How to judge the quality of laser processing

Abstract

To determine a plurality of defective states by laser processing.SOLUTION: A method for determining the quality of laser processing: generates a first type image including a first image which is obtained by imaging a laser processing place by one or more temperature measurement cameras, and indicates temperature distribution of a first temperature range including a melting point of a base material constituting the processing place, and a second image indicating temperature distribution of a second temperature range including a temperature range higher than the upper limit value of the first temperature range; generates the first type image a plurality of times while laser processing is performed on the one processing place as a target; obtains an average value of the sum of brightness values of pixels having brightness values in a prescribed range among pixels included in each image of the plurality of sets of first type images, for each time section including the photographic time of the respective images; acquires a set of first average values; determines the quality of the processing place on the basis of comparison of the set of the first average values with the set of the first average values in a prescribed non-defective; and outputs the result of the determination.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to laser processing. [Background technology]

[0002] Conventionally, a welding method using laser processing (hereinafter referred to as laser welding) has been used as a method for lap-welding two metal plates, as described in Patent Document 1. This laser welding is a welding method in which two metal plates are melted and joined by irradiating an object with a high-energy density laser beam.

[0003] Laser welding is divided into two types: keyhole type and heat conduction type, depending on the heating method used to heat the weld. Keyhole type laser welding is a welding method that uses an increased energy density of laser light to create a depression called a keyhole in the metal plate. A keyhole is formed by the repulsive force generated on the surface of the metal plate when the metal evaporates. By forming a depression, the keyhole type allows for deeper penetration than the heat conduction type. On the other hand, the heat conduction type uses a lower energy density of the laser light than the keyhole type, allowing the metal to absorb the energy. Therefore, with the heat conduction type, the low energy density prevents metal evaporation. In other words, keyholes do not occur with the heat conduction type. The heat conduction type melts the metal by propagating energy through the metal.

[0004] Defects in laser welding are primarily classified as burn-through, insufficient area, and penetration. Burn-through is a condition in which a hole or dent appears in the metal sheet irradiated with the laser beam. For example, consider a case in which there is a gap between the metal sheets to be welded. In this case, the molten metal must overcome the gap between the metal sheets to join them. Therefore, this welding process involves melting the metal sheets to overcome this distance. However, compared to when there is no gap, melting the metal sheets increases the risk of burn-through. Insufficient area refers to an insufficient joining area between the two metal sheets. Insufficient joining area reduces the mechanical strength of the joint. For example, when the laser beam output is low, the metal does not melt sufficiently, resulting in an insufficient joining area. Penetration refers to a hole created in the two metal sheets due to melting of the metal. For example, when the laser beam output is high, the metal melts excessively, resulting in a hole that penetrates the two metal sheets.

[0005] Patent Document 1 describes a method for determining whether a product is defective due to penetration. The method in Patent Document 1 utilizes the brightness distribution of light emitted from the weld. This light is emitted due to a temperature rise in the weld. The method in Patent Document 1 captures this light emitted from a position coaxial with the laser light, thereby obtaining an image of the weld that emits a circular light. The brightness distribution of the weld changes depending on the state of the keyhole. Therefore, the method in Patent Document 1 determines the degree of penetration from this brightness distribution. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Japanese Patent Application Laid-Open No. 2010-115680 Summary of the Invention [Problem to be solved by the invention]

[0007] However, as mentioned above, defects in laser welding are not limited to penetration. In non-penetration states, burn-through and insufficient area may occur. Because burn-through and insufficient area occur due to phenomena different from penetration, it is difficult to apply the method described in Patent Document 1.

[0008] Specifically, in the case of burn-through caused by a gap between metal sheets, it is difficult to judge whether the weld is good or bad based on the brightness distribution. In this burn-through, the metal sheets are melted excessively because molten metal flows into the gap. However, even when molten metal is present in the gap, a keyhole may be formed normally. Therefore, it is difficult to judge whether a weld is good or bad based on the brightness distribution that indicates the state of the keyhole, regarding a poor weld caused by the occurrence of burn-through.

[0009] In the case of insufficient area due to insufficient laser beam output, the keyhole may be formed normally depending on the level of laser beam output. For this reason, as with the burn-through example, it is difficult to judge the quality of the weld from the brightness distribution that indicates the state of the keyhole.

[0010] Therefore, in order to determine whether a weld has multiple defects, including penetration, it is necessary to check the characteristics of multiple welds. [Means for solving the problem]

[0011] The present disclosure can be realized in the following forms.

[0012] (1) According to one aspect of the present disclosure, there is provided a method for determining the quality of laser processing, which includes: (a) a step of capturing an image of a processed portion of the laser processing using one or more temperature measurement cameras to generate a first type of image, the first type of image including a first image representing a temperature distribution in a first temperature range including the melting point of a material constituting the processed portion, and a second image representing a temperature distribution in a second temperature range including a temperature range higher than an upper limit of the first temperature range; (b) a step of performing the step (a) multiple times while the laser processing is being performed on one of the processed portions; and (c) a step of performing the step (b) on a plurality of sets of the first temperature distributions obtained by the step (b). (d) performing an averaging process for the first image and the second image of a seed image, for each time interval including the time at which each image was taken, to obtain an average value of the sum of brightness values ​​of pixels contained in each of the first image and the second image, the average value being within a predetermined range; and (d) determining whether the processed area is good or bad based on a comparison between the set of first average values ​​as a plurality of feature quantities and the set of first average values ​​for a predetermined good product as a plurality of feature quantities that serve as a reference, and outputting the result of the determination. By adopting this configuration, it is possible to acquire feature values ​​corresponding to temperature changes over time at the weld using temperature measurement cameras with different temperature ranges. Furthermore, by acquiring feature values ​​based on pixel brightness values, it is possible to acquire feature values ​​for each part of the weld that shows different brightness values. Therefore, by acquiring multiple feature values ​​for each part of the weld, it is possible to determine the pass / fail status of multiple defects in the weld, including penetration. (2) In the determination method of the above embodiment, in step (c), the brightness values ​​of pixels contained in each of the first image and the second image that have brightness values ​​within a predetermined range may be within the range of all values ​​that the brightness values ​​of pixels contained in each of the first image and the second image can take. By adopting this configuration, the brightness values ​​of the pixels are not changed, and therefore the feature amount of the weld as a whole can be acquired, which makes it easier to determine whether the weld is in a defective state where the features tend to appear across the entire weld. (3) In the determination method of the above embodiment, in step (c), the brightness values ​​of pixels contained in each of the first image and the second image that have brightness values ​​within a predetermined range may be in a range that does not include the maximum possible brightness values ​​of pixels contained in each of the first image and the second image. This configuration allows luminance values ​​that do not reflect temperature differences to be excluded from consideration when determining feature values, thereby enabling feature values ​​that reflect temperature differences to be determined, thereby facilitating the determination of defective welds whose characteristics tend to appear around the welds. (4) The above-described form of the judgment method may further include: (e) a step of performing an averaging process for the first images and the second images of the plurality of sets of first-type images obtained by step (b) to obtain an average value of the sum of brightness values ​​of pixels having brightness values ​​in a predetermined other range among pixels included in each of the first images and the second images for each of the time periods including the shooting time of each image, thereby obtaining a set of second average values, wherein the predetermined other range of brightness values ​​includes the entire range of brightness values ​​of pixels included in each of the first images and the second images; and step (d) may further include a step of making the judgment on the pass / fail of the processed portion based on a comparison between the set of second average values ​​as the plurality of feature amounts and a set of second average values ​​for a predetermined good product as a reference plurality of feature amounts, and outputting the judgment result. This configuration makes it possible to acquire feature amounts for the entire weld and its periphery, making it possible to determine whether a weld is good or bad even in cases where features tend to appear only in the entire weld or only in the periphery. (5) In the determination method of the above embodiment, if the processed part is defective, (f) a step of identifying the type of defect of the processed part based on a comparison between a set of defect features as the plurality of features previously associated with the type of defect and a plurality of feature values ​​including at least one of the set of the first average values ​​and the set of the second average values ​​as the plurality of feature values ​​obtained by imaging the processed part, and outputting the identified type of defect can be adopted. By adopting such an embodiment, the type of defect can be identified from the feature amount of the processed portion. (6) In the determination method of the above embodiment, the set of defect features further includes a cause of the defect associated with at least one of the type of defect and the set of defect features associated with the type of defect, and the step (f) can include a step of identifying the cause of the defect of the processed part based on a comparison of the set of defect features with a plurality of feature values ​​including at least one of the set of first average values ​​and the set of second average values ​​as the plurality of feature values ​​obtained by imaging the processed part, and outputting the identified cause of the defect. By adopting such an embodiment, the cause of the defect can be identified from the feature amount of the processed portion. [Brief explanation of the drawings]

