Plate Shape Judgment Device and Plate Shape Judgment Method

Abstract

Provided are a plate shape determination device and a plate shape determination method capable of enhancing the determination accuracy of the plate shape of a metal strip as compared with the conventional art. 【Solution means】Rotating shafts extend in the width direction of the metal strip 1 and are installed, and loopers 71, 72, 73, 74 that lift the metal strip 1 upward, and cameras 61, 62, 63, 64 that capture an image including the lifted region of the metal strip 1 lifted upward by the loopers 71, 72, 73, 74, and a luminance data acquisition unit 81 that acquires the luminance value of B (blue) from the R (red), G (green), and B (blue) luminance data of the pixels of the image captured by the cameras 61, 62, 63, 64, and using only the luminance data of the luminance value of B (blue) acquired by the luminance data acquisition unit 81, a reflected light region where the luminance value of B (blue) is equal to or greater than a certain value is obtained in the image, and an image processing unit 82 that determines a defect in the plate shape based on the coordinates of the boundary position of the reflected light region.

Description

Technical Field

[0001] The present invention relates to a plate shape determination device and a plate shape determination method.

Background Art

[0002] As an example of a defect determination device and a defect determination method capable of easily determining a defect in the plate surface shape of a metal strip without using a special light source such as a rod shape, Patent Document 1 discloses that a roll having a rotation axis extending in the width direction of a steel sheet to be rolled and lifting the steel sheet to be rolled upward, a camera that captures an image including the lifted area of the steel sheet to be rolled lifted upward by the roll, and a control device that determines a defect in the plate surface shape of the steel sheet to be rolled based on the image captured by the camera.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In the above-mentioned Patent Document 1, as a method for specifying a reflected light region where illumination light reflected on the surface of a metal strip is visible, from an image of the steel sheet to be rolled projected by a camera, the luminance of the metal strip surface is set as a threshold value, and the region of the plate surface of the metal strip showing a luminance value higher than the set threshold luminance is defined as the reflected light region. From the pixel coordinates of the image, two-dimensional coordinate data of the boundary lines on the upstream side and the downstream side of the reflected light region is acquired, and an evaluation value (index) of the plate shape is obtained from the coordinate data.

[0005] In Patent Document 1, there is no specific description on how to obtain luminance. Specifically, an RGB color image is converted into a grayscale black-and-white image, and the luminance in the black-and-white image is used to determine the boundary line between the upstream side and the downstream side of the reflected light region. Using the two-dimensional coordinate data of the boundary line, a plate shape evaluation is performed by performing a predetermined mathematical process. That is, the RGB luminance is decomposed into three types: the luminance of the luminance value of R (red), the luminance of the luminance value of G (green), and the luminance of the luminance value of B (blue), and the evaluation is performed using the average value of the three types of luminance.

[0006] In the technology of Patent Document 1 above, in order to accurately evaluate the plate shape of the rolled metal strip, it is important to accurately extract the change in the coordinates of the boundary line of the reflected light region. As a result of the inventors' intensive studies, it has become clear that there is room for improvement in the coordinate extraction accuracy of the boundary line of the reflected light region in the technology described in Patent Document 1.

[0007] The present invention provides a plate shape determination device and a plate shape determination method capable of improving the determination accuracy of the plate shape of a metal strip compared to the prior art.

Means for Solving the Problem

[0008] The present invention includes a plurality of means for solving the above problems. For example, a plate shape determination device for determining a defect in the plate shape of a metal strip rolled by a rolling mill, wherein a rotation axis extends in the width direction of the metal strip and is installed, a roll for lifting the metal strip upward, a camera for photographing an image including the lifted region of the metal strip lifted upward by the roll, a luminance data acquisition unit for acquiring the luminance value of B (blue) from the luminance data of the pixels of the image photographed by the camera, and using only the luminance data of the luminance value of B (blue) acquired by the luminance data acquisition unit, obtaining a reflected light region in the image where the luminance value of the luminance value of B (blue) is equal to or greater than a certain value, and an image processing unit for determining the defect in the plate shape based on the coordinates of the boundary position of the reflected light region.

Effect of the Invention

[0009] According to the present invention, the accuracy of determining the plate shape of a metal strip can be improved more than before. Problems, configurations, and effects other than those described above will be clarified by the following description of the embodiments.

Brief Description of the Drawings

[0010]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Embodiment for Carrying Out the Invention

[0011] Examples of the plate shape determination device and the plate shape determination method of the present invention will be described with reference to FIGS. 1 to 12. In the drawings used in this specification, the same or corresponding components are denoted by the same or similar reference numerals, and repeated descriptions of these components may be omitted.

[0012] The metal strip of the material to be rolled in the present invention is generally a strip of a metal material capable of being rolled, and its type is not particularly limited. In addition to steel plates, it can be a non-ferrous material such as aluminum or copper.

[0013] First, the overall configuration of the rolling equipment including the plate shape inspection and determination device for the metal strip will be described with reference to FIGS. 1 to 3. FIG. 1 is a schematic diagram showing the configuration of the plate shape inspection and determination device for the metal strip of this embodiment and the rolling equipment equipped with the same. FIGS. 2 and 3 are diagrams showing an example of a state in which the reflected light of illumination appears in a band shape on the surface of the metal strip at the looper section between stands during operation in the rolling equipment.

[0014] The rolling equipment 100 for rolling the metal strip 1 shown in FIG. 1 includes an F1 stand 10, an F2 stand 20, an F3 stand 30, an F4 stand 40, an F5 stand 50, cameras 61, 62, 63, 64, loopers 71, 72, 73, 74 for tension control, an image processing computer 80, a database 83, a control device 84, a monitor 85, etc.

[0015] These F1 stand 10, F2 stand 20, F3 stand 30, F4 stand 40, F5 stand 50, cameras 61, 62, 63, 64, image processing computer 80, database 83, control device 84, and monitor 85 are connected by a communication line 90.

[0016] Among these, a plate shape inspection and determination device for determining defects in the plate shape of the metal strip 1 rolled by the rolling mill of the present invention is constituted by the cameras 61, 62, 63, 64, the loopers 71, 72, 73, 74, and the image processing computer 80.

[0017] Note that the rolling equipment 100 is not limited to the form in which five rolling stands as shown in FIG. 1 are installed, and it may have at least two stands or more.

