Learning model-based dimensional measuring apparatus and method

Abstract

The dimension measurement device and method according to an embodiment of the present invention can obtain dimension information of a specific object in an image of interest using an already trained learning model, thereby obtaining highly reliable dimension measurement result data with improved accuracy.

Description

TECHNICAL FIELD

[0001] This application claims the benefit of the filing dates of Korean Patent Application No. 10-2022-0109769 filed with the Korean Intellectual Property Office on August 31, 2022 and Korean Patent Application No. 10-2023-0103918 filed with the Korean Intellectual Property Office on August 9, 2023, and the entire contents disclosed in the documents of said Korean patent applications are incorporated herein by reference.

[0002] The present invention relates to a dimension measuring apparatus and method, and more particularly, to a dimension measuring apparatus and method that obtain dimension information of a specific object in an image of interest using a pre-trained learning model. BACKGROUND ART

[0003] Due to the depletion of fossil fuels, the price of energy sources has risen, and interest in environmental pollution has increased, so the demand for secondary batteries as an eco-friendly alternative energy source is rapidly increasing.

[0004] Among secondary batteries, lithium batteries have been applied in many industrial fields from mobile application devices to automobiles, robots, energy storage devices and the like as a countermeasure against recent environmental regulations and high crude oil prices.

[0005] Such lithium batteries are generally classified into cylindrical, prismatic or pouch types according to the shape of an exterior member in which an electrode assembly is accommodated.

[0006] Among these, the cylindrical battery is provided in the structure of a battery pack (Cell To Pack, CTP) composed of a plurality of battery cells. In other words, the cylindrical battery is provided with a structure assembled by winding an electrode composed of a separator between a positive electrode and a negative electrode and inserting the wound electrode into a battery can.

[0007] Such cylindrical batteries undergo a quality inspection after the assembly process.

[0008] Generally, in quality inspection of cylindrical batteries, the dimensions of the relevant area are measured using an image of the object being measured, in order to determine the location of the weld, the tensile strength of the weld, or whether there are any defects in the battery structure that occur during the manufacturing process.

[0009] Conventional dimensional measuring devices detect the boundary (edge) of the object to be measured, corresponding to a pre-set threshold brightness value, from a target image which is a grayscale image (gray model), and extract the pixel information of the detected boundary to obtain the dimensional information of the object to be measured.

[0010] However, conventional dimensional measuring devices had a drawback: if an error occurred in the programmed logic designed by the engineer, the boundaries of the object being measured on the target image would be misdetected, reducing the reliability of the dimensional information. [Overview of the project] [Problems that the invention aims to solve]

[0011] The objective of the present invention, which aims to solve the above-mentioned problems, is to provide a dimensional measuring device.

[0012] Another objective of the present invention, in order to solve the problems described above, is to provide a method for measuring dimensions. [Means for solving the problem]

[0013] A learning model-based dimensional measuring device according to one embodiment of the present invention for achieving the above objective includes a memory and a processor that executes at least one instruction stored in the memory, wherein the at least one instruction includes an instruction to acquire an image of interest from a target image that includes the object to be measured, and an instruction to input the image of interest into a learning model that has already been trained, and output the dimensional information of the object to be measured as result data.

[0014] Here, the pre-trained model mentioned above may be a deep learning-based model.

[0015] The command to output the above dimensional information as result data may include a command to segment the image of interest into physical objects using the already learned model, and to label the dimensional measurement points of the measured objects.

[0016] On the other hand, the command to acquire the above-mentioned image of interest can be pre-trained using a rule-based algorithm to acquire the image of interest containing the measurement target from the above-mentioned target image.

[0017] Furthermore, the command to acquire the above-mentioned image of interest can be pre-trained to perform image preprocessing on the target image and then acquire the image of interest.

[0018] A learning model-based dimensional measurement method according to another embodiment of the present invention for achieving the above objective includes the steps of: acquiring an image of interest from a target image that includes the object to be measured; and inputting the image of interest into a learning model that has already been trained, and outputting the dimensional information of the object to be measured as result data.

[0019] Here, the already trained model mentioned above may be a deep learning-based model.

