Welding inspection apparatus and method

Abstract

A welding inspection apparatus according to one embodiment of the present invention may comprise at least one processor, and a memory for storing at least one instruction that is executed using the at least one processor. Here, the at least one instruction can include the instructions of: collecting a welding image in which a welding portion is captured, and welding state data measured during welding; inputting the welding image into a pre-trained first diagnosis model, and inputting the welding state data into a pre-trained second diagnosis model; and outputting a final diagnosis result for the welding portion on the basis of a first diagnosis result diagnosed by the first diagnosis model and / or a second diagnosis result diagnosed by the second diagnosis model.

Description

Welding inspection device and method

[0001] This application claims the benefit of Korean Patent Application No. 10-2024-0104524 filed with the Korean Intellectual Property Office on August 6, 2024, the entire contents of which are incorporated herein by reference.

[0002] The present invention relates to a data processing device and method, and more particularly, to a welding inspection device and method for diagnosing whether a welding part is defective using an artificial intelligence-based diagnostic model.

[0003] Secondary batteries are batteries that can be reused by charging even after discharge, and can be used as an energy source for small devices such as mobile phones, tablet PCs, and vacuum cleaners, and are also used as an energy source for medium and large devices such as automobiles and ESS (Energy Storage Systems) for smart grids.

[0004] Secondary batteries are batteries that can be reused by charging even after discharge, and can be used as an energy source for small devices such as mobile phones, tablet PCs, and vacuum cleaners, and are also used as medium- to large-scale energy sources such as personal mobility, automobiles, and ESS (Energy Storage Systems) for smart grids.

[0005] A cylindrical battery cell includes a battery can and a jelly-roll-shaped electrode assembly housed within the can. Here, an electrode tab electrically connected to the electrode assembly is welded to the inner surface of the battery can, and a weld mark may be visible at the welded area.

[0006] In general, to inspect welding quality, a destructive inspection method is used to determine whether there is a welding defect based on the tensile force when forcibly pulling off the electrode tab, or a non-destructive inspection method is used to determine whether there is a welding defect based on physical quantities such as welding current, welding voltage, and resistance measured during the welding process, or a non-destructive inspection method is used to determine whether there is a welding defect based on an image of the welding area.

[0007] In the case of the destructive inspection method, there is a problem that a full inspection is impossible and the inspection target must be discarded, and in the case of the non-destructive method, although the disadvantages of the destructive inspection do not appear, there is a problem that the inspection reliability is low due to a high over-inspection rate.

[0008] As a related prior art document, there is Korean Patent Publication No. 10-2023-0021474.

[0009] The purpose of the present invention to solve the above problems is to provide a welding inspection device that diagnoses whether a welding area is defective using an artificial intelligence-based diagnostic model.

[0010] Another object of the present invention to solve the above problems is to provide a welding inspection method using such a welding inspection device.

[0011] Another object of the present invention to solve the above problems is to provide a welding inspection system including such a welding inspection device.

[0012] A welding inspection device according to one embodiment of the present invention for achieving the above purpose may include at least one processor; and a memory storing at least one command executed through the at least one processor.

[0013] Here, the at least one command may include a command for collecting a welding image in which a welding portion is photographed, and welding status data measured during a welding process; a command for inputting the welding image into a pre-learned first diagnostic model, and a command for inputting the welding status data into a pre-learned second diagnostic model; and a command for outputting a final diagnostic result for the welding portion based on at least one of a first diagnostic result by the first diagnostic model and a second diagnostic result by the second diagnostic model.

[0014] The above welding image may include an image of the welding area of ​​the electrode tab after welding is completed.

[0015] The above welding status data may include one or more of welding current data, welding voltage data, dynamic resistance data, and output voltage data of the welding device.

[0016] The first diagnostic model may be defined so that, when the welding image is input, it outputs one of normal, over-welding, and weak welding as the first diagnostic result. In addition, the second diagnostic model may be defined so that, when the welding status data is input, it outputs one of normal, over-welding, and weak welding as the second diagnostic result.

[0017] The command for outputting the final diagnosis result for the above welding area may include a command for confirming the first diagnosis result by the first diagnosis model; and a command for outputting the final diagnosis result as overwelding, regardless of the second diagnosis result, when the first diagnosis result is overwelding.

