CN115310486A - Intelligent detection method for welding quality - Google Patents
Intelligent detection method for welding quality Download PDFInfo
- Publication number
- CN115310486A CN115310486A CN202210951200.XA CN202210951200A CN115310486A CN 115310486 A CN115310486 A CN 115310486A CN 202210951200 A CN202210951200 A CN 202210951200A CN 115310486 A CN115310486 A CN 115310486A
- Authority
- CN
- China
- Prior art keywords
- welding
- vibration
- welding quality
- module
- product
- 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.)
- Granted
Links
Images
Landscapes
- Investigating Or Analyzing Materials By The Use Of Ultrasonic Waves (AREA)
Abstract
The invention discloses an intelligent detection method for welding quality, which comprises the following steps: the method comprises the following steps: arranging a vibration excitation point and a plurality of vibration sensors on a welding product; step two: enabling a welded product to vibrate, collecting vibration signals passing through a welding position, and forming multi-channel vibration signals by the vibration signals collected by a plurality of vibration sensors; obtaining a classification label of the welding quality of a current welding product, and forming welding quality data by the multi-channel vibration signal and the classification label; step three: judging whether the quantity of the acquired welding quality data reaches a set threshold value or not; if so, obtaining a data set, and executing a fifth step; if not, executing the fourth step; step four: arranging an excitation point and a vibration sensor at the same position of another welding product, and executing the second step; step five: constructing an artificial intelligence algorithm model, and training the artificial intelligence algorithm model by using a data set; step six: and carrying out on-line detection on the welding quality of the welding product by using an artificial intelligence algorithm model.
Description
Technical Field
The invention belongs to the technical field of welding, and particularly relates to an intelligent detection method for welding quality.
Background
The welding process is used as a main connecting process method and is widely applied to important goods such as high-speed trains, automobiles and the like. For example, there are more than 5000 welds on a car. Therefore, the quality of welding is directly related to the quality of the commodity and even the safety problem. At present, the detection means aiming at the welding quality can be mainly divided into manual judgment and equipment detection. The manual detection is mainly judged by experience of a welder, and the welding quality is judged by observing the welding quality and the flatness of a welding area; the welding quality detection method based on manual experience is often inaccurate in judgment result due to the fact that standards are not uniform. The equipment detection mainly detects the welding part through equipment such as an ultrasonic flaw detector and the like; the welding quality detection method based on the equipment is often low in efficiency and high in equipment cost.
Chinese patent publication No. CN113909667B discloses a welding quality evaluation method of an ultrasonic welding machine based on vibration data, in which a first vibration acceleration sensor is disposed under a welding seat and a second vibration acceleration sensor is disposed on a side surface of the welding seat, in a welding process, the first vibration acceleration sensor and the second vibration acceleration sensor are used to collect vibration data in the welding process, a confidence interval of a vibration index is obtained by using vibration data of normal welding, and the vibration index of the current welding vibration data is compared with the confidence interval, thereby evaluating the welding quality.
Although the welding quality evaluation method of the ultrasonic welding machine based on the vibration data can meet the evaluation requirement on the welding quality to a certain extent, the method can only evaluate standardized welding operation, and the vibration data in the welding process is acquired. When unfavorable working conditions such as improper input pressure, overlapping of adapter plates, interference of a welding head and the like occur in the welding process, the adverse working conditions can be reflected in vibration data, so that the welding quality is evaluated, and the measurement evaluation is not directly carried out on a finally formed welding structure. Since there is no necessary correlation between the occurrence of the unfavorable condition and the final weld quality, the evaluation result of the method has a large error.
Disclosure of Invention
In view of this, the present invention provides an intelligent detection method for welding quality, which directly detects a finally formed welding product and can effectively improve detection accuracy and reliability.