[0013] [Figure 1] FIG. 2 is an explanatory diagram showing an overview of pass / fail judgment of laser processing according to the first embodiment of the present disclosure. [Figure 2] FIG. 6 is a block diagram showing the functional configuration of a computer 600. [Figure 3] An image of the temperature distribution in the low temperature region is shown. [Figure 4] An image of the temperature distribution in the high temperature region is shown. [Figure 5] 10 is a flowchart showing the process of laser processing and capturing a temperature distribution image. [Figure 6] 10 is a flowchart showing a process of calculating the sum of pixel brightness values ​​from an image of the temperature distribution at a welding point. [Figure 7] This is an image in which the brightness range of the low temperature image in FIG. 3 has been changed to 0 to 254. [Figure 8] This is an image in which the brightness range of the image in the high temperature range of FIG. 4 has been changed to 0 to 254. [Figure 9] 10 is a flowchart illustrating a process for generating a set of average values. [Figure 10] This is a waveform showing the relationship between the sum of brightness values ​​and the shooting time based on an image in the low temperature range. [Figure 11] This is a waveform showing the relationship between the sum of brightness values ​​and the shooting time based on an image of the high temperature range. [Figure 12] 10 is a flowchart showing a process for determining whether a welded portion is good or bad. [Figure 13] 10 is a table showing the pass / fail judgment results based on images in the low temperature range. [Figure 14] 10 is a table showing the pass / fail judgment results based on the image of the high temperature range. [Figure 15] 10 is a flowchart for identifying a defect mode and a cause of the defect in a processed portion. [Figure 16] 10 is a table showing a "High / Low matrix" based on a plurality of feature quantities in the low temperature range. [Figure 17] 10 is a table showing a "High / Low matrix" based on multiple feature quantities in the high temperature range. DETAILED DESCRIPTION OF THE INVENTION

[0014] A. First embodiment: A-1.Device configuration: FIG. 1 is an explanatory diagram showing an overview of a method for determining the quality of laser processing according to a first embodiment of the present disclosure. In FIG. 1, an XYZ Cartesian coordinate system is set up to facilitate understanding of the technique. The X and Y axes are assumed to be along a horizontal plane, and the Z axis is assumed to be along a vertical line. An apparatus 10 is used to determine the quality of laser processing. The apparatus 10 performs laser processing and determines the quality of the processing location Pa. The apparatus 10 includes a laser oscillator 100, a laser scanner 200, a low-temperature measuring camera 300, a high-temperature measuring camera 400, a quality display unit 500, and a computer 600.

[0015] The laser oscillator 100 outputs a laser beam LB10. The laser oscillator 100 is, for example, a fiber laser. The laser oscillator 100 is connected to the laser scanner 200 via an optical isolator 110. The laser oscillator 100 includes the optical isolator 110.

[0016] The optical isolator 110 passes the laser light LB10 emitted from the laser oscillator 100. The optical isolator 110 blocks light of the same wavelength as the laser light LB10 from entering the laser oscillator 100. The specific function will be explained later. The optical isolator 110 is disposed at the exit of the laser oscillator 100.

[0017] The laser scanner 200 welds the workpiece Ob by irradiating it with a laser beam LB10. The laser scanner 200 controls the mirrors necessary to capture an image of the processing location Pa. More specifically, in FIG. 1, the laser scanner 200 welds the workpiece Ob1 and the workpiece Ob2 by irradiating them with a laser beam LB10 from the laser oscillator 100. At this time, the workpiece Ob emits light due to heating by the laser beam LB10. The light Ir10 emitted at the processing location Pa of the workpiece Ob includes light Ir110 in a wavelength range required for temperature measurement by the low-temperature measuring camera 300 and light Ir121 in a wavelength range required for temperature measurement by the high-temperature measuring camera 400. The light Ir10 also includes light in a wavelength range not required for temperature measurement, which will be described in detail later. The temperature measurement ranges used by various cameras to determine this wavelength range will be described later. The laser scanner 200 guides light Ir10 emitted from the processing location Pa to the low-temperature measuring camera 300 and the high-temperature measuring camera 400 via a first dichroic mirror 220, a second dichroic mirror 250, and a first reflecting mirror 260. The laser scanner 200 includes a control unit 210, the first dichroic mirror 220, a focusing system 230, a protective glass 240, the second dichroic mirror 250, the first reflecting mirror 260, and an optical filter 270.

[0018] The first dichroic mirror 220 guides the laser light LB10 emitted from the laser oscillator 100 to the focusing system 230. The first dichroic mirror 220 reflects only light in the wavelength range of the laser light LB10. As shown in FIG. 1, the first dichroic mirror 220 reflects the laser light LB10, thereby guiding the laser light LB10 to the focusing system 230. On the other hand, the first dichroic mirror 220 passes light Ir10 from the processing area Pa. This passed light is called light Ir100 because it does not include the wavelength range of the laser light LB10. Between the focusing system 230 and the first dichroic mirror 220, the optical axes of the laser light LB10 and the light Ir10 emitted from the processing area Pa are aligned. This alignment of the optical axes will be described in detail later.

[0019] The focusing system 230 reflects the laser light LB10 from the first dichroic mirror 220 and focuses it on the processing location Pa. Furthermore, the focusing system 230 magnifies the light Ir10 emitted from the processing location Pa. The focusing system 230 reflects the light Ir10 reflected from the processing location Pa in the direction where the second dichroic mirror 250 and the first reflecting mirror 260 are arranged. The focusing system 230 is composed of a second reflecting mirror and a focusing lens. To facilitate understanding of the technology, the second reflecting mirror and the focusing lens are not shown in the drawings.

[0020] The second reflecting mirror is, for example, a galvanometer mirror. The second reflecting mirror changes the reflection direction of various types of light reaching the focusing system 230 by moving in the direction of arrow Am10. Therefore, the second reflecting mirror irradiates the processing area Pa with the laser light LB10 by changing the irradiation position of the laser light LB10 in the direction of arrow Am20. Meanwhile, the second reflecting mirror receives light Ir10 from the processing area Pa along the same optical path as the laser light LB, and reflects the light in the direction where the second dichroic mirror 250 and the first reflecting mirror 260 are located.

[0021] The condenser lens is, for example, a double-sided convex lens. The condenser lens refracts the laser light LB incident as parallel light, thereby focusing the light on the processing location Pa. On the other hand, the condenser lens allows light Ir10 diffused from the processing location Pa to enter from the surface opposite to the surface onto which the laser light LB is incident. Therefore, the condenser lens converts the light Ir10 into parallel light, thereby magnifying the light Ir10. The magnified light Ir10 is guided to the low-temperature measuring camera 300 and the high-temperature measuring camera 400 via the second reflecting mirror.

