KR20210065677A - Method of detecting defect in welded joint using thermal imaging camera - Google Patents

Abstract

Disclosed is a method for detecting a defect in a welding part by using a thermal imaging camera. A welding part defect detection method according to an embodiment of the present invention is a method for detecting defects in a welding part of two steel members, the method comprising: a step of acquiring a plurality of temperature values corresponding to a plurality of points on one surface of the welding part to be inspected; a step of calculating an index value by calculating a standard deviation of the plurality of temperature values; and a step of comparing the index value with a preset defect range to determine whether a defect exists in the welding part to be inspected.

Description

A method for detecting welding defects using a thermal imaging camera {METHOD OF DETECTING DEFECT IN WELDED JOINT USING THERMAL IMAGING CAMERA}

The present invention relates to a method for detecting a defect in a weld, and more particularly, to a method for detecting a defect in a weld using a thermal imaging camera.

With the development of technology and the improvement of living standards, the size of the construction market has been continuously increasing. The size of the domestic construction market is about 120 trillion won, representing the size of the world's top 10 market. In particular, steel structure is being used according to the demand of consumers for the introduction of new technology and beautiful appearance, and it is used in buildings of various sizes, from general houses to high-rise buildings.

On the other hand, as the standard of living increases, the importance of safety of buildings is also being emphasized. It is expected that interest and market size for non-destructive testing safety diagnosis will increase further in line with the promotion of the safety diagnosis industry among government policies related to safety innovation.

The main part of the safety inspection of steel structures includes determining the presence or absence of defects, and in particular, the detection of defects in the welding state at the junction of steel members is essential. Such defect inspection can be divided into visual inspection and non-destructive inspection. Visual inspection is a method of visually measuring defects in welds such as overlap, undercut, and cracks, and non-destructive inspection is a method of measuring defects in welds using ultrasonic inspection and magnetic particle inspection.

In the case of inspection for defects in welds by manpower, the majority of measurement methods relying on the naked eye are limited to measuring defects and deformations in local areas. In order to visually measure the defects of the weld, it is inconvenient to carry additional equipment such as measuring equipment accessories and coating agents, and in the case of high-rise buildings, the inspection work itself is very difficult and very dangerous, aside from the enormous manpower and cost. can do.

As a type of non-destructive testing, ultrasonic probe inspection involves the process of mounting ultrasonic probes on both sides of an object to be inspected. That is, a device for emitting an ultrasonic wave is installed on one side of the object to be examined, and a device for receiving an ultrasonic wave is installed on the other side. Magnetic particle inspection is a method of detecting defects by magnetizing the object to be inspected and checking the part where the magnetic force is strong by spraying magnetic powder that responds to the magnetic field on the surface. As such, ultrasonic inspection and magnetic particle inspection also require very cumbersome preliminary preparation, and the larger the structure to be inspected, the greater the manpower and cost required.

As an alternative to this, a thermal imaging inspection method has been proposed. The thermal imaging inspection method includes a passive inspection method that detects the intrinsic infrared energy emitted by all objects and an active inspection method that uses external energy incident on the object to detect radiation energy that varies depending on the internal condition of the object. However, the thermal imaging inspection method according to the prior art only determines whether a crack has occurred in the structure and cannot measure the degree of crack, that is, the depth of the crack. Since the degree of cracking is related to future structural changes and predictions, not only understanding the existence of cracks but also a technique for understanding the extent of cracks is required.

Korean Patent Registration No. 10-0553124

One aspect of the present invention is to provide a welding part defect detection method capable of determining the type and degree of a defect in addition to determining the presence or absence of a defect in a steel frame welding part using a thermal imaging camera.

Another aspect of the present invention is to provide a welding part defect detection method capable of detecting defects in a steel frame welding part in a simple manner while providing high reliability and high accuracy by comprehensively considering the environmental factors of the inspection object.

According to an embodiment of the present invention, there is provided a method for detecting a defect in the weld of two steel members. A welding part defect detection method according to an embodiment of the present invention includes: acquiring a plurality of temperature values corresponding to a plurality of points on one surface of a welding part to be inspected; calculating an index value by calculating a standard deviation of the plurality of temperature values; and comparing the index value with a preset defect range to determine whether a defect exists in the inspection target welding part.

The acquiring of the plurality of temperature values may include photographing the welding part to be inspected with a thermal imaging camera, and the plurality of temperature values may correspond to a plurality of pixel values in the image photographed by the thermal imaging camera. have.