In order to achieve the purpose, the invention provides the following technical scheme:
an intelligent detection method for welding quality comprises the following steps:
the method comprises the following steps: arranging a vibration excitation point and a plurality of vibration sensors on a welding product;
step two: the method comprises the following steps that a vibration exciter is used for outputting exciting force with a set waveform to an exciting point so as to enable a welded product to vibrate, vibration signals passing through a welding part of the welded product are collected through a vibration sensor, and the vibration signals collected by a plurality of vibration sensors form a multi-channel vibration signal;
obtaining a classification label of the welding quality of the current welding product, and forming the welding quality data of the current welding product by the multi-channel vibration signal and the classification label;
step three: judging whether the quantity of the acquired welding quality data reaches a set threshold value or not; if yes, obtaining a data set consisting of a set number of welding quality data, and executing a fifth step; if not, executing the fourth step;
step four: arranging an excitation point and a vibration sensor at the same position of another welding product of the same type, and executing the second step;
step five: constructing an artificial intelligence algorithm model, and training the artificial intelligence algorithm model by using a data set;
step six: inputting a multi-channel vibration signal into the trained artificial intelligence algorithm model, outputting a classification label corresponding to the welding quality of the welding product, and carrying out online detection on the welding quality of the welding product.
Further, the waveform of the exciting force is a sine wave, a square wave, a triangular wave or a rectangular wave.
Further, the data set is divided into a training set, a verification set and a test set which are respectively used for training, verifying and testing the artificial intelligent algorithm model.
Further, the artificial intelligence algorithm model comprises a deep learning model, a machine learning model or a reinforcement learning model.
Further, the artificial intelligence algorithm model adopts a multi-channel-scale-receptive field fusion feature extraction model.
Further, the multi-channel-scale-receptive field fusion feature extraction model comprises an initialization module, a feature fusion layer and a classifier, and a parallel residual error calibration module and a multi-receptive field module are arranged between the initialization module and the feature fusion layer;
the initialization model is used for performing primary data feature extraction on the multi-channel vibration signal and inputting the extracted primary data features into the residual error calibration module and the multi-sensing-field module;
the residual error calibration module realizes further mining of the data characteristics of the multi-channel vibration signal through fine-grained characteristic extraction and characteristic attention calibration on multiple scales;
the multi-receptive field module acquires effective information contained in the signal through different receptive field sizes on a multi-channel;
the characteristic fusion layer is used for fusing the characteristics extracted by the residual error calibration module and the multi-sensing field module and inputting the fused characteristics into the classifier;
and the sorter classifies the fused features to obtain a classification label of the welding quality of the current welding product.
Further, the feature fusion layer fuses the features extracted by the residual error calibration module and the multi-sensing-field module by adopting a linear fusion method:
F＝αF R +βF M
wherein F represents the fused feature; f R Representing features extracted by a residual error calibration module; f M Representing the features extracted by the multi-receptive-field module; α and β represent learnable parameters, respectively.
The invention has the beneficial effects that:
according to the intelligent detection method for the welding quality, the vibration excitation point and the plurality of vibration sensors are arranged on the welding product, the vibration exciter is used for outputting the excitation force with the set waveform to the vibration excitation point, the welding product vibrates, the vibration signal changes through a welding part in the process of propagating in the welding product, the vibration signal at the moment carries the relevant information of the quality condition of the welding part, and then the signal is collected by the vibration sensors; the vibration sensors at different positions can capture different vibration signals, and the multi-channel vibration signals formed by the vibration signals carry richer welding quality information; the artificial intelligence algorithm model obtains welding quality information from the multi-channel vibration signals, removes invalid information from the multi-channel vibration signals, judges the welding quality condition of the current welding part through a classifier, and outputs classification labels of the welding quality, if the welding quality is qualified or unqualified, thereby realizing the technical purpose of online detection of the welding quality of welding products;
specifically, in each measurement, the arrangement positions of the excitation points on the welded product and the vibration sensor are the same, and the excitation force excited by the vibration exciter in each measurement is the same, so that the consistency of each measurement can be ensured, and the detection precision and reliability can be effectively improved.
