Bus bar welding detection method and system, electronic device, and medium

Abstract

The application provides a busbar welding detection method and system, electronic equipment and medium. The method comprises: using a fixed-point camera to collect images of a terminal box, and obtaining a first assembly image comprising a terminal box body and a busbar assembly inside the terminal box; inputting the first assembly image into a pre-trained multi-task deep learning model, and outputting an assembly detection result of the terminal box and an assembly detection result of the busbar, and an outline extraction result of the busbar; positioning the busbar by analyzing the outline extraction result of the busbar, and performing a busbar poking operation; after the busbar is poked, the fixed-point camera is used to collect images of the terminal box again, and a second assembly image is obtained; and the welding state of the busbar is determined by comparing the similarity of the first assembly image and the second assembly image. The application realizes high-precision, high-efficiency and fully-automated detection of the welding state of the busbar, and effectively guarantees the welding quality of the busbar in the terminal box.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of photovoltaic modules, and relates to a busbar welding detection method and system, an electronic device, and a medium. BACKGROUND

[0002] Under the background of continuous adjustment and optimization of energy structure, photovoltaic modules, as a key clean energy equipment, have been widely used in solar power generation systems due to their clean, renewable, and non-geographical limitations.

[0003] However, in the large-scale production process of photovoltaic modules, due to the interference of factors such as welding equipment parameter drift and material thermal deformation, the busbar in the junction box is prone to process defects such as missing welding and false welding. These defects have hidden characteristics, and if they are not detected and effectively solved in time, they may cause a series of chain reactions such as cell efficiency decay and insulation performance decline, thereby affecting the normal operation of the entire solar power generation system, causing huge economic losses, and even causing the risk of energy supply interruption.

[0004] Currently, the welding detection of busbars mainly relies on manual inspection and traditional visual detection technology. Although the manual inspection method can find welding problems to some extent, it requires a large amount of manpower and time cost. In the large-scale production process, the problem of low efficiency is increasingly prominent. In addition, due to the uncertainty of manual operation, when the busbar is picked, uneven picking force may cause false welding problems to be difficult to be discovered in time, or the originally well-welded parts may be accidentally picked off, thereby affecting the accuracy and reliability of the detection results. The traditional machine vision technology is limited by factors such as image imaging quality and calculation speed, and it is difficult to meet the needs of high-precision and high-efficiency actual application scenarios. SUMMARY

[0005] The application provides a busbar welding detection method, system, electronic device, and medium, which are used to solve the problems of insufficient precision and low efficiency in the existing technology during busbar welding detection.

[0006] In a first aspect, a busbar welding detection method is provided, comprising: acquiring a first assembly image of a junction box and a busbar assembly in the junction box by using a fixed-point camera; inputting the first assembly image into a pre-trained multi-task deep learning model to output an assembly detection result of the junction box and an assembly detection result of the busbar, and outputting a contour extraction result of the busbar when the assembly detection result of the junction box and the assembly detection result of the busbar both meet preset requirements; wherein the contour extraction result of the busbar includes a contour shape feature of the busbar and a region coordinate of the busbar in the first assembly image; positioning the busbar by analyzing the contour extraction result of the busbar, and performing a busbar poking operation based on a dynamic self-adaptive poking strategy; after the busbar is poked, acquiring a second assembly image of the junction box by using the fixed-point camera again; and determining a welding state of the busbar by comparing the first assembly image and the second assembly image.

[0007] In an implementation form of the first aspect, inputting the first assembly image into the pre-trained multi-task deep learning model to output the assembly detection result of the junction box and the assembly detection result of the busbar, and outputting the contour extraction result of the busbar when the assembly detection result of the junction box and the assembly detection result of the busbar both meet preset requirements includes: detecting a number of the junction boxes in the first assembly image based on a deep learning target detection algorithm to obtain the assembly detection result of the junction box; detecting a number of the busbars in the junction box based on a convolutional neural network model to obtain the assembly detection result of the busbar; and when the assembly detection result of the junction box and the assembly detection result of the busbar both meet preset requirements, performing semantic segmentation identification on a region where the busbar is located in the first assembly image based on an optimized semantic segmentation model to obtain the contour extraction result of the busbar.

[0008] In an implementation form of the first aspect, the method further comprises: determining whether the junction box is missing according to the assembly detection result of the junction box; if the junction box is not missing, determining that the assembly detection result of the junction box meets the preset requirements; otherwise, determining that the assembly detection result of the junction box does not meet the preset requirements, generating a junction box missing state code, and starting a preset junction box quality detection process.

[0009] In an implementation form of the first aspect, the method further comprises: determining whether the busbar is missing according to the assembly detection result of the busbar; if the busbar is not missing, determining that the assembly detection result of the busbar meets the preset requirements; otherwise, determining that the assembly detection result of the busbar does not meet the preset requirements, triggering a busbar abnormal alarm mechanism, and jumping to a preset busbar quality detection process.