[0018] Each of the F1 stand 10, F2 stand 20, F3 stand 30, F4 stand 40, and F5 stand 50 includes an upper work roll and a lower work roll, an upper backup roll and a lower backup roll that support the upper work roll and the lower work roll by contacting them respectively, screw-down cylinders 11, 21, 31, 41, 51 provided above the upper backup roll, and load detectors 12, 22, 32, 42, 52. Note that a six-stage configuration in which intermediate rolls are further provided between each work roll and each backup roll can be adopted.

[0019] The looper 71 is a roll for tension control installed between the F1 stand 10 and the F2 stand 20. The rotation axis of the looper 71 extends in the width direction of the traveling metal strip 1 so that the traveling metal strip 1 is placed on the roll, and it is installed to lift and hold the metal strip 1 upward. Note that the looper 71 can be, for example, something biased upward by a spring or the like, something lifted by a hydraulic cylinder, or something lifted by motor drive or the like.

[0020] Similarly, a looper 72 for tension control is installed between the F2 stand 20 and the F3 stand 30, a looper 73 for tension control is installed between the F3 stand 30 and the F4 stand 40, and a looper 74 for tension control is installed between the F4 stand 40 and the F5 stand 50.

[0021] The camera 61 is installed to capture an image of the surface of the metal strip 1 that includes a range where a strip-shaped illumination light that traverses in the strip width direction and is called the reflected light region 2 (see FIGS. 2 and 3) can be seen reflected on the surface of the metal strip 1 lifted and curved by the rolling loop 71. Preferably, when viewed from above the metal strip 1, it can be installed outside the metal strip 1 in the strip width direction. The data of the image captured by the camera 61 is transmitted to the image processing computer 80 via the communication line 90.

[0022] In some cases, it is desirable to use a high-speed camera for this camera 61. The image acquisition speed can be, for example, 1000 to 2000000 frames per second (FPS), but this specific upper and lower limits are merely illustrative and not limited to these values. The cameras 62, 63, and 64 described later can also use high-speed cameras in the same manner as the camera 61.

[0023] Also, the camera 62 is positioned to capture an image of the surface of the metal strip 1 that includes a range where a strip-shaped illumination light that traverses in the strip width direction and is called the reflected light region 2 can be seen reflected on the surface of the metal strip 1 lifted by the rolling loop 72. The camera 63 is positioned to capture an image of the surface of the metal strip 1 that includes a range where a strip-shaped illumination light that traverses in the strip width direction and is called the reflected light region 2 can be seen reflected on the surface of the metal strip 1 lifted by the rolling loop 73. The camera 64 is positioned to capture an image of the surface of the metal strip 1 that includes a range where a strip-shaped illumination light that traverses in the strip width direction and is called the reflected light region 2 can be seen reflected on the surface of the metal strip 1 lifted by the rolling loop 74.

[0024] The data of the images captured by the cameras 62, 63, and 64 are transmitted to the image processing computer 80 via the communication line 90.

[0025] Similar to the camera 61, the cameras 62, 63, and 64 are preferably installed outside the metal strip 1 in the strip width direction when viewed from above the metal strip 1.

[0026] A photographing step is executed in which the cameras 61, 62, 63, 64 photograph an image including an area in which strip-shaped reflected light that crosses the rolled metal strip 1 in the strip width direction is captured.

[0027] It is possible to further provide lighting for illuminating the lifted photographing area of ​​the metal strip 1 lifted upward by the rolls, which is mainly photographed by the cameras 61, 62, 63, and 64. This lighting may be general lighting appropriately arranged on the ceiling of the rolling factory where the rolling equipment 100 is installed, and no special new lighting equipment is required in the present invention, but dedicated lighting may be provided.

[0028] The image processing computer 80 is a device that executes various processes for detecting the plate shape of the metal strip 1 based on the images captured by the cameras 61, 62, 63, and 64, and has a brightness data acquisition unit 81 and an image processing unit .

[0029] The luminance data acquisition unit 81 is a part that acquires the luminance value of B (blue) from the luminance data of the pixels of the images captured by the cameras 61, 62, 63, and 64. Preferably, this luminance data acquisition unit 81 is an entity that executes a luminance data acquisition step that acquires the luminance value of B (blue) from the luminance data of the pixels of the images captured by the cameras 61, 62, 63, and 64.

[0030] The image processing unit 82 is a unit that uses only the luminance data of the B (blue) luminance value acquired by the luminance data acquisition unit 81 to find a reflected light area in the image where the luminance value of B (blue) is a certain value or more, and judges whether the plate shape is defective based on the coordinates of the boundary position of the reflected light area. Preferably, this image processing unit 82 is the main unit that executes the image processing step that uses only the luminance data of the B (blue) luminance value acquired by the luminance data acquisition unit 81 to find a reflected light area in the image where the luminance value of B (blue) is a certain value or more, and judges whether the plate shape is defective based on the coordinates of the boundary position of the reflected light area.

[0031] For example, in the image processing of an image including a lifted region, which is a curved vicinity of the metal strip 1 that is rolled and lifted by the loopers 71, 72, 73, 74 as shown in FIG. 2 or FIG. 3, a range including the upstream and downstream boundaries of the portion where the luminance of the reflected light on the surface of the metal strip 1 reflected in the image is greater than a specific luminance value is specified as the reflected light region 2.

[0032] At this time, in the description described in Patent Document 1, in order to extract the coordinates of the boundary line between the upstream side and the downstream side of the reflected light region 2, the color images projected by the cameras 61, 62, 63, 64 are converted into black-and-white images (grayscale), and the grayscale luminance values (0 to 255) are used, and the region showing a luminance higher than a specific threshold value is regarded as the reflected light region.

[0033] Specifically, in the color images projected by the cameras 61, 62, 63, 64, the pixels constituting the image are combined with the luminance values of R (red), G (green), and B (blue) to form the color image. However, in the conversion to a black-and-white image (grayscale), the average value of the sums of the luminance values of R (red), G (green), and B (blue) is the luminance value converted to grayscale.

