A laser-based weld quality inspection system and method

Abstract

This invention discloses a weld quality inspection system and method based on laser technology. The system includes: a laser acquisition module for emitting a laser beam towards the target weld to be inspected and acquiring the beam signal reflected by the target weld; an optical path transmission module for receiving and transmitting the beam signal; wherein the optical path transmission module includes an optical path adjustment unit and a photoelectric receiving unit, the optical path adjustment unit for adjusting the propagation path of the beam signal to propagate to the photoelectric receiving unit, and the photoelectric receiving unit for converting the beam signal into an electrical signal; and a signal processing module for receiving the electrical signal and performing weld quality inspection based on the electrical signal. This invention achieves weld inspection through laser technology, which can improve the effect of weld quality inspection.

Description

Technical Field

[0001] This invention relates to the field of laser inspection technology, and in particular to a weld quality inspection system and method based on laser technology. Background Technology

[0002] Laser processing technology, as one of the key technologies in modern manufacturing, is widely used in the precision machining of metal and non-metal workpieces. Its core principle is to use a laser beam emitted by a laser, which is transmitted through a laser head and focused on the surface of the workpiece. Processing is achieved through the interaction between the laser and the non-transparent material. At the microscopic level, this manifests as a quantum energy exchange process, while at the macroscopic level, it is specifically manifested as the reflection, absorption, heating, melting, and even vaporization of the material. Ultimately, it completes processes such as small hole forming, kerfing, surface cladding, or workpiece joining.

[0003] In laser processing welding applications, the focused energy of the laser beam melts the workpiece material to form a molten pool. This process is accompanied by the dynamic release of multiple characteristic signals, including metal vapor, radiation light signals, and plasma. The variation patterns of these characteristic signals are directly related to the weld quality, and regardless of how the laser beam focus position is adjusted, the position of the workpiece measurement surface can adaptively match the welding process, providing a feasible basis for real-time monitoring of weld quality.

[0004] However, in traditional laser welding, weld quality monitoring mainly relies on manual operation. Operators must continuously observe the melting position and molten pool morphology of the welding area and adjust welding parameters accordingly. This method faces significant limitations in practical applications: firstly, the welding process involves high temperatures, intense light, and harmful gases, posing a direct threat to the operator's health; secondly, manual monitoring is susceptible to subjective experience and fatigue, making continuous and precise monitoring of the welding process difficult. Inadequate monitoring can lead to fluctuations in weld quality, affecting production efficiency and product yield, thus failing to meet the demands of modern manufacturing for automated and high-precision processing. Summary of the Invention

[0005] This invention provides a weld quality inspection system and method based on laser technology, aiming to improve the weld quality inspection effect.

[0006] This invention provides a weld quality inspection system based on laser technology, comprising:

[0007] The laser acquisition module is used to emit a laser beam toward the target weld to be inspected and to acquire the beam signal reflected by the target weld.

[0008] An optical path transmission module is used to receive and transmit the beam signal; wherein, the optical path transmission module includes an optical path adjustment unit and a photoelectric receiving unit, the optical path adjustment unit is used to adjust the propagation path of the beam signal to propagate to the photoelectric receiving unit, and the photoelectric receiving unit is used to convert the beam signal into an electrical signal;

[0009] The signal processing module is used to receive the electrical signal and perform weld quality inspection based on the electrical signal.

[0010] Furthermore, the optical path adjustment unit is provided with:

[0011] A reflector for receiving and reflecting the light beam signal;

[0012] A focusing element for focusing and propagating the beam signal;

[0013] A beam splitter is used to reflect part of the beam signal and refract another part of the beam signal.

[0014] A filter is used to allow light beam signals of a specified wavelength to pass through and to cut off light beam signals of another specified wavelength.

[0015] Furthermore, multiple photoelectric receiving units are provided, and each of the multiple photoelectric receiving units receives beam signals of different wavelengths.

[0016] Furthermore, the optical path transmission module is provided with the following components sequentially along the optical path propagation direction:

[0017] Reflector;

[0018] A focusing element, disposed in the reflection direction of the reflector, is used to focus the light beam signal;

[0019] The first beam splitter is used to reflect and refract the beam signal focused by the focusing element for the first time to generate a first reflected light path and a first refracted light path.

[0020] The first filter element is disposed on the first reflected light path and is used to filter the light beam signal of the first reflection;

[0021] The first photoelectric receiving unit is disposed on the first reflected light path and is used to receive the light beam signal transmitted through the first filter.

[0022] The second beam splitter is disposed on the first refracted light path and is used to reflect and refract the light beam signal refracted in the first refracted light a second time to generate a second reflected light path and a second refracted light path.

[0023] The second filter element is disposed on the second reflected light path and is used to filter the light beam signal of the second reflection.

[0024] The second photoelectric receiving unit is disposed on the second reflected light path and is used to receive the light beam signal transmitted through the second filter.

