Battery image processing method, battery detection method and device

Abstract

The invention provides a battery image processing method and device and a battery detection method and device, and belongs to the technical field of batteries. The image processing method comprises the following steps: performing frequency domain transformation on a battery image to obtain a low-frequency information image containing low-frequency information in the battery image and a high-frequency information image containing high-frequency information in the battery image; performing first filtering processing on the low-frequency information image to enhance pole piece features in the low-frequency information image; performing second filtering processing on the high-frequency information image to improve the signal-to-noise ratio of the high-frequency information image; carrying out time domain inverse transformation on a low-frequency filtering image obtained by carrying out first filtering processing on the low-frequency information image and a high-frequency filtering image obtained by carrying out second filtering processing on the high-frequency information image so as to obtain an intermediate processing image corresponding to the battery image; and performing spatial filtering processing on the intermediate processing image to obtain a battery detection image.

Description

Technical Field

[0001] The present application relates to the field of battery technology, and in particular to a battery image processing method, a battery detection method and a battery detection device. Background Art

[0002] Energy conservation and emission reduction are the key to the sustainable development of the automobile industry. Electric vehicles have become an important part of the sustainable development of the automobile industry due to their advantages in energy conservation and environmental protection. For electric vehicles, battery technology is an important factor in their development.

[0003] During the production process of batteries, it is necessary to detect defects in the batteries. In the related art, battery images can be used to perform non-destructive testing on batteries. Although this can achieve non-destructive testing, the accuracy of the test results still needs to be improved. Summary of the invention

[0004] The present application aims to solve at least one of the technical problems in the background technology. To this end, one object of the present application is to provide a battery image processing method, a battery detection method and a device to improve the accuracy of defect detection based on battery images.

[0005] An embodiment of the first aspect of the present application provides a battery image processing method, including: performing frequency domain transformation on a battery image to obtain a low-frequency information image containing low-frequency information in the battery image and a high-frequency information image containing high-frequency information in the battery image; performing a first filtering process on the low-frequency information image to enhance the pole piece features in the low-frequency information image; performing a second filtering process on the high-frequency information image to improve the signal-to-noise ratio of the high-frequency information image; performing a time domain inverse transform on a low-frequency filtered image obtained after the first filtering process on the low-frequency information image and a high-frequency filtered image obtained after the second filtering process on the high-frequency information image to obtain an intermediate processed image corresponding to the battery image; and performing spatial filtering process on the intermediate processed image to obtain a battery detection image.

[0006] In the technical solution of the embodiment of the present application, a first filtering process is performed on a low-frequency information image containing low-frequency information in a battery image to enhance the pole piece features in the low-frequency information image, and a second filtering process is performed on a high-frequency information image containing high-frequency information in the battery image to improve the signal-to-noise ratio of the high-frequency information image, that is, different processing methods are used for the low-frequency information image and the high-frequency information, so that the processing method is more targeted, so as to obtain a clear battery image, so that the cathode pole piece and the anode pole piece in the battery image become clearly visible.

[0007] In some embodiments, performing frequency domain transformation on a battery image to obtain a low-frequency information image containing low-frequency information in the battery image and a high-frequency information image containing high-frequency information in the battery image includes: performing wavelet transformation on the battery image to obtain a low-frequency information image containing low-frequency information in the battery image and a high-frequency information image containing high-frequency information in the battery image. By performing wavelet transformation on the battery image, the wavelet transformation can greatly reduce or remove the correlation between the low-frequency information image containing low-frequency information in the battery image and the high-frequency information image containing high-frequency information in the battery image, so that the obtained low-frequency information image containing low-frequency information in the battery image and the high-frequency information image containing high-frequency information in the battery image are more in line with the requirements for subsequent processing of the battery image.

[0008] In some embodiments, wavelet transform is performed on the battery image to obtain a low-frequency information image containing low-frequency information in the battery image and a high-frequency information image containing high-frequency information in the battery image, including: low-pass filtering the battery image in the pixel row direction and / or the pixel column direction to obtain a low-frequency information image containing low-frequency information in the battery image; high-pass filtering the battery image in the pixel row direction and / or the pixel column direction to obtain a high-frequency information image containing high-frequency information in the battery image. By low-pass filtering the battery image in the pixel row direction and / or the pixel column direction to obtain a low-frequency information image, and high-pass filtering the battery image in the pixel row direction and / or the pixel column direction to obtain a high-frequency information image, features in different directions of the battery image can be extracted, thereby better extracting the pole piece features in the battery image, so that the low-frequency information image and the high-frequency information image can be processed in a targeted manner.

[0009] In some embodiments, the high-frequency information image includes a first high-frequency information image, a second high-frequency information image, and a third high-frequency information image, wherein high-pass filtering the battery image in the pixel row direction and / or pixel column direction to obtain a high-frequency information image containing high-frequency information in the battery image includes: low-pass filtering the battery image in the pixel row direction and high-pass filtering the battery image in the pixel column direction to obtain a first high-frequency information image containing high-frequency information in the battery image; high-pass filtering the battery image in the pixel row direction and low-pass filtering the battery image in the pixel column direction to obtain a second high-frequency information image containing high-frequency information in the battery image; high-pass filtering the battery image in the pixel row direction and the pixel column direction to obtain a third high-frequency information image containing high-frequency information in the battery image. Thus, by performing frequency domain transformation processing on the battery image in different high-pass filtering methods, different features can be extracted, thereby better extracting the pole piece features in the battery image, so that the processed battery image has clearer pole piece features.

[0010] In some embodiments, the first filtering process includes filtering the low-frequency information image in the direction of the pixel row to reduce random noise in the low-frequency information image. By filtering the low-frequency information image in the direction of the pixel row, the random noise in the low-frequency information image can be reduced, and the grayscale value of the anode electrode piece and the cathode electrode piece can be reduced, so as to improve the clarity of the battery image through the subsequent battery image processing step.

[0011] In some embodiments, the first filtering process further includes filtering the low-frequency information image in the direction of the pixel columns to reduce random noise in the low-frequency information image. By filtering the low-frequency information image in the direction of the pixel columns, the random noise in the low-frequency information image can be further reduced, thereby enhancing the characteristics of the cathode and anode pole pieces and improving the overall degree of the cathode and anode pole pieces.

[0012] In some embodiments, the second filtering process includes performing an image thresholding process on the high-frequency information image. By performing an image thresholding process on the high-frequency information image, noise signals similar to real signals can be separated, thereby improving the signal-to-noise ratio of the high-frequency information image.

[0013] In some embodiments, performing image thresholding processing on the high-frequency information image includes: taking the absolute value of the grayscale values ​​of all pixels in the high-frequency information image, and assigning the pixel value of the pixel whose absolute value of the grayscale value of the pixel is greater than the preset threshold as the preset threshold. By taking the absolute value of the grayscale values ​​of all pixels in the high-frequency information image, and assigning the pixel value of the pixel whose absolute value of the grayscale value of the pixel is greater than the preset threshold as the preset threshold, the noise signal similar to the real signal in the high-frequency information image is separated, thereby improving the signal-to-noise ratio of the inner circle at the battery pole piece in the high-frequency information image, making the high-frequency information image clearer.

[0014] In some embodiments, performing spatial filtering on the intermediate processed image to obtain the battery detection image includes: performing a third filtering on the intermediate processed image to reduce high-frequency noise of the intermediate processed image. By performing the third filtering on the intermediate processed image to reduce high-frequency noise of the intermediate processed image, the pole piece features in the intermediate processed image are further enhanced, making the intermediate processed image clearer.

[0015] In some embodiments, the third filtering process is to perform mean filtering on the intermediate processing object. Performing mean filtering on the intermediate processing object can make the intermediate processing image smoother.

[0016] In some embodiments, performing spatial filtering on the intermediate processed image to obtain the battery detection image further includes: performing a fourth filtering on the intermediate processed image to reduce random noise in the intermediate processed image, wherein the fourth filtering is filtering the intermediate processed image in the pixel column direction and / or the pixel row direction. Performing filtering on the intermediate processed image in at least one of the pixel column direction and the pixel column direction to reduce random noise in the intermediate processed image can eliminate random noise in the intermediate processed image again, thereby improving the overall degree of the cathode electrode piece and the anode electrode piece in the intermediate processed image.