The invention also has the following advantages:
(1) The invention can realize the rapid quality detection of the welding product before the welding product is put into use, and does not need to design a complex welding quality detection device;
(2) The invention can be suitable for any welding product, including complex curved surface welding parts and assembly welding parts, namely the invention has the advantages of wide application range and high transportability, and solves the problem that the welding quality of the complex welding parts is difficult to judge by manual experience;
(3) The invention is suitable for the online measurement of welding products of different welding modes, including but not limited to spot welding, arc welding and the like;
(4) The method adopts an artificial intelligence algorithm model, and the artificial intelligence algorithm model can automatically extract the characteristics in the multi-channel vibration signal and carry out data characteristic mining; in addition, the artificial intelligence algorithm model also has the advantages of wide coverage range, good adaptability and data driving: the artificial intelligence algorithm can be mapped to any function; because the artificial intelligence algorithm model is data-driven, as long as the data volume is enough, the very good performance effect can be achieved, and in addition, the artificial intelligence algorithm model based on data-driven avoids the design work of characteristics or kernel functions, reduces the dependence degree on expert experience, and greatly improves the efficiency.
Drawings
In order to make the object, technical scheme and beneficial effect of the invention more clear, the invention provides the following drawings for explanation:
FIG. 1 is a flow chart of the intelligent detection method for welding quality of the present invention;
FIG. 2 is a pictorial view of a welded automotive body-in-white;
FIG. 3 is a flow chart of data acquisition;
FIG. 4 is a three-dimensional schematic view of a rear door of a vehicle body by welding;
FIG. 5 is a schematic illustration of the location and number of vibration excitation points and vibration sensors disposed on a rear door of a vehicle body;
FIG. 6 is a schematic structural diagram of a multi-channel-scale-receptive field fusion feature extraction model;
fig. 7 is a schematic structural diagram of a residual error calibration module.
Detailed Description
The present invention is further described with reference to the following drawings and specific examples so that those skilled in the art can better understand the present invention and can practice the present invention, but the examples are not intended to limit the present invention.
As shown in fig. 1, the intelligent detection method for the welding quality of the embodiment includes the following steps:
the method comprises the following steps: an excitation point and a plurality of vibration sensors are arranged on a welding product. Specifically, for the same welding product, the positions of the excitation points measured each time are the same, and the arrangement positions and the number of the vibration sensors are also the same.
Step two: the vibration exciter is used for outputting exciting force with set waveform to the exciting points to enable the welded product to vibrate, the vibration sensors are used for collecting vibration signals passing through the welding positions of the welded product, and the vibration signals collected by the vibration sensors form a multi-channel vibration signal. And acquiring a classification label of the welding quality of the current welding product, and forming the welding quality data of the current welding product by using the multi-channel vibration signal and the classification label.
Specifically, for the same welding product, the excitation force waveforms generated by the vibration exciters each time must be the same, so that the welding quality information carried in the signals received by the vibration sensor is more obvious. The excitation force signal may be selected to include, but is not limited to, a sine wave, a square wave, a triangular wave, a rectangular wave, and the like. Of course, different excitation force waveforms may be selected for different types of welding products. For example, as shown in fig. 2, for an automobile welding body-in-white, an excitation point is selected at a certain part of the body-in-white, a plurality of excitation sensors are arranged at other positions of the body-in-white, and the welding quality of the whole body-in-white can be obtained by analyzing the characteristics in multi-channel signal data formed by the excitation sensors.
Specifically, as shown in fig. 3, in this embodiment, the acquisition process of the vibration signal is as follows:
(1) Selecting a certain welding product, selecting corresponding positions of excitation points, the number and the positions of vibration sensors on the welding product, and setting the magnitude and the waveform of an excitation force;
(2) Generating an excitation force signal;
(3) The signal collector obtains an excitation force signal;
(4) Converting the exciting force signal from a digital signal to an analog signal through a D/A converter, transmitting the analog signal to a vibration exciter after power amplification, and generating a specific waveform exciting force by using the vibration exciter;
(5) The vibration sensor collects vibration signals passing through a welding part;
(6) And converting the analog signals acquired by the vibration sensor into digital signals through A/D conversion to obtain vibration signals.
Step three: judging whether the quantity of the acquired welding quality data reaches a set threshold value or not; if yes, obtaining a data set consisting of a set number of welding quality data, and executing a fifth step; if not, executing the step four. In this way, enough welding quality data can be contained in the data set to improve the performance effect of the artificial intelligence algorithm model.
