Apparatus and method for inspecting battery electrode

Abstract

An apparatus for inspecting a battery electrode according to an embodiment of the present invention may: extract an inspection object image including a region suspected of having a defect based on an electrode image acquired from a camera; and different types of learning models for determining whether or not an electrode corresponding to the inspection target image has a defect are set and applied according to the amount of the training data, thereby shortening the application time due to the initial process application and the change in the process environment. And obtaining result data with improved precision and high reliability by retraining the learning model based on the updated training data.

Description

Technical Field

[0001] This application claims priority to and the benefit of Korean Patent Application No. 10-2022-0109765 filed in the Korean Intellectual Property Office on August 31, 2022, and Korean Patent Application No. 10-2023-0103916 filed in the Korean Intellectual Property Office on August 9, 2023, the entire contents of which are incorporated herein by reference.

[0002] The present invention relates to an apparatus and method for inspecting an electrode of a battery, and more particularly, to an apparatus and method for inspecting an electrode of a battery using a pre-trained learning model to determine whether the electrode is defective due to a surface defect based on an electrode image. Background Art

[0003] As the depletion of fossil fuels leads to rising energy prices and people's concerns about environmental pollution increase, the demand for secondary batteries as environmentally friendly alternative energy sources is rapidly increasing.

[0004] As a response to today's environmental regulations and high oil prices, among secondary batteries, lithium batteries are used in various industrial fields ranging from mobile application devices to automobiles, robots, and energy storage devices.

[0005] These lithium secondary batteries are generally classified into a cylindrical type, a prismatic type, or a pouch type according to the shape of an exterior material accommodating an electrode assembly.

[0006] Among them, the cylindrical battery can adopt a cell-to-pack (CTP) structure composed of a plurality of battery cells. In other words, the cylindrical battery is provided in a structure assembled by winding an electrode including a separator between a negative electrode and a positive electrode and inserting it into a battery can.

[0007] In the electrode preparation process of cylindrical batteries, battery electrode inspection equipment is used to detect defects on the electrode surface caused by foreign matter, scratches, etc.

[0008] Generally, the battery electrode inspection equipment analyzes the electrode image obtained from the camera installed in the process equipment to determine whether there are any defects in the electrode surface. Therefore, in the battery electrode inspection equipment, the result data may be affected by the on-site situation or environment (such as equipment vibration, process condition changes, etc.).

[0009] Therefore, conventional battery electrode inspection equipment applies program logic designed based on defect images obtained from cameras installed on site during initial application of the process.

[0010] Therefore, the conventional battery electrode inspection equipment has a disadvantage of requiring a long time when the process is first applied.

[0011] Furthermore, conventional battery electrode inspection equipment also has the disadvantage of requiring engineers to constantly redesign program logic when any changes occur in field conditions or environments. Summary of the invention

[0012] Technical issues

[0013] To solve one or more problems in the related art, an embodiment of the present disclosure provides a system for inspecting an electrode of a battery.

[0014] To solve one or more problems in the related art, embodiments of the present disclosure also provide an apparatus for inspecting an electrode of a battery.

[0015] To solve one or more problems in the related art, embodiments of the present disclosure also provide a method for inspecting an electrode of a battery.

[0016] Technical Solution

[0017] To achieve the purpose of the present disclosure, a device for inspecting battery electrodes based on an artificial intelligence model may include: at least one processor and a memory, wherein the memory is configured to store at least one instruction executed by the at least one processor, and the at least one instruction may include: an instruction to obtain an inspection object image from which a suspected defective area is extracted from an electrode image taken of at least one electrode surface; and an instruction to output result data related to whether the battery electrode is defective by inputting the inspection object image into a pre-trained learning model.

[0018] Here, the instruction to output the result data may include an instruction to determine any one of a plurality of learning models according to an amount of training data.

[0019] The instruction to determine any one learning model may include an instruction to output the result data using a pre-trained first learning model based on machine learning when the amount of the training data is less than a predetermined reference value.

[0020] The instruction to determine any one learning model may include an instruction to output the result data using a pre-trained second learning model based on deep learning when the amount of the training data is greater than or equal to the predetermined reference value.

[0021] The instruction to output the result data using the first learning model may include: an instruction to extract at least one image feature value from the inspection object image; and an instruction to obtain the result data by inputting the at least one image feature value into the first learning model.

