CN110009606A - A kind of crack propagation dynamic monitoring method and device based on image recognition - Google Patents
A kind of crack propagation dynamic monitoring method and device based on image recognition Download PDFInfo
- Publication number
- CN110009606A CN110009606A CN201910222081.2A CN201910222081A CN110009606A CN 110009606 A CN110009606 A CN 110009606A CN 201910222081 A CN201910222081 A CN 201910222081A CN 110009606 A CN110009606 A CN 110009606A
- Authority
- CN
- China
- Prior art keywords
- image
- crack propagation
- crack
- test
- fatigue
- 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
Links
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30164—Workpiece; Machine component
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Investigating Strength Of Materials By Application Of Mechanical Stress (AREA)
Abstract
The present invention relates to a kind of crack propagation dynamic monitor and method based on image recognition, step are as follows: (1) carry out fatigue crack propagation test, acquire test specimen crack Propagation image；(2) crack image is pre-processed in the way of noise reduction, gray processing etc.；(3) Threshold segmentation binary conversion treatment is carried out to image using based on maximum variance between clusters；(4) based on the processing of morphologic crackle skeletonizing and the calculating of length；(5) processing obtains crack Propagation related data.The present invention combines image processing techniques with the detection of fatigue crack propagation test, can effectively avoid the disadvantages of nowadays crack detecting methods precision such as ocular estimate, potentiometry, magnetic particle method, ultrasonic method, ray method, electromagnetic detection method is not high or detection is cumbersome, higher to environmental requirement.
Description
Technical field
The present invention is a kind of crack propagation dynamic monitoring method and device based on image recognition, and Matlab figure is utilized in it
As handling implement case, obtained fatigue crack pattern is handled, and crack Propagation related data is calculated, be one
The technology of kind automatic measurement FATIGUE CRACK PROPAGATIONBY DYNAMIC MONITORING situation.
Background technique
In fields such as metallurgy, transport, machine-building and aerospaces, need largely to use metal material, in various machines
In the fracture failure accident of tool structure, 80% or so is due to caused by fatigue failure.Fatigue fracture is to annoying large number of rows
The problem of industry, especially in aerospace field, once fracture failure occurs, consequence will be catastrophic.
The test and analysis of fatigue crack initiation and extension are the main tasks of Structural Metallic Fatigue design and life prediction
One of.Nowadays, carry out fatigue crack propagation test both at home and abroad, the method detected to fatigue crack has: ocular estimate, surface
Complex technology, potentiometry, magnetic particle method, ultrasonic method, ray method, electromagnetic detection method, acoustic-emission, Modal Acoustic Emission method etc..But mesh
It is preceding it is various measurement fatigue crack method or be precision is not high or the device is complicated, implement it is cumbersome, to test environmental requirement
It is high.
Image recognition technology has high, low to test environmental requirement, contactless, achievable real time monitoring of precision etc. excellent
Point.Image recognition technology is applied to the measurement of surface fatigue crackle, it can be achieved that Poul Dorset Sheep, and easily operated, cost
It is low, precision is high, to realize that the automation of fatigue crack test, intelligence lays the foundation.
Summary of the invention
The technology of the present invention solves the problems, such as: overcoming the deficiencies of the prior art and provide a kind of crack propagation based on image recognition
Dynamic monitoring method and device combine image processing techniques with the detection of fatigue crack propagation test, can effectively avoid as
The crack detecting methods precision such as modern ocular estimate, potentiometry, magnetic particle method, ultrasonic method, ray method, electromagnetic detection method are not high or detection is numerous
The disadvantages of trivial, higher to environmental requirement.
The technology of the present invention solution: a kind of crack propagation dynamic monitoring method based on image recognition realizes step such as
Under:
The first step carries out fatigue crack propagation test, acquires a test specimen crack Propagation image at regular intervals.
It is mentioned in fatigue crack propagation test reference " GB/T 6398-2000 Fatigue Crack Growth Rate of Metallic Materials test method [S] "
Method carry out.The main device of the dynamic monitoring of fatigue crack propagation test is as shown in Fig. 2, aobvious comprising fatigue machine, focal length
Micro mirror, image capturing system and image processing system.Substantially steps are as follows for test:
(1) prefabricating fatigue crack, using wire cutting mode in sample incision precrack.
(2) CT part size is measured, CT part is as shown in Fig. 3, need to measure width W, thickness B, test specimen notch and precrack
Length an。
(3) setting fatigue crack is tested relevant parameter and is tested.Set largest loop load pmax, loading frequency f with
And stress ratio R, picture shooting interval is set, starts to test.
Second step pre-processes crack Propagation image in the way of noise reduction, gray processing etc..Specific image is located in advance
Steps are as follows for reason:
(1) there may be noises for the crack Propagation image shot, and it is unfavorable that noise can generate the processing of subsequent pictures
It influences, influences result precision, thus firstly the need of progress noise reduction process.The present invention carries out noise reduction using median filtering method, utilizes
Median filtering method sets all in the point neighborhood window for the gray value of pixel each in crack Propagation image
The intermediate value of pixel gray value, to eliminate the noise spot isolated in crack Propagation image；
(2) gray processing handles image.Original image is RGB color image, cannot reflect the morphological feature of image, for convenience
The grey scale change of image is obtained, converts crack Propagation bianry image for crack Propagation image progress Threshold segmentation,
Gray level image need to be converted by original RGB image using the function rgb2gray in Matlab.