The determining whether the defect exists may include comparing the index value with a plurality of defect ranges and determining the type of the defect based on the defect range to which the index value belongs.

The acquiring the plurality of temperature values may include measuring one or more external factors while acquiring the plurality of temperature values, and the defect range may be set differently according to values of the external factors.

Here, the external factor may include at least one of an external temperature, humidity, wind speed, solar radiation, and a surface temperature of a member adjacent to the welding part to be inspected when the plurality of temperature values are acquired.

The welding part defect detection method according to an embodiment of the present invention may further include the step of setting the defect range before the step of acquiring the plurality of temperature values.

Here, the step of setting the defect range may include collecting standard deviation examples calculated from a plurality of temperature values of the weld zone where the defect exists for each of one or more defect types.

In this case, the step of setting the defect range may include building the collected standard deviation cases into a database and inputting the database into a deep learning algorithm.

In addition, the standard deviation case is the width of the weld, the temperature of the top, the thickness of the member to be welded, the type of defect, the depth of the defect, the number of temperature values used in the calculation of the standard deviation, outdoor temperature, humidity, wind speed, insolation, It may be classified and collected according to one or more criteria of the surface temperature of the member adjacent to the welded portion and the angle between the members on both sides of the welded portion.

The acquiring of the plurality of temperature values may include: acquiring the plurality of temperature values in a first time period and simultaneously measuring one or more external factors; and obtaining a plurality of temperature values corresponding to the plurality of points in a second time period in which the external factor is changed, wherein the index value is the first calculated value from the temperature values obtained in the first time period. It may be calculated based on one standard deviation and a rate of change of the second standard deviation calculated from the temperature value obtained in the second time period.

In this case, the one or more external factors may include a surface temperature of a member adjacent to the welding part to be inspected, and obtaining a plurality of temperature values in the second time period may include applying heat to the inspection object. can

According to an embodiment of the present invention, there is provided a welding part defect detection method capable of determining the type and degree of a defect in addition to determining only the presence or absence of a defect in a steel frame welded part using a thermal imaging camera.

In addition, according to an embodiment of the present invention, there is provided a welding part defect detection method capable of detecting defects in a steel frame welding part in a simple way while providing high reliability and high accuracy by comprehensively considering the environmental factors of the inspection object. .

As such, by more accurately detecting defects in the welding part while considering various environmental factors, it is possible to prevent safety accidents that may occur in the future and to reduce the maintenance cost of the structure.

1 is a flowchart schematically illustrating a welding defect detection method according to an embodiment of the present invention.
2 is a conceptual diagram schematically illustrating each step of a method for detecting a welding defect according to an embodiment of the present invention.
3 is a conceptual diagram schematically illustrating a deep learning algorithm that can be used in a method for detecting a welding defect according to an embodiment of the present invention.
4 is a view showing a test object used in the step of collecting standard deviation cases of the welding defect detection method according to an embodiment of the present invention.
5 is a conceptual diagram illustrating the types of defects classified in a method for detecting a welding defect according to an embodiment of the present invention.
6 is a graph illustrating examples of standard deviations collected according to a welding defect detection method according to an embodiment of the present invention.
7 and 8 are diagrams illustrating a process of acquiring a plurality of temperature values by a method for detecting defects in a welding part according to an embodiment of the present invention.

The above-described objects, features and advantages will be described below in detail with reference to the accompanying drawings, and accordingly, those skilled in the art to which the present invention pertains will be able to easily implement the technical idea of the present invention. In describing the present invention, if it is determined that a detailed description of a known technology related to the present invention may unnecessarily obscure the gist of the present invention, the detailed description will be omitted.

Hereinafter, some embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the drawings, the same reference numerals are used to indicate the same or similar components.

1 is a flowchart schematically illustrating a method for detecting a weld defect according to an embodiment of the present invention, and FIG. 2 is a conceptual diagram schematically illustrating each step of the method for detecting a weld defect according to an embodiment of the present invention.

1 and 2, the welding defect detection method according to an embodiment of the present invention includes a step of largely setting a defect range (S100), a plurality of temperature values corresponding to a plurality of points on one surface of a welding part to be inspected. Acquiring (S200), calculating an index value by calculating the standard deviation of a plurality of temperature values (S300), and comparing the index value with a defect range to determine whether a defect exists in the welding part to be inspected (S400) ) may be included.

