An intelligent AI-based lithium battery diaphragm production process optimization method and system

Abstract

This application relates to a method and system for optimizing the production process of lithium battery separators based on intelligent AI. The method includes acquiring a quality inspection image of the lithium battery separator and identifying feature regions in the image; determining the center point of the feature region; establishing an analysis reference line segment based on the center point of the feature region; acquiring pixel values ​​at the intersection points and constructing a filtering domain using the obtained pixel values; extracting the quality inspection image using the obtained filtering domain to obtain distinguishing regions; calculating the ratio of the area of ​​the distinguishing region to the area of ​​the quality inspection image to obtain porosity data; comparing the porosity data with porosity reference data; and providing process optimization directions based on the comparison results. The intelligent AI-based method and system for optimizing the production process of lithium battery separators disclosed in this application combines feature recognition and secondary reconstruction to identify voids in the lithium battery separator and calculate porosity data, providing data reference for adjusting the separator process.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to a method and system for optimizing the production process of lithium battery separators based on intelligent AI. Background Technology

[0002] The lithium-ion battery separator is one of the core internal components of a lithium-ion battery, playing a dual role as a "safety barrier" and an "ion channel." Its core function is to achieve physical isolation between the positive and negative electrodes of the battery (preventing short circuits) while allowing lithium ions to pass freely, ensuring the normal progress of electrochemical reactions.

[0003] Scanning electron microscopy (SEM) is a core technology for characterizing the surface morphology of lithium battery separators. Its working principle is based on the interaction between a high-energy electron beam and the sample surface, generating high-resolution microscopic morphology images by collecting secondary electrons or backscattered electron signals.

[0004] Porosity directly affects the final battery quality. It is generally required to be between 30% and 60%. However, this value will vary under different battery performance requirements, and may even be fixed within a certain range, because it is necessary to balance conductivity and strength (for every 10% increase in porosity, the tensile strength of the separator usually decreases by 20%-30%).

[0005] The commonly used threshold segmentation method has a large error in processing low-contrast images. A threshold deviation of 5% can even lead to a difference of 8% in the porosity calculation results. At the same time, the ambiguity at the edges of the pores will further exacerbate the difference in the calculation results. These differences cannot serve as a positive reference for adjusting the membrane process. Summary of the Invention

[0006] This application provides a method and system for optimizing the production process of lithium battery separators based on intelligent AI. It combines feature recognition and secondary reconstruction to identify voids on the lithium battery separator and calculate porosity data, providing data reference for adjusting the separator process.

[0007] The above-mentioned objective of this application is achieved through the following technical solution:

[0008] In a first aspect, this application provides a method for optimizing the lithium battery separator manufacturing process based on intelligent AI, including:

[0009] Acquire quality inspection images of lithium battery separators and identify feature regions in the quality inspection images;

[0010] Determine the center point of the feature region;

[0011] An analysis reference line segment is established based on the center point of the feature region, and the analysis reference line segment intersects with the edge of the feature region.

[0012] Obtain the pixel values ​​at the intersection points and use the obtained pixel values ​​to construct the filtering area;

[0013] The obtained filtering domain is used to extract the distinguishing regions from the quality inspection image;

[0014] The porosity data is obtained by calculating the ratio of the area of ​​the distinguishing region to the area of ​​the quality inspection image;

[0015] The porosity data is compared with the porosity reference data, and the direction of process optimization is given based on the comparison results.

[0016] In one possible implementation of the first aspect, identifying feature regions in a quality inspection image further includes:

[0017] Multiple local features are extracted from the quality inspection image;

[0018] Calculate the shape feature area of ​​each local feature;

[0019] The obtained local features are sorted according to the area of ​​the shape features;

[0020] Use the first local feature in the sequential sequence as the background feature;

[0021] When sorting the obtained local features, the numerical value of the shape feature area tends to decrease in the sequential sequence.

[0022] In one possible implementation of the first aspect, for the remaining local features, it also includes:

[0023] Establish analysis reference lines on the remaining local features. The analysis reference lines include horizontal analysis reference lines and vertical analysis reference lines.

[0024] Use the pixels on the analysis reference line to create the analysis curve;

[0025] The analysis curve is decomposed in the time domain to obtain multiple sets of location points. Each set of location points includes two location points and two location point pixel values.

