CN107240086B - A kind of fabric defects detection method based on integral nomography - Google Patents
A kind of fabric defects detection method based on integral nomography Download PDFInfo
- Publication number
- CN107240086B CN107240086B CN201610180226.3A CN201610180226A CN107240086B CN 107240086 B CN107240086 B CN 107240086B CN 201610180226 A CN201610180226 A CN 201610180226A CN 107240086 B CN107240086 B CN 107240086B
- Authority
- CN
- China
- Prior art keywords
- image
- gradient energy
- distribution
- fabric
- indefectible
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 39
- 239000004744 fabric Substances 0.000 title claims abstract description 39
- 230000007547 defect Effects 0.000 title claims abstract description 37
- 238000009826 distribution Methods 0.000 claims abstract description 43
- 238000000034 method Methods 0.000 claims abstract description 36
- 230000004069 differentiation Effects 0.000 claims abstract description 4
- 238000012360 testing method Methods 0.000 claims description 11
- 238000012549 training Methods 0.000 claims description 8
- 230000010354 integration Effects 0.000 claims description 5
- 238000002813 epsilometer test Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 239000004753 textile Substances 0.000 abstract description 10
- 230000003044 adaptive effect Effects 0.000 abstract description 2
- 230000011218 segmentation Effects 0.000 abstract description 2
- 238000011897 real-time detection Methods 0.000 abstract 1
- 230000006870 function Effects 0.000 description 16
- 238000004422 calculation algorithm Methods 0.000 description 13
- 230000008859 change Effects 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 238000000605 extraction Methods 0.000 description 4
- 238000001914 filtration Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 241001269238 Data Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 206010016256 fatigue Diseases 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000005304 joining Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000004064 recycling Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000011179 visual inspection Methods 0.000 description 1
- 238000013316 zoning Methods 0.000 description 1
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
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/30124—Fabrics; Textile; Paper
Abstract
A kind of fabric defects detection method based on integral nomography, Defect Detection is used for using integral nomography rapidly extracting gradient energy statistical nature, first by carrying out image study to indefectible template, count its gradient energy feature distribution, distribution peaks are extracted, then adaptively seek differentiation of the threshold parameter for subsequent flaw；Then, the gradient energy of window where seeking each pixel by integrating nomography to image to be detected, in conjunction with the threshold parameter, determine current pixel point whether fault, determine whether present image is flaw fabric by counting the fault sum of entire image.One aspect of the present invention accelerates the principle of operation based on integrogram, the gradient energy feature distribution of rapidly extracting textile image, it realizes the real-time detection of fabric defects, on the other hand solves distribution peaks and obtain adaptive flaw decision threshold parameter, realize the accurate segmentation of fabric defects.The method of the present invention not only can guarantee real-time but also have higher accuracy.
Description
Technical field
The present invention relates to machine vision and technical field of video image processing, specially a kind of quickly based on integral graphic calculation
The fabric defects detection method of method.
Background technique
The Defect Detection of traditional textile industry is mostly based on visual inspection, however human eye vision fatiguability causes to leak
Inspection, artificial observation low efficiency, human cost cost are big, this is extremely uncoordinated with large-scale industrial production, utilizes computer vision
Fabric defects detection is carried out automatically with image processing algorithm, can effectively solve this problem.
Defect Detection is carried out in frequency domain extraction cloth textured feature based on the method for image filtering, such as Gabor filtering, small
Wave conversion, since flaw dimension is uncertain, when filtering, generally requires to extract the result of multiple scale multiple directions as feature
Vector, even with PCA dimension reduction method, single-frame images detection time still needs to tens of seconds；Based on the method for signal statistics in space
Domain counts fabric gray distribution features to identify flaw, such as local binary patterns (Local Binary Pattern, LBP) spy
Sign, gray level co-occurrence matrixes feature, rule band (Regular Band, RB) feature etc., statistical nature has preferable robustness, but unites
It generally requires when counting feature extraction using the multiple pixel datas of neighborhood, if do not used suitable acceleration strategy to will lead to integral operation
Measure abruptly increase；The method for dividing flaw to the direct thresholding of textile image is easy to operate, operation is very fast, but only to plain weave, tiltedly
The uniform gray levels such as line, the fabric of texture-free pattern are effective, and vulnerable to noise jamming.For the industry for realizing automatic flaw detection algorithm
Change application, real-time and accuracy all need to meet, and according to statistics, only small part algorithm can satisfy real-time, and wherein examine
It is less higher than 90% algorithm to survey accuracy.