[0020] The step of outputting the above dimensional information as result data may include the step of segmenting the image of interest into physical objects using the already learned model and labeling the dimensional measurement points of the measured objects.

[0021] On the other hand, the step of acquiring the image of interest may include the step of acquiring the image of interest containing the object to be measured from the target image using a rule-based algorithm.

[0022] Further, the step of acquiring the region-of-interest image may include the step of acquiring the region-of-interest image by performing image preprocessing on the target image. Effects of the Invention

[0023] The dimension measuring apparatus and method according to embodiments of the present invention can obtain highly reliable dimension measurement result data with improved accuracy by acquiring dimension information of a specific target in a region-of-interest image using an already trained learning model. Brief Description of the Drawings

[0024] [Figure 1] It is a misdetection image of a welded portion measured by a general dimension measuring apparatus. [Figure 2] It is a block diagram of a dimension measuring apparatus according to an embodiment of the present invention. [Figure 3] It is a flow diagram of a dimension measuring method using the dimension measuring apparatus according to an embodiment of the present invention. [Figure 4] It is a diagram for explaining a region-of-interest image of the dimension measuring apparatus according to an embodiment of the present invention. [Figure 5] It shows a first current collector plate image of a dimension measuring apparatus according to an embodiment of the present invention, and a region-of-interest image of a welded portion extracted from the first current collector plate image. [Figure 6] It is an image preprocessed to extract the region-of-interest image of Fig. 5 according to an embodiment of the present invention. [Figure 7] It shows a second current collector plate image according to another embodiment of the present invention, and a region-of-interest image extracted from the second current collector plate image. [Figure 8] It is an image preprocessed to extract the region-of-interest image of Fig. 7 according to another embodiment of the present invention. [Figure 9] It is a diagram for explaining a dimension measuring method for an electrode part using an already trained learning model in a dimension measuring apparatus according to still another embodiment of the present invention. Mode for Carrying Out the Invention

[0025] The present invention can be modified in various ways and has many embodiments; therefore, specific embodiments are illustrated in the drawings and described in detail in the detailed description. However, this should be understood not as limiting the present invention to specific embodiments, but rather as including all modifications, equivalents, or substitutes that fall within the spirit and technical scope of the present invention. Similar reference numerals are used for similar components in the description of each drawing.

[0026] Terms such as First, Second, A, B, etc., may be used to describe various components, but the components should not be limited by such terms. The terms are used solely for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the First component may be named the Second component, and similarly, the Second component may be named the First component. The term "and / or" includes a combination of multiple related items or one of multiple related items.

[0027] When it is stated that one component is "linked" or "connected" to another component, it should be understood that this may mean that it is directly linked or connected to that other component, but that there may also be another component in between. Conversely, when it is stated that one component is "directly linked" or "directly connected" to another component, it should be understood that there is no other component in between.

[0028] The terms used in this application are used solely to describe specific embodiments and are not intended to limit the invention. Singular expressions include plural expressions unless they are clearly different in context. In this application, terms such as “includes” or “having” are intended to specify the presence of features, figures, steps, actions, components, parts, or combinations thereof as described in the specification, and should not be understood to preemptively exclude the presence or possibility of adding one or more other features, figures, steps, actions, components, parts, or combinations thereof.

[0029] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as those generally understood by a person of ordinary skill in the art to which this invention pertains. Terms as defined in commonly used dictionaries should be interpreted as having the meaning consistent with their meaning in the context of the relevant art, and not as ideal or overly formal unless explicitly defined herein.

[0030] Figure 1 shows an image of a welded joint that was misdetected by a typical dimensional measuring device.

[0031] As shown in Figure 1, a typical dimensional measuring device can inspect the quality of a finished battery after the assembly process.

[0032] Generally, during battery quality inspection, dimensional information on at least one measurement target is required to confirm the location and tensile strength of the welds, or to determine whether there are any defects in the battery structure that occurred during the manufacturing process.

[0033] Therefore, a typical dimensional measuring device acquires an image containing the object to be measured, and then obtains the boundary (edge) of the object to be measured by applying a pre-designed program logic to the image. Subsequently, the dimensional measuring device measures the dimensions of the object based on the pixel information of the boundary.

