Metal strip shape determination device, rolling mill and determination method

Abstract

The present invention provides a sheet shape determination device for metal strip that is less susceptible to minor disturbances or sudden small obstacles compared to the past, as well as a rolling mill and determination method. The sheet shape determination device for the rolled metal strip (1) includes: cameras (61, 62, 63, 64) configured to capture an image of the rolled metal strip (1) including a region reflecting a strip of reflected light transversely along the width direction; and an image processing computer (80) that determines the sheet shape of the metal strip (1) based on the image captured by the cameras (61, 62, 63, 64). The image processing computer (80) divides the region in the image into multiple partitions along the width direction of the metal strip (1), and sends information corresponding to the distribution of the strip waviness in the rolling direction along the width direction as a signal based on index information representing the size of the region in each partition.

Description

Technical Field

[0001] This invention relates to a sheet shape determination device for metal strips, as well as a rolling mill and determination method. Background Technology

[0002] Patent Document 1 describes a defect detection device and method that can easily determine surface shape defects of a metal strip without using a special light source such as a rod. The device includes: a roller with its rotation axis extending along the width direction of the rolled steel sheet to lift the rolled steel sheet upward; a camera that captures an image of the rolled steel sheet lifted upward by the roller, including the lifted area; and a control device that determines surface shape defects of the metal strip 1 based on the image captured by the camera.

[0003] Existing technical documents

[0004] Patent Literature

[0005] Patent Document 1: Japanese Patent No. 6808888 Summary of the Invention

[0006] Regarding the quality of the shape of the metal strip rolled by the rolling mill, such as whether it has a waviness, there are many known techniques that rely on linear or rod-shaped reflected light along the width of the metal strip for judgment.

[0007] This judgment is based on the following: when a part of the plate develops a wavy shape and the shape of the plate changes, the shape of the reflected light, which was originally linear or rod-shaped, is no longer a regular shape, and a part of it will move or shift.

[0008] However, the following problem exists: because the area of ​​reflected light in the rolling direction at each position in the width direction of the plate is narrow, the reflected light in the linear or rod-shaped directions is significantly affected by subtle interference caused by, for example, small obstacles, which can easily lead to misjudgment.

[0009] When the inventors of this case used the strip-shaped reflected light disclosed in Patent Document 1 to make judgments, they found that they could reduce the influence of such interference. As a result, they conducted research and came up with the invention of this case, which further makes full use of the characteristics of the strip-shaped reflected light.

[0010] The present invention provides a sheet shape determination device, a rolling mill, and a determination method for metal strips that are less susceptible to minor disturbances and sudden small obstacles compared to the past.

[0011] This invention includes several solutions to the aforementioned problems. One example is a sheet shape determination device for rolled metal strip, comprising: a camera configured to capture an image of a region on the surface of the rolled metal strip that reflects a strip of light transversely along the strip width direction; and an image processing unit that determines the sheet shape of the metal strip based on the image captured by the camera. The sheet shape determination device is characterized in that, when the image processing unit divides the region in the image into multiple partitions along the strip width direction and sets a value indicating the position of each partition in the strip width direction as a variable (x), it is suitable for determining the strip width of the region in the image. The plate width direction position within the range is converted into a range of -1≤x≤1, and the index information representing the size of the above-mentioned region in each of the above-mentioned partitions is set as the Chebyshev polynomial of the distribution E(x) of each of the above-mentioned partitions, which consists only of the 0th, 1st, 2nd and 4th degree terms of x. The coefficients (C0', C1', C2', C4') of the above-mentioned Chebyshev polynomial are obtained as information corresponding to the distribution of plate wavy shape in the rolling direction in the plate width direction. One or more of the above-mentioned coefficients, namely the 0th degree coefficient (C0'), the 1st degree coefficient (C1'), the 2nd degree coefficient (C2') and the 4th degree coefficient (C4'), are sent as the judgment result signal of the plate wavy shape distribution in the plate width direction. E(x)=C0'+C1'×x+C2'×(2x 2 -1)+C4'×(8x 4 -8x 2 +1), where -1≤x≤1···(1)

[0012] Invention Effects

[0013] According to the present invention, a sheet shape determination device, a rolling mill, and a determination method for metal strips are provided that are less susceptible to minor disturbances and sudden small obstacles compared to conventional methods. Other issues, structures, and effects beyond those described above will become clear through the following description of embodiments. Attached Figure Description

[0014] Figure 1 This is a schematic diagram showing a rolling mill equipped with a sheet shape determination device for metal strip according to an embodiment of the present invention.

