Method for inspecting surface of metal material, device for inspecting surface of metal material, metal material, method for manufacturing metal material, method for managing quality of metal material, and facility for manufacturing metal material

Abstract

To provide a method for inspecting a surface of a metal material and a device for inspecting a surface of a metal material that can detect a defect in a surface of a metal material accurately without omission.SOLUTION: A method for inspecting a surface of a metal material according to the present invention is to detect a defect in a surface of a metal material, from a plurality of images obtained by picking up images of light radiated to and then reflected from the surface of the metal material, in two or more different wavelength ranges, and the method for inspecting a surface of a metal material includes: a first detection step of creating an image of the difference or ratio in the luminance between at least two images obtained from a predetermined area relative to the surface of the metal material, and detecting a first defect present on the surface of the metal material from the created image; a second detection step of extracting a place brighter than a peripheral area from one image or an image obtained by combining a plurality of images obtained from the predetermined area relative to the surface of the metal material, and thereby detecting a second defect present on the surface of the metal material; and a determination step of detecting a defect by using the first defect and the second defect.SELECTED DRAWING: Figure 13

Description

[Technical field]

[0001] The present invention relates to a surface inspection method for metallic materials for optically detecting defects on the surface of a metallic material, a surface inspection device for metallic materials, a metallic material, a manufacturing method for metallic materials, a quality control method for metallic materials, and manufacturing equipment for metallic materials. [Background technology]

[0002] In recent years, in the manufacturing process of metal materials, particularly steel products, it is required to detect surface defects of steel materials in hot or cold from the viewpoint of improving yield by preventing large-scale nonconformity. The steel materials mentioned here refer to steel products, including seamless steel pipes, welded steel pipes, hot-rolled steel sheets, cold-rolled steel sheets, thick plates, and other steel plates and shaped steel, and semi-finished products such as slabs generated in the manufacturing process of these steel products. For this reason, a method has been proposed as a method for detecting surface defects of steel materials, in which light is irradiated onto a billet in the manufacturing process of seamless steel pipes, the reflected light is received, and the presence or absence of surface defects is determined based on the amount of reflected light (see Patent Document 1). In addition, a method has also been proposed in which visible light of multiple wavelength ranges that do not mutually affect the spontaneous emission emitted from hot steel materials and do not affect each other is irradiated from mutually symmetrical oblique directions with respect to the normal line of the hot steel material surface, an image by the composite reflected light and an image by each reflected light are obtained in the normal direction of the hot steel material surface, and surface defects are detected from a combination of these images (see Patent Document 2). In addition, a method has been proposed in which differentiable illumination light is applied from mutually symmetrical inclined directions relative to the normal to the steel surface, and an arrangement of light and dark areas corresponding to concave shapes is extracted from the difference image between two images taken of the inspection target area illuminated from each direction, thereby detecting surface defects (see Patent Document 3). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 11-37949 [Patent Document 2] Japanese Patent Application Publication No. 59-52735 [Patent Document 3] Patent No. 6079948 Summary of the Invention [Problem to be solved by the invention]

[0004] According to the method described in Patent Document 1, since the reflectance of scale and harmless patterns is different from that of the base steel portion, there is a possibility that scale and harmless patterns occurring in a sound part that is not a surface defect may be erroneously detected as a surface defect. For this reason, in the method described in Patent Document 1, the surface defect (surface flaw) of the billet is distinguished from the surface defect by utilizing the fact that the shape of the surface defect (surface flaw) is linear. However, the surface defects of steel materials are not limited to linear ones, and there are various shapes such as circular ones. For this reason, it is difficult to apply the method described in Patent Document 1 to the detection process of surface defects. On the other hand, in the method described in Patent Document 2, since there are a huge number of types of surface defects, scale, harmless patterns, etc., it is difficult to distinguish between the scale and harmless patterns and the surface defect by simply combining images. In addition, it is practically difficult to construct a detection logic that can handle a huge number of image combinations. In addition, according to the method described in Patent Document 3, the above problem can be solved and concave defects on the steel material surface can be detected with high accuracy. However, surface defects of steel materials are not limited to concave ones, and defects that do not appear as concaves or convexes from the surface (hereinafter referred to as non-open defects) can also occur when rolling is performed with a foreign object attached thereto or when scale is formed on a concave defect. For this reason, it is difficult to detect all surface defects of steel materials using only the method described in Patent Document 3.

[0005] The present invention has been made in view of the above problems, and an object of the present invention is to provide a surface inspection method and a surface inspection device for metal materials that are capable of detecting defects on the surface of the metal material with high accuracy without omission. Another object of the present invention is to provide a high-quality metal material that is free of surface defects, a manufacturing method for a metal material, a quality control method for a metal material, and a manufacturing facility for a metal material. [Means for solving the problem]

[0006] [1] The surface inspection method for metallic materials according to the present invention is a surface inspection method for metallic materials that detects defects on the surface of a metallic material from multiple images obtained by capturing reflected light from the surface of a metallic material by irradiating light in two or more different wavelength bands, and includes the following steps: a first detection step of generating an image of the difference or ratio in brightness between at least two images obtained from a specified area on the surface of the metallic material, and detecting a first defect present on the surface of the metallic material from the generated images; a second detection step of detecting a second defect present on the surface of the metallic material by extracting a bright area compared to the surrounding area from an image obtained by combining one image or multiple images obtained from the specified area on the surface of the metallic material; and a determination step of detecting the defect using the first defect and the second defect.

