Ceramic capacitor defect detection method and device, terminal and storage medium

Abstract

The invention relates to the field of quality detection of electronic components, and discloses a defect detection method and device for a ceramic capacitor, a terminal and a storage medium, and the method comprises the following steps: obtaining a CT image and an appearance image of a ceramic capacitor tray; segmenting and marking a defect area based on the CT image, and obtaining the type, position and size of a defect; constructing a mechanical stress field model of the internal defect of the ceramic capacitor based on an elastic mechanics theory, and calculating a stress concentration coefficient of the defect; constructing an electric field distribution model based on dielectric constant distribution in the ceramic capacitor, and calculating the influence degree of defects on electric field consistency; calculating a comprehensive defect score through a comprehensive scoring model in combination with mechanical analysis and electric field analysis results; and judging the qualification of the ceramic capacitor according to the scoring result, and screening unqualified products. According to the method, multi-dimensional analysis and quantification of defect detection can be realized, the method is suitable for automatic detection in batch production, and the quality control level and the production efficiency of the ceramic capacitor are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of quality inspection of electronic components, and particularly to a method, device, terminal and storage medium for defect detection of ceramic capacitors. Background Art

[0002] Ceramic capacitors are key components widely used in electronic devices, mainly for filtering, decoupling and noise suppression in circuits. However, due to the brittleness of ceramic materials and their complex manufacturing processes, ceramic capacitors are prone to various internal defects during production, such as cracks, voids, delamination, etc. These defects may cause the ceramic capacitors to fail during use, thereby affecting the performance and reliability of electronic devices, and even potentially causing serious safety hazards.

[0003] Existing ceramic capacitor defect detection technologies usually use withstand voltage testing or appearance inspection to identify unqualified products. However, the withstand voltage test can only detect defects that have seriously affected the capacitor performance and cannot effectively identify potential hidden defects. In addition, the appearance inspection is mainly used to judge whether there are visible defects on the surface of the capacitor, and in-depth analysis of internal defects cannot be carried out. This detection method is prone to missing detection of minor but potentially dangerous defects, thus allowing unreliable products to enter the market. In addition, existing detection technologies usually only evaluate defects in a single dimension (such as mechanical properties or electrical properties), and do not comprehensively consider the comprehensive impact of defects on the overall performance of ceramic capacitors.

[0004] Since the production of ceramic capacitors usually adopts large-scale assembly line operation, the existing detection methods are difficult to meet the requirements of modern production in terms of efficiency. For example, the method of detecting ceramic capacitors one by one results in low detection efficiency and cannot achieve the goal of batch detection. At the same time, the lack of support for automated analysis and screening technologies further limits the detection accuracy and speed. These problems are particularly prominent in the current electronic component market, and there is an urgent need for a detection method that can quickly detect internal defects and comprehensively evaluate the impact of defects in multiple dimensions. Summary of the Invention

[0005] Aiming at the deficiencies of the prior art, the present invention provides a method, device, terminal and storage medium for defect detection of ceramic capacitors, which can efficiently and accurately identify the internal defects of ceramic capacitors and comprehensively evaluate the impact of defects on their mechanical properties and electrical properties, so as to realize batch and automated defect detection and screening.

[0006] To achieve the above objectives, the present invention is realized through the following technical solutions: A method for defect detection of ceramic capacitors, comprising the following steps: Obtain the CT image and appearance image of the ceramic capacitor tray; Segment and label the defect area based on CT images to obtain the type and location of the defects; Construct a mechanical stress field model for the internal defects of ceramic capacitors based on the theory of elasticity, and calculate the stress concentration coefficient of the defects; based on the dielectric constant distribution inside the ceramic capacitors, construct an electric field distribution model, and calculate the degree of influence of the defects on the electric field uniformity; through a comprehensive scoring model, combine the results of mechanical stress analysis and electric field distribution analysis to calculate the comprehensive defect score of each ceramic capacitor; Determine the qualification of ceramic capacitors according to the comprehensive defect score, and screen out unqualified products.

[0007] Preferably, the acquisition steps of the CT images and appearance images include: Use industrial CT equipment to acquire three-dimensional internal images of ceramic capacitor trays; Use a high-resolution camera to acquire the appearance images of the ceramic capacitors in the tray; Perform edge detection and texture feature extraction on the CT images to segment potential defect areas; Perform salient object recognition on the appearance images to determine the coordinates and pose information of each ceramic capacitor.

[0008] Preferably, the steps for constructing the mechanical stress field model based on the theory of elasticity include: Perform mesh division on the defect area of the ceramic capacitor; Apply an external compressive load simulating the working state of the capacitor; Calculate the local stress concentration in the defect area through finite element analysis, and obtain the stress concentration coefficient of the defect; When the stress concentration coefficient of the defect is greater than the preset threshold, mark the defect as a dangerous area.

[0009] Preferably, the steps for constructing the electric field distribution model based on the dielectric constant distribution inside the ceramic capacitor include: Construct a partial differential equation model for describing the electric potential distribution according to the internal dielectric constant distribution of the ceramic capacitor; Set boundary conditions, including the working voltage applied to the upper electrode and the grounded electric potential of the lower electrode; Solve the electric potential distribution through numerical calculation methods to obtain the actual electric field in the defect area; Calculate the degree of influence of the defect on the electric field uniformity by comparing the actual electric field and the ideal electric field in the defect area.

