Weld defect detection method and device, computer equipment and storage medium

Abstract

The invention discloses a weld defect detection method and device, computer equipment and a storage medium, and the method comprises the steps: obtaining a weld image of a welded component; respectively inputting the welding seam image into a pre-trained welding seam identification model and a morphological identification model for prediction to obtain welding seam size information and welding seam shape information; based on the welding seam size information, whether the welding seam size of the welding component is abnormal is determined; based on the weld form information, whether the weld form of the welded component has flaws is determined; and if the welding seam size of the welding component is abnormal and / or the welding seam shape has flaws, warning information is output. The welding seam size and the welding seam form are respectively predicted by collecting the welding seam image of the welding component, and whether the welding seam has size abnormity and / or form flaws is respectively determined based on the welding seam size and the welding seam form, so that the welding seam defect condition does not need to be manually detected, a large amount of labor cost can be saved, and the manual false detection rate is reduced; and the detection efficiency and accuracy are improved.

Description

Technical Field

[0001] The present application relates to the field of welding technology, and in particular to a weld defect detection method, device, computer equipment and storage medium. Background Art

[0002] Traditional welding defect detection methods mainly rely on manual inspection and evaluation of welding images. However, the naked eye is easily affected by external factors, resulting in a high false detection rate. Long-term inspection is prone to fatigue, resulting in a decrease in accuracy. It also requires a lot of manpower and time and inevitably brings many subjective errors.

[0003] With the development of image processing technology, the automatic detection technology of weld defects has been widely studied and has made great progress. However, the current automatic detection technology of weld defects only performs simple analysis on the weld X-ray detection image at a two-dimensional level, and can only search for the approximate location of the weld defect, but it is impossible to know what kind of defects exist in the weld, resulting in the inability to effectively solve the weld defects. Summary of the invention

[0004] Based on this, it is necessary to provide a weld defect detection method, device, computer equipment and storage medium to address the above technical issues, so as to solve at least one problem existing in the above-mentioned prior art.

[0005] In a first aspect, a weld defect detection method is provided, comprising:

[0006] Acquire weld images of welded components;

[0007] The weld image is respectively input into a pre-trained weld recognition model and a morphological recognition model for prediction to obtain weld size information and weld morphology information;

[0008] Based on the weld size information, determining whether there is an abnormality in the weld size of the welded component;

[0009] Based on the weld morphology information, determining whether there is a defect in the weld morphology of the welded component;

[0010] If there is an abnormality in the weld size of the welded component and / or there is a defect in the weld shape, an alarm message is output.

[0011] In one embodiment, the weld recognition model is trained in the following manner:

[0012] Acquire a weld sample data set, wherein the weld sample data set includes a plurality of weld sample images with labels;

[0013] Constructing an initial weld identification model;

[0014] The weld sample image is input into the initial weld recognition model for iterative training until a first preset convergence condition is met, thereby obtaining the trained weld recognition model.

[0015] In one embodiment, the morphological recognition model is trained in the following manner:

[0016] Acquire a surface defect sample data set, wherein the surface defect sample data set includes a plurality of surface defect sample images with labels;

[0017] Constructing an initial morphological recognition model;

[0018] The surface defect sample image is input into the initial morphological recognition model for iterative training until a second preset convergence condition is met, thereby obtaining the trained morphological recognition model.

[0019] In one embodiment, determining whether there is an abnormality in the weld size of the welded component based on the weld size information includes:

[0020] Constructing a standard weld size data set, wherein the standard weld size data set includes a plurality of standard weld size information;

[0021] Comparing the weld size information with the corresponding standard weld size information to determine the difference between the weld size information and the standard weld size information;

[0022] When the difference exceeds a preset error range, the weld size of the welded component is abnormal.

[0023] In one embodiment, after determining whether the weld morphology of the welded component has defects, the method includes:

[0024] If the weld size of the welded component is abnormal, and / or the weld shape is defective, obtaining the welding angle and welding melting point temperature of the current welding gun;

[0025] Determine the welding angle range and the welding melting point temperature range corresponding to the welding component;

[0026] Determining whether the welding angle exceeds the welding angle range, and / or determining whether the welding melting point temperature exceeds the welding melting point temperature range;

[0027] If so, the welding angle is adjusted based on the welding angle range, and / or the welding melting point temperature is adjusted based on the welding melting point temperature range.