[0017] In some embodiments, performing spatial filtering on the intermediate processed image to obtain the battery detection image further includes: blurring the filtered intermediate processed image. Blurring the filtered intermediate processed image removes interference parts other than the battery pole piece in the intermediate processed image, thereby obtaining a clear and visible battery detection image.

[0018] In some embodiments, the method further includes: performing contrast enhancement processing on the battery detection image based on the battery image. The battery detection image obtained after the above processing has a significant change in the grayscale of the pole piece relative to the initial battery image. Therefore, by performing contrast enhancement on the battery detection image based on the battery image, the pole piece of the battery detection image can be clearly visible, and while improving the contrast of the battery detection image, the quality of the battery detection image can also be guaranteed.

[0019] In some embodiments, contrast enhancement processing is performed on a battery detection image based on a battery image, including: subtracting values ​​of multiple pixels of the battery detection image from values ​​of multiple pixels of the battery image one by one; and performing image enhancement processing on the battery detection image after the subtraction operation. By subtracting values ​​of multiple pixels of the battery detection image from values ​​of multiple pixels of the battery image one by one, and performing image enhancement processing on the battery detection image after the subtraction operation, the battery electrode image in the battery detection image can be clearly visible, ensuring that the detection model can accurately identify the cathode electrode and the anode electrode in the battery detection image.

[0020] An embodiment of the second aspect of the present application provides a battery image processing device, the device comprising: a frequency domain transformation module, configured to perform frequency domain transformation on a battery image to obtain a low-frequency information image containing low-frequency information in the battery image and a high-frequency information image containing high-frequency information in the battery image; a first filtering processing module, configured to perform a first filtering processing on the low-frequency information image to enhance the pole piece features in the low-frequency information image; a second filtering processing module, configured to perform a second filtering processing on the high-frequency information image to improve the signal-to-noise ratio of the high-frequency information image; a time domain inverse transformation module, configured to perform a time domain inverse transformation on a low-frequency filtered image obtained after the low-frequency information image is subjected to the first filtering processing and a high-frequency filtered image obtained after the high-frequency information image is subjected to the second filtering processing to obtain an intermediate processed image corresponding to the battery image; and a spatial filtering processing module, configured to perform a spatial filtering processing on the intermediate processed image to obtain a battery detection image.

[0021] The third aspect of the present application provides a battery detection method, which includes: acquiring a battery image; processing the battery image using the method of the aforementioned embodiment to obtain a battery detection image; and performing defect detection on the battery based on the battery detection image. By processing the battery image using the battery image processing method in the aforementioned embodiment to obtain a battery detection image, and performing defect detection on the battery based on the battery detection image, the problem of being unable to detect battery defects due to the battery image being unable to provide sufficiently clear and complete cathode and anode plates can be solved, thereby improving the accuracy of the battery defect detection results.

[0022] In some embodiments, the battery detection method further includes: in response to the defect detection result of the battery indicating that the battery is unqualified, removing the unqualified battery. The battery production method is suitable for large-scale batch production detection, and can prevent defective batteries from entering the market to a certain extent.

[0023] In some embodiments, the defect detection result includes: the misalignment between adjacent cathode and anode pole pieces; wherein, performing defect detection on the battery based on the battery detection image includes: inputting the battery detection image into the battery detection model, and obtaining the misalignment between adjacent cathode and anode pole pieces output by the battery detection model. By inputting the battery detection image into the battery detection model, and obtaining the misalignment between adjacent cathode and anode pole pieces output by the battery detection model, the efficiency of defect detection on the battery can be improved.

[0024] In some embodiments, the training process of the battery detection model includes: obtaining a sample battery detection image; marking the real defect information in the sample battery detection image; inputting the sample battery detection image into the battery detection model, and obtaining the predicted defect information of the sample battery output by the battery detection model; calculating the loss value based on the real defect information and the predicted defect information; and adjusting the parameters of the battery detection model based on the loss value. After the loss value is calculated based on the real defect information and the predicted defect information, the defect detection model is trained according to the loss value, and the relevant parameters in the defect detection model are adjusted, so that the defect detection model learns the difference between the predicted defect information and the real defect information according to the loss value during the iteration process, so that the defect information predicted by the model is infinitely close to the real defect information.

[0025] An embodiment of the fourth aspect of the present application provides a battery detection device, comprising: an image acquisition module, configured to acquire a battery image; an image processing module, configured to process the battery image using the method of the aforementioned embodiment to obtain a battery detection image; and a defect detection module, configured to perform defect detection on the battery based on the battery detection image.

[0026] An embodiment of the fifth aspect of the present application provides an electronic device, comprising at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the battery image processing method in the above-mentioned embodiment or the battery detection method in the above-mentioned embodiment.

[0027] An embodiment of the sixth aspect of the present application provides a battery production device, characterized in that it includes the electronic device in the above embodiment.

[0028] An embodiment of the seventh aspect of the present application provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the battery image processing method in the above embodiment or the battery detection method in the above embodiment.

[0029] An embodiment of the eighth aspect of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the battery image processing method in the above embodiment or the battery detection method in the above embodiment.

[0030] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In the accompanying drawings, unless otherwise specified, the same reference numerals throughout the multiple drawings represent the same or similar parts or elements. These drawings are not necessarily drawn to scale. It should be understood that these drawings only depict some embodiments disclosed in the present application and should not be regarded as limiting the scope of the present application.

[0032] Figure 1 A flowchart of a battery image processing method according to some embodiments of the present application;

[0033] Figure 2 A schematic diagram of a battery image of some embodiments of the present application;

[0034] Figure 3 A schematic diagram of grayscale value distribution of a battery image in some embodiments of the present application;

[0035] Figure 4 A schematic diagram of a low-frequency information image in some embodiments of the present application;

[0036] Figure 5 A schematic diagram of a first high-frequency information image in some embodiments of the present application;

[0037] Figure 6 A schematic diagram of a second high-frequency information image in some embodiments of the present application;

[0038] Figure 7 A schematic diagram of a third high-frequency information image according to some embodiments of the present application;

[0039] Figure 8 A schematic diagram of a low-frequency information image after filtering in some embodiments of the present application;

[0040] Fig. 9 This is a schematic diagram of a first high-frequency information image after filtering processing in some embodiments of the present application;

[0041] Fig.10 This is a schematic diagram of a second high-frequency information image after filtering processing in some embodiments of the present application;

[0042] Fig.11 This is a schematic diagram of a third high-frequency information image after filtering processing in some embodiments of the present application;

[0043] Fig.12 A schematic diagram of grayscale value distribution of an intermediate processed image in some embodiments of the present application;

[0044] Fig.13 A schematic diagram of grayscale value distribution of an intermediate processed image after filtering in some embodiments of the present application;

[0045] Fig.14 A schematic diagram of a battery detection image according to some embodiments of the present application;

[0046] Fig.15 A schematic diagram of a processed battery image according to some embodiments of the present application;

[0047] Fig.16 A schematic diagram of a battery image processed in the related art of some embodiments of the present application;

[0048] Fig.17 A block diagram of a battery image processing device according to some embodiments of the present application;

[0049] Fig.18 A flowchart of a battery detection method according to some embodiments of the present application;

[0050] Fig.19 A flowchart of a battery detection model training process for some embodiments of the present application;

[0051] Fig. 20 A block diagram of a battery detection device according to some embodiments of the present application;

[0052] Fig.21 Flowchart of a battery image detection method according to some other embodiments of the present application. DETAILED DESCRIPTION

[0053] The following embodiments of the technical solution of the present application are described in detail in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present application, and are therefore only used as examples, and cannot be used to limit the scope of protection of the present application.

[0054] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" in the specification and claims of this application and the above-mentioned figure descriptions and any variations thereof are intended to cover non-exclusive inclusions.

[0055] In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "multiple" is more than two, unless otherwise clearly and specifically defined.