[0034] For example, a typical dimensional measuring device applies pre-designed program logic to a brightness image (Gray Model) of the object to be measured. This allows the device to detect pixels corresponding to pre-set detection condition values ​​(e.g., threshold brightness values) and, based on these, obtain the boundary of the object to be measured. Subsequently, the device measures the dimensions of the object based on the pixel information of the boundary. Here, the pre-designed program logic may be a rule-based algorithm model. Furthermore, at least one pixel information may include at least one of the following: pixel height, width, peak white, peak dark, and aspect ratio.

[0035] On the other hand, the program logic applied to general dimensional measuring devices can be pre-designed based on a pre-secured target image. However, general dimensional measuring devices that use pre-designed program logic have the disadvantage that if the brightness of the target image changes or the position of the object to be measured within the target image changes, the boundary of the object to be measured may be misdetected or the measurement range of the boundary may be incorrectly specified, reducing the reliability of the measured dimensional information.

[0036] Therefore, in order to solve these problems, the present invention describes a dimensional measuring device and method with improved reliability using a pre-trained machine learning model.

[0037] Preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0038] Figure 2 is a block diagram of a dimensional measuring device according to an embodiment of the present invention.

[0039] Referring to Figure 2, the dimensional measuring device according to an embodiment of the present invention can be used during quality inspections conducted after the assembly process of a cylindrical battery. However, it is not limited to what is described and can be applied to any process in which dimensional measurement is required.

[0040] The above-described dimensional measuring device can acquire dimensional information of the object to be measured by inputting a target image into a pre-learned model. Here, the target image may be an image of at least a part of the inside of a completed battery.

[0041] More specifically, the above-mentioned dimensional measuring device can acquire an image of the object to be measured, and extract an image of interest from the image of the object to be measured, which includes the region of interest (ROI) of the object to be measured.

[0042] Subsequently, the dimensional measuring device can input the image of interest into the already learned model to acquire dimensional information of the object to be measured.

[0043] To describe the above-mentioned dimensional measuring device in detail by its configuration, the dimensional measuring device may include a memory 100, a processor 200, a transmitting / receiving device 300, an input interface device 400, an output interface device 500, and a storage device 600.

[0044] According to the embodiment, the respective components 100, 200, 300, 400, 500, and 600 included in the dimensional measuring device are connected by a bus 700 and can communicate with each other.

[0045] Of the above configurations 100, 200, 300, 400, 500, and 600, the memory 100 and the storage device 600 can be composed of at least one of a volatile storage medium and a non-volatile storage medium. For example, the memory 100 and the storage device 600 can be composed of at least one of a read-only memory (ROM) and a random access memory (RAM).

[0046] Among these, memory 100 may contain at least one instruction executed by processor 200.

[0047] According to the embodiment, at least one command may include a command to obtain an image of interest containing the object to be measured from a target image, and a command to input the image of interest into a previously trained model and output the dimensional information of the object to be measured as result data.

[0048] Here, the already trained model mentioned above may be a deep learning-based model.

[0049] The command to output the above dimensional information as result data may include a command to segment the image of interest into physical objects using the already learned model, and to label the dimensional measurement points of the measured objects.

[0050] On the other hand, the command to acquire the above-mentioned image of interest can be pre-trained using a rule-based algorithm to acquire the image of interest containing the measurement target from the above-mentioned target image.

[0051] Furthermore, the command to acquire the above-mentioned image of interest can be pre-trained to perform image preprocessing on the target image and then acquire the image of interest.

[0052] On the other hand, processor 200 can mean a central processing unit (CPU), a graphics processing unit (GPU), or a dedicated processor on which the method according to the embodiment of the present invention is performed.

[0053] As described above, the processor 200 can execute at least one program command stored in memory 100.

[0054] Figure 3 is a flowchart of a dimension measurement method using a dimension measuring device according to an embodiment of the present invention.

[0055] Referring to Figure 3, the dimensional measuring device according to an embodiment of the present invention can acquire a target image (S1000). In this case, the target image may be an image in which at least a part of the assembled battery is photographed, and which includes the object to be measured for dimensional measurement.