[0010] Preferably, the comprehensive scoring model calculates the comprehensive defect score based on the following parameters: the stress concentration coefficient of the defect and the degree of influence of the defect on the electric field uniformity; The comprehensive defect score is calculated by the following formula: D score =w1 ·SCF + w 2 ·E mismatch Among them, D score is the comprehensive defect score, w 1 and w 2 are weight parameters, SCF is the stress concentration factor of the defect, and E mismatch is the influence degree of the defect on the electric field uniformity. The SCF and E mismatch are calculated respectively by the following formulas: Among them, σ max is the maximum stress in the defect area, and σ nom is the nominal stress of the whole ceramic capacitor; Among them, V in is the applied working voltage, E defect is the actual electric field in the defect area, E ideal is the ideal electric field in the defect-free area, and Ω defoct is the integration domain of the defect area.

[0011] Preferably, the determination steps for the qualification of the ceramic capacitor include: Comparing the comprehensive defect score with a preset score threshold; When the comprehensive defect score is higher than the preset threshold, determining that the ceramic capacitor is a non-conforming product; Marking the non-conforming products and removing the non-conforming products from the production line through an automated device.

[0012] Preferably, the defect types include cracks, voids, delaminations, and nodules, and are classified by the texture features of CT images; among them, the texture features are calculated based on pixel gray-scale distribution and edge characteristics.

[0013] The present invention also provides a defect detection device for a ceramic capacitor, including: An acquisition module for acquiring the CT image and the appearance image of the ceramic capacitor tray; An identification module for defect segmentation and annotation based on the CT image, and obtaining the defect type and location; A mechanical analysis module for constructing a mechanical stress field model of the defect and calculating the stress concentration factor of the defect; An electric field analysis module for constructing an electric field distribution model and calculating the influence degree of the defect on the electric field uniformity; A screening module for determining the qualification of the ceramic capacitor according to the comprehensive scoring model and marking the non-conforming products.

[0014] The present invention also provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the above-mentioned method is implemented.

[0015] The present invention also provides a storage medium, on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned method is implemented.

[0016] The present invention provides a method, device, terminal, and storage medium for defect detection of ceramic capacitors. It has the following beneficial effects: 1. The present invention combines industrial CT images and high-resolution appearance images. Through a high-precision segmentation algorithm and machine learning annotation technology, it can accurately identify tiny defects (such as cracks, voids, delamination, etc.) in ceramic capacitors and extract their type, position, and size information. By constructing a mechanical stress field and an electric field distribution model, it comprehensively evaluates the specific impact of defects on the capacitance performance, significantly improving the accuracy of defect detection.

[0017] 2. The present invention simultaneously models and quantitatively analyzes the impact of defects on ceramic capacitors from two dimensions: mechanics and electrical performance. Through the calculation of the stress concentration coefficient and the degree of electric field mismatch, it comprehensively evaluates the dual impact of defects on the structural integrity and electric field consistency of the capacitor. Compared with single-dimensional detection methods, the evaluation results of the present invention are more comprehensive and reliable.

[0018] 3. By optimizing the data acquisition process, introducing an automated scoring model, and linking mechanical devices, the present invention can achieve rapid detection of multiple ceramic capacitors in a tray and automatic rejection of unqualified products, meeting the on-line quality control requirements of large-scale production lines, effectively improving the detection efficiency and reducing production costs.

[0019] The comprehensive score and detailed defect data generated by the present invention are both output in a structured form, including the position, type, score value, and screening status of the defects. This output is not only used for real-time screening but can also be stored to support subsequent quality traceability and production improvement, helping to discover potential problems in the process and optimize the production process. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a schematic flow chart of the method of the present invention; Figure 2 is a schematic structural diagram of the device of the present invention; Figure 3 is a schematic structural diagram of the computer device of the present invention.

[0021] Among them, 100 is an acquisition module; 200 is an identification module; 300 is a mechanical analysis module; 400 is an electric field analysis module; 500 is a screening module; 40 is a computer device; 41 is a processor; 42 is a memory; 43 is a storage medium. Detailed implementation manner

[0022] Please refer to the attached Figure 1 The present invention provides a method for detecting defects of ceramic capacitors, which can efficiently and accurately detect internal defects of ceramic capacitors and evaluate their impact on capacitor performance.

[0023] As Figure 1 shown, the method for detecting defects of the ceramic capacitor may include the following steps: S1. Obtain the CT image and appearance image of the ceramic capacitor tray; S2. Based on the CT image, perform segmentation and annotation of the defect area to obtain the type and location of the defect; S3. Based on the theory of elasticity mechanics, construct a mechanical stress field model of internal defects of the ceramic capacitor, and calculate the stress concentration coefficient of the defect; S4. Based on the dielectric constant distribution inside the ceramic capacitor, construct an electric field distribution model, and calculate the influence degree of the defect on the electric field uniformity; S5. Through a comprehensive scoring model, combine the results of mechanical stress analysis and electric field distribution analysis to calculate the comprehensive defect score of each ceramic capacitor; S6. Determine the qualification of the ceramic capacitor according to the comprehensive defect score, and screen out unqualified products.

[0024] The following will respectively elaborate on each step of the method of the present invention in detail.

[0025] For step S1, in this embodiment, step S1 mainly involves obtaining the CT image and appearance image of the ceramic capacitor tray, providing basic data support for the identification, modeling, and evaluation of ceramic capacitor defects in subsequent steps.

[0026] Specifically, in this embodiment, first, the ceramic capacitor tray is scanned by an industrial CT device to obtain three-dimensional internal structure information of the ceramic capacitor. The CT device needs to have high-resolution performance and be able to capture fine defects with a size of micron level, and the resolution should preferably be set to ≤1μm. To ensure the scanning quality, during the scanning process, the ceramic capacitor tray needs to be fixed on the rotating platform of the CT device, and multi-view projection data is gradually obtained through 360-degree rotation. As an option, the filtered back projection algorithm (Filtered Back Projection, FBP) can be used to reconstruct the CT scan data to generate a three-dimensional volume image, and the image data format can adopt DICOM or other common industrial standard formats.

