Star flash mainboard welding quality detection method and system
Through the camera, high-definition images of welding points are taken and defect prediction models are used to extract the characteristics of welding points for analysis, which solves the problem that the existing technology is difficult to detect internal defects, and achieves comprehensive inspection of the welding quality of Starflash motherboard and product quality improvement.
Patent Information
- Application Number
- CN202411033660.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-30
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2044-07-30
AI Technical Summary
The prior art is difficult to detect special defects such as cracks or interspersed pore bubbles in the welding of Starflash motherboards, resulting in insufficient comprehensive welding quality inspection.
Use the camera to capture high-definition images, identify the solder joint areas, extract geometric features, texture features, color features and frequency domain features of solder joint images, and input the defect prediction model to predict the probability value of internal defects of solder joints. Defect prediction constructed by ConvNeXt-YOLO5 Models are analyzed.
It can effectively determine whether the motherboard has welding quality defects, which improves the accuracy of welding quality inspection and the quality of production motherboard products.
Smart Images

Figure CN119027378A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of quality inspection, and in particular to a method and system for inspecting the welding quality of a Star Flash motherboard. Background Art
[0002] Huawei's NearLink technology is a new type of short-range wireless communication technology. Through innovative upgrades, it has significantly improved the speed, latency, transmission distance, security and reliability compared to traditional Bluetooth and Wi-Fi technologies. It is called an enhanced version of the "Wi-Fi + Bluetooth" hybrid. Motherboards that support NearLink technology have great market prospects.
[0003] The motherboard contains a large number of components such as resistors, diodes, transistors and Star Flash chips. If the components are not welded firmly, the computer will malfunction after a period of use and need to be repaired. Therefore, after the computer motherboard is welded, it is usually necessary to test the welding condition of the motherboard. Only motherboards that pass the test can be subsequently assembled and produced.
[0004] There are some welding quality detection methods in the prior art, which use cameras to take photos of the motherboard and then process the photos through algorithms to identify whether there are defects such as cold welding, short circuits, and open circuits at the welding points. However, the above schemes are difficult to detect for some special welding defects, such as cracks inside the solder joints or pores and bubbles. Summary of the invention
[0005] In this regard, the present invention provides a Star Flash motherboard welding quality detection method, system, electronic equipment, computer storage medium and computer program product to solve at least one of the above-mentioned technical problems.
[0006] The present invention provides a method for detecting the welding quality of a Star Flash motherboard, the method comprising the following steps:
[0007] Use a camera to capture a high-definition image of the motherboard to be inspected, identify each solder joint area in the high-definition image, and obtain a solder joint area image;
[0008] Extracting solder joint features from each solder joint area image and inputting the solder joint features into a defect prediction model; wherein the solder joint features include geometric features, texture features, color features and solder joint image frequency domain features;
[0009] The defect prediction model outputs a predicted probability value related to internal defects in solder joints. When the predicted probability value is higher than a probability threshold, it is determined that there are soldering quality defects in the mainboard; wherein the internal defects refer to internal cracks and / or bubbles.
[0010] Optionally, the step of using a camera to capture a high-definition image of the motherboard to be inspected, identifying each solder joint area in the high-definition image, and obtaining a solder joint area image includes:
[0011] Using a camera to capture a first high-definition image of the motherboard to be inspected at a first focal length, binarizing the first high-definition image, and identifying the size and structure type of each bright spot area in the binarized image of the first high-definition image; the structure type includes a shadow type and a shadow-free type;
[0012] Determine a second focal length based on the size of each bright spot area, and use a camera to take at least one second high-definition image of each bright spot area to be inspected at the second focal length as the solder joint area image; the second focal length is positively correlated with the size, and the second focal length is greater than the first focal length;
[0013] Wherein, when the structural type of the bright spot area is a shadow-free type, the number of the second high-definition images is one; when the structural type of the bright spot area is a shadow-containing type, the number of the second high-definition images is at least two.
[0014] Optionally, before binarizing the first high-definition image, the method further includes:
[0015] Obtain the motherboard color and welding material type of the motherboard to be inspected in this batch through the production management system, and determine the soldering point color of the motherboard to be inspected according to the welding material type;
[0016] If the mainboard color is a dark type and the solder joint colors are all light types, a binarization processing signal is generated, and the binarization processing signal is used to trigger binarization processing of the first high-definition image.