Step four: arranging an excitation point and a vibration sensor at the same position of another welding product of the same type, and executing a second step;
step five: and constructing an artificial intelligence algorithm model, and training the artificial intelligence algorithm model by using the data set. In this embodiment, the data set is divided into a training set, a verification set, and a test set, which are respectively used for training, verifying, and testing the artificial intelligence algorithm model.
Specifically, the artificial intelligence algorithm model may adopt a deep learning model, a machine learning model, or a reinforcement learning model. The artificial intelligence algorithm model of the embodiment adopts a multi-channel scale receptive field fusion feature extraction model (MMDFF Module). Specifically, the multi-channel-scale-receptive field fusion feature extraction model of the embodiment includes an initialization Module (initialization Module), a feature fusion layer (fusion layer), and a classifier (classifier), and a parallel residual error calibration Module (ResCASB Module) and a multi-receptive field Module (MD Module) are disposed between the initialization Module and the feature fusion layer.
Specifically, the initialization model is used for performing preliminary data feature extraction on the multi-channel vibration signal, and inputting the extracted preliminary data features into the residual error calibration module and the multi-sensing-field module. The initialization module of this embodiment consists of two layers of one-dimensional convolution.
And the residual error calibration module realizes further mining of the data characteristics of the multi-channel vibration signal through fine-grained characteristic extraction and characteristic attention calibration on multiple scales. In the residual error calibration module, the multi-scale residual error and attention calibration are mainly composed.
The multi-receptive field module acquires effective information stored in the signal through different receptive field sizes on multiple channels.
The characteristic fusion layer is used for fusing the characteristics extracted by the residual error calibration module and the multi-sensing-field module and inputting the fused characteristics into the classifier. The characteristic fusion layer adopts a linear fusion method to fuse the characteristics extracted by the residual error calibration module and the multi-sensing-field module:
F＝αF R +βF M
wherein F represents the fused feature; f R Representing the features extracted by the residual error calibration module; f M Representing the features extracted by the multi-receptive-field module; α and β represent learnable parameters, respectively.
And classifying the fused features by the classifier to obtain a classification label of the welding quality of the current welding product.
Step six: inputting a multi-channel vibration signal into the trained artificial intelligence algorithm model, outputting a classification label corresponding to the welding quality of the welding product, and carrying out online detection on the welding quality of the welding product.
The following describes a specific embodiment of the welding quality intelligent detection method according to the present embodiment, taking a body-in-white rear door assembly as an example.
As shown in fig. 4, is a three-dimensional schematic view of a rear door of a vehicle body obtained by welding. Fig. 5 is a schematic diagram of the arrangement positions of the excitation points and the vibration sensors. As shown in fig. 5, a certain point on the welded member was selected as the excitation application point, and 8 vibration sensors were placed at the edge end and the end of the door to be welded.
The computer generates an excitation force signal, and the excitation signal selects a sine wave with the frequency of 100Hz; the excitation force signal is acquired by a signal acquisition device; converting the exciting force signal from a digital signal to an analog signal through a D/A converter, amplifying the power of the analog signal, and transmitting the amplified analog signal to a vibration exciter to generate a sine-wave exciting force; the vibration exciter works for 10s continuously, and the acquisition time of all the vibration sensors is also set to be 10s. Ensuring that the vibration sensor can receive all vibration signals. And carrying out A/D conversion on the analog signals acquired by the vibration sensor, and converting the analog signals into digital signals to obtain vibration signals. And finally, forming the processed vibration signals into an eight-channel vibration signal.
In this embodiment, 10000 weld data are collected, and a training data set is set, and the ratio of the verification data set to the test data set is 6. And carrying out artificial intelligence algorithm model training by utilizing the training data set and the verification data set, and judging the performance of the model by utilizing the test data set. As shown in fig. 6, in the present embodiment, a multi-channel-scale-receptive field fusion feature extraction model (MMDFF Module) is selected to extract eight-channel vibration signal data features. The MMDFF Module model structure is shown in FIG. 6, and includes an initialization Module, a parallel residual error calibration Module (ResCASB Module) and a multi-sense field Module (MD Module), a feature fusion layer and a classifier.