[0056] According to one embodiment, the target image may be a video image of at least one weld that is the object of measurement. For example, the weld may be a weld formed on a first or second current collector in a battery. Here, the first current collector and the second current collector may be a negative electrode current collector and a positive electrode current collector, respectively.

[0057] In another embodiment, the target image may be an X-ray image of at least one electrode portion that is the object of measurement. For example, the electrode portion may be the first or second electrode of a battery.

[0058] Figure 4 is an image illustrating the image of interest of the dimensional measuring device according to an embodiment of the present invention.

[0059] Referring to Figures 3 and 4, the dimensional measuring device can acquire an image of interest from the target image (S3000). Here, the image of interest may be a partial image obtained by extracting a portion of the region of interest (ROI) of the object to be measured from the target image. According to the embodiment, the dimensional measuring device can extract the image of interest after preprocessing the target image using a rule-based algorithm.

[0060] Subsequently, the dimensional measuring device can acquire dimensional information about the object to be measured from the image of interest using the already trained model (S5000). In other words, the dimensional measuring device can acquire dimensional information about the object to be measured by inputting the image of interest into the already trained model. Here, the machine learning model may be a classification learning model using deep learning.

[0061] To explain in more detail with an example, the above-mentioned dimensional measuring device can segment the input data (Feature), which is an image of interest, according to the object to be measured (Class) using a classification learning model that has been pre-trained with training data. Subsequently, the dimensional measuring device can label the dimensional measurement points of the object to be measured, measure the pixel distance between two labeled points, and output this as result data.

[0062] Figures 5 and 6 are images illustrating a method for measuring the dimensions of a welded joint using a welded joint dimension measuring device according to one embodiment of the present invention. More specifically, Figure 5 is an image of a first current collector plate and an image of interest of the welded joint extracted from the first current collector plate image, and Figure 6 is an image preprocessed to extract the image of interest of Figure 5.

[0063] Referring to Figure 5, the dimensional measuring device according to one embodiment of the present invention can acquire a first current collector plate image, which is a target image taken of a completed battery, as described above. For example, the first current collector plate may be a negative electrode current collector plate.

[0064] Subsequently, the dimensional measuring device can preprocess the first current collector plate image, which is the target image, to obtain an image of interest. Here, the image of interest may be an image that includes a weld, which is a region of interest (ROI) within the first current collector plate image.

[0065] For example, if the first current collector plate image includes multiple welds, the dimensional measuring device can individually extract multiple images of interest, each containing one of the welds.

[0066] Referring to Figure 6, the method for extracting the image of interest can be described in more detail. The dimensional measuring device can preprocess the first current collector image in order to detect the position of at least one weld from the first current collector image.

[0067] For example, in the case of a cylindrical battery after the assembly process is complete, multiple electrode tabs can be provided in contact with a region on the first current collector plate in a superimposed state by electrodes wound inside the can. At this time, at least one weld can be formed on the first current collector plate for the electrical contact of the electrode tabs. At this time, at least one of the welds can be formed diagonally on the first current collector plate (see Figure 5). This allows the dimensional measuring device to perform a preprocessing step of rotating the image of the first current collector plate so that the weld is positioned horizontally or vertically.

[0068] Subsequently, the dimensional measuring device can confirm the position of the weld, which is the region of interest, based on the rotated image of the first current collector plate.

[0069] To illustrate in more detail with the embodiment, the dimensional measuring device can convert a rotated first current collector image into a gray value image using a pre-defined rule-based algorithm. Subsequently, the dimensional measuring device can input pre-defined condition values ​​to obtain the weld boundary (edge) within the first current collector image. For example, the pre-defined condition value may be a pre-defined pixel brightness value. In other words, the dimensional measuring device can extract at least one pixel having a pre-defined pixel brightness value as the weld boundary. Subsequently, the dimensional measuring device can confirm the position of the weld based on the information of at least one pixel extracted as the weld boundary.

[0070] Subsequently, the dimensional measuring device can convert the first current collector plate image, which is a gray value image, into a binary image. Then, as shown in Figure 6, the dimensional measuring device can acquire the start and end coordinate points (Start Point, End Point) of the welded section.

