CN109003275B - Segmentation method of weld defect image - Google Patents
Segmentation method of weld defect image Download PDFInfo
- Publication number
- CN109003275B CN109003275B CN201710417514.0A CN201710417514A CN109003275B CN 109003275 B CN109003275 B CN 109003275B CN 201710417514 A CN201710417514 A CN 201710417514A CN 109003275 B CN109003275 B CN 109003275B
- Authority
- CN
- China
- Prior art keywords
- image
- curve
- weld
- defect image
- column
- 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
- 230000007547 defect Effects 0.000 title claims abstract description 87
- 238000000034 method Methods 0.000 title claims abstract description 48
- 230000011218 segmentation Effects 0.000 title claims abstract description 23
- 238000007781 pre-processing Methods 0.000 claims abstract description 4
- 238000004422 calculation algorithm Methods 0.000 claims description 30
- 238000003466 welding Methods 0.000 claims description 19
- 238000012545 processing Methods 0.000 claims description 9
- 230000000877 morphologic effect Effects 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 abstract description 10
- 238000000605 extraction Methods 0.000 abstract description 8
- 230000009286 beneficial effect Effects 0.000 abstract description 2
- 239000000284 extract Substances 0.000 abstract 2
- 230000000694 effects Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 4
- 238000003709 image segmentation Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000004088 simulation Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 230000002950 deficient Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 210000000653 nervous system Anatomy 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
- G06T5/30—Erosion or dilatation, e.g. thinning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- 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
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
- G06T2207/20032—Median filtering
-
- 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
- G06T2207/30152—Solder
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Quality & Reliability (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a method for dividing a weld defect image. After preprocessing an image, the method for dividing the weld defect image firstly carries out gray level curve fitting on each column of pixel points on the image to obtain a column gray level curve, then extracts minimum value points on the curve, determines the size of a structural element according to the corresponding characteristics of the minimum value points, then simulates a weld background image by utilizing closed operation in morphology, and extracts the weld defect image by a digital subtraction technique. The method for segmenting the weld defect image can effectively solve the problem that the defect part of the weld defect image is difficult to extract due to complex textures of the weld defect image, realizes more reliable, more accurate and more universal segmentation and extraction of the weld defect image, and is beneficial to realizing automatic detection of the weld defect image.
Description
Technical Field
The invention relates to an image defect segmentation technology, in particular to a segmentation method of a weld defect image.
Background
The traditional welding defect detection method mainly relies on manual judgment on welding digital images (such as X-ray welding images or ultrasonic welding images and the like), and has the problems of low efficiency and high false detection rate. With the development of image processing technology, defect detection on digital welding images has become an important means for judging the quality of welding products. Welding digital images usually have low contrast, large background fluctuation and small noise, and are easy to cause missed judgment and misjudgment of defects.
In the conventional method, the detection of the welding defect mainly includes four processes. The first procedure is image acquisition. The second procedure is image preprocessing, which is mainly divided into two parts of image noise reduction and image enhancement, wherein the purpose of image noise reduction is to eliminate noise in a welding line image and filter invalid information in the image. The purpose of image enhancement is to improve the visual effect of the image, purposefully emphasize the whole or partial characteristics of the image, enlarge the differences between the characteristics of different parts in the image, improve the image quality and enrich the information. In general, when there is a certain degree of distinction between the defect and the background when the gray level distribution in the image is uniform, the third process may be performed. The third step is image segmentation or extraction segmentation of the image of the defective portion. In the following description of several prior art welding defect detection methods, the differences are mainly focused on this part, which is also a key step from image processing to image analysis. Regardless of the algorithm used, the principle is that the image segmentation aims at segmenting defects in the weld digital image so that the defects are easy to accurately identify by a computer. The last procedure is defect identification. In the detection of welding defects, the effect of image defect segmentation is good or bad, and the effect of welding seam digital image processing is most directly influenced.
The welding defect detection method with relatively good effect in the prior art mainly adopts a threshold segmentation method, a model method, an edge detection algorithm and the like aiming at image segmentation or defect image extraction parts. The Zhang Xiaoguang and Shafeek et al study the defects without removing the background by using a threshold segmentation method, and have good segmentation effect on images with larger contrast between the defects and the background gray level, but the weld edge is easy to be mistakenly regarded as a welding defect when the gray level of the weld edge is close to the defect gray level in the case of weld images with uneven gray level distribution. The model method mainly comprises a Pulse Coupled Neural Network (PCNN) model and a Chan-Vese (CV) model, wherein the PCNN is a novel neural network algorithm, the image processing result of the PCNN is more in line with the human visual nervous system, but the obtained defect edges are often rough due to the fact that the optimal iteration times of the PCNN are difficult to determine, and the result is easy to be interfered by noise. The CV model can effectively utilize the prior information knowledge of the image, but is sensitive to initial conditions and has low calculation efficiency. Luo Aimin the size of morphological structure elements is analyzed, a watershed-based adaptive mathematical morphology algorithm is provided, defect extraction of an X-ray welding image is achieved, and edge information of the defect can be extracted accurately, but omission is easy to generate.