As described above, the thermal imaging inspection method provides many advantages in inspecting defects of structures. Compared to methods such as visual inspection, ultrasonic inspection, and magnetic particle inspection, the thermal imaging inspection method can be performed much more conveniently. In particular, when the thermal imaging camera is mounted on an unmanned aerial vehicle (UAV), it is difficult for the operator to directly access it. Or it becomes possible to inspect dangerous locations. However, the existing thermal imaging inspection method captures a portion with abruptly different temperatures in an image taken with a thermal imaging camera, and judges only the presence or absence of defects without understanding the degree. On the other hand, the welding part defect detection method according to an embodiment of the present invention can quantitatively determine not only the presence or absence of a defect but also the type and degree of the defect by using the standard deviation value of the obtained temperature data.

The surface temperature appears relatively uniform in the portion without defects in the weld (hereinafter referred to as the normal portion). On the other hand, the surface temperature is non-uniform in the defective portion of the weld (hereinafter referred to as defective portion). In particular, the non-uniformity of the temperature value is different for each type of defect, and it may also be affected by environmental factors surrounding the temperature value acquisition. For example, the outdoor temperature, humidity, wind speed, and solar radiation when the temperature value is obtained affect the entire structure, and thus may also affect radiant energy radiated from the top and defect of the weld. In addition, the surface temperature of the defective part may vary depending on other heat sources in the structure, the size of the welded steel member, and the size of the welded part itself. In one embodiment of the present invention, after calculating the index value of the inspection target by using the phenomenon in which the surface temperature is non-uniformly distributed in the defective part of the welded portion of the steel member as described above, the welding part defect can be detected based on this.

To this end, the welding part defect detection method according to an embodiment of the present invention may include first setting a defect range ( S100 ). This is a step of providing a standard for comparing the index values of the inspection target.

According to an embodiment of the present invention, the step of setting the defect range ( S100 ) may include the step of collecting and analyzing a plurality of standard deviation cases ( S110 ). Here, the standard deviation case may mean a standard deviation value calculated for each sample set after measuring the surface temperature value of a predetermined sample size for the normal part and the defective part of the welding part under various circumstances. As will be described later, according to an embodiment, a plurality of standard deviation values may be used as one standard deviation example.

The defect range is for determining whether the inspection object is identified as having a defect through the index value of the inspection object, and may be expressed, for example, as a range of index values corresponding to the defect. In an embodiment in which one standard deviation value itself is used as an index value, the defect range may be, for example, a lower limit or a combination of a lower limit and an upper limit for defining a predetermined range. For example, when the index value of the inspection target is 11 in a state where the lower limit of the defect range is set to 10, it may be determined that the inspection target has a defect.

The collected standard deviation cases can be classified according to the type of defect that the welding defect detection method intends to detect. For example, FIG. 5 shows five types of defects that may occur in welds of a steel frame, namely, overlap, crack, undercut, crater, and fish eye. Assuming that all of the five defects are considered according to an embodiment of the present invention, each standard deviation case to be collected will be classified into six types and collected according to which type of defect or normal part it corresponds to. can Of course, the number of types of defects may be reduced or increased according to embodiments.

According to an embodiment of the present invention, the collected standard deviation cases may be built into a database. Each standard deviation case input to the database may include the standard deviation value of the temperature value in the designated area and values corresponding to factors to be considered. For example, each standard deviation case input to the database may include measurements of external factors at the time the temperature value was measured. These external factors include ambient temperature, humidity, wind speed, insolation, and proximity to the measured weld. It may include one or more of the surface temperature of the member. Measurement values related to other factors may be included as needed for more accurate detection and determination. For example, each standard deviation case entered in the database is the width of the weld at the time of measuring the temperature value, the temperature of the top, the thickness of the welded member, the type of defect, the depth of the defect, and the temperature value used to calculate the standard deviation. It may include a number and the like. In addition, since most of the thermal imaging cameras are equipped with a visible light camera function, the external image of the welding part photographed by the thermal imaging camera can also be collected and the visual inspection can be performed simultaneously. As such, the standard deviation cases stored in the database can be classified according to each item.

Such a database may be input to a predetermined machine learning algorithm, for example, an artificial neural network or a deep learning algorithm. 3 illustrates the concept of a deep learning algorithm as an example of the aforementioned machine learning algorithm. Each set of standard deviation examples stored in the database may include values such as the outside temperature at the time, the temperature of the top, and the surface temperature of adjacent members, in addition to the temperature values of the top or the defective part of the weld as described above, and a machine learning algorithm can derive the characteristics of each defect type to be discriminated based on the above data.