[0034] However, as a result of actually measuring the luminance of the reflected light region 2 formed on the plate surface of the rolled metal strip 1 by the inventors with a luminance meter, as shown in FIG. 4, it was found that the chromaticity, that is, the relative ratios of R (red), G (green), and B (blue), were differently distributed. That is, it was observed that the reflection of light on the surface of the metal strip 1 illuminated by the illumination had different intensities (luminances) of reflected light for each of R (red), G (green), and B (blue). From the chromaticity results in FIG. 4, it was found that the ratio of R (red) was approximately 45%, the ratio of G (green) was approximately 39%, and the ratio of B (blue) was approximately 16%.

[0035] Now, in the determination of the plate shape, the change in the reflection light region 2 that appears bright as the illumination light is reflected on the surface of the metal strip 1 is converted into a numerical value and extracted by image processing. By mathematically processing the coordinates of the boundary lines on the upstream side and the downstream side of the extracted reflection light region 2, the plate shape of the rolled strip is determined.

[0036] The determination of the plate shape is made from the first-order, second-order, and fourth-order components of the Chebyshev polynomial shown in the following equation (1). Specifically, it is determined by the coefficients (C 1 、C 2 、C 4 ). x represents the position in the plate width direction of the rolled metal strip. In equation (1), x is normalized and defined in the range from (-1) to (+1). That is, the plate width end position on the drive side (DS: drive side) is shown as (-1), and the plate width end position on the work side (WS: work side) is shown as (+1).

[0037] E(x)= C 0 +C 1 ×x+C 2 ×(2x 2 -1)+C 4 ×(8x 4 -8x 2 +1) ··· (1) In the Chebyshev polynomial, for example, as an index of the plate elongation distribution in the plate width direction due to rolling, assuming that E(x) represents the length in the rolling direction of the reflection light region 2 on the surface of the metal strip 1 at the loopers 71, 72, 73, 74, and also represents the magnitude corresponding to the plate elongation due to rolling. Note that C 0 、C 1 、C 2 、C 4It shows the values of the coefficients when the Chebyshev polynomial, which is assumed to represent the magnitude corresponding to the plate elongation distribution in the plate width direction, is separated into components of the 0th, 1st, 2nd, and 4th components of x. x represents the normalized position in the plate width direction. For example, x = -1 represents the plate width end position on the drive side (DS), and x = 1 represents the plate width end position on the working side (WS). That is, in this case, if the coefficient (C 1 ) of the 1st component of x in the Chebyshev polynomial is a positive value, it indicates that the plate elongation on the working side (WS) is large. Also, if the coefficient (C 2 ) of the 2nd component of x is a positive value, it indicates that the elongation at the plate width ends is larger than that at the plate width center. Further, if the coefficient (C 4 ) of the 4th component of x is a negative value, it indicates that the elongation in the quarter region in the plate width direction is large.

[0038] The reflection light region 2 is divided in the plate width direction by a specific number. For example, the rolling direction length of the reflection light region 2, which is used as an index of the plate elongation distribution in the plate width direction in each divided area, is averaged within the divided area, and this averaged value is assumed to represent the magnitude corresponding to the plate elongation distribution in the plate width direction, that is, the Chebyshev polynomial E(x) is set as the index.

[0039] Regarding the determination of the plate shape of the rolled metal strip 1, since it is important how accurately the coordinates of the boundary line between the upstream side and the downstream side of the reflection light region 2 formed on the surface of the metal strip 1 projected by the cameras 61, 62, 63, 64 can be extracted, the inventors focused on the coefficient C 1 , which is the coefficient of the 1st component of x in equation (1), and investigated the methods for obtaining the coordinates of the boundary line between the upstream side and the downstream side of the reflection light region, including "[1] Determining the boundary line coordinates of the reflection light region 2 using the combined luminance of RGB", "[2] Determining the boundary line coordinates of the reflection light region 2 using only the luminance value of R (red)", "[3] Determining the boundary line coordinates of the reflection light region 2 using only the luminance value of G (green)", and "[4] Determining the boundary line coordinates of the reflection light region 2 using only the luminance value of B (blue)".

[0040] In the hot rolling line, the leveling of the rolling mill on the upstream side where the looper is installed is intentionally changed to investigate the change in the coefficient C of the formula (1) which is an index of the primary component of the plate shape. 1 The change of this coefficient C 1 For the calculation of, the methods [1] to [4] above were applied, and the values of the respective coefficients C 1 were compared. Fig. 5 shows the change in the leveling value and the change in the coefficient C 1 together.

[0041] Here, the definition of the leveling amount is shown.

[0042] As shown in Figs. 6 and 7, the leveling amount is defined as the value (Gd - Gw) obtained by subtracting the distance Gw between the upper and lower work roll axes at the position of the work side (WS: Work Side) pressure cylinder from the distance Gd between the upper and lower work roll axes at the position of the drive side (DS: Drive Side) pressure cylinder.

[0043] When changing the leveling amount, it is assumed that the plate thickness at the center position of the plate width is not changed. Therefore, as shown in Fig. 7, when the leveling amount increases compared to the state shown in Fig. 6, the plate thickness on the drive side (DS) becomes thicker, the plate thickness on the work side (WS) becomes thinner, and the plate elongation on the work side (WS) becomes larger. The value of the primary component of the Chebyshev coefficient (C 1 ) indicates that when it is a positive value, the plate elongation on the work side (WS) is large, and when it is a negative value, the plate elongation on the drive side (DS) is large.

[0044] In Fig. 5, three time zones where the leveling value is stable are selected as region A, region B, and region C, and the average of the leveling values in that time zone is shown in Fig. 8, and the average of the values of the coefficient C 1 obtained by four methods in that time zone is shown in Fig. 9, and the change in the leveling value and the change in the waviness amount at that time are shown together in Fig. 10.

[0045] Here, the amount of lateral displacement is defined as follows. Measure the amount of deviation of the center position of the width of the metal strip 1 at the looper position upstream of the rolling mill stand from the center position of the pass line width, and measure the amount of deviation of the center position of the width of the metal strip 1 at the looper position downstream of the rolling mill stand from the center position of the pass line width. The value obtained by subtracting the amount of deviation of the center position of the width of the metal strip 1 at the looper position upstream from the center position of the pass line width from the amount of deviation of the center position of the width of the metal strip 1 at the looper position downstream from the center position of the pass line width is defined as the amount of lateral displacement.

[0046] Furthermore, in FIG. 10, the average lateral displacement amounts at three locations, namely regions A, B, and C, which are time zones where the leveling value is stable, are shown in FIG. 11.

