATH_IMAGE030
for the exact transformed gray values of the abnormal gray values,
Figure 129409DEST_PATH_IMAGE031
as a function of the number of the coefficients,
Figure 442579DEST_PATH_IMAGE032
for an accurate average of the abnormal gray values,
Figure 549075DEST_PATH_IMAGE033
the gray scale value is the original abnormal gray scale value,
Figure 416537DEST_PATH_IMAGE031
has a value of
Figure 30796DEST_PATH_IMAGE034
，
Figure 932893DEST_PATH_IMAGE035
Is the smallest gray value among the exact normal gray values,
Figure 475870DEST_PATH_IMAGE036
the maximum gray value in the accurate abnormal gray values;
the piecewise linear transformation function for constructing the normal gray value is:
Figure 361786DEST_PATH_IMAGE037
in the formula (I), the compound is shown in the specification,
Figure 517086DEST_PATH_IMAGE038
for an accurate normal gray value transformed gray value,
Figure DEST_PATH_IMAGE039
as a function of the number of the coefficients,
Figure 601586DEST_PATH_IMAGE040
is an accurate average of the normal gray values,
Figure 315464DEST_PATH_IMAGE041
the gray level value is the original normal gray level value,
Figure 656053DEST_PATH_IMAGE039
has a value of
Figure 113579DEST_PATH_IMAGE042
；
And enhancing the accurate abnormal gray value in the metal surface image to be detected by using the segmented linear transformation function of the abnormal gray value, and inhibiting the accurate normal gray value in the metal surface image to be detected by using the segmented linear transformation function of the normal gray value.
CN202211009351.XA 2022-08-23 2022-08-23 Metal surface defect detection method Active CN115082482B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211009351.XA CN115082482B (en) 2022-08-23 2022-08-23 Metal surface defect detection method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211009351.XA CN115082482B (en) 2022-08-23 2022-08-23 Metal surface defect detection method

Publications (2)

Publication Number Publication Date
CN115082482A true CN115082482A (en) 2022-09-20
CN115082482B CN115082482B (en) 2022-11-22

Family

ID=83244719

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211009351.XA Active CN115082482B (en) 2022-08-23 2022-08-23 Metal surface defect detection method

Country Status (1)

Country Link
CN (1) CN115082482B (en)

Cited By (22)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115239738A (en) * 2022-09-26 2022-10-25 南通鑫生派智能科技有限公司 Intelligent detection method for automobile part configuration
CN115272322A (en) * 2022-09-28 2022-11-01 南通先端精密机械有限公司 Alloy defect quantitative detection method based on data processing
CN115272347A (en) * 2022-09-30 2022-11-01 中威泵业(江苏)有限公司 Bearing defect identification method
CN115272334A (en) * 2022-09-29 2022-11-01 江苏美克美斯自动化科技有限责任公司 Method for detecting micro defects on surface of steel rail under complex background
CN115294140A (en) * 2022-10-09 2022-11-04 南通永卓金属制品有限公司 Hardware part defect detection method and system
CN115294141A (en) * 2022-10-10 2022-11-04 惠智赋能(滨州)信息科技服务有限公司 Deep sea fishing net detection method based on sonar image
CN115311283A (en) * 2022-10-12 2022-11-08 山东鲁玻玻璃科技有限公司 Glass tube drawing defect detection method and system
CN115311270A (en) * 2022-10-11 2022-11-08 南通至顺聚氨酯材料有限公司 Plastic product surface defect detection method
CN115375588A (en) * 2022-10-25 2022-11-22 山东旗胜电气股份有限公司 Power grid transformer fault identification method based on infrared imaging
CN115375688A (en) * 2022-10-25 2022-11-22 苏州德斯米尔智能科技有限公司 Automatic detection method for belt type conveying equipment
CN115423813A (en) * 2022-11-05 2022-12-02 江苏惠汕新能源集团有限公司 Method for detecting welding defects on surface of welded pipe
CN115457035A (en) * 2022-11-10 2022-12-09 山东鲁旺机械设备有限公司 Machine vision-based construction hanging basket welding quality detection method
CN115601365A (en) * 2022-12-15 2023-01-13 无锡康贝电子设备有限公司（Cn） Bearing detection method for numerical control machine tool
CN115841491A (en) * 2023-02-24 2023-03-24 杭州电子科技大学 Quality detection method of porous metal material
CN116071387A (en) * 2023-03-06 2023-05-05 曲阜市巨力铁路轨道工程股份有限公司 Sleeper rail production quality detection method based on machine vision
CN116091499A (en) * 2023-04-07 2023-05-09 山东中胜涂料有限公司 Abnormal paint production identification system
CN116703898A (en) * 2023-08-03 2023-09-05 山东优奭趸泵业科技有限公司 Quality detection method for end face of precision mechanical bearing
CN116862908A (en) * 2023-09-01 2023-10-10 菏泽学院 Metal gear meshing defect detection method based on image processing
CN117173187A (en) * 2023-11-03 2023-12-05 南通银河水泵有限公司 Intelligent valve fault detection system
CN117173188A (en) * 2023-11-03 2023-12-05 泸州通鑫显示科技有限公司 Glass scar identification method
CN117437223A (en) * 2023-12-20 2024-01-23 连兴旺电子(深圳)有限公司 Intelligent defect detection method for high-speed board-to-board connector
CN117611589A (en) * 2024-01-24 2024-02-27 深圳市格瑞邦科技有限公司 Tablet personal computer quality detection method and system

