Welding detection device and welding detection method

Abstract

The welding detection apparatus of the present disclosure comprises: a scanner for photographing a battery to obtain a two-dimensional image and a three-dimensional image; a data processing unit which generates fusion data based on the two-dimensional image and the three-dimensional image; the object recognition unit inputs the fusion data into an artificial intelligence model trained to recognize an object, and recognizes a welding area from the fusion data; and a welding determination unit that determines whether or not there is a defect in the welded portion of the battery on the basis of the welding region.

Description

Technical Field

[0001] The present disclosure relates to a secondary battery, and more particularly, to a welding inspection device and a welding inspection method. Background Art

[0002] In recent years, secondary batteries can be reused multiple times through charging and discharging. Due to the economic and environmental protection characteristics of secondary batteries, secondary batteries are widely used in various industries. In order to manufacture batteries, a welding process may be performed. Therefore, there is a need for a device and method that can accurately detect welding defects in batteries. Summary of the Invention

[0003] (I) Technical Problem to be Solved

[0004] Embodiments of the present disclosure aim to provide a welding inspection device and a welding inspection method for detecting welding defects in batteries.

[0005] The present disclosure can be widely applied to green technology fields such as solar power generation and wind power generation. In addition, the present disclosure can also be applied to environmental protection devices such as electric vehicles and hybrid vehicles that prevent climate change by suppressing air pollution and greenhouse gas emissions.

[0006] (II) Technical Solution

[0007] The welding inspection device according to an embodiment of the present disclosure includes: a scanner that photographs a battery to obtain a two-dimensional image and a three-dimensional image; a data processing unit that generates fusion data based on the two-dimensional image and the three-dimensional image; an object recognition unit that inputs the fusion data into an artificial intelligence model trained to recognize an object and recognizes a welding area from the fusion data; and a welding determination unit that determines whether there is a defect in the welded part based on the welding area.

[0008] In an embodiment, the three-dimensional image may include a plurality of pixels mapping first height values within a first unit, and the two-dimensional image may include a plurality of pixels mapping first brightness values within a second unit.

[0009] In an embodiment, the data processing unit may perform a scaling process to reduce the number of pixels in the two-dimensional image and the three-dimensional image, and generate fusion data using the scaled two-dimensional image and three-dimensional image.

[0010] In an embodiment, the data processing unit may generate fusion data using a second brightness value obtained by normalizing the first brightness value and a second height value obtained by normalizing the first height value.

[0011] In an embodiment, the data processing unit may perform a weighted operation on the corresponding second brightness value and second height value to generate fusion data.

[0012] In an embodiment, each of the second height value and the second brightness value may be a value within a third unit, and the third unit is smaller than the first unit and the second unit.

[0013] In an embodiment, the second height value may be proportional to the ratio of a first difference and a second difference, where the first difference is the difference between a first height value and a set lower limit value of the first height value, and the second difference is the difference between a set upper limit value and the set lower limit value of the first height value.

[0014] In an embodiment, the second height value may be a value obtained by multiplying the ratio by the maximum value within the third unit, and the third unit is smaller than the first unit.

[0015] In an embodiment, the set upper limit value may be a value obtained by adding a first set value to a middle value, where the middle value is the middle value of the first height values of each pixel included in the three-dimensional image, and the set lower limit value may be a value obtained by subtracting a second set value from the middle value.

[0016] In an embodiment, the second brightness value may be proportional to the ratio of a first brightness value and a value obtained by adding 1 to the maximum value of the second unit.

[0017] In an embodiment, the second brightness value may be a value obtained by multiplying the ratio by the maximum value within the third unit, and the third unit is smaller than the second unit.

[0018] In an embodiment, the welding area may include a weld area, and the welding determination unit may determine whether there is a defect in the welded part based on whether the weld area exists within a preset region of interest.

[0019] In an embodiment, the welding area may include a raised area, and the region of interest may be the raised area.

[0020] In an embodiment, the welding area may include a weld area, and the welding determination unit may determine whether there is a defect in the welded part based on the length of the weld area.

[0021] In an embodiment, the welding determination unit may determine whether there is a defect in the welded part based on the first height value of the three-dimensional image corresponding to the weld area.

