CN114418921A - Industrial image crack detection method - Google Patents

Industrial image crack detection method Download PDF

Info

Publication number
CN114418921A
CN114418921A CN202011088177.3A CN202011088177A CN114418921A CN 114418921 A CN114418921 A CN 114418921A CN 202011088177 A CN202011088177 A CN 202011088177A CN 114418921 A CN114418921 A CN 114418921A
Authority
CN
China
Prior art keywords
crack
neighborhood
image
detection method
information entropy
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.)
Pending
Application number
CN202011088177.3A
Other languages
Chinese (zh)
Inventor
陈允杰
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.)
Nanjing Xindingyun Technology Co ltd
Original Assignee
Nanjing Xindingyun 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 Nanjing Xindingyun Technology Co ltd filed Critical Nanjing Xindingyun Technology Co ltd
Priority to CN202011088177.3A priority Critical patent/CN114418921A/en
Publication of CN114418921A publication Critical patent/CN114418921A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/0002Inspection of images, e.g. flaw detection
    • G06T7/0004Industrial image inspection
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01NINVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84Systems specially adapted for particular applications
    • G01N21/88Investigating the presence of flaws or contamination
    • G01N21/8851Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00Image enhancement or restoration
    • G06T5/20Image enhancement or restoration using local operators
    • G06T5/30Erosion or dilatation, e.g. thinning
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/10Segmentation; Edge detection
    • G06T7/12Edge-based segmentation
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/60Analysis of geometric attributes
    • G06T7/62Analysis of geometric attributes of area, perimeter, diameter or volume
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01NINVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84Systems specially adapted for particular applications
    • G01N21/88Investigating the presence of flaws or contamination
    • G01N21/8851Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
    • G01N2021/8887Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Quality & Reliability (AREA)
  • Signal Processing (AREA)
  • Geometry (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • General Health & Medical Sciences (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Image Analysis (AREA)
  • Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)

Abstract

The invention discloses an industrial image crack detection method, which relates to the field of industrial image flaw detection and comprises the steps of enhancing an image crack area according to an image and reducing the image of an offset field; calculating the width of the crack; the length of the crack was calculated. And carrying out quantitative analysis on the flaw and the crack on the basis so as to judge whether the workpiece belongs to the flaw product.