[0027] It should be noted that, in a possible implementation, to improve the data acquisition efficiency, CT scans can be performed on multiple ceramic capacitors in a ceramic capacitor tray simultaneously. For example, if 10 ceramic capacitors are placed on a tray, the three-dimensional structure information of all ceramic capacitors can be obtained simultaneously through one scan, thus improving the detection efficiency.

[0028] In some embodiments, to facilitate subsequent processing, it is also necessary to obtain the two-dimensional appearance image of the ceramic capacitor. Specifically, a high-resolution industrial camera can be used to capture the appearance image of the ceramic capacitor in the tray, and the shooting resolution is preferably set to 0.01 mm to ensure that the external contour features of the ceramic capacitor can be clearly captured. The appearance image is mainly used to calibrate the position and coordinates of each ceramic capacitor in the tray, providing a basis for the spatial positioning of ceramic capacitor defects in subsequent steps.

[0029] It should be noted that, during the process of obtaining the appearance image, a saliency object detection algorithm can be used to identify and mark the ceramic capacitors in the appearance image. As an option, object detection can be achieved through a convolutional neural network (CNN). For example, the YOLO or Faster R-CNN model can be used to detect the ceramic capacitor objects in the image, generating corresponding anchor boxes and position coordinates.

[0030] Exemplarily, for the results output by object detection, it includes the two-dimensional coordinates (such as x, y) of each ceramic capacitor and the corresponding anchor box area. These data can be combined with the three-dimensional structure data of the CT image to provide support for the positioning accuracy in subsequent defect identification steps.

[0031] In a possible implementation, the CT image and the appearance image need to be aligned spatially. For this purpose, image registration can be achieved through the calibration points on the tray. For example, multiple calibration points are set at fixed positions on the ceramic capacitor tray, and these calibration points can appear in both the CT image and the appearance image simultaneously, thus realizing the spatial registration of the two images. Specifically, affine transformation or other geometric transformation methods can be used for registration calculation to ensure that the defects detected in the CT image can be accurately corresponding to the position information in the appearance image.

[0032] It can be understood that, in this embodiment, by obtaining the CT image and the appearance image, a high-quality data basis is provided for the segmentation and annotation of the defect area in subsequent steps. Among them, the CT image is used to identify the internal defects of the ceramic capacitor, while the appearance image is used to accurately locate the position and pose of each ceramic capacitor, thus realizing the organic combination of internal structure analysis and external space calibration.

[0033] It should be further noted that the acquisition of CT scan data and appearance image data can be carried out in parallel or in steps, depending on the actual requirements of pipeline inspection. For example, in some pipeline inspection scenarios, CT scanning can be completed at a dedicated scanning station, while the shooting of appearance images can be completed on the conveyor belt, thereby further improving the overall inspection efficiency.

[0034] This step lays the foundation for the subsequent inspection of ceramic capacitor defects by combining the dual data sources of CT images and appearance images, and ensures the optimal balance between data quality and efficiency.

[0035] For step S2, in this embodiment, the core content of step S2 is to segment and label the defective areas of the ceramic capacitor based on the CT image data, and further extract the types and spatial positions of the defects, so as to provide data support for subsequent mechanical modeling and performance evaluation.

[0036] Specifically, this step first preprocesses the CT image obtained in step S1 to improve the contrast and clarity of the image. The preprocessing includes denoising, enhancement, and gray normalization operations. As an option, the Gaussian filtering method can be used to smooth the CT image to remove random noise in the image while retaining edge details. It should be noted that the preprocessed CT image is the basis for subsequent segmentation and labeling, and its quality directly affects the accuracy of defect recognition.

[0037] In a possible implementation, an edge detection algorithm is used to extract the boundaries of the defective areas from the preprocessed CT image. Exemplarily, the Canny edge detection algorithm can be used, which detects edges by calculating the gradient change rate of the image and can accurately separate the boundaries of internal defects. The detected boundary data is stored in the form of pixel points and is used for subsequent region segmentation.

[0038] Specifically, based on the boundary data, a region growing algorithm based on gray threshold is used to segment the defective areas of the CT image. The region growing algorithm starts from the boundary pixels and gradually expands to adjacent pixels according to the similarity of gray values until all pixels that meet the threshold conditions are labeled as the same defective area. It should be noted that the setting of the gray threshold can be obtained through experimental calibration to ensure that the segmentation result can cover all possible defective areas.

[0039] In some embodiments, in order to further improve the segmentation accuracy, the watershed algorithm can be combined to optimize the region segmentation result. The watershed algorithm is based on the gradient distribution of the image and further refines the segmentation region, which is particularly suitable for the accurate segmentation of defects with complex shapes. For example, for the case of overlapping boundaries or adjacent defects, the watershed algorithm can effectively distinguish the overlapping regions and label them as independent defects.

[0040] It should be noted that after the defect area is segmented, it is necessary to label the defect area, including the type, location and size of the defect. Exemplarily, the defect type can be automatically labeled through texture feature extraction and machine learning classification models. Specifically, the defect texture features can include contrast, entropy and uniformity features in the gray-level co-occurrence matrix (GLCM). These features are input into a trained support vector machine (SVM) model, and the class labels of the defects are output, including cracks, voids, delaminations, nodules, etc.