[0017] Optionally, extracting solder joint features from each solder joint area image and inputting the solder joint features into a defect prediction model comprises:
[0018] Extracting geometric features, texture features, and color features of the solder joints from the solder joint area images;
[0019] The solder joint area image is converted into a frequency domain representation by Fourier transform, and the frequency domain component analysis is performed on the converted frequency domain image to obtain the frequency domain features of the solder joint image, wherein the frequency domain features of the solder joint image include frequency peak features and energy spectrum features;
[0020] The geometric features, the texture features, the color features and the frequency domain features of the solder joint image are constructed into a feature matrix, that is, the solder joint features are obtained, and the solder joint features are input into a defect prediction model.
[0021] Optionally, the frequency domain features of the welding spot image further include at least one of the width and shape of the spectrum, the frequency response function features, the abnormal signal features in the frequency domain, and the statistical features of the frequency domain features.
[0022] Optionally, the defect prediction model is specifically:
[0023] The defect prediction model is constructed based on ConvNeXt-YOLO5, and the defect prediction model includes a backbone network, a feature extraction network and a detection network; the feature extraction network includes a backbone network attention extraction layer, a multi-scale feature layer with an attention module and an adaptive spatial feature fusion layer.
[0024] Among them, ECA modules are respectively set at the output positions of the reciprocal N layers of the backbone network and the multi-scale feature layer; and an adaptive fusion module is also set after the ECA module of the multi-scale feature layer, and the adaptive fusion module is used for weighted summation of multiple feature information outputs of the multi-scale feature layer.
[0025] The present invention also provides a Xingshan motherboard welding quality detection system, the system includes a solder joint recognition module, a feature extraction module, and an internal defect analysis module;
[0026] The solder joint identification module is used to use a camera to capture a high-definition image of the motherboard to be inspected, identify each solder joint area in the high-definition image, and obtain a solder joint area image;
[0027] The feature extraction module is used to extract solder joint features from each solder joint area image and input the solder joint features into the defect prediction model; wherein the solder joint features include geometric features, texture features, color features and solder joint image frequency domain features;
[0028] The internal defect analysis module is used for the defect prediction model to output a predicted probability value related to the internal defects of the solder joint. When the predicted probability value is higher than a probability threshold, it is determined that the mainboard has a soldering quality defect; wherein the internal defect refers to internal cracks and / or bubbles.
[0029] The present invention also provides an electronic device, which includes: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the computer program implements any of the methods described above when executed by the processor.
[0030] The present invention also provides a computer storage medium, which stores a computer program executable by a processor, and the computer program implements any of the above methods when executed by the processor.
[0031] The present invention also provides a computer program product, which includes a computer program executable by a processor, and the computer program implements any of the above methods when executed by the processor.
[0032] The beneficial effect of the present invention is that the present invention predicts the probability of internal defects of the solder joint based on its appearance characteristics, overcomes the defect that the existing motherboard quality inspection method cannot analyze internal defects, and can greatly improve the product quality of the produced motherboard. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.
[0034] Figure 1 It is a flow chart of a method for detecting the welding quality of a Star Flash motherboard disclosed in an embodiment of the present invention;
[0035] Figure 2 is a schematic diagram of a mainboard with different types of solder joints disclosed in an embodiment of the present invention;
[0036] Figure 3 It is a structural schematic diagram of the Xingshan motherboard welding quality detection system disclosed in an embodiment of the present invention. DETAILED DESCRIPTION
[0037] The following is a description of the implementation of the present application by specific specific embodiments. People familiar with the technology can easily understand other advantages and effects of the present application from the contents disclosed in this specification. 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 the field without creative work are within the scope of protection of the present application.
[0038] In addition, the technical features involved in the different embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.
