Method for detecting pore size of polyfluoro film based on machine vision

Abstract

The application discloses a machine vision-based aperture detection method for polyfluorinated film, and belongs to the technical field of visual detection and visual measurement. The method comprises the following steps: acquiring a first image and a second image corresponding to the same detection field of view; acquiring a differential enhancement image with enhanced contrast between holes and a film base; acquiring a definition-optimized image; performing hole segmentation and connected hole separation processing on the definition-optimized image to obtain hole contour data; extracting hole geometric features corresponding to the hole contour data, and calculating a comprehensive equivalent aperture according to the hole contour data and a hole circularity coefficient; statistically analyzing the comprehensive equivalent apertures of all holes to obtain aperture distribution parameters, and identifying abnormal holes and position coordinates of the abnormal holes; establishing a correlation between the aperture distribution parameters, the position coordinates of the abnormal holes and production process parameters, generating a process adjustment suggestion according to a correlation analysis result, and adjusting the production process parameters.

Description

Technical Field

[0001] This application relates to the fields of visual inspection and visual measurement technology, and more specifically, to a method for detecting the pore size of polyfluorinated films based on machine vision. Background Technology

[0002] Fluorinated polyvinyl chloride (PVC) films are widely used in electronic packaging, new energy, and filtration separation due to their excellent chemical stability, weather resistance, and dielectric properties. The uniformity and yield of their pore size distribution are key indicators for evaluating film product quality. In the actual production process of PVC films, fluctuations in the process can easily lead to abnormal pores on the film surface, such as inconsistent pore sizes, uneven distribution, or even localized aggregation. Therefore, high-precision and high-efficiency pore size detection of PVC films during production is urgently needed to adjust the production process in a timely manner and ensure product quality stability. Currently, machine vision-based pore size detection methods have become the main technology in this field, but existing technologies still have many shortcomings in the application scenarios of PVC film pore size detection.

[0003] Patent document CN110441318A discloses a machine vision-based method for detecting defects in the orifices of chemical fiber spinnerets. This method acquires images of the spinneret orifices using an industrial camera, and through steps such as grayscale transformation, region of interest extraction, connected component processing, edge enhancement, and circle fitting, it determines the defects in the spinneret orifices and calculates the orifice diameter and number. The spinneret orifices targeted by this method are discrete through-holes on a metal substrate. The orifices and the substrate exhibit significant contrast in brightness even under a single visible light band, thus requiring only single-band imaging to obtain clear orifice outlines. However, polyfluorinated film materials have high transmittance in the visible light band, and the grayscale difference between the orifice region and the film substrate region is weak under conventional single-band transmission imaging. Therefore, it is difficult to directly use this method to obtain high-contrast images that meet the requirements for accurate orifice segmentation. Furthermore, this method does not address the separation of densely distributed, interconnected orifices.

[0004] Patent document CN110763688A discloses a blind hole foreign object detection system and method. This method uses a coaxial light source to illuminate and image a single blind hole, obtains a blurred outline of the foreign object through local ambiguity calculation, and then determines the presence of foreign objects within the blind hole through morphological closure calculation and quantitative scoring. This method is designed to determine the presence or absence of foreign objects within a single hole location. Its detection targets are mechanically machined blind holes, where the number and density of holes are limited. It does not address the equivalent pore size calculation and statistical distribution analysis of numerous densely packed pores on the film surface, nor does it establish a correlation between pore size detection results and production process parameters. Therefore, it cannot achieve feedback control of the production process based on the detection results, making it difficult to meet the closed-loop adjustment requirements in the production of fluoropolymer films.

[0005] Patent document CN116342682A discloses a method, system, and device for detecting aperture, aperture spacing, and aperture position based on machine vision. This method involves a moving component carrying a detection device to acquire images. After Gaussian filtering preprocessing, an edge detection algorithm is used to obtain the maximum inscribed circle, fitted circle, and minimum circumscribed circle of the aperture to be tested. The aperture, aperture spacing, and aperture position are calculated using the center coordinates of these circles, and the aperture quality is graded and evaluated using a deep learning model. This method primarily targets discrete apertures in the machining field that are limited in number, relatively large in diameter, and have significant optical contrast with the substrate. It employs a single-band image acquisition method and does not address the difficulties in detecting thin-film materials with low optical contrast between pores and the substrate, or the challenges of densely packed and easily adhered pores. Furthermore, it does not involve statistical distribution analysis of the comprehensive equivalent aperture of all pores, identification and location of abnormal pores, or the establishment of a correlation analysis and closed-loop control mechanism between aperture detection results and production process parameters.