[0041] As an option, the classification accuracy can be further improved by combining a convolutional neural network (CNN) on the basis of the SVM classification model. By automatically extracting deep features, CNN can effectively distinguish defect types with similar textures but different morphologies. Exemplarily, for crack defects with a long length and void defects with an approximately spherical shape, CNN can accurately output their classification results.

[0042] In a possible implementation, the spatial position and geometric size of the labeled defect area are extracted, including parameters such as the center point coordinates (x, y, z), length, width and depth. The position coordinates are directly obtained by parsing the three-dimensional volume data of the segmented area, while the size data is obtained by calculating the boundary range of the segmented area.

[0043] It should be further noted that the spatial position of the defect is represented in a reference coordinate system relative to the whole ceramic capacitor, and this reference coordinate system is calibrated by a fixed point during CT scanning. To ensure the position accuracy, the segmentation result of the CT image can be aligned with the position information in the appearance image, and the coordinate differences between the two data sources are corrected through affine transformation.

[0044] In some embodiments, for the convenience of implementing modeling in subsequent steps, three-dimensional visualization processing is also required for the labeled defect area. For example, the segmented defect area is labeled with different colors and superimposed on the CT image for display, so as to visually verify the accuracy of segmentation and labeling.

[0045] It can be understood that in this step, through segmentation and labeling, structured defect data is finally output, including the type, three-dimensional position and size of the defect. These data will be directly used to construct mechanical and electric field models in subsequent steps and provide important basic data support for comprehensive performance evaluation.

[0046] For step S3, in this embodiment, the core of step S3 is to utilize the theory of elastic mechanics and combine the finite element analysis method to construct a mechanical stress field model for the internal defect area of the ceramic capacitor, and calculate the stress concentration factor (SCF) of the defect based on this model to evaluate the structural risk that the defect may pose to the capacitor performance.

[0047] Specifically, this step takes the defect data obtained from segmentation and annotation in step S2 as input, including information such as the three-dimensional shape, position, and size of the defects. These data are first imported into finite element analysis software (such as ANSYS or COMSOL) to generate an overall finite element model of the ceramic capacitor. As an option, the ceramic capacitor model should include the complete capacitor geometric structure and the segmented defect regions, and the mesh division of the defect regions needs to be refined to ensure the modeling accuracy.

[0048] In a possible implementation, to improve the calculation efficiency, this step adopts a non-uniform mesh division strategy for the model. Exemplarily, for the overall structure of the ceramic capacitor, a coarser mesh division can be used, while for the defect regions and their surrounding areas, a high-density mesh division is required, and the mesh element size is preferably set to ≤0.01 mm. This hierarchical mesh division strategy can reduce the use of computing resources while ensuring the calculation accuracy.

[0049] It should be noted that the material parameters of the ceramic capacitor are crucial for the construction of the mechanical model. In this embodiment, the typical values of the elastic modulus E and Poisson's ratio v of the ceramic material are respectively set to E = 120 GPa and v = 0.25. For the defect regions, different material properties are set according to the defect types. For example, for void defects, the elastic modulus is approximately set to zero; for crack defects, their boundaries can be set as weak contact interfaces to simulate the behavior of crack closure or propagation.

[0050] In this embodiment, the setting of the load conditions and boundary conditions is an important part of constructing the mechanical model. Specifically, a uniform compressive load is applied to the top of the ceramic capacitor to simulate the mechanical stress caused by voltage under the actual working conditions. The magnitude of the compressive load is determined according to the rated voltage and actual use conditions of the capacitor, and can be set to 10 MPa for example. At the same time, the bottom of the ceramic capacitor is set as a fixed support to provide the boundary constraint conditions for the mechanical model.

[0051] In some embodiments, to more accurately simulate the stress concentration phenomenon, the load direction needs to be combined with the geometric distribution of the defect regions. For example, for the case where a crack propagates in a certain direction, the load application direction can be perpendicular to the crack plane to maximize the accuracy of the stress concentration region.

[0052] It should be further noted that the stress concentration factor in this embodiment is calculated by the following formula: where σ max represents the maximum stress in the defect region, and σ nom represents the nominal stress of the overall structure of the ceramic capacitor.

[0053] In a specific implementation, the finite element analysis software can directly output the stress distribution in the defect area. By analyzing the stress distribution diagram, the maximum stress value σ is extracted. max Meanwhile, the nominal stress value σ is calculated through the overall load distribution. nom As an option, an isogram of the stress distribution can be used to visually display the stress concentration phenomenon in the defect area, and the position of the maximum stress point is marked on the diagram.

[0054] Exemplarily, for a spherical cavity defect, when a compressive load of 10 MPa is applied to the top of the ceramic capacitor, the finite element analysis results show that the maximum stress value at the defect boundary is σ max = 30 MPa, and the nominal stress value is σ nom = 30 MPa. Therefore, the stress concentration factor is calculated as: It should be particularly noted that the value of SCF can directly reflect the danger of the defect. In this embodiment, the set danger threshold is SCF threshold = 3.0. When SCF > 3.0, the defect is determined to be a high-risk defect.

[0055] In some embodiments, for the convenience of comprehensive performance evaluation in subsequent steps, the analysis results of this step are output in the form of structured data, including key information such as defect location, defect type, stress concentration factor, and maximum stress position. Meanwhile, the stress distribution in the defect area can be displayed through a three-dimensional visualization tool, providing an intuitive analysis reference for engineering personnel.

[0056] It can be understood that in this step, based on the finite element analysis of the elastic mechanics theory, an internal stress field model of the ceramic capacitor is constructed, and the stress concentration factor of the defect is accurately calculated. These results provide an important mechanical basis for the electric field distribution analysis and comprehensive performance evaluation in subsequent steps.