[0039] See also Figure 1 The embodiment of the present invention discloses a method for detecting the welding quality of a Star Flash motherboard, the method comprising the following steps:
[0040] Use a camera to capture a high-definition image of the motherboard to be inspected, identify each solder joint area in the high-definition image, and obtain a solder joint area image;
[0041] Extracting solder joint features from each solder joint area image and inputting the solder joint features into a defect prediction model; wherein the solder joint features include geometric features, texture features, color features and solder joint image frequency domain features;
[0042] The defect prediction model outputs a predicted probability value related to internal defects in solder joints. When the predicted probability value is higher than a probability threshold, it is determined that there are soldering quality defects in the mainboard; wherein the internal defects refer to internal cracks and / or bubbles.
[0043] The present invention extracts the features of the welding points on the Star Flash motherboard through the captured high-definition images, including geometric features, texture features, color features and frequency domain features of the welding point images, which constitute the appearance features of each welding point. These appearance features are related to the internal defects of the welding points, such as cracks and bubbles. By using a trained defect prediction model to analyze these appearance features, it is possible to predict the predicted probability value of the existence of the above defects inside the welding point. When the predicted probability value is higher than the probability threshold, it is determined that the motherboard has welding quality defects. Therefore, the present invention predicts the probability of defects existing inside the welding point based on the appearance features of the welding point, which overcomes the defect that the existing motherboard quality inspection method cannot analyze internal defects, and can greatly improve the product quality of the produced motherboards.
[0044] Optionally, the step of using a camera to capture a high-definition image of the motherboard to be inspected, identifying each solder joint area in the high-definition image, and obtaining a solder joint area image includes:
[0045] Using a camera to capture a first high-definition image of the motherboard to be inspected at a first focal length, binarizing the first high-definition image, and identifying the size and structure type of each bright spot area in the binarized image of the first high-definition image; the structure type includes a shadow type and a shadow-free type;
[0046] Determine a second focal length based on the size of each bright spot area, and use a camera to take at least one second high-definition image of each bright spot area to be inspected at the second focal length as the solder joint area image; the second focal length is positively correlated with the size, and the second focal length is greater than the first focal length;
[0047] Wherein, when the structural type of the bright spot area is a shadow-free type, the number of the second high-definition images is one; when the structural type of the bright spot area is a shadow-containing type, the number of the second high-definition images is at least two.
[0048] In this embodiment, the present invention first captures a first high-definition image of a motherboard with a large viewing angle at a relatively small first focal length, and performs a binarization process on the image. Since the solder joint area is usually reflective or light in color, it is brighter in the binarized image and is represented as white, while the color of the motherboard is usually darker and is processed as black in the binarized image. In this way, the solder joint area can be determined by a simple bright spot recognition algorithm. Compared with the traditional method of performing solder joint recognition on the original image, the solution of the present invention has a faster solder joint recognition rate and a lower recognition cost.
[0049] After the solder joint area is identified, the size and structure type of the solder joint are also identified. According to the size of the solder joint and the corresponding positive correlation function, a suitable and larger second focal length can be calculated, and at least one second high-definition image, i.e., a detail image, of the solder joint is continuously photographed at the second focal length. The second high-definition image is photographed at different angles, so that more comprehensive appearance feature information of the solder joint can be obtained based on multiple images, such as the texture depth of the solder joint surface, the features of certain shadow areas, etc.
[0050] At the same time, the solder joints on the motherboard may be independent or two or more solder joints that are very close to each other. Figure 2 A region in the figure), a single image can capture all the appearance features of the solder joint, while for the latter (see Figure 2 In the example of area B in the figure, shadows may appear on some surface areas of the solder joint due to occlusion, resulting in the inability to accurately extract certain appearance features. In view of this, the present invention further provides that when the structural type of the bright spot area is shadowed, multiple high-definition images (at different tilt angles) of the corresponding bright spot are taken for subsequent feature extraction and fusion; and when the structural type of the bright spot area is shadowless, only one high-definition image of the corresponding bright spot is taken (preferably taken vertically from above).
[0051] Optionally, before binarizing the first high-definition image, the method further includes:
[0052] Obtain the motherboard color and welding material type of the motherboard to be inspected in this batch through the production management system, and determine the soldering point color of the motherboard to be inspected according to the welding material type;
[0053] If the mainboard color is a dark type and the solder joint colors are all light types, a binarization processing signal is generated, and the binarization processing signal is used to trigger binarization processing of the first high-definition image.