[0025] The third filter element is disposed on the second refracted light path and is used to filter the light beam signal refracted in the second refracted light.

[0026] The third photoelectric receiving unit is disposed on the second refracted light path and is used to receive the light beam signal transmitted through the third filter.

[0027] Furthermore, a first aperture, a first calibration element, and a first light-diffusing element are disposed between the first filter element and the first photoelectric receiving unit;

[0028] A second aperture, a second calibration element, and a second light-diffusing element are disposed between the second filter element and the second photoelectric receiving unit;

[0029] A third aperture, a third calibration element, and a third light-diffusing element are provided between the third filter element and the third photoelectric receiving unit.

[0030] This invention also provides a laser-based weld quality inspection method, applied to the laser-based weld quality inspection system described in any of the preceding embodiments, the method comprising:

[0031] A laser beam is emitted toward the target weld to be inspected using a laser acquisition module, and the beam signal reflected by the target weld is acquired.

[0032] The beam signal is received and transmitted through the optical path transmission module, and the beam signal is converted into an electrical signal through the photoelectric receiving unit in the optical path transmission module.

[0033] The electrical signal is received by the signal processing module, and the weld quality is inspected based on the electrical signal.

[0034] Furthermore, the step of receiving the electrical signal through the signal processing module and performing weld quality inspection based on the electrical signal includes:

[0035] The wavelet threshold denoising algorithm is used to denoise the received electrical signal, and the denoised electrical signal is then normalized.

[0036] Weld quality feature parameters are extracted from the normalized electrical signal; wherein, the weld quality feature parameters include grayscale features and geometric features;

[0037] The weld quality characteristic parameters are classified and predicted using a classification model to obtain the corresponding classification prediction results, and the classification prediction results are output as the weld quality inspection results.

[0038] Furthermore, after the step of receiving the electrical signal through the signal processing module and performing weld quality inspection based on the electrical signal, the method further includes:

[0039] When the weld quality inspection result indicates the presence of weld defects, the coordinate position of the weld defects in a preset three-dimensional coordinate system is calculated using geometric optics, based on the beam propagation path. The three-dimensional coordinate system has the origin at the starting point of the laser beam emitted by the laser acquisition module, the positive Z-axis is the direction of laser beam emission, the horizontal axis perpendicular to the direction of laser beam emission is the X-axis, and the vertical axis perpendicular to the direction of laser beam emission is the Y-axis.

[0040] The coordinate positions are correlated with the target weld, and a defect location distribution map is generated.

[0041] Furthermore, after the step of receiving the electrical signal through the signal processing module and performing weld quality inspection based on the electrical signal, the method further includes:

[0042] When the weld quality inspection result indicates the presence of weld defects, a Fourier transform is performed on the electrical signal to convert the time-domain signal corresponding to the electrical signal into a frequency-domain signal.

[0043] A blind source separation algorithm is used to separate the signal components corresponding to the weld defect from the frequency domain signal;

[0044] A separation matrix is ​​constructed based on the separated signal components, and the separation matrix is ​​solved using the negative entropy maximization criterion to obtain the defect signal;

[0045] The defect signal is subjected to inverse Fourier transform to obtain the corresponding time-domain defect signal, and the size information of the weld defect is obtained based on the time-domain defect signal.

[0046] Furthermore, the step of calculating the coordinate position of the weld defect in a preset three-dimensional coordinate system by combining the beam propagation path and using geometric optics includes:

[0047] The coordinate position is calculated according to the following formula:

[0048] ;

[0049] ;

[0050] ;

[0051] Where, x d y d z dThe coordinates of the x-axis, y-axis, and z-axis in the coordinate position are respectively, x1, y1, and z1 are the position coordinates of the reflector in the optical path transmission module, θ1 is the reflection angle of the light beam in the XZ plane, θ2 is the reflection angle of the light beam in the YZ plane, v is the speed of light, and t is the propagation time of the light beam.

[0052] This invention provides a laser-based weld quality inspection system and method. The system emits a laser beam and collects the reflected signal using a laser acquisition module. An optical path transmission module converts the beam signal into an electrical signal, which is then processed by a signal processing module to perform weld quality inspection. This achieves automated and high-precision weld quality inspection. Compared to traditional manual weld quality monitoring methods, this invention reduces the influence of subjective factors on the monitoring results, enabling continuous and accurate monitoring of the welding process. This improves the effectiveness and stability of weld quality inspection, thereby increasing production efficiency and product qualification rate. It meets the demands of modern manufacturing for automated and high-precision processing, and features consistent weld formation, data traceability, and ease of operation. Attached Figure Description

[0053] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0054] Figure 1 A system architecture diagram of a weld quality inspection system based on laser technology provided in an embodiment of the present invention;

[0055] Figure 2 A schematic diagram of the optical path transmission module in a laser-based weld quality inspection system provided in this embodiment of the invention;

[0056] Figure 3 A calibration schematic diagram of a weld quality inspection system based on laser technology provided in an embodiment of the present invention;

[0057] Figure 4 This is a schematic flowchart of a weld quality inspection method based on laser technology provided in an embodiment of the present invention. Detailed Implementation

[0058] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0060] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0061] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0062] Please see below. Figure 1 This invention provides a weld quality inspection system based on laser technology, comprising:

[0063] The laser acquisition module 100 is used to emit a laser beam toward the target weld to be inspected and to acquire the beam signal reflected by the target weld.

