A method and system for determining the specific capacitance of a corrosion foil

Abstract

The application provides a method and system for measuring the specific capacitance of etching foil, relates to the technical field of etching foil performance detection, and comprises the following steps: cutting and surface cleaning of the etching foil sample, pretreatment of the collected image, extraction of the boundary contour of each etching hole from the pretreated image, calculation of the geometric characteristic parameters of the hole, introduction of the parallel plate capacitor theory, conversion of the geometric parameters of each etching hole into corresponding capacitance values by using a formula, further calculation of the local specific capacitance of each hole, accumulation and averaging of the local specific capacitance of all holes in the image area, calculation of the average specific capacitance value of the area, and finally construction of a prediction model, in which multiple area images are taken as input, and the corresponding average specific capacitance value is taken as an output label; after training, the model can directly infer and predict the etching foil image, so that the specific capacitance value per unit area of the whole etching foil can be accurately predicted without the need of formation experiment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field, in particular to a method and system for measuring specific capacitance of etching foil. BACKGROUND

[0002] Etching foil is a functional material widely used in aluminum electrolytic capacitors, and its performance directly affects the energy storage capacity and stability of the capacitor. Among the performance indicators of etching foil, the specific capacitance per unit area is a core parameter for evaluating its quality and process level. Currently, the electrochemical test method is commonly used in industry to detect the specific capacitance of etching foil. This method mainly measures the charge and discharge characteristics of the etching foil sample by applying a voltage to it, to obtain the specific capacitance. However, this testing method has obvious limitations: first, the testing process is tedious, as the sample needs to be connected to the test circuit and soaked in electrolyte for a period of time, which not only takes a long time, but also cannot achieve online detection; second, physical contact with the electrode is required during the test, which is easily disturbed by factors such as poor contact of the clamp, fluctuations in electrolyte concentration, and causes measurement errors; in addition, for large-area etching foil materials, multiple sampling points need to be taken and averaged, further increasing the risk of human intervention and error accumulation.

[0003] With the development of image processing technology, current methods focus on the two-dimensional features of etching holes, and the key parameters that affect the specific capacitance, such as etching hole depth and three-dimensional distribution structure, are not considered. This image analysis method that ignores the etching hole depth can achieve a certain degree of non-contact evaluation, but it is difficult to align with traditional physical measurement results, limiting its credibility and application range in actual industrial detection. More importantly, most of these methods fail to establish a physical mapping relationship from microstructure features to macroscopic specific capacitance, and fail to form a generalizable specific capacitance prediction model, making it impossible to achieve generalized application under new samples, new processes, or complex image conditions.

[0004] The prior art discloses a method for determining the specific capacitance of etched foil, which comprises the following steps: collecting the original image of the etched foil, adjusting the size and normalizing the original image, and then forming a standardized image data; measuring the actual specific capacitance of the corresponding etched foil sample by a traditional formation experiment method; and corresponding the image and the physical test data to form a data sample of image-specific capacitance pairs. However, the image data only depends on the two-dimensional surface topography, lacks modeling of the key factors affecting the specific capacitance, i.e. the etching hole depth and the three-dimensional structure, and is not conducive to real-time deployment and expansion in a production environment due to the large number of parameters of the convolutional neural network and the high consumption of computing resources in the training process. Training parameters such as loss function, optimizer, learning rate and performance evaluation function are set during the network compilation process. During the network training process, the learning rate is automatically adjusted by a callback function, and the model parameters are constantly optimized using the training set. Finally, the model performance is evaluated by the test set, and after training, the network model selected by evaluation is used as the final specific capacitance prediction model.

[0005] The above information disclosed in the background section is only intended to enhance the understanding of the background of the present disclosure, and therefore it can include information known to those of ordinary skill in the art. SUMMARY

[0006] The purpose of the present application is to provide a method and system for determining the specific capacitance of etched foil to solve the problems raised in the background.