The present invention proposes a kind of rapid fabric flaw detection method based on integrogram.Using integrogram by arbitrary size
Summation operation abbreviation in image block is add operation three times, rapidly extracting gradient energy statistical nature, when greatly reducing operation
Between expense, and obtain adaptive flaw decision threshold using kernel function fitting non-symmetrical features distribution, realize defect areas
Accurate segmentation.
Summary of the invention
The problem to be solved in the present invention is: existing fabric defects detection system relies on eye-observation, inefficiency；It is existing logical
The method operand for crossing the fabric defects detection that various complicated algorithms carry out high accuracy is larger, is unsatisfactory for industrial real-time
Property require；The existing method that can quickly detect fabric defects is only capable of reply simple fabric image, to complex texture fabric effects
It is poor.To sum up, existing method is difficult to the compatibility of high real-time and high accuracy.
The technical solution of the present invention is as follows: a kind of fabric defects detection method based on integral nomography, utilizes integral graphic calculation
Summation operation abbreviation in the image block of arbitrary size is add operation three times by method, is used for rapidly extracting gradient energy feature
Defect Detection, specifically: first by carrying out image study to indefectible template, count its gradient energy feature distribution, gained
Feature distribution is asymmetric, is distributed using kernel function fit characteristic, the peak value of distribution is extracted in conjunction with average drifting method, then certainly by peak value
Threshold parameter is sought in adaptation, and the threshold parameter is used for the differentiation of subsequent flaw；Then, to image to be detected, pass through integrogram
Whether the gradient energy of detection window where algorithm seeks each pixel determines current pixel point in conjunction with the threshold parameter
Fault determines whether present image is flaw fabric by counting the fault sum of entire image,
Wherein, the extracting method of gradient energy feature are as follows: the gradient map G (x, y) of original image F (x, y) is sought first, then
The gradient energy characteristic pattern E (x, y) of G (x, y) is sought using integral nomography, to any pixel (x, y), energy feature is
Size centered on point (x, y) is the pixel integration in the window area of dw*dh.
The specific steps of the gradient energy characteristic pattern E (x, y) of G (x, y) are sought using integral nomography are as follows:
1) the integrogram I (x, y) of G (x, y) is sought
I (x, y)=I (x-1, y)+I (x, y-1)-I (x-1, y-1)+G (x, y) (1)
2) according to the characteristic of integrogram, its gradient energy characteristic pattern E (x, y) is sought to arbitrary point (x, y):
Specific step is as follows for the fabric defects detection method:
1) indefectible template image is utilized, training study is used for the threshold parameter of Defect Detection
Training set is established with the image of indefectible fabric, to each width image, obtains the gradient energy characteristic pattern E of imagetrain
(x, y) is fitted E using kernel functiontrainGradient energy distribution in (x, y), obtains cuclear density probability distribution P (e), using equal
Value drift method iteration seeks the extreme value of cuclear density probability distribution P (e), obtains the peak value position of gradient energy distribution, is denoted as
Energy distribution is divided into left and right two according to peak value, calculates separately two variances sigmas by μ1、σ2, to the same kind fabric in training set
Every indefectible image seek parameter μ, σ1、σ2Afterwards, then correspondence takes its average value Most as the kind fabric
Threshold parameter during whole Defect Detection；
2) to each image to be detected, its gradient energy characteristic pattern E is obtainedtest(x, y), if EtestUnder (x, y) meets
Then decision-point (x, y) is fault to formula:
Wherein,For the threshold parameter obtained in step 1), α is control coefrficient；
Total fault number is counted, finally if more than the threshold value T of settingdThen determine that the figure has flaw.
Cuclear density probability distribution P (e) are as follows:
Wherein, N is the sum of all pixels of the indefectible image, { en| n=1,2,3...N } it is each point in indefectible image
Gradient energy data, b be kernel function bandwidth, c0For normalization coefficient, k (z) is Gaussian kernel profile function, i.e. k (z)=exp
(- z/2), z >=0.
