A method for automatic imaging and defect detection of capacitors using multi-size screen frames

Abstract

This invention discloses an automatic imaging and defect detection method for capacitor wire mesh applicable to multiple sizes, belonging to the field of industrial automation inspection technology. The method includes the following steps: generating a corresponding imaging path based on the parameter information of the capacitor wire mesh frame to be inspected; controlling the imaging unit to acquire images of the capacitor wire mesh according to the imaging path; identifying regular rectangular wire mesh regions based on region positioning images, and constructing multiple region masks based on these regions; switching between foreground detection regions and background suppression regions between different region masks based on predefined defect types to be detected, and performing defect detection and classification within the corresponding foreground detection regions; generating an inspection report based on structured data; achieving refined partitioning detection through multi-region masks, effectively improving defect detection efficiency and avoiding human subjective errors, reducing the probability of missed and false detections, and accurately identifying various defects such as wire mesh blockage, expansion and contraction, latex peeling, and dirt.

Description

Technical Field

[0001] This invention relates to the field of industrial automation inspection technology, and in particular to an automatic imaging and defect detection method for capacitor wire mesh applicable to multiple sizes of wire frames. Background Technology

[0002] In capacitor manufacturing, the screen is a crucial mold used for printing electrodes. The dimensional accuracy and shape integrity of its mesh openings, as well as the presence of broken or clogged threads, directly determine the quality of the printed electrodes. Currently, defect detection of the screen relies heavily on offline, sampling inspection methods, primarily using tool microscopes or conventional image measuring instruments for manual inspection. However, this traditional defect detection method has the following drawbacks:

[0003] 1) Low inspection efficiency: Manual defect location, focusing, and measurement are slow and cannot meet the requirements of full inspection.

[0004] 2) Large subjective error in detection: The measurement results and defect type location depend on the operator's experience and judgment, resulting in poor repeatability.

[0005] 3) Unable to automate: Offline operations cannot be linked with the production line, making it difficult to trace and manage quality data.

[0006] 4) Poor adaptability: Different specifications of wire mesh require changing the clamps and recalibrating, which is a complicated process.

[0007] Therefore, there is an urgent need to provide an automatic imaging and defect detection method for capacitor wire mesh applicable to multiple sizes, which can quickly focus, achieve high-definition imaging, and realize precise dimensional measurement and defect detection for capacitor wire mesh of different specifications and sizes. Summary of the Invention

[0008] To overcome the shortcomings of existing technologies, this invention provides an automatic imaging and defect detection method for capacitor wire mesh applicable to multiple sizes of wire frames. It achieves compatible detection of multiple sizes of capacitor wire mesh, eliminating the need for fixture changes and repeated calibration, thus simplifying the workflow. It enables automated image acquisition, effectively improving defect detection efficiency and meeting the full inspection requirements of the production line. Through multi-area mask-based refined partitioning detection, combined with targeted defect type identification, it effectively avoids human subjective errors, reduces the probability of missed or false detections, and accurately identifies various defects such as wire clogging, expansion and contraction, latex peeling, and dirt. Simultaneously, it converts the detection results into structured data and generates a detection report, establishing a data link between the detection and production systems. This enables full traceability, statistical analysis, and standardized management of quality data, effectively solving the technical problems of low efficiency, large errors, insufficient automation, and poor adaptability of traditional manual inspection.

[0009] To achieve the above-mentioned objectives, the present invention adopts the following technical solution: This application provides an automatic imaging and defect detection method for capacitor wire mesh applicable to multiple sizes, including the following steps: S101. Generate the corresponding imaging path based on the parameter information of the wire mesh frame of the capacitor to be detected; S102. The control imaging unit acquires the wire mesh image of the capacitor to be tested according to the imaging path, and performs image preprocessing on the wire mesh image of the capacitor to be tested to obtain a detail-preserving image for defect detection and a region positioning image for region identification. S103. Based on the area positioning image, identify a regular rectangular screen mesh area, and construct multiple area masks based on the regular rectangular screen mesh area; the area mask includes at least a complete screen mesh core area mask, an incomplete screen mesh area mask, a buffer zone mask, and a latex effective detection area mask; S104. Based on a predefined type of defect to be detected, switch between foreground detection areas and background suppression areas in different mask regions, perform defect detection and classification in the corresponding foreground detection area, and generate defect detection results; the defect detection results include at least one defect among mesh blockage, expansion and contraction, latex peeling, and dirt. S105. Encapsulate the defect detection results into structured data, and generate a detection report based on the structured data.