[0071] According to one embodiment, when the welded portion in the first current collector plate image is located horizontally, the starting coordinate point may be a pixel coordinate value located at the left edge of the welded portion with respect to the x-axis, and the ending coordinate point may be a pixel coordinate value located at the right edge of the welded portion with respect to the x-axis.

[0072] According to another embodiment, when the welded portion in the first current collector plate image is located vertically, the starting coordinate point may be the pixel coordinate value of the welded portion located at the uppermost end with respect to the y-axis, and the ending coordinate point may be the pixel coordinate value of the lowermost end with respect to the y-axis.

[0073] According to the embodiment, the dimension measuring device can acquire the start and end coordinate points (Start Point, End Point) in the first current collector image based on a predefined pixel brightness value. For example, if the welded portion in the first current collector image is located horizontally, the dimension measuring device can check the pixel brightness value for each coordinate in the x-axis direction and compare it with a pre-set pixel brightness value. In this case, the dimension measuring device can set the first coordinate point having a brightness value that is greater or less than the pre-set pixel brightness value as the start coordinate point (Start Point). The dimension measuring device can also acquire the last coordinate point having a brightness value that is greater or less than the pre-set pixel brightness value as the end coordinate point (End Point).

[0074] On the other hand, if the welded portion in the first current collector image is located vertically, the dimension measuring device can check the pixel brightness value for each coordinate in the y-axis direction and compare it with a previously set pixel brightness value. At this time, the dimension measuring device can set the first coordinate point having a brightness value that is greater or less than the previously set pixel brightness value as the start point. The dimension measuring device can also acquire the last coordinate point having a brightness value that is greater or less than the previously set pixel brightness value as the end point.

[0075] Subsequently, the dimension measuring device can acquire at least one dimension measurement point based on the extracted weld boundary information and start and end coordinate points (Start Point, End Point). For example, the dimension measuring device can acquire at least one dimension measurement point of the weld using a subpixel algorithm. Here, the subpixel algorithm may be a method that uses interpolation to determine points with a brightness value of 1 pixel or less as corners of an object. In this way, the dimension measuring device can finally acquire the boundary of the weld in the first current collector image based on the information of at least one dimension measurement point.

[0076] Therefore, the dimensional measuring device according to the embodiment of the present invention can extract an image of interest from the first current collector plate image based on the boundary information of the welded part and the dimensional measurement point information obtained in the preprocessing process. Subsequently, the extracted image of interest can be processed (overlaid) by the dimensional measuring device.

[0077] Subsequently, the dimensional measuring device can input the processed image of interest into a pre-trained deep-learning-based classification model.

[0078] Subsequently, the dimensional measuring device can use the pre-trained classification learning model to segment the image of interest into physical units, namely the welded portions of the first current collector plate, and then label the dimensional measurement points of the welded portions. For example, the dimensional measuring device can classify objects in the image of interest on a pixel-by-pixel basis and then label the dimensional measurement points of the welded portions.

[0079] Subsequently, the dimensional measuring device can obtain the dimensions of the object to be measured as result data by measuring the pixel distance between two labeled dimensional measurement points based on a pre-trained deep-learning-based classification learning model. Here, the classification learning model may be a model that transforms at least one training data into various forms and then uses it as training data in order to increase the amount of training data. For example, the training data can be transformed in various ways, such as vertical or horizontal inversion, changes in brightness, saturation or hue, changes in contrast, rotation or translation, and the addition of noise.

[0080] Figures 7 and 8 are images illustrating a method for measuring the dimensions of a welded joint using a dimensional measuring device according to another embodiment of the present invention. More specifically, Figure 7 is an image of a second current collector plate and an image of interest extracted from the second current collector plate image, and Figure 8 is an image preprocessed to extract the image of interest from Figure 7.

[0081] Referring to Figures 7 and 8, the dimensional measuring device according to another embodiment of the present invention can acquire the second current collector plate image, which is the target image, as described above. Here, the second current collector plate image is an image taken after the assembly process and may be an image of the positive electrode current collector plate.