Disclosure of Invention
The invention aims to solve the technical problems that in the prior art, an image defect segmentation method adopted for welding defect detection is easy to cause errors, extracted defect edges are too rough, segmentation results are easy to be interfered by noise, segmentation effect reliability is insufficient, universality is not high enough, and provides a weld defect image segmentation method.
The invention solves the technical problems by the following technical proposal:
the invention provides a method for dividing a weld defect image, which is characterized by comprising the following steps of:
step one, performing column gray scale curve fitting on each column of pixel points in a welding seam image to obtain a column gray scale curve corresponding to each column of pixel points;
step two, extracting minimum value points on the column gray scale curve;
step three, determining the corresponding structure element size based on the minimum value point according to the column gray scale curve corresponding to each column of pixel points, wherein,
when the number of minimum value points extracted from the column gray level curve is zero, determining that the size of a structural element is zero;
when the number of the minimum value points extracted from the column gray scale curve is not zero, finding out the steep drop point of the curve nearest to the left side and the vertex of the curve nearest to the right side of each minimum value point, calculating the distance between the steep drop point and the vertex of the curve nearest to the right side, taking the maximum value in the distances corresponding to all the minimum value points as the size of the structural element,
wherein, definition of the steep drop point of the curve meets the following conditions: the difference between the gray value of the curve abrupt-drop point and the gray value of the pixel point which is positioned on the right side of the curve abrupt-drop point on the column gray level curve and is at the preset abrupt-drop point step distance from the curve abrupt-drop point is not smaller than the preset gray value difference;
step four, obtaining a simulated weld background image according to morphological closing operation based on the determined size of the structural element;
and fifthly, carrying out digital silhouette on the weld joint image and the simulated weld joint background image to obtain difference information images of the two images.
Preferably, the weld image is a digital image of the weld formed after pretreatment.
Preferably, the preprocessing includes image noise reduction processing and image enhancement processing using a median filtering method.
Preferably, the second step further includes, after extracting the minimum value points on the column gray scale curves, excluding the minimum value points with gray scale fluctuation values smaller than a preset gray scale fluctuation threshold value in a neighborhood range of a preset step length.
Preferably, the method for segmenting the weld defect image further comprises the following steps:
and step six, segmenting the defect image in the difference information image by adopting a fuzzy K-means algorithm.
Preferably, in the sixth step, a fuzzy K-means algorithm based on kernel improvement is used to segment the defect image in the difference information image, wherein a membership degree adjustment coefficient and a constraint coefficient are introduced into the fuzzy K-means algorithm based on kernel improvement.
Preferably, the method for segmenting the weld defect image further comprises the following steps:
and step seven, adopting an expansion algorithm to process the defect image obtained by the step six segmentation.
On the basis of conforming to the common knowledge in the field, the above preferred conditions can be arbitrarily combined to obtain the preferred examples of the invention.
The invention has the positive progress effects that:
the method for segmenting the weld defect image can effectively solve the problem that the defect part of the weld defect image is difficult to extract due to complex textures of the weld defect image, realizes more reliable, more accurate and more universal segmentation and extraction of the weld defect image, and is beneficial to realizing automatic detection of the weld defect image.
Drawings
Fig. 1 is a schematic view of a weld image in an application example of a segmentation method of a weld defect image according to a preferred embodiment of the present invention.
FIG. 2 is a schematic diagram of an exemplary minimum point on a column gray scale curve with its left nearest curve dip point and its right nearest curve vertex.
Fig. 3 is a schematic diagram of a weld background image obtained by simulation based on the segmentation method of the weld defect image of fig. 1 according to the preferred embodiment of the present invention.
Fig. 4 is a schematic diagram of a difference information image obtained by using a digital silhouette technique based on the method for dividing a weld defect image according to the preferred embodiment of the present invention shown in fig. 1 and 3.