[0010] In an implementation form of the first aspect, the performing the bus bar poking operation based on the dynamic adaptive poking strategy comprises: obtaining a resolution attribute of the first assembly image, and establishing a mapping relationship between an image coordinate system and a physical coordinate system based on the resolution attribute; converting a region coordinate of the bus bar in the first assembly image into a corresponding physical coordinate based on the mapping relationship between the image coordinate system and the physical coordinate system; controlling a mechanical poking device to move to the physical coordinate of the bus bar; dynamically adjusting a size and a direction of a poking force acting on the mechanical poking device according to a profile feature of the bus bar; and driving the mechanical poking device to poke the bus bar based on the adjusted poking force.

[0011] In an implementation form of the first aspect, the determining the welding state of the bus bar by performing a similarity comparison on the first assembly image and the second assembly image comprises: cropping the first assembly image to retain a bus bar region in the first assembly image to obtain a first region image; cropping the second assembly image to retain a bus bar region in the second assembly image to obtain a second region image; calculating a pixel difference between the first region image and the second region image; comparing the calculated pixel difference with a preset pixel difference threshold; if the calculated pixel difference exceeds the preset pixel difference threshold, determining that the bus bar is abnormally welded, and feeding back an abnormal instruction to a software visual interface; and if the calculated pixel difference is less than or equal to the preset pixel difference threshold, determining that the bus bar is normally welded, and feeding back a normal instruction to the software visual interface.

[0012] In an implementation form of the first aspect, after determining that the bus bar is abnormally welded, the method further comprises: calculating an angle of the bus bar being poked based on the pixel difference; if the angle of the bus bar being poked is greater than a preset value, determining that the bus bar is missed-welded; and if the angle of the bus bar being poked is less than or equal to the preset value, determining that the bus bar is fake-welded.

[0013] In a second aspect, the present application provides a bus bar welding detection system, comprising: a photovoltaic module comprising photovoltaic cells and a plurality of junction boxes; each of the junction boxes is arranged on the surface of the photovoltaic cells and electrically connected with the photovoltaic cells, and comprises a junction box body and an internal bus bar assembly thereof; a bus bar welding detection device connected with the photovoltaic module, used for performing the bus bar welding detection method according to any one of the above aspects; and a mechanical poking device connected with the bus bar welding detection device, used for performing a bus bar poking operation under the driving of the bus bar welding detection device.

[0014] In a third aspect, the present application provides an electronic device, comprising: a memory for storing a computer program; and a processor for executing the computer program stored in the memory, so that the electronic device performs the bus bar welding detection method according to any one of the above aspects.

[0015] In a fourth aspect, the present application provides a computer readable storage medium, having stored thereon a computer program, which when executed by a processor implements the busbar welding detection method according to any one of the preceding aspects.

[0016] As described above, the busbar welding detection method, system, electronic device and medium according to the present application achieve high-precision, high-efficiency and full-automation detection of the welding state of the busbar, and can quantitatively distinguish between virtual welding and missed welding of the busbar, thereby effectively ensuring the welding quality of the busbar in the junction box. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 A structural schematic diagram of a busbar welding detection system according to an embodiment of the present application is shown.

[0018] Figure 2 A distribution schematic diagram of a junction box on a photovoltaic module according to an embodiment of the present application is shown.

[0019] Figure 3 A flowchart of a busbar welding detection method according to an embodiment of the present application is shown.

[0020] Figure 4 A flowchart for judging the assembly detection result of a junction box according to an embodiment of the present application is shown.

[0021] Figure 5 A flowchart for judging the assembly detection result of a busbar according to an embodiment of the present application is shown.

[0022] Figure 6 A flowchart of performing a busbar poking operation based on a dynamic adaptive poking strategy according to an embodiment of the present application is shown.

[0023] Figure 7 A flowchart of determining the welding state of a busbar according to an embodiment of the present application is shown.

[0024] Figure 8 A flowchart of determining the welding state of a busbar according to another embodiment of the present application is shown.

[0025] Figure 9 A structural schematic diagram of a busbar welding detection device according to an embodiment of the present application is shown.

[0026] Figure 10 A structural schematic diagram of an electronic device according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0027] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, unless otherwise specified, the following embodiments and features in the embodiments can be combined with each other.

[0028] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. Therefore, the drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0029] This application provides a busbar welding inspection method, system, electronic device, and medium, applicable to photovoltaic module production line scenarios. It achieves high-precision, high-efficiency, and fully automated inspection of busbar welding status, and can quantitatively distinguish between incomplete and missing welds in busbars, effectively ensuring the welding quality of busbars in junction boxes.

[0030] The following will describe in detail the principle and implementation of a busbar welding inspection method, system, electronic device and medium of this embodiment, so that those skilled in the art can understand the busbar welding inspection method, system, electronic device and medium of this embodiment without creative labor.