[0056] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0057] In the description of the embodiments of the present application, the term "and / or" is only a description of the association relationship of the associated objects, indicating that there may be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0058] In the description of the embodiments of the present application, the term "multiple" refers to more than two (including two). Similarly, "multiple groups" refers to more than two groups (including two groups), and "multiple pieces" refers to more than two pieces (including two pieces).

[0059] In the description of the embodiments of the present application, the technical terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, which are only for the convenience of describing the embodiments of the present application and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the embodiments of the present application.

[0060] In the description of the embodiments of the present application, unless otherwise clearly specified and limited, technical terms such as "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, and it can be the internal connection of two elements or the interaction relationship between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the embodiments of the present application can be understood according to the specific circumstances.

[0061] At present, from the perspective of market development, the application of power batteries is becoming more and more extensive. Power batteries are not only used in energy storage power systems such as hydropower, thermal power, wind power and solar power stations, but also widely used in electric vehicles such as electric bicycles, electric motorcycles, electric cars, as well as military equipment and aerospace and other fields. With the continuous expansion of the application field of power batteries, the market demand is also constantly expanding.

[0062] During the production process of batteries, it is necessary to detect defects in the batteries. Normally, the height of the anode electrode of the battery should be higher than that of the cathode electrode. If the height of the anode electrode is lower than that of the cathode electrode, it may cause excessive lithium ions to be released from the active material of the cathode electrode during the battery charging process, and they cannot be fully inserted into the active material of the anode electrode, resulting in lithium deposition on the surface of the anode electrode, which will cause safety hazards to the battery and bring safety risks.

[0063] In the related art, the method for defect detection of a battery can be: obtaining a battery image through X-rays, the battery image including an anode electrode piece and a cathode electrode piece, preprocessing the battery image using methods such as spatial domain filtering, and then using traditional image algorithms or deep learning detection methods to identify the cathode electrode piece vertices and the anode electrode piece vertices, and determine the height difference between each cathode electrode piece and anode electrode piece.

[0064] However, as the thickness of the battery continues to increase, the penetration ability of X-rays in the middle position of the battery cell along the thickness direction becomes lower and lower. The obtained battery image cannot provide sufficiently clear cathode and anode plates for the algorithm to identify, resulting in the inability to complete battery defect detection.

[0065] In view of the above problems, the present invention proposes a battery image processing method and device, a battery detection method and device, which process the battery image into a state recognizable by a detection model, and then input the image into the detection model for detection to complete battery defect detection.

[0066] The battery image processing method disclosed in the embodiment of the present disclosure can be applied to battery defect detection but is not limited thereto, and can also be applied to the production process of batteries. The battery image processing method of the present application can enable the battery detection image to provide sufficiently clear cathode and anode plates.

[0067] Figure 1 A flowchart of a battery image processing method provided in some embodiments of the present application. Figure 1 As shown, the battery image processing method 100 includes:

[0068] Step S110: performing frequency domain transformation on the battery image to obtain a low-frequency information image containing low-frequency information in the battery image and a high-frequency information image containing high-frequency information in the battery image.

[0069] Step S120: performing a first filtering process on the low-frequency information image to enhance the pole piece features in the low-frequency information image.

[0070] Step S130: performing a second filtering process on the high-frequency information image to improve the signal-to-noise ratio of the high-frequency information image.

[0071] Step S140: performing time domain inverse transform on the low-frequency filtered image obtained after the first filtering process on the low-frequency information image and the high-frequency filtered image obtained after the second filtering process on the high-frequency information image to obtain an intermediate processed image corresponding to the battery image.

[0072] Step S150: performing spatial filtering processing on the intermediate processed image to obtain a battery detection image.

[0073] In some embodiments of the present application, the battery image may be a battery image acquired by X-ray. Figure 2 Schematic diagram of battery images of some embodiments of the present application, such as Figure 2 As shown, the battery image may include the cathode electrode piece and the anode electrode piece of the battery, but the cathode electrode piece and the anode electrode piece of the battery in some positions are blurred and cannot be accurately identified. Figure 3 Schematic diagram of grayscale value distribution of battery images of some embodiments of the present application, in which the horizontal axis is the pixel coordinates in the row direction of the battery image, and the vertical axis is the grayscale value corresponding to the pixel. Figure 3 As shown, the grayscale value in the battery image changes dramatically, that is, the battery image includes noise.

[0074] In some embodiments of the present application, a frequency domain transform is performed on the battery image, the image is transformed from the spatial domain to the frequency domain, and then the battery image can be processed in the frequency domain. The frequency domain transform method used can be a fast Fourier transform method or a wavelet transform method, etc.

[0075] In some embodiments of the present application, the frequency of an image refers to an indicator of the degree of drastic change in the grayscale value of the image, and is the gradient of the grayscale in the plane space. Low-frequency information, i.e., low-frequency components, represent areas in the battery image where the grayscale value changes slowly, corresponding to areas in the battery image where the changes are gentle. Low-frequency information is mainly a comprehensive measure of the intensity of the battery image, i.e., the general overview and outline of the battery image, and is approximate information of the battery image. High-frequency information, i.e., high-frequency components, represent areas in the battery image where the grayscale value changes drastically, corresponding to the edges, outlines, and noise of the battery image. In other words, the low-frequency information image containing the low-frequency information in the battery image contains most of the information of the cathode and anode plates, and the high-frequency information image containing the high-frequency information in the battery image contains the edges and outlines of the cathode and anode plates, as well as some noise in the battery image.

[0076] In some embodiments of the present application, according to the aforementioned embodiments, the electrode pieces in the battery image include cathode electrode pieces and anode electrode pieces, and the low-frequency information image includes cathode electrode pieces and anode electrode pieces. In addition, the low-frequency information image also includes some random noise, which makes the electrode piece features of the cathode electrode pieces and anode electrode pieces unclear. The noise in the low-frequency information image can be filtered out by the first filtering process, and the electrode piece features of the cathode electrode pieces and anode electrode pieces in the low-frequency information image can be enhanced. Exemplarily, the battery width is 42mm, the size of the battery image is 3008*2496, and the length corresponding to one pixel is 9.85um. After the frequency domain transformation, the resolution of the low-frequency information image becomes half of the original to 1504*1248, and the anode electrode pieces and cathode electrode pieces of the battery image are retained in the low-frequency information image. Therefore, the random noise should be eliminated on this image, and the features of the cathode electrode pieces and anode electrode pieces should be enhanced.

[0077] In some embodiments of the present application, the first filtering process may be filtering the low-frequency information image in the direction of pixel rows, or filtering the low-frequency information image in the direction of pixel columns. The filtering method of the first filtering process may be any filtering method such as mean filtering, Gaussian filtering, etc., and the embodiments of the present disclosure do not impose any restrictions on this.

[0078] In some embodiments of the present application, the signal-to-noise ratio of an image is a measure of the quality of the displayed image, and the higher the signal-to-noise ratio, the better the image quality. In the signal-to-noise ratio of an image, the signal refers to the information to be expressed in the image, and the noise refers to the noise introduced by the interference of the signal during the transmission process, such as image blur, color distortion, or strange pixels on the image. The noise in the high-frequency information image can be filtered out by the second filtering process to improve the signal-to-noise ratio of the high-frequency information image. The second filtering process can be filtering the high-frequency information image in the direction of the pixel row, or filtering the high-frequency information image in the direction of the pixel column, or filtering the high-frequency information image in the direction of the pixel row and the direction of the pixel column at the same time. The filtering method of the second filtering process can be any filtering method such as median filtering and Gaussian filtering. These filtering methods can reduce the noise in the image and smooth the image, thereby improving the signal-to-noise ratio of the high-frequency information image.

[0079] In some embodiments of the present application, after performing time domain inverse transform on the low-frequency filtered image obtained after the low-frequency information image is subjected to the first filtering processing and the high-frequency filtered image obtained after the high-frequency information image is subjected to the second filtering processing, the low-frequency filtered image obtained after the first filtering processing and the high-frequency filtered image obtained after the high-frequency information image is subjected to the second filtering processing can be converted from the frequency domain to the spatial domain to obtain an intermediate processed image corresponding to the battery image.

