Detection method and detection device for battery material

Abstract

The embodiment of the invention provides a detection method and a detection device for a battery material. The detection method comprises the following steps: acquiring a battery material image; the battery material image is input into a first contour recognition model to obtain middle contour information, and the middle contour information comprises first position information of an external shape of an impurity contour in the battery material image; according to the middle contour information, first contour information is obtained, and the first contour information comprises second position information of the impurity contour; and determining that the battery material is qualified according to the first contour information. The objective of the invention is to accurately detect whether impurities in a battery material image are qualified or not while accurately identifying contour information of the impurities.

Description

Technical Field

[0001] This application relates to the field of battery technology, and more specifically, to a method, testing device, chip, and computer program product for detecting battery materials. Background Technology

[0002] Energy conservation and emission reduction are key to the sustainable development of the automotive industry, and electric vehicles, due to their energy-saving and environmentally friendly advantages, have become an important component of this sustainable development. For electric vehicles, battery technology is a crucial factor in their development.

[0003] During battery charging and discharging, active ions (such as lithium ions) repeatedly insert and extract between the positive and negative electrodes. When the number or size of impurities in the positive electrode material is high, it reduces the proportion of active material in the electrode and affects battery safety performance. Therefore, it is necessary to detect and control impurities in the positive electrode material. How to solve the problem of low detection accuracy in the process of detecting impurities in the positive electrode material is an urgent technical problem to be solved. Summary of the Invention

[0004] This application provides a method, device, chip, and computer program product for detecting battery materials. The aim is to obtain the outer shape of the impurity contour in the battery material by inputting the battery material image into a first contour recognition model, and then obtain the position information of the impurity contour. This improves the detection accuracy of impurities in battery materials and accurately determines whether the impurities in the battery material image are qualified.

[0005] In a first aspect, this application provides a method for detecting battery materials, comprising: acquiring an image of the battery material; inputting the image of the battery material into a first contour recognition model to obtain intermediate contour information, the intermediate contour information including first position information of the outer shape of the impurity contour in the image of the battery material; obtaining first contour information based on the intermediate contour information, the first contour information including second position information of the impurity contour; and determining that the battery material is qualified based on the first contour information.

[0006] In the technical solution of this application embodiment, by inputting the battery material image into the first contour recognition model to obtain the outer shape of the impurity contour in the battery material, and then obtaining the position information of the impurity contour, the detection accuracy of impurities in the battery material can be improved, and the quality of impurities in the battery material image can be accurately determined.

[0007] In some embodiments, determining that the battery material is qualified based on the first contour information includes: obtaining the area value and length value of the impurity based on the first contour information, wherein the length value is the maximum length of the first contour information; and determining that the battery material is qualified if the area value and / or length value meet preset conditions.

[0008] In the technical solution of this application embodiment, impurities are allowed in the battery material. By determining whether the impurities in the battery material are qualified based on the area value and / or length value of the impurities.

[0009] In some embodiments, determining that the battery material is qualified when the area value and / or length value meet preset conditions includes: determining that the battery material is qualified when the area value is less than or equal to a first preset condition; or, determining that the battery material is qualified when the length value is less than or equal to a second preset condition; or, determining that the battery material is qualified when the area value is less than or equal to the first preset condition and the length value is less than or equal to the second preset condition.

[0010] In the technical solution of this application embodiment, the quality of impurities in the battery material is determined by specifically setting a first preset condition and a second preset condition.

[0011] In some embodiments, the impurities include at least one of the following: iron, copper, zinc, copper-zinc alloys, and stainless steel.

[0012] In some embodiments, X-rays are used to irradiate the battery material in a single battery cell to obtain an image of the battery material.

[0013] In the technical solution of this application embodiment, the battery material image is obtained by irradiating the battery material with X-rays, which does not damage the battery and can accurately obtain all impurities in the battery material.

[0014] In some embodiments, the first contour recognition model is trained by: acquiring first sample data, the first sample data including multiple sets of battery material images; obtaining the first contour recognition model based on the first sample data; wherein the multiple sets of battery material images are configured to be generated from the first battery material images without impurities.

[0015] In the technical solution of this application embodiment, by training the neural network and using multiple sets of battery material images, the outer shape of the impurity contour in the battery material image is obtained, which facilitates the subsequent obtaining of accurate impurity contour information based on this.

[0016] In some embodiments, multiple sets of battery material images are configured to generate battery materials from a first battery material image free of impurities, including: generating impurities in the first battery material image according to a first function to form multiple sets of battery material images.

[0017] In the technical solution of this application embodiment, impurities are randomly generated by a first function. Compared with the prior art using real battery material images, this can reduce the workload of manual annotation while obtaining a large amount of battery material data, making it easier to train a better model.

[0018] In some embodiments, the size range of the impurities generated according to the first function is from 3cm×3cm to 50cm×50cm.

[0019] In the technical solution of this application embodiment, during the process of randomly generating impurities, the size range of the impurities is set according to the size of the impurities in the actual battery materials, so as to better simulate the impurities in the battery materials.

[0020] In some embodiments, the grayscale value of the impurity is determined by an image of the first battery material free of impurities.

[0021] In the technical solution of this application embodiment, in order to better simulate the grayscale of impurities in battery materials, the grayscale value of impurities is determined based on a real image of impurity-free battery materials.

[0022] In some embodiments, the grayscale value of an impurity is determined by an image of a first battery material free of impurities, including: obtaining the average grayscale value of the first battery material image corresponding to the impurity within the range of the first battery material image corresponding to the impurity; and obtaining the grayscale value of the impurity based on the average grayscale value.

[0023] In some embodiments, obtaining first contour information based on intermediate contour information includes: inputting the intermediate contour information into a second contour recognition model to obtain the first contour information.

[0024] In some embodiments, the first contour recognition model is further trained by: acquiring second sample data, the second sample data including multiple sets of battery material images obtained by irradiating the battery material with X-rays; and obtaining the first contour recognition model based on the second sample data.