[0026] All location point groups are merged, and only the two outermost location points are retained;

[0027] A dynamic numerical range is constructed using the pixel values ​​of the retained location points, and this dynamic numerical range is then used to extract the quality inspection image.

[0028] In one possible implementation of the first aspect, when constructing a dynamic numerical range using the pixel values ​​of the reserved location points, each location point pixel value is located at the midpoint of an independent sub-dynamic numerical range.

[0029] In one possible implementation of the first aspect, identifying feature regions in a quality inspection image includes:

[0030] Using color and shape, the hole areas are identified in the quality inspection image.

[0031] The perforated areas are classified into surface perforated areas and three-dimensional perforated areas, based on color difference and area.

[0032] Use the three-dimensional hole region as a feature region;

[0033] The surface hole area is expanded at the edges and then used as a feature area.

[0034] In one possible implementation of the first aspect, when using the three-dimensional hole region as a feature region, it further includes:

[0035] Determine the transition region between two adjacent three-dimensional hole regions;

[0036] Connect the outer contours of two adjacent three-dimensional hole regions to maximize the area of ​​the region formed by the transition region between the two adjacent three-dimensional hole regions.

[0037] The newly added edges of the constituent regions are used as the edges of the transition regions;

[0038] Calculate whether the edge of the transition region is related to the edges of the two adjacent three-dimensional hole regions. If there is a relationship, merge the transition region and the two adjacent three-dimensional hole regions.

[0039] In one possible implementation of the first aspect, using the surface hole region as a feature region after edge expansion includes:

[0040] Locate the highlight areas near the surface pores and determine the edges of the highlight areas;

[0041] The edges of the highlight areas are merged with the edges of the surface hole areas to maximize the area of ​​the surface hole areas.

[0042] Secondly, this application provides a lithium battery separator manufacturing process optimization device based on intelligent AI, comprising:

[0043] The image acquisition unit is used to acquire quality inspection images of lithium battery separators and identify feature regions in the quality inspection images.

[0044] The first processing unit is used to determine the center location point of the feature region;

[0045] The second processing unit is used to establish an analysis reference line segment based on the center point of the feature region, and the analysis reference line segment intersects with the edge of the feature region.

[0046] The third processing unit is used to obtain the pixel values ​​at the intersection and use the obtained pixel values ​​to construct the filtering domain;

[0047] The extraction processing unit is used to extract the quality inspection image using the obtained filtering domain to obtain the distinguishing region;

[0048] The computational processing unit is used to calculate the ratio of the area of ​​the distinguishing region to the area of ​​the quality inspection image to obtain porosity data;

[0049] The results output unit is used to compare porosity data with porosity reference data and provide directions for process optimization based on the comparison results.

[0050] Thirdly, this application provides a lithium battery separator manufacturing process optimization system based on intelligent AI, the system comprising:

[0051] One or more memories for storing instructions; and

[0052] One or more processors are configured to call and execute the instructions from the memory to perform the methods described in the first aspect and any possible implementation thereof.

[0053] Fourthly, this application provides a computer-readable storage medium, the computer-readable storage medium comprising:

[0054] The program, when run by a processor, is executed as described in the first aspect and any possible implementation thereof.

[0055] Fifthly, this application provides a computer program product, including program instructions that, when run by a computing device, execute the method described in the first aspect and any possible implementation thereof.

[0056] Sixthly, this application provides a chip system including a processor for implementing the functions involved in the foregoing aspects, such as generating, receiving, transmitting, or processing the data and / or information involved in the foregoing methods.

[0057] This chip system can consist of chips or include chips and other discrete components.

[0058] In one possible design, the chip system also includes a memory for storing necessary program instructions and data. The processor and the memory can be decoupled and located on different devices, connected via wired or wireless means, or the processor and the memory can be coupled to the same device.

[0059] The beneficial technical effects of this application are as follows:

[0060] The method and system for optimizing the production process of lithium battery separators based on intelligent AI disclosed in this application combine feature recognition and secondary reconstruction to identify voids on the lithium battery separator and calculate porosity data. This method can calculate the porosity of lithium battery separators in continuous production processes and provide the calculation results, providing data reference for separator process adjustment. Attached Figure Description

[0061] Figure 1 This is a flowchart illustrating the steps of an AI-based method for optimizing the production process of lithium-ion battery separators, as provided in this application.

[0062] Figure 2 This is a schematic diagram of a quality inspection image provided in this application.

[0063] Figure 3 This is a schematic diagram of a feature region provided in this application.