In this way, in order to collect standard deviation cases and build a database as needed, the process of collecting standard deviation cases may include the process of manufacturing one or more mimics and performing experiments on the mimics.

4 shows a test object 50 that can be used to conduct such an experiment. In the test object 50 of FIG. 4 , a plurality of iron members 1 to 4 are welded, and welding portions 10 , 20 , and 30 are formed between each member. The intermediate welding part 20 of the test object 50 of FIG. 4 was carefully welded so that there were no defects in the welding, while the left welded part 10 and the right welded part 30 of the test object 50 were intentionally defective in the welding part. defects were formed. 5 shows the shape of the defect intentionally implemented in the welding part 10 . In addition, in the example shown in FIG. 4 , the welded portion 10 and the member 3 and the member 4 formed when the member 1 and the member 2 are welded in parallel (eg, butt welding) are The weld 30 formed in the case of welding at right angles (eg, fillet welding) is shown. Of course, in another embodiment, the member on both sides of the weld may be implemented by making a test specimen including more various angles in addition to 0 degrees and 90 degrees to implement a weld without defects and a weld with defects intentionally formed.

Temperature values of a plurality of points on each of the welds 10 , 20 , and 30 in the manufactured test object 50 may be measured. This can also be done by thermal imaging. Depending on the analysis technique used and the accuracy sought, in addition to the temperature value for calculating the temperature deviation in each defect type, the outside air temperature, humidity, wind speed, solar radiation at the time of shooting, the surface temperature of the member adjacent to the welded part, the angle between the members on both sides of the welded part, One or more of the following factors may be measured together: the width of the weld, the temperature of the top, the thickness of the member to be welded, the type of defect, the depth of the defect, and the number of temperature values used to calculate the standard deviation. Here, separate equipment such as a hygrometer and anemometer may be required to measure some factors such as humidity and wind speed, but there are also factors that can be simultaneously acquired by a thermal imaging camera without separate equipment, such as the surface temperature of adjacent members. In particular, since the surface temperature of the member adjacent to the weld is particularly highly correlated with the temperature values of the normal and defective parts of the weld, the accuracy of the inspection can be increased without sacrificing convenience by using the above factor.

The defective portion of the weld may have a non-uniform temperature value compared to the normal portion, and may have a relatively high temperature or a low temperature depending on the surface condition. By comprehensively considering external variables such as atmospheric temperature, humidity, and wind speed, the relationship between the temperature difference between the defective part and the forearm and the crack depth can be identified through BIG data analysis, and crack depth information can be obtained using this. have.

In the above, a case in which standard deviation cases are collected in a database and analyzed by machine learning has been described as an example, but the present invention is not limited thereto, and various analysis methods may be used. In addition, the step of setting the defect range ( S100 ) does not necessarily have to be set based on an experiment, and may be set based on data received from the outside.

1 shows an embodiment in which the step (S100) of setting the defect range is performed before the step (S200) of obtaining a plurality of temperature values from the welding part to be inspected, but the step of setting the defect range (S100) is Then, those skilled in the art will understand that it only needs to be performed before the step (S400) of comparing the index value calculated from the temperature values of the welding part to be inspected with the defect range.

In order to inspect the welding part of the actual steel structure, the welding part defect detection method according to an embodiment of the present invention may include acquiring a plurality of temperature values corresponding to a plurality of points on one surface of the welding part to be inspected (S200). have.

Various methods may be used to obtain temperature values corresponding to a plurality of points in the welding part, but in an embodiment of the present invention, the step (S200) of obtaining the plurality of temperature values may be performed by thermal imaging. (S210). That is, one surface of the welding part to be inspected may be photographed with a thermal imaging camera.

7 depicts this process. Defects on the welded part appear in a different form than the normal part in which infrared (IR) energy is emitted. A thermal imaging camera can detect the emitted IR energy and process the received signal for display on a display.

In the case of using a thermal imaging camera, a required number of temperature values may be acquired from the corresponding welding part by selecting an area having a predefined size from the acquired thermal image ( S220 ).