[0025] In the technical solution of this application embodiment, after training with the first sample data, the initial network is trained again with the second sample data so that the final first contour recognition model can better detect the contour information of impurities.

[0026] It should be understood that the first sample data and the second sample data can also be fed into the initial network for training at the same time to obtain the final first contour recognition model. This application does not impose any restrictions on the order of the first sample data and the second sample data.

[0027] In some embodiments, inputting intermediate contour information into a second contour recognition model to obtain first contour information includes: obtaining a sliding window of the second contour recognition model, the sliding window sliding on the intermediate contour information; and obtaining the first contour information based on the sliding window.

[0028] In the technical solution of this application embodiment, after obtaining the outline of the bounding rectangle of the impurity, in order to obtain the true outline information of the impurity further, a sliding window is used to slide on the middle outline information to obtain the true outline information of the impurity.

[0029] In some embodiments, obtaining the first contour information based on the sliding window includes: obtaining a first threshold corresponding to the sliding window based on the sliding window; and obtaining the first contour information when the gray value of each pixel in the intermediate contour information is less than the first threshold.

[0030] In some embodiments, obtaining a first threshold corresponding to a sliding window based on the sliding window includes: traversing each pixel of the intermediate contour information through the sliding window to obtain the standard deviation of the intermediate contour information in the sliding window corresponding to each pixel; and obtaining the first threshold based on the standard deviation.

[0031] In the technical solution of this application embodiment, a first threshold is obtained based on the standard deviation of the intermediate contour information corresponding to each pixel point obtained during the sliding window process. Based on the first threshold, the true contour information of the impurities is accurately obtained.

[0032] In some embodiments, obtaining a first threshold based on the standard deviation includes: obtaining the first threshold according to a first formula, wherein the first formula is:

[0033]

[0034] Where S is the standard deviation. is the standard deviation coefficient, and M is the minimum standard deviation threshold.

[0035] In some embodiments, Set it to 0.3, and M to 1.

[0036] In some embodiments, the size of the sliding window ranges from 5cm×5cm to 7m×7cm.

[0037] In some embodiments, before inputting the intermediate contour information into the second contour recognition model, the method further includes: obtaining the pixel grayscale value variance corresponding to the intermediate contour information based on the intermediate contour information; and determining and excluding noise information in the intermediate contour information based on the pixel grayscale value variance.

[0038] In the technical solution of this application embodiment, by calculating the variance of pixel grayscale values ​​corresponding to the intermediate contour information, noise information caused by other substances in the battery material during the X-ray imaging process or in the battery material is eliminated, so as to more accurately detect impurities in the battery material image.

[0039] In some embodiments, excluding noise information in intermediate contour information based on pixel gray value variance includes: determining that the pixel gray value variance is noise information and excluding the noise information when the pixel gray value variance is less than a second threshold.

[0040] Secondly, this application provides a method for generating a contour recognition model, the method comprising: acquiring first sample data, the first sample data including multiple sets of battery material images; training an initial network to obtain a first contour recognition model based on the first sample data; wherein, the multiple sets of battery material images generate a second battery material image containing impurities from the first battery material image without impurities.

[0041] In some embodiments, generating a second battery material image containing impurities from a first battery material image free of impurities includes: generating impurities in the first battery material image according to a first function to form a plurality of battery material images.

[0042] In some embodiments, the grayscale value of the impurity is determined by an image of the first battery material free of impurities.

[0043] In some embodiments, the grayscale value of the impurity is determined by an image of the first battery material without impurities, including: obtaining an average grayscale value of the first battery material image without impurities within the range of the image of the first battery material without impurities corresponding to the impurity; and obtaining the grayscale value of the impurity based on the average grayscale value.

[0044] In some embodiments, after training with the first sample data, the method further includes: acquiring second sample data, the second sample data including multiple sets of battery material images obtained by irradiating the battery material with X-rays; and training the initial network to obtain a first contour recognition model based on the second sample data.

[0045] Thirdly, this application provides a battery material detection device, including an acquisition unit and a processing unit. The acquisition unit is used to acquire a battery material image; the processing unit is used to input the battery material image into a first contour recognition model to obtain intermediate contour information, the intermediate contour information including first position information of the outer shape of the impurity contour in the battery material image; the processing unit is further used to obtain first contour information based on the intermediate contour information, the first contour information including second position information of the impurity contour; the processing unit is further used to determine that the battery material is qualified based on the first contour information.

[0046] Fourthly, this application provides an energy storage system, which includes a detection device as described in the third aspect or any possible implementation thereof.

[0047] Fifthly, this application provides a chip, including: a processor for calling and running a computer program from a memory, causing a device on which the chip is mounted to perform a detection method as described in the first aspect or any possible implementation thereof.

[0048] Sixthly, this application provides a computer-readable storage medium for storing a computer program that, when executed by a computer, causes the computer to implement the detection method as described in the first aspect or any possible implementation thereof.

[0049] In a seventh aspect, this application provides a computer program product, comprising: computer program instructions, which, when executed by a computer, cause the computer to implement the detection method as described in the first aspect or any possible implementation thereof. Attached Figure Description

[0050] Figure 1 A schematic flowchart of a detection system according to an embodiment of this application is shown;

[0051] Figure 2 A schematic flowchart of a battery material detection method provided in an embodiment of this application is shown;

[0052] Figure 3 This illustration shows a schematic diagram of intermediate contour information provided in an embodiment of this application;

[0053] Figure 4 A schematic flowchart of a training method for a first contour recognition model provided in an embodiment of this application is shown;

[0054] Figure 5 A schematic diagram showing an image of a battery material provided in an embodiment of this application is illustrated;

[0055] Figure 6 A schematic flowchart of another training method for a first contour recognition model provided in an embodiment of this application is shown;

[0056] Figure 7 A schematic flowchart of another battery material detection method provided in an embodiment of this application is shown;

[0057] Figure 8 A schematic flowchart of another battery material detection method provided in an embodiment of this application is shown;

[0058] Figure 9 A schematic diagram of impurities in a battery material provided in an embodiment of this application is shown;

[0059] Figure 10 A schematic block diagram of a battery material detection device provided in an embodiment of this application is shown;

[0060] Figure 11 A schematic block diagram of a contour recognition model generation apparatus provided in an embodiment of this application is shown;

[0061] Figure 12 A schematic block diagram of a battery material detection device provided in an embodiment of this application is shown;

[0062] The accompanying drawings are not drawn to scale. Detailed Implementation

[0063] The embodiments of this application will be described in further detail below with reference to the accompanying drawings and examples. The detailed description of the following embodiments and the accompanying drawings are used to illustrate the principles of this application by way of example, but should not be used to limit the scope of this application, that is, this application is not limited to the described embodiments.