[0057] For step S4, in this embodiment, the core of step S4 is to simulate the influence of the defect on the internal electric field uniformity of the ceramic capacitor and quantify the degree of this influence based on the dielectric constant distribution inside the ceramic capacitor by constructing an electric field distribution model, providing a basis for subsequent performance evaluation and screening.

[0058] Specifically, this step constructs an electric field distribution model based on the defect area data marked in step S2 and the overall structure of the ceramic capacitor. In a possible implementation, the electric field distribution model is represented by a partial differential equation of the electric potential distribution, and its control equation is: Where: ε is the spatial distribution of the dielectric constant, including the dielectric constant ε of the complete ceramic materialnormal and the dielectric constant ε of the defect region d e f ec t ; φ is the electric potential distribution function; is the gradient operator.

[0059] It should be noted that the dielectric constant ε of the ceramic capacitor normal is usually in the range of 1000 - 5000, while the dielectric constant of the defect region depends on the specific type of the defect. For example, for a void defect, the inside is air and the dielectric constant is approximately ε air = 1; for a crack defect, its dielectric constant value can be set as ε crack << ε normal .

[0060] In a possible implementation, this step uses the finite element analysis (FEA) or the finite difference method (FDM) to solve the above partial differential equation. Specifically, the ceramic capacitor model is divided into multiple finite elements, and a higher density of mesh division is used for the defect region to capture the local influence of the defect on the electric field distribution. The size of the mesh element is preferably set to ≤0.01 mm to ensure the accuracy of the calculation results.

[0061] As an option, the boundary conditions of the electric field model in this embodiment are set as follows: Apply the working voltage V in to the upper electrode of the ceramic capacitor, and its typical value can be taken as 100 - 1000 V to simulate the actual working conditions; The lower electrode is set to be grounded and the electric potential value is zero.

[0062] Specifically, by solving the above electric potential distribution equation, the electric field distribution E inside the ceramic capacitor can be obtained, and its calculation formula is: where E is the electric field strength vector, representing the electric field direction and magnitude of each unit.

[0063] In some embodiments, in order to analyze the specific influence of the defect on the electric field distribution, it is necessary to calculate the actual electric field E defect of the defect region and the electric field E ideal when there is no defect ideally. The ideal electric field distribution is simulated by ignoring the dielectric constant difference of the defect region.

[0064] It should be noted that the influence degree of the defect on the electric field consistency is quantified by the following formula: where: Emismatch Indicates the degree of electric field mismatch; Ω defoct is the integration domain of the defect area; E defect is the actual electric field of the defect area; E ideal is the ideal electric field in the case of no defect.

[0065] In a possible implementation, the integral calculation is completed by numerical methods, specifically including accumulating the electric field values of all finite elements in the defect area and performing normalization processing.

[0066] Exemplarily, for a spherical cavity defect contained in a ceramic capacitor, at the operating voltage V in = 500V, the actual electric field distribution E defect of the defect area is obtained through finite element analysis, and its maximum value is 1.2×10 6 V / m, while the maximum value of the ideal electric field value E ideal is 1.0×10 6 V / m. According to the above formula calculation, the degree of electric field mismatch is: It should be noted that the magnitude of E mismatch can directly reflect the potential impact of the defect on the capacitance performance. In this embodiment, the determination threshold of the degree of electric field mismatch can be set according to actual needs. For example, when E mismatcg > 0.5V / m, it is considered that the defect has a significant impact on the capacitance performance.

[0067] In some embodiments, for the convenience of visualizing the analysis results, this step displays the electric field distribution through an electric field contour map or a vector map. For example, by superimposing electric field vectors in a three-dimensional model, the degree of perturbation of the defect on the local electric field direction and intensity can be intuitively observed. This visual analysis can not only verify the rationality of the calculation results, but also provide a basis for subsequent defect evaluation and optimization.

[0068] It can be understood that this step quantitatively analyzes the influence degree of the defect on the performance of the ceramic capacitor by constructing and solving the electric field distribution model, laying a foundation for the calculation of the comprehensive performance score in the subsequent steps.

[0069] For step S5, in this embodiment, the core of step S5 is to quantify the defect influence of each ceramic capacitor through a comprehensive scoring model based on the mechanical stress analysis result in step S3 and the electric field distribution analysis result in step S4, and generate a comprehensive defect score D score . The comprehensive defect score is used to evaluate the overall impact of the defect on the performance of the ceramic capacitor, providing a direct determination basis for subsequent screening.

[0070] Specifically, in this step, the comprehensive defect score is calculated by the following formula: D score = w 1 ·SCF + w 2 ·E mismatcg Where: D score represents the comprehensive defect score; w 1 and w 2 are the weight parameters for mechanical stress analysis and electric field mismatch analysis, respectively; SCF is the stress concentration factor calculated in step S3; E mismatch is the degree of electric field mismatch calculated in step S4.

[0071] As an option, the settings of the weight parameters w 1 and w 2 can be obtained by calibrating with experimental data. Specifically, based on the actual working environment of the ceramic capacitor and the failure cases of typical defects, the values of w 1 and w 2 can be determined through multiple tests. For example, in some application scenarios of ceramic capacitors, if the failures caused by mechanical stress account for a relatively high proportion, the value of w 1 can be appropriately increased; for capacitors under high-voltage working conditions, the weight of w 2 needs to be increased.