[0047] As shown in FIG. 11, the average lateral displacement amounts all show negative values in regions A, B, and C. The fact that the lateral displacement amount is a negative value means that the metal strip 1 is laterally displaced toward the DS (drive side). That is, it means that the metal strip 1 is stretched more on the WS (work side) than on the DS (drive side). That is, the coefficient C of the first-order component of x of the Chebyshev polynomial 1 It is important in terms of accuracy that the value be positive in all of regions A, B, and C.

[0048] The above coefficient C 1 The fact that the value of must be positive in all of regions A, B, and C means that, looking at the value of the coefficient C in FIG. 9 1 it can be judged that the determination based on the luminance value of the B (blue) luminance value is appropriate. That is, in the determination using the luminance of the R (red) luminance value or the luminance value of the G (green) luminance value, in region A, the coefficient C of the first-order component of x of the Chebyshev polynomial 1 shows a negative value, so it is presumed to be the cause of an error in the evaluation of the coefficient C 1 1 .

[0049] ​Therefore, in the conventional method of luminance determination of RGB (blue) luminance values, the luminance value of R (red) and the luminance value of G (green) were included, resulting in a coefficient C 1 It seems that there is still room for improvement in the evaluation of . This is also true for the method using the luminance values ​​of R (red) and G (green), the method using the luminance values ​​of R (red) and B (blue), and the method using the luminance values ​​of G (green) and B (blue).

[0050] From these results, the coefficient C 1 In extracting the boundary coordinates of the reflected light area for calculation, a luminance determination method using only the luminance value of B (blue) is considered appropriate.

[0051] Therefore, the luminance data acquisition unit 81 excludes the luminance values ​​of R (red) and G (green) from the luminance data of the pixels of the images captured by the cameras 61, 62, 63, and 64, and acquires only the luminance values ​​of B (blue). Then, the image processing unit 82 uses only the luminance values ​​of B (blue) acquired by the luminance data acquisition unit 81 to find reflected light areas in the image where the luminance values ​​of B (blue) are equal to or greater than a certain value, and determines whether the plate shape is defective based on the coordinates of the boundary positions of the reflected light areas, thereby obtaining a coefficient C 1 Improve the calculation accuracy.

[0052] Furthermore, in the reflected light region 2 where a band-like reflected light can be seen transversely to the width direction of the metal band 1, if the distribution of elongation in the rolling direction (longitudinal direction of the band) at each position in the width direction of the metal band 1 is uniform, the surface of the metal band 1 is flat, the shape is good, and the steepness in the width direction is approximately the same and small, so that the distribution difference in the length in the rolling direction of the reflected light region 2 at each position in the width direction is small, as shown in Figure 2. For this reason, the boundary of the reflected light region 2 by illumination is such that the distance between the upstream boundary line 2A and the downstream boundary line 2B is approximately uniform and approximately parallel in the width direction, and parameters such as the area value of each zone and the average length in the rolling direction when the reflected light region 2 is equally divided into a plurality of zones in the width direction are approximately uniform in all zones.

[0053] On the other hand, when there are differences in elongation in the rolling direction (e.g., edge elongation, middle elongation) depending on the position in the strip width direction, the metal strip is not flat, the strip shape is defective, and due to differences in steepness in the strip width direction caused by strip waves or the like, the regions reflected on the strip surface of the reflected light are different. As shown in FIG. 3, the boundary line of the reflected light region 2 by illumination is either one or both of the upstream boundary line 2A and the downstream boundary line 2B that undulate, and the interval between the upstream boundary line 2A and the downstream boundary line 2B of the reflected light region 2 becomes non-uniform in the strip width direction. For this reason, for example, parameters such as the area value of each region when the reflected light region 2 is evenly divided into a plurality of regions in the strip width direction and the average length in the rolling direction are non-uniform in each region.

[0054] Furthermore, when fitting the Chebyshev polynomial to the strip width direction distribution of the maximum length in the rolling direction of the reflected light region 2 reflected on the surface of the metal strip 1 near the curved portion of the metal strip 1 lifted by the loopers 71, 72, 73, 74 installed between the rolling mill stands of the rolling line to determine the strip elongation distribution due to rolling in the strip width direction, the image processing unit 82 obtains the coordinates of the pixel points regarding the boundary positions in the strip longitudinal direction of the reflected light region. When the value indicating the strip width direction position is taken as the variable (x), the position in the strip width direction within the strip width range of the reflected light region in the image is normalized to the range of -1 ≦ x ≦ 1, and from the change in the strip longitudinal direction coordinates of the pixel points of the boundary positions of the reflected light region, an index value related to the strip width direction elongation distribution of the metal strip 1 is calculated. The index values at least at four or more positions of (x) are obtained, and the Chebyshev polynomial E(x) = C 0 +C 1 ×x + C 2 ×(2x 2 -1) + C 4 ×(8x 4 -8x 2 +1) (where -1 ≦ x ≦ 1) is applied for curve fitting, and the coefficients (C 0 , C 1 , C 2 , C 4 ) of the Chebyshev polynomial are obtained, and it is desirable to determine the defectiveness of the strip shape based on the coefficient (C 1 ) of the term of the first-order component.

[0055] Hereinafter, the method for extracting the "index value related to the elongation distribution in the strip width direction of the metal strip 1" will be briefly described.

[0056] The index value related to the elongation distribution in the strip width direction of the metal strip 1 is specifically any one of "the maximum value of the distance between the upstream boundary line 2A and the downstream boundary line 2B of the reflected light region 2", "the average value of the maximum fluctuation amplitude of the upstream boundary line 2A and the maximum fluctuation amplitude of the downstream boundary line 2B of the reflected light region 2", and "the average value of the oscillation frequencies of the upstream boundary line 2A and the downstream boundary line 2B of the reflected light region 2".

[0057] In the image processing unit 82 of the image processing computer 80 in this embodiment, for each of the acquired images, using the acquired images, preferably, the calculation process of the index value related to the elongation distribution in the strip width direction of any one of the following three metal strips 1 is performed.

[0058] The method for obtaining "the maximum value of the distance between the upstream boundary line 2A and the downstream boundary line 2B of the reflected light region 2" is as follows.