Threshold parameter μ, σ1、σ2Circular are as follows:
1) initial value μ is set0For the gradient energy mean value of current indefectible template image；
2) μ is enabled1=m (μ0)+μ0, wherein m (μ0) it is μ0The average drifting amount at place
In above formula, function g (z)=- k ' (z), k (z) is Gaussian kernel profile function；
If 3) | μ1-μ0| < ε enables parameter μ=μ1, carry out in next step, otherwise enabling μ0=μ1And return to previous step and be iterated, ε
Indicate dimensionless；
4) threshold parameter σ is calculated according to formula (6)1, wherein S is the pixel sum on the right side of gradient energy distribution map peak value.
es∈{en| n=1,2...N } and es≥μ (6)
5) threshold parameter σ is calculated according to formula (7)2, wherein Z is the pixel sum on the left of gradient energy distribution map peak value.
ez∈{en| n=1,2...N } and ez≤μ (7)。
The present invention proposes a kind of rapid fabric flaw detection method based on integrogram, had not only met real-time but also had had higher
Accuracy.Its innovative point is: 1) being put forward for the first time and used using the gradient energy feature of integral nomography rapidly extracting textile image
In flaw differentiate, by integrogram by the summation operation abbreviation in the image block of arbitrary size be add operation three times, greatly reduce
Run expense；Meanwhile when using integral nomography, the present invention is first to do gradient map to do integrogram again, rather than directly integrate
Figure, can remove the interference of illumination variation in this way, further raising detection accuracy；2) present invention is fitted non-right using kernel function
The feature distribution of title extracts distribution peaks in conjunction with average drifting method, adaptively seeks threshold parameter, realizes automatic, accurate point
Cut defect areas.
Detailed description of the invention
Fig. 1 is the flow diagram of the method for the present invention.
Fig. 2 is the schematic diagram based on integrogram zoning integral in the method for the present invention.
Fig. 3 is textile image to be detected of the embodiment of the present invention.
Fig. 4 is gradient energy of embodiment of the present invention characteristic pattern.
Fig. 5 is Defect Detection of embodiment of the present invention result figure.
Fig. 6 is the testing result comparison of the method for the present invention and other art methods.
Specific embodiment
The present invention provides a kind of new fabric defects detection methods, fast and automatically change and realize high-precision fabric defects
Detection.The method of the present invention mainly includes the gradient energy feature extraction based on integrogram, the study of the threshold parameter based on kernel function
And three parts of Defect Detection.
As shown in Figure 1, the present invention counts its gradient energy feature point first by carrying out image study to indefectible template
Cloth, gained feature distribution is asymmetric, is distributed using kernel function fit characteristic, and the peak value of distribution is extracted in conjunction with average drifting method, then
Threshold parameter is adaptively sought by peak value, the threshold parameter is used for the differentiation of subsequent flaw；Then, to image to be detected, lead to
The gradient energy for crossing detection window where integral nomography seeks each pixel determines current picture in conjunction with the threshold parameter
Vegetarian refreshments whether fault, the fault sum by counting entire image determines whether present image is flaw fabric, is embodied
Mode is as follows:
1, the gradient energy feature extraction based on integrogram:
In general, flawless textile image texture is uniformly distributed in the period, to wherein arbitrary point (x, y), it is with it
Energy under center calculation fixed size dw*dh window, ideally, the energy feature should not change with (x, y) and be become
Change, and the appearance of flaw will break its periodical, uniformity, but also Energy distribution changes.Accordingly, window can be used
Energy value describes to detect flaw as feature, but in view of energy value is vulnerable to illumination effect, therefore, the present invention uses gradient energy
Amount first seeks the gradient map G (x, y) of original image F (x, y), then extracts energy feature E (x, y) to G (x, y).Seek feature E
During (x, y), need to its neighborhood territory pixel of each pixels statistics and, if do not used accelerating algorithm, real-time will be difficult to protect