[0064] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0065] The term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. Additionally, the character " / " in this text generally indicates that the preceding and following related objects have an "or" relationship.

[0066] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0067] Unless otherwise specified, all embodiments and optional embodiments of this application can be combined to form new technical solutions.

[0068] Unless otherwise specified, all technical features and optional technical features of this application may be combined to form new technical solutions.

[0069] Unless otherwise specified, all steps of this application may be performed sequentially or randomly, preferably sequentially. For example, the method includes steps (a) and (b), indicating that the method may include steps (a) and (b) performed sequentially, or it may include steps (b) and (a) performed sequentially. For example, the mention that the method may also include step (c) indicates that step (c) may be added to the method in any order; for example, the method may include steps (a), (b), and (c), or it may include steps (a), (c), and (b), or it may include steps (c), (a), and (b), etc.

[0070] In this embodiment of the application, the battery cell can be a secondary battery, which refers to a battery cell that can be recharged to activate the active materials and continue to be used after the battery cell has been discharged.

[0071] The battery cell can be a lithium-ion battery, sodium-ion battery, sodium-lithium-ion battery, lithium metal battery, sodium metal battery, lithium-sulfur battery, magnesium-ion battery, nickel-metal hydride battery, nickel-cadmium battery, lead-acid battery, etc., and the embodiments of this application are not limited to this.

[0072] A single battery cell typically includes an electrode assembly. The electrode assembly includes a positive electrode, a negative electrode, and a separator, with the separator positioned between the positive and negative electrodes. During the charging and discharging process of a single battery cell, active ions (such as lithium ions) repeatedly insert and extract between the positive and negative electrodes. The separator, positioned between the positive and negative electrodes, prevents short circuits while allowing active ions to pass through.

[0073] In some embodiments, the positive electrode may be a positive electrode sheet, which may include a positive electrode current collector and a positive electrode active material disposed on at least one surface of the positive electrode current collector.

[0074] As an example, the positive current collector has two surfaces opposite each other in its own thickness direction, and the positive active material is disposed on either or both of the two opposite surfaces of the positive current collector.

[0075] As an example, the positive current collector can be a metal foil, a conductive polymer material, a carbon material, or a composite current collector. For example, as a metal foil, pure metals, alloys, or surface-treated metals can be used, including but not limited to stainless steel, copper, aluminum, nickel, titanium, or silver. The composite current collector may include a polymer material base layer and a metal layer. The composite current collector can be formed by forming a metal material (aluminum, aluminum alloys, nickel, nickel alloys, titanium, titanium alloys, silver, and silver alloys, etc.) on a polymer material substrate (such as a substrate of polypropylene, polyethylene terephthalate, polybutylene terephthalate, polystyrene, polyethylene, etc.).

[0076] As an example, the positive electrode active material may include at least one of the following materials: lithium phosphate, lithium transition metal oxide, and their respective modified compounds. However, this application is not limited to these materials, and other conventional materials that can be used as battery positive electrode active materials may also be used. These positive electrode active materials may be used alone or in combination of two or more. Examples of lithium phosphate may include, but are not limited to, at least one of lithium iron phosphate (such as LiFePO4 (also referred to as LFP)), lithium iron phosphate and carbon composites, lithium manganese phosphate (such as LiMnPO4), lithium manganese phosphate and carbon composites, lithium iron manganese phosphate, and lithium iron manganese phosphate and carbon composites. Examples of lithium transition metal oxide may include, but are not limited to, lithium cobalt oxide (such as LiCoO2), lithium nickel oxide (such as LiNiO2), lithium manganese oxide (such as LiMnO2, LiMn2O4), lithium nickel cobalt oxide, lithium manganese cobalt oxide, lithium nickel manganese oxide, and lithium nickel cobalt manganese oxide (such as LiNi). 1 / 3 Co 1 / 3 Mn 1 / 3 O2 (also known as NCM) 333 LiNi 0.5 Co 0.2 Mn 0.3 O2 (also known as NCM) 523 LiNi 0.5 Co 0.25 Mn 0.25 O2 (also known as NCM) 211 LiNi 0.6 Co 0.2 Mn 0.2 O2 (also known as NCM) 622 LiNi 0.8 Co 0.1 Mn 0.1 O2 (also known as NCM) 811 At least one of the following: lithium nickel cobalt aluminum oxides (such as LiNi0.8Co0.15Al0.05O2) and their modified compounds. Modified compounds refer to substances obtained by doping or coating, etc., based on the aforementioned substances.

[0077] In some embodiments, the negative electrode may be a negative electrode sheet, and the negative electrode sheet may include a negative electrode current collector.

[0078] As an example, the negative electrode current collector can be a metal foil, a conductive polymer material, a carbon material, or a composite current collector. For example, as a metal foil, pure metals, alloys, or surface-treated metals can be used, including but not limited to stainless steel, copper, aluminum, nickel, titanium, or silver. The composite current collector may include a polymer material substrate and a metal layer. The composite current collector can be formed by forming a metal material (copper, copper alloys, nickel, nickel alloys, titanium, titanium alloys, silver, and silver alloys, etc.) on a polymer material substrate (such as a substrate of polypropylene, polyethylene terephthalate, polybutylene terephthalate, polystyrene, polyethylene, etc.).

[0079] As an example, the negative electrode sheet may include a negative electrode current collector and a negative electrode active material disposed on at least one surface of the negative electrode current collector.