[0018] The command for outputting the final diagnosis result for the above welding area may include a command for confirming the first diagnosis result by the first diagnosis model; and a command for deriving the final diagnosis result based on the first diagnosis result and the second diagnosis result when the first diagnosis result is normal or weak welding.

[0019] The command for deriving the final diagnosis result may include a command for applying a first weight predefined to the first diagnosis result and a second weight predefined to the second diagnosis result; and a command for determining the final diagnosis result based on a sum value of the first diagnosis result to which the first weight is reflected and the second diagnosis result to which the second weight is reflected.

[0020] The sum of the first weight and the second weight is defined as 1, and the second weight can be defined as 0.6 or more and 0.9 or less.

[0021]

[0022] A welding inspection method using a welding inspection device according to one embodiment of the present invention for achieving the above other objects may include the steps of collecting a welding image in which a welding portion is photographed and welding status data measured during a welding process; inputting the welding image into a pre-learned first diagnostic model and inputting the welding status data into a pre-learned second diagnostic model; and outputting a final diagnostic result for the welding portion based on at least one of a first diagnostic result by the first diagnostic model and a second diagnostic result by the second diagnostic model.

[0023] The above welding image may include an image of the welding area of ​​the electrode tab after welding is completed.

[0024] The above welding status data may include one or more of welding current data, welding voltage data, dynamic resistance data, and output voltage data of the welding device.

[0025] The first diagnostic model may be defined so that, when the welding image is input, it outputs one of normal, over-welding, and weak welding as the first diagnostic result. In addition, the second diagnostic model may be defined so that, when the welding status data is input, it outputs one of normal, over-welding, and weak welding as the second diagnostic result.

[0026] The step of outputting the final diagnosis result for the welding portion may include the step of confirming the first diagnosis result by the first diagnosis model; and, if the first diagnosis result is overwelding, the step of outputting the final diagnosis result as overwelding regardless of the second diagnosis result.

[0027] The step of outputting the final diagnosis result for the welding area may include a step of confirming the first diagnosis result by the first diagnosis model; and a step of deriving the final diagnosis result based on the first diagnosis result and the second diagnosis result when the first diagnosis result is normal or weak welding.

[0028] The step of deriving the final diagnosis result may include a step of applying a predefined first weight to the first diagnosis result and a predefined second weight to the second diagnosis result; and a step of determining the final diagnosis result based on a sum value of the first diagnosis result to which the first weight is reflected and the second diagnosis result to which the second weight is reflected.

[0029]

[0030] A welding inspection system according to one embodiment of the present invention for achieving the above-described further object may include a photographing device that generates a welding image for a welding portion; a welding monitoring device that generates welding status data during a welding process; and a welding inspection device that inspects the welding portion using the welding image and the welding status data.

[0031] Here, the welding inspection device inputs the welding image into a pre-learned first diagnostic model, inputs the welding status data into a pre-learned second diagnostic model, and can output a final diagnostic result for the welding portion based on at least one of a first diagnostic result by the first diagnostic model and a second diagnostic result by the second diagnostic model.

[0032] According to the above-described embodiment of the present invention, the diagnostic accuracy for the welding area can be further improved.

[0033] Figure 1 is a block diagram of a welding inspection system according to an embodiment of the present invention.

[0034] Figure 2 is a block diagram of a welding inspection device and a diagnostic model generation device according to an embodiment of the present invention.

[0035] Figure 3 is an example of a welding image by defect type according to an embodiment of the present invention.

[0036] Figure 4 is an example of welding status data according to an embodiment of the present invention.

[0037] Figure 5 is an operation flowchart of a welding inspection method according to an embodiment of the present invention.

[0038] Figure 6 is an operation flowchart of a welding inspection method according to another embodiment of the present invention.

[0039] FIG. 7 is an operational flowchart illustrating a method for determining a final diagnosis result according to an embodiment of the present invention.

[0040] Figure 8 is a block diagram of a welding inspection device according to an embodiment of the present invention.