[0010] Furthermore, generating the corresponding imaging path based on the parameter information of the wire mesh frame of the capacitor to be detected includes the following steps: Select the corresponding specification of positioning top post according to the size of the wire mesh frame of the capacitor to be tested, and fix the wire mesh frame of the capacitor to be tested on the carrying platform by the adjustable clamp; wherein, the positioning top post includes at least one of large-size top post, medium-size top post and small-size top post to adapt to the wire mesh frame of the capacitor to be tested of different sizes; the adjustable clamp is adjusted along at least one of the diagonal direction, the transverse direction or the longitudinal direction to clamp and position the wire mesh frame of the capacitor to be tested of different sizes. An imaging path is generated based on at least one parameter from the size of the wire mesh frame of the capacitor to be tested, product specifications, user-specified scanning area, or laser calibration area; wherein, the imaging path includes at least one of the following: global scanning path, local scanning path, linear scan camera scanning path, and area scan camera sampling point.

[0011] Furthermore, controlling the imaging unit to acquire the wire mesh image of the capacitor to be inspected according to the imaging path, and performing image preprocessing on the wire mesh image of the capacitor to be inspected to obtain a detail-preserving image for defect detection and a region localization image for region identification includes the following steps: Before acquiring the image of the capacitor mesh to be tested, the height information of the surface of the capacitor mesh to be tested is obtained by a laser displacement sensor, and the Z-axis motion mechanism is controlled to complete the autofocus based on the height information. The imaging unit includes a linear array camera and / or an area array camera; The screen image of the capacitor to be tested includes at least one of front-light imaging, back-light imaging, or alternating front-light and back-light imaging.

[0012] Furthermore, controlling the imaging unit to acquire the wire mesh image of the capacitor to be inspected according to the imaging path, and performing image preprocessing on the wire mesh image of the capacitor to be inspected to obtain a detail-preserving image for defect detection and a region localization image for region identification includes the following steps: Background trend estimation and background correction are performed on the original image to obtain a detail-preserving image for detecting defects such as mesh blockage, expansion and contraction, latex peeling and dirt. The detail-preserving image is smoothed to obtain an image for locating the identification area of ​​a regular rectangular screen mesh region.

[0013] Furthermore, based on the region-localized image, a regular rectangular screen printing area is identified, and multiple region masks are constructed based on the regular rectangular screen printing area; the region mask includes at least a complete screen printing core area mask, an incomplete edge screen printing area mask, a buffer zone mask, and a latex effective detection area mask, comprising the following steps: Based on the region localization image, high-transmittance candidate regions are extracted, and the high-transmittance candidate regions are obtained through the following threshold model: , in, For regional positioning images, For threshold coefficient, The absolute median difference of the image is used to locate the region; gray values ​​greater than The pixel region was identified as a candidate region for high light transmittance; Based on the high-transmittance candidate regions, a connected component analysis is performed to obtain a set of candidate regions, and the area, bounding rectangle width, bounding rectangle height, aspect ratio, and rectangularity of each candidate region are calculated. Wherein, the rectangularity satisfies: , In the formula, For the first Area of ​​each candidate region and The first The width and height of the bounding rectangle of each candidate region; Candidate areas whose area, aspect ratio, and rectangularity meet the preset conditions are defined as regular rectangular wire mesh areas.

[0014] Furthermore, based on the region-localized image, a regular rectangular screen printing area is identified, and multiple region masks are constructed based on the regular rectangular screen printing area; the region mask includes at least a complete screen printing core area mask, an incomplete edge screen printing area mask, a buffer zone mask, and a latex effective detection area mask, comprising the following steps: Based on the regular rectangular screen mesh area, an inward shrinkage process is performed to obtain the complete screen mesh core area mask, the expression of which is as follows: For the first A regular rectangular wire mesh area After corrosion treatment, the following was obtained: , in, Indicates morphological corrosion, Indicates an indented structural element; all The union of these elements constitutes the complete core area mask of the wire mesh; The edge-incomplete mesh area mask is obtained by the difference between the regular rectangular mesh area and the complete mesh core area mask; wherein, the edge-incomplete mesh area mask is used to characterize the mesh unit area located at the edge of the rectangular mesh area, which is structurally incomplete or truncated by the boundary. The outer expansion area is obtained by expanding the regular rectangular wire mesh area. The difference between the extended region and the regular rectangular wire mesh region is the buffer mask; The area outside the aforementioned extended area within the effective detection zone is the latex effective detection zone mask; The buffer mask is used to isolate the regular rectangular screen area from the latex effective detection area mask, so as to reduce misjudgment caused by the boundary transition area.

[0015] Furthermore, based on a predefined type of defect to be detected, the foreground detection region and background suppression region are switched between different mask regions. Defect detection and classification are performed within the corresponding foreground detection region, and defect detection results are generated. The defect detection results include at least one defect among mesh blockage, expansion and contraction, latex peeling, and dirt, and include the following steps: When detecting wire mesh blockage and expansion / contraction defects, the complete wire mesh core area mask is used as the foreground detection area, and the incomplete wire mesh core area is used as the background suppression area. When detecting latex peeling defects, the effective latex detection area mask is used as the foreground detection area, and the regular rectangular screen area, the screen area with incomplete edges mask, and the buffer area mask are used as the background suppression area. When detecting contamination defects, the effective detection area mask of the latex, the buffer mask, or the union of both are used as the foreground detection area, and the complete screen core area mask is used as the background suppression area.