[0022] The protective glass 240 protects the internal devices of the laser scanner 200. The protective glass 240 can transmit the laser light LB10 from the focusing optical system 230. The protective glass 240 is disposed at the output port of the laser scanner 200 from which the laser light LB10 is emitted.

[0023] The second dichroic mirror 250 guides the light Ir110 of the light Ir10 reflected from the focusing system 230 to the low-temperature camera 300. The light Ir110 is only the light in the wavelength range required for temperature measurement by the low-temperature camera 300 out of the light Ir100 that arrives via the first dichroic mirror 220. That is, the second dichroic mirror 250 guides the light Ir110 to the low-temperature camera 300 by reflecting only the light Ir110 in the wavelength range required for temperature measurement by the low-temperature camera 300. On the other hand, the second dichroic mirror 250 transmits light other than the light Ir110 in the wavelength range required for temperature measurement by the low-temperature camera 300. This transmitted light is called light Ir120 because it is the light in the wavelength range required for temperature measurement by the low-temperature camera 300.

[0024] The first reflecting mirror 260 guides the light Ir120 that has arrived via the second dichroic mirror 250 to the pyrometric camera 400 via the optical filter 270. The first reflecting mirror 260 is a total reflection mirror. That is, the first reflecting mirror 260 functions to reflect light of all wavelength ranges in the light Ir120 that has passed through the second dichroic mirror 250.

[0025] The second dichroic mirror 250 and the first reflecting mirror 260 are arranged coaxially with the optical axis of the laser light LB10 between the condensing system 230 and the first dichroic mirror 220. That is, the optical axes of the light Ir10 emitted from the processing area Pa and the laser light LB10 coincide between the condensing system 230 and the first dichroic mirror 220. In FIG. 1, this optical axis is an axis extending along the x-axis. This configuration allows the low-temperature measurement camera 300 and the high-temperature measurement camera 400 to capture images of the processing area Pa from the same axis as the optical axis of the laser light LB10. This prevents light emitted from components other than the processing area from entering the field of view of each temperature measurement camera. That is, this configuration prevents a decrease in the accuracy of pass / fail determination of the processing area Pa due to noise in the feature values ​​necessary for pass / fail determination.

[0026] The optical filter 270 passes only the light Ir121 in the wavelength range required for temperature measurement by the high-temperature camera 400, out of the light Ir120 reflected by the first reflecting mirror 260. On the other hand, the optical filter 270 blocks light in wavelength ranges other than the light Ir121. The optical filter 270 is disposed in the optical path from the first reflecting mirror 260 to the high-temperature camera 400.

[0027] The control unit 210 includes a processor, RAM, and ROM. The processor in the control unit 210 executes various programs stored in the RAM to realize the functions of each unit of the laser scanner 200. The processor in the control unit 210 uses the RAM to store information necessary for processing. To facilitate understanding of the technology, explanations of the processor, RAM, and ROM in the control unit 210 will be omitted.

[0028] The low-temperature measuring camera 300 and the high-temperature measuring camera 400 are temperature measuring cameras. A temperature measuring camera detects infrared rays emitted from an object, converts the detected amount of infrared energy into temperature, and generates a temperature distribution image Im that represents the temperature distribution. For example, this is thermography.

[0029] The low-temperature measurement camera 300 generates a first image Im11 of the temperature distribution in a first temperature range that includes the melting point of the material that constitutes the processing area Pa. The first temperature range is also called the low-temperature range. Specifically, the low-temperature range is a temperature range of ±200 K based on the melting point of the workpiece Ob. That is, the low-temperature measurement camera 300 captures an image of the processing area Pa during laser processing and generates a two-dimensional first image Im11 that represents the temperature distribution in the low-temperature range. The low-temperature measurement camera 300 is connected to the laser scanner 200 and the computer 600. The low-temperature measurement camera 300 captures an image of the processing area Pa by receiving light Ir110 in the wavelength range required for temperature measurement via the laser scanner 200. The low-temperature measurement camera 300 sends the captured first image Im11 to the computer 600.

[0030] The pyrometric camera 400 generates a second image Im12 of the temperature distribution in a second temperature range that includes a temperature range higher than the upper limit of the low temperature range. The second temperature range is also called the high temperature range. Specifically, the high temperature range is a temperature measurement range of 400 K, with the upper limit of the low temperature range as its lower limit. That is, the pyrometric camera 400 captures an image of the processing location Pa during laser processing and generates a two-dimensional second image Im12 that represents the temperature distribution in the high temperature range. The pyrometric camera 400 is connected to the laser scanner 200 and the computer 600. The pyrometric camera 400 captures an image of the processing location Pa by receiving light Ir121 in the wavelength range required for temperature measurement via the laser scanner 200. The pyrometric camera 400 sends the captured second image Im12 to the computer 600.

[0031] The pass / fail display unit 500 displays information relating to the pass / fail judgment of the processed portion Pa. The pass / fail display unit 500 is connected to the computer 600. The pass / fail display unit 500 is, for example, a display. The pass / fail display unit 500 displays the information relating to the pass / fail judgment of the processed portion Pa sent from the computer 600, thereby conveying the information to the user.

[0032] FIG. 2 is a block diagram showing the functional configuration of the computer 600. The computer 600 controls each component involved in determining the quality of the processed portion Pa. The computer 600 includes a processor 610, a RAM 620, and a ROM 630. The ROM 630 is a read-only semiconductor memory that stores in advance a control program for controlling each component involved in determining the quality of the processed portion Pa. The RAM 620 includes a main memory, which is a semiconductor memory, and an auxiliary storage device, such as a hard disk or solid state drive. The RAM 620 stores information necessary for determining the quality of the processed portion Pa. The stored information will be described later.

[0033] The processor 610 realizes the functions of each unit by executing various programs stored in the ROM 630. The processor 610 uses the RAM 620 to store information necessary for processing. The processor 610 includes a camera control unit 611, a brightness value adjustment unit 612, a brightness value calculation unit 613, and a determination unit 614. The specific functions of each unit will be described in detail later.

[0034] A-2. Operation of laser scanner and temperature measurement camera: In FIG. 1, laser processing and temperature measurement image capture by the apparatus 10 are performed as follows. The laser scanner 200 guides the laser beam LB10 emitted from the laser oscillator 100 to the first dichroic mirror 220 via the optical isolator 110. The laser beam LB10 is irradiated onto the workpiece Ob via the first dichroic mirror 220, the focusing system 230, and the protective glass 240. The focusing system 230 irradiates the laser beam LB10 onto the processing location Pa by changing the orientation of the second reflecting mirror. A keyhole is formed at the processing location Pa by heating with the laser beam LB10. The workpieces Ob1 and Ob2 are joined at the processing location Pa by melting their metals together. In other words, the workpieces Ob1 and Ob2 are lap-welded.

[0035] As the processing area Pa becomes hot, it emits light. Of this light, light Ir10, which is coaxial with the optical axis of the laser beam LB10, reaches the focusing system 230. The light Ir10 that reaches the focusing system 230 is reflected by the second reflecting mirror toward the first dichroic mirror 220. The first dichroic mirror 220 passes light Ir100, which has a wavelength range other than the wavelength of the laser beam LB10. On the other hand, the first dichroic mirror 220 reflects light in the same wavelength range as the laser beam LB10, toward the laser oscillator 100. This light reaches the optical isolator 110 before reaching the laser oscillator 100, and is therefore blocked by the optical isolator 110. To facilitate understanding of the technology, illustration of the light in the same wavelength range as the laser beam LB10 is omitted.

