CN115294376A - Weld defect detection method based on fusion of ultrasonic waveform and ultrasonic image characteristics - Google Patents
Weld defect detection method based on fusion of ultrasonic waveform and ultrasonic image characteristics Download PDFInfo
- Publication number
- CN115294376A CN115294376A CN202210433330.4A CN202210433330A CN115294376A CN 115294376 A CN115294376 A CN 115294376A CN 202210433330 A CN202210433330 A CN 202210433330A CN 115294376 A CN115294376 A CN 115294376A
- Authority
- CN
- China
- Prior art keywords
- feature
- ultrasonic
- fusion
- vector
- vectors
- 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
- 230000004927 fusion Effects 0.000 title claims abstract description 43
- 230000007547 defect Effects 0.000 title claims abstract description 37
- 238000001514 detection method Methods 0.000 title claims abstract description 34
- 239000013598 vector Substances 0.000 claims abstract description 115
- 238000011176 pooling Methods 0.000 claims abstract description 20
- 230000003044 adaptive effect Effects 0.000 claims abstract description 8
- 238000000034 method Methods 0.000 claims abstract description 6
- 238000007500 overflow downdraw method Methods 0.000 claims abstract description 6
- 238000000354 decomposition reaction Methods 0.000 claims description 15
- 238000003466 welding Methods 0.000 claims description 10
- 229910000831 Steel Inorganic materials 0.000 abstract description 4
- 239000010959 steel Substances 0.000 abstract description 4
- 230000035515 penetration Effects 0.000 description 3
- 239000011148 porous material Substances 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000002592 echocardiography Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000009659 non-destructive testing Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000013441 quality evaluation Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 239000002893 slag Substances 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Investigating Or Analyzing Materials By The Use Of Ultrasonic Waves (AREA)
Abstract
The invention discloses a weld defect detection method based on fusion of ultrasonic waveform and ultrasonic image characteristics, which solves the problems of long detection time and low efficiency of a manual ultrasonic detection method in the prior art. The method can quickly and accurately detect the weld defects, and improve the accuracy of ultrasonic detection of the weld defects of the steel pipe under the complex background. The invention comprises the following steps: step 1, extracting 3-layer wavelet packet characteristics of ultrasonic waveforms, and calculating the sample entropy of each node characteristic signal to obtain an energy characteristic vector; step 2, extracting multi-scale convolution characteristics of the ultrasonic image by using an improved IncepotionV 2 network, and then carrying out global average pooling and full connection to obtain multi-scale convolution characteristic vectors; step 3, constructing a feature adaptive fusion method, fusing the wavelet packet feature vector and the multi-scale convolution feature vector to obtain a fusion feature vector; and 4, inputting the fusion characteristic vector into a SoftMax classifier, and identifying the type of the weld defects.
Description
The technical field is as follows:
the invention belongs to the technical field of steel pipe weld defect detection, and relates to a weld defect detection method based on fusion of ultrasonic waveform and ultrasonic image characteristics.
Background art:
the welding component is widely applied to the industrial fields of boilers, pressure vessels, railways, marine shipbuilding, oil pipelines, aerospace and the like. The accurate detection of the defects at the welding seam has great significance for ensuring the quality of the welding seam and the industrial safety production. The welding seam defects are of various types, including pores, cracks, inclusions, incomplete fusion, incomplete penetration and the like, and have great potential safety hazards in normal use of the steel pipe, so that great economic loss and safety accidents can be caused. Ultrasonic nondestructive testing is one of the most common methods for testing because of its advantages of high penetration, accurate defect localization, low cost, no pollution, high sensitivity, etc. At present, a manual ultrasonic detection method is still widely adopted for in-service welding structures in the industrial field. The method has the advantages of long detection time, low efficiency, different detection results according to different people, incapability of realizing automatic storage of detection data, low detection efficiency and easiness in detection omission. Especially for defect type identification has been a difficult problem of conventional ultrasonic detection methods. The convolutional neural network has remarkable characteristic extraction capability, is widely applied to the field of image detection and identification, and provides a new idea for detecting and identifying the welding defects.
