Inspection method, inspection device, and rolling device control method

Abstract

To provide an inspection method which can highly accurately detect a minute concave-convex defect generated on the surface of a metal rolled plate.SOLUTION: An inspection method comprises: a step A of preparing a metal plate WP that is rolled along the first direction; a step B of projecting the striped illumination in which a plurality of bright and dark streaks extending in the second direction alternately appear along the third direction intersecting the second direction on the surface of the metal plate so as to have an angle equal to or greater than about 80° and equal to or less than about 100° between the first direction and the second direction; a step C of acquiring a two-dimensional image of a striped pattern formed on the surface of the metal plate; and a step D of extracting a defective section image in a predetermined size or smaller from the two-dimensional image.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a method for inspecting the surface of a rolled metal plate, an inspection device, and a method for controlling a rolling device, and more particularly to a technique for detecting defects occurring on the surface of a rolled metal plate. [Background technology]

[0002] Generally, minute surface defects (irregular defects) may be formed on the surface of a rolled metal plate due to scratches on the surface of the rolls of the rolling equipment. Minute irregular defects on the surface of a rolled metal plate have been visually inspected by utilizing the difference in the way light is reflected between the irregular defect and its surroundings. In order to increase the difference in brightness between the irregular defect and its surroundings, the surface may be polished with a grinding stone.

[0003] When a rolled metal sheet is manufactured through multiple rolls, unevenness defects caused by surface flaws on rolls prior to the final roll are rolled until the final product is produced, so the height of the unevenness becomes small, making it difficult to detect with the naked eye.

[0004] Therefore, Patent Document 1 discloses an inspection method that makes it easier to find uneven defects on the steel sheet surface by making the reduction rate of the final roll smaller only in the inspection target portion than in the product portion. Patent Document 1 describes that a convex defect about 3 μm in height that could only be found by conducting five conventional grinding inspections can be found in one inspection by reducing the reduction rate of the inspection target portion to 50%.

[0005] However, visual inspection requires skill and the reliability of the results varies. Therefore, inspection methods using image processing have been developed. For example, Patent Document 2 discloses an inspection device that can detect even slight surface defects by using an inspection image obtained by photographing the surface of a workpiece illuminated with a predetermined pattern (e.g., a striped pattern) and multiple images of non-defective products photographed at the same location as the inspection image. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] JP 2018-27549 A [Patent Document 2] JP 2018-4324 A Summary of the Invention [Problem to be solved by the invention]

[0007] Demands for quality of appearance are becoming stricter, and there is a demand for highly accurate inspection of minute irregular defects less than 3 μm in height, and the techniques described in Patent Documents 1 and 2 may not be sufficient to meet this demand. In particular, on the surface of a metal rolled plate, there are lines called rolling streaks that are parallel to the rolling direction, and these lines disturb the reflection of light. Furthermore, a cold-rolled plate in a state where correction processing or the like is not performed has low uniformity in plate shape, and for example, the thickness at the center of the plate width may be larger (the center is convex) or smaller (the center is concave) than near the edges of the plate width, and this plate shape variation varies greatly depending on the longitudinal position (the position in the rolling direction) within the metal rolled plate or the difference in the continuous operation time of the rolling device. Due to these factors, even if an image processing technique that requires a good product image as described in Patent Document 2 is applied, it is not possible to detect minute irregular defects on the surface of a metal rolled plate with sufficient accuracy.

[0008] The present invention has been made to solve the above problems, and aims to provide an inspection method and an inspection device that can detect minute uneven defects occurring on the surface of a metal rolled plate with high accuracy. Another aim of the present invention is to provide a method of controlling a rolling device using such an inspection method. [Means for solving the problem]

[0009] According to an embodiment of the present invention, there are provided solutions as described in the following items.