[0016] Furthermore, based on a predefined type of defect to be detected, the foreground detection region and background suppression region are switched between different mask regions. Defect detection and classification are performed within the corresponding foreground detection region, and defect detection results are generated. The defect detection results include at least one defect among mesh blockage, expansion and contraction, latex peeling, and dirt, and include the following steps: Based on the calculation of the grayscale median and absolute median difference of the complete core area mask of the wire mesh, a standardized response of the wire mesh area is constructed. Based on the standardized response of the screen area, dark anomaly candidate regions are extracted, and the mesh energy retention rate within the local window where the dark anomaly candidate region is located is calculated; when the local window simultaneously satisfies the dark anomaly condition and the mesh energy decay condition, the region is determined to be a screen blockage defect. The transverse and / or longitudinal cycles of the wire mesh are estimated based on a local window of the complete wire mesh core area mask, and the expansion / contraction rate is calculated based on the change of the transverse and / or longitudinal cycles relative to the normal cycle. When the expansion / contraction rate exceeds a preset expansion / contraction threshold, it is determined that there is an expansion / contraction defect in the area corresponding to the local window.

[0017] Furthermore, in encapsulating the defect detection results into structured data and generating a detection report based on the structured data, the structured data includes detection task information, single image detection results, detailed defect information, and global statistical information; wherein, the detailed defect information includes at least the defect number, defect type, defect type name, pixel position of the defect in a single image, position of the defect in the overall screen coordinates, absolute coordinates of the defect in the machine coordinate system, defect area, and defect bounding rectangle.

[0018] Furthermore, encapsulating the defect detection results into structured data and generating a detection report based on the structured data includes the following steps: The structured data is sent to the host computer via TCP communication; The host computer parses the detection task information, defect type, defect image path, single image pixel coordinates, overall wire mesh coordinates, and machine absolute coordinates, and generates a detection report. The detection report includes at least the defect image, defect type, defect location coordinates, defect area, defect circumscribed rectangle, defect statistics, and measurement data.

[0019] The beneficial effects of this application are as follows: It enables compatible inspection of multi-size capacitor wire mesh frames without the need to change fixtures or repeat calibration, simplifying the work process; it enables automated image acquisition, effectively improving defect detection efficiency and meeting the full inspection requirements of the production line; through multi-area mask-based refined partitioned inspection, combined with targeted identification and judgment of defect types, it effectively avoids human subjective errors, reduces the probability of missed or false detections, and accurately identifies various defects such as wire mesh blockage, expansion and contraction, latex peeling, and dirt; at the same time, it converts the inspection results into structured data and generates inspection reports, opening up the data link between the inspection and production systems, realizing full traceability, statistical analysis, and standardized management of quality data, effectively solving the technical problems of low efficiency, large errors, insufficient automation, and poor adaptability of traditional manual inspection. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram illustrating the steps of an automatic imaging and defect detection method for capacitor wire mesh applicable to multiple sizes, according to the present invention. Detailed Implementation

[0022] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0023] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0024] Example 1 An automatic imaging and defect detection method for capacitor wire mesh applicable to multiple sizes includes the following steps: S101. Generate the corresponding imaging path based on the parameter information of the wire mesh frame of the capacitor to be detected.

[0025] By combining multi-specification positioning top posts with adjustable fixtures, it is possible to quickly adapt to wire mesh frames of capacitors to be tested of different sizes, achieving accurate and rapid positioning of the wire mesh frames without frequent fixture changes or repeated calibrations. This effectively simplifies the pre-testing process and greatly improves the compatibility with wire mesh frames of different specifications. Furthermore, it can plan the imaging movement trajectory of the wire mesh frame to be tested based on parameters such as size and product specifications, providing reliable trajectory support for subsequent automated image acquisition.

[0026] Generating the corresponding imaging path based on the parameter information of the wire mesh frame of the capacitor to be tested includes the following steps: Select the appropriate positioning pins based on the size of the screen frame of the capacitor to be tested, and fix the screen frame to the platform using adjustable clamps. The positioning pins include at least one of large, medium, and small pins to accommodate screen frames of different sizes. The positioning pins employ an inward offset design (e.g., for medium-sized screen frames, positioning is achieved by the right and upper left pins) to ensure the screen frame is parallel to the platform plane and prevent clamping deformation. The adjustable clamps are adjusted along at least one of the diagonal, lateral, or longitudinal directions to clamp and position screen frames of different sizes. The adjustable clamps are diagonally adjustable and equipped with graduations, supporting manual or electric fine-tuning to ensure even distribution of clamping force and prevent frame displacement during high-speed scanning.

[0027] An imaging path is generated based on at least one parameter from the size of the wire mesh frame of the capacitor to be tested, product specifications, user-specified scanning area, or laser calibration area.