The invention content is as follows:
the invention aims to provide a weld defect detection method based on fusion of ultrasonic waveform and ultrasonic image characteristics, which solves the problems of long detection time and low efficiency of a manual ultrasonic detection method in the prior art. The method can quickly and accurately detect the weld defects, and improve the accuracy of ultrasonic detection of the weld defects of the steel pipe under the complex background.
In order to realize the purpose, the invention adopts the technical scheme that:
a weld defect detection method based on ultrasonic waveform and ultrasonic image feature fusion is characterized in that: the method comprises the following steps:
step 1, extracting 3-layer wavelet packet characteristics of ultrasonic waveforms, and calculating the sample entropy of each node characteristic signal to obtain an energy characteristic vector;
step 2, extracting multi-scale convolution characteristics of the ultrasonic image by using an improved IncepotionV 2 network, and then carrying out global average pooling and full connection to obtain multi-scale convolution characteristic vectors;
step 3, constructing a feature adaptive fusion method, fusing the wavelet packet feature vector and the multi-scale convolution feature vector to obtain a fusion feature vector;
and 4, inputting the fusion feature vector into a SoftMax classifier, and identifying the type of the weld defects.
The step (1) comprises the following steps:
performing 3-layer wavelet packet decomposition on the welding seam ultrasonic waveform to obtain 8 node signals of the 3 rd layer, and recording the signals as S 3,i (i =0,1.., 7), then calculating the energy of each node, normalizing the calculated energy, and constructing a feature vector
x i,k To reconstruct the signal S 3,i Discrete point amplitude of (i =0,1.., 7), n is wavelet decomposition sequence length, E j,i Representing the energy value of the ith node at decomposition level j.
The step (2) comprises the following steps:
the improved Inception V2 network is composed of convolution 3 x 3, maximum pooling 2 x 2, convolution 3 x 3, maximum pooling 2 x 2, improved Inception V2, global average pooling 2 x 2 and full connection, wherein the improved Inception V2 module is formed by adding an average pooling layer in the Inception V2 module, widening each layer network by a 1 x 1 convolution kernel, a 3 x 3 convolution kernel and the maximum pooling layer, reducing the number of channels through 1 x 1 convolution and performing feature clustering, stacking and outputting features in the depth direction, and obtaining feature vectors with the same dimensionality as the wavelet packet features through the global average pool and the full connection layer.
In the step (3), the feature adaptive fusion method comprises the following steps:
step 1, inputting the feature vector extracted by the wavelet packet into a shared full-connection layer 1, and outputting ultrasonic waveform feature importance vectors with the same dimensionality when outputting the feature vector;
step 2, inputting the feature vectors extracted by the improved Inception V2 network into the shared full-connection layer 2, and outputting ultrasonic image feature importance vectors with the same dimensionality when outputting the feature vectors;
step 3, normalizing the importance vectors of the two types of characteristics through a Sigmoid function to enable the sum of the importance vectors and the Sigmoid function to be 1, and calculating an ultrasonic image characteristic importance coefficient and an ultrasonic waveform characteristic importance coefficient;
and 4, weighting and adding the ultrasonic waveform characteristic vector and the ultrasonic image characteristic vector to obtain a self-adaptive fusion characteristic vector.
The step (3) comprises the following steps:
fusing the feature vectors obtained in the step 1 and the step 2 by using a feature self-adaptive fusion classification module; inputting the feature vectors extracted from the wavelet packets into the shared full-connection layer 1, and outputting ultrasonic waveform feature importance vectors with the same dimensionality when outputting the feature vectors; inputting the multi-scale feature vector extracted by the improved Inception V2 network into the shared full-connection layer 2, and outputting the ultrasonic image feature importance vector with the same dimension when outputting the feature vector; normalizing the importance vectors of the two types of characteristics through a Sigmoid function to enable the sum of the importance vectors and the Sigmoid function to be 1, and calculating an ultrasonic image characteristic importance coefficient and an ultrasonic waveform characteristic importance coefficient; and multiplying the ultrasonic image characteristic vector and the ultrasonic shape characteristic vector and adding to obtain a self-adaptive fusion characteristic vector: energy characteristics of wavelet packetThe eigenvector is
And corresponding importance coefficient alpha i (i =0,1.., 7.) image multi-scale convolution feature vector is denoted as v i (i =0,1, 7) and corresponding importance coefficient is β i (i =0,1, ·, 7); the fusion feature vector obtained in the fusion module is recorded as f i (i =0,1, 7), then f i ＝α i t i +β i v i (i =0,1,. 7), the resulting fused feature vector is F = (F =) (F) 0 ,f 1 ,...,f 7 )，E j,i Representing the energy value of the ith node at decomposition level j.