Third step carries out Threshold segmentation binary conversion treatment to image using based on maximum variance between clusters.For a Zhang Jing
Pretreated crack Propagation gray level image is crossed, gray scale is significantly different between crackle and surface of test piece, converts gray images into
When bianry image, given threshold is needed, in order to distinguish crackle and surface of test piece.Maximum variance between clusters are a kind of adaptive
The method that determines of threshold value, basic principle is as follows:
For the crack Propagation gray level image I (x, y) that second step obtains, prospect (i.e. crackle) and background (i.e. image
In remove fatigue crack test specimen other parts) segmentation threshold be denoted as T, the pixel number for belonging to prospect accounts for the ratio of entire image
It is denoted as w0, average gray μ0；The ratio that background pixel points account for entire image is w1, average gray μ1；Entire image is put down
Equal gray scale is denoted as μ, and inter-class variance is denoted as g.Assuming that image size is M × N, the gray value of pixel is less than the picture of threshold value T in image
Plain number is N0, number of pixels of the pixel grey scale greater than threshold value T is N1, then:
N0+N1=MxN
w0+w1=1
μ=w0*μ0+w1*μ1
G=w0(μ0-μ)2+wl(μl-μ)2
G=w0w1(μ0-μ1)2
The maximum threshold value T of inter-class variance g is made using the method for traversal, it is as required.
4th step, the crack Propagation two-value that step (3) is obtained based on morphologic crackle skeletonizing processing method
Change image and carry out continuous corrosion and opening operation processing, obtains the crackle skeleton image that crack width is a pixel；According to chain
Code method acquires crackle backbone length (pixel), and carries out scale and convert to obtain actual crack length (mm)；
Specific step is as follows:
(1) crack Propagation image is subtracted into initial test specimen image (image when i.e. recurring number is 0), utilizes Matlab
In function imsubstract.I.e. respective pixel does subtraction to image subtraction between the two images, by mutually cutting algorithm,
Precrack information can be removed, obtained new images are the crack image extended after n times recycle.
(2) it is handled based on morphologic crackle skeletonizing, Morphological skeleton describes the shape and directional information of object.It has
There are the properties such as translation invariance, inverse dilatancy and idempotence, is a kind of effective shape description method.The form of bianry image X
Continuous corrosion and opening operation can be carried out to X to acquire by selecting suitable structural element B by learning skeleton.If S (X) indicates X's
Skeleton then seeks the expression formula of the skeleton of image X are as follows:
Wherein, Sn(X) n-th of skeleton subset for being X, N are that (X Θ nB) operation corrodes X at the last time before empty set
The number of iterations, i.e. N=max n | (X Θ nB) ≠ Φ }.(X Θ nB) indicates that continuous n times corrode X with B, it may be assumed that
(X Θ nB)=((... (X Θ B) Θ B) Θ ...) Θ B
(3) counting crack length in the way of chain code.It being handled by skeletonizing, the width of crack image is a pixel,
Pixel is considered as and puts and encodes one by one, its length can be calculated.Pixel is encoded using 8- directional chain-code, specifically
Process is:
Using some point arbitrarily chosen in image boundary as starting point, since the coordinate of the point, horizontal direction coordinate
It is divided into the grid of identical size with vertical direction coordinate, for the line segment in each grid, with an immediate direction code
It indicates, finally these direction codes is connected along boundary according to counter clockwise direction, obtain chain code.Assuming that the boundary chain of crackle
Code is { u1,u2,…,un, if each yard of section ui(i=1,2 ..., n), required correspondence line segment is Δ l after operationi, then
The length of required zone boundary figure are as follows:
In above formula, NcAnd NoIt is the number of verso section and odd number code section in chain code sequence respectively.When chain code value is even number
When, length 1, when chain code value is odd number, length value is
According to chain code principle, actual crack length are as follows:
L=H*l
In above formula, l is the image length (pixel) of crackle, and H is the calibration coefficient (mm/pixel) of camera system；L is to split
The physical length (mm) of line.
5th step, processing obtain crack Propagation related data.The processing respectively walked by front, available difference are tired
The corresponding counting crack length a of labor crack propagation imagei(i=0,1,2 ..., n)；Before on-test, image taking is set
Time interval, different fatigue crack propagation image correspond to different time ti(i=0,1,2 ..., n)；Before on-test, setting
Fatigue tester loading frequency f, frequency are multiplied with the time, the corresponding circulation of available different fatigue crack propagation image plus
Carry times Ni(i=0,1,2 ..., n).According to " GB/T 6398-2000 Fatigue Crack Growth Rate of Metallic Materials test method
The calculation method that [S] " is related to uses incremental polynomials method to carry out local fit derivation to determine fatigue crack growth rate
With the match value of crack length.It is each n point in front and back, total a consecutive numbers strong point (2n+1), using as follows to any test data point i
Quadratic polynomial is fitted derivation.N value of counting desirable 2,3,4, generally takes 3.