[0022] The instruction for extracting at least one image feature value from the inspection object image may include an instruction for extracting the at least one image feature value from the inspection object image using a rule-based algorithm.

[0023] The at least one image feature value may include data related to one or more of pixel height, pixel width, pixel maximum value (Peak White), pixel minimum value (Peak Dark), aspect ratio, and roundness extracted from the inspection object image.

[0024] The first learning model may be a random forest-based learning model.

[0025] The instruction to output the result data using the pre-trained deep learning-based second learning model may include an instruction to obtain the result data by inputting at least one inspection object image into the second learning model.

[0026] The second learning model may be a learning model based on a convolutional neural network (CNN).

[0027] The at least one instruction may also include an instruction to retrain the pre-trained learning model by using the result data as learning data.

[0028] According to another embodiment of the present disclosure, a method for inspecting battery electrodes using an apparatus for inspecting battery electrodes based on an artificial intelligence model may include the following steps: obtaining an inspection object image from which areas suspected of being defective are extracted from an electrode image taken of at least one electrode surface; and outputting result data related to whether the battery electrode is defective by inputting the inspection object image into a pre-trained learning model.

[0029] Here, outputting the result data related to whether the battery electrode is defective includes: determining any one of a plurality of learning models according to an amount of at least one training data.

[0030] Determining any one of the multiple learning models may include: when the amount of the training data is less than a predetermined reference value, using a pre-trained first learning model based on machine learning to output the result data.

[0031] Determining any one of the multiple learning models may include: when the amount of the training data is greater than or equal to the predetermined reference value, using a pre-trained second deep learning-based learning model to output the result data.

[0032] Outputting the result data using the first learning model may include: extracting at least one image feature value from the inspection object image; and obtaining the result data by inputting the at least one image feature value into the first learning model.

[0033] Extracting the at least one image feature value from the inspection object image may include extracting the at least one image feature value from the inspection object image using a rule-based algorithm.

[0034] The at least one image characteristic value may include data related to one or more of pixel height, pixel width, pixel maximum value (white peak), pixel minimum value (dark peak), aspect ratio, and circularity extracted from the inspection object image.

[0035] The first learning model may be a random forest-based learning model.

[0036] Outputting the result data using the pre-trained deep learning-based second learning model may include obtaining the result data by inputting at least one inspection object image into the second learning model.

[0037] The second learning model may be a learning model based on a convolutional neural network (CNN).

[0038] The method may further include retraining the pre-trained learning model by using the result data as learning data.

[0039] According to another embodiment of the present disclosure, a system for inspecting battery electrodes based on an artificial intelligence model may include: a camera configured to generate an electrode image by photographing the surface of at least one electrode; and a battery electrode inspection device configured to obtain an inspection object image from the electrode image with areas suspected of being defective extracted, input the inspection object image into a pre-trained learning model, and output result data related to whether the battery electrode is defective.

[0040] Beneficial Effects

[0041] According to the battery electrode inspection device and method of the embodiments of the present invention, and the system including the battery electrode inspection device, it is possible to obtain improved inspection reliability by applying different types of learning models according to the amount of learning data to check whether the inspection object image has defects, and retraining the learning model based on the learning data updated using the result data. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1is a conceptual diagram of the process of applying detection logic in a general battery electrode inspection device.

[0043] Figure 2 is a conceptual diagram of the operation of a universal battery electrode inspection device when an error occurs.

[0044] Figure 3 is a block diagram of a battery electrode inspection system according to an embodiment of the present invention.

[0045] Figure 4 is a block diagram of a battery electrode inspection apparatus according to an embodiment of the present invention.

[0046] Figure 5 is an operation flowchart of a battery electrode inspection method using the battery electrode inspection apparatus according to an embodiment of the present invention.

[0047] Figure 6 The images are used to explain the step of extracting an inspection object image in the battery electrode inspection method according to the embodiment of the present invention.

[0048] Figure 7 1 is an operational flowchart for explaining a step of determining whether an electrode fails in a battery electrode inspection method according to an embodiment of the present invention.

[0049] Figure 8 1 is a conceptual diagram for explaining a battery electrode inspection method using a first learning model of the battery electrode inspection method according to an embodiment of the present invention.

[0050] Fig. 9 is a conceptual diagram illustrating a battery electrode inspection method using a second learning model of the battery electrode inspection method according to another embodiment of the present invention.