[0022] The welding detection method according to an embodiment of the present disclosure may include the following steps: photographing a battery to obtain a two-dimensional image and a three-dimensional image; generating fusion data based on the two-dimensional image and the three-dimensional image; identifying a welding area from the fusion data based on an artificial intelligence model trained to identify an object; and determining whether there is a defect in the welded part of the battery based on the welding area.

[0023] In an embodiment, the step of generating the fusion data may include: a step of performing a scaling process to reduce the number of pixels in the two-dimensional image and the three-dimensional image; and a step of generating the fusion data by using the scaled two-dimensional image and three-dimensional image.

[0024] In an embodiment, the step of generating the fusion data may include: a step of normalizing a first luminance value of one pixel among a plurality of pixels mapped to the two-dimensional image to obtain a second luminance value, and normalizing a first height value of a pixel corresponding to one pixel among a plurality of pixels of the three-dimensional image to obtain a second height value; and a step of generating the fusion data by using the second luminance value and the second height value.

[0025] In an embodiment, the maximum value of the bits having the second height value may be less than the maximum value of the bits having the first height value, and the maximum value of the bits having the second luminance value may be less than the maximum value of the bits having the first luminance value.

[0026] In an embodiment, in the step of generating the fusion data by using the second luminance value and the second height value, a weighted operation may be performed on the second luminance value and the second height value to obtain a result value, and the fusion data including the result value may be generated.

[0027] (III) Beneficial Effects

[0028] Embodiments of the present disclosure may provide a welding detection device and a welding detection method for detecting welding defects of a battery.

[0029] Embodiments of the present disclosure may provide a welding detection device and a welding detection method that are not affected by noise.

[0030] Embodiments of the present disclosure may provide a welding detection device and a welding detection method for accurately detecting welding defects of a battery.

[0031] Embodiments of the present disclosure may provide a welding detection device and a welding detection method for quickly detecting welding defects of a battery. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 is a block diagram for explaining a welding detection device according to an embodiment.

[0033] Figure 2 is a diagram for explaining a welded portion of a battery according to an embodiment.

[0034] Figure 3 is a diagram for explaining a welding detection method of a welding detection device according to an embodiment.

[0035] Figure 4 is a diagram for explaining a data processing procedure according to an embodiment.

[0036] Figure 5 is a diagram for explaining the generation process of fused data according to an embodiment.

[0037] Figure 6 is a diagram for explaining the normalization of height values and luminance values according to an embodiment.

[0038] Figure 7 is a diagram for explaining the welding areas identified in the fused data according to an embodiment.

[0039] Explanation of reference numerals:

[0040] 100: Welding detection device 110: Scanner

[0041] 120: Processor 121: Data processing unit

[0042] 123: Object recognition unit 125: Welding judgment unit Detailed implementation manners

[0043] The description of the structure or function of the embodiments disclosed in this specification or application is only an example to illustrate the embodiments according to the technical idea of the present invention. The embodiments according to the technical idea of the present invention can be implemented in various forms other than the embodiments disclosed in this specification or application, and should not be construed as limiting the technical idea of the present invention to the embodiments described in this specification or application.

[0044] Figure 1 is a block diagram for explaining the welding detection device according to an embodiment. Figure 2 is a diagram for explaining the welding part of the battery according to an embodiment.

[0045] Referring to Figure 1 and Figure 2 , the welding detection device 100 according to an embodiment can detect the welding part 250 of the battery 200. The welding detection device 100 can determine whether there are defects in the welding part 250 of the battery 200.

[0046] The battery 200 can be a secondary battery capable of being charged and discharged multiple times. The type of the battery 200 can be classified into battery cells, battery modules, battery packs, etc. according to units. The battery 200 of the present disclosure is not limited to its type and can be applied to various types. According to an embodiment Figure 2The battery 200 may be a battery module. The battery module may include a plurality of battery cells and a bus bar 210 electrically connecting the plurality of battery cells. In an embodiment, the bus bar 210 may include a raised portion (embo) 220 having a higher height than the surroundings, a slit hole 230 formed through a region of the bus bar 210 (or the raised portion 220), and a welding portion 250. Here, the height may refer to the length in the height direction (e.g., the Z-axis direction). On the other hand, the raised portion 220 may be omitted. The welding portion 250 may be formed by welding in a state where the electrode tab of the battery cell is inserted into the slit hole 230. On the other hand, this is only one embodiment, and the welding portion 250 may be formed at different positions within the battery 200.