[0007] [2] In the method for inspecting a surface of a metallic material according to the present invention, in the method for inspecting a surface of a metallic material according to [1], the determination step determines the first defect or the second defect to be a harmful defect.

[0008] [3] The method for inspecting a surface of a metal material according to the present invention is the method for inspecting a surface of a metal material according to [1] or [2], wherein at least one of the two or more different wavelength bands is a wavelength band of 500 nm or less.

[0009] [4] In the method for inspecting a surface of a metal material according to the present invention, which is any one of the methods for inspecting a surface of a metal material [1] to [3], the wavelength band of at least one image used in the first detection step is the same as the wavelength band of the image used in the second detection step.

[0010] [5] The method for inspecting a surface of a metal material according to the present invention is any one of the methods for inspecting a surface of a metal material according to [1] to [4], wherein the first defect is a pattern defect and the second defect is a red scale peeling defect.

[0011] [6] The surface inspection device for metal materials according to the present invention is a surface inspection device for metal materials that detects defects on the surface of a metal material from a plurality of images obtained by capturing light reflected from the surface of a metal material by irradiating light in two or more different wavelength bands, and includes a first detection means that generates an image of the difference or ratio in brightness between at least two images obtained from a predetermined area on the surface of the metal material, and detects a first defect present on the surface of the metal material from the generated image; The system further includes a second detection means for detecting a second defect present on the surface of the metal material by extracting a bright area compared to the surrounding area from an image obtained from the specified area on the surface of the metal material, and a determination means for detecting the defect using the first defect and the second defect.

[0012] [7] The surface inspection apparatus for metallic materials according to the present invention is the surface inspection apparatus for metallic materials according to [6], wherein the determination means determines the first defect or the second defect to be a harmful defect.

[0013] [8] The surface properties of the metallic material according to the present invention are guaranteed using any one of the surface inspection methods for metallic materials [1] to [5].

[0014] [9] The method for manufacturing a metallic material according to the present invention comprises inspecting the surface properties of a metallic material using any one of the methods for inspecting a metallic material surface among [1] to [5], and manufacturing the metallic material based on the inspection results.

[0015]

[10] The quality control method of the present invention involves inspecting the surface properties of a metallic material using any one of the surface inspection methods for metallic materials [1] to [5], and controlling the quality of the metallic material based on the inspection results.

[0016]

[11] The manufacturing equipment for metal materials according to the present invention is provided with the surface inspection device for metal materials according to [6] or [7]. Effect of the Invention

[0017] The surface inspection method and surface inspection device for metallic materials according to the present invention can accurately detect defects on the surface of a metallic material without omission. Also, the metallic material according to the present invention can provide a high-quality metallic material with no surface defects, a manufacturing method for a metallic material, a quality control method for a metallic material, and a manufacturing facility for a metallic material. [Brief description of the drawings]

[0018] [Figure 1] FIG. 1 is a diagram showing an example of a pattern defect and a harmless pattern. [Diagram 2] FIG. 2 is a schematic diagram showing the configuration of the device used in the test. [Diagram 3] FIG. 3 is a diagram showing the results of a comparison of the relationship between signal intensity and wavelength for a pattern defective portion and a healthy portion. [Figure 4] FIG. 4 is a diagram showing an example of a difference image. [Diagram 5] FIG. 5 is a diagram showing the results of a comparison of the relationship between the incident angle of illumination light, the signal intensity difference between a pattern-like defective portion and a healthy portion, and the wavelength. [Figure 6] FIG. 6 is a diagram showing a non-opening defect and a difference image of a non-opening defect. [Figure 7] FIG. 7 shows a non-opening defect in which the red scale has peeled off. [Figure 8] FIG. 8 is a diagram showing a red scale peeling defect and a difference image of the red scale peeling defect. [Figure 9] FIG. 9 is a diagram showing images of a red scale peeling defect captured by the R channel, G channel, and B channel of a color camera. [Figure 10] FIG. 10 is a diagram showing a non-opening defect. [Figure 11] FIG. 11 is a schematic diagram showing the configuration of a surface inspection device for a metallic material according to the first embodiment of the present invention. [Figure 12] FIG. 12 is a schematic diagram showing the configuration of a surface inspection device for a metallic material according to the second embodiment of the present invention. [Figure 13] FIG. 13 is a flowchart showing the flow of a surface inspection process according to an embodiment of the present invention. [Figure 14] FIG. 14 is a flowchart showing a process of the first detection step shown in FIG. [Figure 15] FIG. 15 is a flowchart showing a process of the second detection step shown in FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0019] In thick steel plates, which are a type of steel product, the surface is covered with an oxide film called black scale, and patterns called harmless patterns that do not affect the quality of the steel product may occur. In addition, pattern defects such as scabs also occur on the surface of thick steel plates, but as shown in FIG. 1, it is difficult to distinguish between pattern defects and harmless patterns based on the difference in brightness of the captured images. Therefore, the inventors of the present invention focused on the fact that minute red scales (Fe2O3) are likely to be generated on pattern defects, and that the pattern defects appear reddish due to this red scale, and conducted a test to detect pattern defects using spectral reflection characteristics. The configuration of the device used in the test is shown in FIG. 2. As shown in FIG. 2, in this test, a thick steel plate sample SA having a pattern defect was placed on a linear stage 1, illumination light L having a wide band of wavelengths was irradiated from a xenon light source 2 to the surface of the thick steel plate sample SA, and spectral images of each wavelength were captured using a spectroscopic camera 3 having a one-dimensional field of view.