Figure 8 depicts this process. In the example shown in FIG. 8 , a temperature value corresponding to the pixel was measured by selecting 4x2 pixels from the acquired thermal image. As shown, the difference in temperature values was not large in the designated area of the image photographed with the normal part, whereas the temperature of some points was significantly insufficient compared to the average in the designated area of the image photographed with the defective part and the temperature of some points was average It can be seen that a large excess That is, the temperature values appear relatively uniformly in the normal part, whereas the temperature values appear non-uniformly in the defective part, so that the standard deviation of the measured temperature values appears large.

In FIG. 8 , a case in which the standard deviation is calculated from temperature values of eight points is illustrated as an example, but the present invention is not limited thereto. In another embodiment of the present invention, the region may be selected to have various sizes, and the number of points at which the temperature value is selected may also vary within the selected region.

The process of obtaining a temperature value (S200) is as described above so that the process of determining the presence or absence of defects in the weld can derive a more accurate result, as described above, outside air temperature, humidity, wind speed, insolation, the surface temperature of the member adjacent to the weld, Measure at least one of the following factors: the angle between the members on both sides of the weld, the width of the weld, the temperature of the top, the thickness of the member to be welded, the type of defect, the depth of the defect, and the number of temperature values used to calculate the standard deviation can do. As described above, since the surface temperature of the member adjacent to the welding part can be simultaneously acquired through thermal imaging, it provides an advantage that accuracy can be increased without additional equipment. Of course, some factors such as outside temperature, humidity, wind speed, and solar radiation may be measured using separate equipment or may be obtained based on data received from outside such as the Meteorological Administration.

After obtaining a plurality of temperature values from the welding part to be inspected (S200), a standard deviation of the plurality of obtained temperature values may be calculated to calculate a predetermined index value (S300). Here, the index value may be the calculated standard deviation itself, or may be another value calculated by further processing the standard deviation.

If the temperature value obtained by thermal imaging is used as it is, the possibility of a defect can be determined by detecting a portion having a temperature difference greater than or equal to a threshold value compared to the normal portion. However, a temperature change may occur in both the normal part and the defective part due to external factors such as temperature, humidity, wind speed, solar radiation, and the like, and the temperature change in the normal part and the defective part may not increase or decrease equally. In addition, in the case of the defective part, as shown in FIG. 8 , since the temperature is very non-uniform according to the location, when the defect is detected based on the temperature difference, different results may be derived depending on the selected point.

On the other hand, if a temperature value is extracted from a plurality of points, a standard deviation is calculated for it, and a defect is determined based on the standard deviation value, the determination result is derived with higher reliability because it is not based on the temperature value of any one point. can be

In addition, since one standard deviation value is calculated from several temperature values and an index value is calculated based on this, it is not necessary to use the same number of measured temperature values in all cases. That is, the number of temperature values used in the calculation of standard deviation in the standard deviation cases collected to set the defect range is different, or the temperature value used in the calculation of the standard deviation in the standard deviation cases in the database and the measurement of the welding part to be inspected. Even if the number of is different, it is possible to derive a judgment result based on the corresponding standard deviation values.

In the present specification, it is assumed that a case in which the step of acquiring the temperature value is performed by thermal imaging has been described as an example, but the present invention is not necessarily limited thereto. For example, one or more of the standard deviation cases obtained in the step of collecting standard deviation cases for setting the defect range and obtaining the temperature value from the welding part to be inspected may be directly taken using a general thermometer, etc. have.

When the index value is calculated, it is possible to determine whether a defect exists in the welding part to be inspected by comparing the index value with the defect range ( S400 ).

As described above, the defect range may be set differently for each of the defect types depicted in FIG. 5 . If the database that collects standard deviation cases is input to the machine learning algorithm, the algorithm can learn the characteristics of the standard deviation of each type of temperature value to be determined through BIG DATA analysis. Accordingly, in an embodiment of the present invention, it is possible to determine the type and degree of a defect from the measured temperature values.

In the welding part defect detection method according to an embodiment of the present invention, after the target welding part is inspected as described above, the result value may be added to the above-described database. That is, the normal and defective parts detected in the acquired thermal imaging image and other factors simultaneously measured may be added to the database. Accordingly, as time passes, input values collected in the database can be continuously accumulated, and the reliability and accuracy of the detection result of the inspection method can be continuously increased.

Hereinafter, a method for detecting a weld defect according to another embodiment of the present invention will be described. In the present embodiment, two or more standard deviation values form one index, not just one standard deviation value of the temperature values for which the index value for comparison with the defect range is measured.