[0007] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0008] A method for determining the specific capacitance of etched foil, comprising the following specific steps:

[0009] Step 1: cutting and dividing the etched foil into regions and cleaning it, collecting the etching hole image on the surface of the etched foil, and pre-processing the etching hole image, which includes removing the noise of the etching hole image by using Gaussian filtering and enhancing the contrast of the filtered etching hole image by histogram equalization;

[0010] Step 2: extracting each pre-processed etching hole image, extracting the boundary contour of the enhanced etching hole image by edge detection, calculating the etching hole area by binary image connected domain pixel integration, and calculating the depth by gray linear mapping;

[0011] Step 3: converting the geometric parameters of each etching hole into a single hole capacitance value according to the parallel plate capacitance theory, and converting the single hole capacitance value into a single hole local specific capacitance;

[0012] Step 4, accumulate all the local specific volumes of the holes by using weighted average, and obtain the average specific volume value of the whole image area, and repeat the calculation of the average specific volume value for each region after cropping;

[0013] Step 5, construct a non-destructive prediction model of the specific volume value of the etching foil, input multiple region images of the etching foil, and take the calculated average specific volume value of each region of the etching foil as the label to train the prediction model, and output the specific volume value per unit area of the whole etching foil.

[0014] Further, the etching hole image is preprocessed, and the specific steps are as follows:

[0015] The etching hole image collected on the surface of the etching foil is denoised by Gaussian filtering, a two-dimensional Gaussian kernel is constructed, the weight distribution is adjusted according to the standard deviation, each pixel point in the image is convolved and weighted averaged, and the high-frequency noise interference is suppressed; the denoised image is histogram equalized, the cumulative distribution function is calculated by counting the histogram of pixel gray scale distribution, the original gray value is mapped to the dynamic range, the gray difference between the holes and the background in the image is stretched, and the contrast is enhanced.

[0016] Further, the boundary contour of the enhanced etching hole image is extracted by edge detection, and the specific steps are as follows:

[0017] The Sobel operator is used for gradient calculation, and the horizontal direction gradient calculation formula is:

[0018] ;

[0019] The vertical direction gradient calculation formula is:

[0020] ;

[0021] In the formula, is the pixel value matrix of the input image; represents the horizontal direction gradient; represents the vertical direction gradient;

[0022] The gradient amplitude and direction calculation formula is:

[0023] ;

[0024] In the formula, is the gradient amplitude of the image at the current point, is the angle of the gradient direction of the current pixel point;

[0025] The local maximum pixel point in the gradient direction is retained, the image is region filled, and a binary mask image of the hole region is formed, and the calculation formula is:

[0026] ;

[0027] wherein, represents a binary mask image at the pixel point .

[0028] Further, the binary image connected domain pixel integration is used to calculate the corrosion hole area, and the gray linear mapping is used to calculate the depth, and the specific steps include:

[0029] A certain corrosion hole is defined as a connected region in a binary image, and the area of the region is the total number of all pixels in the region, and the calculation formula is:

[0030] ;

[0031] In the formula, is the pixel coordinate of the image belonging to the th corrosion hole region, is the connected region of the th corrosion hole in the image, is the physical area corresponding to a single pixel, and the physical area corresponding to a single pixel;

[0032] The depth of the corrosion hole is calculated by gray inversion, assuming that the original image gray value is , and the calculation formula is:

[0033] ;

[0034] In the formula, is the maximum gray value of the background region of the image, is the physical depth corresponding to a unit gray difference, is the number of pixel points of the th corrosion hole.

[0035] Further, according to the parallel plate capacitor theory, the geometric parameters of each corrosion hole are converted into single-hole capacitance value, and the single-hole local specific volume is converted according to the single-hole capacitance value, wherein the specific formula of the parallel plate capacitor theory is:

[0036] ;

[0037] In the formula, is the vacuum dielectric constant, is the relative dielectric constant of the medium in the hole;

[0038] The calculation formula for converting the single-hole capacitance value into the single-hole local specific volume is:

[0039] ;

[0040] In the formula, is the local specific volume of the th corrosion hole, is the geometric volume of the hole;

[0041] The calculation method of the hole volume is as follows:

[0042]

[0043] In the formula, and are the hole area and depth, respectively.

[0044] Further, the average specific volume value of each region after cutting is repeatedly calculated, and the calculation formula of the average specific volume value of the image region is as follows:

[0045]

[0046] In the formula, is the total number of detected corrosion holes in the image region.

