Welding defect recognition method and device, storage medium and electronic device

Abstract

This invention provides a method, apparatus, storage medium, and electronic device for identifying welding defects. The method includes: acquiring an original time-series infrared thermal image of the weld region using an infrared thermal imager while welding a thin plate workpiece along the extension direction of the weld region; extracting the region of interest (ROI) from the original time-series infrared thermal image and calibrating the ROI to obtain a time-series temperature distribution image, wherein the time-series temperature distribution image characterizes the temperature value of each pixel; extracting thermal feature parameters from the time-series temperature distribution image, wherein the thermal feature parameters include temperature peak features, temperature gradient features, and thermal diffusion features; predicting the back weld width of the weld region based on the thermal feature parameters; and identifying welding defects in the weld region using the thermal feature parameters and the back weld width. This embodiment solves the technical problem of the inability to detect welding defects in thin plate workpieces in the prior art, significantly improving the efficiency and accuracy of thin plate welding quality inspection.

Description

Technical Field

[0001] This invention relates to the field of vehicle technology, and more specifically, to a method and apparatus for identifying welding defects, a storage medium, and an electronic device. Background Technology

[0002] Among related technologies, thin-plate laser welding is widely used in fields such as automotive battery housings, aerospace thin-walled parts, and precision sheet metal for home appliances due to its concentrated heat input, small deformation, and high precision.

[0003] In related technologies, laser butt welding of ultra-thin plates of 0.1–1.0 mm has the characteristics of shallow penetration (0.2–1.0 mm), easy burn-through, and hidden incomplete penetration defects. Traditional inspection methods have obvious limitations: visual inspection can only identify obvious defects such as surface spatter, undercut, and burn-through, and cannot identify hidden defects such as internal incomplete penetration and insufficient back weld width; destructive inspection (such as metallographic sectioning and tensile testing) is inefficient and costly, and cannot achieve batch online inspection; conventional non-destructive testing technologies (such as ultrasound and X-ray) have poor compatibility, the coupling agent of ultrasonic probes is easy to contaminate thin plate precision components, X-ray inspection has radiation risks and insufficient resolution for small incomplete penetration; existing infrared thermal imaging inspection technology is mostly for thick plate welding, and only judges defects through a single temperature threshold, with poor anti-interference ability and cannot accurately identify micron-level penetration depth changes.

[0004] No efficient and accurate solution has yet been found to address the aforementioned issues in the relevant technologies. Summary of the Invention

[0005] This invention provides a method and apparatus for identifying welding defects, a storage medium, and an electronic device to solve technical problems in related technologies.

[0006] According to an embodiment of the present invention, a method for identifying welding defects is provided, comprising: when welding a thin plate workpiece along the extension direction of the weld region, acquiring an original time-series infrared thermal image of the weld region collected by an infrared thermal imager, wherein the infrared thermal imager is vertically positioned directly above the weld region, and the original time-series infrared thermal image includes multiple original infrared thermal images arranged in time sequence; extracting a region of interest from the original time-series infrared thermal image and performing temperature calibration on the region of interest to obtain a time-series temperature distribution image, wherein the time-series temperature distribution image is used to characterize the temperature value of each pixel; extracting thermal feature parameters from the time-series temperature distribution image, wherein the thermal feature parameters include temperature peak features, temperature gradient features, and thermal diffusion features; predicting the back weld width of the weld region based on the thermal feature parameters; and identifying welding defects in the weld region using the thermal feature parameters and the back weld width.

[0007] Optionally, extracting the region of interest from the original time-series infrared thermal image includes: denoising the original time-series infrared thermal image using Gaussian filtering to obtain an intermediate time-series infrared thermal image; denoising the intermediate time-series infrared thermal image using a median filtering algorithm to obtain a target time-series infrared thermal image; identifying the weld center region and the heat-affected zone surrounding the weld center region in the target time-series infrared thermal image, and outputting the weld center region and the heat-affected zone as regions of interest.

[0008] Optionally, extracting the thermal feature parameters of the time-series temperature distribution image includes: extracting the weld center pixel in the time-series temperature distribution image, and determining the temperature peak value of the weld center pixel in multiple temperature distribution images; recording the occurrence time and duration of the temperature peak value in multiple temperature distribution images; and determining the temperature peak value, the occurrence time, and the duration as the temperature peak value feature.

[0009] Optionally, extracting the thermal feature parameters of the time-series temperature distribution image includes: for every two adjacent temperature distribution images in the time-series temperature distribution image, starting from the starting pixel of the initial temperature distribution image, selecting a first pixel and a second pixel along the welding direction, and a third pixel and a fourth pixel along the vertical direction of the welding direction, respectively; calculating a first temperature change value between the first pixel and the second pixel, calculating a horizontal length value between the first pixel and the second pixel, calculating a second temperature change value between the third pixel and the fourth pixel, and calculating a vertical length value between the third pixel and the fourth pixel; calculating a horizontal temperature gradient based on the ratio of the first temperature change value to the horizontal length value, and calculating a vertical temperature gradient based on the ratio of the second temperature change value to the vertical length value; and determining the horizontal temperature gradient and the vertical temperature gradient as the temperature gradient features.

[0010] Optionally, extracting the thermal feature parameters of the time-series temperature distribution image includes: reading the property parameters of the welding material of the thin plate workpiece, wherein the property parameters include thermal conductivity, material density, and specific heat capacity at constant pressure; calculating the thermal diffusivity of the welding material using the property parameters; determining the heat source application time of the welding material in the time-series temperature distribution image; calculating the thermal diffusivity radius of the heat-affected zone using the thermal diffusivity and the heat source application time; and determining the thermal diffusivity and the thermal diffusivity radius as the thermal diffusivity feature.

[0011] Optionally, predicting the back weld width of the weld region based on the thermal characteristic parameters includes calculating the back weld width of the weld region using the following formula. : ;in, The temperature peak value in the aforementioned temperature peak characteristic. The mean value of the horizontal temperature gradient in the temperature gradient feature is given. Here, a0, a1, a2, and a3 represent the thermal diffusivity coefficients in the aforementioned thermal diffusivity characteristics, and a0, a1, a2, and a3 are all pre-fitted regression coefficients. This is the residual term.

[0012] Optionally, identifying welding defects in the weld region using the thermal feature parameters and the back weld width includes: constructing input data using the thermal feature parameters and the back weld width; inputting the input data into a pre-trained convolutional neural network model, and outputting the welding defects in the weld region, wherein the convolutional neural network model is a welding defect prediction model trained based on a sample dataset.