Compared with the prior art, the invention has the following advantages and effects:
the invention relates to a weld defect detection method based on fusion of ultrasonic waveform and ultrasonic image characteristics, which fully utilizes ultrasonic waveforms and ultrasonic images, realizes effective detection of weld defects through data fusion of the ultrasonic waveform characteristics and the ultrasonic image characteristics of defect echoes, can be used for detecting weldments with typical welding defects of pores, slag inclusion, cracks, incomplete penetration and incomplete fusion, is simple and easy to realize, has higher detection accuracy, can provide technical support for an actual weld defect detection system, and is beneficial to weld quality evaluation.
Description of the drawings:
FIG. 1 is a flowchart of a weld defect detection method based on fusion of ultrasonic waveforms and ultrasonic image features according to an embodiment of the present invention;
fig. 2 is an improved inclusion v2 network in accordance with an embodiment of the present invention;
figure 3 is a modified inclusion v2 module according to embodiments of the present invention;
fig. 4 is a feature adaptive fusion classification module according to an embodiment of the present invention.
The specific implementation mode is as follows:
in order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The invention relates to a weld defect detection method based on ultrasonic waveform and ultrasonic image feature fusion, which comprises the following steps:
step 1, extracting 3-layer wavelet packet characteristics of ultrasonic waveforms, and calculating the sample entropy of each node characteristic signal to obtain an energy characteristic vector;
step 2, extracting multi-scale convolution characteristics of the ultrasonic image by using an improved IncepotionV 2 network, and then carrying out global average pooling and full connection to obtain multi-scale convolution characteristic vectors;
step 3, constructing a feature adaptive fusion method, fusing the wavelet packet feature vector and the multi-scale convolution feature vector to obtain a fusion feature vector;
step 4, inputting the fusion characteristic vector into a SoftMax classifier, and identifying the type of the weld defects;
the invention specifically comprises the following steps:
step 1, extracting wavelet packet characteristic vectors of ultrasonic waveforms. And carrying out wavelet packet decomposition on the ultrasonic waveform to obtain two node signals of low frequency (S) and high frequency (d) on the 1 st layer, respectively carrying out wavelet packet decomposition on the low frequency (S) and high frequency (d) signals again to obtain 4 node signals on the 2 nd layer, and then carrying out decomposition to obtain 8 node signals on the 3 rd layer. Then calculating energy normalization characteristic vectors of decomposed signals of each order after wavelet packet decomposition;
and 2, extracting the multi-scale convolution characteristic vector of the ultrasonic image. And inputting the ultrasonic image into an improved IncepotionV 2 network to obtain a convolution map, and then performing global maximum pooling and full connection to obtain a multi-scale convolution feature vector.
And 3, constructing a feature self-adaptive fusion classification module, and fusing the extracted wavelet packet feature vector and the multi-scale convolution feature vector. Inputting the feature vectors extracted from the wavelet packets into the shared full-connection layer 1, and outputting ultrasonic waveform feature importance vectors with the same dimensionality when outputting the feature vectors; inputting the multi-scale feature vector extracted by the improved Inception V2 network into the shared full-connection layer 2, and outputting the ultrasonic image feature importance vector with the same dimension when outputting the feature vector; normalizing the importance vectors of the two types of characteristics through a Sigmoid function to enable the sum of the importance vectors and the Sigmoid function to be 1, and calculating an ultrasonic image characteristic importance coefficient and an ultrasonic waveform characteristic importance coefficient; and the ultrasonic image characteristic vector and the ultrasonic shape characteristic vector are added in a weighting mode to obtain a self-adaptive fusion characteristic vector.