[0028] The imaging path includes at least one of the following: a global scanning path, a local scanning path, a linear scan camera scanning path, and a sampling point of an area scan camera. The laser calibration area is obtained by controlling the focusing laser spot to move to at least two calibration points in the area to be detected. A rectangular local detection area is generated based on the calibration points, and a bow-shaped scanning path is generated based on the rectangular local detection area.

[0029] S102. The control imaging unit acquires the wire mesh image of the capacitor to be tested according to the imaging path, and performs image preprocessing on the wire mesh image of the capacitor to be tested to obtain a detail-preserving image for defect detection and a region positioning image for region identification.

[0030] By controlling the imaging unit to automatically acquire images of the capacitor mesh to be inspected along the imaging path, and performing differentiated preprocessing on the original images to generate detail-preserving images and region-localization images, this effectively replaces manual positioning, focusing, and shooting operations, greatly improving image acquisition efficiency and meeting the full inspection requirements of the production line. It not only preserves the microscopic details of the mesh to ensure defect detection accuracy but also highlights the region contours to improve positioning efficiency. Simultaneously, by standardizing image acquisition and processing, it effectively eliminates image quality differences caused by manual operation, providing a stable, high-quality image data source for subsequent defect detection processes. Furthermore, since the capacitor mesh contains both high-transmittance small rectangular areas and low-transmittance latex areas including latex holes and latex black areas, traditional threshold segmentation is easily affected by uneven backlighting, local reflections, and mesh density. Therefore, image preprocessing is required for the acquired capacitor mesh images to be tested. This involves brightness correction, detail preservation, and smoothing branch construction to obtain a background-corrected detail-preserved image and a smoothed region localization image, providing stable input for subsequent regular rectangular mesh region recognition, region mask construction, and defect detection.

[0031] Before acquiring the image of the capacitor mesh to be tested, the height information of the surface of the capacitor mesh to be tested is obtained by a laser displacement sensor, and the Z-axis motion mechanism is controlled to complete the autofocus according to the height information; the imaging unit includes a line scan camera and / or an area scan camera; the image of the capacitor mesh to be tested includes at least one of positive light imaging, backlight imaging, or alternating positive light and backlight imaging.

[0032] The control imaging unit acquires images of the wire mesh of the capacitor to be inspected according to the imaging path, and performs image preprocessing on the wire mesh images to obtain a detail-preserving image for defect detection and a region localization image for region identification, including the following steps: Background trend estimation and background correction are performed on the original image to obtain a detail-preserving image for detecting defects such as mesh blockage, expansion and contraction, latex peeling and dirt. When performing background correction, the background trend term of the original image is first estimated using large-scale median filtering, quantile filtering, or low-frequency filtering. The detail-preserving image is obtained using the following expression: , in, The original image grayscale values, To preserve image grayscale values ​​for detail, To prevent constants with a denominator of zero, This represents the average value of the background trend items; This process normalizes the brightness of different imaging positions to a relatively consistent scale, which is beneficial for subsequent rectangle recognition and defect and anomaly judgment. The detail-preserving image is smoothed to obtain an image for locating the identification area of ​​a regular rectangular screen printing region; The region positioning image is obtained by applying Gaussian smoothing, mean smoothing, median smoothing, or low-pass filtering to the detail-preserving image; wherein, the region positioning image is used to weaken the fine mesh pattern inside the rectangular screen area, while preserving the overall light transmission difference between the rectangular screen area and the latex area.

[0033] S103. Based on the area positioning image, identify a regular rectangular screen mesh area, and construct multiple area masks based on the regular rectangular screen mesh area; the area mask includes at least a complete screen mesh core area mask, an incomplete screen mesh area mask, a buffer zone mask, and a latex effective detection area mask.

[0034] By relying on region-localized image recognition to identify regular rectangular wire mesh areas, and by constructing various types of region masks, such as complete wire mesh core area masks, incomplete edge wire mesh area masks, buffer zone masks, and effective latex detection area masks, it is possible to effectively distinguish structural and process differences in different locations, avoid mutual interference between different region features, effectively reduce the probability of missed and false detections, and ensure the accuracy of subsequent defect detection. Simultaneously, based on the region-localized image, intelligent region division eliminates the need for manual selection of the detection range, adapting to the detection needs of multi-size wire mesh frames and ensuring uniform region division standards and consistent detection apertures for wire meshes of different sizes.