[0080] As an example, the negative electrode current collector has two surfaces opposite each other in its own thickness direction, and the negative electrode active material is disposed on either or both of the two opposite surfaces of the negative electrode current collector.

[0081] As an example, negative electrode active materials can be filled or / and deposited within the negative electrode current collector.

[0082] In some embodiments, the positive current collector can be made of aluminum, and the negative current collector can be made of copper.

[0083] During the charge-discharge cycles of a battery, as ions are inserted into or extracted from the positive and negative electrode active materials, the battery's performance is closely related to the quality of the positive electrode material. Excessive content of impurities such as iron (Fe), copper (Cu), zinc (Zn), copper-zinc alloys, and stainless steel in the positive electrode material severely damages battery performance and noise levels. Furthermore, excessive metallic impurities reduce the proportion of active materials in the electrode and can catalyze side reactions between the electrode material and the electrolyte. Large metal particles may also puncture the separator, causing internal short circuits and posing safety hazards.

[0084] To alleviate the problem of internal short circuits in batteries caused by excessive impurities in the cathode material of individual battery cells, this application proposes a method for detecting battery materials. By inputting an image of the battery material into a first contour recognition model, intermediate contour information of impurities in the battery material image is obtained. This intermediate contour information includes first position information of the outer shape of the impurity contour. Based on the intermediate contour information, first contour information is obtained, which includes second position information of the impurity contour. Based on the first contour information, the impurities in the battery material image are determined to be acceptable.

[0085] By inputting the image of the battery material into the first recognition model, the outline information of the circumscribed rectangle of the impurities in the battery material is obtained, and the true outline information of the impurities is obtained based on the outline information of the circumscribed rectangle.

[0086] After obtaining precise contour information of the impurities, it is possible to accurately determine whether the impurities in the battery material image are qualified based on the contour information.

[0087] It should be understood that the battery materials described above refer to the positive or negative electrode materials in the battery cell. Excessive impurities in the positive or negative electrode materials will affect the performance of the battery.

[0088] In this application, the battery material can be the positive electrode material, negative electrode material, electrolyte, and separator of the battery. In this application, taking the positive electrode material of the battery as an example, the positive electrode material is usually a lithium-containing transition metal oxide or other materials that can perform lithium ion insertion and extraction. These materials usually exist in powder form. The technical solutions described in the embodiments of this application are applicable to various laboratories or power battery production lines.

[0089] like Figure 1 The diagram shown is a structural schematic of a detection system 1 according to an embodiment of this application. Taking a power battery production line 2 as an example, the detection system includes a conveyor belt 10, a sensor 20, a detection unit 30, and battery materials 40. The battery cell winding process involves sequentially stacking the positive electrode material, separator, and negative electrode material, and then winding them into a cell via the conveyor belt 10. During the transport of the battery materials 40 on the conveyor belt 10, the sensor 20 collects images of the positive electrode material and uploads these images to the detection unit 30, which performs detection on the battery materials. The sensor 20 can also be a photographic device; this application does not limit its use in this regard.

[0090] The detection unit 30 can be a dedicated processor such as a digital signal processor, a field-programmable gate array (FPGA), a graphics processing unit (GPU), a microcontroller unit (MCU), and a distributed computing node, etc. This application embodiment does not impose any special limitations on the detection unit 30 described above.

[0091] Figure 2 A schematic flowchart of a battery material detection method 100 provided in an embodiment of this application is shown.

[0092] According to some embodiments of this application, such as Figure 2 As shown, the detection method 100 for the battery material includes the following:

[0093] S110: Acquire images of battery materials.

[0094] S120: Input the battery material image into the first contour recognition model to obtain intermediate contour information, which includes the first position information of the outer shape of the impurity contour in the battery material image.

[0095] S130: Based on the intermediate contour information, the first contour information is obtained, which includes the second position information of the impurity contour.

[0096] S140: Based on the first profile information, the battery material is determined to be qualified.

[0097] Specifically, the execution entity of this method embodiment can be a detection unit; optionally, the detection unit in this application can be... Figure 1 The detection unit 30 in the middle.

[0098] In the above S110, the battery material image acquired by the detection unit can be obtained through a camera, sensor, video capture device, scanning device, etc., and this application does not limit it in any way.

[0099] Optionally, images of the battery materials can be obtained by irradiating the battery cell with X-rays. X-ray technology allows direct irradiation of the positive electrode material of the battery cell without damaging the battery, and it can accurately identify all impurities in the battery material with high accuracy.

[0100] Specifically, X-rays are generated by an X-ray generator in the sensor. The generated X-rays pass through the positive electrode material of the battery cell. The X-rays that pass through the positive electrode material of the battery cell are then received by the sensor and processed to obtain an image of the battery material.

[0101] It should be understood that, in the embodiments of this application, the battery material described takes the positive electrode material in a battery cell as an example. The positive electrode material is usually a lithium-containing transition metal oxide or other materials that can perform lithium ion insertion and extraction. These materials are usually in powder form.

[0102] In S120 above, the detection unit inputs the battery material image into the first contour recognition model, and learns intermediate contour information from the first contour recognition model. The intermediate contour information includes the first position information of the outer shape of the impurity contour. More specifically, the first contour information is the position information of the outer rectangle of the true impurity contour, which can be referenced... Figure 3 The rectangle is the bounding rectangle of the impurity, used to describe the approximate location of the impurity.

[0103] It should be understood that the location information of the impurity contour typically refers to the boundary, shape, and specific coordinates of the impurity in the image coordinate system of the battery material image. Specifically, this information can include boundary point coordinates, contour lines, and the geometric center. Boundary point coordinates are the (x, y) coordinates of each boundary point in the impurity, representing the specific position of each point on the impurity contour. The contour line is the line segment or curve connecting all boundary points, representing the edge of the impurity. The geometric center is the coordinate of the geometric center point of the impurity contour, usually obtained by calculating the average of all boundary point coordinates. The circumscribed shape of the impurity contour refers to the smallest geometric shape that can completely enclose all foreground objects or pixels in the impurity. This shape can be rectangular, circular, or polygonal, for example. In this application, the circumscribed shape of the impurity contour obtained through the first contour recognition model is a rectangle.