[0020] Figure 3 shows the results of comparing the relationship between the signal intensity and wavelength of the pattern defect and healthy parts using spectral images. As shown in Figure 3, the signal intensity characteristics on the short wavelength side and the long wavelength side are relatively different between the pattern defect and healthy parts. This suggests that it may be possible to detect pattern defects that are difficult to detect due to the signal of healthy parts using only a single wavelength by comparing the signal intensity between different wavelengths. Here, an example is shown in which a pattern defect in a thick steel plate is detected by the difference between spectral images of two different wavelengths. The pattern defect in a thick steel plate has a form that is pressed onto the surface of the thick steel plate during the manufacturing process of the thick steel plate, and there are no uneven parts when viewed from the surface of the thick steel plate. For this pattern defect, after performing brightness correction so that the average brightness value is constant for the spectral image of the 415 nm wavelength band (415 nm image) and the spectral image of the 750 nm wavelength band (750 nm image), an image obtained by taking the difference in brightness values ​​of the two spectral images is shown in Figure 4. As shown in Figure 4, the signals of healthy parts such as black scale are canceled by the difference, and only the signals of the pattern defect are emphasized.

[0021] It was confirmed that the pattern defect can be detected with high accuracy by utilizing the characteristics of the spectral reflectance spectra of the pattern defect and the healthy portion to generate a difference image between a spectral image captured with light on the long wavelength side and a spectral image captured with light on the short wavelength side. Depending on the manufacturing conditions of the thick steel plate, a surface scale having a surface reflectance spectrum similar to that of the pattern defect may be generated in the healthy portion. However, even in such a case, the pattern defect can be detected with high accuracy while suppressing overdetection by discriminating the pattern defect extracted from the difference image using a judger created by a machine learning technique.

[0022] From the above, it was confirmed that pattern defects with different spectral reflectance characteristics from healthy areas can be detected with high accuracy by comparing luminance information between multiple different wavelengths. Figure 5 also shows the results of comparing the relationship between the incidence angle of illumination light L (angle with respect to the surface normal vector of the thick steel plate), the signal intensity difference between pattern defects and healthy areas, and the wavelength. The angles of incidence listed in the legend of Figure 5 as 20°, 30°, etc., are the incidence angles. As shown in Figure 5, the larger the incidence angle of illumination light L, the larger the signal intensity difference between pattern defects and healthy areas. Therefore, in order to efficiently detect the spectral reflectance characteristics of the target, i.e., the difference in color, it is advisable to increase the incidence angle of illumination light L.

[0023] The two spectral images used to generate the above-mentioned difference image may be captured by two of the three channels of a color camera. In this way, a single color camera can simultaneously detect pattern defects and red scale peeling defects. That is, a difference image is generated using an image captured by the B channel or G channel and an image captured by the R channel to detect pattern defects. On the other hand, the image captured by the B channel or G channel is used to detect red scale peeling defects. Here, as the wavelength of the captured image used for detecting red scale peeling defects, that is, the second detection step described later, it is preferable not to use a wavelength of 600 nm or more, and it is even more preferable to use a wavelength of 550 nm or less. It is more preferable to use a green (wavelength 490 to 550 nm) or blue (wavelength 430 to 490 nm) captured image, and it is most preferable to use a blue (wavelength 430 to 490 nm) captured image.

[0024] Although the above description has been made on pattern defects in thick steel plates, the present invention can be applied to the detection of other defects that occur on the surface of metal materials and have different spectral reflectance characteristics from healthy parts. In addition, the above description has been made on thick steel plates, but it goes without saying that the present invention can also be applied to other steel materials in which red scale can occur, such as slabs, hot-rolled steel plates, welded steel pipes, electric resistance welded pipes, seamless steel pipes, steel sections, bar stock, and wire stock. Furthermore, the present invention can be similarly applied to any metal product that generates red rust on its surface, similar to steel products.

[0025] Incidentally, among the non-opening defects in which no unevenness is apparent from the surface due to the adhesion of foreign matter or the generation of scale as shown in FIG. 6(a), there are some in which the red scale attached to the surface during the transportation of the steel material has peeled off together with the black scale. In FIG. 6(a), the symbols BM indicate the base steel, MS indicate the black scale, RS indicate the red scale, and DI indicate the interface of the non-opening defect, and FIG. 6(b) shows an example of a difference image of the non-opening defect shown in FIG. 6(a). Among such defects, those in which the interface of the non-opening defect is exposed and forms a concave shape as shown in FIG. 7 can be detected using, for example, the method described in Patent Document 3. However, it is difficult to detect those in which only the red scale and black scale of the surface layer have peeled off and foreign matter EP and the like remain in the concave portion, resulting in almost no unevenness and no exposure of the interface DI of the non-opening defect. This is because such defects are often not reddish compared to the surrounding area and do not appear bright in the difference image. In the following, for convenience, such a non-opening defect in which the red scale has peeled off and there is almost no unevenness will be referred to as a red-scale peeling defect. Figure 8(b) shows an example of a difference image of a red-scale peeling defect.

[0026] Here, images of such a red scale peeling defect captured by the R (red) channel, G (green) channel, and B (blue) channel of a color (RGB) camera are shown in Fig. 9(a) to (c), respectively. Fig. 9(a) is an image captured by the R (red) channel, Fig. 9(b) is an image captured by the G (green) channel, and Fig. 9(c) is an image captured by the B (blue) channel. As shown in Fig. 9(a) to (c), since the reflectance of the red scale peeling defect part is high, the red scale peeling defect part is captured brightly in the image in all channels of the R channel, G channel, and B channel, and is particularly noticeable in the images captured by the G channel and B channel. Therefore, the red scale peeling defect is detected by extracting the part that appears brighter than the surroundings from the image captured by the color camera, rather than using the difference image.