[0047] The welding inspection device 100 may include a scanner 110 and a processor 120.

[0048] The scanner 110 may capture the battery 200 to obtain a two-dimensional image and a three-dimensional image. Here, the capture area of the scanner 110 may be an area including the welding portion 250 of the battery 200. The two-dimensional image may be data representing the brightness of each position (or coordinate), and the three-dimensional image may be data representing the height of each position (or coordinate).

[0049] The processor 120 may calculate and process data. For example, the processor 120 may process the two-dimensional image and the three-dimensional image to determine whether there is a defect in the welding portion 250. In an embodiment, the processor 120 may include at least one of a Digital Signal Processor (DSP), a Microprocessor, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), an Accelerated Processing Unit (APU), an Application Processor (AP), a Neural Processing Unit (NPU), and a controller.

[0050] In an embodiment, the processor 120 may include a data processing unit 121, an object recognition unit 123, and a welding determination unit 125. The data processing unit 121 may generate fusion data based on the two-dimensional image and the three-dimensional image. The object recognition unit 123 may recognize the welding area in the fusion data based on an artificial intelligence model. Here, the artificial intelligence model may be pre-trained to recognize objects. The welding determination unit 125 may determine whether there is a defect in the welding portion 250 based on the welding area.

[0051] Figure 3 This is a diagram for explaining the welding inspection method of the welding inspection device according to an embodiment.

[0052] Refer to Figures 1 to 3 , the welding inspection method of the welding inspection device 100 includes: photographing the battery 200 to obtain a two-dimensional image and a three-dimensional image (S110); generating fusion data based on the two-dimensional image and the three-dimensional image (S130); identifying the welding area based on an artificial intelligence model (S150); and determining whether there is a defect in the welding part 250 of the battery 200 (S170).

[0053] In an embodiment, the scanner 110 can photograph the battery 200 to obtain a two-dimensional image and a three-dimensional image S110. In an embodiment, the scanner 110 can include at least one of an optical scanner, a structured light scanner, and a time-of-flight (ToF) scanner. The optical scanner can receive light (such as laser, visible light, infrared light, etc.) reflected from the surface of the battery 200 to obtain a two-dimensional image and / or a three-dimensional image. The structured light scanner can irradiate structured light with a specific pattern onto the surface of the battery 200, and then use the degree of distortion of the reflected light pattern to obtain a three-dimensional image. The ToF scanner can use the time difference between the time when the laser irradiates the surface of the battery 200 and the time when the laser reflected from the surface of the battery 200 is received to obtain a three-dimensional image. However, this is only one embodiment, and the type of the scanner 110 is not limited thereto.

[0054] In an embodiment, the data processing unit 121 can generate fusion data (or a fusion image) based on the two-dimensional image and the three-dimensional image (S130). In an embodiment, the data processing unit 121 can synthesize the two-dimensional image and the three-dimensional image to generate fusion data. In another embodiment, the data processing unit 121 can preprocess at least one of the two-dimensional image and the three-dimensional image and then synthesize them to generate fusion data. Here, the preprocessing can include at least one of resizing processing, normalizing processing, and bit unit conversion processing.

[0055] In a specific embodiment, the step of generating fusion data can include: performing a resizing process to reduce the number of pixels in the two-dimensional image and the three-dimensional image; and using the resized two-dimensional image and three-dimensional image to generate fusion data.

[0056] In an embodiment, the step of generating the fusion data may include: normalizing a first luminance value of one pixel among a plurality of pixels mapped to a two-dimensional image to obtain a second luminance value, and normalizing a first height value of a pixel corresponding to one pixel among a plurality of pixels of a three-dimensional image to obtain a second height value; and generating the fusion data by using the second luminance value and the second height value.

[0057] In an embodiment, the maximum value of the bits of the second height value may be less than the maximum value of the bits of the first height value, and the maximum value of the bits of the second luminance value may be less than the maximum value of the bits of the first luminance value.