[0028] In one embodiment, after determining whether the weld morphology of the welded component has defects, the method includes:

[0029] If there is no abnormality in the weld size of the welded component and there is no defect in the weld shape, detecting whether there is a defect inside the weld of the welded component;

[0030] If there are defects inside the weld of the welded component, an alarm message is output.

[0031] In one embodiment, obtaining a weld image of a welded component includes:

[0032] Acquire a welding component image, wherein the welding component image includes a weld area and a non-weld area;

[0033] Eliminating the non-weld area to obtain a weld area image;

[0034] The weld area image is subjected to image noise reduction processing and image enhancement processing to obtain the preprocessed weld image.

[0035] In a second aspect, a weld defect detection device is provided, comprising:

[0036] A weld image acquisition unit, used to acquire a weld image of a welded component;

[0037] A prediction unit, used for inputting the weld image into a pre-trained weld recognition model and a morphological recognition model for prediction, to obtain weld size information and weld morphology information;

[0038] A first judgment unit, configured to determine whether a weld size of the welding component is abnormal based on the weld size information;

[0039] A second judgment unit is used to determine whether there is a defect in the weld morphology of the welded component based on the weld morphology information;

[0040] The alarm information output unit is used to output alarm information if there is an abnormality in the weld size of the welded component and / or there is a defect in the weld shape.

[0041] In a third aspect, a computer device is provided, comprising a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor, wherein the processor implements the steps of the weld defect detection method as described above when executing the computer-readable instructions.

[0042] In a fourth aspect, a readable storage medium is provided, wherein the readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by a processor, the steps of the weld defect detection method as described above are implemented.

[0043] The above-mentioned weld defect detection method, device, computer equipment and storage medium, the method of which is implemented, includes: obtaining a weld image of a welding component; inputting the weld image into a pre-trained weld recognition model and a morphological recognition model for prediction, and obtaining weld size information and weld morphology information; based on the weld size information, determining whether the weld size of the welding component is abnormal; based on the weld morphology information, determining whether the weld morphology of the welding component is defective; if the weld size of the welding component is abnormal, and / or the weld morphology is defective, outputting an alarm message. In the embodiment of the present application, by collecting the weld image of the welding component, the weld size and weld morphology are predicted respectively, and based on the weld size and weld morphology, it is determined whether it has a size abnormality or a morphological defect. If so, an alarm message can be output to stop the machine in time and notify manual processing to avoid causing a large number of welding products with weld defects, and no manual inspection is required, which can save labor costs, reduce manual false detection rate, and improve detection efficiency and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments of the present application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0045] Figure 1 This is a schematic diagram of a process of a weld defect detection method in one embodiment of the present application;

[0046] Figure 2 This is a schematic diagram of the structure of a weld defect detection device in one embodiment of the present application;

[0047] Figure 3 It is a schematic diagram of a computer device in one embodiment of the present application. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0049] In one embodiment, if Figure 1 As shown, a weld defect detection method is provided, comprising the following steps:

[0050] In step S110, a weld image of a welded component is acquired;

[0051] In the embodiment of the present application, the weld image of the welded component can be collected in real time, periodically or regularly by a pre-set image acquisition device. The image acquisition device can be a camera, a video camera or other device that can be used to collect images. It should be noted that in order to make the collected image clear and reduce the influence of burrs, a suitable lens and fill light equipment can be used to improve the accuracy of image prediction.

[0052] It can be understood that when an image of a welded component is collected, it may include non-weld areas as well as weld areas, so the non-weld areas may be removed to retain the weld areas for subsequent prediction.