[0080] In some embodiments of the present application, spatial filtering is to enhance the image by filtering, and its purpose is to remove high-frequency noise and interference. The spatial filtering method may include low-pass filtering, high-pass filtering, band-pass filtering, and blur processing. The intermediate processed image is subjected to spatial filtering to filter out noise and interference in the intermediate processed image, and the edge of the intermediate processed image is enhanced to obtain a battery detection image.

[0081] According to an embodiment of the present application, a first filtering process is performed on a low-frequency information image containing low-frequency information in a battery image to enhance the pole piece features in the low-frequency information image, and a second filtering process is performed on a high-frequency information image containing high-frequency information in the battery image to improve the signal-to-noise ratio of the high-frequency information image, that is, different processing methods are used for the low-frequency information image and the high-frequency information, so that the processing method is more targeted, so as to obtain a clear battery image, so that the cathode pole piece and the anode pole piece in the battery image become clearly visible.

[0082] According to some embodiments of the present application, the above step S110 includes: performing wavelet transform on the battery image to obtain a low-frequency information image containing low-frequency information in the battery image and a high-frequency information image containing high-frequency information in the battery image.

[0083] In some embodiments of the present application, the wavelet transform can decompose the image to be processed into wavelets of different frequency bands, and realize multi-scale refinement analysis of the image by performing operations such as scaling, translation and amplification of the wavelet basis, which is characterized by being able to fully highlight the statistical characteristics of the image in local aspects. The wavelet transform can arbitrarily decompose and reconstruct the signal, ensuring that the image is analyzed simultaneously in the frequency domain and the time domain, and can better analyze the image at different resolutions.

[0084] In some embodiments of the present application, for example, a two-dimensional discrete wavelet transform can be used to perform a wavelet transform on a battery image, and a Haar function can be used as a wavelet basis for the wavelet transform to perform a wavelet transform on the battery image, so as to obtain a low-frequency information image containing low-frequency information in the battery image and a high-frequency information image containing high-frequency information in the battery image. Figure 4 This is a schematic diagram of a low-frequency information image in some embodiments of the present application, such as Figure 4 As shown, the low-frequency information image includes the general appearance of the battery image, etc. In addition, the high-frequency information image includes the edge, contour and noise of the battery image, etc.

[0085] In some embodiments of the present application, by performing wavelet transform on the battery image, the wavelet transform can greatly reduce or remove the correlation between the low-frequency information image containing the low-frequency information in the battery image and the high-frequency information image containing the high-frequency information in the battery image, thereby making the low-frequency information image containing the low-frequency information in the battery image and the high-frequency information image containing the high-frequency information in the battery image more in line with the requirements of subsequent processing of the battery image.

[0086] According to some embodiments of the present application, performing a wavelet transform on a battery image to obtain a low-frequency information image containing low-frequency information in the battery image and a high-frequency information image containing high-frequency information in the battery image includes: performing a low-pass filtering on the battery image in a pixel row direction and / or a pixel column direction to obtain a low-frequency information image containing the low-frequency information in the battery image; performing a high-pass filtering on the battery image in at least one of a pixel row direction and a pixel column direction to obtain a high-frequency information image containing the high-frequency information in the battery image.

[0087] In some embodiments of the present application, a low-pass filter may be used to perform low-pass filtering on the battery image in the pixel row direction and / or pixel column direction. Using a low-pass filter for low-pass filtering can suppress or prevent the passage of high-frequency components, thereby removing noise in the battery image and smoothing the edge of the battery image. For example, the low-pass filter may be The low-frequency information image may include most of the information in the battery image, including cathode and anode pole piece information. It is understandable that, to obtain a low-frequency information image containing low-frequency information in the battery image, for example, the battery image may be low-pass filtered only in the pixel row direction to obtain a low-frequency information image containing low-frequency information in the battery image; the battery image may be low-pass filtered in the pixel row direction and low-pass filtered in the pixel column direction to obtain a low-frequency information image containing low-frequency information in the battery image; the battery image may also be low-pass filtered in the pixel row direction and the pixel column direction to obtain a low-frequency information image containing low-frequency information in the battery image.

[0088] In some embodiments of the present application, a high-pass filter may be used to perform high-pass filtering on the battery image in the pixel row direction and / or pixel column direction. Using a high-pass filter for high-pass filtering can suppress or prevent the passage of low-frequency components, thereby enhancing the high-frequency components in the battery image, eliminating the low-frequency components in the battery image, and enhancing the edge of the battery image. For example, the high-pass filter may be It can be understood that to obtain a high-frequency information image containing high-frequency information in the battery image, for example, the battery image can be low-pass filtered only in the pixel row direction to obtain a high-frequency information image containing the high-frequency information in the battery image; the battery image can also be high-pass filtered in the pixel row direction and low-pass filtered in the pixel column direction to obtain a high-frequency information image containing the high-frequency information in the battery image; the battery image can also be high-pass filtered in both the pixel row direction and the pixel column direction to obtain a high-frequency information image containing the high-frequency information in the battery image.

[0089] In some embodiments of the present application, a low-frequency information image is obtained by low-pass filtering the battery image in the pixel row direction and / or pixel column direction, and a high-frequency information image is obtained by high-pass filtering the battery image in the pixel row direction and / or pixel column direction. It is possible to extract features in different directions of the battery image, thereby better extracting the pole piece features in the battery image, so that the low-frequency information image and the high-frequency information image can be processed in a targeted manner.

[0090] According to some embodiments of the present application, a high-frequency information image includes a first high-frequency information image, a second high-frequency information image, and a third high-frequency information image, wherein high-pass filtering the battery image in the pixel row direction and / or the pixel column direction to obtain a high-frequency information image containing high-frequency information in the battery image includes: low-pass filtering the battery image in the pixel row direction, and high-pass filtering the battery image in the pixel column direction to obtain a first high-frequency information image containing high-frequency information in the battery image; high-pass filtering the battery image in the pixel row direction, and low-pass filtering the battery image in the pixel column direction to obtain a second high-frequency information image containing high-frequency information in the battery image; high-pass filtering the battery image in the pixel row direction and the pixel column direction to obtain a third high-frequency information image containing high-frequency information in the battery image.

[0091] In some embodiments of the present application, a low-pass filter is used to perform low-pass filtering on the battery image in the direction of pixel rows, and then a high-pass filter is used to perform high-pass filtering on the battery image in the direction of pixel columns to obtain a first high-frequency information image containing high-frequency information in the battery image. Figure 5 This is a schematic diagram of a first high-frequency information image in some embodiments of the present application, such as Figure 5 As shown, the first high-frequency information image includes the horizontal characteristics of the battery image.

[0092] In some embodiments of the present application, a high-pass filter is used to perform high-pass filtering on the battery image in the direction of pixel rows, and then a low-pass filter is used to perform low-pass filtering on the battery image in the direction of pixel columns to obtain a second high-frequency information image containing high-frequency information in the battery image. Figure 6This is a schematic diagram of a second high-frequency information image in some embodiments of the present application, such as Figure 6 As shown, the second high-frequency information image includes the vertical characteristics of the battery image.

[0093] In some embodiments of the present application, a high-pass filter is used to perform high-pass filtering on the battery image in the pixel row direction and the pixel column direction to obtain a third high-frequency information image containing high-frequency information in the battery image. Figure 7 This is a schematic diagram of a third high-frequency information image in some embodiments of the present application, such as Figure 7 As shown, the third high-frequency information image includes the characteristics of the battery image in the diagonal direction.

[0094] In some embodiments of the present application, for example, the low-pass filter may be The high pass filter can be

[0095] According to the embodiments of the present application, by performing frequency domain transformation processing on the battery image through different high-pass filtering methods, different features can be extracted, thereby better extracting the pole piece features in the battery image, so that the processed battery image has clearer pole piece features.

[0096] According to some embodiments of the present application, the first filtering process includes filtering the low-frequency information image in a pixel row direction to reduce random noise in the low-frequency information image.