[0041] 10: Welding target

[0042] 20: Welding device

[0043] 100: Welding Inspection System

[0044] 110: Filming device

[0045] 120: Welding monitoring device

[0046] 130: Welding inspection device

[0047] 200: Diagnostic model generation device

[0048] 800: Welding Inspection Device

[0049] The present invention is susceptible to various modifications and embodiments. Specific embodiments are illustrated in the drawings and described in detail in the detailed description. However, this is not intended to limit the present invention to specific embodiments, but rather to encompass all modifications, equivalents, and alternatives falling within the spirit and technical scope of the present invention. Throughout the description of each drawing, similar reference numerals have been used to designate similar components.

[0050] Terms such as "first," "second," "A," and "B" may be used to describe various components, but these components should not be limited by these terms. These terms are used solely to distinguish one component from another. For example, without departing from the scope of the present invention, the first component could be referred to as the "second component," and similarly, the second component could also be referred to as the "first component." The term "and / or" includes any combination of multiple related items listed or any one of multiple related items listed.

[0051] When a component is referred to as being "connected" or "connected" to another component, it should be understood that it may be directly connected or connected to that other component, but that there may be other components intervening. Conversely, when a component is referred to as being "directly connected" or "connected" to another component, it should be understood that there are no other components intervening.

[0052] The terminology used in this application is only used to describe specific embodiments and is not intended to limit the present invention. The singular expression includes the plural expression unless the context clearly indicates otherwise. In this application, it should be understood that the terms "comprise" or "have" indicate the presence of a feature, number, step, operation, component, part, or combination thereof described in the specification, but do not exclude in advance the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0053] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. Terms defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and shall not be interpreted in an idealized or overly formal sense unless explicitly defined herein.

[0054] Hereinafter, various embodiments of the present invention will be described in detail with reference to the attached drawings.

[0055]

[0056] Figure 1 is a block diagram of a welding inspection system according to an embodiment of the present invention.

[0057] The battery welding inspection system (100) according to an embodiment of the present invention can diagnose whether there is a defect in the welding area of ​​the welding target (10).

[0058] The welding target (10) may be a battery, and the welding area may refer to at least a portion of an electrode tab. However, the welding inspection system (100) according to the present invention can perform welding inspection on various objects other than batteries, and the scope of the present invention is not limited to the type, size, and shape of the welding target (10).

[0059] The welding device (20) is a device that welds a specific part of a welding target (10). For example, the welding device (20) can weld the lower inner surface of a battery can and the electrode tab of an electrode assembly.

[0060] The welding device (20) may include a first welding rod connected to the positive pole of the power supply and a second welding rod connected to the negative pole of the power supply. Here, the first welding rod may contact one side of the welding area, and the second welding rod may contact the other side of the welding area. At this time, when power is supplied by the power supply, the welding area is heated and welding can be performed.

[0061] The battery welding inspection system (100) may include a photographing device (110), a welding monitoring device (120), and a welding inspection device (130).

[0062] The photographing device (110) is a device that photographs a welding area to create a welding image. For example, the photographing device (110) may be a vision sensing device including an image sensor.

[0063] The photographing device (110) is placed on the upper or lower surface of the welding area and can photograph the welding area after the welding work on the welding target (10) is completed. For example, the photographing device (110) can photograph the welding area of ​​the electrode tab after the welding is completed to create a welding image.

[0064] The photographing device (110) is connected to the welding inspection device (130) through a network, and can transmit the generated welding image to the welding inspection device (130).

[0065] A welding monitoring device (120) is a device that measures at least one of the status of a welding device and the status of a welding area during a welding operation on a welding target (10).

[0066] The welding monitoring device (120) can generate welding status data measured during the welding process. Here, the welding status data can include one or more of welding current data, welding voltage data, dynamic resistance data, and output voltage data of the welding device.

[0067] The welding current data may refer to time series data indicating the current value applied through the first welding electrode and the second welding electrode, and the welding voltage data may refer to time series data indicating the voltage value applied through the first welding electrode and the second welding electrode. In addition, the dynamic resistance data may refer to time series data indicating the resistance value of the welding part that changes during the welding process, and the output voltage data of the welding device may refer to time series data indicating the voltage value output from the power supply unit of the welding device.

[0068] The welding monitoring device (120) is connected to the welding inspection device (130) through a network, and can transmit the generated welding status data to the welding inspection device (130).