[0104] Optionally, the impurities in this application include iron (Fe), copper (Cu), zinc (Zn), alloys, etc. The alloys may include copper-zinc alloys, stainless steel, etc. If the impurity content is too high or the area is too large, it will seriously affect the performance of the battery. For example, if the impurity content is too high or the area is too large, it will catalyze the side reactions of the electrolyte in the electrode material. Large metal particles may also puncture the separator, causing a short circuit in the battery and causing safety hazards.

[0105] In S130 above, the detection unit obtains first contour information based on the intermediate contour information. The first contour information includes the second position information of the impurity contour. For example, if the impurity is elliptical, after passing through the first contour recognition model, a rectangular contour is obtained. This rectangle is the circumscribed rectangle of the impurity. Then, based on the rectangular contour, an elliptical impurity shape is obtained, which is the true shape of the impurity.

[0106] In S140 above, the detection unit determines whether the battery material is qualified based on the first contour information. By detecting whether the battery material is qualified, the performance and safety of the battery can be improved.

[0107] It should be understood that the qualification of battery materials as referred to in this application means that the qualification of battery materials is determined by first judging whether the impurities in the battery material image are qualified.

[0108] Optionally, based on the first contour information, the area and length values ​​of the impurity are obtained. If the area and / or length values ​​meet preset conditions, the impurity in the battery material image is determined to be qualified. The area value is obtained by converting the line integral on the contour into the area of ​​the region enclosed by the contour. The length value is the maximum length of the first contour information. The maximum length value refers to the longest segment among the continuous, uninterrupted line segments on the impurity contour. The impurity contour consists of multiple continuous line segments, and the length of the longest line segment is the maximum length value.

[0109] Specifically, if the area value is less than or equal to the first preset condition, the battery material is determined to be qualified; if the area value is greater than the first preset condition, the battery material is determined to be unqualified. Alternatively, if the length value is less than or equal to the second preset condition, the battery material is determined to be qualified. Or, if the area value is less than or equal to the first preset condition and the length value is less than or equal to the second preset condition, the battery material is determined to be qualified.

[0110] One method to determine whether battery materials are qualified is by measuring the area of ​​impurities, or by measuring the maximum length, or by using both, namely the area and length, to more accurately determine whether battery materials are qualified.

[0111] It should be understood that the numerical settings for the first and second preset conditions can be set according to specific circumstances, such as actual production requirements, and this application does not impose any limitations on them.

[0112] The technical solution of this application embodiment uses a first identification model to initially locate impurities, and then uses a second identification model to accurately identify them. After obtaining the precise contour information of the impurities, the battery material's qualification can be accurately determined based on this contour information.

[0113] Optionally, the first contour recognition model can be a neural network model used to identify the contour information of the detected battery material image. For example, YOLOv5 can be used for training to obtain the first contour recognition model. The specific training process is as follows:

[0114] Figure 4 A schematic flowchart of a training method 200 for a first contour recognition model provided in an embodiment of this application is shown.

[0115] According to some embodiments of this application, such as Figure 4 As shown, the training method 200 for the first contour recognition model includes the following:

[0116] S210: Acquire first sample data, which includes multiple sets of battery material images.

[0117] S220: Based on the first sample data, a first contour recognition model is obtained; wherein, multiple sets of battery material images are configured to be generated from the first battery material image without impurities.

[0118] Specifically, in this embodiment of the method, the detection unit can also be a cloud platform, a server, or a computer device, etc. Optionally, the execution entity in this method can also be... Figure 1 The detection unit 30 in the middle.

[0119] It should be understood that the training method 200 of the first contour recognition model described above can also be described as a method 200 for generating a contour recognition model.

[0120] In S210 above, the detection unit acquires first sample data, which includes multiple sets of battery material images.

[0121] Optionally, during training, if the first sample data is selected as real battery material images obtained through X-ray technology, training requires a large amount of data, while real battery material images obtained through X-ray technology are relatively few. Furthermore, impurities in the real battery material images need to be labeled before training, a very time-consuming process. Therefore, in this application, impurities are generated in the first battery material images according to a first function to form a large number of second battery material images. This eliminates the need for extensive manual labeling while providing a large amount of training data for training. The first function is a random function, such as a uniformly distributed random function or a normally distributed random function.

[0122] Specifically, an image of the impurity-free battery material is obtained using X-ray technology, namely the first battery material image described above, which can be referenced. Figure 5 (a) Based on the image of the first battery material, impurities are randomly generated using a first function. During the generation of impurities using the first function, the size and grayscale values ​​of the impurities need to be considered to ensure consistency with the size and grayscale of impurities in the real battery material, thus better fitting the size of impurities in the real battery material.

[0123] The second powder image generated by the first function can be referenced. Figure 5 (b) in the middle.

[0124] It should be understood that the purpose of labeling the initial sample data before training is to provide the ground truth for the initial network, i.e., the exact location and category of each object. These ground truth values ​​serve as the benchmark used to guide the initial network in learning target features during training. Furthermore, the labeled data also acts as a benchmark for evaluating model performance. By comparing the model's predictions with the labeled ground truth values, various performance metrics can be calculated, such as precision, recall, and average precision (AP).

[0125] It should also be understood that the initial network described in this application refers to the state of the first contour recognition model before training, that is, the initial configuration of all components of the first contour recognition model, such as its structure, parameters, activation function, loss function, and optimizer, before training begins. This initial state is the starting point of the training process. As training progresses, the network parameters will be gradually adjusted according to the training data to eventually form the first contour recognition model.

[0126] Optionally, the size range of the impurities generated according to the first function can be from 3cm×3cm to 50cm×50cm to better fit the actual impurity size. For example, a size of 50cm×50cm, or 3cm×3cm, or 10cm×10cm can be selected during the random generation process.