[0006] In summary, existing machine vision-based aperture detection technologies are primarily designed for scenarios where the optical contrast between pores and the substrate is significant, and the number and sparse distribution of pores are limited. These technologies struggle to address the insufficient imaging contrast caused by the slight difference in transmitted light intensity between pores and the substrate in fluoropolymer films. Furthermore, they fail to effectively solve the problems of pore separation and accurate calculation of the equivalent aperture when pores are densely distributed and adhered to each other on the film surface. Additionally, they lack the ability to establish a correlation between pore distribution parameters, abnormal pore characteristics, and production process parameters, thus hindering closed-loop feedback control of the production process based on pore detection results. Therefore, there is an urgent need for a machine vision-based fluoropolymer film aperture detection solution to address the technical problems of low detection contrast, difficulty in separating densely adhered pores, and the disconnect between detection results and production processes in existing technologies. Summary of the Invention

[0007] To overcome a series of shortcomings in existing technologies, the purpose of this application is to provide a machine vision-based method for detecting the pore size of fluoropolymer films, comprising the following steps: Transmission imaging of the polyfluoropolymer film to be tested is performed using a visible light source and a specific absorption light source to obtain a first image and a second image corresponding to the same detection field of view. Differential enhancement processing is performed on the first image and the second image to obtain a differentially enhanced image with enhanced contrast between the holes and the thin film substrate; The sharpness-optimized image is obtained by performing sharpness optimization processing on the differential enhancement image; The image with improved clarity is processed for hole segmentation and separation of adhered holes to obtain hole contour data. Extract the geometric features of the holes corresponding to the hole contour data, and calculate the comprehensive equivalent hole diameter based on the hole contour data and the hole roundness coefficient; The pore size distribution parameters are obtained by statistical analysis of the comprehensive equivalent pore size of all pores. Abnormal pores and their location coordinates are then identified by combining the pore size distribution parameters, pore geometric features and spatial distribution features. Establish the correlation between aperture distribution parameters, abnormal hole location coordinates, and production process parameters. Based on the correlation analysis results, generate process adjustment suggestions and adjust the production process parameters accordingly.

[0008] In some embodiments, the method for performing differential enhancement processing on the first image and the second image is as follows: Image registration processing is performed on the first image and the second image to obtain the registered first image and the second image; Flat field correction is performed on the registered first and second images to obtain the corrected first and second images. Logarithmic transformation is performed on the corrected first and second images respectively to obtain the corresponding logarithmically transformed images; The initial difference image is generated by performing a difference operation on the gray values ​​of the corresponding pixel positions in the logarithmically transformed image; The initial difference image is subjected to grayscale enhancement processing to obtain the difference-enhanced image.

[0009] In some embodiments, the method for sharpness optimization processing of differentially enhanced images is as follows: The edge-preserving smoothing process is applied to the difference-enhanced image to obtain a smoothed image; Extract the gray-level gradient features of each local region in the smoothed image; The sharpening intensity of each local region is determined based on the gray-level gradient characteristics. Adaptive sharpening is applied to the smoothed image based on the sharpening intensity to obtain a sharpness-optimized image.

[0010] In some embodiments, the method for performing hole segmentation and adhesion hole separation processing on a sharpness-optimized image is as follows: An adaptive threshold distribution is generated based on the local grayscale features of the image to optimize its sharpness. Based on the adaptive threshold distribution, hole segmentation is performed on the sharpness-optimized image to obtain candidate hole regions; Connectivity analysis is performed on the candidate hole regions to obtain the connected components of each candidate hole; Generate hole separation marker information based on each candidate hole connectivity domain; Based on the hole separation marking information, the adhered holes are separated to obtain the hole contour data corresponding to the independent holes.