[0013] Optionally, identifying welding defects in the weld region using the thermal characteristic parameters and the back weld width includes: determining whether there is an abnormal abrupt change in the temperature gradient based on the temperature gradient characteristics, and determining whether the back weld width is less than a first preset thickness; if there is no abnormal abrupt change in the temperature gradient, and the back weld width is greater than or equal to the first preset thickness, determining that there is no welding defect in the weld region; determining whether there is an abnormal abrupt change in the temperature gradient based on the temperature gradient characteristics, and determining whether the back weld width is less than the first preset thickness; if there is an abnormal abrupt change in the temperature gradient, and the back weld width is less than the first preset thickness, determining whether the back weld width is greater than or equal to a second preset thickness; if the back weld width is greater than or equal to the second preset thickness, determining that the welding defect is an incomplete penetration defect; determining whether the back weld width is greater than a third preset thickness, and determining whether the welding temperature peak exceeds the material melting point threshold of the thin plate workpiece based on the temperature peak characteristics, wherein the third preset thickness > the thickness of the thin plate workpiece > the first preset thickness > the second preset thickness; if the back weld width is greater than the third preset thickness, and the welding temperature peak exceeds the material melting point threshold of the thin plate workpiece, determining that the welding defect is a burn-through defect.

[0014] According to another embodiment of the present invention, a welding defect identification device is provided, comprising: an acquisition module, configured to acquire an original time-series infrared thermal image of the weld region collected by an infrared thermal imager when welding a thin plate workpiece along the extension direction of the weld region, wherein the infrared thermal imager is vertically disposed directly above the weld region, and the original time-series infrared thermal image includes a plurality of original infrared thermal images arranged in time sequence; a first extraction module, configured to extract a region of interest from the original time-series infrared thermal image and perform temperature calibration on the region of interest to obtain a time-series temperature distribution image, wherein the time-series temperature distribution image is used to characterize the temperature value of each pixel; a second extraction module, configured to extract thermal feature parameters of the time-series temperature distribution image, wherein the thermal feature parameters include temperature peak features, temperature gradient features, and thermal diffusion features; a prediction module, configured to predict the back weld width of the weld region based on the thermal feature parameters; and an identification module, configured to identify welding defects in the weld region using the thermal feature parameters and the back weld width.

[0015] Optionally, the first extraction module includes: a first denoising unit, used to denoise the original time-series infrared thermal image using Gaussian filtering to obtain an intermediate time-series infrared thermal image; The second denoising unit is used to denoise the intermediate time-series infrared thermal image using a median filtering algorithm to obtain the target time-series infrared thermal image; the recognition unit is used to recognize the weld center region and the heat-affected zone around the weld center region in the target time-series infrared thermal image, and output the weld center region and the heat-affected zone as regions of interest.

[0016] Optionally, the second extraction module includes: an extraction unit, configured to extract the weld center pixel in the time-series temperature distribution image and determine the temperature peak value of the weld center pixel in multiple temperature distribution images; a recording unit, configured to record the occurrence time and duration of the temperature peak value in multiple temperature distribution images; and a determination unit, configured to determine the temperature peak value, the occurrence time, and the duration as the temperature peak value feature.

[0017] Optionally, the second extraction module includes: a selection unit, configured to, for each pair of adjacent temperature distribution images in the time-series temperature distribution image, starting from the starting pixel of the initial temperature distribution image, select a first pixel and a second pixel from the two adjacent temperature distribution images along the welding direction, and select a third pixel and a fourth pixel along the vertical direction of the welding direction; a first calculation unit, configured to calculate a first temperature change value between the first pixel and the second pixel, calculate a horizontal length value between the first pixel and the second pixel, calculate a second temperature change value between the third pixel and the fourth pixel, and calculate a vertical length value between the third pixel and the fourth pixel; a second calculation unit, configured to calculate a horizontal temperature gradient based on the ratio of the first temperature change value to the horizontal length value, and calculate a vertical temperature gradient based on the ratio of the second temperature change value to the vertical length value; and a first determination unit, configured to determine the horizontal temperature gradient and the vertical temperature gradient as the temperature gradient features.

[0018] Optionally, the second extraction module includes: a reading unit for reading the property parameters of the welding material of the thin plate workpiece, wherein the property parameters include thermal conductivity, material density, and specific heat capacity at constant pressure; a third calculation unit for calculating the thermal diffusivity of the welding material using the property parameters; a second determination unit for determining the heat source action time of the welding material in the time-series temperature distribution image; a fourth calculation unit for calculating the thermal diffusivity radius of the heat-affected zone using the thermal diffusivity and the heat source action time; and a third determination unit for determining the thermal diffusivity and the thermal diffusivity radius as the thermal diffusivity feature.

[0019] Optionally, the prediction module includes a calculation unit for calculating the back weld width of the weld region using the following formula. : ;in, The temperature peak value in the aforementioned temperature peak characteristic. The mean value of the horizontal temperature gradient in the temperature gradient feature is given. Here, a0, a1, a2, and a3 represent the thermal diffusivity coefficients in the aforementioned thermal diffusivity characteristics, and a0, a1, a2, and a3 are all pre-fitted regression coefficients. This is the residual term.

[0020] Optionally, the identification module includes: a construction unit for constructing input data from the thermal feature parameters and the back weld width; and an output unit for inputting the input data into a pre-trained convolutional neural network model and outputting welding defects in the weld area, wherein the convolutional neural network model is a welding defect prediction model trained based on a sample dataset.

[0021] Optionally, the identification module includes: a first identification unit, configured to determine whether there is an abnormal abrupt change in the temperature gradient based on the temperature gradient characteristics, and to determine whether the back weld width is less than a first preset thickness; if there is no abnormal abrupt change in the temperature gradient, and the back weld width is greater than or equal to the first preset thickness, determine that there is no welding defect in the weld area; a second identification unit, configured to determine whether there is an abnormal abrupt change in the temperature gradient based on the temperature gradient characteristics, and to determine whether the back weld width is less than the first preset thickness; if there is an abnormal abrupt change in the temperature gradient, and the back weld width is less than the first preset thickness, determine whether the back weld width is greater than or equal to a second preset thickness; if the back weld width is greater than or equal to the second preset thickness, determine that the welding defect is an incomplete penetration defect; a third identification unit, configured to determine whether the back weld width is greater than a third preset thickness, and to determine whether the welding temperature peak exceeds the material melting point threshold of the thin plate workpiece based on the temperature peak characteristics, wherein the third preset thickness > the thickness of the thin plate workpiece > the first preset thickness > the second preset thickness; if the back weld width is greater than the third preset thickness, and the welding temperature peak exceeds the material melting point threshold of the thin plate workpiece, determine that the welding defect is a burn-through defect.