FIG. 5 is a schematic view of a weld defect image extracted by a fuzzy K-means algorithm based on kernel improvement based on the segmentation method of the weld defect image of FIG. 4 according to the preferred embodiment of the present invention.
Fig. 6 is a schematic diagram of a weld defect image processed by an expansion algorithm based on the segmentation method of the weld defect image of fig. 5 according to the preferred embodiment of the present invention.
Detailed Description
The following detailed description of the preferred embodiments of the invention, taken in conjunction with the accompanying drawings, is given by way of illustration and not limitation, and any other similar situations are intended to fall within the scope of the invention.
In the following detailed description, directional terms, such as "left", "right", "upper", "lower", "front", "rear", etc., are used with reference to the directions described in the drawings. The components of embodiments of the present invention can be positioned in a number of different orientations and the directional terminology is used for purposes of illustration and is in no way limiting.
According to the preferred embodiment of the invention, the method for segmenting the weld defect image comprises the following steps:
step one, performing column gray scale curve fitting on each column of pixel points in a welding seam image to obtain a column gray scale curve corresponding to each column of pixel points;
extracting minimum value points on the gray scale curves, and preferably, removing the minimum value points with gray scale fluctuation values smaller than a preset gray scale fluctuation threshold value in a neighborhood range of a preset step length from the minimum value points after extracting the minimum value points, so as to remove partial minimum value points generated by pseudo defects;
step three, determining the corresponding structure element size based on the minimum value point according to the column gray scale curve corresponding to each column of pixel points, wherein,
when the number of minimum value points extracted from the column gray level curve is zero, determining that the size of a structural element is zero;
when the number of the minimum value points extracted from the column gray scale curve is not zero, finding out the steep drop point of the curve nearest to the left side and the vertex of the curve nearest to the right side of each minimum value point, calculating the distance between the steep drop point and the vertex of the curve nearest to the right side, taking the maximum value in the distances corresponding to all the minimum value points as the size of the structural element,
wherein, definition of the steep drop point of the curve meets the following conditions: the difference between the gray value of the curve abrupt-drop point and the gray value of the pixel point which is positioned on the right side of the curve abrupt-drop point on the column gray level curve and is at the preset abrupt-drop point step distance from the curve abrupt-drop point is not smaller than the preset gray value difference;
step four, obtaining a simulated weld background image according to morphological closing operation based on the determined size of the structural element;
and fifthly, carrying out digital silhouette on the weld joint image and the simulated weld joint background image to obtain difference information images of the two images.
And preferably, the method for segmenting the weld defect image may further include, after the fifth step:
step six, segmenting a defect image in the difference information image by adopting a fuzzy K-means algorithm based on kernel improvement, wherein a membership degree adjusting coefficient and a constraint coefficient are introduced into the fuzzy K-means algorithm based on kernel improvement;
and step seven, adopting an expansion algorithm to process the defect image obtained by the step six segmentation.
According to the preferred embodiment of the invention, the problem of difficult segmentation of weld defects caused by low contrast and large background fluctuation of most of weld digital images can be well solved. An application example of the segmentation method of the weld defect image according to the above preferred embodiment of the present invention will be described in more detail below.
Referring to fig. 1, which shows a preprocessed weld image, according to the first step, first, the pixel points of each column in the weld image are extracted, and the column gray scale curves thereof are fitted by using a Cftool kit such as matlab software. And then, extracting minimum value points in the column gray scale curve obtained after fitting according to the second step and the third step, and further determining the sizes of the structural elements. A specific method of determining the size of the structural element corresponding to the column gray scale curve can be as follows.
First, the obtained gray-scale sequence curve is discretized, and each point of the obtained curve is (x 1 ，y 1 )，(x 2 ，y 2 )…(x n ，y n ) Wherein n is greater than or equal to 2; x and y are the pixel position and pixel value (i.e., the gray value of the pixel) of each point, respectively. The minimum point is obtained by the following formula (1).
y i -y i-1 < 0 and y i -y i+1 ＜0 (1)
In the formula (1), y i Is the pixel value of the ith point in the curve, point (x i ，y i ) Namely, isMinimum point of the curve.
Then, through the setting of the preset gray level fluctuation threshold value and the preset step length, the minimum value points are screened by the following formula (2), namely, the minimum value points generated by the pseudo defects are eliminated.
y i+α -y i ＞β (2)
In the formula (2), alpha is a preset step length, and beta is a gray scale fluctuation threshold value.
The following step three exemplifies a method for determining the size of the structural element.