[0064] Figure 4 This is a schematic diagram of establishing an analytical reference line provided in this application.

[0065] Figure 5 This is a schematic diagram illustrating the principle of obtaining location points using analysis curves, as provided in this application.

[0066] Figure 6 This is a schematic diagram of a three-dimensional perforated region and a surface perforated region provided in this application.

[0067] Figure 7 This is a schematic diagram of the edge of a transition region and the edge of a highlight region provided in this application. Detailed Implementation

[0068] The technical solutions in this application will be further described in detail below with reference to the accompanying drawings.

[0069] This application discloses a method for optimizing the lithium battery separator manufacturing process based on intelligent AI. Please refer to [link / reference]. Figure 1 In some examples, the intelligent AI-based lithium battery separator manufacturing process optimization method disclosed in this application includes the following steps:

[0070] S101, acquire a quality inspection image of the lithium battery separator and identify the feature regions in the quality inspection image;

[0071] S102, Determine the center location point of the feature region;

[0072] S103, establish an analysis reference line segment based on the center point of the feature region, and the analysis reference line segment intersects with the edge of the feature region;

[0073] S104, obtain the pixel values ​​at the intersection and use the obtained pixel values ​​to construct the filtering field;

[0074] S105, Use the obtained filtering domain to extract the quality inspection image to obtain the distinguishing region;

[0075] S106, calculate the ratio of the area of ​​the distinguishing region to the area of ​​the quality inspection image to obtain porosity data;

[0076] S107 compares the porosity data with the porosity reference data and provides directions for process optimization based on the comparison results.

[0077] Overall, the intelligent AI-based lithium battery separator manufacturing process optimization method provided in this application dynamically optimizes the separator process based on porosity data. For example, a specific value or range of values ​​is first defined, which is the porosity reference data. When the actual detected porosity data is greater than or less than the porosity reference data, the process optimization direction needs to be given. The purpose of the process optimization direction is to make the porosity data equal to the porosity reference data or within the required range of the porosity reference data.

[0078] Alternatively, it can be interpreted that the purpose of this application is to detect whether the porosity data meets the requirements, and to indicate whether it is below or above the requirements.

[0079] Of course, when the porosity data meets the requirements, it is also necessary to combine parameters such as pore size distribution and tortuosity for comprehensive judgment. That is, the technical solution in this application is part of a judgment system, which is mainly responsible for solving the detection of porosity data.

[0080] The judgment system described above is a neural network, or intelligent AI. This system, after training, can analyze received data and provide directions for process optimization. Furthermore, the technical solution provided in this application can also be part of an image processing model (intelligent AI). When an image is given to this image processing model (intelligent AI), the model (intelligent AI) uses this technical solution to process the received image.

[0081] In step S101, a quality inspection image of the lithium battery separator is first acquired. Figure 2(As shown) and identify characteristic regions in the quality inspection image. Here, the characteristic regions refer to pores on the lithium battery separator, such as... Figure 3 As shown, the center position point of the feature region is determined in step S102, and then an analysis reference line segment is established based on the center position point of the feature region in S103.

[0082] For the analysis reference line segment, it is required that the analysis reference line segment intersects with the edge of the feature area. Of course, it is also necessary to set a termination condition for the extension of the analysis reference line segment. In this application, the analysis reference line segment is limited to 1.1-1.3 times the average diameter of the holes on the lithium battery separator.

[0083] The specific method for calculating the average diameter of the cavities is to calculate the ratio of the total area of ​​the cavities to the total number of cavities, obtain the average area of ​​the cavities, and then calculate the average diameter of the cavities based on this ratio.

[0084] The average diameter of the pores on the lithium battery separator is calculated by averaging the characteristic regions identified in the quality inspection image. It is a dynamic value, not a fixed value.

[0085] Next, in step S104, the pixel values ​​at the intersection are obtained and the obtained pixel values ​​are used to construct the filtering domain, that is, the actual pixel values ​​on the image are used to determine which values ​​to use for filtering.

[0086] The specific way to construct a filtering range using pixel values ​​is to use the pixel values ​​as a reference and take the positive and negative error values ​​as endpoints to construct the filtering range. For example, if the pixel value is N, the filtering range is [NM, N+M], where M is a number greater than zero.

[0087] In step S105, the obtained filtering domain is used to extract the quality inspection image to obtain the distinguishing region. Here, some curve segments are first obtained. After merging the obtained curve segments, a closed region, which is the distinguishing region, can be obtained.