[0022] According to another aspect of the embodiments of this application, a storage medium is also provided, the storage medium including a stored program that executes the above steps when the program is run.

[0023] According to another aspect of the embodiments of this application, an electronic device is also provided, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; wherein: the memory is used to store computer programs; and the processor is used to execute the steps in the above method by running the programs stored in the memory.

[0024] This application also provides a computer program product containing instructions that, when run on a computer, cause the computer to perform the steps in the above-described method.

[0025] The beneficial effects of this invention are: 1. Strong adaptability: Specifically designed for laser butt welding of ultra-thin plates from 0.1 to 1.0 mm, it solves the problem of insufficient penetration depth and latent defect identification of thin plates by conventional inspection technology, and is compatible with common thin plate materials such as stainless steel and aluminum alloy. 2. Non-contact and non-destructive: Data is collected only through front infrared thermal imaging, without contact with the workpiece or the need to place sensors on the back. There is no radiation risk, the workpiece structure is not damaged, and the thin plate precision components are not contaminated. 3. High detection efficiency: Single-point detection time ≤ 0.1s, supports batch workpiece online detection, and is more than 100 times more efficient than traditional metallographic section detection; 4. High accuracy: By combining a dedicated transient heat transfer inversion model with a lightweight neural network, the back weld width prediction error is ≤0.02 mm, and the defect judgment accuracy is ≥98%, which can accurately identify hidden defects such as incomplete penetration and burn-through. 5. Closed-loop support: It can be linked with the welding machine control system to adjust parameters such as laser power and welding speed in real time based on the penetration depth prediction results, so as to achieve closed-loop quality control and further improve the stability of welding quality. Attached Figure Description

[0026] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings: Figure 1 This is a hardware structure block diagram of a computer according to an embodiment of the present invention; Figure 2 This is a flowchart of a welding defect identification method according to an embodiment of the present invention; Figure 3 This is a system setup diagram according to an embodiment of the present invention; Figure 4 These are comparison images of infrared thermal images of different defect types in embodiments of the present invention; Figure 5 This is a flowchart of the defect detection method in an embodiment of the present invention; Figure 6 This is a structural block diagram of a welding defect identification device according to an embodiment of the present invention. Detailed Implementation

[0027] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, and not all of them. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present application. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of the present application can be combined with each other.

[0028] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0029] Example 1 The method embodiment provided in Embodiment 1 of this application can be executed in a computer, welding controller, processor, welding robot, or similar processing device. Taking running on a computer as an example, Figure 1 This is a hardware structure block diagram of a computer according to an embodiment of the present invention. For example... Figure 1 As shown, a computer may include one or more ( Figure 1 Only one is shown in the diagram. A processor 102 (which may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.) and a memory 104 for storing data are also shown. Optionally, the computer may further include a transmission device 106 for communication functions and an input / output device 108. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the computer described above. For example, the computer may also include components that are larger than... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0030] The memory 104 can be used to store computer programs, such as application software programs and modules, like the computer program corresponding to a method for identifying welding defects in a computer according to an embodiment of the present invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, thereby implementing the aforementioned method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the computer via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0031] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by a computer's communication provider. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a Radio Frequency (RF) module used for wireless communication with the Internet.

[0032] This embodiment provides a method for identifying welding defects. Figure 2 This is a flowchart of a welding defect identification method according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps: Step S201: When welding a thin plate workpiece along the extension direction of the weld area, acquire the original time-series infrared thermal image of the weld area collected by an infrared thermal imager, wherein the infrared thermal imager is vertically positioned directly above the weld area, and the original time-series infrared thermal image includes multiple original infrared thermal images arranged in time sequence. Optionally, the thin sheet workpiece is an ultra-thin sheet with a thickness of 0.1 to 1.0 mm.

[0033] The detection system in this embodiment includes a high-speed infrared thermal imager, a synchronous control module, a data processing terminal, and a display module. The high-speed infrared thermal imager is installed on the side of the laser welding machine head, with the lens vertically aimed at the weld area of ​​the thin plate. It has a data acquisition frequency of ≥500 frames / second, a temperature measurement range of 0~500℃, a temperature measurement accuracy of ±0.5℃, and is suitable for acquiring transient heat conduction characteristics of thin plates of 0.1~1.0 mm. The synchronous control module is linked with the welding machine control system, triggering the infrared thermal imager to synchronously acquire thermal images during the laser welding process, ensuring the time synchronization of data and welding process. The data processing terminal has a built-in transient heat transfer inversion model and a lightweight neural network prediction model specifically for thin plates, used for data preprocessing, feature extraction, and defect judgment. The display module can display thermal images, predicted weld depth, and defect level judgment results in real time, and supports historical data traceability and quality statistics.

[0034] Figure 3 This is a system setup diagram of an embodiment of the present invention, including: a high-speed infrared thermal imager 1, a synchronous control module 2, a data processing terminal 3, a display module 4, a laser welding machine head 5, a thin plate workpiece to be welded 6, and a weld seam area 7. The infrared thermal imager 1 is installed on the side of the welding machine head 5 and aligned with the weld seam area 7 of the workpiece 6. The synchronous control module 2 is connected to the welding machine and the infrared thermal imager 1 respectively. The data processing terminal 3 receives the data from the infrared thermal imager 1 and outputs the results to the display module 4.

[0035] Step S202: Extract the region of interest from the original time-series infrared thermal image and perform temperature calibration on the region of interest to obtain a time-series temperature distribution image, wherein the time-series temperature distribution image is used to characterize the temperature value of each pixel. Step S203: Extract the thermal feature parameters of the time-series temperature distribution image, wherein the thermal feature parameters include temperature peak features, temperature gradient features, and thermal diffusion features; Step S204: Predict the back weld width of the weld region based on the thermal characteristic parameters; Step S205: Use the thermal characteristic parameters and the back weld width to identify welding defects in the weld area.