[0059] The image processing unit 82 numbers the acquired images in order as 1, 2, 3, ···, k, ···. In the (k)-th image, within each divided area (i) (i = 1 to j) obtained by dividing the reflected light region 2 in the strip width direction, the rolling direction position Pumin(k)i of the pixel that is the most upstream among the pixels constituting the upstream boundary line 2A of the reflected light region 2 is obtained. Similarly, the rolling direction position Pdmax(k)i of the pixel that is the most downstream among the pixels constituting the downstream boundary line 2B of the reflected light region 2 is obtained. This data is stored in the database 83.

[0060] Further, for each of the F images from image number (k-F+1) to image number (k) stored in the database 83, the image processing unit 82 extracts the first position data that is the most upstream position in the rolling direction within the area of the divided area (i) from among the F data from Pumin(k-F+1)i to Pumin(k)i, and designates it as Pure_min(k)i as the data of the (k)th image. Also, the second position data that is the most downstream position in the rolling direction within the area of the divided area (i) is extracted from among the F data from Pdmax(k-F+1)i to Pdmax(k)i, and designates it as Pdre_max(k)i as the data of the (k)th image. These are stored in the database 83.

[0061] The difference [Pdre_max(k)i - Pure_min(k)i] of this position data in the rolling direction is calculated for each divided area (i), and the obtained j values of [Pdre_max(k)i - Pure_min(k)i] are represented as D(k)i as an index A (the maximum length in the rolling direction of the reflected light area 2) corresponding to the elongation of the rolled plate of the (k)th image. When the value indicating the position in the plate width direction is made the variable (x) corresponding to the center position of each divided area (i) in the plate width direction, D(k)i corresponding to the position x normalized to the range of -1 ≦ x ≦ 1 of the position in the plate width direction within the plate width range is defined as E(xi) of the (k)th image. Note that xi is the center position in the plate width direction of the divided area (i) indicated in the (x) notation.

[0062] The method for obtaining the "average value of the maximum fluctuation amplitude of the upstream boundary line 2A and the maximum fluctuation amplitude of the downstream boundary line 2B of the reflected light area 2" is as follows.

[0063] The image processing unit 82 numbers the acquired images in order as 1, 2, 3, ···, k, ···. In the (k)-th image, within each divided area (i) (i = 1 to j) where the reflected light area 2 is divided into j in the plate width direction, the rolling direction position Pumin(k)i of the pixel that is the most upstream among the positions of the pixels constituting the upstream boundary line 2A of the reflected light area 2 is obtained. The rolling direction position Pumax(k)i of the pixel that is the most downstream among the positions of the pixels constituting the upstream boundary line 2A of the reflected light area 2 is obtained. The rolling direction position Pdmin(k)i of the pixel that is the most upstream among the positions of the pixels constituting the downstream boundary line 2B of the reflected light area 2 is obtained. And the rolling direction position Pdmax(k)i of the pixel that is the most downstream among the positions of the pixels constituting the downstream boundary line 2B of the reflected light area 2 is obtained. This data is stored in the database 83.

[0064] Also, for each of the F images from the image number (k - F + 1) to the image number (k) stored in the database 83, the image processing unit 82 extracts the first position data that is the most upstream in the rolling direction within the area of the divided area (i) from among the F pieces of data from Pumin(k - F + 1)i to Pumin(k)i, and sets it as Pure_min(k)i as the data of the (k)-th image. Also, the second position data that is the most downstream in the rolling direction within the area of the divided area (i) is extracted from among the F pieces of data from Pumax(k - F + 1)i to Pumax(k)i, and set as Pure_max(k)i as the data of the (k)-th image. Also, the third position data that is the most upstream in the rolling direction within the area of the divided area (i) is extracted from among the F pieces of data from Pdmin(k - F + 1)i to Pdmin(k)i, and set as Pdre_min(k)i as the data of the (k)-th image. Also, the fourth position data that is the most downstream in the rolling direction within the area of the divided area (i) is extracted from among the F pieces of data from Pdmax(k - F + 1)i to Pdmax(k)i, and set as Pdre_max(k)i as the data of the (k)-th image. These data are stored in the database 83.

[0065] Furthermore, the image processing unit 82 calculates the difference [Pure_max(k)i - Pure_min(k)i] in the rolling direction position data of the upstream boundary line 2A in the (k)-th image for each divided region (i), and obtains the obtained j values of [Pure_max(k)i - Pure_min(k)i] as the maximum fluctuation amplitude "Au(k)i" of the upstream boundary line 2A of the (k)-th image. At the same time, the difference [Pdre_max(k)i - Pdre_min(k)i] in the rolling direction position data of the downstream boundary line 2B in the (k)-th image is calculated for each divided region (i), and the obtained j values of [Pure_max(k)i - Pure_min(k)i] are obtained as the maximum fluctuation amplitude "Ad(k)i" of the downstream boundary line 2B of the (k)-th image. Then, the average value of the maximum fluctuation amplitudes "Au(k)i" and "Ad(k)i" in the rolling direction of the upstream boundary line 2A and the downstream boundary line 2B is calculated for each divided region (i) and represented as A(k)i. A(k)i is defined as the index B (average maximum fluctuation amplitude in the divided region (i) of the upstream boundary line 2A and the downstream boundary line 2B of the reflected light region 2) of the (k)-th image. When the value of A(k)i is corresponded to the position at the center in the plate width direction of the divided region (i) and the value indicating the position in the plate width direction is the variable (x), A(k)i corresponding to the normalized position x in the range of -1 ≦ x ≦ 1 for the position in the plate width direction within the plate width range is defined as E(xi) of the (k)-th image.

[0066] The method for obtaining the "average value of the vibration frequencies of the upstream boundary line 2A and the downstream boundary line 2B of the reflected light region 2" is as follows.

[0067] In the image processing unit 82, for each of the acquired images, the following five analysis processes are performed using the acquired images.

[0068] The image processing unit 82 numbers the acquired images in order as 1, 2, 3, ···, k, ···, and in the (k)-th image, the pixel position Pu(k)p (p = 1 to Ru, where Ru is the number of pixels constituting the upstream boundary line 2A of the reflected light region 2) in the rolling direction of the upstream pixel points constituting the upstream boundary line 2A of the reflected light region 2, and the pixel position Pd(k)q (q = 1 to Rd, where Rd is the number of pixels constituting the downstream boundary line 2B of the reflected light region 2) in the rolling direction of the downstream pixel points constituting the downstream boundary line 2B of the reflected light region 2 are obtained. This data is preferably stored in the database 83.