[0027] The reason for the low contrast of the image captured with the R channel shown in Fig. 9(a) is that some thick steel plates serving as base materials have surfaces that appear reddish due to red scale adhering all over, and these surfaces appear bright overall in the image captured with the R channel, so even if red scale peeling defects appear bright, they will blend in with the texture. For this reason, it is preferable to detect red scale peeling defects by extracting areas that appear brighter than their surroundings from the images captured with the B channel in Fig. 9(b) and the G channel in Fig. 9(c).

[0028] Also, the images captured in the B channel and the G channel may be used directly, or an image obtained by combining the pixels of the images captured in the B channel and the G channel with a specific weight may be used. Here, an image obtained by combining the images captured in the B channel, the G channel, and the R channel with a specific weight may be used, but it is preferable to make the weight of the image captured in the R channel small for the reasons described above. Even in the case of a surface that appears reddish as described above, as shown in FIG. 10, for a non-opening defect in which the red scale RS has not peeled off, a red scale RS2 is further generated on the red scale RS, so that the defect can be detected with high accuracy by the difference image.

[0029] Furthermore, the base steel portion is also exposed when the surface of the steel material rubs against structures such as transport rolls, and is called a scratch. Such scratches occur on the hard surface of the steel material when it is cooled and has sufficient strength, and are often not deep. As the surface is rubbed, the base steel portion is exposed, similar to red scale peeling defects, and they have a unique shape that differs from red scale peeling defects. Even when such scratches are present, it is possible to extract only harmful defects using a classifier created by machine learning techniques.

[0030] Hereinafter, a surface inspection device for metal materials, which is one embodiment of the present invention that has been conceived based on the above-mentioned technical concept, will be described.

[0031] [Surface inspection equipment] [First embodiment] Fig. 11 is a schematic diagram showing the configuration of a surface inspection device for metallic materials according to a first embodiment of the present invention. As shown in Fig. 11, a surface inspection device 10 for metallic materials according to a first embodiment of the present invention is a device for detecting surface defects of a plate-shaped steel material S transported in the direction of the illustrated arrow. The surface inspection device 10 for metallic materials according to the first embodiment of the present invention comprises, as main components, a light source 11, an encoder and pulse generator 12, an area sensor 13 capable of capturing spectral images in a plurality of wavelength bands, an image processing device 14, and a monitor 15.

[0032] The light source 11 irradiates the inspection target portion on the surface of the steel material S with illumination light L in accordance with a trigger signal output by the pulse generator every time a certain number of pulse signals are transmitted from the encoder. The light source 11 is preferably arranged so that the irradiation direction of the illumination light L is inclined within a range of 60° or more and less than 90° with respect to the normal direction of the steel material S surface. This allows surface defects to be detected with high accuracy from the difference image. In addition, although one light source 11 is arranged in this embodiment, multiple light sources 11 may be arranged. In addition, although a xenon light source is adopted as the light source 11 in this embodiment, as long as it contains components of multiple different wavelength bands, light sources having broadband characteristics such as metal halide light sources, halogen light sources, mercury lamps, incandescent lamps, etc., or light sources having specific narrowband characteristics such as LEDs and lasers may be combined. In addition, in cases where the conveying speed of the steel material S is fast, or the positional variation of the steel material S from the determined conveying position (path line) is large, a flash light source or pulse lighting may be used to prevent imaging blur.

[0033] The area sensor 13 captures spectral images of different wavelength bands at approximately the same position of the steel material S. The captured spectral images are preferably coaxial, but may be aligned by image processing. The area sensor 13 may be of the Bayer type, in which filters that transmit different wavelength bands are attached to each element so as to be nested, and then generate multiple images, or of the prism type, in which a prism and multiple elements are used to adjust the elements so as to be coaxial. It is preferable to use an RGB color camera because it is inexpensive, but a multiband area sensor with two or more channels may be used. Furthermore, depending on the spectral reflection characteristics of the healthy part and the surface defect part to be discriminated, a wavelength selection filter may be installed on the optical path in front of the area sensor 13 or the light source 11, etc., to make the color more prominent and improve the detection ability. In addition, when the difference in the spectral reflection characteristics of the healthy part and the surface defect part appears in a narrow band, the wavelength band received by the wavelength selection filter or the like may be narrow if there is a margin of light. In addition, when detecting pattern defects, it is preferable that at least one of the plurality of different wavelength bands includes a narrow wavelength band of 650 nm or more, and at least one of the different wavelength bands includes a narrow wavelength band of 500 nm or less.

[0034] The area sensor 13 captures the spectral images in synchronization with the light source 11 according to a trigger signal output from the pulse generator. The luminance values ​​of each channel of the spectral images are not saturated except for the spectral images in which the end of the steel material S is captured. In addition, although a two-channel area sensor that captures light in the long wavelength band and the short wavelength band is used in this embodiment, a device configuration in which three or more wavelength bands are captured by three or more channel area sensors may be used. In addition, it is preferable that the area sensor 13 receives reflected light so that the light receiving angle with respect to the normal direction of the surface of the steel material S is within a range of 0° or more and less than 20°.