Card, therefore, the present invention propose to accelerate using integrogram.To image G (x, y), defining integration figure I (x, y) are as follows:
It can be seen that the numerical value of any point (x, y) is square formed by the upper left corner to current point (x, y) in source images G in integrogram I
The sum of the pixel value of all the points in shape frame.In order to accelerate operation, can the algorithm shown in formula (1) be quickly obtained integrogram:
I (x, y)=I (x-1, y)+I (x, y-1)-I (x-1, y-1)+G (x, y) (1)
After existing integrogram I (x, y), the pixel integration of any rectangular area can quickly be counted by I (x, y) on image G (x, y)
It calculates, as shown in Fig. 2, enabling I (x1,y1)=R (A), I (x2,y2)=R (A)+R (B), I (x3,y3)=R (A)+R (C), I (x4,y4)=
R (A)+R (B)+R (C)+R (D), R () function representation domain integral, therefore, the integral R (D) in the region D can calculate as follows:
R (D)=I (x4,y4)-I(x2,y2)-I(x3,y3)+I(x1,y1)
All only need operation three times that can seek wherein the sum of pixel value regardless of D area in rectangular area is much according to above formula,
Greatly reduce calculation amount.Accordingly, the specific steps of the gradient energy feature extracting method based on integrogram are as follows:
1) to textile image F (x, y), gradient map G (x, y) is sought；
2) the integrogram I (x, y) of G (x, y) is sought according to formula (1)；
3) its gradient energy figure E (x, y) is sought to arbitrary point (x, y), i.e. the size centered on point (x, y) is dw*dh
Window
Pixel integration in the domain of mouth region:
2, the threshold parameter study based on kernel function:
Ideally, to indefectible image, gradient energy E (x, y) should not change with (x, y) and be changed, and practical
The texture of different zones can not be completely the same on middle fabric, and noise may also be introduced in image acquisition process, therefore, gradient energy
Measure E (x, y) often in certain distribution, in order to obtain suitable threshold parameter to divide flaw, a kind of most straightforward approach is benefit
It is fitted the distribution of E (x, y) with Gauss model, and sets parameter to the mean value and variance of the Gaussian Profile.However, E (x, y)
Distribution be not always that different attenuation characteristics may be presented in the two sides of density peaks with symmetry, so if apparatus
There is the Gauss model fitting of symmetry to be easy to generate deviation.For this purpose, the present invention proposes to be fitted gradient energy using kernel function
Distribution, obtains its cuclear density probability distribution P (e):
Wherein, N is the pixel number of the indefectible image, { en| n=1,2,3...N } it is to the ladder of indefectible image zooming-out
Spend energy feature figure EtrainThe gradient energy data of each point in (x, y), b are kernel function bandwidth, c0For normalization coefficient, k (z)
For core profile function, the present invention selects Gaussian kernel profile function, i.e. k (z)=exp (- z/2), and z >=0 substitutes into formula (4) and solves.
For the peak value position for obtaining gradient energy distribution, the extreme value of cuclear density probability distribution P (e) is sought using average drifting method,
Energy distribution is divided into left and right two according to density peaks, calculates separately calculating of each portion's variance for Defect Detection threshold value, tool
Body step are as follows:
1) initial value μ is set0For the gradient energy mean value of current indefectible template image；
2) μ is enabled1=m (μ0)+μ0, wherein m (μ0) it is μ0The average drifting amount at place:
In above formula g (z)=- k ' (z), i.e. the derivation of Gaussian kernel profile function.
If 3) | μ1-μ0| < ε stops recycling and enables threshold parameter μ=μ1；Otherwise μ is enabled0=μ1And previous step is returned to, ε is indicated
Dimensionless；
4) threshold parameter σ is calculated according to formula (6)1, wherein S is the pixel sum on the right side of gradient energy distribution map peak value.
es∈{en| n=1,2...N } and es≥μ (6)
5) threshold parameter σ is calculated according to formula (7)2, wherein Z is the pixel sum on the left of gradient energy distribution map peak value.
ez∈{en| n=1,2...N } and ez≤μ (7)
To improve algorithm robustness, every indefectible image in training set is sought joining according to formula (5), (6), (7)
Number μ, σ1、σ2, then respectively take the average value of all indefectible imagesAs the threshold value ginseng during final Defect Detection
Number.