[0069] Here, as described above, since the boundary line of the reflected light region 2 varies during rolling, the image processing unit 82 calculates the moving distance Mu(k)p (= Pu(k)p - Pu(k - 1)p) in the rolling direction of the pixel points on the upstream boundary line 2A of the reflected light region 2 from the pixel position Pu(k - 1)p of the image number (k - 1) and the pixel position Pu(k)p of the image number (k) stored in the database 83, and calculates the moving distance Md(k)q (= Pd(k)q - Pd(k - 1)q) in the rolling direction of the pixel points on the downstream boundary line 2B of the reflected light region 2 from the pixel position Pd(k - 1)q of the image number (k - 1) and the pixel position Pd(k)q of the image number (k).

[0070] Furthermore, the image processing unit 82 reconsiders the moving distance Mu(k)p in the rolling direction of the pixel points on the upstream boundary line 2A of the reflected light region 2 within the range of the divided area (i) (i = 1 to j) obtained by dividing the reflected light region 2 in the plate width direction, and obtains the average value of the moving distance Mu(k)p in the rolling direction of the pixel points on the upstream boundary line 2A existing within the range of the divided area (i), and sets it as the average moving distance Mu(k)i for the pixel points constituting the upstream boundary line 2A of the reflected light region 2 in the divided area (i). Similarly, the moving distance Md(k)q in the rolling direction of the pixel points on the downstream boundary line 2B is obtained as the average value of Md(k)q existing within the range of the divided area (i), and is set as the average moving distance Md(k)i for the pixel points constituting the downstream boundary line 2B of the reflected light region 2 in the divided area (i).

[0071] Next, the image processing unit 82 divides the reflected light region 2 into j parts in the plate width direction, and divides the average movement distance Mu(k)i of the pixel points constituting the upstream boundary line 2A of the reflected light region 2 in the divided area (i) by the time Ts required from the (k-1)-th image acquisition to the k-th image acquisition, and the average movement distance Md(k)i of the pixel points constituting the downstream boundary line 2B of the reflected light region 2 in the divided area (i). The average movement speed Vu(k)i (=Mu(k)i / Ts) of the pixel points constituting the upstream boundary line 2A of the reflected light region 2 for each divided area (i), and the average movement speed Vd(k)i (=Md(k)i / Ts) of the pixel points constituting the downstream boundary line 2B are calculated.

[0072] Next, the image processing unit 82 examines how many times the positive and negative signs (+ / -) of the average movement speed Vu(k)i of the pixel points constituting the upstream boundary line 2A and the average movement speed Vd(k)i of the pixel points constituting the downstream boundary line 2B have reversed during the time Tf required from the (k-F+1)-th image acquisition to the k-th image acquisition, and determines the number of sign reversals Nu(k)i of the average movement speed Vu(k)i of the pixel points constituting the upstream boundary line 2A of the reflected light region 2 in the divided area (i), and the number of sign reversals Nd(k)i of the average movement speed Vd(k)i of the pixel points constituting the downstream boundary line 2B of the reflected light region 2.

[0073] After that, the image processing unit 82 obtains the average fluctuation frequency Qu(k)i ([Nu(k)i / Tf] / 2) of the pixel points constituting the upstream boundary line 2A of the reflected light region 2 and the average fluctuation frequency Qd(k)i ([Nd(k)i / Tf] / 2) of the pixel points constituting the downstream boundary line 2B from the number of sign reversals Nu(k)i and Nd(k)i, and further, as the average fluctuation frequency of the pixel points constituting the upstream boundary line 2A and the downstream boundary line 2B forming the reflected light region 2 for each divided area (i), Q(k)i ([Qu(k)i+Qd(k)i] / 2) is obtained.

[0074] Furthermore, the image processing unit 82 sets the value of the obtained information Q(k)i as an index C (the average value of the vibration frequencies of the upstream boundary line 2A and the downstream boundary line 2B of the reflected light region 2) corresponding to the elongation of the rolled plate in the (k)-th image, and when the value indicating the position in the plate width direction is the variable (x), it normalizes the position in the plate width direction within the plate width range to the range of -1 ≦ x ≦ 1, and corresponds the central position of each divided area (i) in the plate width direction to xi, and sets Q(k)i as the Chebyshev polynomial E(xi) = Q(k)i of Equation (1) at the time of the (k)-th image.

[0075] For the index A, or index B, or index C corresponding to the elongation of the rolled plate in the above-mentioned (k)-th image, the image processing unit 82 applies the values of the central position xi in the plate width direction and E(xi) (E(xi) = D(k)i, A(k)i, Q(k)i) of each divided area (i) of the (k)-th image to the Chebyshev polynomial of Equation (1) for each divided area (i), performs curve fitting using the least squares method, and from the equation of the approximate curve, obtains the coefficients (C 0 、C 1 、C 2 、C 4 ) of the degree of x of the Chebyshev polynomial, and transmits the first-order coefficient (C 1 ) as a detection result signal of the elongation distribution of the rolled plate in the first-order component in the plate width direction.

[0076] Preferably, in addition to the first-order coefficient (C 1 ), the image processing unit 82 can further transmit, as a detection result signal of the elongation distribution of the rolled plate in the second-order component or the fourth-order component in the plate width direction, any one or more of the second-order coefficient (C 2 ) or the fourth-order coefficient (C 4 ).

[0077] The database 83 also functions as a recording medium in which various parameters used when operating the rolling facility 100 are recorded.

[0078] Here, regarding the method of dividing the reflected light region 2, each divided area (i) (i = 1 to j) divided into j in the plate width direction may be equally divided into j, or each of the j may be unevenly divided into an arbitrary width.

[0079] Returning to FIG. 1, the control device 84 is a device that controls the operation of each device in the rolling facility 100. In this embodiment, it is a device that executes various controls according to the detection of the plate shape of the metal strip 1 by the image processing computer 80.

[0080] These image processing computer 80, database 83, and control device 84 can be configured by a computer having a monitor 85 such as a liquid crystal display, an input device, a storage device, a CPU, a memory, etc., which will be described later. They may be configured by one computer, or may be configured by separate computers, and are not particularly limited.