[0127] Optionally, the grayscale value of the impurity generated according to the first function is determined by an image of the first battery material without impurities.

[0128] Specifically, given a fixed impurity size, the average grayscale value of the corresponding first battery material image is obtained within the range of the impurity's corresponding first battery material image. Based on this average grayscale value, the grayscale value of the impurity is obtained. The grayscale value of the impurity is 40% to 60% of the average grayscale value of the first powder image; for example, the grayscale value of the impurity is set to 40%, 50%, or 60% of the average grayscale value of the first powder image. Obtaining the grayscale value of the impurity from an impurity-free first battery material image simplifies the operation and allows for a better fit to the actual impurity's grayscale value.

[0129] In the above S220, during the process of the detection unit training the initial network to obtain the first contour recognition model based on the first sample data, it is necessary to adjust the parameters of the initial network so that the loss function value of the initial network during the training process is as low as possible, thereby improving the prediction accuracy of the first contour recognition model.

[0130] Specifically, the detection unit adjusts the parameters of the initial network based on the first sample data to train the first contour recognition model. The parameters include the number of samples per batch, the maximum number of iterations, the type of optimizer, the initial learning rate, the momentum parameter, and the weight decay coefficient, etc.

[0131] In some embodiments, the number of samples per batch can be set to 16, the maximum number of iterations can be set to 350, the optimizer can be an adaptive moment estimator optimizer, the initial learning rate can be set to 0.001, the momentum parameter can be 0.937, and the weight decay coefficient can be 0.0005. It should be understood that the above parameter values ​​can be adjusted according to actual training. For example, the choice of optimizer can be adjusted according to actual conditions, and this application does not impose any limitations on this.

[0132] Optionally, the first contour recognition model can be obtained through two training iterations. The first iteration is the training process of the aforementioned first contour recognition model training method 200, which involves training the initial network by inputting randomly generated impurities. To make the final generated first contour recognition model more accurate, a second iteration is performed by inputting real images of battery materials containing impurities, as detailed below:

[0133] Figure 6 A schematic flowchart of another training method 300 for a first contour recognition model provided in an embodiment of this application is shown.

[0134] According to some embodiments of this application, such as Figure 6 As shown, the training method 300 for the first contour recognition model includes the following:

[0135] S310: Acquire second sample data, which includes multiple sets of battery material images obtained by irradiating the battery material with X-rays.

[0136] S320: Based on the second sample data, obtain the first contour recognition model.

[0137] Specifically, the execution entity in this method embodiment may be a detection unit; optionally, the execution entity in this method may be... Figure 1 The detection unit 30 in the middle.

[0138] It should be understood that the training process of the contour recognition model training method 200 and the contour recognition model training method 300 is the same, and will not be described again in this application.

[0139] It should also be understood that the first sample data and the second sample data can be fed into the initial network for training at the same time to obtain the final first contour recognition model. This application does not impose any restrictions on the training order of the first sample data and the second sample data.

[0140] Optionally, the first contour information obtained after the battery material image is processed by the first contour recognition model may also contain some noise information, such as noise introduced during X-ray imaging or other substances in the battery material. During the detection process, the noise information needs to be eliminated to obtain more accurate contour information of the impurities, as follows:

[0141] Figure 7 A schematic flowchart of another battery material detection method 400 provided in an embodiment of this application is shown.

[0142] According to some embodiments of this application, such as Figure 7 As shown, the detection method 400 for the battery material includes the following:

[0143] S410: Based on the intermediate contour information, obtain the variance of the pixel grayscale values ​​corresponding to the intermediate contour information.

[0144] S420: Based on the variance of pixel grayscale values, determine and exclude noise information in the intermediate contour information.

[0145] Specifically, the execution entity of this method embodiment can be a detection unit; optionally, the execution entity of this method embodiment can be... Figure 1 The detection unit 30 in the middle.

[0146] In the above S410, the detection unit obtains the pixel gray value variance corresponding to the contour information of each impurity based on the intermediate contour information.

[0147] Specifically, based on the outline of the impurities, the gray values ​​of the pixels within the outline are calculated, and the average value of the gray values ​​of all pixels within the outline is calculated. For the gray value of each pixel, the square of the difference between it and the mean is calculated, and then the average of these squared differences is obtained, which is the variance.

[0148] In S420 above, the detection unit determines and excludes noise information in the intermediate contour information based on the pixel grayscale value variance. Specifically, if the pixel grayscale value variance is less than a second threshold, the pixel grayscale value variance is determined to be noise information and excluded. For example, the second threshold is set to 10 based on the grayscale value of impurities in the battery material. More specifically, taking the center point of each detection box as the center point of a box the size of the impurity, the variance of all pixels in the intermediate contour information corresponding to each box is calculated. The variance of impurities is at least 10, and target boxes with a variance below 10 are excluded as noise by default.

[0149] Optionally, after the battery material image is processed by the first contour recognition model to obtain intermediate contour information, the intermediate contour information is then input into the second contour recognition model to obtain the true contour information of the impurities, as follows:

[0150] Figure 8 A schematic flowchart of another battery material detection method 500 provided in an embodiment of this application is shown.

[0151] According to some embodiments of this application, such as Figure 8 As shown, the detection method 500 for the battery material includes the following:

[0152] S510: Obtain the sliding window of the second contour recognition model, and slide the sliding window over the intermediate contour information.

[0153] S520: Obtain the first contour information based on the sliding window.

[0154] Specifically, the execution entity in this method embodiment can be a detection unit; optionally, the execution entity in this method embodiment can be... Figure 1 The detection unit 30 in the middle.

[0155] In the above S510, the sliding window in the first contour recognition model refers to the convolution kernel, that is, by setting the size of the convolution kernel, it slides on the intermediate contour information, that is, it performs convolution operation with the intermediate contour information.