[0031] It should be noted that the busbar welding inspection method described above can be applied to various types of hardware devices. The hardware device can be a computer including components such as a memory, memory controller, one or more processing units (CPUs), peripheral interfaces, RF circuits, audio circuits, speakers, microphones, input / output (I / O) subsystems, displays, other output or control devices, and external ports; the computer includes, but is not limited to, personal computers such as desktop computers, laptops, tablets, smartphones, smart TVs, and personal digital assistants (PDAs). In other embodiments, the hardware device can also be a local server or a cloud server. The server can be deployed on one or more physical servers depending on factors such as function and load, or it can be composed of distributed or centralized server clusters; this embodiment does not impose any limitations.

[0032] This embodiment provides a busbar welding inspection system. Please refer to [link / reference]. Figure 1 The diagram shows the architecture of a busbar welding inspection system according to an embodiment of this application. Figure 1As shown, the busbar welding inspection system provided in this embodiment includes a photovoltaic module, a busbar welding inspection device, and a mechanical actuation device.

[0033] Specifically, the photovoltaic module includes a photovoltaic cell and multiple junction boxes; each junction box is disposed on the surface of the photovoltaic cell and electrically connected to the photovoltaic cell, and includes a junction box body and its internal busbar assembly.

[0034] In this embodiment, the photovoltaic cell is used to convert solar energy into electrical energy. The junction box serves to connect the photovoltaic cell to an external solar charging control device. Figure 1 (Not shown) functions to enable the power output and line protection of photovoltaic modules, acting as the "current regulation center" for photovoltaic modules.

[0035] The busbar welding inspection device is connected to the photovoltaic module and is used to perform the busbar welding inspection method described in this application.

[0036] The mechanical actuation device is connected to the busbar welding inspection device and is used to perform busbar actuation operation under the drive of the busbar welding inspection device.

[0037] Please see Figure 2 The image shows a schematic diagram illustrating the distribution of junction boxes on a photovoltaic module according to an embodiment of this application. Figure 2 As shown, a junction box is located on the left, middle, and right sides of the photovoltaic module. This layout fully considers factors such as the area of ​​the photovoltaic module, current distribution, and actual application scenarios, thereby effectively achieving balanced current transmission and efficient management.

[0038] Each junction box contains a fixed-configuration busbar assembly. Figure 2 (Not shown). According to industry standards, each junction box is typically equipped with two busbars: a main busbar and a backup busbar. These two busbars work together to perform crucial tasks such as current collection, distribution, and transmission. The busbars are made of copper and are arranged in a longitudinally parallel pattern with constant spacing to ensure electrical safety.

[0039] Please see Figure 3 The above is a flowchart illustrating a busbar welding inspection method according to an embodiment of this application. Figure 3 As shown, this embodiment provides a busbar welding inspection method, including the following steps S100 to S500.

[0040] In step S100, a fixed-point camera is used to acquire images of the junction box, obtaining a first assembly image including the junction box body and its internal busbar assembly.

[0041] Specifically, one industrial fixed-point camera can be arranged on the left side, the middle and the right side of the photovoltaic assembly respectively, and each camera is fixed on the optimal observation position at a 45-degree overhead angle with the corresponding junction box through a precise support. This distributed visual system design can comprehensively cover all key detection points of the assembly and avoid image distortion caused by wide-angle shooting of a single camera.

[0042] During image acquisition, high-definition original images captured by the left, middle and right cameras are acquired simultaneously through synchronous trigger control, so as to ensure that the subtle features of the busbar assembly are captured.

[0043] In an embodiment of the present application, the acquired first assembly image is a standard two-dimensional planar image stored in an RGB color space, which completely contains clear images of the junction box body structure and the internal busbar assembly. The pixel proportion of the busbar in the image is not less than 15%, and all the welding point areas need to present a clear focusing state to meet the accuracy requirements of the subsequent image processing algorithm for feature extraction.

[0044] In the present implementation, a stable visual reference is established by fixing the camera position, so as to ensure that the images acquired before and after the execution of the dialing operation by the mechanical dialing device have completely consistent viewing angles and imaging parameters, thereby eliminating the image difference interference caused by the change of the shooting angle and making the subsequent image comparison and analysis more accurate and reliable.

[0045] In step S200, the first assembly image is input into the pre-trained multi-task deep learning model, and the assembly detection result of the junction box and the assembly detection result of the busbar are output. When the assembly detection result of the junction box and the assembly detection result of the busbar both meet the preset requirements, the contour extraction result of the busbar is output.

[0046] Specifically, the assembly detection result of the junction box includes the number of junction boxes. The assembly detection result of the busbar includes the number of busbars. The contour extraction result of the busbar includes the contour shape feature of the busbar and the area coordinates of the busbar in the first assembly image.

[0047] Please refer to Figure 4 , which shows a flowchart for judging the assembly detection result of the junction box according to an embodiment of the present application. As Figure 4 shown, the busbar welding detection method provided by the present application further includes steps S201 and S203.

[0048] In step S201, it is judged whether the junction box is missing according to the assembly detection result of the junction box.

[0049] In step S202, if the junction box is not missing, it is determined that the assembly detection result of the junction box meets the preset requirements.