[0051] Fig.10 This is a conceptual diagram for explaining a relearning step in a battery electrode inspection method according to another embodiment of the present invention.

[0052] 1000: Camera 5000: Battery electrode inspection equipment

[0053] 100: Memory 200: Processor

[0054] 300: transceiver device 400: input interface device

[0055] 500: output interface device 600: storage device

[0056] 700: Bus DETAILED DESCRIPTION

[0057] The present invention can be modified in various forms and have various embodiments, and its specific embodiments are shown by way of example in the accompanying drawings and will be described in detail below. However, it should be understood that it is not intended to limit the present invention to specific embodiments, but on the contrary, the present invention will cover all modifications, equivalents and substitutions that fall within the spirit and technical scope of the present invention. In the description of the entire drawings, the same reference numerals represent the same elements.

[0058] It should be understood that although various elements may be described using terms such as first, second, A, B, etc., these elements should not be limited by these terms. These terms are only used to distinguish one element from another element. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element without departing from the scope of the present invention. As used herein, the term "and / or" includes a combination of any of a plurality of associated listings or a plurality of associated listings.

[0059] It should be understood that when an element is referred to as being "coupled" or "connected" to another element, it can be directly coupled or connected to the other element, or intervening elements may exist. In contrast, when an element is referred to as being "directly coupled" or "directly connected" to another element, there are no intervening elements.

[0060] The terms used herein are only used for the purpose of describing specific embodiments and are not intended to limit the present invention. As used herein, the singular forms "one", "an" and "the" are also intended to include plural forms unless the context clearly indicates otherwise. It should also be understood that the terms "include", "comprise", "include", "include" and / or "have" are used herein to clearly indicate the presence of the features, integers, steps, operations, constituent elements, parts and / or combinations thereof, but do not exclude the presence or addition of one or more other features, integers, steps, operations, constituent elements, parts and / or combinations thereof.

[0061] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those skilled in the art to which the present invention belongs. It will be further understood that terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant art, and will not be interpreted in an idealized or overly formal sense unless explicitly so defined herein.

[0062] Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

[0063] Figure 1 is a conceptual diagram of the process of applying detection logic in a general battery electrode inspection device.

[0064] Reference Figure 1 , the battery electrode inspection equipment can detect defects on the surface of the electrode and determine whether the electrode has failed.

[0065] The universal battery electrode inspection device uses a pre-designed detection logic to determine the defect type and fault of at least one electrode image obtained from a camera installed at the site. Here, the detection logic can be a logic previously designed by an engineer using image features of at least one electrode image of which the defect type has been confirmed.

[0066] Figure 2 is a conceptual diagram of the operation of a universal battery electrode inspection device when an error occurs.

[0067] Reference Figure 2 , the detection logic of the universal battery electrode inspection device is designed based on the defect image of the electrode obtained from the camera installed on site during the initial application of the process. Therefore, the universal battery electrode inspection device has the following disadvantages: when at least one of the on-site environmental factors or process conditions changes, detection errors may occur due to distortion or deformation of the electrode image, which can reduce the accuracy of defect detection. Therefore, the universal electrode inspection device has the following disadvantages: the detection logic should be constantly modified by engineers.

[0068] Furthermore, when modifying the inspection logic to prevent this problem, it is necessary to modify the inspection logic by acquiring defect images generated from cameras installed on site to take the above-mentioned process factors into consideration, which has the disadvantage of taking a long time.

[0069] In addition, the general battery electrode inspection device has the following disadvantages: when the detection logic is modified, it is difficult to determine the classification accuracy of at least one defective image that has been inspected before the modification. Here, the defective image can be an image of an electrode that is determined to be faulty due to the inspection result of the inspection logic being a defect.

[0070] Therefore, the present invention will describe an electrode inspection apparatus and method that extracts an inspection object image suspected of having a surface defect of the electrode from an electrode image using a pre-trained learning model, and determines whether the electrode fails based on this.

[0071] Figure 3 is a block diagram of a battery electrode inspection system according to an embodiment of the present invention.

[0072] Reference Figure 3 , a battery electrode inspection system can detect electrodes with surface defects during the electrode preparation process and determine whether the electrode has failed.

[0073] In more detail, the battery electrode inspection system may include a camera 1000 and a battery electrode inspection device 5000 .