[0064] An optical path transmission module 200 is used to receive and transmit the beam signal; wherein, the optical path transmission module 200 includes an optical path adjustment unit and a photoelectric receiving unit, the optical path adjustment unit is used to adjust the propagation path of the beam signal to propagate to the photoelectric receiving unit, and the photoelectric receiving unit is used to convert the beam signal into an electrical signal;

[0065] The signal processing module is used to receive the electrical signal and perform weld quality inspection based on the electrical signal.

[0066] The weld quality inspection system described in this embodiment includes a laser acquisition module 100, an optical path transmission module 200, and a signal processing module. The laser acquisition module 100 emits a laser beam and collects the reflected signal. The optical path transmission module 200 converts the beam signal into an electrical signal, and the signal processing module performs weld quality inspection based on the electrical signal, achieving automation and high precision in weld quality inspection. Compared with traditional manual weld quality monitoring methods, this embodiment reduces the influence of subjective factors on the monitoring results, enabling continuous and accurate monitoring of the welding process. This improves the effectiveness and stability of weld quality inspection, thereby increasing production efficiency and product qualification rate, meeting the demands of modern manufacturing for automated and high-precision processing. Furthermore, it features consistent weld formation, data traceability, and ease of operation. Therefore, the weld quality inspection system provided in this embodiment can be widely applied in laser processing fields across multiple industries, covering laser welding needs in automotive manufacturing, consumer electronics, new energy development, and sheet metal processing.

[0067] In one embodiment, the optical path adjustment unit is provided with:

[0068] A reflector for receiving and reflecting the light beam signal;

[0069] A focusing element for focusing and propagating the beam signal;

[0070] A beam splitter is used to reflect part of the beam signal and refract another part of the beam signal.

[0071] A filter is used to allow light beam signals of a specified wavelength to pass through and to cut off light beam signals of another specified wavelength.

[0072] Accordingly, multiple photoelectric receiving units are provided, and each of the multiple photoelectric receiving units receives beam signals of different wavelengths.

[0073] The optical path adjustment unit described in this embodiment precisely adjusts the beam signal by incorporating components such as reflectors, focusing elements, beam splitters, and filters, ensuring that beam signals of different wavelengths can be accurately received by the corresponding photoelectric receiving units. Specifically, the reflector changes the beam's propagation direction, guiding it along a predetermined path; the focusing element ensures the beam's concentrated propagation, preventing beam diffusion during transmission and thus protecting signal quality; the beam splitter separates the beam signal, providing a basis for signal processing in different wavelengths; and the filter selects the desired wavelength beam signal, eliminating interference.

[0074] By setting up multiple photoelectric receiving units, beam signals of different wavelengths can be received separately, thus acquiring more comprehensive and detailed physical quantity information. For example, a photoelectric sensor in the 450-750nm wavelength band can capture specific optical signal characteristics, reflecting the physical properties of the weld in that wavelength band; a sensor in the 830-1180nm wavelength band can acquire another part of related information; and a sensor in the wavelength band greater than 1200nm can supplement other aspects of physical quantity characteristics. Through this multi-band, multi-sensor acquisition method, the true condition of the weld can be reflected more accurately.

[0075] This embodiment achieves non-contact acquisition of physical quantity information through a laser acquisition module 100. Physical quantities exist in different characteristic optical signal forms and can be acquired by multiple photoelectric receiving units. Specifically, three different photoelectric receiving units can be used, with wavelength ranges of 450~750nm, 830~1180nm, and >1200nm, respectively. Once these three photoelectric receiving units identify their respective characteristic optical signals, they can be converted into electrical signals by a sensor board.

[0076] In a specific embodiment, combined with Figure 2 The optical path transmission module 200 is provided with the following components sequentially along the optical path propagation direction:

[0077] Reflector 201;

[0078] The focusing element 202 is disposed in the reflection direction of the reflecting element 201 and is used to focus the beam signal;

[0079] The first beam splitter 203 is used to perform a first reflection and refraction on the beam signal focused by the focusing element 202 to generate a first reflected light path and a first refracted light path.

[0080] The first filter element 204 is disposed on the first reflected light path and is used to filter the light beam signal of the first reflection;

[0081] The first photoelectric receiving unit 205 is disposed on the first reflected light path and is used to receive the light beam signal transmitted through the first filter 204;

[0082] The second beam splitter 206 is disposed on the first refracted light path and is used to reflect and refract the first refracted beam signal a second time to generate a second reflected light path and a second refracted light path.