And 4, classifying the defect types by using a SoftMax classifier.
Example (b):
referring to fig. 1, the invention provides a weld defect detection method based on ultrasonic waveform and ultrasonic image feature fusion, comprising the following steps:
step 1, performing 3-layer wavelet packet decomposition on the welding seam ultrasonic waveform to obtain 8 node signals of a 3 rd layer, and recording the 8 node signals as S 3,i (i =0,1.., 7), then calculating the energy of each node, normalizing the calculated energy, and constructing a feature vector
x i,k For reconstructing the signal S 3,i Discrete point amplitude of (i =0,1, 7), n is wavelet decomposition sequence length, E j,i Representing the energy value of the ith node at decomposition time j.
Step 2, referring to fig. 2, constructing an improved inclusion v2 network, specifically: the improved IncepotionV 2 network is composed of convolution 3 x 3, maximum pooling 2 x 2, convolution 3 x 3, maximum pooling 2 x 2, improved IncepotionV 2, global average pooling 2 x 2 and full connection, features are output in a superposition mode along the depth direction, and feature vectors with the same dimension as the wavelet packet features are obtained through the global average pools and the full connection layers. Referring to fig. 3, in the improved inclusion v2 module, an average pooling layer is added to the inclusion v2 module, and then each layer of network is widened by a 1 × 1 convolution kernel, a 3 × 3 convolution kernel and a maximum pooling layer, and the 1 × 1 convolution is used to reduce the number of channels and perform feature clustering;
and inputting the ultrasonic image into the trained improved IncepotionV 2 network, and obtaining a feature vector with the same dimension as the wavelet packet feature through a global average pool and a full connection layer.
And 3, referring to fig. 4, fusing the feature vectors obtained in the steps 1 and 2 by using a feature adaptive fusion classification module. Inputting the feature vectors extracted from the wavelet packets into the shared full-connection layer 1, and outputting ultrasonic waveform feature importance vectors with the same dimensionality when outputting the feature vectors; inputting the multi-scale feature vector extracted by the improved Inception V2 network into the shared full-connection layer 2, and outputting the ultrasonic image feature importance vector with the same dimension when outputting the feature vector; normalizing the importance vectors of the two types of characteristics through a Sigmoid function to enable the sum of the importance vectors and the Sigmoid function to be 1, and calculating an ultrasonic image characteristic importance coefficient and an ultrasonic waveform characteristic importance coefficient; and multiplying the ultrasonic image characteristic vector and the ultrasonic shape characteristic vector and adding to obtain a self-adaptive fusion characteristic vector: let the wavelet packet energy eigenvector be
And corresponding importance coefficient alpha i (i =0,1, 7), image multiscale convolution feature vector denoted v i (i =0,1.., 7) and corresponding importance coefficient is β i (i =0,1, 7). The fusion feature vector obtained in the fusion module is recorded as f i (i =0,1, 7), then f i ＝α i t i +β i v i (i =0,1.., 7), the resulting fused feature vector is F = (F =) (F) 0 ,f 1 ,...,f 7 )；
Step 4. F = (F) 0 ,f 1 ,...,f 7 ) And inputting the defect type into a SoftMax classifier for defect type classification.
Updating model parameters by using a loss function gradient and random gradient descent method to reduce residual errors between model output and label values so as to complete the training of the model; and testing and parameter adjustment are carried out on the model by utilizing the artificial defects on the comparison sample.
The above description is only for the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention, and all equivalent structural changes made by using the contents of the specification and the drawings of the present invention should be included in the scope of the present invention.