[0097] In some embodiments of the present application, Gaussian filtering can be used to filter the low-frequency information image in the pixel row direction. For example, based on the examples of the aforementioned embodiments, since the thickness of the pole piece is approximately 10 pixels on the original image, the size of the convolution kernel is selected by a traversal method, and the convolution kernel size is selected to be 15*1, and the sigma value is 2.

[0098] According to an embodiment of the present application, by filtering the low-frequency information image horizontally in the pixel row direction, random noise in the low-frequency information image can be reduced, and the grayscale values ​​of the anode and cathode electrodes can be reduced, so as to improve the clarity of the battery image through subsequent battery image processing steps.

[0099] According to some embodiments of the present application, the first filtering process further includes filtering the low-frequency information image in a pixel column direction to reduce random noise in the low-frequency information image.

[0100] In some embodiments of the present application, Gaussian filtering can be used to filter the low-frequency information image in the pixel column direction. For example, a longitudinal Gaussian filtering method is used, with a longitudinal convolution kernel size of 1*35 and a sigma value of 5 to filter the low-frequency information image in the pixel column direction. Figure 8 Schematic diagram of low-frequency information image after filtering in some embodiments of the present application, such as Figure 8 As shown, after the first filtering process is performed on the low-frequency information image, the pole piece features in the low-frequency information image are enhanced.

[0101] According to an embodiment of the present application, by filtering the low-frequency information image in the pixel column direction, the random noise in the low-frequency information image can be further reduced, thereby enhancing the characteristics of the cathode pole piece and the anode pole piece and improving the overall degree of the cathode pole piece and the anode pole piece.

[0102] According to some embodiments of the present application, the second filtering process includes performing image thresholding processing on the high-frequency information image.

[0103] In some embodiments of the present application, thresholding the high-frequency information image may include selecting a threshold value for processing, or processing through a threshold range. Thresholding the high-frequency information image is intended to extract the cathode and anode poles in the high-frequency information image, and distinguish the cathode and anode poles from the noise contained in the high-frequency information image. Specifically, the grayscale values ​​of the pixels in the high-frequency information image that are greater than the threshold value are retained, and the grayscale values ​​of the pixels in the high-frequency information image that are less than the threshold value are set to a specific value.

[0104] In the embodiment of the present application, the image thresholding processing method can be any one of binary thresholding, inverse binary thresholding, truncation thresholding, inverse thresholding to 0 or thresholding to 0.

[0105] According to an embodiment of the present application, by performing image thresholding processing on the high-frequency information image, noise signals similar to real signals can be separated, thereby improving the signal-to-noise ratio of the high-frequency information image.

[0106] According to some embodiments of the present application, image thresholding processing of the high-frequency information image includes: taking the absolute value of the grayscale value of all pixels in the high-frequency information image, and assigning the pixel value of the pixel whose absolute value of the grayscale value is greater than a preset threshold to the preset threshold.

[0107] In some embodiments of the present application, the closer the grayscale value of a pixel point in the high-frequency information image is to 0, the smaller the grayscale difference between adjacent pixels is, which is a useful signal. The larger the grayscale value of a pixel point is, the greater the grayscale change of the pixel point in the battery image is, which may be at the edge or there is a noise point. By taking the absolute value of the grayscale value of all pixels in the high-frequency information image and assigning the pixel value of the pixel point whose absolute value of the grayscale value of the pixel point is greater than the preset threshold as the preset threshold, the noise signal similar to the real signal can be separated, thereby further improving the signal-to-noise ratio of the high-frequency information image. Fig. 9This is a schematic diagram of a first high-frequency information image after filtering in some embodiments of the present application. Fig.10 This is a schematic diagram of a second high-frequency information image after filtering in some embodiments of the present application. Fig.11 Schematic diagram of the third high-frequency information image after filtering in some embodiments of the present application, such as Fig. 9 , Fig.10 and Fig.11 As shown, after the first high-frequency information image, the second high-frequency information image and the third high-frequency information image are subjected to image thresholding processing, the noise in the image is reduced and the battery pole pieces in the image are clearer.

[0108] Exemplarily, the grayscale values ​​of all pixels in the high-frequency information image are first processed in absolute values ​​to ensure that the grayscale values ​​of all pixels in the high-frequency image are positive numbers. Secondly, all grayscale values ​​in the image greater than 10 are converted to 10, where the threshold value 10 is an empirical parameter obtained through experiments.

[0109] According to an embodiment of the present application, by taking the absolute value of the grayscale value of all pixels in the high-frequency information image, and assigning the pixel value of the pixel whose absolute value of the grayscale value is greater than the preset threshold to the preset threshold, the noise signal similar to the real signal in the high-frequency information image is separated, thereby improving the signal-to-noise ratio of the inner circle at the battery pole in the high-frequency information image, making the high-frequency information image clearer.

[0110] According to some embodiments of the present application, performing spatial filtering processing on the intermediate processed image to obtain the battery detection image includes: performing a third filtering processing on the intermediate processed image to reduce high-frequency noise of the intermediate processed image.

[0111] In some embodiments of the present application, Fig.12 Schematic diagram of grayscale value distribution of the intermediate processed image of some embodiments of the present application, in which the horizontal axis is the coordinate of the pixel point in the row direction of the intermediate processed image, and the vertical axis is the grayscale value corresponding to the pixel point, such as Fig.12 As shown, the grayscale value in the intermediate processed image changes dramatically, that is, the intermediate processed image includes noise. Fig.13 This is a schematic diagram of the grayscale value distribution of the intermediate processed image after filtering in some embodiments of the present application. After the intermediate processed image is subjected to the third filtering process, the grayscale value variation in the intermediate processed image becomes smooth, that is, the noise in the intermediate processed image is significantly reduced.

[0112] In some embodiments of the present application, after performing time domain inverse transform on the low-frequency filtered image obtained after the low-frequency information image is subjected to the first filtering processing and the high-frequency filtered image obtained after the high-frequency information image is subjected to the second filtering processing, and the corresponding intermediate processed image of the battery image is obtained, the high-frequency noise included in the low-frequency filtered image obtained after the first filtering processing will be displayed in the intermediate processed image. At this time, performing low-frequency filtering on the intermediate processed image can reduce the high-frequency noise of the intermediate processed image, thereby enhancing the electrode features of the cathode electrode and the anode electrode in the intermediate processed image.

[0113] According to an embodiment of the present application, a third filtering process is performed on the intermediate processed image to reduce high-frequency noise of the intermediate processed image, further enhance the pole piece features in the intermediate processed image, and make the intermediate processed image clearer.

[0114] According to some embodiments of the present application, the third filtering process is to perform mean filtering on the intermediate processing object.

[0115] In some embodiments of the present application, mean filtering can be to average the pixels in the neighborhood, which is a low-pass filter. Mean filtering can average the expected pixel values, reduce the original sharp parts of the image, and suppress the high frequency of the image. Exemplarily, a convolution kernel of size 3*1 can be used to perform mean filtering on the intermediate processed image.

[0116] In the embodiment of the present application, mean filtering is performed on the intermediate processing object to make the intermediate processing image smoother.

[0117] According to some embodiments of the present application, performing spatial filtering on the intermediate processed image to obtain a battery detection image also includes: performing a fourth filtering on the intermediate processed image to reduce random noise in the intermediate processed image, wherein the fourth filtering is filtering the intermediate processed image in a pixel column direction and / or a pixel column direction.

[0118] In some embodiments of the present application, the intermediate processed image can be filtered in the direction of pixel rows by using Gaussian filtering. For example, a longitudinal Gaussian filtering method is used, a transverse convolution kernel size is 15*1, and a sigma value of 2 is used to filter the intermediate processed image in the direction of pixel rows. At the same time, the intermediate processed image can also be filtered in the direction of pixel columns by using Gaussian filtering. For example, a longitudinal Gaussian filtering method is used, a longitudinal convolution kernel size is 1*35, and a sigma value of 5 is used to filter the intermediate processed image in the direction of pixel columns.