Description

Industrial image crack detection method
Technical Field
The invention relates to the field of industrial image flaw detection, in particular to an industrial image crack detection method.
Background
With the continuous progress of society and the continuous development of economy, modern industrial production is prosperous, industrial visual inspection technology is more and more common in more and more modern industrial production, and defects in the production process can be better inspected through an industrial visual inspection system. The surface flaws of industrial products seriously affect the quality of the products, how to avoid the surface flaws to carry out quality control is always the biggest problem faced by production enterprises, and the traditional manual detection has the defects of high cost, easy fatigue of detection personnel, easy flaws and missed detection, and the like. The qualification rate of the quality of the product is improved, the product with poor quality is reduced to be sold to customers, and the public praise of a company is effectively maintained. In addition, industrial visual inspection can also help enterprises to improve productivity, thereby improving turnover of enterprises.
The current machine vision flaw detection system integrates multiple advanced technical applications in the field of machine vision, quickly integrates an innovative detection concept, and can realize detection in different stations or single station by adopting automatic feeding and discharging and automatic removing mechanisms. Therefore, the method can be applied to industries such as automobiles, electronics and the like.
Crack detection is a common defect detection method, however, due to the image of noise, weak boundary, uneven illumination and other factors, it is difficult for the conventional gradient operator to obtain an ideal analysis result. Deep learning methods can better capture crack regions, however, such methods require large amounts of well-calibrated data. For industrial detection, objects to be analyzed are different, so that a deep learning network needs to learn again, and the problem that a large amount of calibrated data is difficult to obtain is solved. Therefore, how to acquire the cracks in the images by using the image processing technology has important research significance.
Disclosure of Invention
In order to overcome the defects in the prior art, the invention provides an industrial image crack detection method which can detect parameters such as the length and the width of a workpiece crack in an image in time so as to judge whether the workpiece belongs to a qualified product.
In order to achieve the above object, the present invention provides an industrial image crack detection method, which includes the following steps:
s1, reading the image and graying;
s2, calculating the neighborhood information entropy of each pixel point:
Figure RE-GDA0003114343430000011
in the formula (I), the compound is shown in the specification,
Figure RE-GDA0003114343430000012
the neighborhood of current point x, with size k x k, p (y) is the approximate probability of point y being in the neighborhood:
Figure RE-GDA0003114343430000021
where mean is the neighborhood mean. Since the crack zone is often lower in intensity than the non-crack zone. Therefore, when the current neighborhood contains cracks, the neighborhood mean value is smaller than the normal area gray level. In order to highlight the crack region information, the probability of the part with the gray scale greater than the mean value region is defined as 0. The larger the proportion of crack regions in the neighborhood is, the larger the information entropy is.
And S3, binarizing the information entropy matrix:
since the image contains a small number of cracks, the crack zone values are higher in the entropy matrix. Therefore, the information entropy matrix can be well binarized by using the mean value and the standard deviation in the whole information entropy:
Figure RE-GDA0003114343430000022
mean is the integral mean value of the information entropy matrix, and std is the integral standard deviation of the information entropy matrix.
S4, performing morphological corrosion operation on the obtained binary image;
since the neighborhood information is used in S2, the crack region included in the binarized information entropy matrix BW obtained in S3 is larger than the normal crack region range. This step thus etches BW using a morphological etching operation. The size is k × k.
And S5, calculating the length and the width of the crack area.
Due to interference of factors such as noise and uneven illumination, the acquired BW contains a false target. The target area in part of the image to be detected contains an area with lower gray level, so that the BW contains a pseudo target. This step therefore contains the following subdivision steps.
S5-01: each connected region in BW was extracted and area calculated using the following algorithm
Algorithm S5-01
Step 1: constructing a matrix BW2 with the same size as BW, wherein each point takes the value of 0, and defining a counting variable index as 0;
step 2: the BW is traversed in row order, and the first point (denoted as x) equal to 1 is found, and index is index +1, and a stack is built.
Step 3: BW (x) is 0, and BW2(x) is 1.
Step 4: and taking out the 4 neighborhood points of x, and if the value of the 4 neighborhood points in BW is 1 and the value of the 4 neighborhood points in BW2 is 0, pushing the points.
Step 5: if the current stack is not empty, then pop, note x, go to Step 3. If the current stack is empty, go to Step2 until all points in BW are 0.
Step 6: at this point, each connected region corresponds to an index in BW 2. The area of each connected region is calculated.
S5-02: and eliminating the areas with the area smaller than the threshold value and the area larger than the threshold value in the BW2, and setting the corresponding position value to be 0.
S5-03: the width of each connected region is calculated using the following algorithm
Algorithm S5-03
Step 1: a matrix BW3 equal to BW2 is constructed, with each point taking the value 0. Constructing a vector
Wide, recording the width of each connected region. Let index be 1.
Step 2: wide (index) ═ 1, and the region with index in BW2 is obtained
Figure RE-GDA0003114343430000031
Step 3: performing corrosion operation on BW3, wherein the radius is 1;
step 4: if BW3 contains a non-zero element, wide (index) ═ wide (index) +1, go to Step 3. If BW3 does not contain non-zero elements. The width of the region is denoted as wide (index) ═ wide (index) × 2.
Step 5: index is index + 1; if index is less than the number of connected regions, Step2 is executed. Otherwise, the algorithm is ended.
S5-04: the length of each connected region is calculated using the following algorithm
Algorithm S5-04
Step 1: a matrix BW3 equal to BW2 is constructed, with each point taking the value 0. Constructing a vector
Length, record the Length of each connected region, and let index be 1.
Step 2: (ii) a Wide (index) ═ 1, and the region with index in BW2 is obtained
Figure RE-GDA0003114343430000032