Citations (29)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103149214A (en) * 2013-02-27 2013-06-12 上海交通大学 Method for detecting flaw on surface of fruit
JP2015135540A (en) * 2014-01-16 2015-07-27 大日本印刷株式会社 Image feature point extraction method, defect inspection method, and defect inspection device
CN105869135A (en) * 2015-01-19 2016-08-17 青岛软控机电工程有限公司 Tire defect detection method and device
KR20170022346A (en) * 2015-08-20 2017-03-02 충북대학교 산학협력단 System and Method for Classification of PCB fault and Type of Fault
CN106770323A (en) * 2016-12-15 2017-05-31 常州大学 Based on the textile flaw detection method that hierarchical clustering and Gabor are filtered
CN107918216A (en) * 2017-12-13 2018-04-17 深圳Tcl新技术有限公司 Image Mura defects appraisal procedure, system and readable storage medium storing program for executing
CN109544506A (en) * 2018-10-17 2019-03-29 潍坊路加精工有限公司 The detection method and device of workpiece appearance defects
CN109540925A (en) * 2019-01-23 2019-03-29 南昌航空大学 Complicated ceramic tile surface defect inspection method based on difference shadow method and local variance measurement operator
CN110149486A (en) * 2019-05-17 2019-08-20 凌云光技术集团有限责任公司 A kind of automatic testing method, bearing calibration and the system of newly-increased abnormal point
CN112477438A (en) * 2020-11-13 2021-03-12 深圳汉弘软件技术有限公司 Printing method, printing device, ink-jet printer and storage medium
JP2021128097A (en) * 2020-02-14 2021-09-02 株式会社ブイ・テクノロジー Defect detector, defect detection method, and defect correction device
CN113570605A (en) * 2021-09-28 2021-10-29 深圳市绘晶科技有限公司 Defect detection method and system based on liquid crystal display panel
CN113674247A (en) * 2021-08-23 2021-11-19 河北工业大学 X-ray weld defect detection method based on convolutional neural network
CN113934190A (en) * 2021-12-17 2022-01-14 武汉春田纸品包装有限公司 Corrugated board production quality regulation and control method based on machine vision
CN113963042A (en) * 2021-12-21 2022-01-21 派立锐汽车零部件(武汉)有限公司 Metal part defect degree evaluation method based on image processing
CN114359274A (en) * 2022-03-16 2022-04-15 布鲁奇维尔通风设备启东有限公司 Ventilation equipment blade quality detection method, device and system based on image processing
CN114359270A (en) * 2022-03-09 2022-04-15 山东华硕汽车配件科技有限公司 Computer vision-based automobile engine oil way copper bush defect detection method
CN114359416A (en) * 2022-03-17 2022-04-15 山东水利建设集团有限公司 Building outer wall hollowing leakage abnormity detection and positioning method
CN114419025A (en) * 2022-01-27 2022-04-29 江苏泰和木业有限公司 Fiberboard quality evaluation method based on image processing
CN114494210A (en) * 2022-01-28 2022-05-13 江苏豪尚新材料科技有限公司 Plastic film production defect detection method and system based on image processing
CN114523203A (en) * 2022-03-13 2022-05-24 扬州沃盛车业制造有限公司 Intelligent laser welding method and system
CN114757949A (en) * 2022-06-15 2022-07-15 济宁市海富电子科技有限公司 Wire and cable defect detection method and system based on computer vision
CN114757900A (en) * 2022-03-31 2022-07-15 启东新朋莱纺织科技有限公司 Artificial intelligence-based textile defect type identification method
CN114758185A (en) * 2022-06-16 2022-07-15 南通倍佳机械科技有限公司 Injection molding parameter control method and system based on gray level chromatic aberration
CN114842009A (en) * 2022-07-04 2022-08-02 江苏奥派电气科技有限公司 Cable defect detection optimization method based on gray level run matrix
CN114862855A (en) * 2022-07-07 2022-08-05 南通盈赛纺织品有限公司 Textile defect detection method and system based on template matching
CN114862862A (en) * 2022-07-11 2022-08-05 江苏大田阀门制造有限公司 Pump body cold shut defect identification method and system based on image processing
CN114881960A (en) * 2022-04-29 2022-08-09 南通百杭纺织品有限公司 Feature enhancement-based cloth linear defect detection method and system
CN114913365A (en) * 2022-04-22 2022-08-16 海门王巢家具制造有限公司 Artificial board quality classification method and system based on machine vision