[0119] Then, the same Gaussian filtering method as the low-frequency signal is used to perform filtering twice in the pixel row direction and the pixel column direction, which serves to eliminate the random noise in the image again and improve the overall degree of the pole piece.

[0120] According to an embodiment of the present application, the intermediate processed image is filtered in the pixel row direction and / or pixel column direction to reduce random noise in the intermediate processed image. The random noise in the intermediate processed image can be eliminated again, thereby improving the overall degree of the cathode electrode and the anode electrode in the intermediate processed image.

[0121] According to some embodiments of the present application, performing spatial filtering on the intermediate processed image to obtain the battery detection image further includes: performing blurring on the filtered intermediate processed image.

[0122] In some embodiments of the present application, blurring is to remove non-essential features in the intermediate processed image, such as noise, interference lines, etc., and at the same time remove the interference parts in the intermediate processed image, so that the necessary features in the intermediate processed image are clearer and the recognition of defects is improved.

[0123] In some embodiments of the present application, Fig.14 This is a schematic diagram of a battery detection image in some embodiments of the present application, such as Fig.14 As shown, the interference parts except the anode electrode piece and the cathode electrode piece in the intermediate processing image are removed to obtain a battery detection image. The battery detection image is more convenient for contrast enhancement of the battery image in subsequent steps to obtain a clear and visible battery image.

[0124] According to an embodiment of the present application, the filtered intermediate processed image is blurred to remove interference portions of the intermediate processed image except for the battery pole pieces, thereby obtaining a clearly visible battery image.

[0125] According to some embodiments of the present application, the battery image processing method further includes: performing contrast enhancement processing on the battery detection image based on the battery image.

[0126] In some embodiments of the present application, different algorithms can be used based on the battery image to perform local contrast enhancement processing or global contrast enhancement processing on the battery detection image to obtain a contrast-enhanced battery detection image. The grayscale value of the cathode electrode and the anode electrode in the contrast-enhanced battery detection image is greater than the grayscale value of the non-electrode position. Among them, different algorithms may include but are not limited to: histogram equalization and its improved algorithm, piecewise linear transformation, gamma mapping, and logarithmic mapping.

[0127] According to an embodiment of the present application, the grayscale of the pole piece of the battery detection image obtained after the above processing is significantly changed compared to the initial battery image. Therefore, by contrast enhancing the battery detection image based on the battery image, the pole piece of the battery detection image can be clearly visible, and the contrast of the battery detection image can be improved while ensuring the image quality of the battery detection image.

[0128] According to some embodiments of the present application, contrast enhancement processing is performed on a battery detection image based on a battery image, including: subtracting values ​​of multiple pixels of the battery detection image from values ​​of multiple pixels of the image one by one; and performing image enhancement processing on the battery detection image after the subtraction operation.

[0129] In some embodiments of the present application, the grayscale of the cathode electrode and the anode electrode of the battery in the battery detection image obtained after blurring the filtered intermediate processing image has changed significantly, so the values ​​of multiple pixels of the battery detection image are subtracted one by one from the values ​​of multiple pixels of the battery image, and the grayscale of the electrode position in the battery detection image will be greater than the grayscale of the non-pole position, but at this time, the grayscale difference between the grayscale of the electrode position and the grayscale of the non-pole position in the battery detection image after the subtraction operation is not obvious. The battery detection image after the subtraction operation can be image enhanced. For example, the battery detection image after the subtraction operation is histogram equalized, and the original 16-bit image is converted into an 8-bit image, so as to improve the contrast of the battery detection image and make the cathode electrode and the anode electrode of the battery detection image clearly visible.

[0130] Fig.15 This is a schematic diagram of a processed battery detection image in some embodiments of the present application. Fig.16 Schematic diagram of a battery image processed in the related art of some embodiments of the present application, compared with Fig.15 and Fig.16 It can be seen that the cathode electrode piece and the anode electrode piece of the battery detection image after processing in some embodiments of the present application are clearer.

[0131] According to an embodiment of the present application, by subtracting the values ​​of multiple pixels of a battery detection image from the values ​​of multiple pixels of a battery image one by one, and performing image enhancement processing on the battery detection image after the subtraction operation, the battery electrode images in the battery detection image can be made clearly visible, ensuring that the detection model can accurately identify the cathode electrode and anode electrode in the battery detection image.

[0132] Fig.17 A block diagram of a battery image processing device according to some embodiments of the present application is shown in FIG. Fig.17As shown, the battery image processing device includes: a frequency domain transformation module 1710, which is configured to perform frequency domain transformation on the battery image to obtain a low-frequency information image containing low-frequency information in the battery image and a high-frequency information image containing high-frequency information in the battery image; a first filtering processing module 1720, which is configured to perform a first filtering processing on the low-frequency information image to enhance the pole piece features in the low-frequency information image; a second filtering processing module 1730, which is configured to perform a second filtering processing on the high-frequency information image to improve the signal-to-noise ratio of the high-frequency information image; a time domain inverse transformation module 1730, which is configured to perform a time domain inverse transformation on a low-frequency filtered image obtained after the low-frequency information image is subjected to the first filtering processing and a high-frequency filtered image obtained after the high-frequency information image is subjected to the second filtering processing, so as to obtain an intermediate processing image corresponding to the battery image; a spatial filtering processing module 1740, which is configured to perform a spatial filtering processing on the intermediate processing image to obtain a battery detection image.

[0133] In some embodiments, the frequency domain transformation module 1710 is configured to perform wavelet transformation on the battery image to obtain a low-frequency information image containing low-frequency information in the battery image and a high-frequency information image containing high-frequency information in the battery image.

[0134] In some embodiments, the frequency domain transformation module 1710 includes: a low-frequency transformation module, configured to perform low-pass filtering on the battery image in the pixel row direction and / or the pixel column direction to obtain a low-frequency information image containing low-frequency information in the battery image; a high-frequency transformation module, configured to perform high-pass filtering on the battery image in the pixel row direction and / or the pixel column direction to obtain a high-frequency information image containing high-frequency information in the battery image.

[0135] In some embodiments, the high-frequency information image includes a first high-frequency information image, a second high-frequency information image and a third high-frequency information image; the high-frequency conversion module is configured to low-pass filter the battery image in the pixel row direction and high-pass filter the battery image in the pixel column direction to obtain a first high-frequency information image containing high-frequency information in the battery image; high-pass filter the battery image in the pixel row direction and low-pass filter the battery image in the pixel column direction to obtain a second high-frequency information image containing high-frequency information in the battery image; high-pass filter the battery image in the pixel row direction and the pixel column direction to obtain a third high-frequency information image containing high-frequency information in the battery image.

[0136] In some embodiments, the first filtering processing module 1720 is configured to filter the low-frequency information image in the pixel row direction to reduce random noise in the low-frequency information image.

[0137] In some embodiments, the first filtering processing module 1720 is configured to filter the low-frequency information image in the direction of pixel columns to reduce random noise in the low-frequency information image.

[0138] In some embodiments, the second filtering processing module 1730 is configured to perform image thresholding processing on the high-frequency information image.

[0139] In some embodiments, the second filtering processing module 1730 is configured to take the absolute value of the grayscale values ​​of all pixels in the high-frequency information image, and assign the pixel value of the pixel whose absolute value of the grayscale value is greater than the preset threshold to the preset threshold.

[0140] In some embodiments, the spatial filtering processing module 1750 is configured to perform a third filtering process on the intermediate processed image to reduce high frequency noise of the intermediate processed image.

[0141] In some embodiments, the third filtering process is to perform mean filtering on the intermediate processing object.

[0142] In some embodiments, the spatial filtering processing module 1750 is configured to perform a fourth filtering process on the intermediate processed image to reduce random noise in the intermediate processed image, wherein the fourth filtering process is to perform filtering process on the intermediate processed image in at least one of a pixel column direction and a pixel row direction.

[0143] In some embodiments, the spatial filtering processing module 1750 is configured to perform blur processing on the filtered intermediate processed image.

[0144] In some embodiments, the battery image processing apparatus further includes: a contrast enhancement processing module configured to perform contrast enhancement processing on the battery image based on the battery detection image.