[0088] In some possible implementations, the value of M is increased when there are breakpoints in the obtained curve segment, and decreased when there are many overlapping areas in the obtained curve segment. The adjustment is made by dynamically changing the value of M.

[0089] In step S106, the ratio of the area of ​​the distinguishing region to the area of ​​the quality inspection image is calculated to obtain porosity data. Finally, in step S107, the porosity data is compared with the porosity reference data, and the process optimization direction is given based on the comparison result (the purpose of the process optimization direction is to make the porosity data equal to the porosity reference data or within the required range of the porosity reference data). This part has been described in the foregoing and will not be repeated here.

[0090] There are difficulties in processing using only image recognition, specifically in the following ways:

[0091] The pore morphology of lithium battery separators is diverse, including irregular polygons, narrow slits, branching, etc., and may contain interconnecting pores (through the thickness of the separator), closed pores (local depressions), pore clusters (densely distributed small pores), etc. Image recognition requires modeling and manual labeling training, which limits its comprehensive coverage and adaptability.

[0092] The low grayscale difference between the membrane material and the pores (e.g., the substrate and pores of a polymer membrane are both light-colored in SEM images, with only a slight grayscale difference), or the pore edges exhibit a "gradual transition" rather than a clear boundary due to the roughness of the material itself (e.g., the interwoven structure of a fibrous membrane), can easily lead to "oversegmentation" or "undersegmentation" during image segmentation.

[0093] In high-porosity membranes, pores may be densely distributed or even overlap each other (in two-dimensional projection), forming "connected regions." It is difficult to accurately separate adjacent pores through image recognition alone, leading to counting errors or deviations in pore size measurement.

[0094] In this application, a secondary reconstruction method is used for processing. This method combines image recognition and reconstruction (the obtained filtering domain is used to extract the quality inspection image), which can further clarify the edge positions of the distinguishing regions, thus obtaining more accurate calculation results.

[0095] In some cases, the following additional steps were taken when identifying feature regions in quality inspection images:

[0096] S201, Extract multiple local features from the quality inspection image;

[0097] S202, calculate the shape feature area of ​​each local feature;

[0098] S203, sort the obtained local features according to the area of ​​the shape features;

[0099] S204, the first local feature in the sequential sequence is used as the background feature;

[0100] When sorting the obtained local features, the numerical value of the shape feature area tends to decrease in the sequential sequence.

[0101] The purpose of steps S201 to S204 is to determine the background color. Specifically, this is achieved by generating a dynamic numerical range domain, then using this domain to extract values ​​from the quality inspection image and ultimately determine the background color. The specific method for generating the dynamic numerical range domain is to first denoise the quality inspection image, for example, using Gaussian filtering and wavelet transform filtering.

[0102] After denoising is completed, determine the specific colors of the pixels in the quality inspection image.

[0103] In some possible implementations, mean filtering is also required, which involves summing the values ​​of multiple pixels (pixel blocks) and then averaging them to determine which colors the pixels in the quality inspection image actually have.

[0104] The values ​​obtained at this point are used to extract multiple local features. For each local feature, the shape feature area needs to be calculated. When sorting the multiple local features by size, the shape feature area values ​​should tend to decrease in the order sequence.

[0105] The first local feature in the sequence is used as the background feature. At this time, the outline of the feature can be obtained by the shape of the local feature. Then, the area outside the background feature can be temporarily included in the pore range.

[0106] In some possible implementations, the obtained value can be expanded. For example, if the pixel value is N, the expanded range is [NM, N+M], where the value of M is generally 2-4.

[0107] The remaining local features are processed as follows:

[0108] Establish analysis reference lines on the remaining local features. The analysis reference lines include horizontal analysis reference lines and vertical analysis reference lines.

[0109] Use the pixels on the analysis reference line to create the analysis curve;

[0110] The analysis curve is decomposed in the time domain to obtain multiple sets of location points. Each set of location points includes two location points and two location point pixel values.

[0111] All location point groups are merged, and only the two outermost location points are retained;

[0112] A dynamic numerical range is constructed using the pixel values ​​of the retained location points, and this dynamic numerical range is then used to extract the quality inspection image.

[0113] In the above method, analytical reference lines are used to determine the partial edge points of the remaining local features, such as... Figure 4 As shown, these edge points have pixel values, and the obtained pixel values ​​are used to construct a dynamic numerical range, which is a set of pixel values.