[0050] At step S203, if the junction box is missing, it is determined that the assembly detection result of the junction box does not meet the preset requirement, a junction box missing status code is generated, and a preset junction box quality detection process is started.

[0051] For example, when it is confirmed that the junction boxes in the left, middle and right areas are all accurately detected, i.e., the number of the junction boxes is complete, it is determined that the assembly detection result of the junction box meets the preset process standard.

[0052] In this embodiment, the format of the junction box missing status code can be set as "region identifier + error type + timestamp", which is used to indicate that the current photovoltaic module has been marked as unqualified due to the missing of the junction box in a certain area.

[0053] The preset junction box quality detection process in this embodiment is a complete automatic disposal scheme, which specifically includes: automatically recording defect position information, calling a mechanical arm for close-range re-inspection, comparing a material database to confirm whether it belongs to missing or misloading, generating a detailed quality inspection report and uploading it to a manufacturing execution system (MES) system, and finally diverting the unqualified photovoltaic module to a repair station.

[0054] Please refer to Figure 5 , which shows a flowchart for judging the assembly detection result of the busbar according to an embodiment of the present application. As shown in Figure 5 , the busbar welding detection method provided by the present application further includes steps S204 and S206.

[0055] At step S204, it is determined whether the busbar is missing according to the assembly detection result of the busbar.

[0056] At step S205, if the busbar is not missing, it is determined that the assembly detection result of the busbar meets the preset requirement.

[0057] At step S206, if the busbar is missing, it is determined that the assembly detection result of the busbar does not meet the preset requirement, an abnormal alarm mechanism of the busbar is triggered, and a preset busbar quality detection process is jumped to.

[0058] For example, when it is confirmed that two parallel arranged busbars are detected inside the junction boxes in the left, middle and right areas, it is determined that the assembly detection result of the busbar meets the preset process standard.

[0059] The preset bus bar quality detection process in the embodiment is a complete automatic treatment scheme, and specifically includes: starting a high-resolution industrial camera to perform multi-angle re-inspection; using a laser range finder to verify the actual existence of the bus bar; comparing a production process parameter database to confirm standard configuration; generating a quality inspection report containing a missing position schematic diagram; and automatically sorting the assembly to a repair line.

[0060] The application adopts a phased detection strategy, first performs configuration detection on the junction box, and then performs configuration detection on the bus bar, which can accurately screen out typical manufacturing defect problems such as junction box missing and bus bar missing, improves the efficiency and accuracy of detection, and provides reliable protection for the quality and performance of the photovoltaic assembly.

[0061] In the implementation, the integrity of the junction box and the bus bar assembly is quickly screened at the image level, which effectively avoids the waste of resources caused by photovoltaic assemblies missing junction boxes or bus bars flowing into subsequent detection links. At the same time, accurate quality data support can be provided for the improvement of the production process. This whole-to-local detection logic significantly improves the efficiency and reliability of photovoltaic assembly quality detection.

[0062] In an embodiment of the application, the first assembly image is input into a pre-trained multi-task deep learning model, and the assembly detection result of the junction box and the assembly detection result of the bus bar are output, and when the assembly detection result of the junction box and the assembly detection result of the bus bar both meet the preset requirements, the step S200 of outputting the contour extraction result of the bus bar can include the following steps S210 to S230.

[0063] In step S210, a deep learning-based target detection algorithm is used to detect the number of junction boxes in the first assembly image to obtain the assembly detection result of the junction box.

[0064] In this embodiment, the deep learning-based target detection algorithm can use YOLO (You Only Look Once). YOLO is a deep learning-based real-time target detection algorithm, and its core idea is to regard the target detection task as a single regression problem, directly mapping from image pixels to bounding box coordinates and class probabilities, thereby realizing end-to-end fast detection.

[0065] The embodiment uses a fine-tuning training method to train the YOLO model using a large amount of industrial data. This training method enables the model to better adapt to the actual scenarios in the photovoltaic assembly production environment, including different lighting conditions, junction box arrangement methods, and possible occlusion conditions, thereby significantly improving the generalization ability and robustness of the model.

[0066] It should be noted that the version of YOLO is not strictly limited in this embodiment, and various modified versions (such as commonly used YOLOv5, YOLOv8, YOLOv10, etc.) can be applied to the present application.

[0067] In the present implementation, by introducing the advanced target detection algorithm YOLO, the accuracy of detection is greatly improved under the premise of ensuring the detection speed. YOLO has the ability to complete all target detection in an image at one time, effectively avoiding the missed detection or false detection problems that may occur in traditional visual detection algorithms. In addition, even in a complex production environment, YOLO can still maintain stable detection performance, effectively reducing the detection errors caused by environmental factors.

[0068] In step S220, the number of busbars inside the junction box is detected based on a convolutional neural network (CNN) model, and the assembly detection result of the busbar is obtained.

[0069] In the present embodiment, the convolutional neural network model can use a classic network architecture, such as LeNet, AlexNet, VGG, GoogleNet, ResNet, and DenseNet, etc. These models can accurately identify and count the number of busbars inside the junction box through a large amount of labeled data training.