[0035] The image processing device 14 detects surface defects in the inspection target area by performing subtraction processing and extraction processing, which will be described later, between the spectroscopic images input from each channel of the area sensor 13. Then, the image processing device 14 outputs the spectroscopic images input from the area sensor 13, the spectroscopic images after subtraction processing, and information related to the detection results of surface defects to the monitor 15. In this specification, the image processing device 14 also corresponds to a determination means.

[0036] Second Embodiment Fig. 12 is a schematic diagram showing the configuration of a surface inspection device for metal materials according to a second embodiment of the present invention. As shown in Fig. 12, a surface inspection device 20 for metal materials according to the second embodiment of the present invention detects surface defects of a plate-shaped steel material S transported in the direction of the illustrated arrow. The difference from the surface inspection device 10 for metal materials according to the first embodiment is that the light source 11 is a line light source 21, and the area sensor 13 is a line sensor 22. By using the line sensor 22, the field of view is limited to one line, and there is an advantage that the optical conditions are stable compared to the first embodiment, but there is also a disadvantage that the steel material S is vulnerable to positional fluctuations from the transport position (path line) during transport.

[0037] The above-mentioned surface inspection devices 10, 20 for metal materials can also be provided as part of a manufacturing facility for metal materials.

[0038] The metallic material surface inspection apparatuses 10, 20 having such a configuration perform the surface inspection process described below to accurately detect defects on the metallic material surface without omission. The operation of the metallic material surface inspection apparatuses 10, 20 when performing the surface inspection processes according to the first to third embodiments of the present invention will be described below.

[0039] [Surface inspection processing] Fig. 13 is a flowchart showing the flow of the surface inspection process according to an embodiment of the present invention. In this embodiment, a pattern-like defect portion, a red scale peeling defect portion, and a healthy portion having a harmless pattern in the inspection target portion are each discriminated. The surface inspection process shown in Fig. 13 starts when a command to execute the surface inspection process is input to the image processing device 14, and the surface inspection process proceeds to step S21.

[0040] Note that an irradiation step and an imaging step may be performed before the surface inspection process shown in FIG. 13. In this case, first, in the irradiation step, light is irradiated onto the surface of the metal material. Then, in the imaging step, a plurality of images are acquired by imaging the reflected light from the surface of the metal material by the light irradiated in the above-mentioned irradiation step in two or more different wavelength bands. The "raw image" in FIG. 13 to FIG. 15 refers to a plurality of images acquired by imaging in two or more different wavelength bands. The first detection step and the second detection step described later are performed on a plurality of images acquired from a predetermined area on the surface of the metal material. The "raw image" can be acquired, for example, by the above-mentioned imaging step, or can be acquired from an image acquired on another occasion. Here, in the irradiation step, for example, the light source 11, the line light source 21, the encoder, and the pulse generator 12 described in the above-mentioned surface inspection device can be used as the irradiation means. Similarly, in the imaging step, for example, the area sensor 13 or the line sensor 22 described in the above-mentioned surface inspection device can be used as the imaging means.

[0041] The surface inspection process of this embodiment is a combination of a first detection step (processing of steps S21 to S24) and a second detection step (processing of steps S21, S25 to S27). In the first detection step, when an image captured by the R channel and an image captured by the G channel or the B channel of a color camera are used as the spectral images of different wavelength bands, a pattern defect portion and a healthy portion are discriminated. On the other hand, in the second detection step, when an image captured by the B channel or the G channel of a color camera is used as the spectral images of different wavelength bands, a red scale peeling defect portion and a healthy portion are discriminated.

[0042] Finally, the image processing device 14 integrates the results of the first and second detection steps with the image information of the inspection target area to generate information about harmful defects on the surface of the steel material (step S28). This is the final judgment step in the present invention. All defects detected in each detection step can be considered as independent defects on the steel plate. Therefore, simply adding up the information on the pattern defects and the red scale peeling defects will comprehensively detect all defects that have occurred on the steel plate.

[0043] Specifically, if the part is judged to be a defective part (called the first defect) in the first detection step and judged to be a healthy part in the second detection step, the image processing device 14 generates information regarding the part to be inspected as a pattern-like defective part. On the other hand, if the part is judged to be a healthy part in the first detection step and judged to be a defective part (called the second defect) in the second detection step, the image processing device 14 generates information regarding the part to be inspected as a red scale peeling defective part. Here, the second defect indicates different information regarding the defective part compared to the first defect. Also, if the part is judged to be a healthy part in both the first and second detection steps, the image processing device 14 generates information regarding the part to be inspected as a healthy part. Then, the image processing device 14 generates information regarding the pattern-like defective part and the red scale peeling defective part as harmful defects.

[0044] In the judgment step S28, if the pattern defect and the red scale peeling defect occur at the same position, it is highly likely that a part of the single pattern defect peels off and is detected as the red scale peeling defect. In this case, it is preferable to treat the two defects as one defect. Therefore, the occurrence position of the pattern defect and the occurrence position of the red scale peeling defect may be compared, and if the difference is smaller than a predetermined distance, they may be regarded as the same defect. The distance calculation method described here may be any index that can objectively quantify the positional relationship of the defects, such as the distance between the centers of gravity of the defects, the shortest distance between each defective area, the distance between the centers of gravity of the smallest circumscribing rectangle, the shortest distance between the smallest circumscribing rectangle, etc. The pattern defect and the red scale peeling defect that are regarded as the same defect are highly likely to be actual defects, and they may be judged separately from the single defect, the pattern defect, and the red scale peeling defect during the guidance by mapping or the pass / fail judgment process.