[0036] The light Ir100 that passes through the first dichroic mirror 220 travels toward the second dichroic mirror 250 and the first reflecting mirror 260, which are arranged on the optical axis of the laser light LB10. The second dichroic mirror 250 reflects only the light Ir110 in the wavelength range required for temperature measurement by the low-temperature measuring camera 300. The light Ir120 in the wavelength range other than the light Ir110 passes through the second dichroic mirror 250 and reaches the first reflecting mirror 260. The first reflecting mirror 260 reflects all of the light Ir120 in the wavelength range toward the high-temperature measuring camera 400. An optical filter 270 is arranged between the first reflecting mirror 260 and the high-temperature measuring camera 400. Therefore, of the light Ir120, only the light Ir121 in the wavelength range required for temperature measurement by the high-temperature measuring camera 400 reaches the high-temperature measuring camera 400. Therefore, by photographing the processed place Pa with the low temperature measuring camera 300 and the high temperature measuring camera 400, a first image Im11 of the temperature distribution in the low temperature region and a second image Im12 of the temperature distribution in the high temperature region are generated.

[0037] FIG. 3 shows an example of a first image Im11 of a temperature distribution in the low temperature range. In FIG. 3, the temperature level is represented as the magnitude of pixel brightness. This pixel brightness represents the shade of the pixel. That is, in the temperature distribution image Im of this embodiment, the pixel with the brightest brightness is white, and the pixel with the darkest brightness is black. The degree of brightness of this pixel is represented as a brightness value. In the drawing showing the temperature distribution image Im of this embodiment, the range of pixel brightness values ​​is indicated by the scale Sc. The range of pixel brightness values ​​is called the brightness range. In FIG. 3, the brightness range is 0 to 255. Therefore, pixel locations with a brightness value of 0 are black, and pixel locations with a brightness value of 255 are white. Most of the pixels in region Mp1 have a brightness value of 255. That is, region Mp1 represents a region where the temperature has reached or exceeded the upper limit of the low temperature range due to heating by laser processing.

[0038] FIG. 4 shows an example of a second image Im12 of the temperature distribution in the high-temperature region. In FIG. 4, the brightness range is 0 to 255. Most of the pixels in region Mp2 have a brightness value of 255. The high-temperature region is a temperature measurement range higher than the upper limit of the temperature measurement range by the low-temperature measurement camera 300. The high-temperature region is a temperature range with a temperature width of 400 K, with the upper limit of the low-temperature region as its lower limit. Therefore, there are fewer pixels with a brightness value of 255 in region Mp2 than in region Mp1 in FIG. 3. The low-temperature measurement camera 300 and the high-temperature measurement camera 400 capture images of the processed area Pa from the same axis as the optical axis of the laser beam LB10. Therefore, in the temperature distribution image Im, the centers of regions Mp1 and Mp2 are the irradiation positions of the laser beam LB10, and the temperature is higher toward the center. Therefore, region Mp2 is smaller than region Mp1 because it represents the central area of ​​region Mp1, where the temperature is higher.

[0039] A-3. Laser processing and temperature distribution image capture: 5 is a flowchart showing the process of laser processing and capturing a temperature distribution image Im by the apparatus 10. In step S110 of FIG. 5, the control unit 210 shown in FIG. 1 causes the laser scanner 200 to irradiate the processing area Pa with laser light LB10. A keyhole is formed in the processing area Pa by heating with the laser light LB10. Furthermore, the workpieces Ob1 and Ob2 are fused together, thereby joining the workpieces Ob1 and Ob2.

[0040] In step S120 of FIG. 5, the processor 610 shown in FIG. 2 performs a process of generating a first-type image Im10 multiple times using one or more temperature measurement cameras. Step S120 is performed while laser processing is being performed on one laser processing location Pa. The first-type image Im10 is an image including a first image Im11 and a second image Im12. Therefore, step S120 generates the first image Im11 and the second image Im12 of multiple sets of first-type images Im10. Specifically, the low-temperature measurement camera 300 and the high-temperature measurement camera 400 capture a first image Im11 of the temperature distribution in the low-temperature region and a second image Im12 of the temperature distribution in the high-temperature region multiple times, as shown in FIGS. 3 and 4. This temperature distribution image Im is captured in a state that includes the capture time Dt. This multiple captures are performed at the same time intervals. Information on the first image Im11 and second image Im12, which are temperature distribution images Im, and the photographing time Dt, are stored in the RAM 620. In Fig. 2, the image corresponding to the temperature distribution image Im is abbreviated as the temperature distribution image Im to facilitate understanding of the technology. In step S120, the functional unit of the processor 610 that controls photographing by the low temperature measuring camera 300 and the high temperature measuring camera 400 is the camera control unit 611 shown in Fig. 2.

[0041] In step S130 of Fig. 5, the control unit 210 shown in Fig. 1 confirms that welding has been completed for all welding points. If there are remaining welding points, the control unit 210 proceeds to step S110. If there are no remaining welding points, the control unit 210 ends the processing.

[0042] A-4. Calculation of the average pixel brightness value as a feature from the temperature distribution image: 6 is a flowchart showing a process of calculating the sum Dl of pixel brightness values ​​from an image Im of the temperature distribution at the welding point, performed by the device 10. In step S210 of FIG. 6, the processor 610 shown in FIG. 2 acquires a plurality of sets of first-type images Im10 in S120 of FIG.

[0043] In step S220 of FIG. 6, the processor 610 generates an image configured by pixels having luminance values ​​within a predetermined other range among the pixels included in each of the first image Im11 and the second image Im12 of the plurality of sets of first-type images Im10. The predetermined range of luminance values ​​is the range of all possible values ​​of the luminance values ​​of the pixels included in each of the first image Im11 and the second image Im12. In other words, the images configured by pixels having these luminance values ​​are the first image Im11 and the second image Im12. On the other hand, the predetermined other range of luminance values ​​is a range that does not include the maximum possible value of the luminance values ​​of the pixels included in each of the first image Im11 and the second image Im12. In other words, the processor 610 changes the luminance range of one of the plurality of first-type images Im10 to one that does not include the maximum possible value of the pixel luminance value.

[0044] More specifically, the processor 610 generates an image by changing the brightness value of 255 from the temperature distribution image Im having a brightness range of 0 to 255 to 0. That is, the processor 610 generates a temperature distribution image Im having a brightness range of 0 to 254. Regarding the temperature distribution image Im having a brightness range of 0 to 254, an image based on the first image Im11 is called a third image Im21, and an image based on the second image Im12 is called a fourth image Im22. Furthermore, an image including the third image Im21 and the fourth image Im22 is called a second-type image Im20. Therefore, the third image Im21 and the fourth image Im22 are images configured by pixels having brightness values ​​in other predetermined ranges. That is, the third image Im21 and the fourth image Im22 are images configured by pixels having brightness values ​​in a range that does not include the maximum possible brightness value of the pixels included in the first image Im11 and the second image Im12, respectively. The second type image Im20 characteristically shows the area around the weld with low brightness values ​​by setting the brightness values ​​of the pixels included in the first image Im11 and the second image Im12 to a range that does not include the maximum possible value. The third image Im21 and the fourth image Im22 are stored in RAM 620.

[0045] Furthermore, the processor 610 also generates an image in which the brightness range of 1 to 254 is changed to 0 from the temperature distribution image Im, which has a brightness range of 0 to 255. That is, the processor 610 generates an image in which the pixel brightness values ​​are only 0 and 255. For images in which the pixel brightness values ​​are only 0 and 255, an image based on the first image Im11 is called a fifth image Im31, and an image based on the second image Im12 is called a sixth image Im32. Furthermore, an image including the fifth image Im31 and the sixth image Im32 is called a third-type image Im30. By using this brightness range, the third-type image Im30 distinctively shows the center of the weld, which has a high brightness value. In step S220, the functional unit of the processor 610 that generates the temperature distribution image Im with a different brightness range is the brightness value adjustment unit 612 shown in FIG. 2.