[0054] In this embodiment, although the above scheme of the present invention can improve the recognition rate of solder joints, it requires that the color of the motherboard to be inspected is dark and the color of the solder joints is light. However, motherboards are generally dark colors such as green and blue, but there are also other light colors or uncoated types; due to different welding materials, solder joints are usually silver-white or golden yellow, but they may also appear green or blue, such as solder pastes with different ingredients and additives. In this case, after binarization processing, some solder joints cannot be fully presented as bright spots in the binary image, resulting in omissions in the recognition of solder joints.
[0055] In this regard, the present invention first obtains the relevant information of the motherboards to be inspected in this batch from the production management system, including the motherboard color and the type of welding material. Different types of welding materials have different colors, and the color of the welding material can be determined. When it is determined that the color of the motherboard is a dark type and the colors of the solder joints are all light types, the solder joints can be identified by binarization. Otherwise, the traditional method is used, that is, the solder joints are directly identified on the original image. This method can be implemented with the help of a convolutional network model.
[0056] Optionally, extracting solder joint features from each solder joint area image and inputting the solder joint features into a defect prediction model comprises:
[0057] Extracting geometric features, texture features, and color features of the solder joints from the solder joint area images;
[0058] The solder joint area image is converted into a frequency domain representation by Fourier transform, and the frequency domain component analysis is performed on the converted frequency domain image to obtain the frequency domain features of the solder joint image, wherein the frequency domain features of the solder joint image include frequency peak features and energy spectrum features;
[0059] The geometric features, the texture features, the color features and the frequency domain features of the solder joint image are constructed into a feature matrix, that is, the solder joint features are obtained, and the solder joint features are input into a defect prediction model.
[0060] In this embodiment, the solder joint features in the present invention include geometric features, texture features, and color features of the solder joints, and these features can be directly extracted from the spatial domain image of the solder joint area image.
[0061] 1) Geometric features include shape, size, symmetry, and concavity:
[0062] Shape: The shape of a solder joint can be affected by internal defects. For example, if there is a crack inside the solder joint, its surface may appear irregular or broken. Contour detection algorithms can be used to identify the edges of the solder joint and analyze its shape.
[0063] Size: The size of the solder joint may be related to insufficient or excessive melting during the welding process. Internal defects may cause changes in the volume of the solder joint. The size of the solder joint is determined by measuring the area or diameter of the solder joint through image processing technology.
[0064] Symmetry: The symmetry of a solder joint can reflect the consistency of the soldering process, while asymmetrical solder joints may indicate internal defects.
[0065] Concavity: The concavity of the solder joint surface may be related to internal bubbles or voids. An uneven surface may indicate internal defects.
[0066] 2) Texture features include texture roughness, texture consistency, and texture directionality.
[0067] Texture roughness: The texture roughness of the solder joint surface may be related to uneven melting during the soldering process. Rough texture may indicate internal defects. Use gray level co-occurrence matrix (GLCM) or local binary pattern (LBP) to extract texture features.
[0068] Texture consistency: The texture consistency of the solder joint surface can reflect the stability of the soldering process. Inconsistent texture may indicate internal defects. Texture consistency can be evaluated by texture analysis algorithms.
[0069] Texture Directionality: The texture directionality of the solder joint surface may be related to the flow direction during the welding process, and changes in directionality may indicate internal defects.
[0070] 3) Color features include color distribution, color contrast, and color moment.
[0071] Color distribution: The color distribution of the solder joint may be related to the melting and mixing of the soldering materials. Uneven distribution of color may indicate internal defects. The color distribution of the solder joint can be obtained through color histogram analysis.
[0072] Color contrast: The contrast of the solder joint surface can be related to the depth of fusion during the soldering process. Changes in contrast can indicate internal defects.
[0073] Color moment: Color moment is a statistic that describes color distribution and can be used to identify the color characteristics of the solder joint surface.