[0011] In some embodiments, the method for extracting the geometric features of the hole is as follows: Generate the outline dimension parameters for each hole based on the hole outline data; Generate fitting ellipse parameters for each hole based on the hole contour data; Generate shape feature parameters for each hole based on the contour size parameters and the fitted ellipse parameters; By integrating the contour size parameters and shape feature parameters, the geometric features of each hole are obtained.

[0012] In some embodiments, the method for statistically analyzing the comprehensive equivalent aperture of all holes is as follows: Calculate the statistical characteristic parameters corresponding to the aperture distribution based on the comprehensive equivalent aperture corresponding to all holes; Based on the comprehensive equivalent aperture corresponding to all holes, an aperture distribution histogram is constructed, and a distribution function is fitted based on the aperture distribution histogram to obtain the distribution function parameters corresponding to the aperture distribution; By integrating statistical characteristic parameters and distribution function parameters, the aperture distribution parameters corresponding to all holes are generated.

[0013] In some embodiments, the method for identifying abnormal holes is as follows: Based on the aperture distribution parameters, the criteria for determining aperture anomalies are determined, and aperture anomaly pores are identified according to the criteria. Based on the spatial distribution characteristics of pores, areas of abnormal pore aggregation are identified, and pores with abnormal aggregation are identified based on these areas. Based on the geometric characteristics of the holes, the criteria for determining irregular anomalies are determined, and irregular holes are identified according to the criteria for determining irregular anomalies; Based on the identification results of various abnormal holes, abnormal hole identifiers are generated, and the pixel coordinates of the corresponding hole contours are extracted.

[0014] In some embodiments, the method for determining the location coordinates of abnormal holes is as follows: Based on the calibration parameters of the transmission imaging system, a mapping relationship between the image pixel coordinate system and the thin film physical coordinate system is established; Based on the cumulative transport displacement output by the position encoder, a longitudinal position reference corresponding to the current detection field of view is generated; Based on the mapping relationship and the vertical position reference, the pixel coordinates corresponding to the abnormal hole are transformed to generate the absolute position coordinates of the abnormal hole in the thin film physical coordinate system.

[0015] In some embodiments, the method for establishing the correlation between pore size distribution parameters, abnormal pore location coordinates, and production process parameters is as follows: Obtain the production process parameters, aperture distribution parameters, and coordinates of abnormal hole locations corresponding to historical production batches; A historical data sample set is constructed based on production process parameters, pore size distribution parameters, and coordinates of abnormal pore locations; Based on historical data sample sets, the correlation between pore size distribution parameters and production process parameters, as well as the correlation between the coordinates of abnormal pore locations and production process parameters, are established.

[0016] In some embodiments, the polyfluoropolymer film pore size detection method further includes: The polyfluorinated film produced after process adjustment is subjected to transmission imaging and pore size detection again to obtain the adjusted pore size distribution parameters. Based on the adjusted pore size distribution parameters, it is determined whether the production process meets the preset pore size qualification standard. If the target is not met, the production process parameters are adjusted and the pore size test is repeated. If the target is met, the process adjustment ends, and the pore size test is performed on the polyfluorinated films produced in the next testing cycle to achieve closed-loop control of the production process.

[0017] Compared with the prior art, this application has the following beneficial effects: Firstly, by using dual-band transmission imaging in the visible light band and a specific absorption band, combined with differential enhancement processing, the characteristic absorption characteristics of carbon-fluorine bonds in polyfluorine film materials for specific bands of light are fully utilized, which significantly improves the difference in transmitted light intensity and image contrast between the hole area and the film substrate area, effectively solving the problems of blurred hole boundaries and low detection signal-to-noise ratio under single-band imaging conditions. Secondly, by combining edge smoothing with adaptive sharpening for clarity optimization, and combining adaptive threshold segmentation with connected component analysis for separating adhered holes, the segmentation accuracy of hole contours in high-density, easily adhered hole scenarios is effectively improved, and the probability of misjudgment and missed judgment of adhered holes is reduced. Third, by combining the basic equivalent diameter of the hole profile with the correction coefficient based on the hole roundness coefficient to calculate the comprehensive equivalent aperture, the aperture characterization results can more accurately reflect the actual geometric shape of the hole, thus improving the accuracy of the aperture detection results. Fourth, by establishing the correlation between pore size distribution parameters, abnormal pore location coordinates and production process parameters, and automatically generating process adjustment suggestions, closed-loop feedback control of the production process based on the test results is realized, which improves the stability and consistency of the pore size quality of polyfluorinated film products and the level of automation in the production process. Attached Figure Description