[0010] [Item 1] A step A of preparing a metal plate rolled along a first direction; A step B of projecting striped illumination, in which a plurality of light and dark streaks extending in a second direction alternate along a third direction intersecting the second direction, onto the surface of the metal plate, such that the second direction forms an angle of about 80° or more and about 100° or less with respect to the first direction; A step C of acquiring a two-dimensional image of the stripe pattern formed on the surface of the metal plate; A step D of extracting an image of a defect portion having a size equal to or smaller than a predetermined size from the two-dimensional image; The inspection method includes the steps of: The third direction may be parallel to the first direction. [Item 2] 2. The inspection method according to item 1, wherein the step D includes a step d1 of generating a high-frequency removed image by removing high-frequency components from the two-dimensional image using a low-pass filter, and a step d2 of generating a difference image between the two-dimensional image and the high-frequency removed image, and extracting the defect image based on the difference image. [Item 3] 2. The inspection method according to item 1, wherein the step D extracts the image of the defect by extracting high-frequency components from the two-dimensional image using a high-pass filter. [Item 4] 4. The inspection method according to any one of items 1 to 3, wherein the defect image has an equivalent circle diameter of 500 μm or less. [Item 5] 5. The inspection method according to any one of items 1 to 4, wherein the first direction and the second direction form an angle of approximately 90°. [Item 6] 6. An inspection method according to any one of items 1 to 5, wherein the plurality of bright and dark streaks of the striped illumination have a wave-like pattern in which luminance varies with a period Pl along the third direction. [Item 7] 7. The inspection method according to item 6, wherein the wave-like pattern is a sine wave. [Item 8] The step D is a step of extracting a plurality of images of the defect portion, 8. An inspection method according to any one of items 1 to 7, further comprising, after step D, a step E of selecting, from the extracted plurality of defect images, a plurality of first defect images that appear periodically along the first direction. [Item 9] 9. The inspection method according to item 8, further comprising, after step E, a step F of determining at least one period Pd in ​​which the multiple first defect portion images appear. [Item 10] Item 9. The inspection method according to item 8, wherein in step F, two or more mutually different periods Pd are obtained. The method may further include a step of outputting and / or displaying the two or more different periods Pd. When the metal plate is formed through rolling by a plurality of rolls, the roll having the surface defect that caused the defective portion image can be identified based on the two or more different periods Pd. [Item 11] repeating the steps B and C on different regions of the surface of the metal plate to obtain a plurality of the two-dimensional images over the entire surface; generating an overall image by combining the two-dimensional images of the entire surface; 11. The inspection method according to any one of items 8 to 10, wherein the step E is performed on the entire image. [Item 12] 1. A method of controlling a rolling mill, comprising the steps of: A step of rolling a metal plate along a first direction; An inspection step of performing the inspection method according to any one of items 8 to 11; a step of stopping the rolling process when the plurality of first defect portion images are obtained in the inspection step; The control method includes: [Item 13] a stage that receives a metal plate rolled along a first direction; an illumination device that projects striped illumination, in which a plurality of light and dark streaks extending in a second direction alternate along a direction intersecting the second direction, onto a surface of the metal plate, such that the second direction forms an angle of approximately 80° or more and approximately 100° or less with respect to the first direction; an imaging device for acquiring a two-dimensional image of the stripe pattern formed on the surface of the metal plate; an image processing device that extracts an image of a defect portion having a size equal to or smaller than a predetermined size from the two-dimensional image; An inspection device having the above features.

[0011] The inspection device is attached to a rolling mill, and the output of the inspection device can be used to control the rolling mill. Effect of the Invention

[0012] According to an embodiment of the present invention, there is provided an inspection method and an inspection device capable of detecting minute unevenness defects occurring on the surface of a rolled metal plate with high accuracy. According to another embodiment of the present invention, there is provided a control method of a rolling device using such an inspection method. [Brief description of the drawings]