In formula:
Coefficient b0、b1、b2It is by least square method in (formula 2) section (even if between crack length observation and match value
Sum of square of deviations is minimum) regression parameter that determines.Match value aiCorrespond to recurring number NiOn fitting crack length.Parameter C1
And C2It is for converting input data, to avoid the numerical value dyscalculia when determining regression parameter.In NiThe crack propagation speed at place
Rate is obtained by (formula 1) derivation:
Using corresponding to NiFitting crack length aiStress intensity factor range Δ K value corresponding with da/dN is calculated,
Δ K are as follows:
In above formula, W is specimen width；B is sample thickness；A is counting crack length；Δ P=Pmax-Pmin, i.e., maximum to carry
The difference of lotus minimum load；α=a/W, above formula are effective for the range of a/W >=0.2.
The present invention is with prior art beneficial effect:
(1) carry out Image Acquisition shooting using long focusing microscope, have it is at low cost, do not contact sample and easily realize etc. it is prominent
Advantage；
(2) the crack Propagation length obtained using image recognition technology more can be anti-closer to the actual length of crackle
Reflect the truth of crack propagation；
(3) any type is suitable for the of less demanding of experimental enviroment using image recognition technology test crack extending length
It number can carry out the fatigue tester of crack Propagation, including electric-liquid type low-frequency fatigue test machine, the examination of electro-hydraulic and electromagnetic type high frequency
Machine etc. is tested, have a wide range of application model；
(4) image recognition technology is based on Matlab image processing toolbox, the calculation method and image which is related to
Processing technique is easily achieved；
(5) On-Line Dynamic Monitoring can be achieved, provide tool for testing staff, saved human cost, to realize
Automation, the intelligence of fatigue crack test lay the foundation；
(6) various metals, the detection on process for un-metal material surface and difficult-to-machine material be may extend to and be difficult to observe
The detection of the surface of deep hole fatigue crack arrived.
Detailed description of the invention
Fig. 1 is a kind of crack propagation dynamic monitoring method flow chart based on image recognition；
Fig. 2 is the crack propagation dynamic monitor figure based on image recognition；
Fig. 3 is CT part assay maps；
Fig. 4 is the test specimen image after the initial test specimen image and n times CYCLIC LOADING of median filter process；Left figure is warp
The initial test specimen image of median filter process is crossed, right figure is the test specimen image after the n times CYCLIC LOADING by median filter process；
Fig. 5 is the test specimen image after initial test specimen image and n times CYCLIC LOADING that gray processing is handled；Left figure is to pass through
The initial test specimen image of gray processing processing, right figure are the test specimen image after the n times CYCLIC LOADING that gray processing is handled；
Fig. 6 is the test specimen image after the initial test specimen image and n times CYCLIC LOADING of binary conversion treatment；Left figure is to pass through
The initial test specimen image of binary conversion treatment, right figure are the test specimen image after the n times CYCLIC LOADING by binary conversion treatment；
Fig. 7 is the crack Propagation bianry image after n times CYCLIC LOADING；
Fig. 8 is the crack Propagation bianry image after the n times CYCLIC LOADING that skeletonizing is handled.
Specific embodiment
With reference to the accompanying drawing, to a kind of crack propagation dynamic monitor and method based on image recognition of the present invention do into
One step explanation.
As shown in Figure 1 and Figure 2, a kind of crack propagation dynamic monitoring method based on image recognition of the present invention realizes step such as
Under:
(1) fatigue crack propagation test is carried out, acquires a test specimen crack Propagation image at regular intervals.Fatigue
Crack expansion test is referring to the side mentioned in " GB/T 6398-2000 Fatigue Crack Growth Rate of Metallic Materials test method [S] "
Method carries out.The main device of the dynamic monitoring of fatigue crack propagation test is as shown in Fig. 2, include fatigue machine, long focusing microscope, figure
As acquisition system and image processing system.
Fatigue machine is used to carry out fatigue crack propagation test to test specimen, so that surface of test piece crackle is extended；Focal length is aobvious
Micro mirror is used for real-time display surface of test piece crack propagation situation, observes test convenient for experimenter and carries out situation；Image capturing system
And image processing system be used for shoot record different moments surface of test piece crack propagation situation, and pass through noise reduction, gray processing,
The modes such as binaryzation, skeletonizing handle crack Propagation image, obtain crack Propagation length and load cycle number
Data.
Substantially steps are as follows for test:
1. prefabricating fatigue crack, using wire cutting mode in sample incision precrack.
2. measuring CT part size, CT part is as shown in figure 3, width W need to be measured, thickness B, test specimen notch and precrack length
an。
3. setting fatigue crack test relevant parameter is simultaneously tested.Set largest loop load pmax, loading frequency f with
And stress ratio R, picture shooting interval is set, starts to test.