[0018] Figure 1 A flowchart illustrating the overall process of the machine vision-based pore size detection method for polyfluorinated films provided in this embodiment of the invention. Figure 2 A flowchart illustrating the method for acquiring a first image and a second image provided in an embodiment of the present invention; Figure 3 This is a flowchart of the differential enhancement processing and sharpness optimization processing method provided in the embodiments of the present invention; Figure 4 A flowchart illustrating the hole segmentation, geometric feature extraction, and aperture calculation method provided in an embodiment of the present invention; Figure 5 The flowchart illustrates the statistical analysis, abnormal hole identification, and process adjustment methods provided in this embodiment of the invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of the embodiments of this invention will be described in more detail below with reference to the accompanying drawings. In the drawings, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The described embodiments are some embodiments of this invention, but not all embodiments.

[0020] Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] The embodiments and directional terms described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0022] Example 1 This embodiment provides a basic implementation method for detecting the pore size of polyfluorinated films based on machine vision, which is used to illustrate the overall technical process framework of the method. Subsequent embodiments will elaborate on the specific implementation methods of each key step based on this.

[0023] like Figure 1 As shown, the machine vision-based method for detecting the pore size of fluoropolymer films includes the following steps: Transmission imaging of the polyfluoropolymer film to be tested is performed using a visible light source and a specific absorption light source to obtain a first image and a second image corresponding to the same detection field of view. Differential enhancement processing is performed on the first image and the second image to obtain a differentially enhanced image with enhanced contrast between the holes and the thin film substrate; The sharpness-optimized image is obtained by performing sharpness optimization processing on the differential enhancement image; The image with improved clarity is processed for hole segmentation and separation of adhered holes to obtain hole contour data. Extract the geometric features of the holes corresponding to the hole contour data, and calculate the comprehensive equivalent hole diameter based on the hole contour data and the hole roundness coefficient; The pore size distribution parameters are obtained by statistical analysis of the comprehensive equivalent pore size of all pores. Abnormal pores and their location coordinates are then identified by combining the pore size distribution parameters, pore geometric features and spatial distribution features. Establish the correlation between aperture distribution parameters and the coordinates of abnormal pore locations and production process parameters. Generate process adjustment suggestions based on the correlation analysis results and adjust the production process parameters accordingly. The polyfluorinated film produced after process adjustment is subjected to transmission imaging and pore size detection again to obtain the adjusted pore size distribution parameters. Based on the adjusted pore size distribution parameters, it is determined whether the production process meets the preset pore size qualification standard. If the target is not met, the production process parameters are adjusted and the pore size test is repeated. If the target is met, the process adjustment ends, and the pore size test is performed on the polyfluorinated films produced in the next testing cycle to achieve closed-loop control of the production process.

[0024] This embodiment realizes a closed-loop detection and control process for pore defects in polyfluorinated films, from image acquisition, image processing, pore detection, statistical analysis to the generation of process adjustment suggestions, providing a basic technical framework for the specific implementation of each step in subsequent embodiments.

[0025] Example 2 Based on Embodiment 1, this embodiment provides a detailed explanation of how the first and second images are acquired, focusing on the construction method of the dual-band transmission imaging system and the selection criteria for the light source band, in order to improve the contrast of transmitted light intensity between the hole region and the thin film substrate region.