[0013] [Figure 1] 1 is a schematic diagram of an inspection apparatus 100 according to an embodiment of the present invention. [Diagram 2] 2 is a schematic diagram showing the hardware configuration of a computer 60 included in the inspection device 100. FIG. [Figure 3A] 1 is a diagram showing a schematic example of the positional relationship of an illumination device 40 and an imaging device 50 with respect to a metal plate WP. [Figure 3B] 1 is a schematic diagram showing a two-dimensional image of striped illumination formed on the surface of a metal plate. [Figure 3C] FIG. 13 is a diagram showing an example of a luminance profile of striped illumination. [Figure 4] 1 is a flowchart showing an outline of the overall procedure of an inspection method according to an embodiment of the present invention. [Diagram 5] 10 is a flowchart showing a procedure of an image acquiring step in the inspection method. [Figure 6]1A and 1B are schematic diagrams showing the operation of the inspection apparatus 100 in the y direction. [Figure 7] 1 is a schematic diagram showing the x-direction operation of the inspection apparatus 100. FIG. [Figure 8] 10 is a flowchart showing a procedure of an image processing step in the inspection method. [Figure 9A] FIG. 13 is a diagram showing an example of an acquired raw reflection image. [Figure 9B] FIG. 13 is a diagram showing an example of a combined image obtained by combining a plurality of raw reflection images. [Figure 10] FIG. 13 is a diagram showing an example of a high frequency removed image. [Figure 11] FIG. 13 is a diagram showing an example of a difference image. [Figure 12] FIG. 13 is a diagram showing an example of a defect candidate image. [Figure 13] FIG. 13 is a diagram showing an example of a combined overall image including periodic defects. [Figure 14] FIG. 13 is a diagram showing an example of a periodic defect detection image. [Figure 15] FIG. 2 is a schematic diagram of a rolling device 200. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] Hereinafter, an inspection method, an inspection device, and a control method for a rolling device according to an embodiment of the present invention will be described with reference to the drawings. Note that the inspection method, the inspection device, and the control method for a rolling device according to the embodiment of the present invention are not limited to the following examples.

[0015] First, an inspection method according to an embodiment of the present invention and an inspection device used therein will be described with reference to Fig. 1. Fig. 1 shows a schematic diagram of an inspection device 100 according to an embodiment of the present invention. The inspection device 100 can inspect, for example, a 2000 mm x 5000 mm rolled aluminum plate (e.g., approximately 1 mm thick). The inspection device according to the embodiment of the present invention is not limited to this, and is not limited to the material or size of the rolled metal plate, and inspection can be performed in a single image capture.

[0016] The inspection device 100 includes a stage 10 that receives the rolled metal plate WP, an illumination device 40 that projects striped illumination onto the surface of the metal plate WP, an imaging device 50 that captures a two-dimensional image of the striped pattern formed on the surface of the metal plate WP, and a computer 60. The inspection device 100 further includes an x-direction driving device 20 (20a and 20b) and a y-direction driving device 30 for moving the illumination device 40 and the imaging device 50 relative to the metal plate WP. The x and y directions are orthogonal to each other, and the z direction is orthogonal to the x and y directions.

[0017] Here, a configuration in which the illumination device 40 and the imaging device 50 are moved relative to the stage 10 has been exemplified, but the present invention is not limited to this, and for example, a configuration in which the stage 10 is moved may also be used. The illumination device 40 and the imaging device 50 are sometimes collectively referred to as the "optical system 50." Here, an example in which the optical system 50 is moved relatively is shown, but if the illumination device 40 can project striped illumination over a sufficiently wide range (if the area of ​​the inspection object is small), there is no need to move the illumination device 40. Of course, depending on the size of the inspection object, there is no need to move the imaging device 50.

[0018] The metal plate WP is placed on the stage 10 so that the rolling direction coincides with the y direction. Here, the width direction of the metal plate WP is the x direction, and the length direction is the y direction, which is the same as the stretching direction. The z direction is the thickness direction of the metal plate WP.

[0019] The lighting device 40 projects striped illumination on the surface of the metal plate WP, in which multiple light and dark streaks extending in the x direction alternate along a direction intersecting the x direction. Here, an example is shown in which multiple light and dark streaks extending approximately in the x direction alternate along the y direction, but the present invention is not limited to this example, and any direction that forms an angle of about 80° to about 100° with respect to the x direction (i.e., within about ±10° with respect to the y direction) may be used. Outside this range, the light reflection may be disturbed by the influence of rolling streaks extending in the y direction, and the desired striped pattern two-dimensional image 42A may not be formed on the surface of the metal plate WP.

[0020] Here, assuming that the rolling direction is the first direction and the direction in which the multiple light and dark streaks of the stripe illumination extend is the second direction, as in the above example, the first direction and the second direction are set to form an angle of about 80° or more and about 100° or less. It is preferable that the first direction and the second direction form an angle of about 90°, and the closer to 90° the better. The closer the angle between the first direction and the second direction is to 90°, the easier it is to design a low-pass filter or high-pass filter that utilizes a fast Fourier transform (FFT) to extract an image of the defect.