[0074] The camera 1000 may be installed at a location where the electrode preparation process is performed. For example, the camera 1000 may be fixedly installed in a device that sequentially moves at least one electrode. Therefore, the camera 1000 may acquire at least one electrode image by photographing at least one electrode individually.

[0075] The battery electrode inspection apparatus 5000 may be connected with at least one camera 1000 to obtain at least one electrode image from the camera 1000. However, the scope of the present invention is not limited thereto, but the battery electrode inspection apparatus 5000 may obtain at least one electrode image by various means other than a camera.

[0076] Thereafter, the battery electrode inspection apparatus 5000 may determine whether a battery electrode fails based on at least one electrode image using a pre-trained learning model.

[0077] The battery electrode inspection apparatus 5000 will be described in more detail below.

[0078] Figure 4 is a block diagram of a battery electrode inspection apparatus according to an embodiment of the present invention.

[0079] Reference Figure 4 As described above, the battery electrode inspection device 5000 according to the embodiment of the present invention can extract at least one image of the inspection object area suspected of being defective from at least one electrode image, and determine whether the electrode is defective based on the extracted inspection image using a pre-trained learning model. Here, the defect may include foreign matter, scratches, cracks, etc.

[0080] Describing the battery electrode inspection device in detail according to the configuration, the battery electrode inspection device 5000 may include a memory 100 , a processor 200 , a transceiver device 300 , an input interface device 400 , an output interface device 500 , and a storage device 600 .

[0081] According to the embodiment, the respective components 100 , 200 , 300 , 400 , 500 , and 600 included in the battery electrode inspection apparatus are connected through a bus 700 and can communicate with each other.

[0082] In components 100, 200, 300, 400, 500, and 600, memory 100 and storage device 600 may include at least one of a volatile storage medium and a nonvolatile storage medium. For example, memory 100 and storage device 600 may include at least one of a read-only memory (ROM) and a random access memory (RAM).

[0083] The memory 100 may include at least one instruction executed by the processor 200 .

[0084] According to an embodiment, at least one instruction may include: an instruction to obtain an inspection object image from which a suspected defective area is extracted from an electrode image taken of at least one electrode surface; and an instruction to output result data on whether the battery electrode is defective by inputting the inspection object image into a pre-trained learning model.

[0085] Here, the instruction to output the result data may include an instruction to determine any one of a plurality of learning models according to an amount of at least one training data.

[0086] The instruction to determine any one learning model may include an instruction to output result data using a pre-trained first learning model based on machine learning when the amount of training data is less than a predetermined reference value.

[0087] The instruction to determine any one learning model may include an instruction to output result data using a pre-trained deep learning-based second learning model when the amount of training data is greater than or equal to a predetermined reference value.

[0088] The instruction to output the result data using the first learning model may include an instruction to extract at least one image feature value from the inspection object image, and an instruction to obtain the result data by inputting the at least one image feature value into the first learning model.

[0089] The instruction to extract at least one image feature value from the inspection object image may include an instruction to extract at least one image feature value from the inspection object image using a rule-based algorithm.

[0090] The at least one image characteristic value may include data related to one or more of pixel height, pixel width, pixel maximum value (white peak), pixel minimum value (dark peak), aspect ratio, and circularity extracted from the inspection object image.

[0091] The first learning model may be a random forest based learning model.

[0092] The instruction to output result data using the pre-trained deep learning-based second learning model may include an instruction to obtain result data by inputting at least one inspection object image into the second learning model.

[0093] The second learning model may be a convolutional neural network (CNN) based learning model.

[0094] The at least one instruction may also include an instruction to retrain the pre-trained learning model by using the result data as learning data.

[0095] In addition, the processor 200 may refer to a central processing unit (CPU), a graphics processing unit (GPU), or a dedicated processor that executes the method according to the embodiment of the present invention.

[0096] As described above, the processor 200 may execute at least one program command stored in the memory 100 .

[0097] Figure 5 is an operation flowchart of a battery electrode inspection method using the battery electrode inspection apparatus according to an embodiment of the present invention.

[0098] Reference Figure 5 , the battery electrode inspection apparatus 5000 may obtain at least one electrode image ( S1000 ).

[0099] Thereafter, the battery electrode inspection apparatus 5000 may extract at least one inspection object image from the at least one electrode image (S3000). Here, the electrode image may be an image of an electrode surface photographed by at least one camera 1000 installed on the electrode preparation process apparatus.