[0069] The welding inspection device (130) is a device that diagnoses whether there is a defect in the welding area. Here, the welding inspection device (130) can diagnose whether there is a defect in the welding area using the welding image generated by the photographing device (110) and the welding status data generated by the welding monitoring device (120).

[0070] The welding inspection device (130) can determine whether a welded area is defective, and if so, the type of defect (over-welding or under-welding). For example, the welding inspection device (130) can output any one of normal, over-welding defect, and under-welding defect as a diagnosis result for the welded area.

[0071] The welding inspection device (130) can diagnose whether there is a defect in the welding area using a predefined machine learning-based diagnostic model. Here, the welding inspection device (130) can diagnose whether there is a defect in the welding area using a first diagnostic model (131) that has been pre-trained using welding images as learning data, and a second diagnostic model (132) that has been pre-trained using welding status data as learning data.

[0072] Specifically, the welding inspection device (130) can input a welding image into a first diagnostic model (131) to check the diagnostic result output from the first diagnostic model (131), and input welding status data into a second diagnostic model (132) to check the diagnostic result output from the second diagnostic model (132). Here, the first diagnostic model (131) and the second diagnostic model (132) can output any one of normal, over-welding defect, and under-welding defect as the diagnostic result.

[0073] Thereafter, the welding inspection device (130) can output a final diagnosis result for the welding area based on at least one of the first diagnosis result and the second diagnosis result.

[0074] For example, if the first diagnosis result is over-welding, the welding inspection device (130) may output the final diagnosis result as over-welding, regardless of the second diagnosis result. As another example, if the first diagnosis result is normal or weak welding, the welding inspection device (130) may output the second diagnosis result as the final diagnosis result.

[0075] The diagnostic model generation device (200) is a device that generates a machine learning-based diagnostic model for welding inspection.

[0076] The diagnostic model generation device (200) can generate a first diagnostic model (131) and a second diagnostic model (132) and provide them to the welding inspection device (130).

[0077]

[0078] Figure 2 is a block diagram of a welding inspection device and a diagnostic model generation device according to an embodiment of the present invention.

[0079] Referring to FIG. 2, the diagnostic model generation device (200) can generate a first diagnostic model (231) and a second diagnostic model (232) and provide them to the welding inspection device (130).

[0080] The diagnostic model generation device (200) can train the first diagnostic model (231) and the second diagnostic model (232) using learning data stored in the storage device (241, 242).

[0081] The first diagnostic model (231) can be trained using welding images as learning data.

[0082] Figure 3 is an example of a welding image according to a defect type according to an embodiment of the present invention. Referring to Figure 3, if the welding area is normal, as in (A), the welding area appears relatively clean, and slight stain marks may appear in areas other than the welding area. On the other hand, if the welding area is defective due to over-welding, as in (B), the shape of the welding area is not clearly visible, and the welding image may appear in an overall dark color. Furthermore, if the welding area is defective due to under-welding, as in (C), the welding area appears very clean, and the welding image may appear in an overall bright color.

[0083] The first diagnostic model (231) can be predefined based on a neural network-based image classification algorithm, and the diagnostic model generation device (200) can train the first diagnostic model (231) using labeled welding images (e.g., A, B, C in FIG. 3) as learning data.

[0084] The first diagnostic model (231) can output a first diagnostic result including any one of normal, over-welding defect, and weak welding defect. Here, the first diagnostic result can further include a probability value for each of the result items. For example, the first diagnostic model (231) can calculate a probability value (e.g., normal 3%, over-welding 2%, weak welding 95%) for each of the result items (normal, over-welding, weak welding defect) for an input welding image, and determine and output a result item (e.g., weak welding) having the highest probability value among them.

[0085] When the learning of the first diagnostic model (231) is completed, the diagnostic model generation device (200) can provide the learned first diagnostic model (231) to the welding inspection device (130).

[0086] The second diagnostic model (232) can be trained using welding status data as learning data.

[0087] Fig. 4 is an example of welding status data according to an embodiment of the present invention. Referring to Fig. 4, the welding status data may include one or more of welding current data, welding voltage data, dynamic resistance data, and output voltage data of a welding device.