[0021] The imaging device 50 is, for example, a digital camera, and has a solid-state imaging element (CCD image sensor or CMOS image sensor) and an optical system, and obtains a two-dimensional image of a desired stripe pattern formed on the surface of the metal plate WP. The imaging device 50 and the computer 60 may be connected by wire or wirelessly.

[0022] The computer 60 functions as an image processing device that extracts images of defective parts having a predetermined size or smaller from the two-dimensional images acquired by the imaging device 50. The hardware configuration of the computer 60 is shown in FIG.

[0023] 2, the computer 60 includes a CPU 61, a memory 62, an image processing circuit 63, a display 64, an input device 65, a storage device 66, an interface 67 for receiving data from the imaging device 50, and a control circuit 68, which are connected by a bus 69 and can transmit and receive data between them. The memory 62 may be built into the CPU 61. Alternatively, a field-programmable gate array (FPGA) can be used as an alternative to the CPU 61 and the image processing circuit 63.

[0024] The CPU 61 is a signal processing circuit that controls the operation of the inspection device 100. Typically, the CPU 61 is a semiconductor integrated circuit. The CPU 61 may also be simply called a "processing circuit." The memory 62 stores a computer program executed by the CPU 61, and is, for example, a rewritable memory (flash memory) or an EEPROM. The memory 62 can also be used as a work memory when the CPU 61 performs calculations. The computer program may be recorded in the storage device 66. The storage device 66 is, for example, an HHD or SSD. Various data (input data, inspection results, etc.) can be recorded in the storage device 66.

[0025] The image processing circuit 63 generates an image to be displayed on the display 64 in accordance with instructions from the CPU 61. The display 64 displays various images.

[0026] The input device 65 is, for example, a keyboard, and is used to input initial data required to control the operation of the inspection device 100. For example, size information of the metal plate WP, position information of the imaging range and imaging start point, x-direction movement distance, y-direction movement distance, and size information of defects to be detected are input from the input device 65. The period of the defects to be detected may also be input.

[0027] The control circuit 68 supplies signals for controlling the driving devices 20 and 30 according to instructions from the CPU 61 .

[0028] Next, striped illumination projected onto the surface of the metal plate WP by the illumination device 40 will be described with reference to FIGS. 3A to 3C.

[0029] Fig. 3A is a schematic diagram showing an example of the arrangement of the illumination device 40 and the imaging device 50 relative to the metal plate WP. Fig. 3A also shows a schematic diagram of a defect DF formed on the surface of the metal plate WP. Fig. 3A shows a surface parallel to the yz plane.

[0030] The minute uneven defect DF formed on the surface of the metal plate WP has a height Hm of, for example, less than 3 μm, 2.0 μm or less, or even 1.0 μm. The shape of the defect DF (when viewed from the normal direction of the surface of the metal plate WP) is indefinite, and when its size is expressed in terms of an area circle equivalent diameter Dm, it can be, for example, 500 μm or less, 300 μm or less, or even 100 μm or less. In addition, the surface shape of the defect DF is a smooth curve, as shown in FIG. 3A, and it is difficult to observe the change in brightness of the reflected light. The inspection method according to the embodiment of the present invention has the advantage of being able to detect minute uneven defects that were difficult to detect with conventional inspection methods with high accuracy, but is not particularly limited by the size of the defect, and it goes without saying that it can detect, for example, a defect with an area circle equivalent diameter Dm of more than 500 μm and / or a defect with a height Hm of 3 μm or more.

[0031] The illumination device 40 irradiates the surface of the metal plate WP with light rays (principal light rays) Li of the stripe illumination from a direction parallel to the y direction. That is, by projecting the stripe illumination so that the light rays Li are parallel to the rolling streaks of the metal plate WP, a two-dimensional image 42A of a clear stripe pattern is formed on the surface of the metal plate WP without being affected by the rolling streaks (see FIG. 3B). The angle θi of the light rays Li of the stripe illumination from the radial direction of the metal plate WP is, for example, greater than 0° and equal to or less than 45°. Here, the angle θi is sometimes referred to as the projection angle θi.