[0083] The second filter 207 is disposed on the second reflected light path and is used to filter the beam signal of the second reflected beam.

[0084] The second photoelectric receiving unit 208 is disposed on the second reflected light path and is used to receive the light beam signal transmitted through the second filter 207;

[0085] The third filter element 209 is disposed on the second refracted light path and is used to filter the light beam signal refracted in the second way.

[0086] The third photoelectric receiving unit 210 is disposed on the second refracted light path and is used to receive the light beam signal transmitted through the third filter 209.

[0087] In addition, a first aperture 211, a first calibration element 212 and a first light-diffusing element 213 are provided between the first filter element 204 and the first photoelectric receiving unit 205;

[0088] A second aperture 214, a second calibration element 215, and a second light-diffusing element 216 are disposed between the second filter element 207 and the second photoelectric receiving unit 208.

[0089] A third aperture 217, a third calibration element 218, and a third light-diffusing element 219 are provided between the third filter element 209 and the third photoelectric receiving unit 210.

[0090] In this embodiment, the beam signal (i.e., the laser beam reflected by the weld target) is first completely reflected by the reflector 201. After complete reflection, the beam signal is focused by the focusing element 202 and converged to the first beam splitter 203. The first beam splitter 203 splits the beam signal into a reflected beam and a refracted beam, thereby forming a first reflected optical path and a first refracted optical path. On the first reflected optical path, the beam signal passes through the first filter 204, the first aperture 211, the first calibration element 212, and the first beam homogenizer 213 before being received by the first photoelectric receiving unit 205. On the first refracted optical path, the beam signal passes through the second beam splitter 206, which again splits the beam signal into a reflected beam and a refracted beam, thereby forming a second reflected optical path and a second refracted optical path. In the second reflected light path, the beam signal passes through the second filter 207, the second aperture 214, the second calibration element 215, and the second homogenizer 216 before being received by the second photoelectric receiving unit 208. In the first refracted light path, the beam signal is transmitted to the third filter 209, and after passing through the third aperture 217, the third calibration element 218, and the third homogenizer 219, it is received by the third photoelectric receiving unit 210. After receiving the corresponding beam signals, the three photoelectric receiving units convert them into corresponding electrical signals and send them to the signal processing module for corresponding signal processing and quality detection.

[0091] In practical applications, the reflector 201 can be a mirror; the focusing element can be a plano-convex lens, which can convert a diverging beam into a focused beam. Here, the first calibration element 212, the second calibration element 215, and the third calibration element 218 can also be plano-convex lenses to achieve beam collimation calibration; the filter can be a bandpass filter or a cutoff attenuator, which can effectively filter out unwanted wavelengths, allowing only specific wavelengths of beam signals to pass through. For example, using a bandpass filter can improve signal purity and reduce the impact of external interference on the detection results; the beam splitter can be a beam splitter prism, which can accurately reflect and refract beam signals according to different optical characteristics, thereby achieving the separation of different optical paths; for the aperture and beam homogenizer, the aperture can precisely control the aperture and shape of the beam, preventing stray light from entering the photoelectric receiving unit and further improving signal quality. The beam homogenizer can make the light intensity distribution of the beam on the cross-section more uniform, making the signal received by the photoelectric receiving unit more stable and reliable. Furthermore, frosted glass can be used to attenuate beam signals of specific wavelengths. Furthermore, the specific angles of the reflectors and beam splitters can be adjusted according to the actual optical path design and detection requirements to ensure that the beam signal can accurately propagate to each photoelectric receiving unit. For example, the beam splitter can be at a 45-degree angle to the incident angle of the beam signal. When installing these optical components, it is necessary to ensure their positional and angular accuracy to avoid deviations in the beam signal propagation path due to installation errors, which would affect the accuracy of the detection results.

[0092] To further improve the stability and reliability of the weld quality inspection system, the optical transmission module 200 can also be regularly maintained and calibrated. For example, check the surface of the reflector for stains or damage, and clean or replace it in time; check whether the focal length of the focusing element has changed, and adjust it if necessary; check whether the filtration effect of the filter element meets the requirements, and prevent interference signals from being introduced due to a decrease in filtration performance.

[0093] In another embodiment, combined Figure 3 The reflector 201 is set as an adjustable reflector. In the process of system application, if the photoelectric receiving unit cannot obtain the beam signal, it may be that the optical axis is deviated from the central axis. Therefore, by deflecting the angle of the reflector, the optical axis is made concentric with the sensor axis, so that the photoelectric receiving unit can effectively capture the required beam signal.