[0035] Based on the region localization image, a regular rectangular screen printing area is identified, and multiple region masks are constructed based on the regular rectangular screen printing area; the region mask includes at least a complete screen printing core area mask, an incomplete edge screen printing area mask, a buffer zone mask, and a latex effective detection area mask, including the following steps: Based on the region localization image, high-transmittance candidate regions are extracted, and the high-transmittance candidate regions are obtained through the following threshold model: , in, For regional positioning images, For threshold coefficient, The absolute median difference of the image is used to locate the region; gray values ​​greater than The pixel region was identified as a candidate region for high light transmittance; Based on the high-transmittance candidate regions, a connected component analysis is performed to obtain a set of candidate regions, and the area, bounding rectangle width, bounding rectangle height, aspect ratio, and rectangularity of each candidate region are calculated. Wherein, the rectangularity satisfies: , In the formula, For the first Area of ​​each candidate region and The first The width and height of the bounding rectangle of each candidate region; Candidate areas whose area, aspect ratio, and rectangularity meet the preset conditions are defined as regular rectangular wire mesh areas.

[0036] Based on the inward shrinkage of a regular rectangular wire mesh area, a complete wire mesh core area mask is obtained for wire mesh blockage detection and expansion / contraction detection. Its expression is as follows: For the first A regular rectangular wire mesh area After corrosion treatment, the following was obtained: , in, Indicates morphological corrosion, Indicates an indented structural element; all The union of these elements constitutes the complete core area mask of the wire mesh; The mask for incomplete mesh areas is obtained by the difference between the mask for regular rectangular mesh areas and the mask for complete mesh core areas. The mask for incomplete mesh areas is used to characterize mesh unit areas located at the edge of rectangular mesh areas that are structurally incomplete or truncated by boundaries. This area cannot be directly treated as a latex area mask or a complete mesh area mask, but is used as an auxiliary judgment area.

[0037] The outer expansion area is obtained by expanding the regular rectangular wire mesh area. The difference between the extended region and the regular rectangular screen area is the buffer mask; The area outside the aforementioned extended area within the effective detection zone is the latex effective detection zone mask; The buffer mask is used to isolate the regular rectangular screen area from the effective latex detection area mask, so as to reduce misjudgment caused by the boundary transition area.

[0038] S104. Based on the predefined type of defect to be detected, switch between the foreground detection area and the background suppression area between different area masks, perform defect detection and classification in the corresponding foreground detection area, and generate defect detection results; the defect detection results include at least one defect among mesh blockage, expansion and contraction, latex peeling and dirt.

[0039] By combining preset defect types, the system flexibly switches between foreground detection areas and background suppression areas within various mask regions. This allows for the automatic identification and classification of defects such as mesh blockage, expansion / contraction, latex peeling, and dirt within a specified range. It eliminates subjective human judgment, effectively removing detection errors caused by human experience and ensuring objective, stable, and highly repeatable results. Furthermore, differentiated detection is performed based on the defect characteristics of different regions, suppressing interference from invalid information. This accurately identifies various defects, effectively improving overall detection precision and meeting the high-precision quality inspection requirements of capacitor wire mesh.

[0040] Before defect detection, the types of defects to be detected are predefined. For example, defects appearing inside the screen area and manifesting as localized decreased light transmittance, darkened grayscale, mesh occlusion, or attenuation of texture energy are defined as screen blockage defects; defects appearing inside the screen area and manifesting as significant changes in the mesh spacing opening size are defined as expansion and contraction defects; defects appearing in the latex area and manifesting as localized increased light transmittance, the formation of abnormally bright areas, or shapes and areas significantly larger than normal latex pores are defined as latex detachment defects; and defects appearing in the latex area or near the boundary area and manifesting as localized dark spots, foreign matter adhesion, or abnormally connected patches are defined as contamination defects. Screen blockage defects, latex detachment defects, expansion and contraction defects, and contamination defects are uniformly coded using defect type numbers; the defect type numbers include the first number for screen blockage, the second number for latex detachment, the third number for expansion and contraction, and the fourth number for contamination.

[0041] Based on a predefined type of defect to be detected, the foreground detection region and background suppression region are switched between different mask regions. Defect detection and classification are performed within the corresponding foreground detection region, and defect detection results are generated. The defect detection results include at least one defect among mesh blockage, expansion and contraction, latex peeling, and dirt. The process includes the following steps: When detecting target defects, detection is performed in predefined areas, with different foreground detection areas used for different defects.

[0042] When detecting mesh blockage and expansion / contraction defects, the complete mesh core area mask is used as the foreground detection area, and the incomplete mesh core area is used as the background suppression area to prevent incomplete mesh rectangles within the complete mesh area from being identified as mesh blockage and expansion / contraction defects.

[0043] Based on the calculation of the grayscale median and absolute median difference of the complete core area mask of the wire mesh, a standardized response of the wire mesh area is constructed.

[0044] The standardized response of the wire mesh area meets the following requirements: , in, The median grayscale value of the core area of ​​the complete screen mesh. This represents the absolute median difference of the core area of ​​the complete wire mesh.

[0045] Based on the standardized response of the screen area, candidate dark anomaly regions are extracted, and the mesh energy retention rate within the local window where the candidate dark anomaly region is located is calculated. When the local window simultaneously satisfies the dark anomaly condition and the mesh energy decay condition, the region is determined to be a screen blockage defect.