[0053] In step S120, the weld image is respectively input into a pre-trained weld recognition model and a morphological recognition model for prediction to obtain weld size information and weld morphology information;

[0054] In the embodiment of the present application, the weld recognition model can be a BP neural network, a deep learning model, etc., and the morphological recognition model can be a convolutional neural network, a feedforward neural network, etc. After the weld image is acquired, the weld image can be preprocessed, and the weld image can be subjected to image denoising and image enhancement to obtain a weld image with higher definition. Then, it is input into the weld recognition model and the morphological recognition model for prediction processing, so that the weld size information and weld morphology information can be obtained.

[0055] It should be noted that the weld recognition model can output multiple weld size features, such as weld width, weld height, weld length, weld foot and other size information. The morphological recognition model can output the morphology of the weld, such as whether there are pores, burrs, deformation, symmetry, unwelded cracks and other morphological information.

[0056] In step S130, based on the weld size information, it is determined whether there is an abnormality in the weld size of the welded component;

[0057] In the embodiment of the present application, a standard weld size data set of a welded component can be constructed. When the weld size information is obtained, it can be compared with the standard weld size stored in the standard weld size data set, for example, the height of the weld is greater than or equal to 0 and less than 2 mm. The weld size information is sequentially compared with the corresponding standard weld size to determine whether it is consistent with the standard weld size. If not, it is determined whether the error is within a preset range. If not, it can be determined that the weld size is abnormal.

[0058] In step S140, based on the weld morphology information, it is determined whether there is a defect in the weld morphology of the welded component;

[0059] In the embodiment of the present application, based on the morphological recognition model, the weld morphology information can be output. For example, if the weld morphology is basically symmetrical and there is a slight inverted cone in the weld, it can be considered that there is a defect. Furthermore, the degree of the defect can be determined. For example, if the weld morphology is asymmetrical and there are unwelded cracks, the defect can be considered to be more serious. If the weld morphology is basically symmetrical but there are slight burrs, the defect can be considered to be lighter. Depending on the degree of the defect, different processing methods can be selected. For example, for more serious defects, the machine needs to be shut down and manual processing needs to be notified. For minor defects, the welding gun, temperature, etc. can be adjusted and the welding operation can continue.

[0060] In step S150, if there is an abnormality in the weld size of the welded component and / or there is a defect in the weld shape, an alarm message is output.

[0061] In an embodiment of the present application, if it is determined that there is an abnormality in the weld size of the welded component, or there are defects in the weld shape, or there are both abnormal weld size and defects in the weld shape, an alarm message can be output, for example, through an audible and visual alarm or by text message, email, etc. to notify the staff so that corresponding processing can be carried out in time to avoid continuing welding according to the current welding parameters, which will lead to a lower welding qualification rate.

[0062] It is understandable that if there are major defects, such as serious flaws in the weld shape and a large difference between the weld size and the standard size, the machine can be stopped in time, that is, the welding operation can be stopped, the fault can be marked, and manual processing can be notified to reduce losses. The weld size, weld shape, type, material, welding gun position, angle, etc. of the current weld are obtained and stored so that they can be directly called when the welder is adjusted later.

[0063] In an embodiment of the present application, a weld defect detection method is provided, comprising: obtaining a weld image of a welding component; inputting the weld image into a pre-trained weld recognition model and a morphological recognition model for prediction, and obtaining weld size information and weld morphology information; determining whether the weld size of the welding component is abnormal based on the weld size information; determining whether the weld morphology of the welding component is defective based on the weld morphology information; if the weld size of the welding component is abnormal, and / or the weld morphology is defective, outputting an alarm message. In an embodiment of the present application, by collecting the weld image of the welding component, the weld size and weld morphology are predicted respectively, and based on the weld size and weld morphology, it is determined whether it has a size abnormality or a morphological defect. If so, an alarm message can be output so as to stop the machine in time and notify manual processing to avoid causing a large number of welding products with weld defects, and no manual inspection is required, which can save labor costs, reduce the manual false detection rate, and improve the detection efficiency and accuracy.