[0094] In practical applications, the weld quality inspection system is tested and calibrated before operation, and real-time dynamic calibration can also be performed during the weld quality inspection process. Specifically, during initial calibration (i.e., before testing), a standard calibration piece can be placed first. For example, a standard weld calibration piece with known dimensions and known defect parameters can be placed at the testing station. The defect parameters of the standard calibration piece include defect type (such as porosity, slag inclusion, crack) and defect size (length, width, depth), and the parameters are metrologically certified. Then, laser parameter calibration is performed, that is, the laser power, laser wavelength, and laser scanning frequency of the laser acquisition module 100 are adjusted to control the laser beam to focus on the weld surface of the standard calibration piece, the reflected beam signal is collected and converted into a calibration electrical signal, and the laser power, laser wavelength, and scanning frequency are adjusted accordingly by comparing the characteristics of the calibration electrical signal with the known parameters of the standard calibration piece. Then, optical path parameter calibration is performed, that is, the angle of the reflector 201 and the position of the focusing element 202 in the optical path adjustment unit are adjusted so that the beam signal can be accurately transmitted to the photoelectric receiving unit. By collecting the reflected beam signals at different positions of the standard calibration piece, the optical path transmission efficiency is calculated so that the optical path transmission efficiency meets the usage requirements.

[0095] During real-time dynamic calibration (in detection), a calibration interval is first set, such as a dynamic calibration interval based on the detection duration and changes in the detection environment. Then, real-time calibration is performed. That is, when the calibration interval is reached, the detection of the target is paused, the laser beam is switched to a preset built-in calibration target point, the reflected beam signal is acquired and converted into a calibration electrical signal, and the deviation between the real-time calibration electrical signal characteristic parameters (such as amplitude and frequency) and the standard characteristic parameters after initial calibration is calculated. Then, the parameters are adjusted based on the deviation. The optical path angle is also adjusted; for example, based on the offset direction of the real-time beam signal, the angle of the reflector 201 is adjusted to ensure that the real-time beam signal transmission path is consistent with the initial calibration path. The adjustment amount can be determined by calculating the beam offset using an image recognition algorithm.

[0096] like Figure 4 As shown, this embodiment of the invention provides a weld quality inspection method based on laser technology, which is applied to the weld quality inspection system based on laser technology as described above. The method specifically includes steps S101 to S103.

[0097] Step S101: Use the laser acquisition module 100 to emit a laser beam toward the target weld to be inspected, and acquire the beam signal reflected by the target weld;

[0098] Step S102: Receive and transmit the beam signal through the optical path transmission module 200, and convert the beam signal into an electrical signal through the photoelectric receiving unit in the optical path transmission module 200;

[0099] Step S103: Receive the electrical signal through the signal processing module and perform weld quality inspection based on the electrical signal.

[0100] This embodiment emits a laser beam and collects the reflected signal through the laser acquisition module 100. The optical path transmission module 200 converts the beam signal into an electrical signal, and the signal processing module performs weld quality inspection based on the electrical signal, thus achieving automation and high precision in weld quality inspection. Compared with traditional manual weld quality monitoring methods, this embodiment reduces the influence of subjective factors on the monitoring results, enables continuous and accurate monitoring of the welding process, improves the effectiveness and stability of weld quality inspection, and thus increases production efficiency and product qualification rate. It meets the needs of modern manufacturing for automated and high-precision processing, and also features consistent weld formation, data traceability, and simple operation.

[0101] In one embodiment, receiving the electrical signal via a signal processing module and performing weld quality inspection based on the electrical signal includes:

[0102] The wavelet threshold denoising algorithm is used to denoise the received electrical signal, and the denoised electrical signal is then normalized.

[0103] Weld quality feature parameters are extracted from the normalized electrical signal; wherein, the weld quality feature parameters include grayscale features and geometric features;

[0104] The weld quality characteristic parameters are classified and predicted using a classification model to obtain the corresponding classification prediction results, and the classification prediction results are output as the weld quality inspection results.

[0105] This embodiment first preprocesses the electrical signal, including noise reduction and normalization. Noise reduction removes noise interference from the electrical signal, making it cleaner and improving the accuracy of subsequent feature extraction. Normalization unifies the numerical range of the electrical signal, preventing large numerical differences from affecting the extraction and analysis of feature parameters. Then, weld quality feature parameters are extracted, specifically grayscale and geometric features. Grayscale features reflect the brightness distribution of the weld area; different weld qualities may correspond to different grayscale features. For example, surface defects in the weld may cause abnormal changes in grayscale values. Geometric features describe the shape and size of the weld; features such as weld width, length, and flatness are closely related to weld quality. Accurate extraction of these feature parameters provides strong data support for subsequent classification and prediction. Classification and prediction are then achieved through a classification model, which can be built based on machine learning algorithms such as support vector machines, decision trees, and neural networks.

[0106] Specifically, when using wavelet threshold denoising algorithm to denoise the received electrical signal, it can be achieved according to the following formula:

[0107] ;

[0108] in, These are the wavelet coefficients after noise reduction. These are the original wavelet coefficients. The noise reduction threshold is set.