[0046] The energy retention rate of the mesh pattern meets the following requirements: , in, This represents the energy within the frequency band corresponding to a normal fine mesh texture. This represents the total energy of the local window.

[0047] The transverse and / or longitudinal cycles of the wire mesh are estimated based on a local window of the complete wire mesh core area mask, and the expansion / contraction rate is calculated based on the change of the transverse and / or longitudinal cycles relative to the normal cycle. When the expansion / contraction rate exceeds a preset expansion / contraction threshold, it is determined that there is an expansion / contraction defect in the area corresponding to the local window.

[0048] When detecting latex peeling defects, the effective latex detection area mask is used as the foreground detection area, while the regular rectangular mesh area mask, the incomplete edge mesh area mask, and the buffer zone mask are used as background suppression areas to prevent incomplete mesh rectangles from being identified as latex peeling defects. Calculations are performed within the effective latex detection area mask: , , Standardized response: , like: , Then extract the candidate region of brightness anomaly. And calculate its area: , Equivalent diameter: , Local light transmittance: , If the area, diameter, or light transmittance exceeds the limit, it is judged as a latex peeling defect.

[0049] When detecting contamination defects, the effective detection area mask of the latex, the buffer mask, or the union of both is used as the foreground detection area, and the complete screen core area mask is used as the background suppression area. Dark anomalies: , Construct a candidate region set And calculate: area: , Average anomaly response: , Roundness or shape index: , By further combining position and shape constraints on the candidate areas, normal latex pores and boundary transition areas are eliminated to determine the dirt defects.

[0050] S105. Encapsulate the defect detection results into structured data, and generate a detection report based on the structured data.

[0051] By encapsulating defect detection results into standardized structured data and automatically generating inspection reports based on this structured data, the system achieves organized storage and visual presentation of inspection information. This solves the problems of missing and difficult-to-trace management of traditional offline inspection data. Completely retained inspection data can be retrieved at any time to trace quality issues and can be integrated with production line automation systems to achieve data-driven quality management. The standardized reports present the inspection situation intuitively, which facilitates subsequent quality analysis and control. There is no need for manual data compilation and organization, which greatly reduces the workload of quality inspection.

[0052] Encapsulating defect detection results into structured data and generating a detection report based on the structured data includes the following steps: Structured data includes detection task information, single image detection results, detailed defect information, and global statistical information; among which, detailed defect information includes at least the defect number, defect type, defect type name, pixel position of the defect in the single image, position of the defect in the overall screen coordinates, absolute coordinates of the defect in the machine coordinate system, defect area, and defect bounding rectangle.

[0053] The pixel location of a defect in a single image is denoted as: , The position of the defect in the overall screen coordinate system is obtained by translation, rotation, scaling, affine transformation or perspective transformation between a single image and the overall screen coordinate system; the absolute coordinates of the defect in the machine coordinate system are calculated based on the overall screen coordinate system, pixel resolution, camera mounting angle and machine reference point.

[0054] The structured data is sent to the host computer via TCP communication; The host computer parses the detection task information, defect type, defect image path, single image pixel coordinates, overall wire mesh coordinates, and machine absolute coordinates, and generates a detection report. The detection report includes at least the defect image, defect type, defect location coordinates, defect area, defect circumscribed rectangle, defect statistics, and measurement data. The single image detection result includes at least: image number, preprocessed or detection result image path, high-resolution original image path, a list of defects within the current image, and defect statistics within the current image.

[0055] After a defect is selected by the host computer, the focusing laser sensor or laser pointer is controlled to move to the corresponding position of the defect according to the absolute coordinates of the defect in the machine coordinate system, so as to indicate the actual position of the defect on the frame.

[0056] The host computer displays the local high-magnification image and provides two-point measurement, distance measurement, or defect contour measurement functions.

[0057] Example 2 The above is an automatic imaging and defect detection method for capacitor wire mesh applicable to multi-size wire frames provided in the embodiments of this application. The following is an automatic imaging and defect detection system for capacitor wire mesh applicable to multi-size wire frames provided in the embodiments of this application.

[0058] The second aspect of this application provides an automatic imaging and defect detection system for capacitor wire mesh applicable to multiple sizes of wire frames, comprising: Frame carrier module: A programmable, controllable translational carrier platform equipped with positioning posts and movable clamps for fixing wire mesh frames of capacitors of different specifications to be tested. This module can achieve the clamping and fixing of wire mesh frames of capacitors of different specifications to be tested.

[0059] High-precision motion and positioning module: This module includes a three-axis high-precision linear module (XYZ) and a servo drive system, used to precisely position the imaging unit in three-dimensional space. The X-axis linear motor slide is mounted on the top of the frame, driven by a high-precision linear guide and servo motor, enabling rapid and smooth lateral movement. The Z-axis lead screw slide is vertically mounted on the X-axis slide, using a precision lead screw and guide mechanism to drive the imaging unit for micron-level precise positioning in the vertical direction. The line scan camera, area scan camera, and focusing laser sensor are integrated into the Z-axis slide via a custom bracket, forming a compact imaging assembly. The Y-axis linear motor slide is horizontally arranged at the bottom of the frame, used to support and move the mesh frame carrier.