Specifically, the initialization module performs preliminary data feature extraction on the eight-channel vibration signals, the initialization module is composed of two layers of one-dimensional convolution, and the sizes of convolution kernels of the initialization module are respectively 7 and 5. And then inputting the preliminarily extracted eight-channel vibration signal data characteristic input into the residual error calibration module and the multi-sensing-field module in parallel.
As shown in fig. 7, the residual error calibration module mainly performs fine-grained feature extraction and feature attention calibration on multiple scales, so as to further mine the eight-channel vibration signal data features. In the residual error calibration module, the multi-scale residual error and attention calibration are mainly composed. For multi-scale residuals, when an eight-channel vibration signal data feature comes in, it will first go through a layer of convolution layers with K =1, and then divide the feature map into 4 parts. The first part of the circuit is simple, x 1 Go directly to y without processing 1 (ii) a Second part of the line, x 2 After K =3 convolution, the signal is divided into two lines, and one line is continuously propagated forwards to y 2 The other to x 3 So that the third line obtains the information of the second line; the third line, the fourth line and so on. After the data features are subjected to multi-scale residual, global sum and cost can be realizedAnd extracting the ground information. For the feature attention module, global Average Pooling (GAP) and Global Maximum Pooling (GMP) are adopted to compress the space dimension of the input features, the most significant part is coded by the Global Maximum Pooling (GMP), another important clue of channel information is represented, the two are fused to deduce a more refined channel attention weight, and finally some effective information of the data features is reserved. The feature mentioned in this block is marked F R 。
The multi-receptive field module is mainly used for acquiring effective information stored in the signal through different receptive field sizes on multiple channels. When the eight-channel vibration signal data features enter the multi-field module, the eight-channel vibration signal data features firstly pass through a convolution layer with K =1, and then the features are input into three channels, the convolution kernel of each channel is 3, and the sizes of the holes are 1,2,3 respectively. In each channel, after convolution feature extraction of eight-channel vibration signals in different receptive fields, feature fusion is carried out in a Flatten layer, so that effective information in different receptive fields is effectively reserved. The feature mentioned in this block is marked F M 。
And fusing the features extracted by the residual error calibration module and the multi-sensing-field module in the feature fusion layer. The characteristic fusion layer adopts a linear fusion mode, and the specific formula is as follows:
F＝αF R +βF M
wherein F represents the fused feature; f R Representing the features extracted by the residual error calibration module; f M Representing the features extracted by the multi-receptive-field module; α and β represent learnable parameters, respectively.
And inputting the feature representation after feature fusion into a classifier for classification, and finally obtaining a classification label of the welding quality of the current welding part.
And carrying out online welding part quality detection by using the trained artificial intelligence algorithm model.
The above-mentioned embodiments are merely preferred embodiments for fully illustrating the present invention, and the scope of the present invention is not limited thereto. The equivalent substitution or change made by the technical personnel in the technical field on the basis of the invention is all within the protection scope of the invention. The protection scope of the invention is subject to the claims.
Claims (7)
1. An intelligent detection method for welding quality is characterized in that: the method comprises the following steps:
the method comprises the following steps: arranging a vibration excitation point and a plurality of vibration sensors on a welding product;
step two: outputting exciting force with set waveform to an exciting point by using a vibration exciter so as to enable a welded product to vibrate, collecting vibration signals passing through a welding part of the welded product by using a vibration sensor, and forming multi-channel vibration signals by using the vibration signals collected by a plurality of vibration sensors;
obtaining a classification label of the welding quality of a current welding product, and forming the welding quality data of the current welding product by using the multi-channel vibration signal and the classification label;
step three: judging whether the quantity of the acquired welding quality data reaches a set threshold value or not; if yes, obtaining a data set consisting of a set number of welding quality data, and executing a fifth step; if not, executing the fourth step;
step four: arranging an excitation point and a vibration sensor at the same position of another welding product of the same type, and executing a second step;
step five: constructing an artificial intelligence algorithm model, and training the artificial intelligence algorithm model by using a data set;
step six: and inputting a multi-channel vibration signal into the trained artificial intelligence algorithm model, outputting a classification label corresponding to the welding quality of the welding product, and carrying out online detection on the welding quality of the welding product.