[0036] Through the above steps, when welding thin plate workpieces along the extension direction of the weld seam area, the original time-series infrared thermal image of the weld seam area is acquired by an infrared thermal imager. The infrared thermal imager is vertically positioned directly above the weld seam area. The original time-series infrared thermal image includes multiple original infrared thermal images arranged in chronological order. A region of interest (ROI) is extracted from the original time-series infrared thermal image, and the ROI is temperature-calibrated to obtain a time-series temperature distribution image, which characterizes the temperature value of each pixel. Thermal feature parameters of the time-series temperature distribution image are extracted, including temperature peak features, temperature gradient features, and thermal diffusion features. The back weld width of the weld seam area is predicted based on the thermal feature parameters. Welding defects in the weld seam area are identified using the thermal feature parameters and the back weld width. This realizes a method and device for identifying welding defects, a storage medium, and an electronic device, achieving a resolution of 0.1–1.0. The single-sided inspection, double-sided judgment, real-time online, high-precision non-destructive testing of ultra-thin sheet laser butt welding has solved the technical problem that existing technologies cannot detect welding defects in thin sheet workpieces, and has greatly improved the efficiency and accuracy of thin sheet welding quality inspection.

[0037] In one example, extracting the region of interest (ROI) from the original time-series infrared thermal image includes: denoising the original time-series infrared thermal image using Gaussian filtering to obtain an intermediate time-series infrared thermal image; denoising the intermediate time-series infrared thermal image using a median filtering algorithm to obtain a target time-series infrared thermal image; identifying the weld center region and the heat-affected zone surrounding the weld center region in the target time-series infrared thermal image, and outputting the weld center region and the heat-affected zone as the ROI.

[0038] Infrared thermal image acquisition and preprocessing require the thin plate workpieces to be welded to adopt butt joint assembly, ensuring that the butt joint gap is ≤0.05mm, the plate thickness ranges from 0.1 to 1.0 mm, and the material is stainless steel or aluminum alloy; start the laser welding machine to perform butt welding, and the synchronous control module triggers the high-speed infrared thermal imager to acquire time-series infrared thermal images of the weld center and the heat-affected zones on both sides at a frequency of 500 to 1000 frames / second, with no less than 100 thermal images acquired at each welding point.

[0039] This example employs a combination of Gaussian filtering and median filtering to denoise the acquired raw thermal image, eliminating image noise caused by welding spatter and ambient light. Through image segmentation algorithms, the weld center (width ≤ 2mm) and the heat-affected zone (width ≤ 5mm) are identified, and invalid background data is removed to extract the region of interest (ROI). In this embodiment, the weld center is the geometric center of the weld area, representing the axis region where the heat source is most concentrated and the temperature gradient is steepest. The heat-affected zone is the area where the material undergoes changes in metallographic structure or mechanical properties under the influence of heat during welding, but has not yet reached a molten state. The width ≤ 5mm is set based on engineering experience in thin-plate welding and represents a reasonable range for the typical heat-affected zone in thin-plate laser welding.

[0040] Furthermore, based on the built-in calibration parameters of the infrared thermal imager, the temperature of the preprocessed image is calibrated to obtain the true temperature distribution matrix of each pixel.

[0041] In one embodiment of this example, extracting the thermal feature parameters of the time-series temperature distribution image includes: extracting the weld center pixel in the time-series temperature distribution image, and determining the temperature peak value of the weld center pixel in multiple temperature distribution images; recording the occurrence time and duration of the temperature peak value in multiple temperature distribution images; and determining the temperature peak value, the occurrence time, and the duration as the temperature peak value feature.

[0042] Temperature peak feature: Extract the temperature peak of the center pixel of the weld. The timing of the temperature peak during welding Duration of temperature peak .

[0043] In one embodiment of this example, extracting the thermal feature parameters of the time-series temperature distribution image includes: for every two adjacent temperature distribution images in the time-series temperature distribution image, starting from the starting pixel of the initial temperature distribution image, selecting a first pixel and a second pixel along the welding direction, and a third pixel and a fourth pixel along the vertical direction of the welding direction, respectively; calculating a first temperature change value between the first pixel and the second pixel, calculating a horizontal length value between the first pixel and the second pixel, calculating a second temperature change value between the third pixel and the fourth pixel, and calculating a vertical length value between the third pixel and the fourth pixel; calculating a horizontal temperature gradient based on the ratio of the first temperature change value to the horizontal length value, and calculating a vertical temperature gradient based on the ratio of the second temperature change value to the vertical length value; and determining the horizontal temperature gradient and the vertical temperature gradient as the temperature gradient feature.

[0044] Calculate the temperature gradient along the welding direction at the weld center. Temperature gradient perpendicular to the welding direction To obtain the uniformity of the temperature field distribution, T: Represents the temperature change, referring to the temperature difference between two adjacent pixels, or the temperature change with distance within a small area. x: Represents the spatial distance increment (pixel pitch or actual physical distance) in the welding direction (usually horizontal / vertical). y: Represents the spatial distance increment in the direction perpendicular to the welding direction (usually the transverse / thickness direction). Gx (temperature gradient in the welding direction): Reflects the rate at which heat is transferred along the weld. If Gx changes drastically, it indicates uneven temperature distribution and the presence of hot spots. Gy (temperature gradient perpendicular to the welding direction): Reflects the ability of heat to conduct into the plate. Uniformity: If the temperature distribution is uniform, then the values ​​of Gx and Gy should remain relatively stable in the central region of the weld without abrupt changes.

[0045] In one embodiment of this example, extracting the thermal feature parameters of the time-series temperature distribution image includes: reading the property parameters of the welding material of the thin plate workpiece, wherein the property parameters include thermal conductivity, material density, and specific heat capacity at constant pressure; calculating the thermal diffusivity of the welding material using the property parameters; determining the heat source application time of the welding material in the time-series temperature distribution image; calculating the thermal diffusivity radius of the heat-affected zone using the thermal diffusivity and the heat source application time; and determining the thermal diffusivity and the thermal diffusivity radius as the thermal diffusivity feature.

[0046] Calculate the heat diffusion radius of the heat-affected zone based on the temperature distribution matrix of the thermal image. Thermal diffusivity This reflects the ability of heat to be conducted into the interior of the sheet material. By establishing a quantitative mapping relationship between thermal characteristics (such as Tmax, Gx, etc.) and the back weld width, an equivalent thermal diffusivity is derived from experimental data using machine learning or numerical fitting methods. This equivalent thermal diffusivity is used to characterize the heat transfer capacity of a specific sheet material and process.

[0047] Where λ is the thermal conductivity and ρ is the material density. Specific heat capacity at constant pressure.

[0048] Wherein, the thermal diffusion radius r: r=2 .