[0058] In an embodiment, in the step of generating the fusion data by using the second luminance value and the second height value, a weighted operation may be performed on the second luminance value and the second height value to obtain a result value, and the fusion data including the result value may be generated.

[0059] In an embodiment, the object recognition unit 123 may identify a welding area based on a trained artificial intelligence model (S150). For example, the artificial intelligence model may be trained by using a large amount of training data to determine a rule (or pattern) for identifying an object. The trained artificial intelligence model of the present disclosure may be a program pre-trained to identify an object from input data. Here, the input data may be the fusion data. The artificial intelligence model of the present disclosure may be trained by the welding detection device 100 or an external device. In an embodiment, the artificial intelligence model may be pre-trained by at least one of supervised training and unsupervised training. Supervised training may be a method of training the artificial intelligence model by using input data and label data indicating the correct answer for the input data. For example, supervised training may be a method of training the relationship between the data input to the artificial intelligence model and the data output from the artificial intelligence model such that when the input data is input to the artificial intelligence model, the result data output from the artificial intelligence model is the same as the label data. Unsupervised training may be a method of training the artificial intelligence model to find patterns from the input data without label data and clustering similar data for prediction. In an embodiment, the welding detection device 100 may further include a storage device for storing the trained artificial intelligence model.

[0060] In an embodiment, the welding determination unit 125 may determine whether there is a defect in the welding portion 250 of the battery 200 based on the identified welding area (S170). In an embodiment, the step S170 of determining whether there is a defect in the welding portion 250 may include at least one of a Region of Interest (ROI) check S171, a welding length check S172, a side welding check S173, and a bead height check S174.

[0061] In an embodiment, the ROI inspection S171 may determine whether there is a defect in the welded part 250 based on whether the welding area identified in the fusion data exists within a preset region of interest. In an embodiment, the welding area may include at least one of a weld bead area, a raised area, and a slit hole area. The region of interest may be a region having a preset size and existing at a preset position in the fusion data. The weld length inspection S172 may determine whether there is a defect in the welded part 250 based on whether the length of the weld area identified in the fusion data is within a reference range. The offside welding inspection S173 may determine whether there is a defect in the welded part 250 based on whether the length of the offside area that is biased to one side in the weld area identified in the fusion data is within a reference range. The weld height inspection S174 may determine whether there is a defect in the welded part 250 based on the first height value of the three-dimensional image corresponding to the weld area identified in the fusion data.

[0062] Figure 4 is a diagram for explaining a data processing procedure according to an embodiment.

[0063] Referring to Figure 1 and Figure 4 , the scanner 110 may acquire a three-dimensional image 410 and a two-dimensional image 420.

[0064] In an embodiment, the three-dimensional image 410 may include a plurality of pixels that map first height values within a first bit unit. Each pixel among the plurality of pixels may represent a position (or coordinate) on a plane (e.g., the XY plane). For example, each pixel may represent a unique XY coordinate that does not repeat with others. A separate first height value of the first bit unit may be mapped to each pixel among the plurality of pixels. For example, a larger first height value may indicate a higher height. The first bit unit may represent the number of bits assigned to one pixel. For example, if the first bit unit is 16 bits, the first height value may be a value in the range of 0 to 65535.

[0065] In an embodiment, the two-dimensional image 420 may include a plurality of pixels that map first luminance values within a second bit unit. Each pixel among the plurality of pixels may represent a position (or coordinate) on a plane (e.g., the XY plane). For example, each pixel may represent a unique XY coordinate that does not repeat with others. A separate first luminance value within the second bit unit may be mapped to each pixel among the plurality of pixels. For example, a larger first luminance value may indicate a brighter luminance. The second bit unit may represent the number of bits assigned to one pixel. For example, if the second bit unit is 10 bits, the first luminance value may be a value in the range of 0 to 1023.