[0026] In some embodiments, such as Figure 2 As shown, the methods for obtaining the first and second images are as follows: Based on the characteristic absorption peak positions of carbon-fluorine bonds in polyfluorine film materials, the center wavelength of a light source in a specific absorption band is determined, and the emission band corresponding to a light source in the visible light band is also determined. Construct a dual-band transmission imaging system that includes a camera and a switching filter device, and establish imaging channels corresponding to visible light band light sources and specific absorption band light sources; The visible light source is controlled to be lit, and the filter device is controlled to switch to the corresponding light transmission state. The polyfluorine film to be tested is subjected to transmission exposure and the first image is obtained. The time interval between two exposure acquisitions is determined based on the conveying speed of the polyfluoropolymer film to be tested, and the displacement of the polyfluoropolymer film to be tested during the two exposure acquisitions is controlled to not exceed the preset pixel threshold. By controlling the illumination of a light source in a specific absorption band and switching the filter device to the corresponding light transmission state, a second image is obtained by performing transmission exposure on the polyfluoropolymer film to be tested. The working sequence of the visible light band light source, the specific absorption band light source, the filter device and the camera is synchronously controlled so that the first image and the second image correspond to the same detection field of view.

[0027] In some embodiments, the center wavelength of the specific absorption band light source is determined based on the characteristic absorption peak of the carbon-fluorine bonds in the polyfluorine film material. This causes significant absorption and energy attenuation when the light beam passes through the film substrate region, while the light energy remains essentially unchanged when passing through the pore region due to the absence of material medium absorption. This creates a significant difference in transmitted light intensity between the pore region and the film substrate region, improving the detection contrast of pore defects. The visible light band light source is selected from the wavelength range where the polyfluorine film material has weaker absorption, in order to obtain higher transmittance and clear imaging information.

[0028] The dual-band imaging method described in this embodiment can obtain two images reflecting the differences in absorption characteristics of the thin film substrate under the same detection field of view, providing high-quality raw image data for the differential enhancement processing step in Embodiment 1.

[0029] Example 3 like Figure 3 As shown, this embodiment, based on Embodiment 1, specifically describes the differential enhancement processing method for the first image and the second image, as well as the sharpness optimization processing method for the differential enhanced image, in order to improve the contrast between the holes and the thin film substrate and the image sharpness.

[0030] In some embodiments, the method for performing differential enhancement processing on the first image and the second image is as follows: Image registration processing is performed on the first image and the second image to obtain the registered first image and the second image; Flat field correction is performed on the registered first and second images to obtain the corrected first and second images. Logarithmic transformation is performed on the corrected first and second images respectively to obtain the corresponding logarithmically transformed images; The initial difference image is generated by performing a difference operation on the gray values ​​of the corresponding pixel positions in the logarithmically transformed image; The initial difference image is subjected to grayscale enhancement processing to obtain the difference-enhanced image.

[0031] In some embodiments, the method for sharpness optimization processing of differentially enhanced images is as follows: The edge-preserving smoothing process is applied to the difference-enhanced image to obtain a smoothed image; Extract the gray-level gradient features of each local region in the smoothed image; The sharpening intensity of each local region is determined based on the gray-level gradient characteristics. Adaptive sharpening is applied to the smoothed image based on the sharpening intensity to obtain a sharpness-optimized image.

[0032] The image processing method described in this embodiment can effectively eliminate the effects of image registration errors and uneven illumination, and improve image clarity while preserving hole edge information, providing a high-quality image foundation for subsequent hole segmentation and feature extraction.

[0033] Example 4 like Figure 4 As shown, this embodiment, based on Embodiment 1, specifically describes the methods for hole segmentation and separation of adhered holes in sharpness-optimized images, the method for extracting hole geometric features, the method for calculating hole roundness coefficient, and the method for calculating comprehensive equivalent aperture.

[0034] In some embodiments, the method for performing hole segmentation and adhesion hole separation processing on a sharpness-optimized image is as follows: An adaptive threshold distribution is generated based on the local grayscale features of the image to optimize its sharpness. Based on the adaptive threshold distribution, hole segmentation is performed on the sharpness-optimized image to obtain candidate hole regions; Connectivity analysis is performed on the candidate hole regions to obtain the connected components of each candidate hole; Generate hole separation marker information based on each candidate hole connectivity domain; Based on the hole separation marking information, the adhered holes are separated to obtain the hole contour data corresponding to the independent holes.