[0049] Optionally, predicting the back weld width of the weld region based on the thermal characteristic parameters includes calculating the back weld width of the weld region using the following formula. : ;in, The temperature peak value in the aforementioned temperature peak characteristic. The mean value of the horizontal temperature gradient in the temperature gradient feature is given. Here, a0, a1, a2, and a3 represent the thermal diffusivity coefficients in the aforementioned thermal diffusivity characteristics, and a0, a1, a2, and a3 are all pre-fitted regression coefficients. This is the residual term.

[0050] When welding ultrathin plates, the heat input from the front laser and the heat conduction from the back are linearly correlated, and the weld width on the back side... The core principle of quantitative mapping relationship with frontal thermal characteristic parameters is to establish a one-dimensional transient heat transfer inversion model adapted to 0.1–1.0 mm ultrathin plates.

[0051] By combining the boundary conditions of thin plates (back side insulation, front side thermal convection), the mapping relationship between thermal characteristic parameters and back side weld width under different plate thicknesses, materials, and welding parameters is fitted through finite element simulation, thus constructing... The multiple regression model, ; An inversion model was pre-constructed by setting multiple sets of temperature peaks, temperature gradients, thermal diffusion characteristic parameters, and the actual back-side melt width. The model was then solved iteratively using the least squares method to obtain a0, a1, a2, a3, ... .

[0052] In one example of this embodiment, identifying welding defects in the weld region using the thermal feature parameters and the back weld width includes: constructing input data using the thermal feature parameters and the back weld width; inputting the input data into a pre-trained convolutional neural network model, and outputting the welding defects in the weld region, wherein the convolutional neural network model is a welding defect prediction model trained based on a sample dataset.

[0053] This example demonstrates how to achieve accurate defect classification by pre-constructing a lightweight CNN neural network model adapted to thin plate thermal feature data. This model requires extensive training with numerous samples. Thin plate welding samples under different welding parameters (laser power, welding speed, defocusing amount) are collected, and corresponding frontal thermal feature parameters and the actual penetration depth and defect type obtained from destructive detection are acquired to construct a training dataset. This dataset is then input into the lightweight CNN model for training, optimizing the model weights to achieve a classification accuracy of ≥98% for incomplete penetration, normal weld penetration, and burn-through defects. Finally, the real-time predicted back weld width and extracted thermal feature parameters are input into the trained lightweight CNN model, which outputs three defect level classifications: normal, incomplete penetration, and burn-through.

[0054] In another example of this embodiment, identifying welding defects in the weld region using the thermal characteristic parameters and the back weld width includes: Scenario 1: Based on the temperature gradient characteristics, determine whether there is an abnormal abrupt change in the temperature gradient, and determine whether the back weld width is less than the first preset thickness; if there is no abnormal abrupt change in the temperature gradient, and the back weld width is greater than or equal to the first preset thickness, determine that there is no welding defect in the weld area; Scenario 2: Based on the temperature gradient characteristics, determine whether there is an abnormal abrupt change in the temperature gradient, and determine whether the back weld width is less than the first preset thickness; if there is an abnormal abrupt change in the temperature gradient, and the back weld width is less than the first preset thickness, determine whether the back weld width is greater than or equal to the second preset thickness; if the back weld width is greater than or equal to the second preset thickness, determine that the welding defect is an incomplete penetration defect; Case 3: Determine whether the back weld width is greater than the third preset thickness, and determine whether the welding temperature peak exceeds the melting point threshold of the thin plate workpiece based on the temperature peak characteristics, wherein the third preset thickness > the thickness of the thin plate workpiece > the first preset thickness > the second preset thickness; if the back weld width is greater than the third preset thickness and the welding temperature peak exceeds the melting point threshold of the thin plate workpiece, determine that the welding defect is a burn-through defect.

[0055] Optionally, the third preset thickness is 1.2 × plate thickness, and the thickness of the thin plate workpiece is 0.8 × plate thickness and 0.5 × plate thickness.

[0056] Normal: Backside weld width ≥ 0.8 × plate thickness, and without abnormal temperature changes; Incomplete penetration: 0.5 × plate thickness ≤ back weld width < 0.8 × plate thickness, and the temperature gradient is abnormally increased; Burn-through: Backside melt width > 1.2 × plate thickness, and the peak temperature exceeds the material's melting point threshold (660℃ for aluminum alloy and 1500℃ for stainless steel).

[0057] Figure 4 These are infrared thermal images comparing different defect types in the embodiments of the present invention: (a) Thermal image of normal weld: uniform temperature distribution with no obvious temperature abrupt change; (b) Thermal image of incomplete weld: the temperature gradient at the center of the weld is significantly increased and the heat diffusion radius is reduced; (c) Thermal image of burn-through weld: the temperature peak exceeds the melting point of the material and the heat diffusion radius is abnormally expanded.

[0058] Optionally, the back weld width prediction value and defect level judgment result can be transmitted to the display module in real time through the data processing terminal, so that operators can intuitively obtain welding quality information. At the same time, the system automatically counts the defect rate and weld depth distribution of batch welded workpieces and generates a quality statistical report to provide data support for welding process optimization.

[0059] Figure 5 This is a flowchart of the defect detection method in this embodiment of the invention. First, the system is built, then thermal image acquisition and preprocessing are performed, then thermal feature extraction is performed, back-side melt width inversion is performed, defect level is determined, and finally real-time display and statistics are performed. The arrows indicate the execution order of each step.