(2) to crack Propagation image preprocessing in the way of noise reduction, gray processing etc., the present invention chooses the initial of shooting
Test specimen image and the crack Propagation image after n times CYCLIC LOADING are illustrated, and steps are as follows for specific image preprocessing:
1. there may be noise, noises to generate unfavorable shadow to the processing of subsequent picture for the crack Propagation image shot
It rings, influences result precision, thus firstly the need of progress noise reduction process.The present invention carries out noise reduction using median filtering method, in utilization
The gray value of pixel each in crack Propagation image is set all pictures in the point neighborhood window by value filtering method
The intermediate value of vegetarian refreshments gray value, to eliminate the noise spot isolated in crack Propagation image；
2. gray processing handles image.Original image is RGB color image, cannot reflect the morphological feature of image, for convenience
The grey scale change of image is obtained, converts crack Propagation bianry image for crack Propagation image progress Threshold segmentation,
Gray level image need to be converted by original RGB image using the function rgb2gray in Matlab, be handled by gray processing initial
Crack Propagation image after test specimen image and n times CYCLIC LOADING is as shown in Figure 5.
(3) Threshold segmentation binary conversion treatment is carried out to image using based on maximum variance between clusters.For one by pre-
The crack Propagation gray level image of processing, gray scale is significantly different between crackle and surface of test piece, converts gray images into two-value
When image, given threshold is needed, in order to distinguish crackle and surface of test piece.Maximum variance between clusters are a kind of adaptive thresholds
It is worth determining method, basic principle is as follows:
For the crack Propagation gray level image I (x, y) that (2) step obtains, prospect (i.e. crackle) and background (i.e. image
In remove fatigue crack test specimen other parts) segmentation threshold be denoted as T, the pixel number for belonging to prospect accounts for the ratio of entire image
It is denoted as w0, average gray μ0；The ratio that background pixel points account for entire image is w1, average gray μ1；Entire image is put down
Equal gray scale is denoted as μ, and inter-class variance is denoted as g.Assuming that image size is M × N, the gray value of pixel is less than the picture of threshold value T in image
Plain number is N0, number of pixels of the pixel grey scale greater than threshold value T is N1, then:
N0+N1=M × N
w0+w1=1
μ=w0*μ0+w1*μ1
G=w0(μ0-μ)2+w1(μ1-μ)2
G=w0w1(μ0-μ1)2
The maximum threshold value T of inter-class variance g is made using the method for traversal, it is as required.
Crack Propagation image such as Fig. 6 institute after the initial test specimen image and n times CYCLIC LOADING of binary conversion treatment
Show.
(4) based on the processing of morphologic crackle skeletonizing and the measurement of length.Specific step is as follows:
1. crack Propagation image is subtracted initial test specimen image (image when i.e. recurring number is 0), Matlab is utilized
In function imsubstract.I.e. respective pixel does subtraction to image subtraction between the two images.By mutually cutting algorithm,
Precrack information can be removed, obtained new images are the crack image extended after n times recycle, are recycled by n times
Crack propagation bianry image after load is as shown in Fig. 7.
2. being handled based on morphologic crackle skeletonizing, Morphological skeleton describes the shape and directional information of object.It has
There are the properties such as translation invariance, inverse dilatancy and idempotence, is a kind of effective shape description method.The form of bianry image X
Continuous corrosion and opening operation can be carried out to X to acquire by selecting suitable structural element B by learning skeleton.If S (X) indicates X's
Skeleton then seeks the expression formula of the skeleton of image X are as follows:
Wherein, Sn(X) n-th of skeleton subset for being X, N are that (X Θ nB) operation corrodes X at the last time before empty set
The number of iterations, i.e. N=max n | (X Θ nB) ≠ Φ }.(X Θ nB) indicates that continuous n times corrode X with B, it may be assumed that
(X Θ nB)=((... (X Θ B) Θ B) Θ ...) Θ B
Crack Propagation bianry image after the n times CYCLIC LOADING of skeletonizing processing is as shown in Fig. 8.
3. the counting crack length in the way of chain code.It is handled by skeletonizing, the width of crack image is a pixel, will
Pixel, which is considered as, to be put one by one and is encoded, its length can be calculated.Pixel is encoded using 8- directional chain-code, specific mistake
Cheng Shi:
Using some point arbitrarily chosen in image boundary as starting point, since the coordinate of the point, horizontal direction coordinate
It is divided into the grid of identical size with vertical direction coordinate, for the line segment in each grid, with an immediate direction code
It indicates, finally connects these direction codes along boundary according to counter clockwise direction, available chain code.Assuming that the side of crackle
Boundary's chain code is { u1,u2,…,un, if each yard of section ui(i=1,2 ..., n), required correspondence line segment is Δ after operation
li, then, the length of required zone boundary figure are as follows:
In above formula, NcAnd NoIt is the number of verso section and odd number code section in chain code sequence respectively.When chain code value is even number
When, length 1, when chain code value is odd number, length value is
According to chain code principle, actual crack length are as follows:
L=H*l
In above formula, l is the image length (pixel) of crackle, and H is the calibration coefficient (mm/pixel) of camera system；L is to split
The physical length (mm) of line.
(5) processing obtains crack Propagation related data.The processing respectively walked by front, available different fatigue are split
The corresponding counting crack length a of line expanded imagesi(i=0,1,2 ..., n)；Before on-test, the time of image taking is set
Interval, different fatigue crack propagation image correspond to different time ti(i=0,1,2 ..., n)；Before on-test, set tired
Labor testing machine loading frequency f, frequency are multiplied with the time, the corresponding CYCLIC LOADING time of available different fatigue crack propagation image
Number Ni(i=0,1,2 ..., n).It is related to according to " GB/T 6398-2000 Fatigue Crack Growth Rate of Metallic Materials test method [S] "
And the calculation method arrived, use incremental polynomials method to carry out local fit derivation to determine fatigue crack growth rate and crackle
The match value of length.It is each n point in front and back, total a consecutive numbers strong point (2n+1), using following secondary more to any test data point i
Item formula is fitted derivation.N value of counting desirable 2,3,4, generally takes 3.