[0035] In some embodiments, the method for extracting the geometric features of the hole is as follows: Generate the outline dimension parameters for each hole based on the hole outline data; Generate fitting ellipse parameters for each hole based on the hole contour data; Generate shape feature parameters for each hole based on the contour size parameters and the fitted ellipse parameters; By integrating the contour size parameters and shape feature parameters, the geometric features of each hole are obtained.

[0036] In some embodiments, the method for calculating the hole roundness coefficient is as follows: Subpixel-level curve fitting is performed on the sequence of contour points corresponding to the hole contour data to obtain a continuous contour curve. Calculate the perimeter of the hole profile based on the continuous profile curve; The roundness coefficient of the hole is calculated based on the pixel area inside the outline corresponding to the hole and the perimeter of the hole outline.

[0037] In some embodiments, the method for calculating the comprehensive equivalent aperture is as follows: Based on the pixel area inside the contour corresponding to each hole, generate the basic equivalent diameter corresponding to each hole; Based on the roundness coefficient of each hole, a correction coefficient corresponding to each hole is generated using a preset correction function. The basic equivalent diameter and correction coefficient are combined and calculated to generate the comprehensive equivalent aperture corresponding to each hole.

[0038] This embodiment achieves precise segmentation and geometric feature extraction of each independent hole in a sharpness-optimized image. By combining the hole roundness coefficient with the basic equivalent diameter, a comprehensive equivalent aperture that more accurately reflects the actual shape of the hole is obtained, providing a data foundation for subsequent statistical analysis of aperture distribution and identification of abnormal holes.

[0039] Example 5 like Figure 5 As shown, this embodiment, based on Embodiment 1, provides a detailed explanation of the statistical analysis method for the comprehensive equivalent aperture of all holes, the identification method for abnormal holes and the determination method for their location coordinates, the method for establishing the correlation between aperture distribution parameters and the location coordinates of abnormal holes and production process parameters, the method for generating process adjustment suggestions, and the preset aperture qualification standard.

[0040] In some embodiments, the method for statistically analyzing the comprehensive equivalent aperture of all holes is as follows: Calculate the statistical characteristic parameters corresponding to the aperture distribution based on the comprehensive equivalent aperture corresponding to all holes; Based on the comprehensive equivalent aperture corresponding to all holes, an aperture distribution histogram is constructed, and a distribution function is fitted based on the aperture distribution histogram to obtain the distribution function parameters corresponding to the aperture distribution; By integrating statistical characteristic parameters and distribution function parameters, the aperture distribution parameters corresponding to all holes are generated.

[0041] In some embodiments, the method for identifying abnormal holes is as follows: Based on the aperture distribution parameters, the criteria for determining aperture anomalies are determined, and aperture anomaly pores are identified according to the criteria. Based on the spatial distribution characteristics of pores, areas of abnormal pore aggregation are identified, and pores with abnormal aggregation are identified based on these areas. Based on the geometric characteristics of the holes, the criteria for determining irregular anomalies are determined, and irregular holes are identified according to the criteria for determining irregular anomalies; Based on the identification results of various abnormal holes, abnormal hole identifiers are generated, and the pixel coordinates of the corresponding hole contours are extracted.

[0042] In some embodiments, the method for determining the location coordinates of abnormal holes is as follows: Based on the calibration parameters of the transmission imaging system, a mapping relationship between the image pixel coordinate system and the thin film physical coordinate system is established; Based on the cumulative transport displacement output by the position encoder, a longitudinal position reference corresponding to the current detection field of view is generated; Based on the mapping relationship and the vertical position reference, the pixel coordinates corresponding to the abnormal hole are transformed to generate the absolute position coordinates of the abnormal hole in the thin film physical coordinate system.

[0043] In some embodiments, the method for establishing the correlation between pore size distribution parameters, abnormal pore location coordinates, and production process parameters is as follows: Obtain the production process parameters, aperture distribution parameters, and coordinates of abnormal hole locations corresponding to historical production batches; A historical data sample set is constructed based on production process parameters, pore size distribution parameters, and coordinates of abnormal pore locations; Based on historical data sample sets, the correlation between pore size distribution parameters and production process parameters, as well as the correlation between the coordinates of abnormal pore locations and production process parameters, are established.