[0015] Figure 2 This is a diagram illustrating an example of the condition of the metal strip between stands during operation in a rolling mill.

[0016] Figure 3 This is another example of the condition of the metal strip between stands during operation in a rolling mill.

[0017] Figure 4This is a diagram illustrating an example of the method for calculating the average length within each segment of the metal strip shape determination device in the embodiment, and the distribution of components of the Chebyshev polynomial.

[0018] Figure 5 This is a diagram illustrating an example of the area calculation method and the component distribution of the Chebyshev polynomial in each segmented partition of the metal strip shape determination device of the embodiment.

[0019] Figure 6 This is a diagram illustrating an example of the method for calculating the median length within each segment of the metal strip shape determination device in the embodiment, and the distribution of components of the Chebyshev polynomial.

[0020] Figure 7 This is a diagram illustrating an example of the distribution of the distance between the upstream and downstream boundary lines of each segment of the reflected light region when the metal strip is divided into seven parts along the width direction in the metal strip shape determination device of the embodiment.

[0021] Figure 8 This is an example of the display screen of the monitor in the metal strip shape determination device of the embodiment.

[0022] Figure 9 This is a diagram illustrating the plate shape control block of the primary component in the plate shape determination device for a metal strip plate according to an embodiment.

[0023] Figure 10 This is a diagram illustrating the plate shape control block of the secondary component in the plate shape determination device for the metal strip plate of an embodiment.

[0024] Figure 11 This is a diagram illustrating the four-component plate shape control block in the plate shape determination device for a metal strip plate according to an embodiment.

[0025] Figure 12 This is a diagram illustrating an example of the configuration relationship between a metal strip and a camera in three-dimensional space.

[0026] Figure 13 It means to Figure 12 A diagram showing the state of the metal strip plate and the camera projected onto the two-dimensional XY plane.

[0027] Figure 14 It means to Figure 12 The diagram shows the state of the metal strip plate and the camera projected onto the two-dimensional plane XZ.

[0028] Figure 15 This is a diagram showing the correction status of the boundary line position of the reflected light area in the metal strip shape determination device of the embodiment.

[0029] Figure 16 This is a diagram illustrating an example of the relationship between the distance (D) from the metal strip (in the case of a strip width of 2m) to the camera and the perspective correction ratio α (0) in the metal strip shape determination device of the embodiment.

[0030] Figure 17 This is a diagram illustrating an example of the relationship between the distance (D) from the metal strip (with a strip width of 1m) to the camera and the perspective correction ratio α (0) in the metal strip shape determination device of the embodiment.

[0031] Figure 18 This is the detection and control process of the plate shape in the plate shape determination device of the metal strip plate in the embodiment.

[0032] Description of Reference Numerals

[0033] 1…Metal strip

[0034] 1A, 1B... Reflected light areas

[0035] 10…F1 rack

[0036] 11, 21, 31, 41, 51… Press down cylinder

[0037] 12, 22, 32, 42, 52… Load detectors

[0038] 20…F2 rack

[0039] 30…F3 rack

[0040] 40…F4 rack

[0041] 50…F5 rack

[0042] Cameras 61, 62, 63, 64…

[0043] 71, 72, 73, 74… Looping (rollers)

[0044] 80… Image Processing Computer (Image Processing Department)

[0045] 82…Control device

[0046] 85…monitor

[0047] 90… communication line

[0048] 100…rolling equipment Detailed Implementation

[0049] use Figures 1 to 18This invention describes embodiments of the metal strip shape determination device, rolling mill, and determination method of the present invention. It should be noted that in the drawings used in this specification, there are instances where the same or corresponding structural elements are labeled with the same or similar reference numerals, and repeated descriptions of these structural elements are omitted.

[0050] First, use Figures 1 to 3 Describe the overall structure of the rolling equipment, including the sheet shape determination device for metal strip. Figure 1 This is a schematic diagram showing the structure of the metal strip shape determination device and the rolling equipment equipped with the shape determination device according to this embodiment. Figure 2 and Figure 3 This is a diagram illustrating an example of the condition of the metal strip surface between stands during operation in a rolling mill.