[0070] Traditional image processing methods usually rely on manually designed feature extractors, which may not be able to fully capture the complex information in the image, thereby limiting the detection accuracy. However, the artificial intelligence convolutional neural network used in the present application can automatically learn the feature representation in the image without human intervention in the feature extraction process. This feature allows the model to capture more comprehensive image information, significantly improving detection accuracy. In addition, the convolutional neural network model also exhibits good generalization ability, which can adapt to complex scenes such as variable lighting conditions and noisy environments, ensuring robustness in practical applications.

[0071] In step S230, when the assembly detection result of the junction box and the assembly detection result of the busbar both meet the preset requirements, the region where the busbar is located in the first assembly image is subjected to semantic segmentation identification based on an optimized semantic segmentation model, and the contour extraction result of the busbar is obtained.

[0072] Specifically, the optimized semantic segmentation model in the present embodiment uses the Unet architecture.

[0073] It should be noted that in order to improve the feature extraction ability and accuracy of the model, the encoder part of Unet is specially improved in this application, that is, the resnet18 structure is used to replace the original module. Resnet18 as a classic deep residual network has strong feature learning ability and good universality.

[0074] In this implementation, by integrating the resnet18 structure into the Unet architecture, the excellent feature extraction ability of Unet in image segmentation tasks is retained, and the gradient vanishing problem of deep network is effectively solved through the residual connection of resnet, thereby significantly improving the recognition accuracy of busbar contour features and the stability of model training.

[0075] In an embodiment of the present application, based on the optimized semantic segmentation model, the semantic segmentation and identification of the area where the busbar is located in the first assembly image includes: inputting the first assembly image into the optimized semantic segmentation model; using the strong feature learning and analysis ability of the optimized semantic segmentation model to identify and classify each element in the image; for the area where the busbar is located, the optimized semantic segmentation model performs detailed semantic segmentation processing; through complex convolution operation and feature mapping, the contour feature of the busbar is accurately identified, and the contour coordinate points of each busbar in the junction box in the first assembly image are output; by recording all the contour coordinate points of each busbar, the area coordinates are formed.

[0076] The optimized semantic segmentation model in this embodiment can adapt to busbar contour tasks of different models and different installation angles, greatly improving the universality and reliability of the detection system.

[0077] In step S300, the busbar is positioned by analyzing the contour extraction result of the busbar, and the busbar poking operation is performed based on the dynamic adaptive poking strategy.

[0078] Please refer to Figure 6 , which shows the flowchart of the busbar poking operation based on the dynamic adaptive poking strategy according to an embodiment of the present application. As Figure 6 shown, the busbar poking operation based on the dynamic adaptive poking strategy can include the following steps S301 to S305.

[0079] In step S301, the resolution attribute of the first assembly image is obtained, and the mapping relationship between the image coordinate system and the physical coordinate system is established based on the resolution attribute.

[0080] In step S302, based on the mapping relationship between the image coordinate system and the physical coordinate system, the area coordinates of the busbar in the first assembly image are converted into corresponding physical coordinates.

[0081] At step S303, the mechanical contact device is controlled to move to the physical coordinates of the bus bar.

[0082] At step S304, the size and direction of the contact force acting on the mechanical contact device are dynamically adjusted according to the profile characteristics of the bus bar.

[0083] At step S305, the mechanical contact device is driven to contact the bus bar based on the adjusted contact force.

[0084] Specifically, in the coordinate conversion process, an affine transformation can be used to establish the mapping relationship between the image coordinate system and the physical coordinate system.

[0085] For example, the resolution of the first assembly image is 1920x1080 pixels, and the corresponding actual physical size is 600mmx400mm. The conversion matrix can be established by obtaining the feature points through the calibration plate. If the top-left corner coordinates of the bus bar in the image are (240, 360), the physical coordinates (75mm, 133.3mm) can be calculated through the conversion matrix. The specific calculation formula is as follows:

[0086]

[0087] where X phy represents the horizontal coordinate in the physical coordinate system, Y phy represents the vertical coordinate in the physical coordinate system, x pix represents the horizontal coordinate in the image coordinate system, and y pix represents the vertical coordinate in the image coordinate system.

[0088] In this embodiment, after the mechanical contact device contacts the bus bar, the size and direction of the contact force need to be adjusted in real time to adapt to the actual working conditions.

[0089] Specifically, during the contact process, the actual contact force is monitored through the force sensor of the mechanical contact device. If the contact force does not reach the expected value (such as excessive local friction), the contact force is dynamically increased. During the contact process, the displacement deviation of the bus bar is analyzed synchronously. If the deviation exceeds the threshold value, the direction of the contact force is adjusted to correct the trajectory.

[0090] At step S400, after contacting the bus bar, the fixed-point camera is used to perform image acquisition on the junction box again to obtain a second assembly image.

[0091] In this embodiment, the arrangement of the fixed-point camera and the image acquisition process are the same as those in step S100, and will not be repeated here.