[0044] In some embodiments, the method for generating process adjustment suggestions is as follows: Identify the corresponding abnormal features based on the aperture distribution parameters and the coordinates of the abnormal hole locations; Based on the abnormal characteristics and correlations, the corresponding production process abnormalities are identified; Based on the abnormalities in the production process, corresponding process adjustment suggestions are generated.

[0045] In some embodiments, the preset aperture qualification criteria include: The arithmetic mean of the comprehensive equivalent aperture in the adjusted aperture distribution parameters is not greater than the preset upper limit of the average aperture. The standard deviation of the comprehensive equivalent aperture in the adjusted aperture distribution parameters is no greater than the preset upper limit of aperture dispersion. The proportion of abnormal holes whose comprehensive equivalent pore diameter exceeds the preset single pore diameter upper limit within a unit detection area shall not exceed the preset abnormal proportion upper limit. There are no local areas within the detection area that have been identified as clustering anomalies.

[0046] When all the judgment conditions are met, the current production process is determined to have reached the preset aperture qualification standard; when any judgment condition is not met, the current production process is determined to have not reached the preset aperture qualification standard, and the production process adjustment continues.

[0047] This embodiment achieves multi-dimensional anomaly identification based on pore size distribution parameters and spatial distribution characteristics of abnormal pores. By establishing the correlation between detection results and production process parameters, targeted process adjustment suggestions are generated. Furthermore, the effect of process adjustment is quantitatively judged by combining preset pore size qualification standards, ultimately realizing closed-loop optimization and control of the polyfluorinated film production process.

[0048] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention 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; and these 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 the present invention.

Claims

1. A machine vision-based method for detecting the pore size of polyfluorinated films, characterized in that, Includes the following steps: Transmission imaging of the polyfluoropolymer film to be tested is performed using a visible light source and a specific absorption light source to obtain a first image and a second image corresponding to the same detection field of view. Differential enhancement processing is performed on the first image and the second image to obtain a differentially enhanced image with enhanced contrast between the holes and the thin film substrate; The sharpness-optimized image is obtained by performing sharpness optimization processing on the differential enhancement image; The image with improved clarity is processed for hole segmentation and separation of adhered holes to obtain hole contour data. Extract the geometric features of the holes corresponding to the hole contour data, and calculate the comprehensive equivalent hole diameter based on the hole contour data and the hole roundness coefficient; The pore size distribution parameters are obtained by statistical analysis of the comprehensive equivalent pore size of all pores. Abnormal pores and their location coordinates are then identified by combining the pore size distribution parameters, pore geometric features and spatial distribution features. Establish the correlation between aperture distribution parameters, abnormal hole location coordinates, and production process parameters. Based on the correlation analysis results, generate process adjustment suggestions and adjust the production process parameters accordingly.

2. The method for detecting the pore size of a polyfluorinated film according to claim 1, characterized in that, The method for performing differential enhancement processing on the first and second images is as follows: Image registration processing is performed on the first image and the second image to obtain the registered first image and the second image; Flat field correction is performed on the registered first and second images to obtain the corrected first and second images. Logarithmic transformation is performed on the corrected first and second images respectively to obtain the corresponding logarithmically transformed images; The initial difference image is generated by performing a difference operation on the gray values ​​of the corresponding pixel positions in the logarithmically transformed image; The initial difference image is subjected to grayscale enhancement processing to obtain the difference-enhanced image.

3. The method for detecting the pore size of a polyfluorinated film according to claim 2, characterized in that, The method for sharpness optimization processing of differentially enhanced images is as follows: The edge-preserving smoothing process is applied to the difference-enhanced image to obtain a smoothed image; Extract the gray-level gradient features of each local region in the smoothed image; The sharpening intensity of each local region is determined based on the gray-level gradient characteristics. Adaptive sharpening is applied to the smoothed image based on the sharpening intensity to obtain a sharpness-optimized image.