Claims (5)
1. A weld defect detection method based on ultrasonic waveform and ultrasonic image feature fusion is characterized in that: the method comprises the following steps:
step 1, extracting 3-layer wavelet packet characteristics of ultrasonic waveforms, and calculating the sample entropy of each node characteristic signal to obtain an energy characteristic vector;
step 2, extracting multi-scale convolution characteristics of the ultrasonic images by using an improved IncepotionV 2 network, and then performing global average pooling and full connection to obtain multi-scale convolution characteristic vectors;
step 3, constructing a feature adaptive fusion method, fusing the wavelet packet feature vector and the multi-scale convolution feature vector to obtain a fusion feature vector;
and 4, inputting the fusion characteristic vector into a SoftMax classifier, and identifying the type of the weld defects.
2. The weld defect detection method based on ultrasonic waveform and ultrasonic image feature fusion according to claim 1, characterized in that: the step (1) comprises the following steps:
performing 3-layer wavelet packet decomposition on the welding seam ultrasonic waveform to obtain 8 node signals of the 3 rd layer, and recording the signals as S 3,i (i =0,1,.. 7), then the energy of each node is calculated, the calculated energies are normalized, and a feature vector T is constructed:
x i,k for reconstructing the signal S 3,i Discrete point amplitude of (i =0,1, 7), n is wavelet decomposition sequence length, E j,i Representing the energy value of the ith node at decomposition level j.
3. The weld defect detection method based on ultrasonic waveform and ultrasonic image feature fusion according to claim 1, characterized in that: the step (2) comprises the following steps:
the improved Inception V2 network is composed of convolution 3 x 3, maximum pooling 2 x 2, convolution 3 x 3, maximum pooling 2 x 2, improved Inception V2, global average pooling 2 x 2 and full connection, wherein the improved Inception V2 module is formed by adding an average pooling layer in the Inception V2 module, widening each layer network by a 1 x 1 convolution kernel, a 3 x 3 convolution kernel and the maximum pooling layer, reducing the number of channels through 1 x 1 convolution and performing feature clustering, stacking and outputting features in the depth direction, and obtaining feature vectors with the same dimensionality as the wavelet packet features through the global average pool and the full connection layer.
4. The weld defect detection method based on ultrasonic waveform and ultrasonic image feature fusion according to claim 1, characterized in that: in the step (3), the feature adaptive fusion method comprises the following steps:
step 1, inputting the feature vector extracted by the wavelet packet into a shared full-connection layer 1, and outputting ultrasonic waveform feature importance vectors with the same dimensionality when outputting the feature vector;
step 2, inputting the feature vectors extracted by the improved Inception V2 network into the shared full-connection layer 2, and outputting ultrasonic image feature importance vectors with the same dimensionality when outputting the feature vectors;
step 3, normalizing the importance vectors of the two types of characteristics through a Sigmoid function to enable the sum of the importance vectors and the Sigmoid function to be 1, and calculating an ultrasonic image characteristic importance coefficient and an ultrasonic waveform characteristic importance coefficient;
and 4, weighting and adding the ultrasonic waveform characteristic vector and the ultrasonic image characteristic vector to obtain a self-adaptive fusion characteristic vector.
5. The weld defect detection method based on ultrasonic waveform and ultrasonic image feature fusion according to claim 2, characterized in that: the step (3) comprises the following steps:
fusing the feature vectors obtained in the step 1 and the step 2 by using a feature self-adaptive fusion classification module; inputting wavelet packet extracted feature vectors into a shared concatenationThe layer 1 is connected, and when the feature vectors are output, ultrasonic waveform feature importance vectors with the same dimensionality are output; inputting the multi-scale feature vector extracted by the improved Inception V2 network into the shared full-connection layer 2, and outputting the ultrasonic image feature importance vector with the same dimension when outputting the feature vector; normalizing the importance vectors of the two types of characteristics through a Sigmoid function to enable the sum of the importance vectors and the Sigmoid function to be 1, and calculating an ultrasonic image characteristic importance coefficient and an ultrasonic waveform characteristic importance coefficient; and multiplying the ultrasonic image characteristic vector and the ultrasonic shape characteristic vector and adding to obtain a self-adaptive fusion characteristic vector: let the wavelet packet energy eigenvector be