[0047] Further, a nondestructive prediction model of the specific volume value of the etching foil is constructed, a plurality of region images of the etching foil are input, the calculated average specific volume value of each region of the etching foil is taken as a label to train the prediction model, and the unit area specific volume value of the whole etching foil is output;

[0048] The training process of the nondestructive prediction model of the specific volume value of the etching foil is as follows: the preprocessed image samples are divided into a training set, a validation set and a test set; the training set and the validation set images are used as the input of the prediction model, and the calculated average specific volume value of each region of the etching foil is taken as a label to train the prediction model.

[0049] The nondestructive prediction model of the specific volume value of the etching foil after training inputs the whole etching foil image, and outputs the overall unit area specific volume prediction value of the etching foil.

[0050] The application also provides a system for measuring the specific volume value of the etching foil, which is used to execute the above-mentioned method for measuring the specific volume value of the etching foil, and comprises:

[0051] An image preprocessing module is used to cut and divide regions of the etching foil, clean the etching foil, collect the corrosion hole images on the surface of the etching foil, and preprocess the corrosion hole images, wherein the preprocessing includes removing the noise of the corrosion hole images by using Gaussian filtering and enhancing the contrast of the filtered corrosion hole images by histogram equalization.

[0052] A corrosion hole geometry extraction module is used to extract each corrosion hole image after preprocessing, extract the boundary contour of the enhanced corrosion hole image by edge detection, calculate the corrosion hole area by binary image connected domain pixel integration, and calculate the depth by gray level linear mapping.

[0053] ​​A single-hole capacitance calculation module is configured to convert the geometric parameters of each etching hole into a single-hole capacitance value according to the parallel-plate capacitance theory, and convert the single-hole capacitance value into a single-hole local specific capacity;

[0054] A region specific capacity accumulation module is configured to accumulate all hole local specific capacities by using a weighted average to obtain an average specific capacity value of the entire image region, and repeatedly calculate the average specific capacity value for each region after cropping;

[0055] A specific capacity prediction modeling module is configured to construct a nondestructive prediction model of the specific capacity value of the etching foil, take multiple region images of the etching foil as input, take the calculated average specific capacity value of each region of the etching foil as a label to train the prediction model, and output the unit area specific capacity value of the entire etching foil.

[0056] Compared with the prior art, the present application has the following advantages:

[0057] The present application cuts and surface-washes the etching foil sample, then pre-processes the collected images, extracts the boundary contour of each etching hole from the pre-processed images, calculates the geometric feature parameters of the hole, such as area, perimeter and depth, introduces the parallel-plate capacitance theory, converts the geometric parameters of each etching hole into corresponding capacitance values by using a formula, further calculates the local specific capacity of each hole, then accumulates and averages the local specific capacities of all holes in the image region to calculate the average specific capacity value of the region, and finally constructs a prediction model, takes multiple region images as input, takes the corresponding average specific capacity value as output label, and after training, the model can directly infer and predict the etching foil image, so as to realize accurate prediction of the unit area specific capacity value of the entire etching foil without chemical conversion experiment. BRIEF DESCRIPTION OF DRAWINGS

[0058] Figure 1 The figure is a schematic diagram of the overall method of the present application;

[0059] Figure 2 The figure is a volume parameter and gray mean value change graph of the traditional technical solution in the embodiment of the present application;

[0060] Figure 3 The figure is a specific capacity value change graph of the traditional technical solution in the embodiment of the present application;

[0061] Figure 4 The figure is a volume parameter and gray mean value change graph of the present application;

[0062] Figure 5 The figure is a specific capacity value change graph of the present application;

[0063] Figure 6 The figure is a specific capacity value change comparison graph of the traditional technical solution in the embodiment of the present application and the present embodiment;

[0064] Figure 7The overall system structure block diagram of the present application. DETAILED DESCRIPTION

[0065] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to specific embodiments.

[0066] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present application should be understood as the common meanings understood by those with ordinary skills in the art to which the present application belongs. The terms "first", "second" and similar terms used in the present application do not represent any order, number or importance, but are only used to distinguish different components. The terms "include" or "contain" and similar terms mean that the elements or objects before the terms cover the elements or objects listed after the terms and their equivalents, and do not exclude other elements or objects. The terms "connect" or "connected" and similar terms are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. The terms "upper", "lower", "left", "right" and the like only represent relative positional relationships, which can change accordingly when the absolute positions of the described objects change.