Patent Citations (29)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103149214A (en) * 2013-02-27 2013-06-12 上海交通大学 Method for detecting flaw on surface of fruit
JP2015135540A (en) * 2014-01-16 2015-07-27 大日本印刷株式会社 Image feature point extraction method, defect inspection method, and defect inspection device
CN105869135A (en) * 2015-01-19 2016-08-17 青岛软控机电工程有限公司 Tire defect detection method and device
KR20170022346A (en) * 2015-08-20 2017-03-02 충북대학교 산학협력단 System and Method for Classification of PCB fault and Type of Fault
CN106770323A (en) * 2016-12-15 2017-05-31 常州大学 Based on the textile flaw detection method that hierarchical clustering and Gabor are filtered
CN107918216A (en) * 2017-12-13 2018-04-17 深圳Tcl新技术有限公司 Image Mura defects appraisal procedure, system and readable storage medium storing program for executing
CN109544506A (en) * 2018-10-17 2019-03-29 潍坊路加精工有限公司 The detection method and device of workpiece appearance defects
CN109540925A (en) * 2019-01-23 2019-03-29 南昌航空大学 Complicated ceramic tile surface defect inspection method based on difference shadow method and local variance measurement operator
CN110149486A (en) * 2019-05-17 2019-08-20 凌云光技术集团有限责任公司 A kind of automatic testing method, bearing calibration and the system of newly-increased abnormal point
JP2021128097A (en) * 2020-02-14 2021-09-02 株式会社ブイ・テクノロジー Defect detector, defect detection method, and defect correction device
CN112477438A (en) * 2020-11-13 2021-03-12 深圳汉弘软件技术有限公司 Printing method, printing device, ink-jet printer and storage medium
CN113674247A (en) * 2021-08-23 2021-11-19 河北工业大学 X-ray weld defect detection method based on convolutional neural network
CN113570605A (en) * 2021-09-28 2021-10-29 深圳市绘晶科技有限公司 Defect detection method and system based on liquid crystal display panel
CN113934190A (en) * 2021-12-17 2022-01-14 武汉春田纸品包装有限公司 Corrugated board production quality regulation and control method based on machine vision
CN113963042A (en) * 2021-12-21 2022-01-21 派立锐汽车零部件(武汉)有限公司 Metal part defect degree evaluation method based on image processing
CN114419025A (en) * 2022-01-27 2022-04-29 江苏泰和木业有限公司 Fiberboard quality evaluation method based on image processing
CN114494210A (en) * 2022-01-28 2022-05-13 江苏豪尚新材料科技有限公司 Plastic film production defect detection method and system based on image processing
CN114359270A (en) * 2022-03-09 2022-04-15 山东华硕汽车配件科技有限公司 Computer vision-based automobile engine oil way copper bush defect detection method
CN114523203A (en) * 2022-03-13 2022-05-24 扬州沃盛车业制造有限公司 Intelligent laser welding method and system
CN114359274A (en) * 2022-03-16 2022-04-15 布鲁奇维尔通风设备启东有限公司 Ventilation equipment blade quality detection method, device and system based on image processing
CN114359416A (en) * 2022-03-17 2022-04-15 山东水利建设集团有限公司 Building outer wall hollowing leakage abnormity detection and positioning method
CN114757900A (en) * 2022-03-31 2022-07-15 启东新朋莱纺织科技有限公司 Artificial intelligence-based textile defect type identification method
CN114913365A (en) * 2022-04-22 2022-08-16 海门王巢家具制造有限公司 Artificial board quality classification method and system based on machine vision
CN114881960A (en) * 2022-04-29 2022-08-09 南通百杭纺织品有限公司 Feature enhancement-based cloth linear defect detection method and system
CN114757949A (en) * 2022-06-15 2022-07-15 济宁市海富电子科技有限公司 Wire and cable defect detection method and system based on computer vision
CN114758185A (en) * 2022-06-16 2022-07-15 南通倍佳机械科技有限公司 Injection molding parameter control method and system based on gray level chromatic aberration
CN114842009A (en) * 2022-07-04 2022-08-02 江苏奥派电气科技有限公司 Cable defect detection optimization method based on gray level run matrix
CN114862855A (en) * 2022-07-07 2022-08-05 南通盈赛纺织品有限公司 Textile defect detection method and system based on template matching
CN114862862A (en) * 2022-07-11 2022-08-05 江苏大田阀门制造有限公司 Pump body cold shut defect identification method and system based on image processing