[0114] Decomposing the analysis curve in the time domain refers to using wavelet decomposition to process the analysis curve. The resulting sub-curves have a start time (position point) and an end time (position point), and each sub-curve corresponds to a set of position points.

[0115] The specific explanation of wavelet decomposition is as follows:

[0116] Wavelet decomposition is a signal analysis method based on wavelet transform. It decomposes the original signal into different frequency components (from high to low frequency) and preserves the temporal localization information of each component, enabling multi-scale, multi-resolution analysis of the signal. Wavelet decomposition relies on the "mother wavelet" and its scaled and shifted "daughter wavelet." The mother wavelet is a function with rapid decay and zero mean; in this application, either the Haar wavelet or the Daubechies wavelet can be used.

[0117] Taking the Fourier Fast Transform (FTT) as an example, the FTT transforms the analysis curve into the frequency domain. The resulting sub-curve only has frequency domain characteristics but lacks time domain characteristics; that is, the sub-curve has neither a start time nor an end time. However, when using wavelet transform for decomposition, the resulting sub-curve has both a start time (position point) and an end time (position point). These position points can then correspond to changes in brightness and darkness in the image.

[0118] Merging all position point groups and retaining only the two outermost position points means placing all position points on an analysis reference line according to their corresponding positions obtained using wavelet decomposition, and then retaining only the two outermost position points.

[0119] Finally, the pixel values ​​of the retained location points are used to construct a dynamic numerical range, and the dynamic numerical range is used to extract the quality inspection image. The purpose here is to use the edges obtained from the analysis for further extraction, which can clarify the edges of the remaining local features.

[0120] In some possible implementations, when constructing a dynamic numerical range using the pixel values ​​of reserved location points, each location point pixel value is located at the midpoint of an independent sub-dynamic numerical range.

[0121] In some possible implementations, when dynamic numerical ranges overlap, the two dynamic numerical ranges are merged.

[0122] When constructing a dynamic numerical range using the pixel values ​​of the retained location points and extracting the quality inspection image using the dynamic numerical range, a complete closed shape may be obtained, or only some curve segments may be obtained. In this case, it is necessary to appropriately increase the dynamic numerical range, or construct an analysis curve here and perform the above steps.

[0123] Analysis of curves and location points, such as Figure 5 As shown.

[0124] This application also provides another method for identifying feature regions in quality inspection images, as follows:

[0125] Using color and shape, the hole areas are identified in the quality inspection image.

[0126] The perforated areas are classified into surface perforated areas and three-dimensional perforated areas, based on color difference and area.

[0127] Use the three-dimensional hole region as a feature region;

[0128] The surface hole area is expanded at the edges and then used as a feature area.

[0129] After obtaining the closed shape, the closed shape can be used to extract the range on the quality inspection image. The content obtained at this time is a part of the quality inspection image. Then, color and shape are used to identify the hole area in the quality inspection image.

[0130] The colors and shapes mentioned here refer to the content obtained through extensive statistics and manual annotation. However, the disadvantage of this method is that it can only know the known types and cannot obtain the location types. For example, the color of the hole area is basically a known quantity, but the shape type of the hole is complex and cannot be known in its entirety.

[0131] After obtaining the hole regions, they are classified into surface hole regions and three-dimensional hole regions, based on color difference and area. Specifically, color difference refers to the color difference between the hole region and its surrounding area, while area is the actual area of ​​the hole region. These two values ​​are fixed settings. If both the color difference and area are greater than the corresponding settings, the hole region is classified as a three-dimensional hole region; otherwise, it is classified as a surface hole region. Figure 6 As shown.

[0132] When using a three-dimensional hole region as a feature region, it also needs to be processed in the following way:

[0133] Determine the transition region between two adjacent three-dimensional hole regions;

[0134] Connect the outer contours of two adjacent three-dimensional hole regions to maximize the area of ​​the region formed by the transition region between the two adjacent three-dimensional hole regions.

[0135] The newly added edges of the constituent regions are used as the edges of the transition regions;

[0136] Calculate whether the edge of the transition region is related to the edges of the two adjacent three-dimensional hole regions. If there is a relationship, merge the transition region and the two adjacent three-dimensional hole regions.