[0072] In a possible implementation, the calculation process of the comprehensive defect score is divided into two main stages: First, combining the analysis results of step S3, the stress concentration factor SCF of each defect is extracted. It should be noted that the magnitude of SCF directly reflects the stress concentration degree in the defect area under the action of external loads. For example, the SCF of crack defects is usually significantly higher than that of void defects and is closely related to the length, width, and direction of the crack.

[0073] Second, combining the electric field analysis results of step S4, the degree of electric field mismatch E mismatch of each defect is calculated. It can be understood that the value of E mismatch reflects the degree of perturbation of the electric field distribution of the ceramic capacitor by the defect. For void defects, the electric field mismatch is mainly manifested as a decrease in the local electric field intensity, while crack defects may cause a significant deviation in the electric field direction.

[0074] It should be noted that in order to ensure the accuracy of the comprehensive score, the results of mechanical analysis and electric field analysis are standardized in this embodiment. Exemplarily, for SCF and E mismatch, the following standardization formulas can be used respectively: Among them, SCF max and E mismatch,max are respectively the maximum values in the corresponding datasets, used to eliminate the differences between different data scales, so as to ensure the rationality of the comprehensive score calculation.

[0075] Exemplarily, for a certain ceramic capacitor, the SCF of its defect is 3.2 and E mismatch = 0.4. Assuming that SCF max = 5.0 and E mismatch,max = 1.0. After standardization, we get: Then, assuming that w 1 = 0.7 and w 2 = 0.3, then the calculation of its comprehensive defect score is: D score = 0.7·0.64 + 0.3·0.4 = 0.448 + 0.12 = 0.568 In some embodiments, the result of the comprehensive score needs to be compared with a preset determination threshold D threshold . It should be noted that the value of D threshold can be determined according to the specific application scenario and quality standard. For example, in some applications with high reliability requirements, D threshold can be set to 0.5; while in applications with general reliability requirements, the threshold can be appropriately increased.

[0076] It can be understood that when D score ≥ D threshold , it is determined that the ceramic capacitor is a non-conforming product; otherwise, it is a conforming product. In this embodiment, the calculation of the comprehensive score not only integrates the multi-dimensional analysis results of mechanical stress and electric field mismatch, but also realizes a unified evaluation standard through standardization and weighting processing, so as to more comprehensively and accurately evaluate the comprehensive impact of defects on the performance of ceramic capacitors.

[0077] In a possible implementation manner, the comprehensive score result will be output in the form of structured data, including the D score value of each ceramic capacitor and its corresponding determination result. These data can be directly transmitted to the subsequent screening system to provide a basis for automatically screening non-conforming products.

[0078] It should be further noted that the implementation of this step depends on the analysis results of the foregoing steps. Its core lies in realizing multi-dimensional defect evaluation through a quantitative comprehensive score model, laying a foundation for product screening in the subsequent steps.

[0079] For step S6, in this embodiment, the core of step S6 is to utilize the comprehensive defect score D calculated in step S5 score , and in combination with a preset score determination threshold D threshold , determine the qualification of the ceramic capacitor, and complete the screening of unqualified products through an automated device.

[0080] Specifically, the comprehensive defect score D of each ceramic capacitor score is calculated by the aforementioned comprehensive scoring model, and the numerical range and distribution of the score depend on the actual quality status of the ceramic capacitor and the weight setting of the scoring model. In this embodiment, the setting of the determination threshold D threshold is crucial, which directly determines the division standard between qualified and unqualified products.

[0081] It should be noted that the determination logic of this step can be expressed by the following formula: In a possible implementation manner, in this step, according to the above determination logic, the determination result of each ceramic capacitor will be bound to its position number, and structured output data will be generated. The output data includes the number of each ceramic capacitor, the comprehensive defect score D score , the determination result, and the screening status. These data will be transmitted to the automated control system for the next screening operation.

[0082] It can be understood that the screening operation is an important part of this step, and its goal is to remove unqualified products from the production line according to the determination result. Specifically, the implementation of the automated screening device can adopt a robotic arm, a pneumatic sorting device, or other suitable mechanical equipment. For example, in one implementation manner, the number and position of each ceramic capacitor on the conveyor belt will be aligned with the aforementioned calibration data through appearance image recognition, the position of the unqualified product will be identified by the automatic control system, and then it will be grabbed and removed by the robotic arm.

[0083] As an option, in the scenario of in-line inspection, the screening rhythm can be adjusted in combination with the speed of the conveying system. For example, if the number of ceramic capacitors detected each time is large (such as 10 ceramic capacitors in a tray), the screening device can complete the screening operation batch by batch to optimize production efficiency.

[0084] It should be further noted that in some embodiments, the screening operation is not limited to removing unqualified products, but can also classify and record unqualified products. For example, classify unqualified products according to the type or severity of the defects. A possible implementation manner is to divide D score into several intervals based on the distribution range of the comprehensive defect score. For example: If D score≥0.8, considered as a serious defect; If 0.5≤D score <0.8, judged as moderate defect; If D score <0.5, considered as a slight defect.

[0085] For example, products with serious defects may be directly rejected; products with moderate defects may be reworked; and products with slight defects may be allowed to pass the quality standards of specific applications.

[0086] It is understandable that the implementation of this step needs to ensure both the accuracy of the determination results and the efficiency and stability of the screening operation. In the design of the screening device, the shape, size and specific characteristics of the ceramic capacitors as well as the tray or conveyor belt should be fully considered. For example, for smaller ceramic capacitors, a vacuum adsorption robot arm can be used to complete the grasping; for larger batches of testing, multiple robots can be used to work together.