In formula:
Coefficient b0、b1、b2It is by least square method in (formula 2) section (even if between crack length observation and match value
Sum of square of deviations is minimum) regression parameter that determines.Match value aiCorrespond to recurring number NiOn fitting crack length.Parameter C1
And C2It is for converting input data, to avoid the numerical value dyscalculia when determining regression parameter.In NiThe crack propagation speed at place
Rate is obtained by (formula 1) derivation:
Using corresponding to NiFitting crack length aiStress intensity factor range Δ K value corresponding with da/dN is calculated,
Δ K are as follows:
In above formula, W is specimen width；B is sample thickness；A is counting crack length；Δ P=Pmax-Pmin, i.e., maximum to carry
The difference of lotus minimum load；α=a/W, above formula are effective for the range of a/W >=0.2.
Above embodiments are provided just for the sake of the description purpose of the present invention, and are not intended to limit the scope of the invention.This
The range of invention is defined by the following claims.It does not depart from spirit and principles of the present invention and the various equivalent replacements made and repairs
Change, should all cover within the scope of the present invention.
Claims (6)
1. a kind of crack propagation dynamic monitoring method based on image recognition, which is characterized in that comprise the following steps that
(1) fatigue crack propagation test is carried out, test specimen crack Propagation image is acquired；The fatigue crack propagation test is
The fatigue crack propagation test carried out using standard compact tension specimen CT test specimen of the fatigue machine to material；The crack Propagation figure
As referring to that the image in the surface of test piece crack Propagation situation of different moments shooting, display foreground are that the fatigue in image is split
Line, image background are that the test specimen other parts of fatigue crack are removed in image；
(2) noise reduction and gray processing are carried out according to the crack Propagation image that noise reduction, gray processing technology obtain step (1) shooting
Pretreatment, obtains crack Propagation gray level image；
(3) Threshold segmentation binary conversion treatment is carried out to the pretreated image of step (2) using based on maximum variance between clusters；
(4) according to the crack Propagation binary picture obtained based on morphologic crackle skeletonizing processing method to step (3)
As carrying out continuous corrosion and opening operation processing, the crackle skeleton image that crack width is a pixel is obtained；According to Chain-Code-Method
Crackle backbone length (pixel) is acquired, and carries out scale and converts to obtain actual crack length (mm)；
(5) local fit derivation is carried out using incremental polynomials method, determines the quasi- of fatigue crack growth rate and crack length
Conjunction value, the actual crack length and relevant parameter obtain to step (4) are handled, and calculation processing obtains different stress intensities
The data of factor range Δ K and corresponding crack growth rate da/dN.
2. a kind of crack propagation dynamic monitoring method based on image recognition according to claim 1, it is characterised in that: institute
It states in step (2), the method for noise reduction is median filtering method, using median filtering method, by picture each in crack Propagation image
The gray value of vegetarian refreshments is set as the intermediate value of all pixels point gray value in the point neighborhood window, to eliminate fatigue crack expansion
Open up the noise spot isolated in image.
3. a kind of crack propagation dynamic monitoring method based on image recognition according to claim 1, it is characterised in that: institute
It states in step (3), is a kind of adaptive Threshold based on maximum variance between clusters, is implemented as follows:
For the crack Propagation gray level image I (x, y) that step (2) obtains, prospect (i.e. crackle) and background are (i.e. in image
Except the test specimen other parts of fatigue crack) segmentation threshold be denoted as T, belong to prospect pixel number account for entire image ratio note
For w0, average gray μ0；The ratio that background pixel points account for entire image is w1, average gray μ1；Entire image is averaged
Gray scale is denoted as μ, and inter-class variance is denoted as g, if image size is M × N, the gray value of the pixel pixel less than threshold value T in image
Number is N0, number of pixels of the pixel grey scale greater than threshold value T is N1, then:
N0+N1=M × N
w0+w1=1
μ=w0*μ0+w1*μ1
G=w0(μ0-μ)2+w1(μ1-μ)2
G=w0w1(μ0-μ1)2
The maximum threshold value T of inter-class variance g is made using the method for traversal, it is as required；After acquiring threshold value T, gray value is greater than
Threshold value T and pixel is obtained less than threshold value T be divided into two classes, carry out binary conversion treatment.