[0137] The purpose of this method is to determine whether the area between two adjacent three-dimensional porous regions belongs to the three-dimensional porous region or to the surface of the lithium battery separator.

[0138] Specifically, the first step is to connect the outer contours of two adjacent 3D hole regions to maximize the area of ​​the transition region between them. Then, the newly added edges forming the transition region are used as the edges of the transition region. Figure 7 As shown, the final step is to calculate whether the edge of the transition region is related to the edges of the two adjacent three-dimensional hole regions.

[0139] The correlation here is calculated as follows:

[0140] The system uses the pixels on the edge of the transition region and the pixels on the edge of the adjacent part of the 3D hole region to form a sequence. Then it calculates the difference sequence or the quadratic difference sequence. Next, it checks whether the slope change point of the curve corresponding to the difference sequence or the quadratic difference sequence coincides with the position of the connection point between the edge of the transition region and the edge of the adjacent part of the 3D hole region.

[0141] If the positions do not coincide or the distance is greater than or equal to a set value, it is considered that the edge of the calculated transition region (at least one) is related to the edges of the two adjacent three-dimensional hole regions; otherwise, there is no relationship.

[0142] In some cases, the surface hole region is used as a feature region after edge expansion, as follows:

[0143] Locate the highlight areas near the surface pores and determine the edges of the highlight areas;

[0144] The edges of the highlight areas are merged with the edges of the surface hole areas to maximize the area of ​​the surface hole areas.

[0145] This is because the highlight areas (which have obvious reflective characteristics and are white or grayish-white) are generally protruding edges, and these protruding edges are part of the surface hole area, so they are included in the scope of the surface hole area.

[0146] This application also provides a lithium battery separator manufacturing process optimization device based on intelligent AI, including:

[0147] The image acquisition unit is used to acquire quality inspection images of lithium battery separators and identify feature regions in the quality inspection images.

[0148] The first processing unit is used to determine the center location point of the feature region;

[0149] The second processing unit is used to establish an analysis reference line segment based on the center point of the feature region, and the analysis reference line segment intersects with the edge of the feature region.

[0150] The third processing unit is used to obtain the pixel values ​​at the intersection and use the obtained pixel values ​​to construct the filtering domain;

[0151] The extraction processing unit is used to extract the quality inspection image using the obtained filtering domain to obtain the distinguishing region;

[0152] The computational processing unit is used to calculate the ratio of the area of ​​the distinguishing region to the area of ​​the quality inspection image to obtain porosity data;

[0153] The results output unit is used to compare porosity data with porosity reference data and provide directions for process optimization based on the comparison results.

[0154] Furthermore, identifying characteristic regions in quality inspection images includes:

[0155] Multiple local features are extracted from the quality inspection image;

[0156] Calculate the shape feature area of ​​each local feature;

[0157] The obtained local features are sorted according to the area of ​​the shape features;

[0158] Use the first local feature in the sequential sequence as the background feature;

[0159] When sorting the obtained local features, the numerical value of the shape feature area tends to decrease in the sequential sequence.

[0160] Furthermore, for the remaining local features, it also includes:

[0161] Establish analysis reference lines on the remaining local features. The analysis reference lines include horizontal analysis reference lines and vertical analysis reference lines.

[0162] Use the pixels on the analysis reference line to create the analysis curve;

[0163] The analysis curve is decomposed in the time domain to obtain multiple sets of location points. Each set of location points includes two location points and two location point pixel values.

[0164] All location point groups are merged, and only the two outermost location points are retained;

[0165] A dynamic numerical range is constructed using the pixel values ​​of the retained location points, and this dynamic numerical range is then used to extract the quality inspection image.

[0166] Furthermore, when constructing the dynamic numerical range using the pixel values ​​of the reserved location points, each location point pixel value is located at the midpoint of an independent sub-dynamic numerical range.

[0167] Furthermore, the feature of identifying the characteristic regions in the quality inspection image includes:

[0168] Using color and shape, the hole areas are identified in the quality inspection image.

[0169] The perforated areas are classified into surface perforated areas and three-dimensional perforated areas, based on color difference and area.

[0170] Use the three-dimensional hole region as a feature region;

[0171] The surface hole area is expanded at the edges and then used as a feature area.

[0172] Furthermore, when using the three-dimensional hole region as a feature region, it also includes:

[0173] Determine the transition region between two adjacent three-dimensional hole regions;

[0174] Connect the outer contours of two adjacent three-dimensional hole regions to maximize the area of ​​the region formed by the transition region between the two adjacent three-dimensional hole regions.