In the first example, the column gray scale curve has a (gray scale) minimum point, and referring to fig. 2, a search is made to the left at the minimum point x to find a curve steep drop point a (x a ，y a ). That is, in this example, the preset steep drop point step distance is 7, and the preset gray value difference is 15. Then, the curve vertex B closest to the nearest curve is searched rightward from the point x (x b ，y b ). From this, it can be determined that the structural element size d (x) =x b -x a 。
|y a+7 -y a |≥15 (3)
It should be appreciated that in other examples, there may be multiple minima points in the column gray scale curve. In this case, the above calculation process may be repeated for each minimum point, and finally, the maximum value of the calculation result therein is taken as the structure element size of the corresponding column gray scale curve.
And simulating the image background by using morphological closing operation to obtain a simulated weld background image according to the sizes of the structural elements of each column of pixel points in the obtained image.
The results of background simulation on the digital image of the weld in fig. 1 are shown in fig. 3. It can be seen from the figure that the defect information in the simulated background image has been completely covered and assimilated with the background information. At this time, the difference information image (for example, shown in fig. 4) can be obtained by performing a difference operation on the preprocessed weld image (for example, shown in fig. 1) and the simulated weld background image (for example, shown in fig. 3) by using a digital silhouette technology, so as to realize extraction of the weld defect.
It should be appreciated that the difference information image obtained so far has enabled the present invention to achieve a more reliable, more accurate, more versatile segmented extraction of weld defect images than the methods of the prior art.
After the digital silhouette is completed, a defect image in the difference information image is segmented by adopting a fuzzy K-means algorithm based on kernel improvement in the step six, so that the weld defect image can be segmented and extracted better.
In view of the fact that the general fuzzy K-means algorithm is relatively high in randomness and is susceptible to isolated points, the algorithm can be improved based on some characteristics of weld defects. The improved algorithm transforms the sample space points into a high-dimensional space for screening and classifying, and the membership degree adjusting coefficient alpha and the constraint coefficient beta are added on the basis of the original algorithm, so that the anti-interference performance of the algorithm on isolated points and the reliability of the algorithm are improved. The specific principle of the improved K-means algorithm is shown in the following formula (4).
In the formula (4), x j As a point of the sample,
from image data space R for sample points s Mapping relation to feature space H, W= [ W ] 1 ，w 2 ，...w c ]And c represents the number of the clustering centers in the feature space H.
The expression of the cluster center in the feature space H is as follows formula (5):
order the
Thereby representing x j Mapped and clustered with center w i Distance in the feature space H. The following are the followingThree situations can be distinguished.
1. If j (1. Ltoreq.j. Ltoreq.n) is absent for any i (1. Ltoreq.i. Ltoreq.k) such that D ij When=0, the following relationship exists:
2. if j (1. Ltoreq.j. Ltoreq.n) is present for any i (1. Ltoreq.i.ltoreq.c) such that D ij When=0, u ij Is any non-negative real number that satisfies the following conditions:
3. when i (1.ltoreq.i.ltoreq.c) is present, D is present for any j (1.ltoreq.j.ltoreq.n) ij When=0, u ij Is any non-negative real number that satisfies the following condition:
calculating according to the formula and eliminating w i The following formula (9) can be obtained,
wherein the method comprises the steps of
Alpha is a membership degree adjusting coefficient, mainly reflects the membership relation between a space point and a clustering center, represents the classification condition of the algorithm on pixel points, beta is a constraint coefficient, mainly reflects the constraint force of mapping from an original data space to a high-dimensional feature space H, and represents the clearing capacity of the algorithm on isolated points.
The fuzzy K-means algorithm based on kernel improvement includes the specific steps of setting membership regulating coefficient alpha and constraint coefficient beta, convergence precision epsilon and fuzzy index m, and initialAnd (5) the membership degree matrix is converted, and the initial value of the iteration number k is 0. D is then calculated according to the above formula (9) ij Then calculate
Obtaining a fuzzy clustering matrix U k Iteration is continuously carried out until the termination condition U is met k -U k-1 ε or i (1. Ltoreq.i.ltoreq.c) is present such that +.>
Until that point.
The defect image extracted by the kernel-based modified fuzzy K-means algorithm is shown in fig. 5. After the defects in the image are completely segmented, the extracted defect image is subjected to expansion processing to smooth the defect boundary and make the defect morphology as close as possible to reality, and finally the extracted defect is shown in fig. 6.