[0087] It should be noted that the implementation result of this step directly affects the final quality screening effect of ceramic capacitors, and its success depends on the accuracy of the aforementioned scoring model and the execution capability of the screening device. This embodiment achieves efficient and reliable rejection of unqualified products through the organic combination of scoring results and automated screening, providing a solid guarantee for the production quality control of ceramic capacitors.

[0088] In general, the present invention combines the acquisition and analysis of CT images and appearance images, constructs a mechanical stress field model based on the theory of elastic mechanics, and constructs an electric field distribution model based on the dielectric constant distribution inside the ceramic capacitor, respectively calculating the stress concentration factor and electric field mismatch degree of the defect, and finally quantifies the impact of the defect on the capacitor performance through a comprehensive scoring model, and determines the eligibility based on the score, thereby realizing the automatic screening of unqualified products. The present invention can efficiently and accurately detect the internal defects of ceramic capacitors, is suitable for mass production detection, and improves production efficiency and product quality reliability.

[0089] The method of the present invention focuses on the detection and screening of internal defects in ceramic capacitors, combines mechanical stress analysis with electric field distribution analysis, quantifies the impact of defects on capacitor performance through comprehensive scoring, and ultimately achieves the screening of unqualified products. The following is a description of the specific workflow of the method: 1. Data Acquisition The present invention first obtains the CT image and appearance image of the ceramic capacitor tray through industrial CT equipment and high-resolution camera. The CT image is used to record the internal three-dimensional structural information of the ceramic capacitor and capture the shape, position and size of internal defects such as cracks, voids and delamination. The appearance image is used to calibrate the two-dimensional coordinates and positional relationship of each ceramic capacitor to ensure the matching of internal data and appearance data in subsequent analysis.

[0090] The data acquisition stage has a batch characteristic, which can scan and process multiple ceramic capacitors on a tray simultaneously, thus laying a foundation for subsequent efficient detection.

[0091] 2. Defect Segmentation and Labeling After data acquisition, the defect regions of ceramic capacitors are segmented and labeled based on CT images. Image segmentation uses region growing algorithms and edge detection algorithms based on gray - level thresholds, combined with texture feature analysis techniques to accurately extract the boundaries of defect regions. Subsequently, machine learning algorithms are used to label the segmented defect regions, identify the types of defects (such as cracks, voids, or delaminations), and extract the three - dimensional spatial information of the defects, including position and size.

[0092] The innovation of this step lies in combining high - precision CT images and automated classification techniques, achieving high - precision identification of ceramic capacitor defects and providing high - quality input data for subsequent mechanical and electric - field modeling.

[0093] 3. Mechanical Stress Field Modeling Using the theory of elasticity and finite - element analysis methods, a mechanical stress field model of internal defects in ceramic capacitors is constructed. In model construction, based on the defect data obtained by segmentation and labeling, by refining the mesh division of the defect region, the stress concentration phenomenon in the defect region is accurately captured.

[0094] This step calculates the local stress value at the defect through finite - element analysis and compares it with the overall nominal stress value to quantify the stress concentration coefficient of the defect. This coefficient is an important index for evaluating the impact of defects on the mechanical properties of capacitors and directly reflects the danger of defects under working loads.

[0095] 4. Electric - Field Distribution Modeling After completing the mechanical modeling, an electric - field distribution model is further constructed based on the dielectric constant distribution inside the ceramic capacitor. This model simulates the influence of defects on the uniformity of the internal electric - field distribution of ceramic capacitors by solving the electric - potential distribution. The dielectric constant of the defect region is set according to the type of defect. For example, the inside of a void is air, resulting in a dielectric constant much lower than that of the complete ceramic material.

[0096] The model outputs the actual electric - field distribution in the defect region and quantifies the degree of perturbation of the defect to the electric - field uniformity by comparing it with the electric - field distribution in the ideal defect - free case. This modeling method innovatively introduces the actual influence of defects on the electric - field distribution and can comprehensively evaluate the potential threat of defects to the electrical properties of capacitors.

[0097] 5. Comprehensive Score Calculation After obtaining the analysis results of the mechanical stress field and the electric field distribution, the influence degree of the defect is quantitatively calculated based on the comprehensive scoring model. The comprehensive scoring model unifies the multi-dimensional defect influence into a scoring value by combining the stress concentration coefficient and the degree of electric field mismatch with weights.

[0098] The innovation of the comprehensive scoring lies in its multi-parameter fusion ability. It can not only consider the influences of both mechanical and electrical properties simultaneously, but also adapt to the requirements of different application scenarios by adjusting the weights. The scoring result directly reflects the threat degree of the defect to the overall performance of the ceramic capacitor.

[0099] 6. Qualification determination and screening According to the comprehensive scoring result, the score of each ceramic capacitor is compared with a preset scoring threshold. If the score is higher than the threshold, the capacitor is determined to be a non-conforming product; otherwise, it is a conforming product. The determination result is output in structured data, including the number, scoring value, and determination status of each capacitor.

[0100] The screening process is completed by an automated mechanical device. For example, a robotic arm is used to remove non-conforming products from the conveyor belt. In some scenarios, non-conforming products can also be classified and recorded (such as distinguishing severe, medium, and minor defects according to the score level). The innovation of this step lies in the linkage between the comprehensive scoring result and the automated screening device, realizing efficient and accurate removal of non-conforming products, and at the same time supporting further quality traceability and improvement.

[0101] The ceramic capacitor defect detection device described below can be correspondingly referred to the ceramic capacitor defect detection method described above.