[0156] In the above S520, based on the second contour recognition model, the sliding window traverses each pixel of the first contour information to obtain the standard deviation of the first contour information in the sliding window corresponding to each pixel; based on the standard deviation, the first threshold is obtained. This can also be described as using a two-dimensional convolution kernel to convolve the first contour information. The size of the convolution kernel (i.e., the sliding window) can be set to 7cm × 7cm, and each pixel in the first contour information is traversed using the center of this convolution kernel.

[0157] More specifically, based on the first formula, the first threshold is obtained. The first formula is:

[0158]

[0159] Where S is the standard deviation. is the standard deviation coefficient, and M is the minimum standard deviation threshold.

[0160] In some embodiments, It can be set to 0.3, and M can be set to 1.

[0161] In some embodiments, the size of the sliding window ranges from 5cm×5cm to 7cm×7cm. For example, it can be set to 5cm×5cm, or 6cm×6cm, or even 7cm×7cm.

[0162] Optionally, the first contour information is obtained if the grayscale value of each pixel in the first contour information is less than a first threshold. That is, the pixel values ​​of the first contour information are obtained according to the second formula, which can be referenced. Figure 9 As shown, impurities and battery materials are distinguished based on grayscale values, with white dots representing impurities and black parts representing battery materials.

[0163] Specifically, the pixel values ​​of the first contour information are obtained according to the second formula, which is:

[0164]

[0165] Where 'a' is the average pixel grayscale value of the battery material corresponding to the convolution kernel.

[0166] Based on the above scheme, by setting a sliding window, the first contour information is filtered to obtain accurate contour information of impurities.

[0167] The method for detecting battery materials according to embodiments of this application has been described in detail above. The following will combine... Figure 10 and Figure 11 The following embodiments of the present application describe a detection device for battery materials and a contour recognition model generation device. The technical features described in the method embodiments are applicable to the following device embodiments.

[0168] Figure 10 This is a schematic block diagram of the battery material detection device 600 provided in an embodiment of this application. Figure 10 As shown, the battery material detection device 600 includes an acquisition unit 610 and a processing unit 620.

[0169] The acquisition unit 610 is used to acquire images of battery materials.

[0170] The processing unit 620 is used to input the battery material image into the first contour recognition model to obtain intermediate contour information, the intermediate contour information including the first position information of the outer shape of the impurity contour in the battery material image.

[0171] The processing unit 620 is further configured to obtain first contour information based on the intermediate contour information, wherein the first contour information includes second position information of the impurity contour.

[0172] The processing unit 620 is also used to determine the qualification of the battery material based on the first contour information.

[0173] Figure 11 This is a schematic block diagram of the contour recognition model generation device 700 provided in an embodiment of this application. Figure 11 As shown, the contour recognition model generation device 700 includes an acquisition unit 710 and a processing unit 720.

[0174] The acquisition unit 710 is used to acquire first sample data, which includes multiple sets of battery material images.

[0175] The processing unit 720 is used to obtain an intermediate contour recognition model based on the first sample data; wherein multiple sets of battery material images are configured to be generated from the first battery material image without impurities.

[0176] Figure 12 A schematic block diagram of a battery material detection device 800 provided in an embodiment of this application is shown. Figure 12 As shown, the detection device 800 includes a processor 810 and a memory 820, wherein the memory 820 is used to store instructions, and the processor 810 is used to read the instructions and execute the detection methods of the various embodiments of the present application based on the instructions.

[0177] The memory 820 can be a separate device independent of the processor 810, or it can be integrated into the processor 810.

[0178] Optionally, such as Figure 12 As shown, the battery material detection device 800 may further include a transceiver 830, and the processor 810 can control the transceiver 830 to communicate with other devices. Specifically, it can send information or data to other devices, or receive information or data sent by other devices.

[0179] It should be understood that the processor in the embodiments of this application may be an integrated circuit chip with signal processing capabilities. In implementation, the steps of the above method embodiments can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor described above can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the detection method disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. The software modules can be located in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0180] It is understood that the memory in the embodiments of this application can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous linked dynamic random access memory (SLDRAM), and direct rambus RAM (DR RAM). It should be noted that the memory used in the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0181] Optionally, embodiments of this application also provide a battery, which includes the detection device or contour recognition model generation device provided in embodiments of this application.

[0182] Optionally, embodiments of this application also provide an electrical device, which includes a battery and a detection device or a contour recognition model generation device provided in embodiments of this application.

[0183] Optionally, embodiments of this application also provide an energy storage system, which includes a battery and a detection device or a contour recognition model generation device provided in embodiments of this application.

[0184] This application also provides a chip including a processor for calling and running computer programs from memory.

[0185] Optionally, a device equipped with the chip can execute the corresponding processes implemented by the battery material detection device or the contour recognition model generation device in the various methods of the embodiments of this application. For the sake of brevity, these will not be described in detail here.

[0186] This application also provides a computer-readable storage medium for storing computer programs.

[0187] Optionally, the computer-readable storage medium can be applied to the battery material detection device or the contour recognition model generation device in the embodiments of this application, and when the computer program is run on the computer, it causes the computer to execute the corresponding processes implemented by the battery material detection device or the contour recognition model generation device in the various methods of the embodiments of this application. For the sake of brevity, it will not be described in detail here.

[0188] This application also provides a computer program product, including computer program instructions.

[0189] Optionally, the computer program product can be applied to the battery material detection device or the contour recognition model generation device in the embodiments of this application. When the computer program instructions are run on the computer, the computer executes the corresponding processes implemented by the battery material detection device or the contour recognition model generation device in the various methods of the embodiments of this application. For the sake of brevity, they will not be described in detail here.

[0190] This application also provides a computer program.

[0191] Optionally, the computer program can be applied to the battery material detection device or the contour recognition model generation device in the embodiments of this application. When the computer program is run on a computer, it causes the computer to execute the corresponding processes implemented by the battery material detection device or the contour recognition model generation device in the various methods of the embodiments of this application. For the sake of brevity, it will not be described in detail here.

[0192] According to some embodiments of this application, see Figures 2 to 9 This application provides a method for detecting battery materials. The detection unit inputs the acquired battery material image into a first contour recognition model and a second contour recognition model to obtain the approximate contour and the true contour of the impurity in sequence. Based on the contour information, the area value and length value of the impurity are obtained to accurately determine whether the impurities in the battery material image are qualified.