[0046] FIG. 7 is an example of a third image Im21 obtained by changing the brightness range of the first image Im11 of the temperature distribution in the low temperature range in FIG. 3 to 0 to 254. That is, FIG. 7 is an image obtained by removing the brightness value 255 by changing the brightness value of pixels included in the first image Im11 of FIG. 3 that have a brightness value of 255 to 0. Because the region Mp3 does not include pixels with a brightness value of 255, the image shows only the periphery of the irradiation position of the laser beam LB10. That is, FIG. 7 more easily shows the characteristics of the state around the irradiation position of the laser beam LB10 than the first image Im11 of FIG. 3.

[0047] FIG. 8 is a fourth image Im22 obtained by changing the brightness range of the second image Im12 of the temperature distribution in the high temperature region in FIG. 4 to 0 to 254. In other words, FIG. 8 is an image in which the brightness value of 255 of the pixels in the second image Im12 in FIG. 4 has been changed to 0, thereby removing the brightness value of 255. Since the region Mp4 does not include the pixel brightness value of 255, the image is smaller than the region Mp3 and displays only the periphery of the irradiation position of the laser beam LB10. Compared to the second image Im12 in FIG. 4, FIG. 8 more easily shows the characteristics of the state in the periphery of the irradiation position of the laser beam LB10.

[0048] In step S230 of FIG. 6, the processor 610 calculates the sum of the brightness values ​​of pixels included in the temperature distribution image Im that have brightness values ​​within a predetermined range. That is, the processor 610 calculates the sum Dl of the brightness values ​​of the pixels in the image before and after the change of the pixel brightness values. Specifically, the images before and after the change of the pixel brightness values ​​are the first image Im11 and the second image Im12 before the change, and the third image Im21 and the fourth image Im22 after the change. The sum Dl of the brightness values ​​is also referred to as the heat distribution intensity. For example, the first image Im11 of FIG. 3 includes many pixels with a brightness value of 255 in the region Mp1. Therefore, the third image Im21 of FIG. 7, which does not include pixels with a brightness value of 255, has a smaller sum Dl of the brightness values ​​than the first image Im11 of FIG. 3. Therefore, the sum Dl of the brightness values ​​indicates the characteristics of each image. The sum Dl of the brightness values ​​is stored in the RAM 620. In step S230, the sum Dl of the luminance values ​​is realized by the luminance value calculation unit 613 as a functional unit of the processor 610.

[0049] 6, processor 610 stores the shooting time Dt and the sum of brightness values ​​Dl for each image in RAM 620. If the processing of step S240 has already been performed, processor 610 adds the newly acquired shooting time Dt and sum of brightness values ​​Dl to the already recorded shooting time Dt and sum of brightness values ​​Dl.

[0050] 6, the processor 610 confirms that processing has been completed for all temperature distribution images Im at the processing location Pa. If there are remaining temperature distribution images Im, the processor 610 proceeds to step S210. If there are no remaining temperature distribution images Im, the processor 610 ends the processing.

[0051] A-5. Generating a set of mean values ​​as multiple features: 9 is a flowchart showing the process of generating a set Dfs of average values ​​as a plurality of feature quantities Dfs by the device 10. In step S310 of FIG. 9, the processor 610 of FIG. 2 obtains the sum Dl of the luminance values ​​and the shooting time Dt from the RAM 620.

[0052] In step S320 of FIG. 9, processor 610 creates a waveform of the relationship between the sum Dl of brightness values ​​and the photographing time Dt in processed area Pa based on the brightness range.

[0053] In step S330 of FIG. 9, the processor 610 obtains, from the waveform of step S320, an average value of the sum of brightness values ​​of pixels included in the temperature distribution image Im that have brightness values ​​within a predetermined range, for each time interval Dti that includes the capture time Dt of the temperature distribution image Im. The sum of brightness values ​​of pixels that have brightness values ​​within a predetermined range is, specifically, the total brightness value Dl. That is, the processor 610 divides the capture time Dt of the processing location Pa into multiple time intervals Dti. The multiple time intervals Dti are stored in the RAM 620. In FIG. 2, to facilitate understanding of the technique, the multiple time intervals Dti described later are represented as time intervals Dti. The processor 610 calculates an average value Df of the total brightness value Dl included in the time interval Dti based on the capture time Dt of the temperature distribution image Im. This process is called averaging.

[0054] The averaging process is performed using the following formula (1). The number of times a temperature distribution image Im is captured within a certain time interval Dti is defined as N. A certain time interval Dti is a time interval determined by dividing the time during which the processing location Pa is captured. Furthermore, the sum Dl of the brightness values ​​of one temperature distribution image Im included in a certain time interval Dti is defined as Sn. n indicates the nth capture in a certain time interval Dti. For example, S2 indicates the sum Dl of the brightness values ​​of the image captured the second time in a certain time interval Dti. n is an integer value between 2 and n. Therefore, the average value X of the sum Dl of the brightness values ​​included in a certain time interval Dti can be calculated using formula (1). X = (ΣSn) / N … (1)

[0055] Figure 10 shows an example of a waveform representing the relationship between the sum of brightness values ​​Dl and the capture time Dt based on the first image Im11, which is an image in the low-temperature region. That is, Figure 10 is a waveform with the heat distribution intensity on the vertical axis and the time on the horizontal axis. Figure 10 shows the heat distribution intensity from the start of welding to the solidification of the molten pool after welding. Therefore, the heat distribution intensity tends to increase in the early stage and decrease in the late stage. The capture time Dt can be divided into the early stage, the middle to late stage, and the post-weld stage, based on the welding stages. These time periods Dti are referred to as time period Dti11, time period Dti12, and time period Dti13, respectively. Furthermore, the period including all the time periods Dti is referred to as time period Dti10. The average value Df of the sum of brightness values ​​Dl in each time period Dti is also referred to as the feature value Df of the processing area Pa.

[0056] FIG. 11 is an example of a waveform showing the relationship between the sum of brightness values ​​Dl and the capture time Dt based on the second image Im12, which is an image of a high-temperature region. That is, FIG. 11 is a waveform with the heat distribution intensity on the vertical axis and the time on the horizontal axis. FIG. 11 was created based on the second image Im12, which was captured of the same processing area Pa as FIG. 10. Therefore, as with FIG. 12, the time intervals Dti can be divided based on the welding stage. In FIG. 11, the time intervals Dti for the early stage, the middle to late stage, and the post-welding stage are referred to as time interval Dti21, time interval Dti22, and time interval Dti23. Furthermore, the interval including all the time intervals Dti is referred to as time interval Dti20. The average value of the sum of brightness values ​​Dl in each time interval Dti is the feature value Df of the processing area Pa.

[0057] In step S340 of FIG. 9, the processor 610 acquires a set of average values ​​Dfs by executing the averaging process of step S330. The set of average values ​​Dfs is a set of average values ​​Df calculated in step S330. This set of average values ​​Dfs is also referred to as a plurality of feature values ​​Dfs. Therefore, a first set of average values ​​Dfs1 is acquired as the set of average values ​​Dfs based on the first-type image Im10. A second set of average values ​​Dfs2 is acquired as the set of average values ​​Dfs based on the second-type image Im20. The plurality of feature values ​​Dfs are stored in the RAM 620. In FIG. 2, to facilitate understanding of the technique, all of the plurality of feature values ​​Dfs are abbreviated as a plurality of feature values ​​Dfs. The functional unit of the processor 610 that realizes the functions of steps S320 to S340 is the brightness value calculation unit 613.

[0058] A-6. Weld quality assessment: 12 is a flowchart showing a process for determining the quality of a welded portion by the apparatus 10. In step S410 of FIG.