[0032] The imaging device 50 is disposed at a position where it can obtain a two-dimensional image 42A of the stripe pattern formed on the surface of the metal plate WP. That is, it is disposed at a position where the light ray Li receives the reflected light Lr reflected on the surface of the metal plate WP. The optical axis of the imaging device 50 (the direction of the reflected light Lr) does not necessarily need to be symmetrical with the direction of the light ray Li of the stripe illumination (the optical axis of the illumination device 40), but as shown in FIG. 3A, if the clockwise direction from the normal line (dashed line) in FIG. 3A is positive and θi is a positive value, θr is a negative value, and the magnitude of θr is, for example, greater than 0° and equal to or less than 60°. Here, the angle θr may be referred to as the light receiving angle θr.

[0033] As shown in FIG. 3B, the illumination device 40 projects striped illumination on the surface of the metal plate WP, in which multiple light and dark stripes extending in the x direction alternate along the y direction, and a clear two-dimensional image 42A of a striped pattern is formed on the surface of the metal plate WP. At this time, the multiple light and dark stripes of the striped illumination preferably have a wavy pattern in which the brightness changes with a period P1, and the wavy pattern preferably has a pattern in which the brightness changes continuously and smoothly. This can improve the accuracy of detecting and extracting the defect image, which is an image of a fine defect DF, by image processing. The wavy pattern is, for example, a sine wave as shown in FIG. 3C.

[0034] As such an illumination device 40, for example, a line illumination device LFX3-100SW-PT-B (pitch 2 mm) manufactured by CCS Inc. can be used. The stripe illumination is, for example, white light, but is not particularly limited. It is sufficient that the imaging device 50 can capture a reflected image.

[0035] Next, FIG. 4 shows a flowchart of an inspection method according to an embodiment of the present invention.

[0036] In accordance with instructions from the computer 60, the inspection device 100 acquires a two-dimensional image of the stripe pattern formed on the surface of the metal plate WP (step SA1). Next, the acquired image is processed (step SA2), an overall image is formed (step SA3), and finally, the inspection result is output (step SA4). Of course, step SA3 may be omitted depending on the size of the inspection object.

[0037] For example, size information of the metal plate WP, position information of the imaging range and imaging start point, x-direction movement distance, y-direction movement distance, and size information of defects to be detected are input from the input device 65 of the computer 60. After these necessary information are input, an inspection start signal is output from the CPU 61 to the control circuit 68. The input information is recorded in the memory 62 or the storage device 66.

[0038] Next, an example of the image acquiring step will be described with reference to Fig. 5 to Fig. 7. Fig. 5 is a flow chart showing the procedure of the image acquiring step in the inspection method, Fig. 6 is a schematic diagram showing the operation of the inspection device 100 in the y direction, and Fig. 7 is a schematic diagram showing the operation of the inspection device 100 in the x direction.

[0039] The control circuit 68, to which the inspection start signal has been input, operates the x-direction driver 20 and the y-direction driver 30 based on the input information, and moves the optical system 50 to the inspection start point (FIG. 5, step SB1).

[0040] After the optical system 50 reaches the inspection start point, the control circuit 68 operates the y-direction driver 30 to move the optical system 50 until it reaches the input y-direction movement distance as shown in FIG. 6. This operation is called the y-direction operation (FIG. 5, step SB2). At this time, the control circuit 68 inputs a trigger signal to the imaging device 50 every time the y-direction driver 30 advances a distance dA, and the imaging device 50 captures an image each time. The distance dA is shorter than the vertical field of view pA of the image 52A captured by the imaging device 50. Therefore, the image 52A(1) and the image 52A(2) captured next partially overlap in the y direction. The images captured by the imaging device 50 are stored in the memory 62 or the storage device 66 each time in association with the position information of the y-direction driver 30 and the x-direction driver 20.

[0041] The image 52A captured by the imaging device 50 is smaller than the two-dimensional image 42A formed on the surface of the metal plate WP by the illumination device 40, and a range in which the luminance distribution is uniform within the two-dimensional image 42A is selected.

[0042] When the above y-direction operation is completed, the control circuit 68 operates the x-direction driver 20 by a distance dB as shown in FIG. 7, and after the movement, a trigger signal is input to the imaging device 50, and the imaging device 50 captures an image. The distance dB is shorter than the width-direction field of view pB of the image 52A captured by the imaging device 50. Therefore, the image 52A(11) and the image 52A(12) captured next partially overlap in the x direction. This operation is called the x-direction operation (FIG. 5, step SB3).