Step 3: extracting all area boundary points in BW3 and recording as R;
step 4: and (3) calculating the distance between two points in the R, wherein the maximum distance is the neighborhood length (index).
Step 5: index is index + 1; if index is less than the number of connected regions, Step2 is executed. Otherwise, the algorithm is ended.
The invention has the following beneficial effects:
the industrial image crack detection method provided by the invention can objectively quantify information such as the width and the length of the crack in the image, and reduce the influence caused by noise and uneven illumination.
Drawings
The present invention will be further described and illustrated with reference to the following drawings.
Fig. 1 is a flowchart of an industrial image crack detection method according to the present invention.
Figure 2 is an image containing a flaw.
Fig. 3 is an information entropy corresponding to the defective area of fig. 2.
Fig. 4 is a binarization result corresponding to the flaw of fig. 2 obtained by using the information entropy calculation.
FIG. 5 shows the result of morphological etching of the binarized image (FIG. 4).
Detailed Description
The technical solution of the present invention will be more clearly and completely explained by the description of the preferred embodiments of the present invention with reference to the accompanying drawings.
Examples
As shown in fig. 1, the industrial image crack detection method provided by the present invention includes the following steps:
s1, reading the image and graying. Fig. 2 is an original image.
S2, calculating the neighborhood information entropy of each pixel point:
Figure RE-GDA0003114343430000041
in the formula (I), the compound is shown in the specification,
Figure RE-GDA0003114343430000042
the neighborhood of current point x, with size k x k, p (y) is the approximate probability of point y being in the neighborhood:
Figure RE-GDA0003114343430000043
where mean is the neighborhood mean. Since the crack zone is often lower in intensity than the non-crack zone. Therefore, when the current neighborhood contains cracks, the neighborhood mean value is smaller than the normal area gray level. In order to highlight the crack region information, the probability of the part with the gray scale greater than the mean value region is defined as 0. The larger the proportion of crack regions in the neighborhood is, the larger the information entropy is. Fig. 3 is the information entropy of fig. 2. From the results, it can be seen that the defective region has a larger information entropy than the other regions.
And S3, binarizing the information entropy matrix:
since the image contains a small number of cracks, the crack zone values are higher in the entropy matrix. Therefore, the information entropy matrix can be well binarized by using the mean value and the standard deviation in the whole information entropy:
Figure RE-GDA0003114343430000051
mean is the integral mean value of the information entropy matrix, and std is the integral standard deviation of the information entropy matrix. Fig. 4 shows the binarization result.
S4, performing morphological corrosion operation on the obtained binary image;
since the neighborhood information is used in S2, the crack region included in the binarized information entropy matrix BW obtained in S3 is larger than the normal crack region range. This step thus etches BW using a morphological etching operation. The size is k × k. FIG. 5 shows the result of morphological etching of the binarized image.
And S5, calculating the length and the width of the crack area.
Due to interference of factors such as noise and uneven illumination, the acquired BW contains a false target. The target area in part of the image to be detected contains an area with lower gray level, so that the BW contains a pseudo target. This step therefore contains the following subdivision steps.
S5-01: each connected region in BW was extracted and area calculated using the following algorithm
Algorithm S5-01
Step 1: constructing a matrix BW2 with the same size as BW, wherein each point takes the value of 0, and defining a counting variable index as 0;
step 2: the BW is traversed in row order, and the first point (denoted as x) equal to 1 is found, and index is index +1, and a stack is built.
Step 3: BW (x) is 0, and BW2(x) is 1.
Step 4: and taking out the 4 neighborhood points of x, and if the value of the 4 neighborhood points in BW is 1 and the value of the 4 neighborhood points in BW2 is 0, pushing the points.
Step 5: if the current stack is not empty, then pop, note x, go to Step 3. If the current stack is empty, go to Step2 until all points in BW are 0.
Step 6: at this point, each connected region corresponds to an index in BW 2. The area of each connected region is calculated.
S5-02: and eliminating the areas with the area smaller than the threshold value and the area larger than the threshold value in the BW2, and setting the corresponding position value to be 0.
S5-03: the width of each connected region is calculated using the following algorithm
Algorithm S5-03
Step 1: a matrix BW3 equal to BW2 is constructed, with each point taking the value 0. And constructing a vector Wide and recording the width of each connected region. Let index be 1.
Step 2: wide (index) ═ 1, and the region with index in BW2 is obtained
Figure RE-GDA0003114343430000061
Step 3: performing corrosion operation on BW3, wherein the radius is 1;
step 4: if BW3 contains a non-zero element, wide (index) ═ wide (index) +1, go to Step 3. If BW3 does not contain non-zero elements. The width of the region is denoted as wide (index) ═ wide (index) × 2.
Step 5: index is index + 1; if index is less than the number of connected regions, Step2 is executed. Otherwise, the algorithm is ended.
S5-04: the length of each connected region is calculated using the following algorithm
Algorithm S5-04
Step 1: a matrix BW3 equal to BW2 is constructed, with each point taking the value 0. Constructing a vector
Length, record the Length of each connected region, and let index be 1.
Step 2: (ii) a Wide (index) ═ 1, and the region with index in BW2 is obtained
Figure RE-GDA0003114343430000062
Step 3: extracting all area boundary points in BW3 and recording as R;
step 4: and (3) calculating the distance between two points in the R, wherein the maximum distance is the neighborhood length (index).
Step 5: index is index + 1; if index is less than the number of connected regions, Step2 is executed. Otherwise, the algorithm is ended.
And S6, judging whether the required result meets the requirement of good products or not, and if not, marking the result as a waste product.
The above detailed description merely describes preferred embodiments of the present invention and does not limit the scope of the invention. Without departing from the spirit and scope of the present invention, it should be understood that various changes, substitutions and alterations can be made herein by those skilled in the art without departing from the spirit and scope of the invention as defined by the appended claims and their equivalents. The scope of the invention is defined by the claims.