[0092] At step S500, the welding state of the bus bar is determined by comparing the similarity of the first assembly image and the second assembly image.

[0093] Please refer to Figure 7 , which shows a flow chart of determining the welding state of the busbar according to an embodiment of the present application. As shown in Figure 7 , determining the welding state of the busbar by similarity comparison between the first assembly image and the second assembly image includes the following steps S510 to S560.

[0094] In step S510, the first assembly image is cropped to retain the busbar region in the first assembly image, obtaining a first region image.

[0095] In step S520, the second assembly image is cropped to retain the busbar region in the second assembly image, obtaining a second region image.

[0096] In step S530, the pixel difference between the first region image and the second region image is calculated.

[0097] In step S540, the calculated pixel difference is compared with a preset pixel difference threshold.

[0098] In step S550, if the calculated pixel difference exceeds the preset pixel difference threshold, it is determined that the busbar welding is abnormal, and an abnormal instruction is fed back to the software visualization interface.

[0099] In step S560, if the calculated pixel difference is less than or equal to the preset pixel difference threshold, it is determined that the busbar welding is normal, and a normal instruction is fed back to the software visualization interface.

[0100] In this embodiment, fixed region cropping logic is adopted, that is, only the busbar region in each junction box is subjected to pixel difference operation. This processing method focusing on key features effectively eliminates the interference of background noise and other irrelevant elements, greatly improving the operation efficiency. In addition, fixed-size image processing ensures the spatial consistency of pixel difference calculation, improving the accuracy of calculation.

[0101] An intuitive and efficient feedback mechanism is also adopted in this embodiment. For example, when the calculated pixel difference exceeds the preset pixel difference threshold, it is determined that the busbar welding is abnormal, and an NG (No Good) instruction is fed back to the software visualization interface to prompt the operator to check and handle in a conspicuous way. Otherwise, an OK instruction is fed back to the software visualization interface to provide clear instructions for smooth production process. The visualization interface, as the core component of the human-machine interaction (HMI) system, can be integrated into the SCADA system of the industrial computer, supporting touch operation and data traceability functions.

[0102] In the present embodiment, the fixed area cropping logic and the pixel difference calculation are used to accurately determine the welding state of the bus bar, thereby improving the accuracy and reliability of the detection. The real-time feedback mechanism of the software visualization interface enables the operator to quickly obtain the detection results and take timely measures, effectively avoiding potential quality problems.

[0103] Referring to Figure 8 , a flowchart for determining the welding state of the bus bar is shown as another embodiment of the present application. As Figure 8 indicated, after determining that the bus bar is abnormally welded, the following steps S551 to S553 are further included.

[0104] In step S551, the angle of the bus bar being pushed is calculated based on the pixel difference.

[0105] In step S552, if the angle of the bus bar being pushed is greater than a preset value, the bus bar is determined to be missing.

[0106] In step S553, if the angle of the bus bar being pushed is less than or equal to the preset value, the bus bar is determined to be fake.

[0107] In the present embodiment, calculating the angle of the bus bar being pushed based on the pixel difference includes: converting the pixel difference into actual physical displacement in combination with the calibration parameters of the camera; and determining the angle of the bus bar being pushed according to the actual physical displacement.

[0108] The angle of the bus bar being pushed directly reflects the bonding strength between the bus bar and the solder strip. Specifically, in the missing state, the bus bar does not form an effective connection with the solder strip, and a slight external force can cause it to be significantly deflected; while in the fake state, due to the weak adhesion of the local weld points, the bus bar will only be partially pushed under a large external force.

[0109] In the present embodiment, objective and quantitative differentiation between fake and missing is achieved, avoiding subjective errors in manual judgment.

[0110] It should be noted that the protection scope of the bus bar welding detection method described in the embodiments of the present application is not limited to the execution order of the steps listed in the embodiments, and any scheme achieved by adding, replacing or replacing steps of the prior art according to the principles of the present application is included in the protection scope of the present application.

[0111] Referring to Figure 9 , a structural schematic diagram of a bus bar welding detection device is shown as an embodiment of the present application. As Figure 9 indicated, the present application provides a bus bar welding detection device, which includes a first image acquisition module, an image processing module, a pushing execution module, a second image acquisition module and an image comparison module.

[0112] The first image acquisition module is configured to acquire a first assembly image of the terminal box and a busbar assembly in the terminal box by using a fixed-point camera.

[0113] The image processing module is configured to input the first assembly image into a pre-trained multi-task deep learning model, output an assembly detection result of the terminal box and an assembly detection result of the busbar, and output a contour extraction result of the busbar when the assembly detection result of the terminal box and the assembly detection result of the busbar both meet preset requirements; the contour extraction result of the busbar includes a contour feature of the busbar and a region coordinate of the busbar in the first assembly image.

[0114] The dialing execution module is configured to locate the busbar by analyzing the contour extraction result of the busbar, and perform a dialing operation on the busbar based on a dynamic self-adaptive dialing strategy.