4. The method for detecting the pore size of a polyfluorinated film according to claim 3, characterized in that, The method for hole segmentation and adhesion hole separation in sharpness-optimized images is as follows: An adaptive threshold distribution is generated based on the local grayscale features of the image to optimize its sharpness. Based on the adaptive threshold distribution, hole segmentation is performed on the sharpness-optimized image to obtain candidate hole regions; Connectivity analysis is performed on the candidate hole regions to obtain the connected components of each candidate hole; Generate hole separation marker information based on each candidate hole connectivity domain; Based on the hole separation marking information, the adhered holes are separated to obtain the hole contour data corresponding to the independent holes.

5. The method for detecting the pore size of a polyfluorinated film according to claim 4, characterized in that, The method for extracting the geometric features of holes is as follows: Generate the outline dimension parameters for each hole based on the hole outline data; Generate fitting ellipse parameters for each hole based on the hole contour data; Generate shape feature parameters for each hole based on the contour size parameters and the fitted ellipse parameters; By integrating the contour size parameters and shape feature parameters, the geometric features of each hole are obtained.

6. The method for detecting the pore size of a polyfluorinated film according to claim 1, characterized in that, The method for statistical analysis of the comprehensive equivalent aperture of all holes is as follows: Calculate the statistical characteristic parameters corresponding to the aperture distribution based on the comprehensive equivalent aperture corresponding to all holes; Based on the comprehensive equivalent aperture corresponding to all holes, an aperture distribution histogram is constructed, and a distribution function is fitted based on the aperture distribution histogram to obtain the distribution function parameters corresponding to the aperture distribution; By integrating statistical characteristic parameters and distribution function parameters, the aperture distribution parameters corresponding to all holes are generated.

7. The method for detecting the pore size of a polyfluorinated film according to claim 6, characterized in that, The method for identifying abnormal holes is as follows: Based on the aperture distribution parameters, the criteria for determining aperture anomalies are determined, and aperture anomaly pores are identified according to the criteria for determining aperture anomalies. Based on the spatial distribution characteristics of pores, areas of abnormal pore aggregation are identified, and pores with abnormal aggregation are identified based on these areas. Based on the geometric characteristics of the holes, the criteria for determining irregular anomalies are determined, and irregular holes are identified according to the criteria for determining irregular anomalies; Based on the identification results of various abnormal holes, abnormal hole identifiers are generated, and the pixel coordinates of the corresponding hole contours are extracted.

8. The method for detecting the pore size of a polyfluorinated film according to claim 7, characterized in that, The method for determining the location coordinates of abnormal holes is as follows: Based on the calibration parameters of the transmission imaging system, a mapping relationship between the image pixel coordinate system and the thin film physical coordinate system is established; Based on the cumulative transport displacement output by the position encoder, a longitudinal position reference corresponding to the current detection field of view is generated; Based on the mapping relationship and the vertical position reference, the pixel coordinates corresponding to the abnormal hole are transformed to generate the absolute position coordinates of the abnormal hole in the thin film physical coordinate system.

9. The method for detecting the pore size of a polyfluorinated film according to claim 1, characterized in that, The method for establishing the correlation between pore size distribution parameters, abnormal pore location coordinates, and production process parameters is as follows: Obtain the production process parameters, aperture distribution parameters, and coordinates of abnormal hole locations corresponding to historical production batches; A historical data sample set is constructed based on production process parameters, pore size distribution parameters, and coordinates of abnormal pore locations; Based on historical data sample sets, the correlation between pore size distribution parameters and production process parameters, as well as the correlation between the coordinates of abnormal pore locations and production process parameters, are established.

10. The method for detecting the pore size of a polyfluorinated film according to any one of claims 1-9, characterized in that, The method for detecting the pore size of polyfluorofilms further includes: The polyfluorinated film produced after process adjustment is subjected to transmission imaging and pore size detection again to obtain the adjusted pore size distribution parameters. Based on the adjusted pore size distribution parameters, it is determined whether the production process meets the preset pore size qualification standard. If the target is not met, the production process parameters are adjusted and the pore size test is repeated. If the target is met, the process adjustment ends, and the pore size test is performed on the polyfluorinated films produced in the next testing cycle to achieve closed-loop control of the production process.

Images