[0074] As for the frequency domain features of the solder joint image, it is necessary to use Fourier transform to convert the solder joint area image to the frequency domain representation, and then perform frequency domain component analysis on the frequency domain image, mainly to analyze the frequency peak characteristics and energy spectrum characteristics in the frequency domain image. The frequency distribution and energy distribution in the frequency domain image are related to the internal defects of the solder joint. The above features are matrixed and fused to obtain the solder joint features, and then the defect prediction model performs in-depth processing and analysis on them to obtain the probability of the corresponding solder joint having internal defects.
[0075] 1) Frequency peak characteristics include the location and size of the frequency peak, which can reflect certain characteristics inside the solder joint. For example, if there is a defect inside the solder joint, an additional peak may appear in the low-frequency area, indicating that the presence of the defect has changed the vibration mode of the solder joint.
[0076] 2) The energy spectrum feature is mainly the concentration or dispersion of the energy spectrum in the frequency range. The energy spectrum can show the distribution of the vibration energy of the solder joint at different frequencies. If there are defects inside the solder joint, the energy spectrum may be abnormally concentrated or dispersed in certain frequency ranges, which is related to the local stiffness change caused by the defect.
[0077] The present invention comprehensively uses the above-mentioned multiple parameters related to the internal defects of the solder joint to analyze the internal defects of the solder joint, which can improve the accuracy of the defect analysis result.
[0078] Optionally, the frequency domain features of the welding spot image further include at least one of the width and shape of the spectrum, the frequency response function features, the abnormal signal features in the frequency domain, and the statistical features of the frequency domain features.
[0079] In this embodiment, in addition to the above-mentioned frequency peak characteristics and energy spectrum characteristics, the present invention may also select at least one of the width and shape of the spectrum, the frequency response function, abnormal signals in the frequency domain, and statistical analysis of frequency domain characteristics.
[0080] Width and shape of the spectrum: The width of the spectrum is often related to the damping characteristics of the solder joint, while changes in shape may indicate inhomogeneities or defects within the solder joint. For example, if the spectrum appears asymmetrical or has multiple peaks, it may indicate material discontinuities or defects within the solder joint.
[0081] The frequency response function (FRF) can reveal the dynamic response characteristics of the solder joint at different frequencies. Defects inside the solder joint may cause significant changes in the FRF, such as a shift in the resonant frequency or a reduction in the peak value.
[0082] In frequency domain analysis, abnormal signals may appear as atypical frequency components, which may be related to defects inside the solder joint. For example, cracks or pores inside the solder joint may introduce new frequency components, resulting in abnormal peaks in the spectrum.
[0083] Statistical analysis of frequency domain features: By performing statistical analysis on frequency domain features, such as calculating the mean and variance of the spectrum, the characteristics of internal defects in the solder joint can be further identified. For example, a high variance in the spectrum may indicate a large inhomogeneity or defect inside the solder joint.
[0084] Optionally, the defect prediction model is specifically:
[0085] The defect prediction model is constructed based on ConvNeXt-YOLO5, and the defect prediction model includes a backbone network, a feature extraction network and a detection network; the feature extraction network includes a backbone network attention extraction layer, a multi-scale feature layer with an attention module and an adaptive spatial feature fusion layer.
[0086] Among them, ECA modules are respectively set at the output positions of the reciprocal N layers of the backbone network and the multi-scale feature layer; and an adaptive fusion module is also set after the ECA module of the multi-scale feature layer, and the adaptive fusion module is used for weighted summation of multiple feature information outputs of the multi-scale feature layer.
[0087] In this embodiment, the defect prediction model of the present invention is constructed based on ConvNeXt-YOLO5, including a backbone network, a feature extraction network and a detection network. The present invention improves the traditional ConvNeXt-YOLO5, and sets ECA modules at the output positions of the reciprocal N layers (e.g., the reciprocal 1st to 4th layers) of the backbone network and the multi-scale feature layer, respectively. The setting of the ECA module does not add too many parameters to the defect prediction model, but can increase the model's feature extraction capability for small targets and increase the detection accuracy of the model. In addition, an adaptive fusion module is also set after the ECA module of the multi-scale feature layer, which is used to weighted sum the multiple feature information outputs of the multi-scale feature layer to enhance the invariance of the feature scale.