[0064] In one embodiment of the present application, the weld recognition model is trained in the following manner:

[0065] Acquire a weld sample data set, wherein the weld sample data set includes a plurality of weld sample images with labels;

[0066] Constructing an initial weld identification model;

[0067] The weld sample image is input into the initial weld recognition model for iterative training until a first preset convergence condition is met, thereby obtaining the trained weld recognition model.

[0068] Specifically, an initial weld recognition model can be constructed, and weld images can be collected through multi-source data sources to construct a sample data set. The weld sizes in the weld sample images in the sample data set can be labeled through manual annotation, and then the initial weld recognition model can be iteratively trained based on the labeled weld sample images. After each training, the loss value can be calculated once through a preset loss function, such as a cross entropy loss function. If the loss value is greater than a preset threshold, a weld sample image is selected for prediction again until the preset convergence conditions are met, such as the number of iterations reaches the preset threshold, or the loss value is less than the preset threshold, then a trained weld recognition model can be obtained.

[0069] Among them, the weld recognition model can be a BP neural network model, a convolutional neural network model, etc.

[0070] It should be noted that the weld recognition model can be evaluated and optimized through the test set, and during the use of the weld recognition model, its network parameters can also be continuously updated to improve its prediction accuracy.

[0071] In one embodiment of the present application, the morphological recognition model is trained in the following manner:

[0072] Acquire a surface defect sample data set, wherein the surface defect sample data set includes a plurality of surface defect sample images with labels;

[0073] Constructing an initial morphological recognition model;

[0074] The surface defect sample image is input into the initial morphological recognition model for iterative training until a second preset convergence condition is met, thereby obtaining the trained morphological recognition model.

[0075] Specifically, an initial morphological recognition model can be constructed, and surface defect images can be collected through multi-source data sources to construct a surface defect sample data set. The surface defects of sample images in the surface defect sample data set can be labeled through manual annotation, and then the initial morphological recognition model can be iteratively trained based on the labeled surface defect sample images. After each training, the loss value can be calculated once through a preset loss function. If the loss value is greater than a preset threshold, a surface defect sample image is selected for prediction again until the preset convergence conditions are met, such as the number of iterations reaches the preset threshold, or the loss value is less than the preset threshold, then a trained weld recognition model can be obtained.

[0076] Among them, the morphological recognition model can be a feedforward neural network model, a convolutional neural network model, etc.

[0077] It should be noted that the morphological recognition model can be evaluated and optimized through the test set, and during the use of the morphological recognition model, its network parameters can also be continuously updated to improve its prediction accuracy.

[0078] In an embodiment of the present application, determining whether there is an abnormality in the weld size of the welding component based on the weld size information includes:

[0079] Constructing a standard weld size data set, wherein the standard weld size data set includes a plurality of standard weld size information;

[0080] Comparing the weld size information with the corresponding standard weld size information to determine the difference between the weld size information and the standard weld size information;

[0081] When the difference exceeds a preset error range, the weld size of the welded component is abnormal.

[0082] Specifically, different welding components may correspond to different standard weld sizes. Therefore, a standard weld size subset corresponding to each welding component may be established. Then, after obtaining the weld size information, the corresponding standard weld size subset may be found according to the welding component, and each weld size may be compared with the corresponding standard weld size in the standard weld size subset to determine whether the weld size information is consistent with the standard weld size information. If not, determine whether the difference between the weld size information and the standard weld size information is within a preset error range. If so, it indicates that the weld size is normal, otherwise, it indicates that the weld size is abnormal.

[0083] The weld size may include weld height, weld width, weld foot, weld penetration, and weld excess. For example, the standard weld excess value may be 0-3 mm, and the standard weld height may be 0-1 mm. If the predicted weld excess is 2 mm and the weld height is 1.5 mm, it can be considered that it does not meet the standard weld size and is abnormal.

[0084] In one embodiment of the present application, after determining whether the weld morphology of the welded component has defects, the method includes:

[0085] If the weld size of the welded component is abnormal, and / or the weld shape is defective, obtaining the welding angle and welding melting point temperature of the current welding gun;

[0086] Determine the welding angle range and the welding melting point temperature range corresponding to the welding component;

[0087] Determining whether the welding angle exceeds the welding angle range, and / or determining whether the welding melting point temperature exceeds the welding melting point temperature range;

[0088] If so, the welding angle is adjusted based on the welding angle range, and / or the welding melting point temperature is adjusted based on the welding melting point temperature range.