[0060] This embodiment aims to address the shortcomings of existing thin-plate laser welding inspection technologies, such as insufficient adaptability, difficulty in identifying latent defects, and low inspection efficiency. It provides a method for identifying welding defects, enabling single-sided inspection, double-sided judgment, real-time online operation, and high-precision non-destructive testing of laser butt welds in ultra-thin plates ranging from 0.1 to 1.0 mm. This significantly improves the efficiency and accuracy of thin-plate welding quality inspection. The following provides complete examples illustrating normal, incomplete penetration, and burn-through defects: Implementation Scenario 1: Laser Butt Weld Inspection of 0.5 mm Aluminum Alloy Sheets Test object: 0.5 mm thick 6061 aluminum alloy sheet, with butt joint assembly, butt joint gap of 0.03 mm, laser welding parameters: laser power of 1200 W, welding speed of 20 mm / s, defocusing amount of 0 mm; System setup: A high-speed infrared thermal imager with a frame rate of 800 frames per second is used and installed on the side of the welding machine head with the lens aimed at the weld area. The synchronous control module is linked with the welding machine control system to ensure synchronous data acquisition. Thermal image acquisition and preprocessing: Welding machine is started and welding is carried out. Time-series infrared thermal images of 800 frames / second are acquired simultaneously. After Gaussian filtering + median filtering for noise reduction, ROI region extraction (1.5 mm at the weld center and 4 mm in the heat-affected zone), and temperature calibration, an effective temperature matrix is ​​obtained. Thermal feature extraction: Extracting the highest temperature at the weld center =420℃, time of peak temperature occurrence =0.02s, duration of highest temperature =0.01s, temperature gradient along the welding direction = 80℃ / mm, temperature gradient perpendicular to the welding direction =60 ℃ / mm, thermal diffusion radius =3.2 mm, thermal diffusivity =8.5×10⁻⁵ m / s; Backside weld width inversion: Input the above thermal characteristic parameters into the thin-plate-specific transient heat transfer inversion model to obtain the backside weld width. =0.42 mm Defect determination: The back weld width and thermal feature parameters were input into the lightweight CNN model, and the model determined the result to be "normal". Verification: Metallographic sectioning revealed the actual back-side weld width. =0.41 mm, defect type is normal, detection error is 0.01 mm, accuracy is 100%.

[0061] Scenario 2: Detection of incomplete penetration in laser butt welding of 0.3 mm stainless steel sheets Test object: 0.3 mm thick 304 stainless steel sheet, butt joint assembly, butt joint gap 0.02 mm, laser welding parameters: laser power 800W, welding speed 15mm / s, defocusing amount -0.5 mm; System setup: A high-speed infrared thermal imager with a frame rate of 1000 frames per second is used, and the rest is the same as in implementation scenario 1; Thermal image acquisition and preprocessing: Same as in Example 1, to acquire an effective temperature matrix; Thermal feature extraction: Extracting the highest temperature at the weld center =380℃, temperature gradient along the welding direction =150℃ / mm, thermal diffusion radius =2.5 mm, thermal diffusivity =7.2×10⁻⁵ m² / s; Backside weld width inversion: Input thermal characteristic parameters into the inversion model to obtain the backside weld width. =0.14 mm; Defect determination: The model's determination result is "incomplete penetration"; Verification: Metallographic section analysis shows the actual back-side weld width =0.13 mm, indicating a clear incomplete weld penetration defect. The detection error was 0.01 mm, and the judgment was accurate.

[0062] Implementation Scenario 3: Laser Butt Burn-through Detection of 1.0 mm Aluminum Alloy Sheets Test object: 1.0 mm thick 6061 aluminum alloy sheet, butt joint assembly, butt joint gap 0.04 mm, laser welding parameters: laser power 1800 W, welding speed 15 mm / s, defocusing amount 0 mm; System setup: Same as implementation scenario 1; Thermal image acquisition and preprocessing: Same as in Example 1; Thermal feature extraction: Extracting the highest temperature at the weld center ==680℃ (exceeding the melting point of aluminum alloy 660℃), temperature gradient along the vertical welding direction =200℃ / mm, thermal diffusion radius =5.8 mm; Backside weld width inversion: Input thermal characteristic parameters into the inversion model to obtain the backside weld width. =1.35 mm; Defect determination: The model determines the result as "burn-through"; Verification: A clear weld-through hole is visible on the back of the workpiece; the actual weld width on the back is... =1.32 mm, detection error 0.03 mm, judgment accurate.

[0063] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0064] Example 2 This embodiment also provides a welding defect identification device, which is used to implement the above embodiments and preferred embodiments, and will not be repeated as already described. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0065] Figure 6 This is a structural block diagram of a welding defect identification device according to an embodiment of the present invention, such as... Figure 6 As shown, the device includes: The acquisition module 61 is used to acquire the original time-series infrared thermal image of the weld area collected by the infrared thermal imager when welding a thin plate workpiece along the extension direction of the weld area. The infrared thermal imager is vertically arranged directly above the weld area, and the original time-series infrared thermal image includes multiple original infrared thermal images arranged in time sequence. The first extraction module 62 is used to extract the region of interest in the original time-series infrared thermal image and to perform temperature calibration on the region of interest to obtain a time-series temperature distribution image, wherein the time-series temperature distribution image is used to characterize the temperature value of each pixel. The second extraction module 63 is used to extract the thermal feature parameters of the time-series temperature distribution image, wherein the thermal feature parameters include temperature peak features, temperature gradient features, and thermal diffusion features. Prediction module 64 is used to predict the back weld width of the weld region based on the thermal characteristic parameters; The identification module 65 is used to identify welding defects in the weld area using the thermal characteristic parameters and the back weld width.

[0066] Optionally, the first extraction module includes: a first denoising unit, used to denoise the original time-series infrared thermal image using Gaussian filtering to obtain an intermediate time-series infrared thermal image; a second denoising unit, used to denoise the intermediate time-series infrared thermal image using a median filtering algorithm to obtain a target time-series infrared thermal image; and an identification unit, used to identify the weld center region and the heat-affected zone around the weld center region in the target time-series infrared thermal image, and output the weld center region and the heat-affected zone as regions of interest.

[0067] Optionally, the second extraction module includes: an extraction unit, configured to extract the weld center pixel in the time-series temperature distribution image and determine the temperature peak value of the weld center pixel in multiple temperature distribution images; a recording unit, configured to record the occurrence time and duration of the temperature peak value in multiple temperature distribution images; and a determination unit, configured to determine the temperature peak value, the occurrence time, and the duration as the temperature peak value feature.

[0068] Optionally, the second extraction module includes: a selection unit, configured to, for each pair of adjacent temperature distribution images in the time-series temperature distribution image, starting from the starting pixel of the initial temperature distribution image, select a first pixel and a second pixel from the two adjacent temperature distribution images along the welding direction, and select a third pixel and a fourth pixel along the vertical direction of the welding direction; a first calculation unit, configured to calculate a first temperature change value between the first pixel and the second pixel, calculate a horizontal length value between the first pixel and the second pixel, calculate a second temperature change value between the third pixel and the fourth pixel, and calculate a vertical length value between the third pixel and the fourth pixel; a second calculation unit, configured to calculate a horizontal temperature gradient based on the ratio of the first temperature change value to the horizontal length value, and calculate a vertical temperature gradient based on the ratio of the second temperature change value to the vertical length value; and a first determination unit, configured to determine the horizontal temperature gradient and the vertical temperature gradient as the temperature gradient features.