[0193] During the training of the first contour recognition model, the initial sample data input to the initial network is an image of impurity-free battery material, with impurities randomly generated. The size and grayscale value of the impurities are set according to the impurities found in the actual battery material. Based on the first sample data, the parameters of the initial network are adjusted to train an intermediate contour recognition model.

[0194] After the image of the battery cell material is processed by the first contour recognition model to obtain the first contour information, the intermediate contour information may contain noise information. The variance of the pixel gray value corresponding to the first contour information is obtained based on the intermediate contour information, and the noise information in the intermediate contour information is determined and eliminated.

[0195] The intermediate contour information is input into the second contour recognition model. The sliding window based on the second contour recognition model slides on the first contour information, and the first contour information is obtained according to the sliding window.

[0196] Although this application has been described with reference to preferred embodiments, various modifications can be made thereto and components can be replaced with equivalents without departing from the scope of this application. In particular, the technical features mentioned in the various embodiments can be combined in any manner, provided there is no structural conflict. This application is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A method for detecting battery materials, characterized in that, include: Acquire images of battery materials; The battery material image is input into the first contour recognition model to obtain intermediate contour information, which includes the first position information of the outer shape of the impurity contour in the battery material image. Based on the intermediate contour information, first contour information is obtained, and the first contour information includes the second position information of the impurity contour; Based on the first contour information, the battery material is determined to be qualified.

2. The detection method according to claim 1, characterized in that, The step of determining that the battery material is qualified based on the second contour information includes: Based on the first contour information, the area value and length value of the impurity are obtained, wherein the length value is the maximum length of the second contour information; If the area value and / or the length value meet preset conditions, the battery material is determined to be qualified.

3. The detection method according to claim 2, characterized in that, Determining that the battery material is qualified when the area value and / or the length value meet preset conditions includes: If the area value is less than or equal to a first preset condition, the battery material is determined to be qualified; or, If the length value is less than or equal to a second preset condition, the battery material is determined to be qualified; or, If the area value is less than or equal to the first preset condition and the length value is less than or equal to the second preset condition, the battery material is determined to be qualified.

4. The detection method according to any one of claims 1 to 3, characterized in that, An image of the battery material is obtained by irradiating the battery material in a single battery cell with X-rays.

5. The detection method according to any one of claims 1 to 4, characterized in that, The first contour recognition model is trained in the following way: Acquire first sample data, which includes multiple sets of battery material images; Based on the first sample data, the first contour recognition model is obtained; The multiple sets of battery material images are configured to be generated from a first battery material image free of impurities.

6. The detection method according to claim 5, characterized in that, The multiple sets of battery material images are configured to generate battery materials from a first, impurity-free battery material image, including: According to the first function, the impurities are generated in the first battery material image to form the multiple sets of battery material images.

7. The detection method according to claim 5 or 6, characterized in that, The grayscale value of the impurity is determined by the image of the first battery material without impurities.

8. The detection method according to claim 7, characterized in that, The grayscale value of the impurity is determined by the image of the impurity-free first battery material, including: Within the range of the first battery material image corresponding to the impurity, the average grayscale value of the corresponding first battery material image is obtained; The gray value of the impurity is obtained based on the average gray value.

9. The detection method according to any one of claims 6 to 8, characterized in that, The first contour recognition model is also trained in the following way: Acquire second sample data, which includes multiple sets of battery material images obtained by irradiating the battery material with X-rays; The first contour recognition model is obtained based on the second sample data.

10. The detection method according to any one of claims 1 to 9, characterized in that, The step of obtaining the first contour information based on the intermediate contour information includes: The intermediate contour information is input into the second contour recognition model to obtain the first contour information.

11. The detection method according to claim 10, characterized in that, The step of inputting the intermediate contour information into the second contour recognition model to obtain the first contour information includes: Obtain a sliding window of the second contour recognition model, wherein the sliding window slides over the intermediate contour information; The first contour information is obtained based on the sliding window.

12. The detection method according to claim 11, characterized in that, Obtaining the first contour information based on the sliding window includes: Based on the sliding window, the first threshold corresponding to the sliding window is obtained; The first contour information is obtained when the gray value of each pixel in the intermediate contour information is less than the first threshold.

13. The detection method according to claim 12, characterized in that, The step of obtaining the first threshold corresponding to the sliding window based on the sliding window includes: The sliding window is used to traverse each pixel of the intermediate contour information to obtain the standard deviation of the intermediate contour information in the sliding window corresponding to each pixel. The first threshold is obtained based on the standard deviation.

14. The detection method according to any one of claims 1 to 13, characterized in that, Before inputting the intermediate contour information into the second contour recognition model, the method further includes: Based on the intermediate contour information, the variance of the pixel grayscale value corresponding to the intermediate contour information is obtained; Based on the variance of the pixel grayscale values, noise information in the intermediate contour information is determined and excluded.

15. The detection method according to any one of claims 1 to 14, characterized in that, The impurities include at least one of the following: iron, copper, zinc, copper-zinc alloys, and stainless steel.

16. A device for detecting battery materials, characterized in that, Includes an acquisition unit and a processing unit. The acquisition unit is used to acquire images of battery materials; The processing unit is used to input the battery material image into a first contour recognition model to obtain intermediate contour information, wherein the intermediate contour information includes the first position information of the outer shape of the impurity contour in the battery material image; The processing unit is further configured to obtain first contour information based on the intermediate contour information, wherein the first contour information includes the second position information of the impurity contour; The processing unit is further configured to determine whether the battery material is qualified based on the first contour information.

17. A chip, characterized in that, include: A processor for retrieving and running a computer program from memory, causing a device on which the chip is mounted to perform the detection method as described in any one of claims 1 to 15.

18. A computer program product, characterized in that, include: Computer program instructions, when executed by a computer, cause the computer to implement the detection method as described in any one of claims 1 to 15.

Images