Non-Patent Citations (6)

* Cited by examiner, † Cited by third party
Title
K. V. MORTIN ET AL: "A System for Detecting and Detecting Defects in Sheet Metal on Grayscale Images", 《INTERNATIONAL RUSSIAN AUTOMATION CONFERENCE》 *
P. JOLLY ET AL: "Analyzing Surface Defects in Apples Using Gabor Features", 《SITIS》 *
Z. LI ET AL: "Metal surface defect detection based on MATLAB", 《IAEAC》 *
刘思: "铸坯表面缺陷自动检测系统的算法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 *
张奔: "基于机器视觉的轴承滚子表面缺陷检测研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 *
蔡南平: "基于图像的钢轨表面缺陷检测", 《中国优秀硕士学位论文全文数据库 (信息科技辑)》 *

Cited By (35)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115239738B (en) * 2022-09-26 2022-12-09 南通鑫生派智能科技有限公司 Intelligent detection method for automobile part configuration
CN115239738A (en) * 2022-09-26 2022-10-25 南通鑫生派智能科技有限公司 Intelligent detection method for automobile part configuration
CN115272322A (en) * 2022-09-28 2022-11-01 南通先端精密机械有限公司 Alloy defect quantitative detection method based on data processing
CN115272334A (en) * 2022-09-29 2022-11-01 江苏美克美斯自动化科技有限责任公司 Method for detecting micro defects on surface of steel rail under complex background
CN115272334B (en) * 2022-09-29 2023-08-25 研索仪器科技(上海)有限公司 Method for detecting tiny defects on surface of steel rail under complex background
CN115272347A (en) * 2022-09-30 2022-11-01 中威泵业(江苏)有限公司 Bearing defect identification method
CN115294140A (en) * 2022-10-09 2022-11-04 南通永卓金属制品有限公司 Hardware part defect detection method and system
CN115294141A (en) * 2022-10-10 2022-11-04 惠智赋能(滨州)信息科技服务有限公司 Deep sea fishing net detection method based on sonar image
CN115311270A (en) * 2022-10-11 2022-11-08 南通至顺聚氨酯材料有限公司 Plastic product surface defect detection method
CN115311283A (en) * 2022-10-12 2022-11-08 山东鲁玻玻璃科技有限公司 Glass tube drawing defect detection method and system
CN115311283B (en) * 2022-10-12 2023-01-24 山东鲁玻玻璃科技有限公司 Glass tube drawing defect detection method and system
CN115375588A (en) * 2022-10-25 2022-11-22 山东旗胜电气股份有限公司 Power grid transformer fault identification method based on infrared imaging
CN115375688A (en) * 2022-10-25 2022-11-22 苏州德斯米尔智能科技有限公司 Automatic detection method for belt type conveying equipment
CN115375588B (en) * 2022-10-25 2023-02-07 山东旗胜电气股份有限公司 Power grid transformer fault identification method based on infrared imaging
WO2023134789A1 (en) * 2022-10-25 2023-07-20 苏州德斯米尔智能科技有限公司 Automatic inspection method for belt-type conveying device
CN115423813A (en) * 2022-11-05 2022-12-02 江苏惠汕新能源集团有限公司 Method for detecting welding defects on surface of welded pipe
CN115423813B (en) * 2022-11-05 2023-03-24 江苏惠汕新能源集团有限公司 Method for detecting welding defects on surface of welded pipe
CN115457035A (en) * 2022-11-10 2022-12-09 山东鲁旺机械设备有限公司 Machine vision-based construction hanging basket welding quality detection method
CN115601365B (en) * 2022-12-15 2023-03-14 无锡康贝电子设备有限公司 Bearing detection method for numerical control machine tool
CN115601365A (en) * 2022-12-15 2023-01-13 无锡康贝电子设备有限公司（Cn） Bearing detection method for numerical control machine tool
CN115841491A (en) * 2023-02-24 2023-03-24 杭州电子科技大学 Quality detection method of porous metal material
CN116071387A (en) * 2023-03-06 2023-05-05 曲阜市巨力铁路轨道工程股份有限公司 Sleeper rail production quality detection method based on machine vision
CN116091499A (en) * 2023-04-07 2023-05-09 山东中胜涂料有限公司 Abnormal paint production identification system
CN116703898A (en) * 2023-08-03 2023-09-05 山东优奭趸泵业科技有限公司 Quality detection method for end face of precision mechanical bearing
CN116703898B (en) * 2023-08-03 2023-10-20 山东优奭趸泵业科技有限公司 Quality detection method for end face of precision mechanical bearing
CN116862908A (en) * 2023-09-01 2023-10-10 菏泽学院 Metal gear meshing defect detection method based on image processing
CN116862908B (en) * 2023-09-01 2023-11-24 菏泽学院 Metal gear meshing defect detection method based on image processing
CN117173187A (en) * 2023-11-03 2023-12-05 南通银河水泵有限公司 Intelligent valve fault detection system
CN117173188A (en) * 2023-11-03 2023-12-05 泸州通鑫显示科技有限公司 Glass scar identification method
CN117173188B (en) * 2023-11-03 2024-01-26 泸州通鑫显示科技有限公司 Glass scar identification method
CN117173187B (en) * 2023-11-03 2024-03-05 南通银河水泵有限公司 Intelligent valve fault detection system
CN117437223A (en) * 2023-12-20 2024-01-23 连兴旺电子(深圳)有限公司 Intelligent defect detection method for high-speed board-to-board connector
CN117437223B (en) * 2023-12-20 2024-02-23 连兴旺电子(深圳)有限公司 Intelligent defect detection method for high-speed board-to-board connector
CN117611589A (en) * 2024-01-24 2024-02-27 深圳市格瑞邦科技有限公司 Tablet personal computer quality detection method and system
CN117611589B (en) * 2024-01-24 2024-04-12 深圳市格瑞邦科技有限公司 Tablet personal computer quality detection method and system