[0145] In some embodiments, the contrast enhancement processing module is configured to subtract values ​​of multiple pixels of the battery image from values ​​of multiple pixels of the battery detection image one by one; and perform image enhancement processing on the battery image after the subtraction operation.

[0146] The third aspect of the present application provides a battery detection method. Fig.18 A flowchart of a battery detection method provided in some embodiments of the present application is as follows: Fig.18 As shown, the battery detection method includes:

[0147] Step S1810: Acquire a battery image.

[0148] Step S1820: Process the battery image using the battery image processing method in the above embodiment to obtain a battery detection image.

[0149] Step S1830: Perform defect detection on the battery based on the battery detection image.

[0150] According to some embodiments of the present application, the operation of performing defect detection on a battery can be performed by means of a defect detection model or a recognition algorithm, and can be selected according to the actual application and needs. Exemplarily, in order to realize defect detection on a battery, a defect detection model can be pre-trained to obtain a pre-trained defect detection model. After acquiring a battery image, the battery image is processed using the battery image processing method in the aforementioned embodiment to obtain a battery detection image, and the battery detection image is input into the pre-trained defect detection model. After being processed by the defect detection model, the defect detection information of the battery is obtained.

[0151] In an embodiment of the present application, by adopting the battery image processing method in the aforementioned embodiment to process the battery image to obtain a battery detection image, and performing defect detection on the battery based on the battery detection image, the problem of being unable to detect battery defects due to the battery image being unable to provide sufficiently clear and complete cathode and anode plates can be solved, thereby improving the accuracy of the battery defect detection results.

[0152] According to some embodiments of the present application, the battery detection method further includes: in response to a defect detection result of the battery indicating that the battery is unqualified, discarding the unqualified battery.

[0153] In the embodiment of the present application, in response to the battery test result indicating that the battery is defective, the defective battery can be transferred to the waste outlet, and in response to the battery test result indicating that the battery is not defective, the non-defective battery can be subjected to further testing. The battery production method is suitable for large-scale batch production testing, and can prevent defective batteries from entering the market to a certain extent.

[0154] According to some embodiments of the present application, the results of defect detection include: the amount of misalignment between adjacent cathode pole pieces and anode pole pieces; wherein, performing defect detection on a battery based on a battery detection image includes: inputting the battery detection image into a battery detection model, and obtaining the amount of misalignment between adjacent cathode pole pieces and anode pole pieces output by the battery detection model.

[0155] In some embodiments of the present application, in the current battery, the separator exceeds the anode pole piece in both the width and length directions, so that the anode pole piece and the cathode pole piece are insulated. The anode pole piece exceeds the cathode pole piece in both the width and length directions, which to a certain extent prevents excessive lithium ions from being released from the active material of the cathode pole piece during charging and cannot be fully inserted into the active material of the anode pole piece, thereby causing lithium precipitation on the surface of the anode pole piece, posing a safety risk.

[0156] In some embodiments of the present application, the battery detection model may be a convolutional neural network (CNN), a recurrent neural network (RNN), or other deep learning networks, machine learning networks, etc.

[0157] In some embodiments of the present application, whether a battery is defective can be indicated based on the amount of misalignment between adjacent cathode pole pieces and anode pole pieces.

[0158] In an embodiment of the present application, by inputting a battery detection image into a battery detection model and obtaining the misalignment between adjacent cathode pole pieces and anode pole pieces output by the battery detection model, the efficiency of defect detection on the battery can be improved.

[0159] Fig.19 A flowchart of the training process of the battery detection model of some embodiments of the present application, such as Fig.19 As shown in the figure, the training process of the battery detection model includes:

[0160] Step S1910: Obtain a sample battery detection image.

[0161] Step S1920: Mark the real defect information in the sample battery inspection image.

[0162] Step S1930: input the sample battery inspection image into the battery inspection model, and obtain the predicted defect information of the sample battery output by the battery inspection model.

[0163] Step S1940: Calculate the loss value based on the actual defect information and the predicted defect information.

[0164] Step S1950: Adjust the parameters of the battery detection model based on the loss value.

[0165] In some embodiments of the present application, illustratively, a sufficient number of battery images of various types of batteries are obtained, and the battery image processing method in the aforementioned embodiment is used to process the battery images to obtain battery detection images, and then the labelme software is used to annotate the battery anode and cathode electrodes in the battery detection images, and a training set, a validation set, and a test set are randomly generated in a ratio of 6:2:2 to train the battery detection model.

[0166] In some embodiments of the present application, after the loss value is calculated based on the real defect information and the predicted defect information, the defect detection model is trained according to the loss value, and the relevant parameters in the defect detection model are adjusted so that the defect detection model learns the difference between the predicted defect information and the real defect information according to the loss value during the iteration process, so that the defect information predicted by the model is infinitely close to the real defect information.

[0167] Some embodiments of the fourth aspect of the present application provide a battery detection device, Fig. 20 FIG. 1 is a block diagram of a battery detection device according to some embodiments of the present application. Fig. 20 As shown, the battery detection device includes: an image acquisition module 2010, configured to acquire a battery image; an image processing module 2020, configured to process the battery image using the battery image processing method of the aforementioned embodiment to obtain a battery detection image; and a defect detection module 2030, configured to perform defect detection on the battery based on the battery detection image.

[0168] Some embodiments of the present application also provide an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the battery image processing method or the battery detection method in the above-mentioned embodiments.

[0169] Various embodiments of the systems and techniques described above in this application can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0170] An embodiment of the present application provides a battery production device, including the electronic device of the above embodiment.

[0171] Some embodiments of the present application further provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the battery image processing method or the battery detection method in the above-mentioned embodiments.

[0172] Computer readable media can be tangible media that can contain or store programs for use by or in conjunction with instruction execution systems, devices or equipment. Machine readable media can be machine readable signal media or machine readable storage media. Machine readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the above. More specific examples of machine readable storage media can include electrical connections based on one or more lines, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above.

[0173] An embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the battery image processing method or the battery detection method in the above-mentioned embodiment.

[0174] Fig.21 Flowchart of a battery image detection method according to another embodiment of the present application, such as Fig.21 As shown, the method includes:

[0175] Step S2101: low-pass filtering is performed on the battery image in the pixel row direction and / or the pixel column direction to obtain a low-frequency information image containing low-frequency information in the battery image.

[0176] Step S2102: low-pass filtering is performed on the battery image in the direction of pixel rows, and high-pass filtering is performed on the battery image in the direction of pixel columns to obtain a first high-frequency information image containing high-frequency information in the battery image.

[0177] Step S2103: performing high-pass filtering on the battery image in the direction of pixel rows, and performing low-pass filtering on the battery image in the direction of pixel columns, so as to obtain a second high-frequency information image containing high-frequency information in the battery image.

[0178] Step S2104: high-pass filtering is performed on the battery image in the pixel row direction and the pixel column direction to obtain a third high-frequency information image containing high-frequency information in the battery image.

[0179] Step S2105: Filter the low-frequency information image in the pixel row direction to reduce random noise in the low-frequency information image.

[0180] Step S2106: filtering the low-frequency information image in the direction of pixel columns to reduce random noise in the low-frequency information image.

[0181] Step S2107: taking the absolute value of the grayscale values ​​of all pixels in the high-frequency information image, and assigning the pixel value of the pixel whose absolute value of the grayscale value is greater than the preset threshold to the preset threshold, so as to improve the signal-to-noise ratio of the high-frequency information image.

[0182] Step S2108: performing time domain inverse transform on the low-frequency filtered image obtained after filtering the low-frequency information image and the high-frequency filtered image obtained after filtering the high-frequency information image to obtain an intermediate processed image corresponding to the battery image.

[0183] Step S2109: performing mean filtering on the intermediate processed image to reduce high-frequency noise of the intermediate processed image.

[0184] Step S2110: performing a fourth filtering process on the intermediate processed image to reduce random noise in the intermediate processed image, wherein the fourth filtering process is filtering the intermediate processed image in at least one of a pixel column direction and a pixel row direction.