[0175] The newly added edges of the constituent regions are used as the edges of the transition regions;

[0176] Calculate whether the edge of the transition region is related to the edges of the two adjacent three-dimensional hole regions. If there is a relationship, merge the transition region and the two adjacent three-dimensional hole regions.

[0177] Furthermore, the surface perforated area is expanded at the edges and used as a feature area, including:

[0178] Locate the highlight areas near the surface pores and determine the edges of the highlight areas;

[0179] The edges of the highlight areas are merged with the edges of the surface hole areas to maximize the area of ​​the surface hole areas.

[0180] In one example, the unit in any of the above devices may be one or more integrated circuits configured to implement the above methods, such as one or more application-specific integrated circuits (ASICs), or one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs), or a combination of at least two of these integrated circuit forms.

[0181] For example, when the units in the device can be implemented through a processing element scheduler, the processing element can be a general-purpose processor, such as a central processing unit (CPU) or other processor capable of calling programs. Alternatively, these units can be integrated together to form a system-on-a-chip (SOC).

[0182] In this application, various objects such as messages / information / devices / network elements / systems / apparatus / actions / operations / processes / concepts may be named. It is understood that these specific names do not constitute a limitation on the relevant objects. The names may be changed depending on the scenario, context, or usage habits. The understanding of the technical meaning of the technical terms in this application should be mainly determined from their functions and technical effects embodied / performed in the technical solution.

[0183] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0184] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0185] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0186] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0187] It should also be understood that in the various embodiments of this application, the terms "first," "second," etc., are merely to indicate that multiple objects are different. For example, a first time window and a second time window are only to indicate different time windows. They should not have any effect on the time windows themselves, and the aforementioned terms "first," "second," etc., should not impose any limitations on the embodiments of this application.

[0188] It should also be understood that, in the various embodiments of this application, unless otherwise specified or in case of logical conflict, the terms and / or descriptions between different embodiments are consistent and can be referenced by each other, and the technical features in different embodiments can be combined to form new embodiments according to their inherent logical relationships.

[0189] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a computer-readable storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned computer-readable storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0190] This application also provides a lithium battery separator manufacturing process optimization system based on intelligent AI, the system comprising:

[0191] One or more memories for storing instructions; and

[0192] One or more processors are configured to retrieve and execute the instructions from the memory, performing the methods described above.

[0193] This application also provides a computer program product including instructions that, when executed, cause the terminal device and the network device to perform operations corresponding to the methods described above.

[0194] This application also provides a chip system including a processor for implementing the functions involved in the above description, such as generating, receiving, transmitting, or processing the data and / or information involved in the above methods.

[0195] This chip system can consist of chips or include chips and other discrete components.

[0196] The processor mentioned above can be a CPU, a microprocessor, an ASIC, or one or more integrated circuits that execute a program to control the method of transmitting the feedback information described above.

[0197] In one possible design, the chip system also includes a memory for storing necessary program instructions and data. The processor and the memory can be decoupled and located on different devices, connected via wired or wireless means to support the chip system in implementing the various functions described in the above embodiments. Alternatively, the processor and the memory can also be coupled to the same device.

[0198] Optionally, the computer instructions are stored in memory.

[0199] Optionally, the memory can be a storage unit within the chip, such as a register or cache. Alternatively, the memory can be a storage unit located outside the chip within the terminal, such as a ROM or other types of static storage devices that can store static information and instructions, such as RAM.

[0200] It is understood that the memory in this application may be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory.

[0201] Non-volatile memory can be ROM, programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory.

[0202] Volatile memory can be RAM, which is used as an external cache. There are many different types of RAM, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus random access memory.

[0203] The embodiments described in this specific implementation are preferred embodiments of this application and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A method for optimizing the production process of lithium battery separators based on intelligent AI, characterized in that, include: Acquire quality inspection images of lithium battery separators and identify feature regions in the quality inspection images; Determine the center point of the feature region; An analysis reference line segment is established based on the center point of the feature region, and the analysis reference line segment intersects with the edge of the feature region. Obtain the pixel values ​​at the intersection points and use the obtained pixel values ​​to construct a filtering region. Using the pixel values ​​as a reference, take the positive and negative error values ​​as endpoints to construct the filtering region. The pixel values ​​are N, and the filtering region is [NM, N+M], where M is a number greater than zero. The obtained filtering domain is used to extract the distinguishing regions from the quality inspection image; The porosity data is obtained by calculating the ratio of the area of ​​the distinguishing region to the area of ​​the quality inspection image; The porosity data is compared with the porosity reference data, and the direction of process optimization is given based on the comparison results.