[0043] For example, the vertical field of view pA is 40 mm, the distance dA is 20 mm, the widthwise field of view pB is 50 mm, and the distance dB is 25 mm. Thus, if pA, dA, pB, and dB are set so as to satisfy the relationships dA = 0.5·pA and dB = 0.5·pB, all locations on the surface of the metal plate WP will be imaged twice, so the possibility of overlooking minute defects can be reduced. Of course, pA, dA, pB, and dB are not limited to this example, and as illustrated in FIGS. 6 and 7, 0.5·pA < dA < pA and 0.5·pB < dB < pB may be used. Alternatively, dA < 0.5·pA and dB < 0.5·pB may be used. Note that the relationship between pA and dA and the relationship between pB and dB can be set independently of each other.

[0044] When the above-described x-direction operation is completed, the control circuit 68 performs the y-direction operation again. The operation direction at this time is opposite to the previous y-direction operation. Thereafter, the y-direction operation and the x-direction operation are repeated to acquire an image of the entire inspection range and the operation is stopped (FIG. 5, step SB4).

[0045] Next, with reference to FIG. 8, the image processing step will be described. FIG. 8 is a flowchart showing the procedure of the image processing step in the inspection method.

[0046] The plurality of images acquired in the image acquisition step are stored in the memory 62 or the storage device 66. The image processing circuit 63 reads the images from the memory 62 or the storage device 66 and performs the same processing on each individual image. Note that image processing may be performed each time one or two or more images are acquired in the image acquisition step. The number of images on which image processing is performed may be determined in consideration of the imaging time and the calculation time so as to shorten the total inspection time.

[0047] FIG. 9A shows an example of a two-dimensional image of the stripe pattern as acquired. Hereinafter, this image is referred to as a raw reflected image. Although multiple light and dark streaks are formed extending in the x direction perpendicular to the direction of the rolling streaks (i.e., the y direction), the multiple streaks are not completely parallel, but are gently curved. It can also be seen that small curvatures exist locally. The gentle curvatures are due to variations in the surface shape of the metal plate to be inspected (variations in the flatness of the metal plate surface), and differ depending on the position where the raw reflected image is acquired. Meanwhile, small curvatures that exist locally correspond to images of defects (referred to as "defect image").

[0048] FIG. 9B shows a combined image obtained by combining reflected images 62A(n) excluding overlapping portions of the acquired raw reflected images 52A(n). For example, the reflected image 62A(4) in FIG. 9B is obtained from the raw reflected image shown in FIG. 9A. As shown in FIG. 9B, the curve of the light and dark streaks included in the reflected image 62A differs depending on the position. This is due to the variation in the surface shape of the metal sheet WP (variation in the flatness of the surface of the metal sheet). For example, the thickness of the central part of the sheet width may be larger (the central part is convex) or smaller (the central part is concave) than that near the edges of the sheet width, and this variation in the sheet shape varies greatly depending on the longitudinal position (the position in the rolling direction) within the metal rolled sheet or the difference in the continuous operation time of the rolling device. Therefore, it is not possible to use image processing techniques that utilize standard images such as non-defective product images.

[0049] In the image processing step of the inspection method according to the embodiment of the present invention, for example, a low-pass filter is applied to the raw reflected image to remove image components with high spatial frequencies to generate a high-frequency removed image (FIG. 8, step SC1). The frequency components to be removed are set according to the size (e.g., area circle equivalent diameter) of the surface defect to be detected. Alternatively, conversely, a high-pass filter may be used to extract only image components corresponding to the size of the surface defect to be detected. An example of using a low-pass filter will be described below. For example, the area circle equivalent diameter of the defect portion image to be extracted is 500 μm or less, but may be 300 μm or less, or may be, for example, 100 μm or less.

[0050] The design of the low-pass filter also depends on the settings of the optical system 50, such as the light-dark cycle (pitch) of the illumination device 40 used and the settings of the imaging device 50. For example, even if a line illumination device with a certain pitch is used, the pitch of the stripe pattern included in the raw reflection image may differ if the projection angle θi and / or the reception angle θr (see FIG. 3A) is different. The low-pass filter is designed with these factors in mind, and is set to remove at least components that are equal to or higher than the spatial frequency of the gently curved stripe pattern (striped pattern) in the raw reflection image.