[0060] Autofocus module: Integrated into the motion module, it includes a laser displacement sensor to quickly acquire the height information of the surface of each capacitor wire mesh frame to be tested, and feed it back to the control system to drive the Z-axis to complete rapid pre-focusing.

[0061] A multi-mode optical imaging module, including: Core camera: A combination of high-resolution, high-depth-of-field industrial line scan and area scan cameras, used to acquire perspective-distortion-free images of the screen of the capacitor under inspection.

[0062] Positive light source system: A line light source located coaxially or to the side of the camera, used to illuminate the surface of the wire mesh frame of the capacitor to be tested, clearly imaging the wire outline, surface contamination or damage.

[0063] Backlight system: A line light source located below the wire mesh frame of the capacitor under test, used for transmitted illumination to clearly image the outline of the mesh and determine whether there is any blockage.

[0064] The system is programmable and can control the individual or alternating triggering of the front light source and back light source to acquire images of two modes in a single measurement process.

[0065] The host computer and control module, including the industrial computer and host computer software interface, is the core of the system. This module is used for: coordinating the motion platform, automatic focusing and light source triggering, and safety protection; inputting and saving the capacitor mesh information of the barcode scanner; setting the scanning path of the line scan camera and the shooting point of the area scan camera; acquiring and storing high-definition images from the camera; running basic capacitor mesh defect detection algorithms; automatically generating inspection reports, recording defect types and location information, querying historical data, displaying magnified defect images, and measuring the mesh size through a measurement interface.

[0066] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the system and unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0067] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. An automatic imaging and defect detection method for capacitor wire mesh applicable to multiple sizes, characterized in that, Includes the following steps: S101. Generate the corresponding imaging path based on the parameter information of the wire mesh frame of the capacitor to be detected; S102. The control imaging unit acquires the wire mesh image of the capacitor to be tested according to the imaging path, and performs image preprocessing on the wire mesh image of the capacitor to be tested to obtain a detail-preserving image for defect detection and a region positioning image for region identification. S103. Based on the area positioning image, identify a regular rectangular screen mesh area, and construct multiple area masks based on the regular rectangular screen mesh area; the area mask includes at least a complete screen mesh core area mask, an incomplete screen mesh area mask, a buffer zone mask, and a latex effective detection area mask; S104. Based on the predefined type of defect to be detected, switch between the foreground detection area and the background suppression area between different area masks, perform defect detection and classification in the corresponding foreground detection area, and generate defect detection results. The defect detection results include at least one of the following defects: clogged mesh, expansion and contraction, latex peeling, and dirt. S105. Encapsulate the defect detection results into structured data, and generate a detection report based on the structured data.

2. The automatic imaging and defect detection method for capacitor wire mesh applicable to multi-size wire frames according to claim 1, characterized in that, Step S101 includes the following steps: Select the corresponding specification of positioning top post according to the size of the wire mesh frame of the capacitor to be tested, and fix the wire mesh frame of the capacitor to be tested on the carrying platform by the adjustable clamp; wherein, the positioning top post includes at least one of large-size top post, medium-size top post and small-size top post to adapt to the wire mesh frame of the capacitor to be tested of different sizes; the adjustable clamp is adjusted along at least one of the diagonal direction, the transverse direction or the longitudinal direction to clamp and position the wire mesh frame of the capacitor to be tested of different sizes. An imaging path is generated based on at least one parameter from the size of the wire mesh frame of the capacitor to be tested, product specifications, user-specified scanning area, or laser calibration area; wherein, the imaging path includes at least one of the following: global scanning path, local scanning path, linear scan camera scanning path, and area scan camera sampling point.

3. The automatic imaging and defect detection method for capacitor wire mesh applicable to multi-size wire frames according to claim 1, characterized in that, In step S102, before acquiring the image of the capacitor mesh to be tested, the height information of the surface of the capacitor mesh to be tested is obtained by a laser displacement sensor, and the Z-axis motion mechanism is controlled to complete the autofocus according to the height information. The imaging unit includes a linear array camera and / or an area array camera; The screen image of the capacitor to be tested includes at least one of front-light imaging, back-light imaging, or alternating front-light and back-light imaging.

4. The automatic imaging and defect detection method for capacitor wire mesh applicable to multi-size wire frames according to claim 1, characterized in that, In step S102, the image preprocessing includes the following steps: Background trend estimation and background correction are performed on the original image to obtain a detail-preserving image for detecting defects such as mesh blockage, expansion and contraction, latex peeling and dirt. The detail-preserving image is smoothed to obtain an image for locating the identification area of ​​a regular rectangular screen mesh region.