Claims (5)

1. An industrial image crack detection method is characterized by comprising the following steps:
s1, acquiring image data and graying;
s2, calculating the information entropy of each pixel point neighborhood;
s3, carrying out binarization by using the information entropy matrix;
s4, performing morphological corrosion operation on the obtained binary image;
s5, calculating the length and the width of the crack area;
s6, judging whether the current workpiece belongs to a qualified product
The industrial image crack detection method of claim 1, wherein in the step S2, the neighborhood information entropy of each point is calculated as:
Figure DEST_PATH_IMAGE001
in the formula (I), the compound is shown in the specification,
Figure DEST_PATH_IMAGE002
is the current point
Figure DEST_PATH_IMAGE003
The neighborhood of (a) is determined,
Figure DEST_PATH_IMAGE004
is a point
Figure DEST_PATH_IMAGE005
Approximate probability within the neighborhood:
Figure DEST_PATH_IMAGE006
Figure DEST_PATH_IMAGE007
is a neighborhood mean
The industrial image crack detection method as claimed in claim 2, wherein in S3, the obtained image has a large number of crack-containing regions, so that the mean and variance of the information entropy matrix are calculated and the information is used to binarize the information entropy matrix:
Figure DEST_PATH_IMAGE008
,
wherein
Figure DEST_PATH_IMAGE009
The average value of the whole body is shown as the average value,
Figure DEST_PATH_IMAGE010
as a whole standard deviation.
2. The industrial image crack detection method of claim 3, wherein in S4, the corrected target region is obtained by using an erosion operation in the psychology, and the psychology size is consistent with the neighborhood size in S2.
3. The industrial image crack detection method of claim 4, wherein in the step S5, the communication area of each region is first extracted, the area of each communication area is calculated, and the region with the area smaller than the threshold value is deleted.
4. The length of the crack is obtained by using a zone growing method.
5. The width of the crack was calculated using a morphological etching operation.
CN202011088177.3A 2020-10-13 2020-10-13 Industrial image crack detection method Pending CN114418921A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011088177.3A CN114418921A (en) 2020-10-13 2020-10-13 Industrial image crack detection method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011088177.3A CN114418921A (en) 2020-10-13 2020-10-13 Industrial image crack detection method