[0045] Information on harmful defects on the surface of the steel material is displayed on the monitor 15 as an image showing the distribution of harmful defects, for example. Furthermore, information on harmful defects on the surface of the steel material is used for guidance to workers, for judging whether the steel material S can be shipped and / or whether maintenance is required by collecting information on the server, and for automatically transporting the steel material to a treatment process such as maintenance. The image showing the distribution of harmful defects displayed on the monitor 15 can be independently mapped on a map using information on the occurrence positions of harmful defects on the steel plate obtained in each detection step, thereby making it possible to visualize the occurrence status of harmful defects. In this case, if it is desired to distinguish and visualize the pattern defects and the red scale peeling defects, the markers on the map may be different. In addition, when the pass / fail judgment of the product is made using the information on the generated harmful defects, the information on the harmful defects generated in each detection step may be totaled and counted to make a judgment. Specifically, if the allowable number of generated defects is 5 or less and 3 pattern defects and 4 red scale peeling defects are detected, the total number of detected defects is 7, and it may be judged as a failure.

[0046] In this way, in the surface inspection process according to the embodiment of the present invention, by combining the first detection step and the second detection step, pattern-like defects, red scale peeling defects, and healthy areas with harmless patterns are discriminated, and the accuracy of defect detection is further improved. Note that the feature amounts obtained by the first detection step and the second detection step for the same surface defect candidate area may be combined to determine whether the area is a surface defect area or a healthy area by machine learning. Also, all images of surface defect candidate areas extracted by either the first or second detection step may be labeled as surface defect areas or healthy areas, and a classifier may be created using a convolutional neural network.

[0047] [First detection step] FIG. 14 is a flow chart showing the flow of the first detection step shown in FIG. 13. In this embodiment, a pattern-like defect part in the inspection target part is discriminated from a healthy part having a harmless pattern. The pattern-like defect described here is a defect in which there is no apparent unevenness from the surface, such as a foreign object being pressed against the surface or a scale being generated on a concave defect. The healthy part having a harmless pattern means a part having a surface coating or surface properties with different optical properties from the base steel part having a thickness of several to several tens of μm, such as black scale, and is a part that becomes a noise factor in the surface inspection process. The first detection step shown in FIG. 14 is started at the timing when an execution command for the first detection step is input to the image processing device 14, and the surface inspection process, which is the first detection step, proceeds to the processing of step S1.

[0048] In the process of step S1, if the positions of the spectral images of different wavelength bands are shifted in pixel units, the image processing device 14 performs a positioning process. If the spectral images cannot be captured coaxially, the spectral images of different wavelength bands need to be aligned. The alignment method varies depending on the form of the positional shift of the spectral images of each wavelength band, and it is preferable to perform processes such as parallel movement, linear conversion, and one-to-one correspondence of pixels as necessary. However, the positions do not have to be exactly the same. Positional shifts within a range in which the processes of steps S14, S24, S27, and S28 can be performed are included in the "same position" of the present invention as an allowable range. Similarly, positional shifts with respect to the "predetermined area" to be processed are also allowed within a range in which the processes of steps S14, S24, S27, and S28 can be performed. This completes the process of step S1, and the surface inspection process proceeds to the process of step S2.

[0049] In the process of step S2, the image processing device 14 executes a first pre-processing such as correction for making the average brightness value constant for a plurality of spectral images of different wavelength bands, correction of brightness unevenness, normalization of signal intensity, etc. With this, the process of step S2 is completed, and the surface inspection process proceeds to the process of step S3. The processes of steps S1 and S2 correspond to the correction process S21 shown in FIG.

[0050] In the process of step S3, the image processing device 14 compares a plurality of spectral images of different wavelength bands, and generates a composite processed image in which only the pattern defects are emphasized by utilizing the difference in the spectral reflection characteristics. Specifically, the image processing device 14 selects two spectral images having a large difference in the spectral reflection characteristics, and generates a composite processed image by calculating the difference or ratio of the luminance of the two spectral images. A case in which a difference image of two spectral images is generated as a composite processed image will be described. In this case, the image processing device 14 calculates a luminance value Id1(x,y) of the differential image Id1 by subtracting a luminance value Ir(x,y) of the second spectral image Ir (e.g., a luminance value of a second channel having a sensitivity characteristic corresponding to a wavelength band on the long wavelength side) from a luminance value Ib(x,y) of the first spectral image Ib (e.g., a luminance value of a first channel having a sensitivity characteristic corresponding to a wavelength band on the short wavelength side), as shown in the following formula (1).

[0051]

number

[0052] The spectral images Ib and Ir are images with a pixel count of X×Y. In addition, in an xy coordinate system with two orthogonal axes set for each of the spectral images Ib and Ir, the x coordinate is 1≦x≦X and the y coordinate is 1≦y≦Y. In this example, a difference image is generated as the composite processed image, but an image of a luminance value ratio using the following formula (2) may be generated as the composite processed image. Furthermore, a threshold process may be performed separately on the luminance values ​​of the two spectral images, and then an AND process may be performed to generate a composite processed image.

[0053]

number

[0054] Although this process can be performed with only two spectral images, the following methods (a) to (c) are particularly effective for three or more spectral images. This completes the process of step S3, and the surface inspection process proceeds to step S4. This process of step S3 corresponds to the synthesis process S22 shown in FIG.