[0081] The control of the operation of each device by the image processing computer 80 and the control device 84 is executed based on various programs recorded in the storage device. Note that the control processing of the operations executed by the image processing computer 80 and the control device 84 may be combined into one program, may be separated into multiple programs respectively, or may be a combination thereof. Also, part or all of the program may be realized by dedicated hardware and may be modularized.

[0082] The monitor 85 is a display device such as a display or an acoustic device such as an alarm. For example, when the image processing computer 80 detects that there is a problem with the plate shape, it is a device for conveying the corresponding countermeasure work to the operator. Therefore, such a monitor 85 is often a display.

[0083] Here, the above-mentioned image processing computer 80 includes a display signal section and transmits signals regarding the content to be displayed on the monitor 85 to the control device 84 and the monitor 85.

[0084] During operation, the operator can confirm the state of the plate shape by visually observing the display screen of the monitor 85, each stand itself, and between each stand.

[0085] Note that the present invention is not limited to a form in which the operator is notified of the occurrence of a problem in the plate shape and the control device 84 automatically performs an operation to improve the problem in the plate shape. Instead, it can be a form in which it is only displayed on the monitor 85, or a form in which the display on the monitor 85 is omitted and the control device 84 automatically performs an operation to improve the problem in the plate shape only.

[0086] Next, the plate shape inspection and judgment device and the flow of the judgment method for the metal strip 1 to be rolled in the present invention will be described with reference to FIG. 12. FIG. 12 shows the judgment method for the plate shape of the present invention in a flowchart.

[0087] As shown in FIG. 12, before the start of rolling or during rolling, in the image processing computer 80, a judgment image number F (for example, F = 20 or F = 1000) based on an index value related to the elongation distribution in the plate width direction of the metal strip 1 in the reflection light region 2 of the reflection light region 2 is set (step S201), and an image number k is set (step S202). Here, the initial value of the image number is k = 1.

[0088] Thereafter, the image processing computer 80 acquires a surface image (image number k) of the metal strip 1 near the curved portion of the metal strip 1 lifted by the loopers 71, 72, 73, 74 photographed by the cameras 61, 62, 63, 64 (step S203). Here, the acquired images are numbered in the order of acquisition as 1, 2, 3, ···, k, ···.

[0089] Next, the luminance data acquisition unit 81 of the image processing computer 80 performs image processing on the image (image number k) photographed by the cameras 61, 62, 63, 64 to extract the reflection light region 2 on the surface of the metal strip 1 near the curved portion of the metal strip 1 (step S204).

[0090] For example, the luminance data acquisition unit 81 of the image processing computer 80 obtains only the luminance values of all B (blue) among the R (red) luminance values, G (green) luminance values, and B (blue) luminance values of the pixels (picture elements) reflected in the selected rolling surface image range among the images captured by the cameras 61, 62, 63, and 64, and determines the pixel coordinates that constitute the upstream boundary line 2A and the downstream boundary line 2B seen in the plate width direction on the upstream side and the downstream side of the reflected light region 2 projected on the surface of the metal strip 1, thereby identifying the reflected light region 2.

[0091] Next, the image processing unit 82 of the image processing computer 80 executes the first-stage calculation process of an index value related to the elongation distribution in the plate width direction of the metal strip 1, preferably index A "the maximum value of the distance between the upstream boundary line 2A and the downstream boundary line 2B of the reflected light region 2", or index B "the average value of the maximum fluctuation amplitude of the upstream boundary line 2A of the reflected light region 2 and the maximum vibration amplitude of the downstream boundary line 2B", or index C "the average value of the vibration frequencies of the upstream boundary line 2A and the downstream boundary line 2B of the reflected light region 2" (step S205).

[0092] Next, the image processing unit 82 of the image processing computer 80 determines whether the image number k < F (step S206). When it is determined that the image number k < F, the process proceeds to step S207, where the image number k is updated (k = k + 1) (step S207), and the process proceeds to step S203, where it waits for the processing of a predetermined number of images to be performed. On the other hand, when it is determined in step S206 that the image number k ≧ F, the process proceeds to step S208.

[0093] Next, the image processing unit 82 of the image processing computer 80 executes the second-stage calculation process of the index value related to the elongation distribution in the plate width direction of the metal strip 1 (step S208). Details thereof are omitted.

[0094] Next, the image processing unit 82 of the image processing computer 80 associates the index value in each divided area (i) when the reflected light area 2 is divided in the plate width direction with the center position (xi) in the plate width direction of the divided area (i), represents E(xi) expressed as E(xi) in terms of (x) to the Chebyshev polynomial of formula (1) for curve fitting, and calculates the coefficients (C 0 、C 1 、C 2 、C 4 ) of the x-degree of the Chebyshev polynomial from the obtained approximate formula (step S209).

[0095] Next, the image processing unit 82 of the image processing computer 80 transmits, as a detection result signal of the plate elongation distribution in the (k)-th image, the Chebyshev coefficients (C 1 、C 2 、C 4 ) obtained in step S209 to, for example, the control device 84 or the monitor 85 (step S210).

[0096] Thereafter, the image processing computer 80 determines whether rolling is continuing (step S211). When it is determined that rolling is continuing, the process returns to step S207, the image number (k) is updated as (k = k + 1), and the plate shape detection process is continued. On the other hand, when it is determined that rolling is completed, the process ends.

[0097] In this embodiment, it is assumed that the third-order component is not adopted in the Chebyshev polynomial. This is because the rolling control mechanism of the rolling mill is not adapted to correct the plate elongation of the third-order component. By omitting the arithmetic processing and corresponding means of the third-order component, it is easier to judge the situation of the rolled plate shape and correct the plate elongation of the separated first-order component, second-order component, and fourth-order component.

[0098] The image processing computer 80 can output a control command signal to the control device 84 to correct the leveling, bending force, pair crossing angle, etc. based on the polynomial approximation result in the plate width direction obtained using the Chebyshev polynomial of formula (1). Further, alternatively, or in addition, by outputting a display command signal for performing a guidance display necessary for correcting the leveling, bending force, pair crossing angle, etc. to the monitor 85, the operator can be informed of correction information such as the leveling, bending force, pair crossing angle, etc.

[0099] Preferably, the image processing computer 80 is the vector of each order term in the Chebyshev polynomial E(x) of the above formula (1) (C 0 、C 1 、C 2 、C 4 ), the 0th component of the function of x [C 0 , the 1st component of x [C 1 ×x], the 2nd component of x [C 2 ×(2x 2 -1)], the 4th component of x [C 4 ×(8x 4 -8x 2 +1)], a signal can be transmitted to the monitor 85 to display the graph of each component term.