[0069] Optionally, the second extraction module includes: a reading unit for reading the property parameters of the welding material of the thin plate workpiece, wherein the property parameters include thermal conductivity, material density, and specific heat capacity at constant pressure; a third calculation unit for calculating the thermal diffusivity of the welding material using the property parameters; a second determination unit for determining the heat source action time of the welding material in the time-series temperature distribution image; a fourth calculation unit for calculating the thermal diffusivity radius of the heat-affected zone using the thermal diffusivity and the heat source action time; and a third determination unit for determining the thermal diffusivity and the thermal diffusivity radius as the thermal diffusivity feature.

[0070] Optionally, the prediction module includes a calculation unit for calculating the back weld width of the weld region using the following formula. : ;in, The temperature peak value in the aforementioned temperature peak characteristic. The mean value of the horizontal temperature gradient in the temperature gradient feature is given. Here, a0, a1, a2, and a3 represent the thermal diffusivity coefficients in the aforementioned thermal diffusivity characteristics, and a0, a1, a2, and a3 are all pre-fitted regression coefficients. This is the residual term.

[0071] Optionally, the identification module includes: a construction unit for constructing input data from the thermal feature parameters and the back weld width; and an output unit for inputting the input data into a pre-trained convolutional neural network model and outputting welding defects in the weld area, wherein the convolutional neural network model is a welding defect prediction model trained based on a sample dataset.

[0072] Optionally, the identification module includes: a first identification unit, configured to determine whether there is an abnormal abrupt change in the temperature gradient based on the temperature gradient characteristics, and to determine whether the back weld width is less than a first preset thickness; if there is no abnormal abrupt change in the temperature gradient, and the back weld width is greater than or equal to the first preset thickness, determine that there is no welding defect in the weld area; a second identification unit, configured to determine whether there is an abnormal abrupt change in the temperature gradient based on the temperature gradient characteristics, and to determine whether the back weld width is less than the first preset thickness; if there is an abnormal abrupt change in the temperature gradient, and the back weld width is less than the first preset thickness, determine whether the back weld width is greater than or equal to a second preset thickness; if the back weld width is greater than or equal to the second preset thickness, determine that the welding defect is an incomplete penetration defect; a third identification unit, configured to determine whether the back weld width is greater than a third preset thickness, and to determine whether the welding temperature peak exceeds the material melting point threshold of the thin plate workpiece based on the temperature peak characteristics, wherein the third preset thickness > the thickness of the thin plate workpiece > the first preset thickness > the second preset thickness; if the back weld width is greater than the third preset thickness, and the welding temperature peak exceeds the material melting point threshold of the thin plate workpiece, determine that the welding defect is a burn-through defect.

[0073] It should be noted that the above modules can be implemented by software or hardware. For the latter, they can be implemented in the following ways, but are not limited to: all the above modules are located in the same processor; or, the above modules are located in different processors in any combination.

[0074] Example 3 Embodiments of the present invention also provide a storage medium storing a computer program, wherein the computer program is configured to execute the steps in any of the above method embodiments when running.

[0075] Optionally, in this embodiment, the storage medium may be configured to store a computer program for performing the following steps: S1, when welding a thin plate workpiece along the extension direction of the weld area, the original time-series infrared thermal image of the weld area is acquired by an infrared thermal imager, wherein the infrared thermal imager is vertically positioned directly above the weld area, and the original time-series infrared thermal image includes multiple original infrared thermal images arranged in time sequence. S2, extract the region of interest from the original time-series infrared thermal image, and perform temperature calibration on the region of interest to obtain a time-series temperature distribution image, wherein the time-series temperature distribution image is used to characterize the temperature value of each pixel; S3, extract the thermal feature parameters of the time-series temperature distribution image, wherein the thermal feature parameters include temperature peak features, temperature gradient features, and thermal diffusion features; S4, predict the back weld width of the weld region based on the thermal characteristic parameters; S5, using the thermal characteristic parameters and the back weld width to identify welding defects in the weld area.

[0076] Optionally, in this embodiment, the storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0077] Embodiments of the present invention also provide an electronic device including a memory and a processor, the memory storing a computer program and the processor being configured to run the computer program to perform the steps in any of the above method embodiments.

[0078] Optionally, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor and the input / output device is connected to the processor.

[0079] Optionally, in this embodiment, the processor can be configured to perform the following steps via a computer program: S1, when welding a thin plate workpiece along the extension direction of the weld area, the original time-series infrared thermal image of the weld area is acquired by an infrared thermal imager, wherein the infrared thermal imager is vertically positioned directly above the weld area, and the original time-series infrared thermal image includes multiple original infrared thermal images arranged in time sequence. S2, extract the region of interest from the original time-series infrared thermal image, and perform temperature calibration on the region of interest to obtain a time-series temperature distribution image, wherein the time-series temperature distribution image is used to characterize the temperature value of each pixel; S3, extract the thermal feature parameters of the time-series temperature distribution image, wherein the thermal feature parameters include temperature peak features, temperature gradient features, and thermal diffusion features; S4, predict the back weld width of the weld region based on the thermal characteristic parameters; S5, using the thermal characteristic parameters and the back weld width to identify welding defects in the weld area.

[0080] Optionally, specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementations, and will not be repeated here.

[0081] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0082] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0083] It should be understood that the terminology used herein is for the purpose of describing particular exemplary embodiments only and is not intended to be limiting. Unless the context clearly indicates otherwise, the singular forms “a,” “an,” and “described” as used herein may also include the plural forms. The terms “comprising,” “including,” “containing,” and “having” are inclusive and therefore indicate the presence of the stated features, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, elements, components, and / or combinations thereof. The method steps, processes, and operations described herein are not construed as requiring them to be performed in a particular order described or illustrated unless the order of performance is explicitly indicated. It should also be understood that additional or alternative steps may be used.

[0084] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method for identifying welding defects, characterized in that, include: When welding a thin plate workpiece along the extension direction of the weld seam area, the original time-series infrared thermal image of the weld seam area is acquired by an infrared thermal imager, wherein the infrared thermal imager is vertically positioned directly above the weld seam area, and the original time-series infrared thermal image includes multiple original infrared thermal images arranged in time sequence. Extract the region of interest from the original time-series infrared thermal image and perform temperature calibration on the region of interest to obtain a time-series temperature distribution image, wherein the time-series temperature distribution image is used to characterize the temperature value of each pixel. Extract the thermal feature parameters of the time-series temperature distribution image, wherein the thermal feature parameters include temperature peak features, temperature gradient features, and thermal diffusion features; Predict the back weld width of the weld region based on the thermal characteristic parameters; Welding defects in the weld area are identified using the thermal characteristic parameters and the back weld width.