[0059] In step S420 of FIG. 12, the processor 610 determines whether the processing location Pa is acceptable or unacceptable based on a comparison between a set Dfs of average values ​​of the multiple feature quantities Dfs at the processing location Pa and a set Dfas of average values ​​of a predetermined non-defective product as a reference multiple feature quantity Dfas. Furthermore, the processor 610 outputs a pass / fail determination result Dr. The set Dfas of average values ​​of a predetermined non-defective product as a reference multiple feature quantity Dfas is determined as a set of average values ​​Dfa of the feature quantities Df at the processing location Pa obtained under the same conditions as the multiple feature quantities Df at the processing location Pa. The same conditions refer to the same camera temperature measurement range, brightness range, and time interval Dti. Furthermore, the average value Dfa of the feature quantities Df at the non-defective product obtained under the same conditions is the average value of the feature quantities Df at the non-defective product obtained under the same conditions. σ denotes the standard deviation of the feature quantities Df at the non-defective product, with the average value Dfa of the feature quantities Df at the non-defective product as the median.

[0060] That is, the processor 610 makes a judgment based on the average value Dfa of non-defective products plus or minus 3σ. If the feature value Df of the processed area Pa exceeds the average value Dfa of non-defective products plus 3σ, the processor 610 judges the processed area Pa to be a defective area. If the feature value Df of the processed area Pa is below the average value Dfa of non-defective products minus 3σ, the processor 610 judges the processed area Pa to be a defective area. If the feature value Df of the processed area Pa is within the average value Dfa of non-defective products plus or minus 3σ, the processor 610 judges the processed area Pa to be a non-defective area. In this embodiment, the set of average values ​​Dfs as the multiple feature values ​​Dfs is a first set of average values ​​Dfs1 and a second set of average values ​​Dfs2. Therefore, in this embodiment, the pass / fail judgment is made by previously determining the set of first average values ​​Dfas1 for non-defective products and the set of second average values ​​Dfas2 for non-defective products as the set of average values ​​Dfas for non-defective products. The average value Dfa of the non-defective products and the set of average values ​​Dfas for the non-defective products are stored in the RAM 620. In Fig. 2, for ease of understanding of the technology, the average value Dfa of all the non-defective products and the set of average values ​​Dfas for the non-defective products are abbreviated and shown as a plurality of feature amounts Dfas. In step S420, the process of determining whether the processed location Pa is good or bad is realized by the determination unit 614 as a functional unit of the processor 610.

[0061] 12, processor 610 stores the pass / fail determination result Dr of processed portion Pa in RAM 620. If the processing of step S420 has already been performed up to that point, processor 610 adds the pass / fail determination result Dr to the pass / fail determination result Dr that has already been recorded.

[0062] 12, processor 610 confirms that the determination has been completed for all feature quantities Df. If there are remaining feature quantities Df in processing location Pa, processor 610 proceeds to step S420. Furthermore, processor 610 outputs the determination result via pass / fail display unit 500. If there are no remaining feature quantities Df, processor 610 ends the process.

[0063] 13 is a table showing the pass / fail judgment results Dr for different defective locations based on images in the low temperature range. Four types of processed locations Pa, from processed location P1 to processed location P4, are processed locations Pa that are classified into defective states (hereinafter referred to as failure modes) of burn-through, penetration, and insufficient area.

[0064] Burn-through is a state in which a hole or dent appears in the workpiece Ob1. In other words, with burn-through, no through-hole occurs between the workpiece Ob1 and the workpiece Ob2. Penetration is a state in which a hole has appeared that penetrates through the workpiece Ob1 and the workpiece Ob2. Insufficient area is a state in which the cross-sectional area where the workpiece Ob1 and the workpiece Ob2 are joined does not meet the required mechanical strength.

[0065] Defective modes are also classified by the cause of the defect. Causes of burn-through defects include, for example, a cause caused by a large gap between the workpieces Ob and a cause caused by a large misalignment of the workpieces Ob.

[0066] The cause of large gaps between workpieces Ob is that in order to join the workpieces Ob across the gaps, excess workpiece Ob is melted, causing burn-through. The cause of large gaps between workpieces Ob is called "large gaps."

[0067] A cause of large deviation of the workpiece Ob is burn-through, which occurs when there is a large deviation in the irradiation position of the laser beam LB10. For example, if the irradiation position is misaligned, the edge of the workpiece Ob may be irradiated with the laser beam LB10, causing the edge to melt excessively. This is because there is little room for the heat to escape from the edge of the workpiece Ob during processing. Therefore, burn-through occurs in the workpiece Ob. A cause of large deviation of the workpiece Ob is called "large workpiece deviation."

[0068] Poor penetration may be caused by, for example, an abnormal focus of the laser beam LB10 or by the output of the laser beam LB10 being too high.

[0069] An abnormality in the focus of the laser beam LB10 is a factor that causes penetration, for example, when the focus of the laser beam LB10 is shifted from normal. Specifically, when the spot diameter of the laser beam LB10 is narrower than normal, the laser beam LB10 with a high energy density is irradiated onto the workpiece Ob, causing penetration. A factor caused by an abnormality in the focus of the laser beam LB10 is called a "focus abnormality."

[0070] The cause of the high output of the laser beam LB10 is that the keyhole is formed deeper than usual, causing penetration. The cause of the high output of the laser beam LB10 is called "high output."

[0071] The causes of the area shortage defect are, for example, a factor due to low output of the laser beam LB10 and a factor due to accumulating fumes. A factor due to low output of the laser beam LB10 is that the melting range of the metal is smaller than usual due to the low output of the laser beam LB10, which causes an area shortage. A factor due to low output of the laser beam LB10 is called "low output."

[0072] The cause of fume retention is a lack of fume blowing, which causes fumes that have accumulated around the processing point Pa to obstruct the laser light LB10, resulting in an insufficient area. Fumes are dust particles that are emitted when the workpiece Ob is heated. Fume blowing is performed to prevent these fumes from accumulating around the processing point Pa. The cause of fume retention is called "fume blowing abnormality."

[0073] In FIG. 13, the feature Df used for pass / fail judgment is the feature Df of the brightness range and time interval Dti in the low temperature range. More specifically, there are three types of brightness range: 0 to 255, 0 to 254, and brightness value of 255 only. These are images that characteristically represent the overall, peripheral, and central states with respect to the irradiation position of the laser beam LB10. Furthermore, there are four types of time intervals Dti: time interval Dti10, time interval Dti11, time interval Dti12, and time interval Dti13. That is, FIG. 13 shows a set of 12 types of average values ​​as multiple feature Dfs. The pass / fail judgment result Dr is indicated as H when the feature Df exceeds the aforementioned judgment criterion. On the other hand, the pass / fail judgment result Dr is indicated as L when the feature Df falls below the aforementioned judgment criterion.

[0074] For example, in the processing location P1, the workpiece Ob is melted excessively compared to the normal state, causing burn-through. However, the processing location P1 is determined to be a good location in the time interval Dti11 in the brightness range of 0 to 254. Therefore, the processing location P1 is not determined to be a defective location under these conditions. The processing location P1 is determined to be a defective location in time intervals Dti other than the time interval Dti11. Therefore, to determine whether the processing location P1 is good or bad, the feature amount Df in the time interval other than the time interval Dti11 in the brightness range of 0 to 254 is required. The same is true for the processing location P2, which has a penetration defect mode.

[0075] In FIG. 13, the processing locations P3 and P4 are in a defective mode classified as an insufficient area. However, the processing locations P3 and P4 are determined to be good locations in the brightness range 255. The processing location P3 is determined to be a defective location in the time interval Dti13 outside the brightness range 255. The processing location P4 is determined to be a defective location under conditions other than the brightness range 255 and the time interval Dti13. Therefore, to determine the pass / fail of the processing locations P3 and P4, feature amounts Df in different brightness ranges are required. In other words, to determine the pass / fail of the processing locations P3 and P4, feature amounts Df in the periphery as well as the center of the processing location Pa are required.