[0067] Embodiment:

[0068] Please refer to Figures 1 to 6 The present application provides a technical solution:

[0069] A method for measuring the specific volume of etching foil, the specific steps comprising:

[0070] Step 1, cutting and dividing the etching foil into regions and cleaning, collecting the etching hole image on the surface of the etching foil, and pretreating the etching hole image, the pretreatment including removing the noise of the etching hole image by using Gaussian filter, and enhancing the contrast of the filtered etching hole image by histogram equalization;

[0071] Cutting and cleaning the etching foil, collecting the etching hole image on the surface of the etching foil, and pretreating the etching hole image, specifically: cutting the etching foil into multiple small square regions; ultrasonic cleaning the surface of the etching foil; collecting high-definition images of the etching holes on the surface of the etching foil;

[0072] The pretreatment of the etching hole image, the specific steps are:

[0073] The corrosion hole image collected on the surface of the corrosion foil is denoised by Gaussian filtering, a two-dimensional Gaussian kernel is constructed, the weight distribution is adjusted according to the standard deviation, convolution and weighted average are performed on each pixel point in the image, and high-frequency noise interference is suppressed; the denoised image is histogram equalized, the cumulative distribution function is calculated by counting the histogram of the pixel gray scale distribution, the original gray value is mapped to the dynamic range, the gray difference between the hole and the background in the image is stretched, and the contrast is enhanced.

[0074] Step 2, extract each corrosion hole image after preprocessing, use edge detection to extract the boundary profile of the enhanced corrosion hole image, use binary image connected domain pixel integration to calculate the area of the corrosion hole, and use gray linear mapping to calculate the depth;

[0075] The edge detection is used to extract the boundary profile of the enhanced corrosion hole image, and the specific steps are:

[0076] The Sobel operator is used for gradient calculation, and the horizontal direction gradient calculation formula is:

[0077] ;

[0078] The vertical direction gradient calculation formula is:

[0079] ;

[0080] In the formula, is the pixel value matrix of the input image; represents the horizontal direction gradient; represents the vertical direction gradient;

[0081] The gradient amplitude and direction calculation formula is:

[0082] ;

[0083] In the formula, is the gradient amplitude of the image at the current point, is the angle of the gradient direction of the current pixel point;

[0084] The local maximum pixel point in the gradient direction is reserved, the image is region filled, the binary mask image of the hole region is formed, and the calculation formula is:

[0085] ;

[0086] In the formula, represents the binary mask image at the pixel point .

[0087] The binary image connected domain pixel integration is used to calculate the area of the corrosion hole, and the gray linear mapping is used to calculate the depth, and the specific steps include:

[0088] A certain corrosion hole is defined as a connected region in a binary image, and the area is the total number of pixels in the region, and the calculation formula is:

[0089] ;

[0090] In the formula, is the pixel coordinate of the image belonging to the th corrosion hole region, is the connected region of the th corrosion hole in the image, is the physical area corresponding to a single pixel, and the physical area corresponding to a single pixel;

[0091] The depth of the corrosion hole is calculated by gray scale inversion, assuming that the original image gray value is , and the calculation formula is:

[0092] ;

[0093] In the formula, is the maximum gray value of the background region in the image, is the physical depth corresponding to a unit gray difference, is the number of pixel points of the th corrosion hole.

[0094] Step 3, according to the parallel plate capacitor theory, the geometric parameters of each corrosion hole are converted into single hole capacitance value, and the single hole local specific volume is converted according to the single hole capacitance value;

[0095] According to the parallel plate capacitor theory, the geometric parameters of each corrosion hole are converted into single hole capacitance value, and the single hole local specific volume is converted according to the single hole capacitance value, wherein the specific formula of the parallel plate capacitor theory is:

[0096] ;

[0097] In the formula, is the vacuum dielectric constant, is the relative dielectric constant of the medium in the hole;

[0098] The calculation formula for converting the single hole capacitance value into the single hole local specific volume is:

[0099] ;

[0100] In the formula, is the local specific volume of the th corrosion hole, is the geometric volume of the hole;

[0101] Wherein the calculation method of the hole volume is:

[0102]

[0103] wherein, and are the hole area and depth, respectively.