[0088] Here, the welding current data may include time series data representing the current value applied through the first welding electrode and the second welding electrode, as in (A), and the welding voltage data may include time series data representing the voltage value applied through the first welding electrode and the second welding electrode, as in (B). In addition, the dynamic resistance data may include time series data representing the resistance value of the welding part that changes during the welding process, as in (C), and the output voltage data of the welding device may include time series data representing the voltage value output from the power supply unit of the welding device, as in (D).

[0089] The second diagnostic model (232) can be predefined based on a machine learning-based classification algorithm (e.g., XGB; Extreme Gradient Boosting, etc.), and the diagnostic model generation device (200) can train the second diagnostic model (232) using welding status data as learning data.

[0090] In the embodiment, the second diagnostic model (232) can preprocess the welding status data according to a predefined preprocessing process, and train the second diagnostic model (232) using the preprocessed welding status data as learning data. For example, the diagnostic model generation device (200) can calculate the average welding current value, the maximum welding current value, the average welding voltage value, the initial resistance value (the resistance value at the time point t1), the final resistance value (the resistance value at the time point t2), and the average output voltage value of the welding device in the sections t1 to t2 of FIG. 3. Thereafter, the diagnostic model generation device (200) can train the second diagnostic model (232) using the welding status data including the calculated values ​​as learning data.

[0091] The second diagnostic model (232) can output a second diagnostic result including any one of normal, over-welding defect, and weak welding defect. Here, the second diagnostic result can further include a probability value for each of the result items. For example, the second diagnostic model (232) can calculate a probability value (e.g., normal 3%, over-welding 2%, weak welding 95%) for each of the result items (normal, over-welding, weak welding defect) for an input welding image, and determine and output a result item (e.g., weak welding) having the highest probability value among them.

[0092] When the learning of the second diagnostic model (232) is completed, the diagnostic model generation device (200) can provide the learned second diagnostic model (232) to the welding inspection device (130).

[0093] The welding inspection device (130) can diagnose whether there is a defect in the welding area and determine the type of defect using the first diagnostic model (131) and the second diagnostic model (132).

[0094] Specifically, the welding inspection device (130) can output a final diagnosis result based on at least one of a first diagnosis result output from the diagnosis model (131) by inputting a welding image generated by the photographing device (110) into the first diagnosis model (131) and a second diagnosis result output from the second diagnosis model (132) by inputting welding status data generated by the welding monitoring device (120) into the second diagnosis model (132).

[0095]

[0096] Figure 5 is an operation flowchart of a welding inspection method according to an embodiment of the present invention.

[0097] The welding inspection device can collect welding images of the welding area and welding status data measured during the welding process (S510). Here, the welding inspection device can receive welding images from the photographing device and welding status data from the welding monitoring device.

[0098] The welding image may include an image of the welded area of ​​the electrode tab after the welding operation is completed.

[0099] The welding status data may include one or more of welding current data, welding voltage data, dynamic resistance data, and output voltage data of the welding device.

[0100] The welding inspection device can input a welding image into a first diagnostic model, check the first diagnostic result output from the first diagnostic model, and input welding status data into a second diagnostic model, and check the diagnostic result output from the second diagnostic model (S520).

[0101] The first diagnostic model may be defined to output one of normal, overwelding, and weak welding as the first diagnostic result when a welding image is input, and the second diagnostic model may be defined to output one of normal, overwelding, and weak welding as the second diagnostic result when welding status data is input. Here, the welding inspection device may receive the first diagnostic model and the second diagnostic model in a state where pre-learning has been completed from the diagnostic model generation device.

[0102] The welding inspection device can output a final diagnosis result for the welding area based on at least one of the first diagnosis result and the second diagnosis result (S530).

[0103] The first diagnostic model, which is pre-trained using welding images as learning data, has a high detection accuracy for overwelding defects. Referring to Figure 3, the welding image classified as overwelding defect (B) appears in an overall dark color, clearly distinguishing it from the welding images classified as normal (A) and weak welding defect (C). Therefore, the first diagnostic model has a relatively high detection accuracy for overwelding defects compared to the second diagnostic model.

[0104] Considering the characteristics of this first diagnostic model, the welding inspection device can output the final diagnostic result as overwelding, regardless of the second diagnostic result, if the first diagnostic result is overwelding. If the first diagnostic result is normal or underwelding, the welding inspection device can output the second diagnostic result as the final diagnostic result.