2. The intelligent detection method for the welding quality according to claim 1, characterized in that: the waveform of the exciting force is a sine wave, a square wave, a triangular wave or a rectangular wave.
3. The intelligent detection method for the welding quality according to claim 1, characterized in that: the data set is divided into a training set, a verification set and a test set which are respectively used for training, verifying and testing the artificial intelligent algorithm model.
4. The intelligent detection method for the welding quality according to claim 1, characterized in that: the artificial intelligence algorithm model comprises a deep learning model, a machine learning model or a reinforcement learning model.
5. The intelligent detection method for the welding quality according to claim 1, characterized in that: the artificial intelligence algorithm model adopts a multi-channel-scale-receptive field fusion characteristic extraction model.
6. The intelligent detection method for the welding quality according to claim 5, characterized in that: the multi-channel-scale-receptive field fusion feature extraction model comprises an initialization module, a feature fusion layer and a classifier, wherein a parallel residual error calibration module and a multi-receptive field module are arranged between the initialization module and the feature fusion layer;
the initialization model is used for performing primary data feature extraction on the multi-channel vibration signal and inputting the extracted primary data features into the residual error calibration module and the multi-sensing-field module;
the residual error calibration module realizes further mining of the data characteristics of the multi-channel vibration signal through fine-grained characteristic extraction and characteristic attention calibration on multiple scales;
the multi-reception-field module acquires effective information stored in the signal through different reception field sizes on multiple channels;
the characteristic fusion layer is used for fusing the characteristics extracted by the residual error calibration module and the multi-sensing-field module and inputting the fused characteristics into the classifier;
and the differentiator classifies the fused features to obtain a classification label of the welding quality of the current welding product.
7. The intelligent welding quality detection method according to claim 6, characterized in that: the characteristic fusion layer fuses the characteristics extracted by the residual error calibration module and the multi-sensing-field module by adopting a linear fusion method:
F＝αF R +βF M
wherein F represents the fused feature; f R Representing features extracted by a residual error calibration module; f M Representing the features extracted by the multi-receptive-field module; α and β represent learnable parameters, respectively.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210951200.XA CN115310486B (en) | 2022-08-09 | 2022-08-09 | Intelligent welding quality detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210951200.XA CN115310486B (en) | 2022-08-09 | 2022-08-09 | Intelligent welding quality detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115310486A true CN115310486A (en) | 2022-11-08 |
| CN115310486B CN115310486B (en) | 2023-09-26 |
Family
ID=83861512
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210951200.XA Active CN115310486B (en) | 2022-08-09 | 2022-08-09 | Intelligent welding quality detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115310486B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110261483A (en) * | 2019-07-31 | 2019-09-20 | 三一汽车起重机械有限公司 | Weld defect detection method, device, detection device and readable storage medium storing program for executing |
| CN111815572A (en) * | 2020-06-17 | 2020-10-23 | 深圳市大德激光技术有限公司 | Method for detecting welding quality of lithium battery based on convolutional neural network |
| CN112184686A (en) * | 2020-10-10 | 2021-01-05 | 深圳大学 | Segmentation algorithm for detecting laser welding defects of safety valve of power battery |
| CN112465790A (en) * | 2020-12-03 | 2021-03-09 | 天津大学 | Surface defect detection method based on multi-scale convolution and trilinear global attention |
| CN113255778A (en) * | 2021-05-28 | 2021-08-13 | 广汽本田汽车有限公司 | Welding spot quality detection method and device based on multi-model fusion and storage medium |
| CN113378725A (en) * | 2021-06-15 | 2021-09-10 | 山东大学 | Cutter fault diagnosis method, equipment and storage medium based on multi-scale-channel attention network |
| US20210390338A1 (en) * | 2020-06-15 | 2021-12-16 | Dalian University Of Technology | Deep network lung texture recogniton method combined with multi-scale attention |