[0100] In addition, the inspection object image may be an image in which at least one region where a defect occurs in at least one electrode image is extracted separately. Here, the defect may include, for example, scratches, foreign matter, deterioration, etc., but is not limited to these.

[0101] Thereafter, the battery electrode inspection apparatus 5000 may determine whether the electrode fails based on the inspection object image using the pre-trained learning model ( S5000 ).

[0102] Thereafter, the battery electrode inspection apparatus 5000 may retrain the learning model based on at least one updated learning data ( S7000 ).

[0103] Figure 6 The images are used to explain the step of extracting an inspection object image in the battery electrode inspection method according to the embodiment of the present invention.

[0104] Reference Figure 6 , the inspection object image may be obtained using a rule-based algorithm. In other words, the battery electrode inspection apparatus 5000 may output the inspection object image by inputting the electrode image and the predefined conditional command into the rule-based algorithm.

[0105] Here, the rule-based algorithm can calculate the average brightness value of the electrode image according to a predefined conditional command, select at least one pixel whose difference exceeds a certain range based on the calculated average brightness, extract a certain area including the corresponding pixel, thereby obtaining and outputting the inspection object image. For example, the inspection object image can be a binary image and can be provided in a size of 128 pixels×128 pixels.

[0106] Figure 7is an operational flowchart for explaining the steps of determining whether an electrode fails in a battery electrode inspection method according to an embodiment of the present invention.

[0107] Reference Figure 7 , the battery electrode inspection device 5000 according to the embodiment of the present invention can check the amount of learning data classified according to the data set in the memory 100 (S5100). Here, at least one learning data may include at least one inspection result data pre-checked by the battery electrode inspection device 5000. In other words, the learning data may include fault data or normal data. For example, the fault data may be classified and stored according to the defect type in the data set.

[0108] Therefore, the battery electrode inspection apparatus 5000 may select the type of pre-trained learning model.

[0109] According to an embodiment, when the amount of learning data is less than a predetermined reference value, the battery electrode inspection apparatus 5000 may select a pre-trained first learning model ( S5300 ).

[0110] Thereafter, the battery electrode inspection apparatus 5000 may extract image feature values ​​from the inspection object image ( S5310 ).

[0111] Thereafter, the battery electrode inspection apparatus 5000 may determine whether the battery electrode fails using the pre-trained first learning model ( S5350 ).

[0112] In addition, according to another embodiment, when the amount of learning data is greater than or equal to a predetermined reference value, the battery electrode inspection apparatus 5000 may select a pre-trained second learning model ( S5500 ).

[0113] Thereafter, the battery electrode inspection apparatus 5000 may determine whether the battery electrode fails using the selected second learning model ( S5700 ).

[0114] Figure 8 1 is a conceptual diagram for explaining a battery electrode inspection method using a first learning model of the battery electrode inspection method according to an embodiment of the present invention.

[0115] Reference Figure 8 , as described above, when the amount of learning data is less than a predetermined reference value, the battery electrode inspection apparatus 5000 may select a pre-trained first learning model.

[0116] Thereafter, the battery electrode inspection apparatus 5000 may extract image feature values ​​from the inspection object image.

[0117] Here, the image feature value may be at least one pixel information extracted from the detection object image. For example, the image feature value may include one or more of pixel height, pixel width, pixel maximum value (white peak), pixel minimum value (dark peak), aspect ratio, and roundness.

[0118] According to an embodiment, when the amount of learning data is less than a predetermined reference value, the battery electrode inspection apparatus 5000 may obtain an image feature value from at least one inspection object image using a rule-based algorithm.

[0119] Thereafter, the battery electrode inspection apparatus 5000 may determine whether the battery electrode fails using the pre-trained first learning model.

[0120] More specifically, the battery electrode inspection device 5000 may receive image feature values ​​extracted from the image to be inspected as input data of the pre-trained first learning model. Thereafter, the battery electrode inspection device 5000 may output result data related to whether the battery electrode has failed based on the inspection object image. Here, the battery electrode may be an electrode corresponding to the inspection object image.

[0121] According to an embodiment, the first pre-trained learning model may be an algorithm based on machine learning. For example, the first learning model may be a bagging-type random forest pre-trained based on at least one learning data pre-classified in the data set of the memory 100 .

[0122] Fig. 9 is a conceptual diagram illustrating a battery electrode inspection method using a second learning model of the battery electrode inspection method according to another embodiment of the present invention.