[0105]

[0106] FIG. 6 is an operation flowchart of a welding inspection method according to another embodiment of the present invention, and FIG. 7 is an operation flowchart for explaining a method for determining a final diagnosis result according to an embodiment of the present invention.

[0107] The welding inspection device can collect welding images of the welding area and welding status data measured during the welding process (S610). Here, the welding inspection device can receive welding images from the photographing device and welding status data from the welding monitoring device.

[0108] The welding inspection device can input a welding image into the first diagnostic model (131) and check the first diagnostic result output from the first diagnostic model (S620).

[0109] If the first diagnosis result is overwelding (Y in S630), the welding inspection device can output the final diagnosis result as overwelding (S660). That is, considering that the first diagnosis model (131) has a high overwelding detection accuracy, if overwelding is determined by the first diagnosis model (131), as illustrated in FIG. 7, the welding inspection device can output the final diagnosis result as overwelding.

[0110] If the first diagnosis result is normal or weak welding (N of S630), the welding inspection device can input welding status data into the second diagnosis model (132) and check the second diagnosis result output from the second diagnosis model (132) (S640).

[0111] Thereafter, the welding inspection device can derive a final diagnosis result based on predefined weights for each of the first diagnosis result and the second diagnosis result.

[0112] More specifically, the welding inspection device may apply a predefined first weight to the first diagnosis result and a predefined second weight to the second diagnosis result (S650). Here, the sum of the first and second weights may be defined as 1, and the second weight may be defined as 0.6 or greater and 0.9 or less.

[0113] Thereafter, the welding inspection device can determine and output a final diagnosis result based on the sum of the first diagnosis result reflecting the first weight and the second diagnosis result reflecting the second weight (S660).

[0114] For example, if the probability value for each of the result items is calculated as [Normal 3%, Overwelding 2%, Weak Welding 95%] by the first diagnostic model (131), the welding inspection device can reflect the preset first weight (W1) in the first diagnostic result (Normal 3*W1%, Overwelding 2*W1%, Weak Welding 95*W1%), as illustrated in FIG. 7. In addition, if the probability value for each of the result items is calculated as [Normal 85%, Overwelding 2%, Weak Welding 13%] by the second diagnostic model (132), the welding inspection device can reflect the preset second weight (W2) in the second diagnostic result (Normal 85*W2%, Overwelding 2*W2%, Weak Welding 13*W2%), as illustrated in FIG. 7. Thereafter, the welding inspection device can add up the first diagnosis result reflecting the first weight and the second diagnosis result reflecting the second weight (normal 3*W1+85*W2%, overwelding 2*W1+2*W2%, underwelding 95*W1+13*W2%) and determine the result item (argmax) showing the highest probability value as the final diagnosis result.

[0115]

[0116] In order to derive the optimal range of weights according to the embodiment of the present invention, a comparative test was performed, and the results are as follows.

[0117] WeightAccuracyW1W2TotalNormalOverWeldWeakWeld100.9780.9940.8390.9330.90.10.9780.9940.8440.9330.80.20.980.9950.8530.9330.70.30.9810.9950.8680.9330.60.40.9820.9950.8770.9 320.50.50.9840.9950.8960.9320.40.60.9850.9940.9110.9320.30.70.9870.9930.9450.9330.20.80.9880.9920.9620.9330.10.90.9890.9920.9540.933010.9890.9940.9350.932

[0118] Referring to Table 1 above, when the first weight (W1) applied to the first diagnosis result is defined as 0.1 or more and 0.3 or less, and the second weight (W2) reflected in the second diagnosis result is defined as 0.7 or more and 0.9 or less, it can be confirmed that the overall detection accuracy and the detection accuracy for overwelding are high.

[0119]

[0120] Figure 8 is a block diagram of a welding inspection device according to an embodiment of the present invention.

[0121] A welding inspection device (800) according to an embodiment of the present invention can be linked with a photographing device that generates a welding image for a welding area; and a welding monitoring device that generates welding status data during a welding process.

[0122] A welding inspection device (800) may include at least one processor (810), a memory (820) storing at least one command executed through the processor, and a transmission / reception device (830) connected to a network to perform communication.