[0100] The operator can check the display screen of the monitor 85 and perform an operation to correct, for example, the leveling, bending force, pair crossing angle (in the case of a pair crossing rolling mill), etc.

[0101] Next, the effects of this embodiment will be described.

[0102] A flatness determination device for determining flatness defects of the metal strip 1 rolled by the rolling mill of the present embodiment described above is installed with a rotating shaft extending in the width direction of the metal strip 1, and loopers 71, 72, 73, 74 that lift the metal strip 1 upward, cameras 61, 62, 63, 64 that capture an image including the lifted area of the metal strip 1 lifted upward by the loopers 71, 72, 73, 74 (imaging step), a luminance data acquisition unit 81 that acquires the luminance value of B (blue) from the luminance data of the pixels of the image captured by the cameras 61, 62, 63, 64 (luminance data acquisition step), and an image processing unit 82 that obtains a specular reflection area where the luminance value of B (blue) is equal to or greater than a certain value in the image using only the luminance data of the luminance value of B (blue) acquired by the luminance data acquisition unit 81, and determines flatness defects based on the coordinates of the boundary position of the specular reflection area (image processing step).

[0103] As a result, compared with the conventional method, the change in the coordinates of the boundary line of the specular reflection area can be accurately extracted, so that the flatness determination accuracy can be improved more than before.

[0104] In addition, the image processing unit 82 obtains the coordinates of the pixel points regarding the boundary position in the strip longitudinal direction of the specular reflection area. When the value indicating the strip width direction position is a variable (x), the position in the strip width direction within the strip width range of the specular reflection area in the image is normalized to the range of -1 ≤ x ≤ 1, and from the change in the strip longitudinal direction coordinates of the pixel points of the boundary position of the specular reflection area, an index value related to the strip width direction elongation distribution of the metal strip 1 is calculated, and the respective index values at at least four or more positions of a plurality of (x) are obtained, and E(x) = C 0 +C 1 ×x + C 2 ×(2x 2 -1) + C 4 ×(8x 4 -8x 2 +1) (where -1 ≤ x ≤ 1) is applied to the Chebyshev polynomial, curve fitting is performed, and the coefficients (C 0 , C 1 , C 2 , C 4 ) of the Chebyshev polynomial are obtained, and the coefficient (C 1By determining the sheet shape defect based on , the change in the sheet elongation distribution can be evaluated with high accuracy.

[0105] Furthermore, the image processing unit 82 further uses the coefficient (C 2 ) of the quadratic component term or the coefficient (C 4 ) of the quartic component term to determine the sheet shape defect, so that it can cope with more changes in the sheet shape.

[0106] <Others> Note that the present invention is not limited to the above embodiments, and various modifications and applications are possible. The above-described embodiments have been described in detail for easy understanding of the present invention, and are not necessarily limited to those having all the configurations described.

Explanation of Signs

[0107] 1... Metal strip 2... Reflection light region 2A... Upstream boundary line of the reflection light region 2B... Downstream boundary line of the reflection light region 10... F1 stand 11, 21, 31, 41, 51... Pressing cylinder 12, 22, 32, 42, 52... Load detector 20... F2 stand 30... F3 stand 40... F4 stand 50... F5 stand 61, 62, 63, 64... Camera 71, 72, 73, 74... Looper 80... Image processing computer 81... Luminance data acquisition unit 82... Image processing unit 83... Database 84... Control device 85... Monitor 90... Communication line 100... Rolling equipment

Claims

1. A plate shape judgment device for judging defects in the plate shape of a metal strip rolled by a rolling mill, comprising: a roll having a rotating shaft extending in a width direction of the metal strip and lifting the metal strip upward; a camera that captures an image including a lifted area of ​​the metal strip lifted upward by the roll; a brightness data acquisition unit that acquires a brightness value of B (blue) from brightness data of R (red), G (green), and B (blue) of a pixel of an image captured by the camera; and an image processing unit that uses only the luminance data of the luminance value of the B (blue) acquired by the luminance data acquisition unit to obtain a reflected light area in the image where the luminance value of the B (blue) is a certain value or more, and judges defects in the plate shape based on the coordinates of the boundary position of the reflected light area. Plate shape judgment device.

2. The plate shape determination device according to claim 1, The image processing unit includes: The coordinates of the pixel points at the boundary positions of the reflected light regions in the plate longitudinal direction are obtained; When a value indicating a position in the sheet width direction is a variable (x), the positions in the sheet width direction within the sheet width range of the reflected light region in the image are normalized to a range of -1≦x≦1, and an index value related to the sheet width direction elongation distribution of the metal strip sheet is calculated from the change in the sheet longitudinal direction coordinate of the coordinates of the pixel point of the boundary position of the reflected light region, and each index value at a plurality of positions (x) of at least four points or more is obtained; E(x) = C ０ +C １ ×x+C ２ × (2x ２ -1) + C ４ × (8x ４ -8x ２ +1) (where -1≦x≦1) and perform curve fitting. The coefficient of the degree of x of the Chebyshev polynomial (C ０ , C １ , C ２ , C ４ ) and calculate the coefficient of the first-order component (C １ ) and judge the plate shape defect based on the Plate shape judgment device.

3. In the plate shape determination device according to claim 2, The image processing unit further calculates the coefficient of the second-order component (C ２ ) or the coefficient of the fourth-order component (C ４ ) to judge the defects of the plate shape. Plate shape judgment device.

4. A method for determining a defect in the shape of a metal strip rolled by a rolling mill, comprising: an imaging step of taking an image including a lifted area of ​​the metal strip lifted upward by a roll that is installed with a rotation axis extending in the width direction of the metal strip and lifts the metal strip upward, using a camera; a brightness data acquisition step of acquiring a brightness value of B (blue) from brightness data of R (red), G (green), and B (blue) of a pixel of the image captured in the imaging step by an image processing computer; and an image processing step of determining a reflected light area in the image where the luminance value of B (blue) is equal to or greater than a certain value by using only the luminance data of the luminance value of B (blue) acquired in the luminance data acquisition step, and determining defects in the plate shape by an image processing computer based on the coordinates of the boundary position of the reflected light area. How to judge plate shape.

Images