[0109] When extracting grayscale features, the mean μ and standard deviation of the grayscale value of the corresponding image can be calculated using the following formulas. :

[0110] ;

[0111] ;

[0112] Where M and N are the number of rows and columns of the image, respectively, and f(i,j) is the gray value of the image at position (i,j).

[0113] When extracting geometric features, specifically the width, reinforcement height, and undercut depth of the weld are extracted. Here, the weld edge coordinates can be determined by an edge detection algorithm, and then the feature values ​​are calculated based on the coordinates.

[0114] When using a classification model to predict the weld quality feature parameters, cross-validation can be used to evaluate and optimize the model. Specifically, the dataset can be divided into training and test sets; the model is trained on the training set and its performance is validated on the test set. Training the classification model requires a large amount of labeled data, which should cover weld samples of different quality levels to ensure the model can learn patterns of various weld quality features. The quality and diversity of the labeled data directly affect the model's generalization ability. Labeled data can be obtained through manual annotation or by combining inspection results from actual production. To further improve the accuracy of classification prediction, ensemble learning can be used on the classification model. Ensemble learning combines multiple weak classifiers to build a strong classifier. For example, the Bagging method can be used to perform multiple samplings with replacement on the original dataset to obtain multiple different training subsets. Then, a classifier is trained on each training subset, and finally, the prediction results of these classifiers are combined.

[0115] In one embodiment, after the step of receiving the electrical signal through the signal processing module and performing weld quality inspection based on the electrical signal, the method further includes:

[0116] When the weld quality inspection result indicates the presence of weld defects, the coordinate position of the weld defects in a preset three-dimensional coordinate system is calculated using geometric optics, based on the beam propagation path. The three-dimensional coordinate system has the origin at the starting point of the laser beam emitted by the laser acquisition module 100, the positive Z-axis as the laser beam emission direction, the horizontal axis perpendicular to the laser beam emission direction as the X-axis, and the vertical axis perpendicular to the laser beam emission direction as the Y-axis.

[0117] The coordinate positions are correlated with the target weld, and a defect location distribution map is generated.

[0118] Specifically, the step of calculating the coordinate position of the weld defect in a preset three-dimensional coordinate system by combining the beam propagation path and using geometric optics includes:

[0119] The coordinate position is calculated according to the following formula:

[0120] ;

[0121] ;

[0122] ;

[0123] Where, x d y d z d The coordinates of the x-axis, y-axis, and z-axis in the coordinate position are respectively, x1, y1, and z1 are the position coordinates of the reflector 201 in the optical path transmission module 200, θ1 is the reflection angle of the light beam in the XZ plane, θ2 is the reflection angle of the light beam in the YZ plane, v is the speed of light, and t is the propagation time of the light beam.

[0124] In this embodiment, when a quality problem is detected in the target weld, i.e. when a weld defect exists, the coordinate position of the weld defect in a preset three-dimensional coordinate system is first calculated by geometric optics. The coordinate system takes the starting point of the laser beam emitted by the laser acquisition module 100 as the origin, the laser beam emission direction as the positive Z-axis, the horizontal direction perpendicular to the laser beam emission direction as the X-axis, and the vertical direction perpendicular to the laser beam emission direction as the Y-axis.

[0125] Next, the calculated coordinates are correlated with the target weld to generate a defect location distribution map. This map clearly shows the specific location of weld defects within the entire weld, allowing maintenance personnel to quickly locate the defects and take targeted repair measures, thus improving repair efficiency and quality.

[0126] Furthermore, to further improve the effectiveness of inspection and repair, the generated defect location distribution map can be compared and analyzed with the weld design drawings. This comparison allows for a clear visual indication of the deviation between the defect location and design requirements, and an assessment of the impact of defects on the overall weld performance. If a significant deviation is found, it may be necessary to re-evaluate the repair plan or even adjust the welding process to ensure the weld quality meets standards. Simultaneously, the defect location distribution map and related inspection data can be stored and managed to establish a weld quality inspection database. This database records detailed information for each weld inspection, including basic weld information, inspection time, inspection results, and defect location distribution map. Analysis and mining of this database can summarize the patterns and trends of weld defect occurrence, providing data support for improving welding processes and enhancing weld quality. For example, if frequent weld defects are found at a specific location, welding process parameters can be optimized for that location, such as adjusting welding speed and current intensity, thereby reducing defect occurrence.

[0127] In another embodiment, after the step of receiving the electrical signal via the signal processing module and performing weld quality inspection based on the electrical signal, the method further includes:

[0128] When the weld quality inspection result indicates the presence of weld defects, a Fourier transform is performed on the electrical signal to convert the time-domain signal corresponding to the electrical signal into a frequency-domain signal.