[0123] The at least one command may include a command for collecting a welding image in which a welding portion is photographed, and welding status data measured during a welding process; a command for inputting the welding image into a pre-learned first diagnostic model, and a command for inputting the welding status data into a pre-learned second diagnostic model; and a command for outputting a final diagnostic result for the welding portion based on at least one of a first diagnostic result by the first diagnostic model and a second diagnostic result by the second diagnostic model.

[0124] The above welding image may include an image of the welding area of ​​the electrode tab after welding is completed.

[0125] The above welding status data may include one or more of welding current data, welding voltage data, dynamic resistance data, and output voltage data of the welding device.

[0126] The first diagnostic model may be defined so that, when the welding image is input, it outputs one of normal, over-welding, and weak welding as the first diagnostic result. In addition, the second diagnostic model may be defined so that, when the welding status data is input, it outputs one of normal, over-welding, and weak welding as the second diagnostic result.

[0127] The command for outputting the final diagnosis result for the above welding area may include a command for confirming the first diagnosis result by the first diagnosis model; and a command for outputting the final diagnosis result as overwelding, regardless of the second diagnosis result, when the first diagnosis result is overwelding.

[0128] The command for outputting the final diagnosis result for the above welding area may include a command for confirming the first diagnosis result by the first diagnosis model; and a command for deriving the final diagnosis result based on the first diagnosis result and the second diagnosis result when the first diagnosis result is normal or weak welding.

[0129] The command for deriving the final diagnosis result may include a command for applying a first weight predefined to the first diagnosis result and a second weight predefined to the second diagnosis result; and a command for determining the final diagnosis result based on a sum value of the first diagnosis result to which the first weight is reflected and the second diagnosis result to which the second weight is reflected.

[0130] The sum of the first weight and the second weight is defined as 1, and the second weight can be defined as 0.6 or more and 0.9 or less.

[0131] The welding inspection device (800) may further include an input interface device (840), an output interface device (850), a storage device (860), etc. Each component included in the welding inspection device (800) may be connected by a bus (870) to communicate with each other. The storage device (860) may store a diagnostic model, a defect detection model, and an anomaly detection model.

[0132] Here, the processor (810) may refer to a central processing unit (CPU), a graphics processing unit (GPU), or a dedicated processor on which methods according to embodiments of the present invention are performed. In addition, the memory may be configured with at least one of a volatile / transitory storage medium and a non-volatile / non-transitory storage medium. For example, the memory may be configured with at least one of a read-only memory (ROM) and a random access memory (RAM), and may include an Electrically Erasable Programmable Read-only Memory (EEPROM).

[0133]

[0134] The operations of the method according to an embodiment of the present invention can be implemented as a computer-readable program or code on a computer-readable recording medium. The computer-readable recording medium may include any type of recording device that stores data readable by a computer system. The computer-readable recording medium may also be distributed across network-connected computer systems, allowing the computer-readable program or code to be stored and executed in a distributed manner.

[0135] The operation of the method according to an embodiment of the present invention may be implemented in various forms related to a program, such as a computer program or code itself or a computer program product.

[0136] Additionally, the computer-readable recording medium may include one or more of a volatile / transitory recording medium and a non-volatile / non-transitory recording medium.

[0137] A computer-readable recording medium may include hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, and flash memory, and may include, for example, various types of servers located on a network. Program instructions may include not only machine language codes, such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like.

[0138] While some aspects of the present invention have been described in the context of a device, they may also represent a description of a corresponding method, wherein a block or device corresponds to a method step or a feature of a method step. Similarly, aspects described in the context of a method may also be described as a corresponding block or item or a feature of a corresponding device. Some or all of the method steps may be performed by (or using) a hardware device, such as, for example, a microprocessor, a programmable computer, or an electronic circuit. In some embodiments, one or more of the most significant method steps may be performed by such a device.

[0139] Although the present invention has been described with reference to preferred embodiments thereof, it will be understood by those skilled in the art that various modifications and changes may be made to the present invention without departing from the spirit and scope of the present invention as set forth in the claims below.

Claims

1. At least one processor; and A memory that stores at least one instruction to be executed through at least one processor, At least one of the above commands, A command to collect welding images in which the welding area is photographed, and welding condition data measured during the welding process; A command for inputting the above welding image into a pre-learned first diagnostic model and inputting the above welding status data into a pre-learned second diagnostic model; and A welding inspection device, comprising a command for outputting a final diagnosis result for the welding area based on at least one of a first diagnosis result by the first diagnosis model and a second diagnosis result by the second diagnosis model.