[0055] (a) Spatial transformation method This is an effective method for 3-channel, particularly color images, and generates an image that extracts the effects of spectral reflectance characteristics (hue) by converting the color space. The color space after conversion can be HSV or HLS, and the spectral reflectance information appears in the hue information. Alternatively, it can be converted to XYZ, L*u*v*, or L*a*b* space, and the luminance value of each color image can be normalized and then threshold processing can be applied.

[0056] (b) Statistical multivariate analysis This is a method in which the luminance value of each wavelength band of each pixel is treated as a feature vector, and parts with different spectral reflectance characteristics are extracted by statistical multivariate analysis. As an example, the target inspection area is divided into sufficiently large parts, and in each area, the luminance value of each wavelength band of all pixels is treated as a feature vector to perform principal component analysis (PCA), and the image is reconstructed using the Mahalanobis distance of each pixel as the representative value of that pixel. Since the pattern-like defective part has a different color, it is expected that the Mahalanobis distance will be larger than that of the healthy part. Although principal component analysis has been explained here, the same effect can be obtained by calculating the deviation from the model in the same way using a Gaussian mixture model, independent component analysis, or regression model.

[0057] (c) Machine learning methods It is the same as the statistical multivariate analysis method up to the point where the luminance value of each wavelength band of each pixel is made into a feature vector, but in this method, each feature vector is pre-supervised to determine whether it is a pattern-like defect or a healthy part, and a judger is created by general supervised learning to judge whether each pixel is a pattern-like defect or a healthy part. As a method for clustering the feature, k-means, kernel method, decision tree method, Gaussian mixture model, regression model, etc. may be used. When using this method, threshold processing described later is not necessary, and pixels that are defect candidates can be directly extracted.

[0058] In the process of step S4, the image processing device 14 generates an image in which surface defects are emphasized by performing a second pre-processing using a frequency filter or the like on the composite processed image. This completes the process of step S4, and the surface inspection process proceeds to the process of step S5.

[0059] In the process of step S5, the image processing device 14 generates a binary image by performing threshold processing on the luminance value of the image obtained by the process of step S4. This completes the process of step S5, and the surface inspection process proceeds to the process of step S6. The processes of steps S4 and S5 correspond to the threshold processing S23 shown in FIG.

[0060] In the process of step S6, the image processing device 14 performs processes such as expansion and contraction on the binarized image to remove connected and isolated points as necessary, and then performs a labeling process in which adjacent pixels are regarded as blobs (lumps) and labeled. The image processing device 14 then regards the blobs extracted by the labeling process as surface defect candidates. This completes the process of step S6, and the surface inspection process proceeds to the process of step S7.

[0061] In the process of step S7, the image processing device 14 judges the type and grade (including harmlessness) of the surface defect for each surface defect candidate portion obtained by the process of step S6. The judgment method may be a judgment method in which judgment rules are manually determined, or may be automatically generated by a general machine learning method using feature quantities such as a regression model (linear regression, logistic regression, multiple regression, support vector machine, nonlinear kernel, etc.), a decision tree model, a random forest, a Bayesian estimation model, a Gaussian mixture model, or a method in which these models are boosted. In this case, by adding each representative value (average value, maximum value) of the luminance values ​​of a plurality of spectral images and a value that allows each representative value to be compared (sum, difference, ratio), etc. to the feature quantity, judgment that takes into account the spectral reflectance characteristics is possible. In addition, if a sufficient number of data N is obtained, a convolutional neural network may be used. In addition, a plurality of machine learning methods may be combined. The judgment result is used in the process of step S28 to be integrated with the judgment result of the second detection step described later as a first defect. This completes the process of step S7, and a series of surface inspection processes are terminated. The processes in steps S6 and S7 correspond to the defect determination step S24 shown in FIG.

[0062] As is clear from the above description, in the first detection step, light is irradiated onto the surface of the steel material S, and the light reflected from the surface of the steel material S by the irradiated light is imaged in two or more different wavelength bands, and pattern-like defects present on the surface of the steel material S are detected from information on the relative signal intensities between the multiple images obtained from the same position on the surface of the steel material S. This makes it possible to accurately detect all surface defects in the steel material S without omission.

[0063] [Second detection step] Fig. 15 is a flow chart showing the flow of the second detection step shown in Fig. 13. In the second detection step, red scale peeling defects in the inspection target area are discriminated from healthy areas with harmless patterns. The red scale peeling defects mentioned here refer to non-opening defects in which the red scale has peeled off and there are no irregularities. The second detection step shown in Fig. 15 starts when an execution command for the second detection step is input to the image processing device 14, and the surface inspection process, which is the second detection step, proceeds to the process of step S21.

[0064] In the second detection step, first, the image processing device 14 performs a correction process on the spectral images of the different wavelength bands (step S21). The specific correction process method is the same as the processes of steps S1 and S2 in the first detection step. The process of step S21 may be omitted, and the corrected image obtained after the process of step S2 in the first detection step may be used as it is as the corrected image after the process of step S21. Next, the image processing device 14 extracts an image for extracting the bare iron part from the corrected image obtained by the process of step S21 (step S25). Here, the image for extracting the bare iron part is a monochrome image with a small R channel component of the color camera. For this purpose, the images captured by the B channel or G channel may be used as they are, or these images may be combined with a specific weight. In addition, depending on the steel material, the image component captured by the R channel may be mixed. Most preferably, the image captured by the B channel of the color camera is used as it is. In order to reduce the processing load of steps S1, S2, S21, and S25, it is preferable that the wavelength band of at least one image used in the first detection step is the same as the wavelength band of the image used in the second detection step.