The invention is based on morphological filtering background simulation, defects can be rapidly and accurately extracted from the weld joint digital image under the complex background through proper determination of the sizes of the structural elements, and the accuracy and universality of defect extraction results are further improved through introducing membership degree adjusting coefficients and constraint coefficients into an improved fuzzy K-means algorithm.
While specific embodiments of the invention have been described above, it will be appreciated by those skilled in the art that these are by way of example only, and the scope of the invention is defined by the appended claims. Various changes and modifications to these embodiments may be made by those skilled in the art without departing from the principles and spirit of the invention, but such changes and modifications fall within the scope of the invention.
Claims (7)
1. The method for segmenting the weld defect image is characterized by comprising the following steps of:
step one, performing column gray scale curve fitting on each column of pixel points in a welding seam image to obtain a column gray scale curve corresponding to each column of pixel points;
step two, extracting minimum value points on the column gray scale curve;
step three, determining the corresponding structure element size based on the minimum value point according to the column gray scale curve corresponding to each column of pixel points, wherein,
when the number of minimum value points extracted from the column gray level curve is zero, determining that the size of a structural element is zero;
when the number of the minimum value points extracted from the column gray scale curve is not zero, finding out the steep drop point of the curve nearest to the left side and the vertex of the curve nearest to the right side of each minimum value point, calculating the distance between the steep drop point and the vertex of the curve nearest to the right side, taking the maximum value in the distances corresponding to all the minimum value points as the size of the structural element,
wherein, definition of the steep drop point of the curve meets the following conditions: the difference between the gray value of the curve abrupt-drop point and the gray value of the pixel point which is positioned on the right side of the curve abrupt-drop point on the column gray level curve and is at the preset abrupt-drop point step distance from the curve abrupt-drop point is not smaller than the preset gray value difference;
step four, obtaining a simulated weld background image according to morphological closing operation based on the determined size of the structural element;
and fifthly, carrying out digital silhouette on the weld joint image and the simulated weld joint background image to obtain difference information images of the two images.
2. The method of segmenting a weld defect image of claim 1, wherein the weld image is a pre-processed digital image of the weld.
3. The method for segmenting the weld defect image according to claim 2, wherein the preprocessing includes an image noise reduction processing and an image enhancement processing using a median filtering method.
4. The method of claim 1, wherein the second step further comprises, after extracting the minimum points on the gray scale curve, excluding the minimum points from which the gray scale fluctuation value is smaller than the preset gray scale fluctuation threshold value in the neighborhood of the preset step.
5. The method of segmenting a weld defect image of claim 1, further comprising the steps of:
and step six, segmenting the defect image in the difference information image by adopting a fuzzy K-means algorithm.
6. The method for segmenting the weld defect image according to claim 5, wherein in the sixth step, a fuzzy K-means algorithm based on kernel improvement is used to segment the defect image in the difference information image, wherein a membership adjustment coefficient and a constraint coefficient are introduced into the fuzzy K-means algorithm based on kernel improvement.
7. The method of segmenting a weld defect image of claim 6, further comprising the steps of:
and step seven, adopting an expansion algorithm to process the defect image obtained by the step six segmentation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710417514.0A CN109003275B (en) | 2017-06-06 | 2017-06-06 | Segmentation method of weld defect image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710417514.0A CN109003275B (en) | 2017-06-06 | 2017-06-06 | Segmentation method of weld defect image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109003275A CN109003275A (en) | 2018-12-14 |
| CN109003275B true CN109003275B (en) | 2023-04-28 |
Family
ID=64573401
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710417514.0A Active CN109003275B (en) | 2017-06-06 | 2017-06-06 | Segmentation method of weld defect image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109003275B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109636800B (en) * | 2019-01-28 | 2022-10-14 | 中国科学院上海硅酸盐研究所 | Method for measuring size of internal defect of object |