[0089] Specifically, if it is determined that there is an abnormality in the weld size of the welded component and / or there are defects in the weld morphology, the current welding angle and welding melting point temperature can be obtained to determine whether the welding angle exceeds the welding angle range and / or whether the welding melting point temperature exceeds the welding melting point temperature range, so as to perform defect analysis on the abnormal weld size and defects in weld morphology and determine the causes of the abnormal size and morphological defects. For example, when there are slight grooves on the weld, it can be considered that the temperature is too low. At this time, the temperature can be increased. If a uniform waist drum-shaped morphology appears, the welding temperature is feasible, but the extrusion amount is too small. At this time, the extrusion amount can be adjusted. If the weld is offset, the welding gun angle can be adjusted, etc.

[0090] In one embodiment of the present application, after determining whether the weld morphology of the welded component has defects, the method includes:

[0091] If there is no abnormality in the weld size of the welded component and there is no defect in the weld shape, detecting whether there is a defect inside the weld of the welded component;

[0092] If there are defects inside the weld of the welded component, an alarm message is output.

[0093] Specifically, since the acquired weld image can only detect whether there is any abnormality on the weld surface, in the actual welding process, defects such as pores may often appear inside the weld due to the welding temperature, the amount of welding material extrusion and other reasons, it is possible to detect whether there are defects inside the weld. For example, the weld area can be scanned by ultrasound, or the weld position can be knocked with a knocking piece to determine the location of the weld defect and record it. At the same time, an alarm message can be output to prompt the staff to avoid false detection problems.

[0094] In one embodiment of the present application, the step of acquiring a weld image of a welded component includes:

[0095] Acquire a welding component image, wherein the welding component image includes a weld area and a non-weld area;

[0096] Eliminating the non-weld area to obtain a weld area image;

[0097] The weld area image is subjected to image noise reduction processing and image enhancement processing to obtain the preprocessed weld image.

[0098] Specifically, when collecting weld images, due to the problems of shooting range and shooting angle, the area around the weld will also be photographed. In order to avoid the interference of non-weld areas on the prediction of weld parts, they can be eliminated and the weld area can be retained. Optionally, the weld edge can be detected to obtain a set of pixel points at the weld edge, and the weld boundary can be determined by fitting. Then, the weld area is extracted based on the weld boundary to obtain a weld area image. Then, the image can be aligned by filtering to reduce the noise. And the image is enhanced by flipping, translating, etc. to obtain a preprocessed weld image. The preprocessed weld image can be used to input into the weld recognition model and the morphological recognition model for prediction processing.

[0099] In an embodiment of the present application, by collecting weld images of welded components, the weld size and weld morphology are predicted respectively, and based on the weld size and weld morphology, it is determined whether there are size abnormalities or morphological defects. If so, an alarm message can be output to stop the machine in time and notify manual processing to avoid causing a large number of welding products with weld defects. There is no need for manual inspection, which can save labor costs, reduce manual false detection rate, and improve detection efficiency and accuracy.

[0100] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0101] In one embodiment, a weld defect detection device is provided, which corresponds one-to-one to the weld defect detection method in the above embodiment. Figure 2 As shown, the weld defect detection device includes a weld image acquisition unit 10, a prediction unit 20, a first judgment unit 30, a second judgment unit 40 and an alarm information output unit 50. Each functional module is described in detail as follows:

[0102] A weld image acquisition unit 10, used to acquire a weld image of a welded component;

[0103] A prediction unit 20 is used to input the weld image into a pre-trained weld recognition model and a morphological recognition model for prediction, so as to obtain weld size information and weld morphology information;

[0104] A first judgment unit 30 is used to determine whether there is an abnormality in the weld size of the welding component based on the weld size information;

[0105] A second judgment unit 40 is used to determine whether there is a defect in the weld morphology of the welded component based on the weld morphology information;

[0106] The alarm information output unit 50 is used to output an alarm message if there is an abnormality in the weld size of the welded component and / or there is a defect in the weld shape.