5. The automatic imaging and defect detection method for capacitor wire mesh applicable to multi-size wire frames according to claim 1, characterized in that, In step S103, the process of identifying the regular rectangular screen area based on the region positioning image includes the following steps: Based on the region localization image, high-transmittance candidate regions are extracted, and the high-transmittance candidate regions are obtained through the following threshold model: ， in, For regional positioning images, For threshold coefficient, The absolute median difference of the image is used to locate the region; gray values ​​greater than The pixel region was identified as a candidate region for high light transmittance; Based on the high-transmittance candidate regions, a connected component analysis is performed to obtain a set of candidate regions, and the area, bounding rectangle width, bounding rectangle height, aspect ratio, and rectangularity of each candidate region are calculated. Wherein, the rectangularity satisfies: ， In the formula, For the first Area of ​​each candidate region and The first The width and height of the bounding rectangle of each candidate region; Candidate areas whose area, aspect ratio, and rectangularity meet the preset conditions are defined as regular rectangular wire mesh areas.

6. The automatic imaging and defect detection method for capacitor wire mesh applicable to multi-size wire frames according to claim 1, characterized in that, In step S103, constructing multiple region masks based on the regular rectangular screen mesh area includes the following steps: Based on the regular rectangular screen mesh area, an inward shrinkage process is performed to obtain the complete screen mesh core area mask, the expression of which is as follows: For the first A regular rectangular wire mesh area After corrosion treatment, the following was obtained: ， in, Indicates morphological corrosion, Indicates an indented structural element; all The union of these elements constitutes the complete core area mask of the wire mesh; The edge-incomplete mesh area mask is obtained by the difference between the regular rectangular mesh area and the complete mesh core area mask; wherein, the edge-incomplete mesh area mask is used to characterize the mesh unit area located at the edge of the rectangular mesh area, which is structurally incomplete or truncated by the boundary. The outer expansion area is obtained by expanding the regular rectangular wire mesh area. The difference between the extended region and the regular rectangular wire mesh region is the buffer mask; The area outside the aforementioned extended area within the effective detection zone is the latex effective detection zone mask; The buffer mask is used to isolate the regular rectangular screen area from the latex effective detection area mask, so as to reduce misjudgment caused by the boundary transition area.

7. The automatic imaging and defect detection method for capacitor wire mesh applicable to multi-size wire frames according to claim 1, characterized in that, In step S104, switching between foreground detection regions and background suppression regions based on predefined defect types includes the following steps: When detecting wire mesh blockage and expansion / contraction defects, the complete wire mesh core area mask is used as the foreground detection area, and the incomplete wire mesh core area is used as the background suppression area. When detecting latex peeling defects, the effective latex detection area mask is used as the foreground detection area, and the regular rectangular screen area, the screen area with incomplete edges mask, and the buffer area mask are used as the background suppression area. When detecting contamination defects, the effective detection area mask of the latex, the buffer mask, or the union of both are used as the foreground detection area, and the complete screen core area mask is used as the background suppression area.

8. The automatic imaging and defect detection method for capacitor wire mesh applicable to multi-size wire frames according to claim 1, characterized in that, In step S104, the defect detection and classification within the corresponding foreground detection area, and the generation of defect detection results, includes the following steps: Based on the calculation of the grayscale median and absolute median difference of the complete core area mask of the wire mesh, a standardized response of the wire mesh area is constructed. Based on the standardized response of the screen area, dark anomaly candidate regions are extracted, and the mesh energy retention rate within the local window where the dark anomaly candidate region is located is calculated; when the local window simultaneously satisfies the dark anomaly condition and the mesh energy decay condition, the region is determined to be a screen blockage defect. The transverse and / or longitudinal cycles of the wire mesh are estimated based on a local window of the complete wire mesh core area mask, and the expansion / contraction rate is calculated based on the change of the transverse and / or longitudinal cycles relative to the normal cycle. When the expansion / contraction rate exceeds a preset expansion / contraction threshold, it is determined that there is an expansion / contraction defect in the area corresponding to the local window.

9. The automatic imaging and defect detection method for capacitor wire mesh applicable to multi-size wire frames according to claim 1, characterized in that, In step S105, the structured data includes detection task information, single image detection results, defect details, and global statistical information; wherein, the defect details include at least the defect number, defect type, defect type name, pixel position of the defect in the single image, position of the defect in the overall screen coordinates, absolute coordinates of the defect in the machine coordinate system, defect area, and defect bounding rectangle.

10. The automatic imaging and defect detection method for capacitor wire mesh applicable to multi-size wire frames according to claim 1, characterized in that, In step S105, generating the detection report based on the structured data includes the following steps: The structured data is sent to the host computer via TCP communication; The host computer parses the detection task information, defect type, defect image path, single image pixel coordinates, overall wire mesh coordinates, and machine absolute coordinates, and generates a detection report. The detection report includes at least the defect image, defect type, defect location coordinates, defect area, defect circumscribed rectangle, defect statistics, and measurement data.

Images