| CN110310295B (en) * | 2019-03-27 | 2021-09-14 | 广东技术师范学院天河学院 | Weld contour extraction method and system |
| CN109993741B (en) * | 2019-04-03 | 2022-10-04 | 南昌航空大学 | Steel rail welding seam contour automatic positioning method based on K-means clustering |
| CN111811437A (en) * | 2020-09-10 | 2020-10-23 | 湖南泰嘉新材料科技股份有限公司 | Band saw blade welding quality detection method and device |
| CN114965516B (en) * | 2022-08-01 | 2022-10-25 | 山东易斯特工程工具有限公司 | Shield cutter head brazing defect detection method based on X-ray flaw detection |
| CN115861307B (en) * | 2023-02-21 | 2023-04-28 | 深圳市百昌科技有限公司 | Fascia gun power supply driving plate welding fault detection method based on artificial intelligence |
| CN116503394B (en) * | 2023-06-26 | 2023-09-08 | 济南奥盛包装科技有限公司 | Printed product surface roughness detection method based on image |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103323526A (en) * | 2013-05-30 | 2013-09-25 | 哈尔滨工业大学 | Welding line defect detection and identification method based on ultrasonic phased array and support vector machine |
| CN104036495A (en) * | 2014-05-22 | 2014-09-10 | 南京航空航天大学 | Welding defect extraction method and welding defect detection method |
| CN104751474A (en) * | 2015-04-13 | 2015-07-01 | 上海理工大学 | Cascade quick image defect segmentation method |
| CN105205821A (en) * | 2015-09-21 | 2015-12-30 | 江苏科技大学 | Weld image segmentation method |
| CN106770318A (en) * | 2016-12-07 | 2017-05-31 | 江南大学 | Weld defect real-time detection method based on improved background subtraction |
-
2017
- 2017-06-06 CN CN201710417514.0A patent/CN109003275B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103323526A (en) * | 2013-05-30 | 2013-09-25 | 哈尔滨工业大学 | Welding line defect detection and identification method based on ultrasonic phased array and support vector machine |
| CN104036495A (en) * | 2014-05-22 | 2014-09-10 | 南京航空航天大学 | Welding defect extraction method and welding defect detection method |
| CN104751474A (en) * | 2015-04-13 | 2015-07-01 | 上海理工大学 | Cascade quick image defect segmentation method |
| CN105205821A (en) * | 2015-09-21 | 2015-12-30 | 江苏科技大学 | Weld image segmentation method |
| CN106770318A (en) * | 2016-12-07 | 2017-05-31 | 江南大学 | Weld defect real-time detection method based on improved background subtraction |
Non-Patent Citations (2)
| Title |
|---|
| 基于B样条曲线的X射线图像焊缝缺陷分割与提取;梁硼等;《焊接学报》;20120725(第07期);全文 * |
| 铝锂合金双激光束双侧同步焊接各向拉伸性能分析;占小红 等;《焊接学报》;20150525;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109003275A (en) | 2018-12-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109003275B (en) | Segmentation method of weld defect image | |
| CN107657279B (en) | Remote sensing target detection method based on small amount of samples | |
| CN109598287B (en) | Appearance flaw detection method for resisting network sample generation based on deep convolution generation | |
| CN106251361B (en) | A kind of rail surface defects image self-adapting division method | |
| CN106778595B (en) | Method for detecting abnormal behaviors in crowd based on Gaussian mixture model | |
| CN107610087B (en) | Tongue coating automatic segmentation method based on deep learning | |
| CN111563893B (en) | Grading ring defect detection method, device, medium and equipment based on aerial image | |
| CN108537751B (en) | Thyroid ultrasound image automatic segmentation method based on radial basis function neural network | |
| CN102567744B (en) | Method for determining quality of iris image based on machine learning | |
| CN105046259B (en) | Coronal mass ejection detection method based on multi-feature fusion | |
| CN107403438A (en) | Improve the ultrasonoscopy focal zone dividing method of fuzzy clustering algorithm | |
| CN110598030A (en) | Oracle bone rubbing classification method based on local CNN framework | |
| CN108550131B (en) | SAR image vehicle detection method based on feature fusion sparse representation model | |
| CN104751474A (en) | Cascade quick image defect segmentation method | |
| CN106340007A (en) | Image processing-based automobile body paint film defect detection and identification method | |
| CN113435460A (en) | Method for identifying brilliant particle limestone image | |
| CN116740044B (en) | Copper pipe milling surface processing method and system based on visual detection and control | |
| CN111460953A (en) | Electrocardiosignal classification method based on self-adaptive learning of countermeasure domain | |
| CN112991271A (en) | Aluminum profile surface defect visual detection method based on improved yolov3 | |
| CN111783616A (en) | Data-driven self-learning-based nondestructive testing method | |
| CN112183237A (en) | Automatic white blood cell classification method based on color space adaptive threshold segmentation | |
| CN112396580B (en) | Method for detecting defects of round part | |
| Nguyen et al. | Pavement crack detection and segmentation based on deep neural network | |
| CN115147613A (en) | Infrared small target detection method based on multidirectional fusion | |
| CN113205540A (en) | Multi-scale automatic anisotropic morphological direction derivative edge detection method |
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 |