| CN114119637A (en) * | 2021-11-29 | 2022-03-01 | 大连大学 | Brain white matter high signal segmentation method based on multi-scale fusion and split attention |
-
2022
- 2022-08-09 CN CN202210951200.XA patent/CN115310486B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110261483A (en) * | 2019-07-31 | 2019-09-20 | 三一汽车起重机械有限公司 | Weld defect detection method, device, detection device and readable storage medium storing program for executing |
| US20210390338A1 (en) * | 2020-06-15 | 2021-12-16 | Dalian University Of Technology | Deep network lung texture recogniton method combined with multi-scale attention |
| CN111815572A (en) * | 2020-06-17 | 2020-10-23 | 深圳市大德激光技术有限公司 | Method for detecting welding quality of lithium battery based on convolutional neural network |
| CN112184686A (en) * | 2020-10-10 | 2021-01-05 | 深圳大学 | Segmentation algorithm for detecting laser welding defects of safety valve of power battery |
| CN112465790A (en) * | 2020-12-03 | 2021-03-09 | 天津大学 | Surface defect detection method based on multi-scale convolution and trilinear global attention |
| CN113255778A (en) * | 2021-05-28 | 2021-08-13 | 广汽本田汽车有限公司 | Welding spot quality detection method and device based on multi-model fusion and storage medium |
| CN113378725A (en) * | 2021-06-15 | 2021-09-10 | 山东大学 | Cutter fault diagnosis method, equipment and storage medium based on multi-scale-channel attention network |
| CN114119637A (en) * | 2021-11-29 | 2022-03-01 | 大连大学 | Brain white matter high signal segmentation method based on multi-scale fusion and split attention |
Non-Patent Citations (1)
| Title |
|---|
| 王辉涛 等: "基于全局时空感受野的高效视频分类方法", 小型微型计算机***, no. 08, pages 1768 - 1775 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115310486B (en) | 2023-09-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110389170B (en) | Train component crack damage detection method and system based on Lamb wave imaging | |
| CN106447040B (en) | Mechanical equipment health state evaluation method based on Heterogeneous Multi-Sensor Data fusion | |
| CN104730109B (en) | A kind of resistance spot welding quality of welding spot detection method based on dynamic resistance curve | |
| CN111896625A (en) | Real-time monitoring method and monitoring system for rail damage | |
| CN102162577B (en) | Pipeline defect surface integrity detection device and detection method | |
| CN103760231B (en) | Weld defect giant magnetoresistance eddy current detection method based on decision tree and detection device | |
| CN106198725B (en) | A kind of butt weld defect detecting system and detection method based on feature guided wave | |
| CN105424364A (en) | Diagnostic method and device of train bearing failure | |
| CN104677751B (en) | Quality detection method for resistance-spot-welding spots on basis of calculation of thermal effect of welding process | |
| CN111754463B (en) | Method for detecting CA mortar layer defects of ballastless track based on convolutional neural network | |
| CN105913059A (en) | Vehicle VIN code automatic identifying system and control method therefor | |
| CN104777222A (en) | Pipeline defect identification and visualization method based on three-dimensional phase trajectory of Duffing system | |
| KR102007494B1 (en) | System for inspecting welding quality of weld zone using ultrasonic | |
| CN113705412A (en) | Deep learning-based multi-source data fusion track state detection method and device | |
| CN104792865A (en) | Recognizing and positioning method of small defects of pipelines through ultrasonic guided waves based on fractal dimensions | |
| CN103760230A (en) | BP neural network-based giant magnetoresistance eddy current testing method for welding defect | |
| CN102519577B (en) | Method and system for identifying road surface in a road | |
| CN103713042A (en) | Eddy-current welding defect detection method based on k-nearest neighbor algorithm | |
| CN103760229A (en) | Welding defect giant magnetoresistance vortexing detection method based on supporting vector machine | |
| CN102419242A (en) | Apparatus and method for testing air bag control unit of vehicle | |
| CN115310486A (en) | Intelligent detection method for welding quality | |
| CN103954628B (en) | Ensemble empirical mode decomposition (EEMD) and approximate entropy combined steel tube damage monitoring method | |
| CN105572230A (en) | Polarity weighting vector fully focusing imaging method for identifying crack type defects quantitatively | |
| CN113256566A (en) | Pipeline weld defect identification method | |
| CN206662523U (en) | A kind of spot welding on-line detecting system |
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 |