And corresponding importance coefficient alpha i (i =0,1.., 7.) image multi-scale convolution feature vector is denoted as v i (i =0,1, 7) and corresponding importance coefficient is β i (i =0,1, ·, 7); the fusion feature vector obtained in the fusion module is recorded as f i (i =0,1, 7), then f i ＝α i t i +β i v i (i =0,1.., 7), the resulting fused feature vector is F = (F =) (F) 0 ,f 1 ,...,f 7 )，E j,i Representing the energy value of the ith node at decomposition level j.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210433330.4A CN115294376A (en) | 2022-04-24 | 2022-04-24 | Weld defect detection method based on fusion of ultrasonic waveform and ultrasonic image characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210433330.4A CN115294376A (en) | 2022-04-24 | 2022-04-24 | Weld defect detection method based on fusion of ultrasonic waveform and ultrasonic image characteristics |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115294376A true CN115294376A (en) | 2022-11-04 |
Family
ID=83821204
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210433330.4A Pending CN115294376A (en) | 2022-04-24 | 2022-04-24 | Weld defect detection method based on fusion of ultrasonic waveform and ultrasonic image characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115294376A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115980298A (en) * | 2023-03-20 | 2023-04-18 | 山东思睿环境设备科技有限公司 | Multi-parameter-based adaptive water quality detection and analysis method and device |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120226159A1 (en) * | 2009-05-08 | 2012-09-06 | Anthony Sinclair | Ultrasonic scanning system and ultrasound image enhancement method |
| CN107727749A (en) * | 2017-08-30 | 2018-02-23 | 南京航空航天大学 | A kind of ultrasonic quantitative detection method based on wavelet packet fusion feature extraction algorithm |
| CN113129266A (en) * | 2021-03-22 | 2021-07-16 | 太原科技大学 | Stainless steel weld defect detection method based on multi-domain expression data enhancement and model self-optimization |
| US20210334587A1 (en) * | 2018-09-04 | 2021-10-28 | Boe Technology Group Co., Ltd. | Method and apparatus for training a convolutional neural network to detect defects |
| CN113658176A (en) * | 2021-09-07 | 2021-11-16 | 重庆科技学院 | Ceramic tile surface defect detection method based on interactive attention and convolutional neural network |
| CN113887667A (en) * | 2021-10-28 | 2022-01-04 | 中国人民解放军空军工程大学 | Composite material defect detection method and system based on infrared and ultrasonic signal fusion |
| CN113945641A (en) * | 2021-10-19 | 2022-01-18 | 合肥工业大学 | Ultrasonic-based prestressed pipeline grouting quality detection method |
| CN114047259A (en) * | 2021-10-28 | 2022-02-15 | 深圳市比一比网络科技有限公司 | Method for detecting multi-scale steel rail damage defects based on time sequence |
| CN114187174A (en) * | 2021-10-26 | 2022-03-15 | 重庆师范大学 | Image super-resolution reconstruction method based on multi-scale residual error feature fusion |
| CN114283117A (en) * | 2021-11-24 | 2022-04-05 | 广西大学 | Insulator defect detection method based on improved YOLOv3 convolutional neural network |
| CN114354751A (en) * | 2021-12-31 | 2022-04-15 | 北京科技大学 | Metal material internal defect ultrasonic detection system and method based on 5G embedded mode |
-
2022
- 2022-04-24 CN CN202210433330.4A patent/CN115294376A/en active Pending
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120226159A1 (en) * | 2009-05-08 | 2012-09-06 | Anthony Sinclair | Ultrasonic scanning system and ultrasound image enhancement method |
| CN107727749A (en) * | 2017-08-30 | 2018-02-23 | 南京航空航天大学 | A kind of ultrasonic quantitative detection method based on wavelet packet fusion feature extraction algorithm |
| US20210334587A1 (en) * | 2018-09-04 | 2021-10-28 | Boe Technology Group Co., Ltd. | Method and apparatus for training a convolutional neural network to detect defects |