[0088] See also Figure 3 , the embodiment of the present invention also discloses a Star Flash motherboard welding quality detection system, the system includes a solder joint recognition module, a feature extraction module, and an internal defect analysis module;
[0089] The solder joint identification module is used to use a camera to capture a high-definition image of the motherboard to be inspected, identify each solder joint area in the high-definition image, and obtain a solder joint area image;
[0090] The feature extraction module is used to extract solder joint features from each solder joint area image and input the solder joint features into the defect prediction model; wherein the solder joint features include geometric features, texture features, color features and solder joint image frequency domain features;
[0091] The internal defect analysis module is used for the defect prediction model to output a predicted probability value related to the internal defects of the solder joint. When the predicted probability value is higher than a probability threshold, it is determined that the mainboard has a soldering quality defect; wherein the internal defect refers to internal cracks and / or bubbles.
[0092] An embodiment of the present invention further discloses an electronic device, which includes: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the computer program implements any of the methods described above when executed by the processor.
[0093] Multiple components in the electronic device are connected to the I / O interface, including: an input unit, an output unit, a storage unit, and a communication unit. The input unit can be any type of device that can input information to the electronic device. The input unit can receive input digital or character information, and generate key signal input related to user settings and / or function control of the electronic device, and can include but is not limited to a mouse, a keyboard, a touch screen, a track pad, a track ball, a joystick, a microphone, and / or a remote controller. The output unit can be any type of device that can present information, and can include but is not limited to a display, a speaker, a video / audio output terminal, a vibrator, and / or a printer. The storage unit can include but is not limited to a disk, an optical disk. The communication unit allows the electronic device to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks, and can include but is not limited to a modem, a network card, an infrared communication device, a wireless communication transceiver, and / or a chipset, such as a Bluetooth™ device, an 802.11 device, a WiFi device, a WiMax device, a cellular communication device, and / or the like.
[0094] An embodiment of the present invention further discloses a computer storage medium, which stores a computer program executable by a processor, and the computer program implements any of the above methods when executed by the processor.
[0095] In the context of the present disclosure, a computer storage medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A computer storage medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
[0096] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).
[0097] An embodiment of the present invention further discloses a computer program product, which includes a computer program executable by a processor, and the computer program implements any of the above methods when executed by the processor.
[0098] The systems and techniques described herein can be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), the Internet, and a blockchain network.
[0099] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in this disclosure can be performed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.
[0100] Although the embodiments or examples of the present disclosure have been described with reference to the accompanying drawings, it should be understood that the above-mentioned methods, systems and devices are merely exemplary embodiments or examples, and the scope of the present invention is not limited by these embodiments or examples, but only by the claims after authorization and their equivalent scope. Various elements in the embodiments or examples can be omitted or replaced by their equivalent elements. In addition, each step can be performed in an order different from that described in the present disclosure. Further, the various elements in the embodiments or examples can be combined in various ways. It is important that with the evolution of technology, many elements described herein can be replaced by equivalent elements that appear after the present disclosure.
Claims
1. A method for detecting the welding quality of a Star Flash motherboard, characterized in that: The method comprises the following steps: Use a camera to capture a high-definition image of the motherboard to be inspected, identify each solder joint area in the high-definition image, and obtain a solder joint area image; Extracting solder joint features from each solder joint area image, and inputting the solder joint features into a defect prediction model; wherein the solder joint features include geometric features, texture features, color features, and solder joint image frequency domain features; The defect prediction model outputs a predicted probability value related to internal defects in solder joints. When the predicted probability value is higher than a probability threshold, it is determined that there are soldering quality defects in the mainboard; wherein the internal defects refer to internal cracks and / or bubbles.
2. A method for detecting the welding quality of a Star Flash motherboard according to claim 1, characterized in that: The method of using a camera to capture a high-definition image of the motherboard to be inspected, identifying each solder joint area in the high-definition image, and obtaining a solder joint area image includes: Using a camera to capture a first high-definition image of the motherboard to be inspected at a first focal length, binarizing the first high-definition image, and identifying the size and structure type of each bright spot area in the binarized image of the first high-definition image; the structure type includes a shadow type and a shadow-free type; Determine a second focal length based on the size of each bright spot area, and use a camera to take at least one second high-definition image of each bright spot area to be inspected at the second focal length as the solder joint area image; the second focal length is positively correlated with the size, and the second focal length is greater than the first focal length; Wherein, when the structural type of the bright spot area is a shadow-free type, the number of the second high-definition images is one; when the structural type of the bright spot area is a shadow-containing type, the number of the second high-definition images is at least two.