[0076] FIG. 14 is a table showing pass / fail judgment results Dr based on images of the high temperature range for different defective locations. The table configuration is the same as in FIG. 13. Furthermore, processing locations P1 to P4 are the same processing location Pa as in FIG. 13. Processing location P3 is judged as a good location for all feature amounts Df. However, as mentioned above, processing location P3 can be judged as a defective location from feature amounts Df in the low temperature range. Therefore, feature amounts Df in the low temperature range are required to judge the pass / fail of processing location P3.

[0077] By adopting such an embodiment, the device 10 can acquire the feature amount Df corresponding to the temperature change over time at the processing location Pa using temperature measurement cameras with different temperature ranges. Furthermore, the device 10 can acquire the feature amount Df for each part of the processing location Pa from images with different brightness value ranges. Therefore, by acquiring multiple feature amounts Dfs, the device 10 can determine the pass / fail of multiple failure modes Ddm, including penetration, at the processing location Pa.

[0078] B. Second embodiment: In the first embodiment, the device 10 only judges the quality of the processed portion Pa. However, the device 10 can identify the failure mode Ddm and the cause of failure Ddf in the processed portion Pa by determining in advance a plurality of feature amounts Dfs in the failure portion.

[0079] FIG. 15 is a flowchart showing the process of the apparatus 10 for identifying a failure mode and a cause of the failure at the processing location Pa. Step S450 in FIG. 15 is a process subsequent to step S440 in FIG. 12. In step S450, the processor 610 identifies the cause and type of the failure at the processing location Pa. This identification is performed based on a comparison between the set of failure features Dfds and a plurality of feature values ​​Dfs including at least one of a set of first average values ​​Dfs1 and a set of second average values ​​Dfs2, which are a plurality of feature values ​​Dfs obtained by imaging the processing location Pa. The set of failure features Dfds is a plurality of feature values ​​Dfs associated with a failure type in advance. Furthermore, the set of failure features Dfds is a plurality of feature values ​​Dfs including a failure cause Ddf associated with at least one of a failure type Ddm and a set of failure features Dfds associated with the failure type Ddm. The failure type Ddm is also referred to as a failure mode Ddm.

[0080] In the second embodiment, the RAM 620 pre-stores a set of defect features Dfds determined for each pair of a failure mode Ddm and a failure factor Ddf. The relationship between the failure mode Ddm, the failure factor Ddf, and the feature Df is called a "High / Low matrix." That is, the "High / Low matrix" represents the set of defect features Dfds pre-associated with the failure mode Ddm and the failure factor Ddf. The processor 610 identifies the failure mode Ddm and the failure factor Ddf of the processed location Pa based on a comparison between the "High / Low matrix" and multiple feature values ​​Dfs obtained by imaging the processed location Pa. This configuration allows the device 10 to identify the failure mode Ddm and the failure factor Ddf of the processed location Pa. The set of defect features Dfds is stored in the RAM 620. In step S450 of FIG. 15, the process of identifying the failure mode Ddm and the cause of the failure Ddf of the processed portion Pa is realized by the determination unit 614 as a functional unit of the processor 610.

[0081] FIG. 16 is a table showing a "High / Low matrix" based on the feature value Df in the low temperature range. FIG. 16 is a table created based on a processed area of ​​the same type as the processed area Pa. For example, consider a case where the multiple feature values ​​Dfs of the processed area Pa are judged to be defective areas in the time intervals Dti10, Dti11, and Dti12 except for the brightness range of 255. In this case, in FIG. 16, the failure mode Ddm and the cause of the failure Ddf for which the pass / fail judgment result Dr matches can be identified as an insufficient area due to a fume blow abnormality.

[0082] However, in the "High / Low matrix" of Fig. 16, the pass / fail judgment results Dr for the burn-through due to "large workpiece displacement" and the penetration due to "large output" are the same. In other words, in the "High / Low matrix" based on the feature amount Df in the low temperature range, the judgment unit 614 cannot identify these failure modes Ddm.

[0083] FIG. 17 is a table showing a "High / Low matrix" based on the feature value Df in the high-temperature range. Similar to FIG. 16, FIG. 17 is a table created based on the same type of machining location as the machining location Pa. In FIG. 17, the pass / fail judgment results Dr for the burn-through due to "large workpiece misalignment" and the penetration due to "high output" are different. Specifically, the burn-through due to "large workpiece misalignment" indicates a defective location only in the time interval Dti22 in the brightness range of 0 to 255. On the other hand, the penetration due to "high output" indicates a defective location in all brightness ranges in the time interval Dti. Therefore, by using the "High / Low matrix" based on the feature value Df in the high-temperature range, the judgment unit 614 can identify these failure modes Ddm and failure factors Ddf.

[0084] C. Other Embodiments: In the above embodiment, the laser processing is performed by forming a keyhole to join the workpieces. That is, the laser processing is performed by keyhole-type laser processing. However, the laser processing is not limited to keyhole-type laser processing. The laser processing may be heat conduction-type laser processing. Heat conduction-type laser processing does not cause metal evaporation due to its low energy density. Therefore, keyholes are not generated in heat conduction-type laser processing. Heat conduction-type laser processing melts metal by propagating energy into the metal. However, even in the heat conduction-type laser processing, the low-temperature measuring camera 300 and the high-temperature measuring camera 400 can generate a temperature distribution image Im. Therefore, since multiple feature values ​​Dfs can be acquired, the above embodiment can be applied to the processed area Pa by heat conduction-type laser processing.

[0085] In the above embodiment, the device 10 is configured with two cameras, the low-temperature measuring camera 300 and the high-temperature measuring camera 400. However, the device 10 may be configured with one temperature measuring camera, or may be configured with three or more temperature measuring cameras. For example, the device 10 may be configured with one temperature measuring camera capable of generating images of both low-temperature and high-temperature ranges. The device 10 may also be configured with three cameras, one capable of generating images of the low-temperature range and two temperature measuring cameras that generate images of different high-temperature ranges.

[0086] In the above embodiment, the low temperature range, which is the temperature measurement range by the low temperature measurement camera 300, is a temperature range of ±200 K based on the melting point of the workpiece Ob. However, this temperature range may be any range that includes the melting point of the workpiece Ob. The temperature range may also be a temperature range other than ±200 K based on the melting point of the workpiece Ob.

[0087] In the above embodiment, the high temperature range, which is the range of temperature measurement by the pyrometric camera 400, is a temperature range of 400 K, with the upper limit of the low temperature range as its lower limit. However, this temperature range may be any range that includes a temperature range higher than the upper limit of the temperature range including the melting point of the workpiece Ob. For example, this temperature range may partially overlap with the temperature range of the low temperature range, or may be separated from it. This temperature range may also be in contact with the temperature range of the low temperature range so as to be continuous with it.

[0088] In the above embodiment, the time interval Dti for which averaging processing is performed is four intervals based on the welding stage. However, the division of the time interval Dti is not limited to these four intervals. The time interval Dti may be divided into two or more intervals. The time interval Dti may be changed to make it easier to acquire the feature value Df. For example, the time interval Dti may be determined experimentally based on processing phenomena such as the keyhole formation time and the molten pool solidification time.

[0089] In the above embodiment, the brightness ranges are classified into three types: an overall brightness range of 0 to 255, a peripheral brightness range of 0 to 254, and a central brightness range of only 255. However, the types of brightness ranges are not limited to these three types. The brightness range may be determined based on a brightness value that represents the characteristics of the processed area Pa.