3, Defect Detection
In detection-phase, to each image to be detected, the gradient energy first, in accordance with first part based on integrogram is special
Extracting method is levied, its gradient energy characteristic pattern E is obtainedtest(x,y).If Etest(x, y) meets following formula, and then decision-point (x, y) is
Fault:
Wherein,For the threshold parameter obtained and to indefectible template image training, α is control system
Number.Total fault number is counted, if more than the threshold value T of settingdThen determine that the figure has flaw, TdThe reality that can be controlled by user according to quality
The adjustment of border demand.
Fig. 3,4,5,6 are implementation result figure of the present invention, and textile image to be detected derives from Hong Kong University's electric and electronic engineering
It is the textile image data set that industrial automation research laboratory provides, wherein window width dw and height dh are disposed as 25,
Control coefrficient α is set as 4.Fig. 3 shows that 3 textile images to be detected in data set, respectively (a), (b), (c), Fig. 4 are
The corresponding gradient energy characteristic pattern extracted, Fig. 5 are testing result of the present invention, and (a), (b), (c) in Fig. 3,4,5 respectively correspond one
Width image.From fig. 4, it can be seen that there is marked difference in the gradient energy feature of defect areas and normal texture region, illustrate constructed
Gradient energy feature can preferably distinguish defect areas and normal texture region.Mentioned algorithm testing result figure as seen from Figure 5
Flaw can be accurately positioned.Fig. 6 is the comparison of the method for the present invention and other methods, and (a) is fabric original image, (b) is present invention side
Method testing result, (c) is the testing result based on LBP feature, (d) is the testing result of RB algorithm.As seen from Figure 6, relative to
Other methods, testing result of the present invention are more accurate.The test statistics of entire data set are shown as threshold value TdIt is selected as 50
When, false retrieval image only 5 width, accuracy up to 97%, and under MATLAB platform the average each image processing time only need 56ms, and
Method of the existing other methods such as based on LBP feature needs up to tens of seconds under MATLAB platform, in existing research speed compared with
The fast method based on RB algorithm also needs 140ms under C language environment, when the mentioned algorithm of the present invention is handled under C language environment
Between will be shorter, realize not only high accuracy but also the detection of high real-time.
Claims (5)
1. a kind of fabric defects detection method based on integral nomography, it is characterized in that using nomography is integrated by arbitrary size
Summation operation abbreviation in image block is add operation three times, is used for Defect Detection with rapidly extracting gradient energy feature, specifically
Are as follows: first by carrying out image study to indefectible template, its gradient energy feature distribution is counted, gained feature distribution is non-right
Claim, be distributed using kernel function fit characteristic, the peak value of distribution is extracted in conjunction with average drifting method, then threshold value is adaptively sought by peak value
Parameter, the threshold parameter are used for the differentiation of subsequent flaw；Then, it to image to be detected, is sought by integral nomography each
The gradient energy of detection window where pixel, in conjunction with the threshold parameter, determine current pixel point whether fault, pass through statistics
The fault sum of entire image determines whether present image is flaw fabric；
Wherein, the extracting method of gradient energy feature are as follows: seek the gradient map G (x, y) of original image F (x, y) first, recycle
Integral nomography seeks the gradient energy characteristic pattern E (x, y) of G (x, y), and to any pixel (x, y), energy feature is with point
Size centered on (x, y) is the pixel integration in the window area of dw*dh.
2. fabric defects detection method according to claim 1, it is characterized in that seeking G's (x, y) using integral nomography
The specific steps of gradient energy characteristic pattern E (x, y) are as follows:
1) the integrogram I (x, y) of G (x, y) is sought
I (x, y)=I (x-1, y)+I (x, y-1)-I (x-1, y-1)+G (x, y) (1)
2) according to the characteristic of integrogram, its gradient energy characteristic pattern E (x, y) is sought to arbitrary point (x, y):
3. fabric defects detection method according to claim 1 or 2, it is characterized in that specific step is as follows:
1) indefectible template image is utilized, training study is used for the threshold parameter of Defect Detection
Training set is established with the image of indefectible fabric, to each width image, obtains the gradient energy characteristic pattern E of imagetrain(x,
Y), E is fitted using kernel functiontrainGradient energy distribution in (x, y), obtains cuclear density probability distribution P (e), utilizes mean value
Drift method iteration seeks the extreme value of cuclear density probability distribution P (e), obtains the peak value position of gradient energy distribution, is denoted as μ,
Energy distribution is divided into left and right two according to peak value, calculates separately two variances sigmas1、σ2, to the same kind fabric in training set
Every indefectible image seeks parameter μ, σ1、σ2Afterwards, then correspondence takes its average value As the final of the kind fabric
Threshold parameter during Defect Detection；
2) to each image to be detected, its gradient energy characteristic pattern E is obtainedtest(x, y), if Etest(x, y) meets following formula then
Decision-point (x, y) is fault:
Wherein,For the threshold parameter obtained in step 1), α is control coefrficient；
Total fault number is counted, finally if more than the threshold value T of settingdThen determine that the figure has flaw.