[0123] Reference Fig. 9 , as described above, when the amount of learning data is equal to or greater than a predetermined reference value, the battery electrode inspection apparatus 5000 may determine whether the battery electrode is defective using the pre-trained second learning model.

[0124] In more detail, the battery electrode inspection device 5000 may receive the inspection object image as input data of the pre-trained second learning model. Therefore, the battery electrode inspection device 5000 may output data as a result of determining whether the battery electrode fails based on the inspection object image.

[0125] According to an embodiment, the pre-trained second learning model may be a pre-trained learning model based on at least one learning data pre-classified in the data set in the memory 100 .

[0126] More specifically, the pre-trained second learning model may be a classification model based on artificial intelligence (AI). In other words, the second learning model may be a classification model based on deep learning, which is pre-trained based on the result data of the battery electrode inspection device 5000 according to pre-stored learning data.

[0127] According to an embodiment, the pre-trained second learning model may be a learning model of an artificial neural network based on a convolutional neural network (CNN). For example, the battery electrode inspection device 5000 may pre-train the second learning model so that the second learning model extracts image feature data of the inspection object image using a convolutional layer, reduces the feature data to a lower dimension through a pooling layer, classifies it according to the defect type through a fully connected layer, and determines whether the battery electrode fails.

[0128] Fig.10 This is a conceptual diagram for explaining a relearning step in a battery electrode inspection method according to another embodiment of the present invention.

[0129] Reference Fig.10 , the battery electrode inspection apparatus 5000 may store result data obtained from the embodiment of the present invention (ie, result data for determining whether an electrode is defective) in the memory 100 and classify it into data sets.

[0130] Thereafter, the battery electrode inspection device 5000 may retrain the learning model based on at least one updated learning data. In other words, the battery electrode inspection device 5000 may retrain the learning model by using the result data of the previous learning as the learning data. Therefore, the amount of learning data increases, thereby improving the inspection reliability of the battery electrode inspection device 5000.

[0131] The battery electrode inspection apparatus and method according to the embodiments of the present invention have been described above.

[0132] The battery electrode inspection device and method according to an embodiment of the present invention can extract an inspection object image including an area suspected of being defective based on an electrode image obtained from a camera, determine and apply different types of learning models to determine whether the electrode corresponding to the inspection object image is defective according to the amount of learning data, thereby shortening the application time caused by the initial process application and the change of the process environment.

[0133] Furthermore, the battery electrode inspection apparatus and method can obtain highly reliable result data with improved accuracy by retraining the learning model based on updated learning data.

[0134] The operation of the method according to an embodiment of the present invention can be implemented as a code or a computer-readable program on a computer-readable recording medium. The computer-readable recording medium includes all types of recording devices for storing data readable by a computer system. In addition, the computer-readable recording medium can be distributed in a networked computer system, so as to store and execute a computer-readable program or code in a distributed manner.

[0135] In addition, the computer-readable recording medium may include hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, etc. The program instructions may include not only machine language codes created by a compiler, but also high-level language codes that can be executed by a computer using an interpreter.

[0136] Although some aspects of the present invention have been described in the context of an apparatus, it can also be represented by a description according to a corresponding method, wherein a block or device corresponds to a method step or a feature of a method step. Similarly, the aspects described in the context of a method can also represent the features of a corresponding block or project or a corresponding device. Some or all of the method steps can be performed by (or using) a hardware device (e.g., a microprocessor, a programmable computer, or an electronic circuit). In some embodiments, one or more of the most important method steps can be performed by such a device.

[0137] The present invention has been described above with reference to exemplary embodiments thereof, but those skilled in the art will appreciate that various corrections and changes may be made to the present invention within the scope without departing from the spirit and scope of the invention described in the appended claims.

Claims

1. A battery electrode inspection device based on an artificial intelligence model, the battery electrode inspection device comprising: at least one processor; as well as a memory configured to store at least one instruction executed by the at least one processor, The at least one instruction includes: an instruction to obtain an inspection object image in which a region suspected of having a defect is extracted from an electrode image in which a surface of at least one electrode is photographed; and An instruction to output result data regarding whether the battery electrode is defective by inputting the inspection object image into a pre-trained learning model.

2. The battery electrode inspection device according to claim 1, wherein: The instruction to output the result data includes an instruction to determine any one of a plurality of learning models according to an amount of at least one training data.