[0115] The second image acquisition module is configured to acquire a second assembly image of the terminal box and the busbar by using the fixed-point camera again after the dialing operation on the busbar.

[0116] The image comparison module is configured to determine a welding state of the busbar by comparing the first assembly image and the second assembly image.

[0117] It should be noted that the first image acquisition module, the image processing module, the dialing execution module, the second image acquisition module, and the image comparison module described in this embodiment correspond to the steps of the busbar welding detection method described above one by one, and thus will not be described again here.

[0118] The busbar welding detection device provided in this embodiment can implement the busbar welding detection method described in this application, but the implementation device of the busbar welding detection method described in this application includes but is not limited to the structure of the busbar welding detection device listed in this embodiment. Any modification and replacement of the existing technology according to the principle of this application are included in the protection scope of this application.

[0119] Please refer to Figure 10 , which shows a structural schematic diagram of an electronic device according to an embodiment of the present application. As shown in Figure 10 , the present application provides an electronic device, which includes a memory and a processor.

[0120] The memory is configured to store a computer program.

[0121] In some possible implementation manners, the memory can include a ROM, a RAM, a disk, a U disk, a memory card, an optical disk, or various media that can store program codes.

[0122] In embodiments of the present application, the memory can include computer system readable media in the form of volatile memory, such as RAM and / or cache memory. The electronic device can further include other removable / non-removable, volatile / non-volatile computer system storage media. The memory can include at least one program product having a set (e.g., at least one) of program modules that are configured to carry out the functions of embodiments of the present application.

[0123] The processor is configured to execute a computer program stored in the memory, so that the electronic device performs the bus bar welding detection method described in any one of the above.

[0124] Exemplarily, the processor can be a general processor, including a central processing unit (CPU), a network processor (NP), etc. In other embodiments, the processor can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component.

[0125] In some implementations, the electronic device provided by embodiments of the present application can further include a display. The display is connected in communication with the memory and the processor, and is configured to display a related graphical user interface (GUI) of the bus bar welding detection method.

[0126] In embodiments of the present application, the display can include a display screen (display panel). In some implementations, the display panel can be configured in the form of a liquid crystal display (LCD), an organic light-emitting diode (OLED), etc. In addition, the display can also be a touch panel (touch screen, touch screen), which can include a display screen and a touch-sensitive surface. When the touch-sensitive surface detects a touch operation on or near it, it is transmitted to the processor to determine the type of touch event, and then the processor provides corresponding visual output on the display device according to the type of touch event.

[0127] In several embodiments provided in the present application, it should be understood that the disclosed system, device or method can be implemented in other manners. For example, the division of the above-described device embodiment is only a logical function division, and there can be another division manner for the actual implementation, for example, multiple devices or multiple units can be combined or integrated into another system, or some characteristics can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections between different parts can be indirect couplings or communication connections through some interfaces, devices or units, and can be electric, mechanical or in other forms.

[0128] The modules / units described as separated parts can or can not be physically separated, and the parts displayed as modules / units can or can not be physical modules, i.e., can be located in one place or distributed on multiple network units. Some or all of the modules / units can be selected according to actual needs to achieve the purpose of the embodiments of the present application. For example, the functional modules / units in the embodiments of the present application can be integrated in one processing module, or each module / unit can be physically present alone, or two or more modules / units can be integrated in one module / unit.

[0129] Those of ordinary skill in the art should further appreciate that the units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, computer software or a combination of both. In order to clearly illustrate the interchangeability of hardware and software, the components and steps of the examples have been described in general terms in the above description. Whether the functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. A person skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0130] The embodiments of the present application further provide a computer readable storage medium, which stores a computer program. The computer program is executed by a processor to implement the bus bar welding detection method. Those skilled in the art can understand that all or part of the steps of the method described in the above embodiments can be instructed by a program to complete the processor, and the program can be stored in a computer readable storage medium. The storage medium is a non-transitory medium, such as a random access memory, a read-only memory, a flash memory, a hard disk, a solid state disk, a magnetic tape, a floppy disk, an optical disc, and any combination thereof. The storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, a data center, and the like, which includes one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a magnetic tape), an optical medium (for example, a digital video disc (DVD)), or a semiconductor medium (for example, a solid state disk (SSD)), and the like.

[0131] The description of the corresponding flow or structure of each of the above figures has its own emphasis. The part not described in detail in a certain flow or structure can be referred to the related description of other flow or structure.

[0132] The above embodiments only exemplarily illustrate the principles and effects of the present application, and are not used to limit the present application. Any person skilled in the art can modify or change the above embodiments without departing from the spirit and scope of the present application. Therefore, all equivalent modifications or changes completed by those skilled in the art without departing from the spirit and technical thought of the present application should be covered by the claims of the present application.