[0090] In the above embodiment, the averaging process may be a process of obtaining the average value of the sum Dl of all luminance values ​​included in each image, or a process of obtaining the average value of the sum Dl of luminance values ​​included in each image within a predetermined range. For example, the predetermined range may be a range excluding the maximum luminance value.

[0091] In the above embodiment, the pass / fail judgment was performed using a plurality of feature quantities Dfs as a set of average values ​​from the temperature distribution image Im captured by the temperature measurement camera. That is, these feature quantities Dfs are obtained from an image whose brightness range is unchanged. However, the plurality of feature quantities Dfs do not necessarily need to include feature quantities Dfs obtained from an image whose brightness range is unchanged. For example, in this embodiment, the plurality of feature quantities Dfs are obtained from a combination of a brightness range of 0 to 255, a brightness range of 0 to 254, and a brightness range of 255 only. However, the plurality of feature quantities Dfs may also be obtained from a combination of a brightness range of 0 to 254 and a brightness range of 255 only. This configuration makes it easier for differences to appear in the pass / fail judgment result Dr compared to when the pass / fail judgment is performed based on a comparison that also includes a plurality of feature quantities Dfs in the brightness range of 0 to 255.

[0092] In the above embodiment, the pass / fail judgment is based on ±3σ of the average value Dfa of the pass / fail products. This average value Dfa of the pass / fail products is an average value determined from the feature values ​​Df of multiple pass / fail products. However, the range of the judgment criteria may be changed based on the actual variation of the pass / fail products. For example, the judgment criteria may be set to the feature value Df obtained from a single pass / fail product. Alternatively, the judgment criteria may be set to a range smaller than 3σ.

[0093] In the above embodiment, the "High / Low matrix" is created based on the feature value Df obtained in advance from the inspection of the processed area Pa. However, the "High / Low matrix" may also be created from the feature value Df determined by estimation based on the processing phenomenon. For example, if the laser output is insufficient, it can be estimated that the temperature will be lower than that of a good product. Therefore, as in this embodiment, the "High / Low matrix" for the cause of the "low output" defect can be created to include the judgment result L.

[0094] The present disclosure is not limited to the above-described embodiments and can be realized in various configurations without departing from the spirit of the present disclosure. For example, the technical features of the embodiments corresponding to the technical features in each aspect described in the Summary of the Disclosure section can be appropriately replaced or combined to solve some or all of the above problems or achieve some or all of the above effects. Furthermore, if a technical feature is not described as essential in this specification, it can be deleted as appropriate. [Explanation of symbols]

[0095] 10...apparatus, 100...laser oscillator, 110...optical isolator, 200...laser scanner, 210...controller, 220...first dichroic mirror, 230...light collecting system, 240...protective glass, 250...second dichroic mirror, 260...first reflecting mirror, 270...optical filter, 300...low temperature measuring camera, 400...high temperature measuring camera, 500...good / bad indicator, 600...computer, 610...processor, 611...camera control unit, 612... Brightness value adjustment unit, 613... brightness value calculation unit, 614... judgment unit, 620... RAM, 630... ROM, Am10... arrow, Am20... arrow, Ddf... cause, Ddm... failure mode, Df... feature amount, Dfa... average value of non-defective products, Dfas... set of average values ​​for non-defective products, Dfds... set of defect feature amounts, Dfas1... set of first average values ​​for non-defective products, Dfas2... set of second average values ​​for non-defective products, Dfs... multiple feature amounts, Dfs1... set of first average values Dfs2...set of second average values, Dl...sum of brightness values, Dr...result of pass / fail judgment, Dt...time of shooting, Dti...time interval, Dti10...time interval, Dti11...time interval, Dti12...time interval, Dti13...time interval, Dti20...time interval, Dti21...time interval, Dti22...time interval, Dti23...time interval, Im...image, Im10...first type image, Im11...first image, Im12...second image, Im20...second type image, Im2 1...3rd image, Im22...4th image, Im30...3rd type image, Im31...5th image, Im32...6th image, Ir10...light, Ir100...light, Ir110...light, Ir120...light, Ir121...light, LB10...laser light, Mp1...area, Mp2...area, Mp3...area, Mp4...area, Ob...workpiece, Ob1...workpiece, Ob2...workpiece, P1...working location, P2...working location, P3...working location, P4...working location, Pa...working location, Sc...scale

Claims

1. A method for determining whether laser processing is good or bad, comprising: (a) a step of capturing an image of a processed portion of the laser processing using one or more temperature measurement cameras to generate a first type of image, the first type of image including a first image representing a temperature distribution in a first temperature range including a melting point of a material constituting the processed portion, and a second image representing a temperature distribution in a second temperature range including a temperature range higher than an upper limit value of the first temperature range; (b) performing the step (a) a plurality of times while the laser processing is being performed on one of the processing locations; (c) performing an averaging process for the first images and the second images of the plurality of sets of first type images obtained in the process (b), to obtain a set of first average values ​​by obtaining an average value of sums of luminance values ​​of pixels having luminance values ​​within a predetermined range among pixels included in each of the first images and the second images for each time interval including the shooting time of each image; (d) determining whether the processed portion is good or bad based on a comparison between the set of first average values ​​as a plurality of feature amounts and a set of first average values ​​of a predetermined good product as a reference plurality of feature amounts, and outputting the result of the determination; A method for determining whether laser processing is good or bad, including:

2. 2. The method for determining whether laser processing is good or bad according to claim 1, In the step (c), A method for determining the quality of laser processing, wherein the brightness values ​​of pixels contained in each of the first image and the second image that have brightness values ​​within a predetermined range are within the range of all values ​​that the brightness values ​​of pixels contained in each of the first image and the second image can take.

3. 2. The method for determining whether laser processing is good or bad according to claim 1, In the step (c), A method for determining the quality of laser processing, wherein the brightness values ​​of pixels contained in each of the first image and the second image that have brightness values ​​within a predetermined range are in a range that does not include the maximum possible brightness values ​​of pixels contained in each of the first image and the second image.

4. The method for determining whether laser processing is good or bad according to claim 3, further comprising: (e) performing an averaging process for the first images and the second images of the plurality of sets of first type images obtained in the process (b), for each of the time periods including the photographing time of each image, to obtain an average value of sums of luminance values ​​of pixels included in each of the first images and the second images, the sums having luminance values ​​in another predetermined range, and acquiring a set of second average values, wherein the other predetermined range of luminance values ​​includes the entire range of luminance values ​​of pixels included in each of the first images and the second images, The step (d) is a step of determining whether the processed area is good or bad based on a comparison between the set of second average values ​​as the plurality of feature quantities and a set of second average values ​​of a predetermined good product as a reference plurality of feature quantities, and outputting the result of the determination.

5. A method for determining whether laser processing is good or bad according to any one of claims 1 to 4, Furthermore, if the processed part is defective, (f) A method for determining the quality of laser processing, comprising a step of identifying the type of defect in the processed area based on a comparison between a set of defect features as the plurality of feature values ​​associated in advance with the type of defect and a plurality of feature values ​​including at least one of the set of first average values ​​and the set of second average values ​​as the plurality of feature values ​​obtained by imaging the processed area, and outputting the identified type of defect.

6. 6. The method for determining whether laser processing is good or bad according to claim 5, the set of defect features further includes a cause of the defect associated with at least one of the type of defect and the set of defect features associated with the type of defect; The step (f) is a method for determining the quality of laser processing, and includes a step of identifying a cause of the defect in the processed area based on a comparison between the set of defect features and a plurality of feature values ​​including at least one of the set of first average values ​​and the set of second average values ​​as the plurality of feature values ​​obtained by imaging the processed area, and outputting the identified cause of the defect.

Images