[0066] The data processing unit 121 can generate fused data 430 based on the three-dimensional image 410 and the two-dimensional image 420. The fused data 430 can include a plurality of pixels representing positions (or coordinates) on a plane (e.g., the XY plane). Values generated based on the first height value of the corresponding pixel of the three-dimensional image 410 and the first luminance value of the corresponding pixel of the two-dimensional image 420 can be mapped to each pixel in the fused data 430. Here, the corresponding pixels can represent pixels at the same position. For example, the data processing unit 121 can generate a result value based on the first height value of the pixel at a specific position in the three-dimensional image 410 and the first luminance value of the pixel at the same position in the two-dimensional image 420, respectively. The data processing unit 121 can generate fused data 430 including the result value of each pixel among the plurality of pixels.

[0067] The object recognition unit 123 can input the fused data 430 into an artificial intelligence model trained to recognize an object, and obtain output data 450 output from the trained artificial intelligence model. Here, the object can be a welding area including at least one of a weld area, a raised area, and a slit hole area. The weld area can be an area representing the welded part of the battery, the raised area can be an area representing the raised part of the battery, and the slit hole area can be an area representing the slit hole of the battery. The output data 450 can be an image, but is not limited thereto, and can also be various types of data. The output data 450 can include recognition information for each of the weld area, the raised area, and the slit hole area. For example, the recognition information can include information such as the boundaries, lengths, and positions of the weld area, the raised area, and the slit hole area, respectively.

[0068] Figure 5 is a diagram for explaining the generation process of fused data according to an embodiment.

[0069] Referring to Figure 1 and Figure 5 , the data processing unit 121 according to an embodiment can generate fused data based on the three-dimensional image 510 and the two-dimensional image 520.

[0070] In an embodiment, the data processing unit 121 can generate fused data after preprocessing the three-dimensional image 510 and the two-dimensional image 520. The preprocessing can include at least one of scaling processes 511, 521, normalization processes 513, 523, and bit unit conversion processes 515, 525. The data processing unit 121 can perform a weighted operation 530 on the preprocessed three-dimensional image and two-dimensional image to generate fused data. In another embodiment, the data processing unit 121 can perform a weighted operation 530 on the three-dimensional image 510 and the two-dimensional image 520 without performing preprocessing.

[0071] In an embodiment, the weighted operation 530 may be a calculation performed according to Mathematical Formula 1 531. Here, H is a height value mapped to a pixel (e.g., a first height value, a second height value, etc.), L is a luminance value mapped to a pixel (e.g., a first luminance value, a second luminance value, etc.), and w may be a weight. w may be a number greater than or equal to 0 and less than or equal to 1. For example, w may be a value such as 0.5, 0.4, etc. FV is a result value mapped to a pixel and may be included in the fusion data. On the other hand, the above example is just one embodiment, and the weighted operation 530 may be implemented with various mathematical formulas.

[0072] In an embodiment, the data processing unit 121 may perform scaling processes 511 and 521 to reduce the number of pixels in the three-dimensional image 510 and the two-dimensional image 520. The three-dimensional image 510 may include a plurality of pixels mapping a first height value, and the two-dimensional image 520 may include a plurality of pixels mapping a first luminance value. The sizes of the three-dimensional image and the two-dimensional image after the scaling processes 511 and 521 have been performed may be reduced. The scaling processes 511 and 521 may utilize interpolation algorithms in various ways. In an embodiment, the sizes of the three-dimensional image 510 and the two-dimensional image 520 may be 3200x1135. Here, 3200 may be the number of pixels in the horizontal direction (e.g., the X-axis direction), and 1135 may be the number of pixels in the vertical direction (e.g., the Y-axis direction). For example, if the magnification factors in the horizontal and vertical directions of the scaling processes 511 and 521 are 0.4 times, the sizes of the three-dimensional image and the two-dimensional image after the scaling processes 511 and 521 have been performed may be 1280x454. That is, the number of pixels in the horizontal and vertical directions may be reduced. Therefore, the amount of data that the welding detection device 100 needs to process may be reduced.

[0073] In an embodiment, the data processing unit 121 may generate fusion data using the three-dimensional image and the two-dimensional image after the scaling processes 511 and 521 have been performed. Specifically, the data processing unit 121 may perform a weighted operation 530 on the height value of the three-dimensional image and the luminance value of the two-dimensional image after the scaling processes 511 and 521 have been performed to generate fusion data. For example, the height value and the luminance value of the pixels corresponding to each other in the three-dimensional image and the two-dimensional image after the scaling processes 511 and 521 have been performed may be input as H and L in the Mathematical Formula 1 531 of the weighted operation 530 to generate fusion data including the result value FV.