4. a kind of crack propagation dynamic monitoring method based on image recognition according to claim 1, it is characterised in that: institute
It states in step (4), the crackle skeletonizing processing method is realized are as follows: the morphology skeleton of bianry image X is by selecting structural elements
Plain B, the form and size that structural element is not fixed are while design form converts algorithm, according to input picture and institute
It needs the morphological feature of information to design, continuous corrosion and opening operation is carried out to acquire to X, if S (X) indicates the skeleton of X, then
Seek the expression formula of the skeleton of image X are as follows:
Wherein, Sn(X) n-th of skeleton subset for being X, N are that (X Θ nB) operation is secondary at the last time iteration before empty set by X corrosion
Number, i.e. N=max { n | (X Θ nB) ≠ Φ }, (X Θ nB) indicate that continuous n times corrode X with B, i.e., (X Θ nB)=((... (X
ΘB)ΘB)Θ…)ΘB。
5. a kind of crack propagation dynamic monitoring method based on image recognition according to claim 1, it is characterised in that: institute
It states in step (5), incremental polynomials method is realized are as follows:
It is each n point in front and back to any test data point i, a consecutive numbers strong point (2n+1), is carried out using following quadratic polynomial altogether
Fitting derivative, points n value take 2,3,4,
In formula:
ai-n≤a≤ai+n(formula 2)
Coefficient b0、b1、b2It is in (formula 2) section by least square method, even if the deviation between crack length observation and match value
The regression parameter that quadratic sum minimum determines, match value aiCorrespond to recurring number NiOn fitting crack length, parameter C1And C2It is
For converting input data, the numerical value dyscalculia when determining regression parameter is avoided；In NiThe crack growth rate at place is by (formula
1) derivation and obtain:
Using corresponding to NiFitting crack length aiCalculate stress intensity factor range Δ K value corresponding with da/dN, Δ K
Are as follows:
In above formula, W is specimen width；B is sample thickness；A is counting crack length；Δ P=Pmax-Pmin, i.e., maximum load is most
The difference of side crops industry；α=a/W, above formula are effective for the range of a/W >=0.2.
6. a kind of crack propagation dynamic monitor based on image recognition characterized by comprising fatigue machine, focal length are micro-
Mirror, image capturing system and image processing system；Fatigue machine is used to carry out fatigue crack propagation test to test specimen, so that test specimen
Face crack is extended；Long focusing microscope is used for real-time display surface of test piece crack propagation situation, observes examination convenient for experimenter
Test carry out situation；Image capturing system and image processing system are used to shoot the surface of test piece crack propagation of record different moments
Situation, and crack Propagation image is handled by noise reduction, gray processing, binaryzation, skeletonizing mode, obtain crack Propagation
The data of length and load cycle number.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910222081.2A CN110009606A (en) | 2019-03-22 | 2019-03-22 | A kind of crack propagation dynamic monitoring method and device based on image recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910222081.2A CN110009606A (en) | 2019-03-22 | 2019-03-22 | A kind of crack propagation dynamic monitoring method and device based on image recognition |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110009606A true CN110009606A (en) | 2019-07-12 |
Family
ID=67167789
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910222081.2A Pending CN110009606A (en) | 2019-03-22 | 2019-03-22 | A kind of crack propagation dynamic monitoring method and device based on image recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110009606A (en) |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110702687A (en) * | 2019-09-29 | 2020-01-17 | 武汉科技大学 | Crack real-time monitoring system for seat framework fatigue test |
| CN110763146A (en) * | 2019-10-31 | 2020-02-07 | 河海大学 | High-precision optical extensometer and measuring method based on double cameras |
| CN110988002A (en) * | 2019-11-27 | 2020-04-10 | 南京航空航天大学 | Rapid imaging method for microcracks of damaged section of foreign object based on image recognition |
| CN111413361A (en) * | 2020-02-24 | 2020-07-14 | 南昌大学 | Thermal fatigue crack simulation test device and method |
| CN111536890A (en) * | 2020-05-09 | 2020-08-14 | 中南大学 | Harris vertex extraction method and system for detecting cellular regularity |
| CN111681217A (en) * | 2020-05-29 | 2020-09-18 | 香港生产力促进局 | Machine vision image intelligent analysis method, device and system for fatigue fracture |
| CN111899228A (en) * | 2020-07-07 | 2020-11-06 | 蒋梦兰 | Porcelain surface crack repairing and identifying system |
| CN112129766A (en) * | 2020-09-24 | 2020-12-25 | 安徽美诺福科技有限公司 | Method, device and equipment for testing reaming crack automatic identification technology and electronic equipment |
| CN112150418A (en) * | 2020-09-08 | 2020-12-29 | 上海交通大学 | Intelligent identification method for magnetic powder inspection |
| CN112215810A (en) * | 2020-09-27 | 2021-01-12 | 武汉大学 | Fatigue test crack monitoring method and device |