[0185] Step S2111: performing blur processing on the filtered intermediate processed image to obtain a battery detection image.

[0186] Step S2112: subtract the values ​​of the multiple pixels of the battery detection image from the values ​​of the multiple pixels of the battery image in a one-to-one correspondence.

[0187] Step S2113: performing image enhancement processing on the battery detection image after the subtraction operation.

[0188] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application, and they should all be included in the scope of the claims and specification of the present application. In particular, as long as there is no structural conflict, the various technical features mentioned in the various embodiments can be combined in any way. The present application is not limited to the specific embodiments disclosed herein, but includes all technical solutions that fall within the scope of the claims.

Claims

1. A battery image processing method, characterized in that: The method comprises: Performing frequency domain transformation on the battery image to obtain a low-frequency information image containing low-frequency information in the battery image and a high-frequency information image containing high-frequency information in the battery image; Performing a first filtering process on the low-frequency information image to enhance the pole piece features in the low-frequency information image; Performing a second filtering process on the high-frequency information image to improve the signal-to-noise ratio of the high-frequency information image; Performing a time domain inverse transform on a low-frequency filtered image obtained after the low-frequency information image is subjected to the first filtering process and a high-frequency filtered image obtained after the high-frequency information image is subjected to the second filtering process, so as to obtain an intermediate processed image corresponding to the battery image; The intermediate processed image is subjected to spatial filtering processing to obtain a battery detection image.

2. The method according to claim 1, characterized in that The performing frequency domain transformation on the battery image to obtain a low-frequency information image containing low-frequency information in the battery image and a high-frequency information image containing high-frequency information in the battery image comprises: The battery image is subjected to wavelet transformation to obtain the low-frequency information image containing the low-frequency information in the battery image and the high-frequency information image containing the high-frequency information in the battery image.

3. The method according to claim 2, characterized in that The performing wavelet transform on the battery image to obtain a low-frequency information image containing low-frequency information in the battery image and a high-frequency information image containing high-frequency information in the battery image comprises: Performing low-pass filtering on the battery image in a pixel row direction and / or a pixel column direction to obtain a low-frequency information image containing low-frequency information in the battery image; The battery image is subjected to high-pass filtering in a pixel row direction and / or a pixel column direction to obtain a high-frequency information image containing high-frequency information in the battery image.

4. The method according to claim 3, characterized in that The high-frequency information image includes a first high-frequency information image, a second high-frequency information image and a third high-frequency information image, The step of high-pass filtering the battery image in the pixel row direction and / or the pixel column direction to obtain a high-frequency information image containing high-frequency information in the battery image includes: Performing low-pass filtering on the battery image in a pixel row direction and performing high-pass filtering on the battery image in a pixel column direction to obtain the first high-frequency information image containing high-frequency information in the battery image; Performing high-pass filtering on the battery image in the direction of pixel rows and performing low-pass filtering on the battery image in the direction of pixel columns to obtain the second high-frequency information image containing high-frequency information in the battery image; The battery image is subjected to high-pass filtering in the pixel row direction and the pixel column direction to obtain the third high-frequency information image containing the high-frequency information in the battery image.

5. The method according to any one of claims 1 to 4, characterized in that The first filtering process includes filtering the low-frequency information image in a pixel row direction to reduce random noise in the low-frequency information image.

6. The method according to any one of claims 1 to 5, characterized in that The first filtering process further includes filtering the low-frequency information image in a pixel column direction to reduce random noise in the low-frequency information image.

7. The method according to any one of claims 1 to 6, characterized in that The second filtering process includes performing an image thresholding process on the high-frequency information image.

8. The method according to claim 7, characterized in that The performing image thresholding processing on the high-frequency information image comprises: The absolute values ​​of the grayscale values ​​of all the pixels in the high-frequency information image are taken, and the pixel values ​​of the pixels whose absolute values ​​of the grayscale values ​​of the pixels are greater than a preset threshold are assigned to the preset threshold.

9. The method according to any one of claims 1 to 8, characterized in that The performing spatial filtering on the intermediate processed image to obtain a battery detection image comprises: A third filtering process is performed on the intermediate processed image to reduce high frequency noise of the intermediate processed image.

10. The method according to claim 9, characterized in that The third filtering process is to perform mean filtering on the intermediate processing object.

11. The method according to claim 9 or 10, characterized in that: The performing spatial filtering on the intermediate processed image to obtain a battery detection image further comprises: A fourth filtering process is performed on the intermediate processed image to reduce random noise in the intermediate processed image, wherein the fourth filtering process is filtering the intermediate processed image in a pixel column direction and / or a pixel row direction.

12. The method according to any one of claims 9 to 11, characterized in that The performing spatial filtering on the intermediate processed image to obtain a battery detection image further comprises: The intermediate processed image that has been subjected to the filtering processing is blurred.

13. The method according to any one of claims 1 to 12, characterized in that The method further comprises: The battery detection image is subjected to contrast enhancement processing based on the battery image.

14. The method according to claim 13, characterized in that The performing contrast enhancement processing on the battery detection image based on the battery image includes: Subtracting the values ​​of the multiple pixels of the battery detection image from the values ​​of the multiple pixels of the battery image one by one; Performing image enhancement processing on the battery detection image after the subtraction operation.

15. A battery image processing device, characterized in that: The device comprises: a frequency domain transformation module configured to perform frequency domain transformation on the battery image to obtain a low frequency information image containing low frequency information in the battery image and a high frequency information image containing high frequency information in the battery image; A first filtering processing module is configured to perform a first filtering process on the low-frequency information image to enhance the pole piece features in the low-frequency information image; A second filtering processing module is configured to perform a second filtering process on the high-frequency information image to improve the signal-to-noise ratio of the high-frequency information image; a time domain inverse transformation module, configured to perform time domain inverse transformation on a low-frequency filtered image obtained after the low-frequency information image is subjected to the first filtering process and a high-frequency filtered image obtained after the high-frequency information image is subjected to the second filtering process, so as to obtain an intermediate processed image corresponding to the battery image; The spatial filtering processing module is configured to perform spatial filtering processing on the intermediate processed image to obtain a battery detection image.

16. A battery detection method, characterized in that: The method comprises: Get battery image; Processing the battery image using the method described in any one of claims 1 to 14 to obtain a battery detection image; Defect detection is performed on the battery based on the battery detection image.

17. The method according to claim 16, characterized in that The method further comprises: In response to the defect detection result of the battery indicating that the battery is unqualified, the unqualified battery is rejected.

18. The method according to claim 16, characterized in that The defect detection results include: the misalignment between adjacent cathode pole pieces and anode pole pieces; The performing defect detection on the battery based on the battery detection image includes: The battery detection image is input into a battery detection model, and the misalignment between the adjacent cathode pole pieces and anode pole pieces output by the battery detection model is obtained.

19. The method according to claim 18, characterized in that The training process of the battery detection model includes: Obtaining a sample battery detection image; Marking the real defect information in the sample battery inspection image; Inputting the sample battery inspection image into the battery inspection model, and obtaining predicted defect information of the sample battery output by the battery inspection model; Calculating a loss value based on the actual defect information and the predicted defect information; and Parameters of the battery detection model are adjusted based on the loss value.

20. A battery detection device, characterized in that: The device comprises: An image acquisition module, configured to acquire a battery image; an image processing module, configured to process the battery image using the method according to any one of claims 1 to 14 to obtain a battery detection image; and The defect detection module is configured to perform defect detection on the battery based on the battery detection image.

21. An electronic device, characterized in that: include at least one processor; as well as a memory communicatively coupled to the at least one processor; in The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the battery image processing method described in any one of claims 1 to 14 or the battery detection method described in any one of claims 16 to 19.

22. A battery production device, characterized in that: Comprising the electronic device as claimed in claim 21.

23. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the battery image processing method according to any one of claims 1 to 14 or the battery detection method according to any one of claims 16 to 19 is implemented.

24. A computer program product, characterized in that The invention comprises a computer program, which, when executed by a processor, implements the battery image processing method according to any one of claims 1 to 14 or the battery detection method according to any one of claims 16 to 19.

Images