[0107] In one embodiment of the present application, the weld recognition model is trained in the following manner:

[0108] Acquire a weld sample data set, wherein the weld sample data set includes a plurality of weld sample images with labels;

[0109] Constructing an initial weld identification model;

[0110] The weld sample image is input into the initial weld recognition model for iterative training until a first preset convergence condition is met, thereby obtaining the trained weld recognition model.

[0111] In one embodiment of the present application, the morphological recognition model is trained in the following manner:

[0112] Acquire a surface defect sample data set, wherein the surface defect sample data set includes a plurality of surface defect sample images with labels;

[0113] Constructing an initial morphological recognition model;

[0114] The surface defect sample image is input into the initial morphological recognition model for iterative training until a second preset convergence condition is met, thereby obtaining the trained morphological recognition model.

[0115] In one embodiment of the present application, the first determination unit 30 is further configured to:

[0116] Constructing a standard weld size data set, wherein the standard weld size data set includes a plurality of standard weld size information;

[0117] Comparing the weld size information with the corresponding standard weld size information to determine the difference between the weld size information and the standard weld size information;

[0118] When the difference exceeds a preset error range, the weld size of the welded component is abnormal.

[0119] In one embodiment of the present application, the device further includes a welding parameter adjustment unit for:

[0120] If the weld size of the welded component is abnormal, and / or the weld shape is defective, obtaining the welding angle and welding melting point temperature of the current welding gun;

[0121] Determine the welding angle range and the welding melting point temperature range corresponding to the welding component;

[0122] Determining whether the welding angle exceeds the welding angle range, and / or determining whether the welding melting point temperature exceeds the welding melting point temperature range;

[0123] If so, the welding angle is adjusted based on the welding angle range, and / or the welding melting point temperature is adjusted based on the welding melting point temperature range.

[0124] In one embodiment of the present application, the device further includes a weld internal defect detection unit, which is used to:

[0125] If there is no abnormality in the weld size of the welded component and there is no defect in the weld shape, detecting whether there is a defect inside the weld of the welded component;

[0126] If there are defects inside the weld of the welded component, an alarm message is output.

[0127] In one embodiment of the present application, the weld image acquisition unit 10 is further used for:

[0128] Acquire a welding component image, wherein the welding component image includes a weld area and a non-weld area;

[0129] Eliminating the non-weld area to obtain a weld area image;

[0130] The weld area image is subjected to image noise reduction processing and image enhancement processing to obtain the preprocessed weld image.

[0131] In an embodiment of the present application, by collecting weld images of welded components, the weld size and weld morphology are predicted respectively, and based on the weld size and weld morphology, it is determined whether there are size abnormalities or morphological defects. If so, an alarm message can be output to stop the machine in time and notify manual processing to avoid causing a large number of welding products with weld defects. There is no need for manual inspection, which can save labor costs, reduce manual false detection rate, and improve detection efficiency and accuracy.

[0132] The specific definition of the weld defect detection device can be found in the definition of the weld defect detection method above, which will not be repeated here. Each module in the above-mentioned weld defect detection device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0133] In one embodiment, a computer device is provided. The computer device may be a terminal device, and its internal structure diagram may be as follows: Figure 3 As shown. The computer device includes a processor, a memory, and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a readable storage medium. The readable storage medium stores computer-readable instructions. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer-readable instructions are executed by the processor, a weld defect detection method is implemented. The readable storage medium provided in this embodiment includes a non-volatile readable storage medium and a volatile readable storage medium.

[0134] In an embodiment of the present application, a computer device is provided, including a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor, and when the processor executes the computer-readable instructions, the steps of the weld defect detection method as described above are implemented.

[0135] In an embodiment of the application, a readable storage medium is provided, wherein the readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by a processor, the steps of the weld defect detection method as described above are implemented.