[0065] Next, the image processing device 14 extracts bright areas from the image that are brighter than the surrounding areas using a threshold value, performs labeling, and extracts surface defect candidate areas (step S26), in the same manner as in the first detection step. Next, the image processing device 14 judges whether each surface defect candidate area obtained in step S26 is a red scale peeling defect or not (step S27). Here, the judgment method may be feature analysis, convolutional neural network, or the like. The judgment result is used in the processing of step S28 to be integrated with the judgment result of the first detection step described above as a second defect. This completes the processing of step S27, and the series of surface inspection processing ends.

[0066] As is clear from the above description, in the second detection step, light is irradiated onto the surface of the steel material S, the light reflected from the surface of the steel material S by the irradiated light is imaged in two or more different wavelength bands, and red scale peeling defects are detected from information extracted from one image or an image obtained by combining multiple images obtained from the same position on the surface of the steel material S and from which bright areas compared to the surrounding areas are extracted. This makes it possible to accurately detect surface defects of the steel material S without omission.

[0067] The surface inspection process according to one embodiment of the present invention can also be applied to a quality control method for metal materials. In this case, the surface inspection process according to one embodiment of the present invention is used to investigate the presence or absence of surface defects in a metal material such as steel material S, and quality control is performed to confirm whether the surface defect occurrence status (occurrence rate, size of defects, etc.) is below a predetermined allowable standard. This makes it possible to provide a metal material with controlled or guaranteed surface properties. For example, by recording whether the surface defect occurrence status is below a predetermined allowable standard in an operation record, inspection book, delivery note, inspection certificate, etc., the manufacturer can control the quality and the consumer can be guaranteed the quality. The surface inspection process according to one embodiment of the present invention can also be applied to a manufacturing method for metal materials. In this case, the surface properties of the metal material are inspected by the surface inspection process according to one embodiment of the present invention, and the metal material is manufactured based on the inspection results.

[0068] Although the embodiments of the present invention have been described above, the present invention is not limited by the descriptions and drawings that form part of the disclosure of the present invention according to the present embodiments. In other words, other embodiments, examples, and operation techniques, etc., made by those skilled in the art based on the present embodiments are all included in the scope of the present invention. [Explanation of symbols]

[0069] 1 Linear Stage 2 Xenon light source 3. Spectroscopic camera 10,20 Metal material surface inspection equipment 11 Light source 12 Encoder and pulse generator 13 Area Sensor 14 Image processing device 15 Monitor 21 Linear light source 22 Line sensor BM Underground Railway D Red scale peeling defect DI Non-opening defect interface EP foreign matter L illumination light MS Black Scale RS, RS2 Red Scale S steel material SA thick steel plate sample

Claims

1. A surface inspection method for a metallic material, which detects defects on the surface of a metallic material from a plurality of images obtained by capturing light reflected from the surface of the metallic material by irradiating the surface with light in two or more different wavelength bands, comprising: a first detection step of generating an image of the difference or ratio of brightness of at least two images obtained from a predetermined area on the surface of the metal material, and detecting a first defect present on the surface of the metal material from the generated image; a second detection step of detecting a second defect present on the surface of the metal material by extracting a bright area compared to its surroundings from one image obtained from the predetermined region on the surface of the metal material or an image obtained by combining a plurality of images; a determining step of detecting the defect using the first defect and the second defect; A method for inspecting a surface of a metal material, comprising:

2. The method for inspecting a surface of a metallic material according to claim 1 , wherein the determining step determines the first defect or the second defect as a harmful defect.

3. The method for inspecting a surface of a metallic material according to claim 1 , wherein at least one of the two or more different wavelength bands is a wavelength band of 500 nm or less.

4. 2. The method for inspecting a surface of a metallic material according to claim 1, wherein a wavelength band of at least one image used in the first detection step is the same as a wavelength band of an image used in the second detection step.

5. 2. The method for inspecting a surface of a metallic material according to claim 1, wherein the first defect is a pattern defect and the second defect is a red scale peeling defect.

6. A surface inspection device for a metallic material, which detects defects on the surface of a metallic material from a plurality of images obtained by capturing light reflected from the surface of the metallic material by irradiating the surface with light in two or more different wavelength bands, comprising: a first detection means for generating an image of a difference or ratio in brightness between at least two images obtained from a predetermined area on the surface of the metal material, and detecting a first defect present on the surface of the metal material from the generated image; a second detection means for detecting a second defect present on the surface of the metal material by extracting a bright area compared to a surrounding area from an image obtained from the predetermined area on the surface of the metal material; a determination means for detecting the defect using the first defect and the second defect; A surface inspection device for a metal material comprising:

7. 7. The surface inspection device for a metallic material according to claim 6, wherein the determining means determines the first defect or the second defect to be a harmful defect.

8. A metallic material, the surface quality of which is guaranteed using the method for inspecting a metallic material surface according to any one of claims 1 to 5.

9. A method for manufacturing a metallic material, comprising inspecting a surface property of a metallic material using the method for inspecting a surface of a metallic material according to any one of claims 1 to 5, and manufacturing the metallic material based on the inspection results.

10. A quality control method for a metallic material, comprising inspecting the surface properties of a metallic material using the method for inspecting a surface of a metallic material according to any one of claims 1 to 5, and controlling the quality of the metallic material based on the inspection results.

11. A manufacturing facility for metallic materials, comprising the surface inspection device for metallic materials according to claim 6 or 7.

Images