3. The battery electrode inspection device according to claim 2, wherein: The instructions for determining any learning model include: When the amount of the training data is less than a predetermined reference value, an instruction to output the result data using a pre-trained first learning model based on machine learning.

4. The battery electrode inspection device according to claim 3, wherein: The instructions for determining any learning model include: When the amount of the training data is greater than or equal to the predetermined reference value, an instruction to output the result data using a pre-trained second learning model based on deep learning.

5. The battery electrode inspection device according to claim 3, wherein: The instruction to output the result data using the first learning model includes: An instruction to extract at least one image feature value from the inspection object image; and Instructions for obtaining the result data by inputting the at least one image feature value into the first learning model.

6. The battery electrode inspection device according to claim 5, wherein: The instruction for extracting at least one image feature value from the inspection object image includes an instruction for extracting the at least one image feature value from the inspection object image using a rule-based algorithm.

7. The battery electrode inspection device according to claim 5, wherein: The at least one image feature value includes data related to one or more of pixel height, pixel width, pixel maximum value (white peak), pixel minimum value (dark peak), aspect ratio, and circularity extracted from the inspection object image.

8. The battery electrode inspection device according to claim 3, wherein: The first learning model is a learning model based on random forest.

9. The battery electrode inspection device according to claim 4, wherein: The instruction to output the result data using a pre-trained deep learning-based second learning model includes an instruction to obtain the result data by inputting at least one inspection object image into the second learning model.

10. The battery electrode inspection device according to claim 4, wherein: The second learning model is a learning model based on a convolutional neural network (CNN).

11. The battery electrode inspection device according to claim 1, wherein: The at least one instruction also includes instructions to retrain the pre-trained learning model by using the result data as learning data.

12. A battery electrode inspection method using a battery electrode inspection device based on an artificial intelligence model, the battery electrode inspection method comprising the following steps: Obtaining an inspection object image from which a region suspected of having a defect is extracted from an electrode image obtained by photographing a surface of at least one electrode; as well as Result data on whether the battery electrode is defective is output by inputting the inspection object image into a pre-trained learning model.

13. The battery electrode inspection method according to claim 12, wherein: The step of outputting the result data regarding whether the battery electrode is defective includes the step of determining any one of a plurality of learning models according to an amount of at least one training data.

14. The battery electrode inspection method according to claim 13, wherein: The step of determining any one of the plurality of learning models comprises the following steps: When the amount of the training data is less than a predetermined reference value, the result data is output using a pre-trained first learning model based on machine learning.

15. The battery electrode inspection method according to claim 14, wherein: The step of determining any one of the plurality of learning models comprises the following steps: When the amount of the training data is greater than or equal to the predetermined reference value, the result data is output using a pre-trained second learning model based on deep learning.

16. The battery electrode inspection method according to claim 14, wherein: The step of outputting the result data using the first learning model comprises the following steps: extracting at least one image feature value from the inspection object image; and The result data is obtained by inputting the at least one image feature value into the first learning model.

17. The battery electrode inspection method according to claim 16, wherein: The step of extracting at least one image feature value from the inspection object image includes extracting the at least one image feature value from the inspection object image using a rule-based algorithm.

18. The battery electrode inspection method according to claim 16, wherein: The at least one image feature value includes data related to one or more of pixel height, pixel width, pixel maximum value (white peak), pixel minimum value (dark peak), aspect ratio, and circularity extracted from the inspection object image.

19. The battery electrode inspection method according to claim 14, wherein: The first learning model is a learning model based on random forest.

20. The battery electrode inspection method according to claim 15, wherein: The step of outputting the result data using a pre-trained deep learning-based second learning model includes obtaining the result data by inputting at least one inspection object image into the second learning model.

21. The battery electrode inspection method according to claim 15, wherein: The second learning model is a learning model based on a convolutional neural network (CNN).

22. The battery electrode inspection method according to claim 12, further comprising the following steps: The pre-trained learning model is retrained by using the result data as learning data.

23. A battery electrode inspection system based on an artificial intelligence model, the battery electrode inspection system comprising: a camera configured to generate an electrode image by photographing a surface of at least one electrode; as well as A battery electrode inspection device is configured to obtain an inspection object image from the electrode image in which a suspected defective area is extracted, input the inspection object image into a pre-trained learning model, and output result data related to whether the battery electrode is defective.

Images