Claims

1. A busbar welding inspection method characterized by comprising: The method comprises the following steps: An image of the terminal box is collected by a fixed-point camera to obtain a first assembly image including the terminal box body and the internal bus bar assembly thereof; The first assembly image is input into a pre-trained multi-task deep learning model to output an assembly detection result of the terminal box and an assembly detection result of the bus bar, and when the assembly detection result of the terminal box and the assembly detection result of the bus bar both meet preset requirements, a contour extraction result of the bus bar is output; wherein the contour extraction result of the bus bar includes the contour shape feature of the bus bar and the region coordinates of the bus bar in the first assembly image; The bus bar is positioned by analyzing the contour extraction result of the bus bar, and a bus bar poking operation is performed based on a dynamic adaptive poking strategy; After the bus bar is poked, the terminal box is collected by the fixed-point camera again to obtain a second assembly image; The welding state of the bus bar is determined by comparing the similarity of the first assembly image and the second assembly image; The first assembly image is input into a pre-trained multi-task deep learning model to output an assembly detection result of the terminal box and an assembly detection result of the bus bar, and when the assembly detection result of the terminal box and the assembly detection result of the bus bar both meet preset requirements, a contour extraction result of the bus bar is output, which comprises: A target detection algorithm based on deep learning is used to detect the number of terminal boxes in the first assembly image to obtain the assembly detection result of the terminal box; A convolutional neural network model is used to detect the number of internal bus bars of the terminal box to obtain the assembly detection result of the bus bar; When the assembly detection result of the terminal box and the assembly detection result of the bus bar both meet preset requirements, an optimized semantic segmentation model is used to perform semantic segmentation identification on the region of the bus bar in the first assembly image to obtain the contour extraction result of the bus bar, wherein the optimized semantic segmentation model uses the Unet architecture, the original encoder part of Unet is replaced by the resnet18 structure, and the preset requirements include that the terminal box is not missing and the bus bar is not missing.

2. The method of claim 1, wherein, Further comprising: determining whether the terminal box is missing according to the assembly detection result of the terminal box; if the terminal box is not missing, it is determined that the assembly detection result of the terminal box meets the preset requirements; otherwise, it is determined that the assembly detection result of the terminal box does not meet the preset requirements, a terminal box missing state code is generated, and a preset terminal box quality detection process is started.

3. The method of claim 1, wherein, Further comprising: determining whether the bus bar is missing according to the assembly detection result of the bus bar; if the bus bar is not missing, it is determined that the assembly detection result of the bus bar meets the preset requirements; otherwise, it is determined that the assembly detection result of the bus bar does not meet the preset requirements, a bus bar abnormal alarm mechanism is triggered, and a preset bus bar quality detection process is jumped to.

4. The method of claim 1, wherein, Performing a bus bar poking operation based on a dynamic adaptive poking strategy comprises: obtaining the resolution attribute of the first assembly image, and establishing a mapping relationship between the image coordinate system and the physical coordinate system based on the resolution attribute; Converting the region coordinate of the bus bar in the first assembly image into corresponding physical coordinate based on the mapping relationship between the image coordinate system and the physical coordinate system; Controlling the mechanical poking device to move to the physical coordinate of the bus bar; According to the profile characteristics of the bus bar, dynamically adjusting the size and direction of the poking force acting on the mechanical poking device; Based on the adjusted poking force, driving the mechanical poking device to poke the bus bar.

5. The method of claim 1, wherein, By comparing the similarity of the first assembly image and the second assembly image, the welding state of the bus bar is determined, including: Cutting the first assembly image to retain the bus bar region in the first assembly image to obtain a first region image; Cutting the second assembly image to retain the bus bar region in the second assembly image to obtain a second region image; Calculating the pixel difference between the first region image and the second region image; Comparing the calculated pixel difference with the preset pixel difference threshold; If the calculated pixel difference exceeds the preset pixel difference threshold, it is determined that the bus bar is abnormally welded, and an abnormal instruction is fed back to the software visualization interface; If the calculated pixel difference is less than or equal to the preset pixel difference threshold, it is determined that the bus bar is normally welded, and a normal instruction is fed back to the software visualization interface.

6. The method of claim 5, wherein, After determining that the bus bar is abnormally welded, further including: Based on the pixel difference, calculating the angle of the bus bar being poked; If the angle of the bus bar being poked is greater than a preset value, it is determined that the bus bar is missed; If the angle of the bus bar being poked is less than or equal to the preset value, it is determined that the bus bar is false.

7. A busbar weld inspection system characterized by, Including: A photovoltaic module includes photovoltaic cells and a plurality of junction boxes; each junction box is arranged on the surface of the photovoltaic cell and electrically connected to the photovoltaic cell, including a junction box body and an internal bus bar assembly thereof; A bus bar welding detection device connected to the photovoltaic module is used to execute the bus bar welding detection method according to any one of claims 1 to 6; A mechanical poking device connected to the bus bar welding detection device is used to execute the bus bar poking operation under the driving of the bus bar welding detection device.

8. An electronic device, comprising: Including: A memory for storing a computer program; A processor for executing the computer program stored in the memory to enable the electronic device to execute the bus bar welding detection method according to any one of claims 1 to 6.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the bus bar welding detection method according to any one of claims 1 to 6.

Images