2. The method according to claim 1, characterized in that, Extracting the region of interest from the original time-series infrared thermal image includes: Gaussian filtering is used to denoise the original time-series infrared thermal image to obtain an intermediate time-series infrared thermal image; The intermediate time-series infrared thermal image is denoised using a median filtering algorithm to obtain the target time-series infrared thermal image. Identify the weld center region and the heat-affected zone surrounding the weld center region in the target time-series infrared thermal image, and output the weld center region and the heat-affected zone as regions of interest.

3. The method according to claim 1, characterized in that, Extracting the thermal feature parameters of the time-series temperature distribution image includes: Extract the weld center pixel from the time-series temperature distribution image, and determine the temperature peak value of the weld center pixel in multiple temperature distribution images; Record the occurrence time and duration of the temperature peak in multiple temperature distribution images; The temperature peak value, the occurrence time, and the duration are defined as the temperature peak value characteristics.

4. The method according to claim 1, characterized in that, Extracting the thermal feature parameters of the time-series temperature distribution image includes: For each two adjacent temperature distribution images in the time-series temperature distribution image, starting from the starting pixel of the initial temperature distribution image, the first pixel and the second pixel are selected from the two adjacent temperature distribution images along the welding direction, and the third pixel and the fourth pixel are selected from the vertical direction along the welding direction. Calculate the first temperature change value between the first pixel and the second pixel, calculate the horizontal length value between the first pixel and the second pixel, calculate the second temperature change value between the third pixel and the fourth pixel, and calculate the vertical length value between the third pixel and the fourth pixel; The horizontal temperature gradient is calculated based on the ratio of the first temperature change value to the horizontal length value, and the vertical temperature gradient is calculated based on the ratio of the second temperature change value to the vertical length value. The horizontal temperature gradient and the vertical temperature gradient are defined as the temperature gradient features.

5. The method according to claim 1, characterized in that, Extracting the thermal feature parameters of the time-series temperature distribution image includes: Read the property parameters of the welding material of the thin plate workpiece, wherein the property parameters include thermal conductivity, material density, and specific heat capacity at constant pressure; The thermal diffusivity of the welding material is calculated using the aforementioned property parameters; Determine the heat source application time of the welding material in the time-series temperature distribution image; The thermal diffusion radius of the heat-affected zone is calculated using the thermal diffusivity and the duration of heat source action. The thermal diffusivity and the thermal diffusivity radius are defined as the thermal diffusivity characteristics.

6. The method according to claim 1, characterized in that, Predicting the back weld width of the weld region based on the thermal characteristic parameters includes: The back weld width of the weld area is calculated using the following formula. : ； in, The temperature peak value in the aforementioned temperature peak characteristic. The mean value of the horizontal temperature gradient in the temperature gradient feature is given. Here, a0, a1, a2, and a3 represent the thermal diffusivity coefficients in the aforementioned thermal diffusivity characteristics, and a0, a1, a2, and a3 are all pre-fitted regression coefficients. This is the residual term.

7. The method according to claim 1, characterized in that, Identifying welding defects in the weld region using the aforementioned thermal characteristic parameters and the back weld width includes: The thermal characteristic parameters and the back melt width are used to construct the input data; The input data is fed into a pre-trained convolutional neural network model, which outputs the welding defects in the weld area. The convolutional neural network model is a welding defect prediction model trained based on a sample dataset.

8. The method according to claim 1, characterized in that, Identifying welding defects in the weld region using the aforementioned thermal characteristic parameters and the back weld width includes: Based on the temperature gradient characteristics, it is determined whether there is an abnormal abrupt change in the temperature gradient, and whether the back weld width is less than the first preset thickness; if there is no abnormal abrupt change in the temperature gradient, and the back weld width is greater than or equal to the first preset thickness, it is determined that there is no welding defect in the weld area. Based on the temperature gradient characteristics, determine whether there is an abnormal abrupt change in the temperature gradient, and determine whether the back weld width is less than the first preset thickness; if there is an abnormal abrupt change in the temperature gradient, and the back weld width is less than the first preset thickness, determine whether the back weld width is greater than or equal to the second preset thickness; if the back weld width is greater than or equal to the second preset thickness, determine that the welding defect is an incomplete penetration defect. Determine whether the back weld width is greater than the third preset thickness, and determine whether the welding temperature peak exceeds the material melting point threshold of the thin plate workpiece based on the temperature peak characteristics, wherein the third preset thickness > the thickness of the thin plate workpiece > the first preset thickness > the second preset thickness; if the back weld width is greater than the third preset thickness and the welding temperature peak exceeds the material melting point threshold of the thin plate workpiece, determine that the welding defect is a burn-through defect.

9. A welding defect identification device, characterized in that, include: The acquisition module is used to acquire the original time-series infrared thermal image of the weld area collected by the infrared thermal imager when welding a thin plate workpiece along the extension direction of the weld area. The infrared thermal imager is vertically arranged directly above the weld area, and the original time-series infrared thermal image includes multiple original infrared thermal images arranged in time sequence. The first extraction module is used to extract the region of interest from the original time-series infrared thermal image and to perform temperature calibration on the region of interest to obtain a time-series temperature distribution image, wherein the time-series temperature distribution image is used to characterize the temperature value of each pixel. The second extraction module is used to extract the thermal feature parameters of the time-series temperature distribution image, wherein the thermal feature parameters include temperature peak features, temperature gradient features, and thermal diffusion features. The prediction module is used to predict the back weld width of the weld region based on the thermal characteristic parameters. An identification module is used to identify welding defects in the weld area using the thermal characteristic parameters and the back weld width.

10. The apparatus according to claim 9, characterized in that, The prediction module includes: The calculation unit is used to calculate the back weld width of the weld region using the following formula. : ;in, The temperature peak value in the aforementioned temperature peak characteristic. The mean value of the horizontal temperature gradient in the temperature gradient feature is given. Here, a0, a1, a2, and a3 represent the thermal diffusivity coefficients in the aforementioned thermal diffusivity characteristics, and a0, a1, a2, and a3 are all pre-fitted regression coefficients. This is the residual term.

11. A storage medium, characterized in that, The storage medium stores a computer program, wherein the computer program is configured to execute the method described in any one of claims 1 to 8 when it is run.

12. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to run the computer program to perform the method as described in any one of claims 1 to 8.

Images