[0129] A blind source separation algorithm is used to separate the signal components corresponding to the weld defect from the frequency domain signal;

[0130] A separation matrix is ​​constructed based on the separated signal components, and the separation matrix is ​​solved using the negative entropy maximization criterion to obtain the defect signal;

[0131] The defect signal is subjected to inverse Fourier transform to obtain the corresponding time-domain defect signal, and the size information of the weld defect is obtained based on the time-domain defect signal.

[0132] In addition to locating weld defects, this embodiment also specifically detects the size information of the weld defects. First, a Fourier transform is performed on the electrical signal to convert the time-domain signal into a frequency-domain signal. This is because the frequency components of the signal are easier to analyze in the frequency domain, which helps to separate the signal corresponding to the weld defect. Next, a blind source separation algorithm is used. This algorithm can separate individual signal components from the mixed signal without knowing the specific information of the signal source and transmission channel, thus separating the signal component corresponding to the weld defect from the frequency-domain signal. Then, a separation matrix is ​​constructed based on the separated signal components. The separation matrix is ​​solved using the negative entropy maximization criterion. Negative entropy is an indicator of the non-Gaussianity of a signal; maximizing negative entropy makes the separated signals as independent as possible, thus obtaining an accurate defect signal. Afterwards, an inverse Fourier transform is performed on the defect signal to convert it back to the time domain, thus obtaining the time-domain defect signal corresponding to the weld defect. The size information of the weld defect can be obtained from the obtained time-domain defect signal. For example, by analyzing the amplitude, width, and other characteristics of the time-domain defect signal, the specific size of the weld defect can be calculated.

[0133] When using a blind source separation algorithm to separate the signal components corresponding to the weld defects from the frequency domain signal, the Independent Component Analysis (ICA) algorithm can be specifically used to separate the signal components corresponding to the internal weld defects from the frequency domain signal. The goal is to find a separation matrix W such that the separated signals S = WX (where X is the observed signal matrix, i.e., the frequency domain signal matrix) maximize independence. The separation matrix W is solved using the negative entropy maximization criterion. The formula for calculating negative entropy is:

[0134] ;

[0135] Where s is the separated signal, and E{·} is the expectation operator.

[0136] When obtaining the size information of the weld defect based on the time-domain defect signal, the specific steps may include:

[0137] Defect length: L = v s ×T, where v s The laser velocity emitted by the detection system is T, and the duration of the time-domain defect signal is T.

[0138] Defect width: Where k1 is the width correction coefficient, which can be obtained through calibration using standard defect samples;

[0139] Defect depth: , where k2 is the depth correction coefficient, which can also be determined by calibration using standard defect samples.

[0140] Furthermore, regarding the obtained dimensional information, if the defect size is small, it indicates minimal impact on the overall weld performance, allowing for simple repair measures. However, if the defect size is large, re-welding or weld reinforcement may be necessary. Additionally, the weld defect size information can be compared with the weld design standards. If the defect size exceeds the allowable range of the design standards, the weld may not meet quality requirements and requires appropriate treatment. Simultaneously, correlating and comparing this defect size information with data previously stored in the weld quality inspection database can further summarize the distribution patterns of different types of weld defect sizes, providing more targeted data support for subsequent welding process optimization. For example, if a certain type of weld frequently exhibits large-sized defects, the welding process for that type of weld can be improved by adjusting welding parameters or replacing welding materials to reduce the probability of large-sized defects and improve the overall weld quality.

[0141] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to in the method section. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of the claims of this application.

[0142] It should also be noted that, in this specification, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

Claims

1. A weld quality inspection system based on laser technology, characterized in that, include: The laser acquisition module is used to emit a laser beam toward the target weld to be inspected and to acquire the beam signal reflected by the target weld. An optical path transmission module is used to receive and transmit the beam signal; wherein, the optical path transmission module includes an optical path adjustment unit and a photoelectric receiving unit, the optical path adjustment unit is used to adjust the propagation path of the beam signal to propagate to the photoelectric receiving unit, and the photoelectric receiving unit is used to convert the beam signal into an electrical signal; A signal processing module is used to receive the electrical signal and perform weld quality inspection based on the electrical signal; The step of receiving the electrical signal and performing weld quality inspection based on the electrical signal includes: The wavelet threshold denoising algorithm is used to denoise the received electrical signal, and the denoised electrical signal is then normalized. Weld quality feature parameters are extracted from the normalized electrical signal; wherein, the weld quality feature parameters include grayscale features and geometric features; The weld quality characteristic parameters are classified and predicted using a classification model to obtain the corresponding classification prediction results, and the classification prediction results are output as the weld quality inspection results. The signal processing module is also used for: When the weld quality inspection result indicates the presence of weld defects, a Fourier transform is performed on the electrical signal to convert the time-domain signal corresponding to the electrical signal into a frequency-domain signal. A blind source separation algorithm is used to separate the signal components corresponding to the weld defect from the frequency domain signal; A separation matrix is ​​constructed based on the separated signal components, and the separation matrix is ​​solved using the negative entropy maximization criterion to obtain the defect signal; The defect signal is subjected to inverse Fourier transform to obtain the corresponding time-domain defect signal, and the size information of the weld defect is obtained based on the time-domain defect signal.