[0102] Please refer to the appendix Figure 2 , the present invention also provides a ceramic capacitor defect detection device, including: An acquisition module 100 for acquiring the CT image and the appearance image of the ceramic capacitor tray; An identification module 200 for performing defect segmentation and annotation based on the CT image, and obtaining the defect type and location; A mechanical analysis module 300 for constructing a mechanical stress field model of the defect and calculating the stress concentration coefficient of the defect; An electric field analysis module 400 for constructing an electric field distribution model and calculating the influence degree of the defect on the electric field consistency; A screening module 500 for determining the qualification of the ceramic capacitor according to the comprehensive scoring model and marking non-conforming products.

[0103] The device in this embodiment can be used to execute the above method embodiment, and its principle and technical effects are similar, which will not be elaborated here.

[0104] Please refer to the appendix Figure 3, the present invention also provides a computer device 40, including: a processor 41 and a memory 42, where the memory 42 stores a computer program executable by the processor, and when the computer program is executed by the processor, it executes the method as described above.

[0105] The present invention also provides a storage medium 43, on which a computer program is stored, and when the computer program is run by the processor 41, it executes the method as described above.

[0106] Among them, the storage medium 43 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (abbreviated as SRAM), electrically erasable programmable read-only memory (abbreviated as EEPROM), erasable programmable read-only memory (abbreviated as EPROM), programmable read-only memory (abbreviated as PROM), read-only memory (abbreviated as ROM), magnetic memory, flash memory, magnetic disk or optical disc.

[0107] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for detecting defects in ceramic capacitors, characterized in that: The following steps are involved: Obtain CT images and appearance images of ceramic capacitor trays; Segment and mark defect areas based on CT images to obtain defect types and locations; Based on the theory of elastic mechanics, a mechanical stress field model of internal defects in ceramic capacitors was constructed to calculate the stress concentration coefficient of the defects. Based on the dielectric constant distribution inside the ceramic capacitor, an electric field distribution model is constructed to calculate the impact of defects on the consistency of the electric field. Through the comprehensive scoring model, combined with the results of mechanical stress analysis and electric field distribution analysis, the comprehensive defect score of each ceramic capacitor is calculated; The qualification of ceramic capacitors is determined based on the comprehensive defect score, and unqualified products are screened out.

2. The defect detection method of ceramic capacitor according to claim 1, characterized in that: The steps of collecting the CT image and the appearance image include: Use industrial CT equipment to acquire 3D internal images of ceramic capacitor trays; Use a high-resolution camera to capture the appearance of ceramic capacitors in a tray; Perform edge detection and texture feature extraction on CT images to segment potential defect areas; Perform salient object recognition on the appearance image to determine the coordinates and posture information of each ceramic capacitor.

3. The defect detection method of ceramic capacitor according to claim 1, characterized in that: The steps of constructing the mechanical stress field model based on elastic mechanics theory include: Mesh the defective area of ​​the ceramic capacitor; Apply an external compressive load simulating the working state of the capacitor; The local stress concentration in the defect area is calculated by finite element analysis, and the stress concentration coefficient of the defect is obtained; When the stress concentration factor of a defect is greater than a preset threshold, the defect is marked as a dangerous area.

4. The defect detection method of ceramic capacitor according to claim 1, characterized in that: The steps of constructing the electric field distribution model based on the dielectric constant distribution inside the ceramic capacitor include: According to the internal dielectric constant distribution of ceramic capacitors, a partial differential equation model is constructed to describe the potential distribution; Setting boundary conditions, including the working voltage applied to the upper electrode and the ground potential of the lower electrode; The potential distribution is solved by numerical calculation method to obtain the actual electric field in the defect area; By comparing the actual electric field in the defect area with the ideal electric field, the influence of the defect on the consistency of the electric field is calculated.

5. The defect detection method of ceramic capacitor according to claim 1, characterized in that: The comprehensive scoring model calculates the comprehensive defect score based on the following parameters: the stress concentration factor of the defect and the degree of influence of the defect on the electric field consistency; The comprehensive defect score is calculated by the following formula: D score =w1·SCF+w2·E mismatch Among them, D score is the comprehensive defect score, w1 and w2 are weight parameters, SCF is the stress concentration factor of the defect, E mismatch is the influence of defects on the electric field consistency, The SCF and E mismatch Calculated by the following formulas: Among them, σ max is the maximum stress in the defect area, σ nom is the nominal stress of the ceramic capacitor as a whole; Among them, V in is the applied working voltage, E defect is the actual electric field in the defect area, E ideal is the ideal electric field in the defect-free region, Ω defoct is the integration domain of the defect area.

6. The defect detection method of ceramic capacitor according to claim 1, characterized in that: The qualification determination steps of the ceramic capacitor include: comparing the composite defect score to a preset score threshold; When the comprehensive defect score is higher than the preset threshold, the ceramic capacitor is judged as a substandard product; Rejected products are marked and automatically removed from the production line.

7. The defect detection method of ceramic capacitor according to claim 1, characterized in that: The defect types include cracks, voids, delaminations and nodules, and are classified by texture features of CT images; wherein the texture features are calculated based on pixel grayscale distribution and edge characteristics.

8. A defect detection device for a ceramic capacitor, used to execute the defect detection method for a ceramic capacitor according to any one of claims 1 to 7, characterized in that: include: An acquisition module, used for acquiring a CT image and an appearance image of a ceramic capacitor tray; The recognition module is used to segment and annotate defects based on CT images and obtain defect types and locations; Mechanical analysis module, used to construct the mechanical stress field model of defects and calculate the stress concentration coefficient of defects; The electric field analysis module is used to construct an electric field distribution model and calculate the impact of defects on the consistency of the electric field; The screening module is used to determine the eligibility of ceramic capacitors based on a comprehensive scoring model and mark unqualified products.

9. A terminal comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images