[0051] Figure 10 shows an example of a high-frequency-removed image generated by applying a low-pass filter to the raw reflected image in Figure 9A. It can be seen that the small localized curvatures present in the raw reflected image have been removed, while the gentle curvatures have been preserved. This high-frequency-removed image can be used as a standard image (non-defective image) that would have been obtained if the metal plate WP to be inspected did not have any minute defects.

[0052] Next, the high frequency removed image in FIG. 10 is used as a standard image to obtain the difference from the raw reflected image in FIG. 9A (FIG. 8, step SC2). This difference process can obtain an image as shown in FIG. 11. This difference image may contain not only a defect image, which is an image of a defect, but also noise components. Note that the difference image may be obtained by subtracting the high frequency removed image from the raw reflected image, or conversely, by subtracting the raw reflected image from the high frequency removed image.

[0053] Next, in order to remove noise components, for example, binarization processing (FIG. 8, step SC3) and area determination (FIG. 8, step SC4) are performed. The difference image may contain noise components other than the defect image. For example, a difference image in which the brightness of bright parts is low and / or the area of ​​bright parts is small can be regarded as noise. Therefore, if necessary, a threshold correction processing is performed prior to the binarization processing (step SC3) (or as part of the binarization processing). The threshold may be different depending on, for example, the position of the image. The binarization processing (step SC3) is performed using the threshold obtained by the threshold correction processing. After that, the area of ​​bright parts is calculated by, for example, a labeling processing, and only parts in which the area of ​​bright parts is equal to or larger than a certain level are left and are stored as a defect candidate image (FIG. 8, step SC5). An example of a defect candidate image is shown in FIG. 12.

[0054] After performing the above-mentioned image processing on all the acquired raw reflected images, an overall image is formed (FIG. 4, step SA3). In the overall image formation step, an overall image is formed by combining all the defect candidate images based on the position information of each of the defect candidate images obtained in the image processing step. When defects having periodicity in the rolling direction are included, for example, an overall image as shown in FIG. 13 is obtained. The overall image and the results obtained in each processing step are stored, for example, in the storage device 66.

[0055] The final inspection results are output in various forms (FIG. 4, step SA4). For example, if the input information does not include the periodicity of the defects to be detected, the display 64 displays the entire image. Alternatively, multiple defect images that appear periodically along the y direction are selected from the coordinates of the center of gravity of the defect image in the entire image, and these are displayed as in the example of FIG. 14. Also, the period Pd at which the defect images appear may be found, and the period Pd may be displayed on the display 64 together with the image or independently of the image. Multiple periods Pd may be found.

[0056] As shown in FIG. 15, for example, a rolling mill 200 having a plurality of rolling rolls 80a, 80b between a pair of tension rolls 90a, 90b is used. The metal sheet MP is rolled by the rolling roll 80a, and then rolled again by the rolling roll 80b. Each of the rolling rolls 80a, 80b is equipped with a hydraulic reduction device and a load meter. If necessary, a thickness meter for measuring the thickness of the sheet after being rolled by each of the rolling rolls 80a, 80b, and a tension meter for measuring the tension of the metal sheet between the rolling rolls 80a, 80b are also provided. The rolling mill 200 is well known, so details will be omitted.

[0057] If the diameter of the rolling roll 80a is different from that of the rolling roll 80b, the period in which defects caused by roll marks appear will be different. In addition, differences in the rolling reduction rate will also appear as differences in the period. Therefore, two or more different periods Pd may be output and / or displayed. When the metal sheet MP is formed through rolling by a plurality of rolls, the roll having the surface defect that caused the defective portion image can be identified based on two or more different periods Pd.

[0058] If the input information indicates periodicity of the defect to be detected, only those bright areas in the overall image whose center of gravity coordinates match the input period are extracted as bright areas, and a periodic defect detection image such as that shown in FIG. 14 is generated and displayed on the display 64.