[0074] In an embodiment, the data processing unit 121 may perform normalization processes 513 and 523 to change the first height value of the three-dimensional image 510 (or the height value of the three-dimensional image that has undergone the scaling process 511) and the first luminance value of the two-dimensional image 520 (or the luminance value of the two-dimensional image that has undergone the scaling process 521) into values within a predetermined range. In this case, the data processing unit 121 may obtain a second height value obtained by normalizing the first height value and may obtain a second luminance value obtained by normalizing the first luminance value. Specific embodiments will be described below with reference to Figure 6 illustrate specific embodiments.

[0075] In an embodiment, the data processing unit 121 may generate fusion data by using the second height value and the second luminance value obtained by performing the normalization processes 513 and 523. In an embodiment, the data processing unit 121 may perform a weighted operation 530 on the corresponding second luminance value and second height value to generate fusion data. For example, the data processing unit 121 may input the height value and the luminance value mapped to corresponding pixels into H and L of the mathematical formula 1 531 of the weighted operation 530 to generate fusion data including the result value FV.

[0076] In an embodiment, the data processing unit 121 may perform bit unit conversion processes 515 and 525 to change the bit unit of the three-dimensional image 510 (or the three-dimensional image that has undergone the normalization process 513) and the two-dimensional image 520 (or the two-dimensional image that has undergone the normalization process 523). Here, the bit unit may represent the representable range of values (or information). For example, if the bit unit is 8 bits, values in the range of 0 to 255 (e.g., height values, luminance values, etc.) may be mapped to each of a plurality of pixels. For example, if the bit unit is 16 bits, values in the range of 0 to 65535 (e.g., height values, luminance values, etc.) may be mapped to each of a plurality of pixels. The 8-bit case has a narrower representable range than the 16-bit case, but can reduce the data processing time. In an embodiment, the second height value and the second luminance value that have undergone the bit unit conversion processes 515 and 525 may be values within a third bit unit.

[0077] In an embodiment, the first bit unit conversion process 515 may change the first bit unit of the three-dimensional image 510 (or the three-dimensional image that has undergone preprocessing (e.g., the scaling process 511 or the normalization process 513)) into a third bit unit. For example, the first bit unit may be N bits and the third bit unit may be M bits. Here, N and M are natural numbers respectively, and N may be greater than M. For example, N may be 16 and M may be 8.

[0078] In an embodiment, the second unit conversion process 525 may change the second unit of the two-dimensional image 520 (or the two-dimensional image that has been preprocessed (e.g., the scaling process 521 or the normalization process 523)) to a third unit. For example, the second unit may be K bits, and the third unit may be M bits. Here, K and M are natural numbers respectively, and K may be greater than M. For example, K may be 10 and M may be 8. In an embodiment, K may be the same as N or different from N.

[0079] In an embodiment, the data processing unit 121 may perform a weighted operation 530 on the three-dimensional image and the two-dimensional image that have undergone the bit unit conversion processes 515 and 525 to obtain fused data.

[0080] Figure 6 FIG. is for explaining the normalization of the height value and the luminance value according to an embodiment.

[0081] Referring to Figure 1 and Figure 6 the data processing unit 121 may perform a normalization process on the first height value H1 mapped to each pixel included in the three-dimensional image to obtain a second height value H2. In an embodiment, the data processing unit 121 may perform a bit unit conversion process to change the first bit unit of the second height value H2 to a third bit unit. The third bit unit may be smaller than the first bit unit. Here, the first bit unit may be N bits, and the third bit unit may be M bits. For example, in the case of N bits, the first height value H1 may be a value in the range of 0 to n. Here, n may be 2 N - 1. For example, in the case of M bits, the second height value H2 may be a value in the range of 0 to m. Wherein, m may be 2 M - 1.