2. In claim 1, The above welding image is, A welding inspection device including an image of the welded area of ​​the electrode tab after welding is completed.

3. In claim 1, The above welding status data is, A welding inspection device comprising at least one of welding current data, welding voltage data, dynamic resistance data, and output voltage data of a welding device.

4. In claim 1, The above first diagnostic model is defined to output one of normal, overwelding, and underwelding as the first diagnostic result when the welding image is input, A welding inspection device, wherein the second diagnostic model is defined to output one of normal, over-welding, and under-welding as a second diagnostic result when the welding status data is input.

5. In claim 4, The command to output the final diagnosis results for the above welding area is: A command to confirm the first diagnosis result by the first diagnosis model; and A welding inspection device, comprising a command to output the final diagnosis result as overwelding, regardless of the second diagnosis result, when the first diagnosis result is overwelding.

6. In claim 4, The command to output the final diagnosis results for the above welding area is: A command to confirm the first diagnosis result by the first diagnosis model; and A welding inspection device, comprising a command for deriving the final diagnosis result based on the first diagnosis result and the second diagnosis result when the first diagnosis result is normal or weak welding.

7. In claim 6, The command to derive the final diagnosis result is: A command for applying a first weight predefined to the first diagnosis result and a second weight predefined to the second diagnosis result; and A welding inspection device, comprising a command for determining the final diagnosis result based on the sum of the first diagnosis result in which the first weight is reflected and the second diagnosis result in which the second weight is reflected.

8. In claim 7, The sum of the first weight and the second weight is defined as 1, The above second weighting is defined as 0.6 or more and 0.9 or less, and the welding inspection device.

9. A welding inspection method using a welding inspection device, A step of collecting a welding image in which the welding part is photographed, and welding condition data measured during the welding process; A step of inputting the above welding image into a pre-learned first diagnostic model and inputting the above welding status data into a pre-learned second diagnostic model; and A welding inspection method, comprising a step of outputting a final diagnosis result for the welding area based on at least one of a first diagnosis result by the first diagnosis model and a second diagnosis result by the second diagnosis model.

10. In claim 9, The above welding image is, A welding inspection method comprising an image of the welded area of ​​the electrode tab after welding is completed.

11. In claim 9, The above welding status data is, A welding inspection method comprising at least one of welding current data, welding voltage data, dynamic resistance data, and output voltage data of a welding device.

12. In claim 9, The above first diagnostic model is defined to output one of normal, overwelding, and underwelding as the first diagnostic result when the welding image is input, A welding inspection method, wherein the second diagnostic model is defined to output one of normal, over-welding, and under-welding as a second diagnostic result when the welding status data is input.

13. In claim 12, The step of outputting the final diagnosis result for the above welding area is: A step of confirming the first diagnosis result by the first diagnosis model; and A welding inspection method, comprising a step of outputting the final diagnosis result as overwelding, regardless of the second diagnosis result, when the first diagnosis result is overwelding.

14. In claim 12, The step of outputting the final diagnosis result for the above welding area is: A step of confirming the first diagnosis result by the first diagnosis model; and A welding inspection method, comprising a step of deriving the final diagnosis result based on the first diagnosis result and the second diagnosis result when the first diagnosis result is normal or weak welding.

15. In claim 14, The steps for deriving the final diagnosis results are: A step of applying a predefined first weight to the first diagnosis result and applying a predefined second weight to the second diagnosis result; and A welding inspection method, comprising a step of determining the final diagnosis result based on the sum of the first diagnosis result in which the first weight is reflected and the second diagnosis result in which the second weight is reflected.

16. A photographing device that creates a welding image of the welding area; A welding monitoring device that generates welding status data during the welding process; and A welding inspection device is included that inspects the welding area using the welding image and the welding status data. The above welding inspection device, A welding inspection system that inputs the welding image into a pre-learned first diagnostic model, inputs the welding status data into a pre-learned second diagnostic model, and outputs a final diagnostic result for the welding area based on at least one of a first diagnostic result by the first diagnostic model and a second diagnostic result by the second diagnostic model.

Images