| CN113008669A (en) * | 2021-01-22 | 2021-06-22 | 天津大学 | Method for dynamically monitoring stress intensity factor of crack tip |
| CN113188975A (en) * | 2021-05-07 | 2021-07-30 | 中南大学 | Rock mass fracture and flying rock motion analysis system and method based on image processing technology |
| CN113446932A (en) * | 2021-05-18 | 2021-09-28 | 西北工业大学 | Non-contact crack measuring method and system |
| CN113533074A (en) * | 2021-07-20 | 2021-10-22 | 华东理工大学 | Material high-temperature fatigue threshold value measuring system and crack length high-precision calibration method |
| CN113865487A (en) * | 2021-09-23 | 2021-12-31 | 北京航空航天大学 | Fatigue crack propagation real-time monitoring method based on structure surface displacement field |
| CN113899746A (en) * | 2021-09-30 | 2022-01-07 | 江苏纹动测控科技有限公司 | DIC-based steel structure fatigue crack propagation form measuring method |
| CN114037705A (en) * | 2022-01-11 | 2022-02-11 | 南通皋亚钢结构有限公司 | Metal fracture fatigue source detection method and system based on moire lines |
| CN114152616A (en) * | 2021-10-14 | 2022-03-08 | 盐城工学院 | Crack image recognition system and use method thereof |
| CN116501000A (en) * | 2023-06-26 | 2023-07-28 | 深圳市鑫典金光电科技有限公司 | Control method and system of composite copper heat dissipation bottom plate production equipment |
| CN117036348A (en) * | 2023-10-08 | 2023-11-10 | 中国石油大学(华东) | Metal fatigue crack detection method based on image processing and crack recognition model |
| CN117329977A (en) * | 2023-11-28 | 2024-01-02 | 中国飞机强度研究所 | Visual characteristic characterization and measurement processing method for structural fatigue crack under complex working condition |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101359365A (en) * | 2008-08-07 | 2009-02-04 | 电子科技大学中山学院 | Iris positioning method based on Maximum between-Cluster Variance and gray scale information |
| CN101413901A (en) * | 2008-12-01 | 2009-04-22 | 南京航空航天大学 | Surface fatigue crack detecting method based on CCD image characteristic |
| CN102692188A (en) * | 2012-05-08 | 2012-09-26 | 浙江工业大学 | Dynamic crack length measurement method for machine vision fatigue crack propagation test |
| CN102937593A (en) * | 2012-10-20 | 2013-02-20 | 山东理工大学 | Ceramic radome crack automatic detection method |
| CN106092785A (en) * | 2016-06-17 | 2016-11-09 | 北京航空航天大学 | The method using asymmetric crackle test fatigue crack growth rate |
-
2019
- 2019-03-22 CN CN201910222081.2A patent/CN110009606A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101359365A (en) * | 2008-08-07 | 2009-02-04 | 电子科技大学中山学院 | Iris positioning method based on Maximum between-Cluster Variance and gray scale information |
| CN101413901A (en) * | 2008-12-01 | 2009-04-22 | 南京航空航天大学 | Surface fatigue crack detecting method based on CCD image characteristic |
| CN102692188A (en) * | 2012-05-08 | 2012-09-26 | 浙江工业大学 | Dynamic crack length measurement method for machine vision fatigue crack propagation test |
| CN102937593A (en) * | 2012-10-20 | 2013-02-20 | 山东理工大学 | Ceramic radome crack automatic detection method |
| CN106092785A (en) * | 2016-06-17 | 2016-11-09 | 北京航空航天大学 | The method using asymmetric crackle test fatigue crack growth rate |
Non-Patent Citations (2)
| Title |
|---|
| 刘国华著: "《HALCON数字图像处理》", 31 May 2018 * |
| 葛茂忠等: "激光熔覆修复对TC4钛合金疲劳裂纹扩展速率的影响", 《材料导报B：研究篇》 * |
Cited By (30)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110702687A (en) * | 2019-09-29 | 2020-01-17 | 武汉科技大学 | Crack real-time monitoring system for seat framework fatigue test |
| CN110763146A (en) * | 2019-10-31 | 2020-02-07 | 河海大学 | High-precision optical extensometer and measuring method based on double cameras |
| CN110988002A (en) * | 2019-11-27 | 2020-04-10 | 南京航空航天大学 | Rapid imaging method for microcracks of damaged section of foreign object based on image recognition |
| CN111413361A (en) * | 2020-02-24 | 2020-07-14 | 南昌大学 | Thermal fatigue crack simulation test device and method |
| CN111536890A (en) * | 2020-05-09 | 2020-08-14 | 中南大学 | Harris vertex extraction method and system for detecting cellular regularity |
| CN111681217A (en) * | 2020-05-29 | 2020-09-18 | 香港生产力促进局 | Machine vision image intelligent analysis method, device and system for fatigue fracture |
| CN111681217B (en) * | 2020-05-29 | 2023-06-20 | 香港生产力促进局 | Machine vision image intelligent analysis method, device and system for fatigue fracture |
| CN111899228A (en) * | 2020-07-07 | 2020-11-06 | 蒋梦兰 | Porcelain surface crack repairing and identifying system |
| CN112150418A (en) * | 2020-09-08 | 2020-12-29 | 上海交通大学 | Intelligent identification method for magnetic powder inspection |
| CN112150418B (en) * | 2020-09-08 | 2022-07-26 | 上海交通大学 | Intelligent identification method for magnetic powder inspection |
| CN112129766A (en) * | 2020-09-24 | 2020-12-25 | 安徽美诺福科技有限公司 | Method, device and equipment for testing reaming crack automatic identification technology and electronic equipment |
| CN112215810A (en) * | 2020-09-27 | 2021-01-12 | 武汉大学 | Fatigue test crack monitoring method and device |
| CN112215810B (en) * | 2020-09-27 | 2022-03-25 | 武汉大学 | Fatigue test crack monitoring method and device |
| CN113008669A (en) * | 2021-01-22 | 2021-06-22 | 天津大学 | Method for dynamically monitoring stress intensity factor of crack tip |
| CN113188975A (en) * | 2021-05-07 | 2021-07-30 | 中南大学 | Rock mass fracture and flying rock motion analysis system and method based on image processing technology |