2. The method for optimizing lithium battery separator production process based on intelligent AI according to claim 1, characterized in that, When identifying feature regions in quality inspection images, the process also includes: Multiple local features are extracted from the quality inspection image; Calculate the shape feature area of ​​each local feature; The obtained local features are sorted according to the area of ​​the shape features; Use the first local feature in the sequential sequence as the background feature; When sorting the obtained local features, the numerical value of the shape feature area tends to decrease in the sequential sequence.

3. The method for optimizing lithium battery separator production process based on intelligent AI according to claim 2, characterized in that, The remaining local features also include: Establish analysis reference lines on the remaining local features. The analysis reference lines include horizontal analysis reference lines and vertical analysis reference lines. Use the pixels on the analysis reference line to create the analysis curve; The analysis curve is decomposed in the time domain to obtain multiple sets of location points. Each set of location points includes two location points and two location point pixel values. All location point groups are merged, and only the two outermost location points are retained; A dynamic numerical range is constructed using the pixel values ​​of the retained location points, and this dynamic numerical range is then used to extract the quality inspection image.

4. The method for optimizing lithium battery separator production process based on intelligent AI according to claim 3, characterized in that, When constructing a dynamic numerical range using the pixel values ​​of the reserved location points, each location point pixel value is located at the midpoint of an independent sub-dynamic numerical range.

5. The method for optimizing lithium battery separator production process based on intelligent AI according to any one of claims 1 to 4, characterized in that, Its features are, Identifying feature regions in quality inspection images includes: Using color and shape, the hole areas are identified in the quality inspection image. The perforated areas are classified into surface perforated areas and three-dimensional perforated areas, based on color difference and area. Use the three-dimensional hole region as a feature region; The surface hole area is expanded at the edges and then used as a feature area.

6. The method for optimizing lithium battery separator production process based on intelligent AI according to claim 5, characterized in that, When using a three-dimensional hole region as a feature region, it also includes: Determine the transition region between two adjacent three-dimensional hole regions; Connect the outer contours of two adjacent three-dimensional hole regions to maximize the area of ​​the region formed by the transition region between the two adjacent three-dimensional hole regions. The newly added edges of the constituent regions are used as the edges of the transition regions; Calculate whether the edge of the transition region is related to the edges of the two adjacent three-dimensional hole regions. If there is a relationship, merge the transition region and the two adjacent three-dimensional hole regions.

7. The method for optimizing lithium battery separator production process based on intelligent AI according to claim 5, characterized in that, Using surface hole areas as feature areas after edge expansion includes: Locate the highlight areas near the surface hole areas and determine the edges of the highlight areas; The edges of the highlight areas are merged with the edges of the surface hole areas to maximize the area of ​​the surface hole areas.

8. A lithium battery separator production process optimization device based on intelligent AI, characterized in that, include: The image acquisition unit is used to acquire quality inspection images of lithium battery separators and identify feature regions in the quality inspection images. The first processing unit is used to determine the center location point of the feature region; The second processing unit is used to establish an analysis reference line segment based on the center point of the feature region, and the analysis reference line segment intersects with the edge of the feature region. The third processing unit is used to obtain the pixel values ​​at the intersection and use the obtained pixel values ​​to construct a filtering domain. The pixel values ​​are used as a reference, and positive and negative error values ​​are used as endpoints to construct the filtering domain. The pixel values ​​are N, and the filtering domain is [NM, N+M], where M is a number greater than zero. The extraction processing unit is used to extract the quality inspection image using the obtained filtering domain to obtain the distinguishing region; The computational processing unit is used to calculate the ratio of the area of ​​the distinguishing region to the area of ​​the quality inspection image to obtain porosity data; The results output unit is used to compare porosity data with porosity reference data and provide process optimization directions based on the comparison results.

9. A lithium battery separator production process optimization system based on intelligent AI, characterized in that, The system includes: One or more memories for storing instructions; and One or more processors are configured to retrieve and execute the instructions from the memory to perform the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes: The program, when run by a processor, executes the method as described in any one of claims 1 to 7.

Images