[0104] 6. The method of claim 1, wherein the average specific capacity value of each region after cutting is repeatedly calculated, and the calculation formula of the average specific capacity value of the image region is:

[0105]

[0106] wherein, is the total number of etching holes detected in the image region.

[0107] Step 5, constructing a non-destructive prediction model of the specific capacity value of the etching foil, the multiple region images of the etching foil are input, the calculated average specific capacity value of each region of the etching foil is taken as the label to train the prediction model, and the unit area specific capacity value of the whole etching foil is output.

[0108] constructing a non-destructive prediction model of the specific capacity value of the etching foil, the multiple region images of the etching foil are input, the calculated average specific capacity value of each region of the etching foil is taken as the label to train the prediction model, and the unit area specific capacity value of the whole etching foil is output;

[0109] wherein the training process of the non-destructive prediction model of the specific capacity value of the etching foil is as follows: the preprocessed image samples are divided into a training set, a validation set and a test set; the training set and the validation set images are used as the input of the prediction model, and the calculated average specific capacity value of each region of the etching foil is taken as the label to train the prediction model;

[0110] The non-destructive prediction model of the specific capacity value of the etching foil after training inputs the whole etching foil image, and outputs the overall unit area specific capacity prediction value of the etching foil.

[0111] Table 1: Data statistics table measured by traditional technology method

[0112]

[0113] As Figures 2-3 ​​As shown, with the increase of volume parameter, the image gray mean value also gradually rises, especially when the area parameter is higher than 1000, the gray scale is concentrated between 110-130, which shows that the image gray scale is associated with the light transmittance of the hole structure, and the increase of area reflects more light transmission path; however, the perimeter distribution is more concentrated, mostly concentrated in the lower volume parameter area, and the corresponding gray value change is not significant, which shows that the contour perimeter cannot effectively reflect the real optical characteristics or depth structure of the hole, and when the specific volume value is in the range of 55-58 μF / cm3, the area parameter fluctuates sharply between 1200-1400, while the total capacitance is relatively concentrated between 750-850 μF. This phenomenon shows that the area and specific volume are not linear or stable mapping, especially in high area samples, there are still low capacitance values, indicating that the image area seriously overestimates the effective energy storage area when there are blind holes or non-through holes, resulting in unreliable specific volume calculated by empirical formula.

[0114] Table 2: Image recognition and modeling prediction method data statistics table

[0115]

[0116] As Figures 4-5 shown, the volume parameter gradually increases from low to high, and the gray mean value is mainly distributed between 90 and 130, wherein with the increase of volume, the gray mean value as a whole shows an upward trend, and the three features-depth, perimeter, and area are presented in the figure with different horizontal lines, which clearly correspond to the gray response of each group of images, wherein the points corresponding to high volume parameters are mostly concentrated in the high gray interval, indicating that the gray feature extracted in this scheme has obvious correlation with the three-dimensional structure volume, which reflects the effectiveness of image semantic information in reflecting microstructure. The performance of image gray mean value and area feature in specific volume prediction process, the horizontal axis is the area parameter, and the vertical axis is the image gray mean value, the points show an obvious clustering trend, it can be seen that in the area above 1200, the gray mean value is concentrated between 110 and 130, showing strong concentration and stability, reflecting the consistency between image gray and hole structure characteristics. This embodiment takes the image gray mean value as a bridge to realize the nondestructive modeling of specific volume by fusing volume parameters, which enhances the physical interpretation and precision of prediction, and shows the significant advantages of fusion modeling method.

[0117] As Figure 6As shown, the specific volume measured by the traditional method fluctuates significantly in the 35 groups of samples, especially drops to 50 at 10, and continues to deteriorate to 50 after 35, reflecting the sensitivity of the traditional process to the concentration fluctuation of the corrosion liquid and the merging of the holes. In contrast, by image preprocessing and boundary contour extraction, the embodiment accurately obtains the hole area and depth, avoids manual measurement errors, and the data range of image recognition and modeling method is only 6, the standard deviation is reduced by 40%, and the interval of 20-30 is ±3 due to the overlapping of the holes. The fluctuation is compressed to ±1 by accumulating the specific volume of single hole and model generalization.