4. fabric defects detection method according to claim 3, it is characterized in that cuclear density probability distribution P (e) are as follows:
Wherein, N is the pixel sum of indefectible image, { en| the N of n=1,2,3 ... } be indefectible image in each point gradient
Energy datum, b are kernel function bandwidth, c0For normalization coefficient, k (z) is Gaussian kernel profile function, i.e. k (z)=exp (- z/2),
z≥0。
5. fabric defects detection method according to claim 3, it is characterized in that threshold parameter μ, σ1、σ2Specific calculating side
Method are as follows:
1) initial value μ is set0For the gradient energy mean value of current indefectible template image；
2) μ is enabled1=m (μ0)+μ0, wherein m (μ0) it is μ0The average drifting amount at place:
In above formula, N is the pixel sum of indefectible image, { en| the N of n=1,2,3 ... } be indefectible image in each point ladder
Energy datum is spent, b is kernel function bandwidth, and function g (z)=- k ' (z), k (z) is Gaussian kernel profile function；
If 3) | μ1-μ0| < ε enables parameter μ=μ1, carry out in next step, otherwise enabling μ0=μ1And return to previous step and be iterated, ε is indicated
Dimensionless；
4) threshold parameter σ is calculated according to formula (6)1:
es∈{en| n=1,2...N } and es≥μ (6)
Wherein, S is the pixel sum on the right side of gradient energy distribution map peak value；
5) threshold parameter σ is calculated according to formula (7)2
ez∈{en| n=1,2...N } and ez≤μ(7)
Wherein, Z is the pixel sum on the left of gradient energy distribution map peak value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610180226.3A CN107240086B (en) | 2016-03-25 | 2016-03-25 | A kind of fabric defects detection method based on integral nomography |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610180226.3A CN107240086B (en) | 2016-03-25 | 2016-03-25 | A kind of fabric defects detection method based on integral nomography |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107240086A CN107240086A (en) | 2017-10-10 |
| CN107240086B true CN107240086B (en) | 2019-05-14 |
Family
ID=59983575
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610180226.3A Active CN107240086B (en) | 2016-03-25 | 2016-03-25 | A kind of fabric defects detection method based on integral nomography |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107240086B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108090894B (en) * | 2017-11-24 | 2020-09-08 | 浙江理工大学 | Fabric defect detection method based on distance matching function and perceptual hash algorithm |
| CN109583306B (en) * | 2018-10-31 | 2021-10-08 | 杭州电子科技大学 | Bobbin residual yarn detection method based on machine vision |
| CN109934287A (en) * | 2019-03-12 | 2019-06-25 | 上海宝尊电子商务有限公司 | A kind of clothing texture method for identifying and classifying based on LBP and GLCM |
| TWI783200B (en) | 2019-12-30 | 2022-11-11 | 緯創資通股份有限公司 | Method and device for determining whether object has defect |
| CN114226262A (en) * | 2020-09-09 | 2022-03-25 | 宜谷京科技实业有限公司 | Flaw detection method, flaw classification method and flaw detection system |
| CN112308854B (en) * | 2020-11-26 | 2023-12-26 | 武汉工程大学 | Automatic detection method and system for chip surface flaws and electronic equipment |
-
2016
- 2016-03-25 CN CN201610180226.3A patent/CN107240086B/en active Active
Non-Patent Citations (5)
| Title |
|---|
| A fast regularity measure for surface defect detection;Du-Ming Tsai 等;《Machine Vision and Applications》;20120104;全文 * |
| A Shift-Tolerant Dissimilarity Measure for Surface Defect Detection;Du-Ming Tsai 等;《IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS》;20120228;第8卷(第1期);全文 * |
| 图像处理在仪表识别和布匹瑕疵检测中的应用;张北航;《中国优秀硕士学位论文全文数据库 信息科技辑》;20090115;全文 * |
| 瑕疵点检测算法研究及应用;王文杰;《中国优秀硕士学位论文全文数据库 信息科技辑》;20110715;全文 * |
| 钢珠瑕疵检测算法的研究;俞哲 等;《仪器仪表学报》;20091031;第30卷(第10期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107240086A (en) | 2017-10-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107240086B (en) | A kind of fabric defects detection method based on integral nomography | |
| Yao et al. | Curvature aided Hough transform for circle detection | |