[0136] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing related hardware through computer-readable instructions, and the computer-readable instructions can be stored in a non-volatile readable storage medium or a volatile readable storage medium. When the computer-readable instructions are executed, they may include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0137] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0138] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A weld defect detection method, characterized in that: The method comprises: Acquire weld images of welded components; The weld image is respectively input into a pre-trained weld recognition model and a morphological recognition model for prediction to obtain weld size information and weld morphology information; Based on the weld size information, determining whether there is an abnormality in the weld size of the welded component; Based on the weld morphology information, determining whether there is a defect in the weld morphology of the welded component; If there is an abnormality in the weld size of the welded component and / or there is a defect in the weld shape, an alarm message is output.

2. The weld defect detection method according to claim 1, characterized in that: The weld recognition model is trained in the following way: Acquire a weld sample data set, wherein the weld sample data set includes a plurality of weld sample images with labels; Constructing an initial weld identification model; The weld sample image is input into the initial weld recognition model for iterative training until a first preset convergence condition is met, thereby obtaining the trained weld recognition model.

3. The weld defect detection method according to claim 1, characterized in that: The morphological recognition model is trained in the following way: Acquire a surface defect sample data set, wherein the surface defect sample data set includes a plurality of surface defect sample images with labels; Constructing an initial morphological recognition model; The surface defect sample image is input into the initial morphological recognition model for iterative training until a second preset convergence condition is met, thereby obtaining the trained morphological recognition model.

4. The weld defect detection method according to claim 1, characterized in that: The step of determining whether there is an abnormality in the weld size of the welded component based on the weld size information includes: Constructing a standard weld size data set, wherein the standard weld size data set includes a plurality of standard weld size information; Comparing the weld size information with the corresponding standard weld size information to determine the difference between the weld size information and the standard weld size information; When the difference exceeds a preset error range, the weld size of the welded component is abnormal.

5. The weld defect detection method according to claim 1, characterized in that: After determining whether the weld morphology of the welded component has defects, the method includes: If the weld size of the welded component is abnormal, and / or the weld shape is defective, obtaining the welding angle and welding melting point temperature of the current welding gun; Determine the welding angle range and the welding melting point temperature range corresponding to the welding component; Determining whether the welding angle exceeds the welding angle range, and / or determining whether the welding melting point temperature exceeds the welding melting point temperature range; If so, the welding angle is adjusted based on the welding angle range, and / or the welding melting point temperature is adjusted based on the welding melting point temperature range.

6. The weld defect detection method according to claim 1, characterized in that: After determining whether the weld morphology of the welded component has defects, the method includes: If there is no abnormality in the weld size of the welded component and there is no defect in the weld shape, detecting whether there is a defect inside the weld of the welded component; If there are defects inside the weld of the welded component, an alarm message is output.

7. The weld defect detection method according to claim 1, characterized in that: The step of obtaining a weld image of a welded component comprises: Acquire a welding component image, wherein the welding component image includes a weld area and a non-weld area; Eliminating the non-weld area to obtain a weld area image; The weld area image is subjected to image noise reduction processing and image enhancement processing to obtain the preprocessed weld image.

8. A weld defect detection device, characterized in that: The device comprises: A weld image acquisition unit, used to acquire a weld image of a welded component; A prediction unit, used for inputting the weld image into a pre-trained weld recognition model and a morphological recognition model for prediction, to obtain weld size information and weld morphology information; A first judgment unit, configured to determine whether a weld size of the welding component is abnormal based on the weld size information; A second judgment unit is used to determine whether there is a defect in the weld morphology of the welded component based on the weld morphology information; The alarm information output unit is used to output alarm information if there is an abnormality in the weld size of the welded component and / or there is a defect in the weld shape.

9. A computer device comprising a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor, characterized in that: When the processor executes the computer-readable instructions, the steps of the weld defect detection method according to any one of claims 1 to 7 are implemented.

10. A readable storage medium storing computer-readable instructions, characterized in that: When the computer-readable instructions are executed by a processor, the steps of the weld defect detection method according to any one of claims 1 to 7 are implemented.

Images