[0082] Subsequently, the dimensional measuring device can perform image preprocessing based on the second current collector plate image to acquire the image of interest.

[0083] More specifically, generally, the welded area on the second current collector image can be arranged in a way that is symmetrical to the origin, as shown in Figure 7, with a specific pattern. This allows the dimensional measuring device to deform the target image (polar space data) so that the welded area, which is the region of interest in the second current collector image, is positioned horizontally.

[0084] Subsequently, the dimensional measuring device can confirm the position of the weld, which is the region of interest, based on the deformed second current collector plate image.

[0085] To illustrate in more detail with the examples, the dimensional measuring device can convert a rotated second current collector image into a gray value image using a pre-defined rule-based algorithm. Subsequently, the dimensional measuring device can input predefined condition values ​​to obtain the weld boundary (edge) within the second current collector image.

[0086] Subsequently, the dimensional measuring device can convert the first current collector plate image, which is a gray value image, into a binary image. After that, the dimensional measuring device can acquire the start and end coordinate points (Start Point, End Point) of the welded portion. According to the embodiment, the dimensional measuring device can acquire the start and end coordinate points (Start Point, End Point) in the second current collector plate image based on a predefined pixel brightness value.

[0087] Subsequently, the dimensional measuring device can acquire at least one dimensional measurement point based on the extracted weld boundary information and start and end coordinate points (Start Point, End Point). This allows the dimensional measuring device to extract an image of interest from the second current collector image based on the weld boundary information and dimensional measurement point information acquired during the preprocessing step.

[0088] Subsequently, the extracted image of interest can be processed (overlaid) by the dimensional measuring device described above.

[0089] Subsequently, the dimensional measuring device can input the processed image of interest into a pre-trained deep-learning-based classification model.

[0090] Subsequently, the dimensional measuring device can use the pre-trained classification learning model to segment the image of interest into the welded portion of the second current collector plate, which is a physical unit.

[0091] Subsequently, the dimensional measuring device can label the dimensional measurement points of the welded joint based on a pre-trained deep-learning-based classification learning model.

[0092] Subsequently, the dimension measuring device can obtain the dimensions of the object to be measured as result data by measuring the pixel distance between two labeled dimension measurement points. Here, the classification learning model may be a model that, in order to increase the amount of training data, transforms at least one training data into various forms and then uses it as training data. For example, the training data can be transformed in various ways, such as vertical or horizontal inversion, changes in brightness, saturation or hue, changes in contrast, rotation or translation, and the addition of noise.

[0093] Figure 9 is an image illustrating a method for measuring the dimensions of an electrode using a pre-learned model within a dimensional measuring device according to another embodiment of the present invention.

[0094] Referring to Figure 9, the dimensional measuring device according to yet another embodiment of the present invention can acquire an image of the electrode portion taken after the assembly process as the target image, as described above. For example, the electrode portion image may be an X-ray image.

[0095] Subsequently, the dimensional measuring device can extract an image of interest from the electrode image, which includes the connection between the positive and negative electrodes, which is the region of interest (ROI).

[0096] Subsequently, the extracted images of interest can be processed (overlaid) by the dimensional measuring device described above.

[0097] Subsequently, the dimensional measuring device can input the processed image of interest into a pre-trained deep learning classification model. This allows the dimensional measuring device to acquire length data of the electrode portion connecting the negative and positive ends of the electrode.

[0098] More specifically, the dimensional measuring device can segment an image of interest input to a pre-trained deep-learning-based classification learning model into physical units. In some embodiments, the classification learning model may be a semantic segmentation model. For example, the dimensional measuring device can segment an image of interest into electrode images.

[0099] Subsequently, the dimensional measuring device can label the dimensional measurement points of the weld based on the classification learning model. For example, the dimensional measurement points within the electrode may be one end of the negative electrode (first point) located in close proximity to the tab region, and one end of the positive electrode (second point) located in close proximity to the electrode body (the region where the electrode holding portion is formed). In this case, the dimensional measuring device can label multiple points to correspond to the number of turns of the electrode.