[0082] In an embodiment, according to Mathematical Formula 2 613, the second height value H2 may be proportional to the ratio of the first difference and the second difference. The first difference may be the difference between the first height value H1 and the set lower limit value HL of the first height value H1. The second difference may be the difference between the set upper limit value HU and the set lower limit value HL of the first height value H1. In an embodiment, the second height value H2 may be the value obtained by multiplying the ratio of the first difference and the second difference by the maximum value m within the third bit unit. For example, if the third bit unit is M bits, the maximum value m within the third bit unit may be 2 M - 1.

[0083] In an embodiment, according to Mathematical Formula 3614, the set upper limit value HU of the first height value H1 can be the value obtained by adding the first set value a to the intermediate value HM. Here, the intermediate value HM can be the intermediate value of the first height value H1 of each pixel included in the three-dimensional image. The set lower limit value HL of the first height value H1 can be the value obtained by subtracting the second set value b from the intermediate value HM. The first set value a and the second set value b can be preset values. In an embodiment, the first set value a and the second set value b can be the same value or different values.

[0084] In an embodiment, the set upper limit value HU of the first height value H1 can be changed to the maximum value m within the third unit of the second height value H2. The set lower limit value HL of the first height value H1 can be changed to the minimum value (0) of the second height value H2.

[0085] In an embodiment, the data processing unit 121 can perform a normalization process on the first luminance value L1 mapped to each pixel included in the two-dimensional image to obtain the second luminance value L2. In an embodiment, the data processing unit 121 can perform a bit unit conversion process to change the second unit of the second luminance value L2 to the third unit. The second unit can be smaller than the first unit. Here, the second unit can be K bits, and the third unit can be M bits. For example, in the case of N bits, the first height value H1 can be a value within the range of 0 to k. Here, k can be K 2 M -1. For example, in the case of M bits, the second luminance value L2 can be a value within the range of 0 to m. Here, m can be

[0086] 2

[0087] -1. [[ID=1\1]]

[0086] In an embodiment, according to Mathematical Formula 4623, the second luminance value L2 can be proportional to the ratio of the first luminance value L1 and the value obtained by adding 1 to the maximum value k of the second unit. In an embodiment, the second luminance value L2 can be the value obtained by multiplying the ratio of the first luminance value L1 and the value obtained by adding 1 to the maximum value k of the second unit by the maximum value m within the third unit.

[0087] In an embodiment, the maximum value k of the second unit of the first luminance value L1 can be changed to the maximum value m of the third unit of the second height value L2. The minimum value (0) of the first height value H1 can be changed to the minimum value (0) of the second height value H2.

[0088] Figure 7 It is a diagram for explaining the welding area identified in the fusion data according to the embodiment.

[0089] Refer to Figure 1 and Figure 7, the object recognition unit 123 can recognize the objects within the welding area in the fused data. In an embodiment, the welding area may include at least one of a raised area 720, a slit hole area 730, and a weld area 750. The raised area may be an area representing the raised part of the battery, the slit hole area may be an area representing the slit hole of the battery, and the weld area 750 may be an area representing the welded part of the battery.

[0090] In an embodiment, the welding determination unit 125 can determine whether there is a defect in the welded part according to whether the weld area 750 exists within a preset region of interest. For example, if the weld area 750 does not exist within the preset region of interest, the welding determination unit 125 can determine that there is a defect in the welded part. In an embodiment, the region of interest may be a preset region. For example, the region of interest may be a region in the fused data that has a preset size and exists at a preset position. In another embodiment, the region of interest may be set as the raised area 720.

[0091] In an embodiment, the welding determination unit 125 can determine whether there is a defect in the welded part of the battery according to the length of the weld area 750. For example, assume that the long side of the welded part is formed along the longitudinal direction (e.g., the Y-axis direction).

[0092] In an embodiment, the welding determination unit 125 can determine whether there is a defect in the welded part according to whether the total length Y1 in the longitudinal direction (e.g., the Y-axis direction) of the weld area 750 is within a reference range. For example, if the total length Y1 of the weld area 750 is not within the reference range, the welding determination unit 125 can determine that there is a defect in the welded part.