| CN113129266A (en) * | 2021-03-22 | 2021-07-16 | 太原科技大学 | Stainless steel weld defect detection method based on multi-domain expression data enhancement and model self-optimization |
| CN113658176A (en) * | 2021-09-07 | 2021-11-16 | 重庆科技学院 | Ceramic tile surface defect detection method based on interactive attention and convolutional neural network |
| CN113945641A (en) * | 2021-10-19 | 2022-01-18 | 合肥工业大学 | Ultrasonic-based prestressed pipeline grouting quality detection method |
| CN114187174A (en) * | 2021-10-26 | 2022-03-15 | 重庆师范大学 | Image super-resolution reconstruction method based on multi-scale residual error feature fusion |
| CN113887667A (en) * | 2021-10-28 | 2022-01-04 | 中国人民解放军空军工程大学 | Composite material defect detection method and system based on infrared and ultrasonic signal fusion |
| CN114047259A (en) * | 2021-10-28 | 2022-02-15 | 深圳市比一比网络科技有限公司 | Method for detecting multi-scale steel rail damage defects based on time sequence |
| CN114283117A (en) * | 2021-11-24 | 2022-04-05 | 广西大学 | Insulator defect detection method based on improved YOLOv3 convolutional neural network |
| CN114354751A (en) * | 2021-12-31 | 2022-04-15 | 北京科技大学 | Metal material internal defect ultrasonic detection system and method based on 5G embedded mode |
Non-Patent Citations (1)
| Title |
|---|
| JING LIU,GUOCHENG XU,ET.,AL.: "Defect intelligent identification in resistance spot welding ultrasonic detection based on wavelet packet and neural network", 《THE INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115980298A (en) * | 2023-03-20 | 2023-04-18 | 山东思睿环境设备科技有限公司 | Multi-parameter-based adaptive water quality detection and analysis method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112232400B (en) | Stainless steel weld ultrasonic defect detection method based on depth feature fusion | |
| CN109783906B (en) | Intelligent analysis system and method for detecting magnetic flux leakage data in pipeline | |
| CN110879254B (en) | Steel rail crack acoustic emission signal detection method based on improved least square generation type countermeasure network | |
| CN110470729B (en) | Eddy current-based nondestructive testing method for oil field casing pipe defects | |
| CN109325544B (en) | Pipeline abnormity classification method based on middle layer characteristics | |
| CN115239632B (en) | Rail surface damage detection method integrating inspection image and ultrasonic image | |
| CN110243934A (en) | A kind of ultrasonic weld seam detection method based on wavelet convolution neural network | |
| CN115294376A (en) | Weld defect detection method based on fusion of ultrasonic waveform and ultrasonic image characteristics | |
| Xu et al. | Rail defect detection method based on recurrent neural network | |
| CN115791969A (en) | Jacket underwater crack detection system and method based on acoustic emission signals | |
| CN113506239A (en) | Strip steel surface defect detection method based on cross-stage local network | |
| CN115656335A (en) | Method for quickly detecting and identifying internal defects of bearing ring | |
| WO2022088226A1 (en) | Heterogeneous field signal-based intelligent inversion method for pipeline defects | |
| CN116773654A (en) | Intelligent pipeline defect detection device and detection method based on circumferential guided waves | |
| Cui et al. | A recognition algorithm to detect pipe weld defects | |
| CN115096996A (en) | Rail transit train welding quality detection method based on improved Mask R-CNN | |
| Han et al. | Damage detection of quayside crane structure based on improved faster R-CNN | |
| CN114018951A (en) | Plastic part weld joint detection method based on DCGAN (DCGAN-based gas diffusion analysis) significant feature sample expansion pretreatment | |
| CN114418929A (en) | Weld defect identification method based on consistency multi-scale metric learning | |
| Mohamed et al. | An adaptive neuro-fuzzy inference system-based approach for oil and gas pipeline defect depth estimation | |
| Zhang et al. | Diagnosis and recognition of pipeline damage defects based on improved convolutional neural network | |
| CN118228772B (en) | Distance measurement method, framework, system and medium for actually measured traveling wave of power transmission line | |
| Shi et al. | CNN based defect recognition model for phased array ultrasonic testing images of electrofusion joints | |
| CN116645371B (en) | Rail surface defect detection method and system based on feature search | |
| CN116151625B (en) | Marine ship risk assessment method based on big data analysis |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20221104 |