| CN113188975B (en) * | 2021-05-07 | 2022-07-15 | 中南大学 | Rock mass fracture and flying rock motion analysis system and method based on image processing technology |
| CN113446932A (en) * | 2021-05-18 | 2021-09-28 | 西北工业大学 | Non-contact crack measuring method and system |
| CN113533074A (en) * | 2021-07-20 | 2021-10-22 | 华东理工大学 | Material high-temperature fatigue threshold value measuring system and crack length high-precision calibration method |
| CN113865487A (en) * | 2021-09-23 | 2021-12-31 | 北京航空航天大学 | Fatigue crack propagation real-time monitoring method based on structure surface displacement field |
| CN113865487B (en) * | 2021-09-23 | 2022-11-25 | 北京航空航天大学 | Fatigue crack propagation real-time monitoring method based on structure surface displacement field |
| CN113899746A (en) * | 2021-09-30 | 2022-01-07 | 江苏纹动测控科技有限公司 | DIC-based steel structure fatigue crack propagation form measuring method |
| CN113899746B (en) * | 2021-09-30 | 2024-05-17 | 江苏纹动测控科技有限公司 | DIC-based steel structure fatigue crack growth morphology measurement method |
| CN114152616A (en) * | 2021-10-14 | 2022-03-08 | 盐城工学院 | Crack image recognition system and use method thereof |
| CN114037705A (en) * | 2022-01-11 | 2022-02-11 | 南通皋亚钢结构有限公司 | Metal fracture fatigue source detection method and system based on moire lines |
| CN116501000A (en) * | 2023-06-26 | 2023-07-28 | 深圳市鑫典金光电科技有限公司 | Control method and system of composite copper heat dissipation bottom plate production equipment |
| CN116501000B (en) * | 2023-06-26 | 2023-09-05 | 深圳市鑫典金光电科技有限公司 | Control method and system of composite copper heat dissipation bottom plate production equipment |
| CN117036348A (en) * | 2023-10-08 | 2023-11-10 | 中国石油大学(华东) | Metal fatigue crack detection method based on image processing and crack recognition model |
| CN117036348B (en) * | 2023-10-08 | 2024-01-09 | 中国石油大学(华东) | Metal fatigue crack detection method based on image processing and crack recognition model |
| CN117329977A (en) * | 2023-11-28 | 2024-01-02 | 中国飞机强度研究所 | Visual characteristic characterization and measurement processing method for structural fatigue crack under complex working condition |
| CN117329977B (en) * | 2023-11-28 | 2024-02-13 | 中国飞机强度研究所 | Visual characteristic characterization and measurement processing method for structural fatigue crack under complex working condition |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110009606A (en) | A kind of crack propagation dynamic monitoring method and device based on image recognition | |
| Xia et al. | material degradation assessed by digital image processing: Fundamentals, progresses, and challenges | |
| Liu et al. | A novel image enhancement algorithm based on stationary wavelet transform for infrared thermography to the de-bonding defect in solid rocket motors | |
| Lin | Applying a co-occurrence matrix to automatic inspection of weaving density for woven fabrics | |
| Martínez et al. | Quality inspection of machined metal parts using an image fusion technique | |
| CN109472822A (en) | Dimension of object measurement method based on depth image processing | |
| Glud et al. | Automated counting of off-axis tunnelling cracks using digital image processing | |
| Shafeek et al. | Assessment of welding defects for gas pipeline radiographs using computer vision | |
| Sun et al. | Real-time automatic detection of weld defects in steel pipe | |
| CN101387493B (en) | Shape and position dimension non-contact photoelectric detection method for pylon component hole | |
| CN110163853A (en) | A kind of detection method of edge defect | |
| Wang et al. | Digital image correlation (DIC) based damage detection for CFRP laminates by using machine learning based image semantic segmentation | |
| CN107564002A (en) | Plastic tube detection method of surface flaw, system and computer-readable recording medium | |
| CN109870461A (en) | A kind of electronic component quality detection system | |
| CN108346141A (en) | Unilateral side incidence type light guide plate defect extracting method | |
| CN102928435A (en) | Aircraft skin damage identification method and device based on image and ultrasound information fusion | |
| CN109685766A (en) | A kind of Fabric Defect detection method based on region fusion feature | |
| CN110334727B (en) | Intelligent matching detection method for tunnel cracks | |
| CN108492312A (en) | Visual tracking method based on reverse rarefaction representation under illumination variation | |
| Yamaguchi et al. | Practical image measurement of crack width for real concrete structure | |
| CN117132947B (en) | Dangerous area personnel identification method based on monitoring video | |
| CN111238927A (en) | Fatigue durability evaluation method and device, electronic equipment and computer readable medium | |
| Sutcliffe et al. | Automatic defect recognition of single-v welds using full matrix capture data, computer vision and multi-layer perceptron artificial neural networks | |
| Jiang et al. | Small infrared target detection algorithm based on mathematical morphology | |
| CN114755236A (en) | System and method for detecting surface defects of electroplated part with revolution curved surface |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190712 |
|
| WD01 | Invention patent application deemed withdrawn after publication |