| CN106251361B (en) | A kind of rail surface defects image self-adapting division method | |
| CN108898610A (en) | A kind of object contour extraction method based on mask-RCNN | |
| CN108090894B (en) | Fabric defect detection method based on distance matching function and perceptual hash algorithm | |
| CN105574527B (en) | A kind of quick object detecting method based on local feature learning | |
| CN106530347A (en) | Stable high-performance circle feature detection method | |
| CN104732536A (en) | Sub-pixel edge detection method based on improved morphology | |
| CN106157303A (en) | A kind of method based on machine vision to Surface testing | |
| Ni et al. | Automatic detection and counting of circular shaped overlapped objects using circular hough transform and contour detection | |
| CN110415208A (en) | A kind of adaptive targets detection method and its device, equipment, storage medium | |
| CN104408711A (en) | Multi-scale region fusion-based salient region detection method | |
| CN112258569B (en) | Pupil center positioning method, pupil center positioning device, pupil center positioning equipment and computer storage medium | |
| Hoseini et al. | Fabric defect detection using auto-correlation function | |
| CN108765409A (en) | A kind of screening technique of the candidate nodule based on CT images | |
| CN109509168B (en) | A kind of details automatic analysis method for picture quality objective evaluating dead leaf figure | |
| Selamet et al. | Automatic detection and classification of defective areas on metal parts by using adaptive fusion of faster R-CNN and shape from shading | |
| CN110276260B (en) | Commodity detection method based on depth camera | |
| CN102279925A (en) | Chain processing face recognition method and system | |
| Liu et al. | Adaptive ideal image reconstruction for bacteria colony detection | |
| CN111047595A (en) | Real-time sea wave segmentation method and device based on self-adaptive threshold frame difference method | |
| Wang et al. | Image segmentation method for an illumination highlight region of interior design effects based on the partial differential equation | |
| CN107730518A (en) | A kind of CANNY threshold algorithms value-acquiring method and device based on FPGA | |
| Santosh et al. | Text detection from documented image using image segmentation | |
| CN115984316B (en) | Industrial image edge extraction method and device for complex environment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20201030 Address after: 226019 No.205, building 6, Nantong University, No.9, Siyuan Road, Nantong City, Jiangsu Province Patentee after: Center for technology transfer, Nantong University Address before: 226019 Jiangsu city of Nantong province sik Road No. 9 Patentee before: NANTONG University |
|
| TR01 | Transfer of patent right | ||
| EE01 | Entry into force of recordation of patent licensing contract | ||
| EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20171010 Assignee: Nantong Shunteng Textile Co.,Ltd. Assignor: Center for technology transfer, Nantong University Contract record no.: X2021980016785 Denomination of invention: A fabric defect detection method based on integral graph algorithm Granted publication date: 20190514 License type: Common License Record date: 20211229 |
|
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: 226001 No.9, Siyuan Road, Chongchuan District, Nantong City, Jiangsu Province Patentee after: Nantong University Technology Transfer Center Co.,Ltd. Address before: 226019 No.205, building 6, Nantong University, No.9, Siyuan Road, Nantong City, Jiangsu Province Patentee before: Center for technology transfer, Nantong University |
|
| EE01 | Entry into force of recordation of patent licensing contract | ||
| EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20171010 Assignee: Yizhong (Nantong) Network Technology Co.,Ltd. Assignor: Nantong University Technology Transfer Center Co.,Ltd. Contract record no.: X2023980053327 Denomination of invention: A Fabric Defect Detection Method Based on Integral Graph Algorithm Granted publication date: 20190514 License type: Common License Record date: 20231221 |