[0059] The inspection device 100 is used, for example, to inspect a metal sheet WP that has passed through the rolling device 200 and is cut to a predetermined length, but can also be attached to the rolling device 200. When a periodic defect image is detected from the inspection results, the output of the inspection device may be used to control the rolling device. That is, when the inspection device detects a periodic defect image, the operation of the rolling device may be controlled to be stopped. That is, the method for controlling a rolling device according to an embodiment of the present invention includes a step of rolling a metal sheet along a first direction, a step of performing the above-mentioned inspection method, and a step of stopping the rolling when a plurality of first defect images are obtained in the inspection step. [Industrial Applicability]

[0060] Embodiments of the present invention are broadly applicable to the inspection and manufacturing of rolled metal sheets. [Explanation of symbols]

[0061] 10: Stage 20: x-direction drive unit 30: y-direction drive unit 40: Lighting equipment 50: Imaging device (optical system) 60: Computer 61: CPU 62: Memory 63: Image processing circuit 64: Display 65: Input device 66: Storage device 67: Interface 68: Control circuit 69: Bus

Claims

1. A step A of preparing a metal plate rolled along a first direction; a step B of projecting striped illumination, in which a plurality of light and dark streaks extending in a second direction alternate along a third direction intersecting the second direction, onto the surface of the metal plate such that the second direction forms an angle of about 80° or more and about 100° or less with respect to the first direction, wherein the striped illumination is projected such that a projection angle θi between a principal ray of the striped illumination and a normal to the surface of the metal plate is greater than 0° and less than 45°; A process C of acquiring a two-dimensional image including a stripe pattern including gentle curvature and localized small curvature formed on the surface of the metal plate and a defect portion image of a defect having a height of less than 3 μm by an imaging device having an optical axis arranged so that a light-receiving angle θr formed with a normal to the surface of the metal plate is more than 0° and not more than 60°; A step D of extracting the defect image from the two-dimensional image; Inclusive of The step D includes a step d1 of generating a high-frequency removed image in which the local small curvature is removed by removing high-frequency components from the two-dimensional image using a low-pass filter, and a step d2 of generating a difference image between the two-dimensional image and the high-frequency removed image, and extracting the defect image based on the difference image.

2. 2. The inspection method according to claim 1, wherein the defect image has an equivalent circle diameter of 500 μm or less.

3. The inspection method according to claim 1 , wherein the first direction and the second direction form an angle of approximately 90°.

4. The inspection method according to claim 1 , wherein the plurality of bright and dark streaks of the striped illumination have a wave-like pattern in which luminance changes with a period Pl along the third direction.

5. The method of claim 4 , wherein the wave-like pattern is a sine wave.

6. The step D is a step of extracting a plurality of images of the defect portion, The inspection method according to any one of claims 1 to 5, further comprising, after step D, a step E of selecting, from the extracted plurality of defect images, a plurality of first defect images that appear periodically along the first direction.

7. 7. The inspection method according to claim 6, further comprising, after said step E, a step F of determining at least one period Pd in ​​which said plurality of first defect portion images appear.

8. The inspection method according to claim 7 , wherein in said step F, two or more mutually different periods Pd are obtained.

9. repeating the steps B and C on different regions of the surface of the metal plate to obtain a plurality of two-dimensional images over the entire surface; generating an overall image by combining the two-dimensional images of the entire surface; The inspection method according to claim 6 , wherein the step E is performed on the entire image.

10. 1. A method of controlling a rolling mill, comprising the steps of: A step of rolling a metal plate along a first direction; An inspection step of carrying out the inspection method according to any one of claims 6 to 9; a step of stopping the rolling process when the plurality of first defect portion images are obtained in the inspection step; The control method includes:

11. a stage for receiving a metal plate rolled along a first direction; an illumination device that projects striped illumination, in which a plurality of light and dark streaks extending in a second direction alternately appear along a direction intersecting the second direction, onto a surface of the metal plate, such that the second direction forms an angle of about 80° or more and about 100° or less with respect to the first direction, and a projection angle θi between a principal ray of the striped illumination and a normal to the surface of the metal plate is greater than 0° and less than 45°; an imaging device that has an optical axis arranged so that a light-receiving angle θr formed with a normal to the surface of the metal plate is greater than 0° and less than 60°, and that acquires two-dimensional images including a stripe pattern including gentle curvatures and localized small curvatures formed on the surface of the metal plate and defect portion images of defects having a height of less than 3 μm; an image processing device that generates a high-frequency removed image in which the local small curvature is removed by removing high-frequency components from the two-dimensional image using a low-pass filter, generates a difference image between the two-dimensional image and the high-frequency removed image, and extracts the defect image from the two-dimensional image based on the difference image.

Images