Publications (1)

Publication Number Publication Date
CN114418921A true CN114418921A (en) 2022-04-29

Family

ID=81260566

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011088177.3A Pending CN114418921A (en) 2020-10-13 2020-10-13 Industrial image crack detection method

Country Status (1)

Country Link
CN (1) CN114418921A (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103048329A (en) * 2012-12-11 2013-04-17 北京恒达锦程图像技术有限公司 Pavement crack detecting method based on active contour model
CN104535589A (en) * 2015-01-20 2015-04-22 广东电网有限责任公司电力科学研究院 Online detection method and device for low-voltage current mutual inductor
CN107132232A (en) * 2017-04-24 2017-09-05 西南交通大学 A kind of crack detecting method of high ferro OCS Messenger Wire support base
CN107527354A (en) * 2017-07-06 2017-12-29 长安大学 A kind of region growing method based on composite diagram
CN111256594A (en) * 2020-01-18 2020-06-09 中国人民解放军国防科技大学 Method for measuring physical characteristics of surface state of aircraft skin

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103048329A (en) * 2012-12-11 2013-04-17 北京恒达锦程图像技术有限公司 Pavement crack detecting method based on active contour model
CN104535589A (en) * 2015-01-20 2015-04-22 广东电网有限责任公司电力科学研究院 Online detection method and device for low-voltage current mutual inductor
CN107132232A (en) * 2017-04-24 2017-09-05 西南交通大学 A kind of crack detecting method of high ferro OCS Messenger Wire support base
CN107527354A (en) * 2017-07-06 2017-12-29 长安大学 A kind of region growing method based on composite diagram
CN111256594A (en) * 2020-01-18 2020-06-09 中国人民解放军国防科技大学 Method for measuring physical characteristics of surface state of aircraft skin

Similar Documents

Publication Publication Date Title
CN114937055B (en) Image self-adaptive segmentation method and system based on artificial intelligence
CN111080582B (en) Method for detecting defects of inner and outer surfaces of workpiece
CN110210477B (en) Digital instrument reading identification method
CN109239073B (en) Surface defect detection method for automobile body
CN115063430B (en) Electric pipeline crack detection method based on image processing
CN116777907A (en) Sheet metal part quality detection method
CN110070523B (en) Foreign matter detection method for bottle bottom
CN111598801B (en) Identification method for weak Mura defect
CN117152129B (en) Visual detection method and system for surface defects of battery cover plate
CN104458766A (en) Cloth surface blemish detection method based on structure texture method
CN115719332A (en) Welding quality detection method
CN117333467B (en) Image processing-based glass bottle body flaw identification and detection method and system
CN116109577A (en) Printing label defect detection system and method
CN110060239B (en) Defect detection method for bottle opening of bottle
CN116381053A (en) Ultrasonic detection method and system for welding metal materials
CN107610119A (en) The accurate detection method of steel strip surface defect decomposed based on histogram
CN112561875A (en) Photovoltaic cell panel coarse grid detection method based on artificial intelligence
CN113129265B (en) Method and device for detecting surface defects of ceramic tiles and storage medium
CN114332081A (en) Textile surface abnormity determination method based on image processing
CN114418921A (en) Industrial image crack detection method
CN112750119A (en) Detection and measurement method for weak defects on surface of white glass cover plate
CN111062939A (en) Method for quickly discriminating quality and automatically extracting defect characteristics of strip steel surface
CN112330606B (en) Defect detection method based on machine learning
Yun et al. Detection of line defects in steel billets using undecimated wavelet transform
CN113870299A (en) 3D printing fault detection method based on edge detection and morphological image processing

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