[0118] As shown in the formula, the present application also provides a system for measuring the specific volume of the corrosion foil, which is used to perform the method for measuring the specific volume of the corrosion foil described above, comprising: Figure 7

[0119] An image preprocessing module is used to crop and divide the corrosion foil into regions, clean the corrosion foil, collect the corrosion hole image on the surface of the corrosion foil, and preprocess the corrosion hole image. The preprocessing includes removing the noise of the corrosion hole image by using Gaussian filtering and enhancing the contrast of the filtered corrosion hole image by histogram equalization.

[0120] A corrosion hole geometry extraction module is used to extract each corrosion hole image after preprocessing, extract the boundary contour of the enhanced corrosion hole image by edge detection, calculate the area of the corrosion hole by pixel integration of the binary image connected domain, and calculate the depth by gray linear mapping.

[0121] A single hole capacitance calculation module is used to convert the geometric parameters of each corrosion hole into a single hole capacitance value according to the parallel plate capacitance theory, and to calculate the single hole local specific volume according to the single hole capacitance value.

[0122] A regional specific volume accumulation module is used to accumulate all the hole local specific volumes by weighted average to obtain the average specific volume value of the entire image region, and to repeatedly calculate the average specific volume value for each region after cropping.

[0123] A specific volume prediction modeling module is used to construct a non-destructive prediction model of the specific volume of the corrosion foil, input multiple regional images of the corrosion foil, take the calculated average specific volume value of each region of the corrosion foil as the label, train the prediction model, and output the unit area specific volume value of the entire corrosion foil.

[0124] The above formulas are dimensionless values calculated, and the formulas are obtained by software simulation of a large amount of data to obtain a formula closest to the actual situation. The preset parameters in the formula are set by a person skilled in the art according to the actual situation.

[0125] ​The above-described embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented by software, the above-described embodiments can be implemented in whole or in part in the form of a computer program product. A person of ordinary skill in the art can be aware that units and algorithm steps of the examples described in connection with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether the functions are performed by hardware or software depends on the specific application and design constraints of the technical solutions.

[0126] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, and can be located in one place or distributed on multiple network units. Part or all of the units can be selected to achieve the purpose of the embodiments according to actual needs.

[0127] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical scope disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A method of determining the specific capacitance of a corrosion foil, characterized by, The specific steps include: Step 1, the etching foil is cut to divide the area and washed, the etching hole image on the surface of the etching foil is collected, and the etching hole image is pretreated, the pretreatment including removing the noise of the etching hole image by using Gaussian filtering, and enhancing the contrast of the filtered etching hole image by histogram equalization; Step 2, each etching hole image after pretreatment is extracted, the boundary profile of the enhanced etching hole image is extracted by edge detection, the area of the etching hole is calculated by binary image connected domain pixel integration, and the depth is calculated by gray linear mapping; Step 3, the geometric parameters of each etching hole are converted into single-hole capacitance value according to the parallel plate capacitor theory, and the single-hole local specific volume is converted according to the single-hole capacitance value; Step 4, the average specific volume value of the whole image area is obtained by using weighted average accumulation of all hole local specific volume, and the average specific volume value of each region after cutting is repeatedly calculated; Step 5, a non-destructive prediction model of etching foil specific volume is constructed, multiple region images of the etching foil are input, the average specific volume value of each region of the etching foil calculated is taken as a label to train the prediction model, and the unit area specific volume value of the whole etching foil is output.

2. A method of determining the specific capacitance of a corrosion foil as claimed in claim 1, characterized in that: The pretreatment of the etching hole image includes the following specific steps: The etching hole image collected on the surface of the etching foil is denoised by Gaussian filtering, a two-dimensional Gaussian kernel is constructed, the weight distribution is adjusted according to the standard deviation, each pixel point in the image is convolved and weighted averaged, and the high-frequency noise interference is suppressed; the denoised image is histogram equalized, the cumulative distribution function is calculated by counting the pixel gray distribution histogram, the original gray value is mapped to the dynamic range, the gray difference between the hole and the background in the image is stretched, and the contrast is enhanced.