[0093] In an embodiment, the welding determination unit 125 can determine whether there is an offset area 755 in the weld area 750, and determine whether there is a defect in the welded part according to whether the length Y2 in the longitudinal direction (e.g., the Y-axis direction) of the offset area 755 is within a reference range. For example, the welding determination unit 125 can determine an area in the total length X1 in the transverse direction (e.g., the X-axis direction) of the weld area 750 that has a length X2 less than a reference value as the offset area 755. In another example, the welding determination unit 125 can determine an area in the weld area 750 that deviates more than a reference value in the transverse direction (e.g., the X-axis direction) with respect to the slit hole area 730 (or the guiding area 735) as the offset area 755. The guiding area 735 may be a preset area. For example, if the length Y2 of the offset area 755 is not within the reference range, the welding determination unit 125 can determine that there is a defect in the welded part.

[0094] In an embodiment, the welding determination unit 125 may determine whether there is a defect in the welded part based on a first height value of a three-dimensional image corresponding to the weld region 750. Here, the first height value may be the height value of the three-dimensional image of the non-fused data. In an embodiment, the welding determination unit 125 may determine whether there is a defect in the welded part according to whether the first height value is within a reference range. For example, if the first height value is not within the reference range, the welding determination unit 125 may determine that there is a defect in the welding.

Claims

1. A welding detection device, comprising: A scanner that photographs a battery to obtain a two-dimensional image and a three-dimensional image; A data processing unit that generates fusion data based on the two-dimensional image and the three-dimensional image; An object recognition unit that, based on an artificial intelligence model trained to recognize an object, recognizes a welding area from the fusion data; And A welding judgment unit that determines whether there is a defect in the welded part of the battery based on the welding area.

2. The welding detection device according to claim 1, wherein The three-dimensional image includes a plurality of pixels that map first height values within a first unit of measure, The two-dimensional image includes a plurality of pixels that map first luminance values within a second unit of measure.

3. The welding detection device according to claim 2, wherein The data processing unit performs a scaling process to reduce the number of pixels in the two-dimensional image and the three-dimensional image, and uses the scaled two-dimensional image and three-dimensional image to generate the fusion data.

4. The welding detection device according to claim 2 or 3, wherein The data processing unit uses a second luminance value obtained by normalizing the first luminance value and a second height value obtained by normalizing the first height value to generate the fusion data.

5. The welding detection device according to claim 4, wherein The data processing unit performs a weighted operation on the corresponding second luminance value and second height value to generate the fusion data.

6. The welding detection device according to claim 4, wherein Each of the second height value and the second luminance value is a value within a third unit of measure, and the third unit of measure is smaller than the first unit of measure and the second unit of measure.

7. The welding detection device according to claim 4, wherein The second height value is proportional to the ratio of a first difference and a second difference, the first difference being the difference between the first height value and a set lower limit value of the first height value, and the second difference being the difference between a set upper limit value of the first height value and the set lower limit value.

8. The welding detection device according to claim 7, wherein The second height value is the value obtained by multiplying the ratio by the maximum value within the third unit of measure, and the third unit of measure is smaller than the first unit of measure.

9. The welding detection device according to claim 7 or 8, wherein The set upper limit value is a value obtained by adding a first set value to a median value, the median value being the median value of the first height values of each of the plurality of pixels included in the three-dimensional image, The set lower limit value is a value obtained by subtracting a second set value from the median value.

10. The welding detection device according to claim 4, wherein The second luminance value is proportional to the ratio of the first luminance value and a value obtained by adding 1 to the maximum value of the second unit of measure.

11. The welding detection device according to claim 10, wherein The second luminance value is the value obtained by multiplying the ratio by the maximum value within the third unit of measure, and the third unit of measure is smaller than the second unit of measure.

12. The welding detection device according to claim 1, wherein The welding area includes a weld area, and the welding determination unit determines whether there is a defect in the welded part based on whether the weld area exists within a preset region of interest.

13. The welding detection device according to claim 12, wherein, the welding area includes a raised area, and the region of interest is the raised area.

14. The welding detection device according to claim 1, wherein, the welding area includes a weld area, and the welding determination unit determines whether there is a defect in the welded part based on the length of the weld area.

15. The welding detection device according to claim 14, wherein, the welding determination unit determines whether there is a defect in the welded part based on a first height value of the three-dimensional image corresponding to the weld area.

Images