6 is a graph illustrating examples of standard deviations collected according to a welding defect detection method according to an embodiment of the present invention. Referring to FIG. 6 , the defective part in the welded portion of the steel structure may exhibit different standard deviation values for each type of defect, and these standard deviation values may also exhibit different aspects depending on the outdoor temperature. Therefore, in the present embodiment, such a rate of change of the standard deviation may be used as an index value.

For example, when collecting standard deviation cases for setting the defect range and/or measuring the temperature values of the welding part to be inspected, at least one of external air temperature, humidity, wind speed, solar radiation and the surface temperature of a member adjacent to the inspection target welding part It is possible to simultaneously measure the external factor of , and measure the temperature values at the same point in the corresponding weld again at different times when the external factors show different values.

Since a predetermined difference will occur between the first standard deviation calculated in the first time period and the second standard deviation calculated in the second time period, in the step of determining whether a defect exists, at least one of the first standard deviation and the second standard deviation The type of defect can be determined based on the value of , and also the rate of change between the first standard deviation and the second standard deviation.

Therefore, for example, when a thermal imaging camera is attached to an unmanned aerial vehicle and a steel structure is inspected, a plurality of temperature values can be obtained from the welding part to be inspected in the first time period, and the standard external factor values are different. It is possible to acquire a plurality of temperature values again at the same point in two time periods. In general, external temperature, humidity, wind speed, insolation, and surface temperature of the member adjacent to the weld will exhibit different values at different times.

In particular, when the surface temperature of the member adjacent to the welding portion is used, the step of heating the member may be performed during or before acquiring the temperature value in the second time period. Accordingly, the surface temperature of the adjacent member as an external factor makes a larger difference, so that the first standard deviation and the second standard deviation calculated in the first time period and the second time period can show a clearer difference.

According to some embodiments of the present invention described above, it is possible to determine whether there is a defect in the welded part of the steel frame by a very simple method using a thermal imaging camera. In particular, unlike conventional methods, the thermal imaging camera can not only judge the presence or absence of defects in the welded part of the steel frame, but also the type and extent of the defect, and allows the inspector to quickly devise an appropriate countermeasure. Safety can be greatly increased.

In addition, according to an embodiment of the present invention, high reliability and high accuracy can be provided by comprehensively considering the environmental factors of an inspection target, and the reliability and accuracy of the inspection method can be further increased as data is accumulated.

As described above, although the present invention has been described with reference to limited embodiments and drawings, the present invention is not limited to the above-described embodiments, which are various modifications and variations from these descriptions for those of ordinary skill in the art to which the present invention pertains. Transformation is possible. Accordingly, the spirit of the present invention should be understood only by the claims described below, and all equivalents or equivalent modifications thereof will fall within the scope of the spirit of the present invention.

Claims (8)

A method for detecting a defect in a weld, comprising:
acquiring a plurality of temperature values corresponding to a plurality of points on one surface of the welding part to be inspected;
calculating an index value by calculating a standard deviation of the plurality of temperature values; and
and determining whether a defect exists in the inspection target welding part by comparing the index value with a preset defect range.
The method of claim 1,
The acquiring of the plurality of temperature values includes photographing the welding part to be inspected with a thermal imaging camera, wherein the plurality of temperature values correspond to a plurality of pixel values in the image photographed by the thermal imaging camera. A method for detecting defects in the welding part.
The method of claim 1,
The determining whether the defect exists comprises comparing the index value with a plurality of defect ranges and determining the type of the defect based on the defect range to which the index value belongs.
4. The method of claim 1 or 3,
wherein acquiring the plurality of temperature values includes measuring one or more external factors while acquiring the plurality of temperature values;
The defect detection method, characterized in that the defect range is set differently according to the value of the external factor.
5. The method of claim 4,
Wherein the external factor includes at least one of an external temperature, humidity, wind speed, insolation, and a surface temperature of a member adjacent to the welding part to be inspected when the plurality of temperature values are acquired.
The method of claim 1,
Before the step of acquiring the plurality of temperature values, the welding part defect detection method further comprising the step of setting the defect range.
7. The method of claim 6,
The step of setting the defect range comprises collecting standard deviation cases calculated from a plurality of temperature values of the weld where the defect exists for each of the one or more defect types.
8. The method of claim 7,
The step of setting the defect range comprises:
Building the collected standard deviation cases into a database, and
Welding defect detection method comprising the step of inputting the database to a deep learning algorithm.

Images