3. A method for detecting the welding quality of a Star Flash motherboard according to claim 2, characterized in that: Before binarizing the first high-definition image, the method further includes: Obtain the motherboard color and welding material type of the motherboard to be inspected in this batch through the production management system, and determine the soldering point color of the motherboard to be inspected according to the welding material type; If the mainboard color is a dark type and the solder joint colors are all light types, a binarization processing signal is generated, and the binarization processing signal is used to trigger binarization processing of the first high-definition image.
4. A method for detecting the welding quality of a Star Flash motherboard according to claim 1, characterized in that: The step of extracting solder joint features from each solder joint area image and inputting the solder joint features into a defect prediction model comprises: Extracting geometric features, texture features, and color features of the solder joints from the solder joint area images; The solder joint area image is converted into a frequency domain representation by Fourier transform, and the frequency domain component analysis is performed on the converted frequency domain image to obtain the frequency domain features of the solder joint image, wherein the frequency domain features of the solder joint image include frequency peak features and energy spectrum features; The geometric features, the texture features, the color features and the frequency domain features of the solder joint image are constructed into a feature matrix, that is, the solder joint features are obtained, and the solder joint features are input into a defect prediction model.
5. A method for detecting the welding quality of a Star Flash motherboard according to claim 4, characterized in that: The frequency domain features of the welding spot image also include at least one of the width and shape of the spectrum, the frequency response function features, the abnormal signal features in the frequency domain, and the statistical features of the frequency domain features.
6. A method for detecting the welding quality of a Star Flash motherboard according to claim 1, characterized in that: The defect prediction model is specifically: The defect prediction model is constructed based on ConvNeXt-YOLO5, and the defect prediction model includes a backbone network, a feature extraction network and a detection network; the feature extraction network includes a backbone network attention extraction layer, a multi-scale feature layer with an attention module and an adaptive spatial feature fusion layer; Among them, ECA modules are respectively set at the output positions of the reciprocal N layers of the backbone network and the multi-scale feature layer; and an adaptive fusion module is also set after the ECA module of the multi-scale feature layer, and the adaptive fusion module is used for weighted summation of multiple feature information outputs of the multi-scale feature layer.
7. A Xingshan motherboard welding quality detection system, characterized by: The system includes a solder joint recognition module, a feature extraction module, and an internal defect analysis module; The solder joint identification module is used to use a camera to capture a high-definition image of the motherboard to be inspected, identify each solder joint area in the high-definition image, and obtain a solder joint area image; The feature extraction module is used to extract solder joint features from each solder joint area image and input the solder joint features into the defect prediction model; wherein the solder joint features include geometric features, texture features, color features and solder joint image frequency domain features; The internal defect analysis module is used for the defect prediction model to output a predicted probability value related to the internal defects of the solder joint. When the predicted probability value is higher than a probability threshold, it is determined that the mainboard has a soldering quality defect; wherein the internal defect refers to internal cracks and / or bubbles.
8. An electronic device, characterized in that: The electronic device comprises: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the computer program implements the method according to any one of claims 1 to 6 when executed by the processor.
9. A computer storage medium, characterized in that: The computer storage medium stores a computer program executable by a processor, and the computer program implements the method according to any one of claims 1 to 6 when executed by the processor.
10. A computer program product, characterized in that: The computer program product comprises a computer program executable by a processor, and the computer program implements the method according to any one of claims 1 to 6 when being executed by the processor.
Citation Information
Patent Citations
Two-dimensional detection method for printing quality of PCB solder paste
CN107966448A
Weld defect detection method and system and storage medium
CN116309409A
Image processing method, device and system
CN117915213A
Welding spot defect detection method and system and storage medium
CN118212206A
Air cleaner for wall installation
KR102206156B1
Cited By
Intelligent welding defect detection method and device for explosion-proof valve of battery cover plate
CN120765570A