3. A method of determining the specific capacitance of a corrosion foil as defined in claim 1, characterized in that: The specific steps of extracting the boundary profile of the enhanced etching hole image by edge detection include: The Sobel operator is used for gradient calculation, the horizontal direction gradient calculation formula is: ； The vertical direction gradient calculation formula is: ； In the formula, is a matrix of pixel values of the input image; represents a horizontal direction gradient; represents a vertical direction gradient; The gradient amplitude and direction calculation formula is: ； In the formula, is the gradient amplitude of the image at the current point, is the angle of the gradient direction of the current pixel point; The local maximum pixel point in the gradient direction is retained, the image is region filled, and the binary mask image of the hole region is formed, and the calculation formula is: ； wherein, represents a binary mask map at the pixel point .

4. A method of determining the specific capacitance of a corrosion foil as defined in claim 1, characterized in that: The specific steps of calculating the area of the etching hole by binary image connected domain pixel integration and calculating the depth by gray linear mapping include: A certain etching hole is defined as a connected region in a binary image, and the area is the total number of all pixels in the region, and the calculation formula is: ； In the formula, is the pixel coordinate belonging to the first corrosion hole region in the image, is the connected region of the first corrosion hole in the image, is the physical area corresponding to a single pixel, and the physical area corresponding to a single pixel; The depth of the corrosion hole is calculated by gray scale inversion. The original image gray scale value is , and the calculation formula is: ； In the formula, is the maximum gray value of the background region in the image, is the physical depth corresponding to a unit gray difference, is the number of pixel points of the first corrosion hole.

5. The method of determining the specific capacitance of a corrosion foil of claim 1, wherein: According to the parallel plate capacitor theory, the geometric parameters of each etching hole are converted into single-hole capacitance value, and the single-hole local specific volume is converted according to the single-hole capacitance value, wherein the specific formula of the parallel plate capacitor theory is: ； wherein is the vacuum permittivity, is the relative permittivity of the medium inside the hole; The calculation formula of the single-hole local specific volume converted from the single-hole capacitance value is: ； wherein is the local specific capacity of the th corrosion pit, is the geometric volume of the pit. The calculation method of the hole volume is: ； wherein and are the hole area and depth, respectively.

6. A method of determining the specific capacitance of a corrosion foil as defined in claim 1, wherein: The calculation formula of the image region average specific volume value is: ； In the formula, is the total number of detected etch holes within the image area.

7. A method of determining the specific capacitance of a corrosion foil as defined in claim 1, wherein: A non-destructive prediction model of etching foil specific volume is constructed, multiple region images of the etching foil are input, the average specific volume value of each region of the etching foil calculated is taken as a label to train the prediction model, and the unit area specific volume value of the whole etching foil is output. The specific training process of the etching foil specific volume value nondestructive prediction model is as follows: the preprocessed image samples are divided into a training set, a verification set and a test set; the training set and the verification set are used as the input of the prediction model, and the average specific volume value of each region of the etching foil obtained by calculation is used as the label to train the prediction model; The etching foil specific volume value nondestructive prediction model after training is input into the whole etching foil image, and the whole etching foil unit area specific volume prediction value is output.

8. A system for determining the specific capacitance of a corrosion foil, characterized by: The determination system is used to perform the determination method of any one of claims 1-7, comprising: An image preprocessing module is configured to crop and divide regions of the etching foil, clean the etching foil, collect etching hole images on the surface of the etching foil, and preprocess the etching hole images, wherein the preprocessing includes removing noise of the etching hole images by using Gaussian filtering and enhancing contrast of the filtered etching hole images by histogram equalization; An etching hole geometry extraction module is configured to extract each etching hole image after preprocessing, extract the boundary contour of the enhanced etching hole image by edge detection, calculate the area of the etching hole by pixel integration of the binary image connected domain, and calculate the depth by gray linear mapping; A single-hole capacitance calculation module is configured to convert the geometric parameters of each etching hole into a single-hole capacitance value according to the parallel plate capacitance theory, and convert the single-hole capacitance value into a single-hole local specific volume; A region specific volume accumulation module is configured to accumulate all hole local specific volumes by weighted average to obtain the average specific volume value of the whole image region, and repeatedly calculate the average specific volume value of each region after cropping; A specific volume prediction modeling module is configured to construct an etching foil specific volume value nondestructive prediction model, input multiple region images of the etching foil, train the prediction model by using the average specific volume value of each region of the etching foil obtained by calculation as the label, and output the unit area specific volume value of the whole etching foil.

Images