[0100] Subsequently, the dimensional measuring device can obtain the length of the electrode being measured as result data by measuring the pixel distance between the labeled first and second points based on a pre-trained deep-learning-based classification learning model. Here, the classification learning model may be a model that transforms at least one training data into various forms and then uses it as training data in order to increase the amount of training data. For example, the training data can be transformed in various ways, such as vertical or horizontal inversion, changes in brightness, saturation or hue, changes in contrast, rotation or movement, and the addition of noise.

[0101] The dimensional measuring apparatus and method according to embodiments of the present invention have been described above.

[0102] The dimensional measuring apparatus and method according to embodiments of the present invention can acquire highly reliable dimensional measurement result data with improved accuracy by acquiring dimensional information of a specific object in an image of interest using a pre-trained learning model.

[0103] The operation of the methods according to the embodiments and experimental examples of the present invention can be embodied as a computer-readable program or code on a computer-readable recording medium. A computer-readable recording medium includes all types of recording devices that store data that can be read by a computer system. Furthermore, computer-readable recording media can be distributed across networked computer systems, allowing computer-readable programs or code to be stored and executed in a distributed manner.

[0104] Furthermore, computer-readable recording media can include hardware devices specially configured to store and execute program instructions, such as ROM, RAM, and flash memory. Program instructions can include not only machine code generated by a compiler, but also high-level language code that can be executed by a computer using an interpreter or the like.

[0105] Some aspects of the present invention have been described in the context of apparatus, but they can also be described by corresponding methods, where a block or apparatus corresponds to a method step or a feature of a method step. Similarly, aspects described in the context of a method can be described by corresponding blocks or items or features of corresponding apparatus. Some or all of the method steps can be carried out by (or using) hardware devices such as, for example, a microprocessor, a programmable computer, or an electronic circuit. In some embodiments, one or more of the most important method steps can be carried out by such devices.

[0106] While preferred embodiments of the present invention have been described above with reference to the present invention, those skilled in the art will understand that the present invention can be modified and altered in various ways without departing from the spirit and scope of the invention as set forth in the following claims. [Explanation of Symbols]

[0107] 100: Memory 200: Processor 300: Transceiver / Receiver 400: Input Interface Device 500: Output interface device 600: Storage device 700: Bus

Claims

1. A learning model-based dimensional measuring device, memory; and Includes a processor that executes at least one instruction stored in the memory, The aforementioned at least one instruction, A command to obtain an image of interest containing the object to be measured, from a target image, by performing image preprocessing using a rule-based algorithm to rotate the portion of the object to be measured included in the image of interest horizontally or vertically, and The aforementioned image of interest is input into a pre-trained classification learning model to segment the image of interest into physical objects. The divided physical object is identified by inputting it into the already trained classification learning model and labeling the two dimensional measurement points of the object. The command includes a command to output dimensional information of the object to be measured, indicating the distance between the two aforementioned dimensional measurement points, as result data. The aforementioned target image includes an image of at least a portion of the inside of a battery, and is a dimensional measuring device.

2. The dimension measuring device according to claim 1, wherein the previously trained model is a deep-learning-based model.

3. The command to acquire the aforementioned image of interest is, The dimensional measuring device according to claim 1, comprising an instruction to acquire an image of interest containing the object to be measured from the target image using a rule-based algorithm.

4. A learning model-based dimensional measurement method, The steps include: obtaining an image of interest containing the object to be measured from a target image by performing image preprocessing using a rule-based algorithm to rotate the portion of the object to be measured included in the image of interest horizontally or vertically; and The aforementioned image of interest is input into a pre-trained classification learning model to segment the image of interest into physical objects. The divided physical object is identified by inputting it into the already trained classification learning model and labeling the two dimensional measurement points of the object. The step includes measuring the distance between the two aforementioned dimensional measurement points and outputting the dimensional information of the object to be measured, which indicates the distance, as result data. The aforementioned target image includes an image of at least a portion of the inside of a battery, and is a method for measuring dimensions.

5. The dimensional measurement method according to claim 4, wherein the previously trained model is a deep-learning-based model.

6. The step of acquiring the aforementioned image of interest is: The dimensional measurement method according to claim 4, further comprising the step of obtaining an image of interest containing the object to be measured from the target image using a rule-based algorithm.

Images