2. The weld quality inspection system based on laser technology according to claim 1, characterized in that, The optical path adjustment unit is equipped with: A reflector for receiving and reflecting the light beam signal; A focusing element for focusing and propagating the beam signal; A beam splitter is used to reflect part of the beam signal and refract another part of the beam signal. A filter is used to allow light beam signals of a specified wavelength to pass through and to cut off light beam signals of another specified wavelength.

3. The weld quality inspection system based on laser technology according to claim 2, characterized in that, The photoelectric receiving unit is provided in multiple ways, and each of the multiple photoelectric receiving units receives beam signals of different wavelengths.

4. The weld quality inspection system based on laser technology according to claim 3, characterized in that, The optical path transmission module is provided with the following components sequentially along the optical path propagation direction: Reflector; A focusing element, disposed in the reflection direction of the reflector, is used to focus the light beam signal; The first beam splitter is used to reflect and refract the beam signal focused by the focusing element for the first time to generate a first reflected light path and a first refracted light path. The first filter element is disposed on the first reflected light path and is used to filter the light beam signal of the first reflection; The first photoelectric receiving unit is disposed on the first reflected light path and is used to receive the light beam signal transmitted through the first filter. The second beam splitter is disposed on the first refracted light path and is used to reflect and refract the light beam signal refracted in the first refracted light a second time to generate a second reflected light path and a second refracted light path. The second filter element is disposed on the second reflected light path and is used to filter the light beam signal of the second reflection. The second photoelectric receiving unit is disposed on the second reflected light path and is used to receive the light beam signal transmitted through the second filter. The third filter element is disposed on the second refracted light path and is used to filter the light beam signal refracted in the second refracted light. The third photoelectric receiving unit is disposed on the second refracted light path and is used to receive the light beam signal transmitted through the third filter.

5. The weld quality inspection system based on laser technology according to claim 4, characterized in that, A first aperture, a first calibration element, and a first light-diffusing element are disposed between the first filter element and the first photoelectric receiving unit; A second aperture, a second calibration element, and a second light-diffusing element are disposed between the second filter element and the second photoelectric receiving unit; A third aperture, a third calibration element, and a third light-diffusing element are provided between the third filter element and the third photoelectric receiving unit.

6. A weld quality inspection method based on laser technology, applied to the weld quality inspection system based on laser technology as described in any one of claims 1-5, characterized in that, The method includes: A laser beam is emitted toward the target weld to be inspected using a laser acquisition module, and the beam signal reflected by the target weld is acquired. The beam signal is received and transmitted through the optical path transmission module, and the beam signal is converted into an electrical signal through the photoelectric receiving unit in the optical path transmission module. The electrical signal is received by the signal processing module, and the weld quality is inspected based on the electrical signal. The step of receiving the electrical signal through a signal processing module and performing weld quality inspection based on the electrical signal includes: The wavelet threshold denoising algorithm is used to denoise the received electrical signal, and the denoised electrical signal is then normalized. Weld quality feature parameters are extracted from the normalized electrical signal; wherein, the weld quality feature parameters include grayscale features and geometric features; The weld quality characteristic parameters are classified and predicted using a classification model to obtain the corresponding classification prediction results, and the classification prediction results are output as the weld quality inspection results. After receiving the electrical signal through the signal processing module and performing weld quality inspection based on the electrical signal, the process includes: When the weld quality inspection result indicates the presence of weld defects, a Fourier transform is performed on the electrical signal to convert the time-domain signal corresponding to the electrical signal into a frequency-domain signal. A blind source separation algorithm is used to separate the signal components corresponding to the weld defect from the frequency domain signal; A separation matrix is ​​constructed based on the separated signal components, and the separation matrix is ​​solved using the negative entropy maximization criterion to obtain the defect signal; The defect signal is subjected to inverse Fourier transform to obtain the corresponding time-domain defect signal, and the size information of the weld defect is obtained based on the time-domain defect signal.

7. The weld quality inspection method based on laser technology according to claim 6, characterized in that, After receiving the electrical signal through the signal processing module and performing weld quality inspection based on the electrical signal, the method further includes: When the weld quality inspection result indicates the presence of weld defects, the coordinate position of the weld defects in a preset three-dimensional coordinate system is calculated using geometric optics, based on the beam propagation path. The three-dimensional coordinate system has the origin at the starting point of the laser beam emitted by the laser acquisition module, the positive Z-axis is the direction of laser beam emission, the horizontal axis perpendicular to the direction of laser beam emission is the X-axis, and the vertical axis perpendicular to the direction of laser beam emission is the Y-axis. The coordinate positions are correlated with the target weld, and a defect location distribution map is generated.

Images