Also Published As

Publication number Publication date
CN115082482B (en) 2022-11-22

Similar Documents

Publication Publication Date Title
CN115082482B (en) Metal surface defect detection method
CN113689428B (en) Mechanical part stress corrosion detection method and system based on image processing
CN115082467B (en) Building material welding surface defect detection method based on computer vision
CN115601364B (en) Golden finger circuit board detection method based on image analysis
CN115511889B (en) Method for detecting welding defects on surface of solar cell panel bracket
CN116091504B (en) Connecting pipe connector quality detection method based on image processing
TW201407154A (en) Integration of automatic and manual defect classification
CN110263795B (en) Target detection method based on implicit shape model and graph matching
CN115249246A (en) Optical glass surface defect detection method
CN114820625B (en) Automobile top block defect detection method
CN115311484B (en) Optimized clustering segmentation method for weld data
CN115861307B (en) Fascia gun power supply driving plate welding fault detection method based on artificial intelligence
CN115330791A (en) Part burr detection method
CN115272336A (en) Metal part defect accurate detection method based on gradient vector
CN114359288B (en) Medical image cerebral aneurysm detection and positioning method based on artificial intelligence
CN115994907A (en) Intelligent processing system and method for comprehensive information of food detection mechanism
CN115908362A (en) Method for detecting wear resistance of skateboard wheel
CN116977230B (en) Scanning electron microscope image optimization enhancement method
CN114742849B (en) Leveling instrument distance measuring method based on image enhancement
CN115082444A (en) Copper pipe weld defect detection method and system based on image processing
CN114972353A (en) Quality evaluation method and system for metal high-pressure special pipe fitting
CN109242823B (en) Reference image selection method and device for positioning calculation and automatic driving system
CN116229438B (en) Spinning quality visual identification system
CN117078679B (en) Automatic assembly line production detection method for cooling fan based on machine vision
CN116894841B (en) Visual detection method for quality of alloy shell of gearbox

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