[0051] Figure 1 The rolling equipment 100 for rolling metal strip 1 shown 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, and 64, loopers 71, 72, 73, and 74 for tension control, an image processing computer 80, a control device 82, and a monitor 85. Furthermore, the F1 stand 10, F2 stand 20, F3 stand 30, F4 stand 40, F5 stand 50, cameras 61, 62, 63, and 64, the image processing computer 80, and the control device 82 are connected via a communication line 90.

[0052] The metal strip shape determination device of the present invention consists of cameras 61, 62, 63, 64, tension control loopers 71, 72, 73, 74, and image processing computer 80.

[0053] Furthermore, the rolling mill 100 is not limited to... Figure 1 The configuration shown has five rolling mill stands, but at least two stands are acceptable.

[0054] Frames F1 (10), F2 (20), F3 (30), F4 (40), and F5 (50) each include upper and lower work rolls, upper and lower support rolls supported by contact with these upper and lower work rolls, pressure cylinders 11, 21, 31, 41, and 51 located above the upper support rolls, and load detectors 12, 22, 32, 42, and 52. Furthermore, a six-stage structure can be configured where intermediate rolls are also provided between each work roll and each support roll.

[0055] Loop 71 is a tension control roller disposed between frame F1 10 and frame F2 20. Loop 71 is configured such that its axis of rotation extends along the width direction of the metal strip 1 to support the traveling metal strip 1, and is configured to lift and hold the metal strip 1 upward.

[0056] In addition, the looper 71 may be designed to use a device that applies upward force using a spring or the like, or to lift it by means of a hydraulic cylinder or motor.

[0057] The camera 61 is configured to capture an image of the rolled metal strip 1, including the area reflecting a strip of reflected light transversely along the width direction. Preferably, it is configured to capture an image of the area of ​​the metal strip 1 that is lifted upwards by the looper 71. In particular, it can be positioned on the outer side of the metal strip 1 in the width direction when viewed from above, and can be positioned at a height of 1m to 5m from the metal strip 1, and a distance of 5m to 40m from the end of the metal strip 1. It is desirable that the position of the camera 61 in the rolling direction is approximately the same as that of the looper 71. The image data captured by the camera 61 is transmitted to the image processing computer 80 via the communication line 90.

[0058] Similarly, the tension control loop 72 is located between frame 20 of F2 and frame 30 of F3, the tension control loop 73 is located between frame 30 of F3 and frame 40 of F4, and the tension control loop 74 is located between frame 40 of F4 and frame 50 of F5.

[0059] Furthermore, camera 62 is positioned to capture images of the area where the metal strip 1 is lifted vertically upwards from looper 72, camera 63 is positioned to capture images of the area where the metal strip 1 is lifted upwards from looper 73, and camera 64 is positioned to capture images of the area where the metal strip 1 is lifted upwards from looper 74. It is desirable that the positions of cameras 62, 63, and 64 in the rolling direction are approximately the same as those of loopers 72, 73, and 74, respectively. The image data captured by cameras 62, 63, and 64 is transmitted to image processing computer 80 via communication line 90.

[0060] Similar to camera 61, cameras 62, 63, and 64 are preferably positioned on the outer side of the metal strip 1 in the width direction when viewed from above. Furthermore, they can be positioned at a height of 1m to 5m from the metal strip 1 and at a distance of 5m to 40m from the end of the metal strip 1.

[0061] These cameras 61, 62, 63, and 64 are used to perform a photographing step, which involves taking pictures of an area containing a strip of reflected light transversely along the width direction of the rolled metal strip 1.

[0062] The system can be further equipped with lighting to illuminate the raised shooting area of ​​the metal strip 1, which is mainly captured by cameras 61, 62, 63, and 64 and is lifted upwards from the rollers. This lighting can be ordinary lighting appropriately configured on the ceiling or the like of a rolling mill where the rolling equipment 100 is installed. In this invention, no particularly novel lighting equipment is required, but dedicated lighting can also be provided.

[0063] The image processing computer 80 performs various processes to determine the shape of the metal strip plate 1 based on images captured by cameras 61, 62, 63, and 64.

[0064] For example, for Figure 2 or Figure 3 The image shown includes the area of ​​the metal strip plate 1 that is lifted upward by the loops 71, 72, 73, and 74. Through image processing, the range (1A, 1B) including the upstream / downstream boundary of the part where the brightness of the reflected light on the plate surface reflected in the image is greater than a certain brightness value is determined as the reflected light region 1A or the reflected light region 1B.

[0065] In this embodiment, the image processing computer 80 divides the reflected light region in the image into multiple partitions along the width direction of the metal strip 1. Based on the area value, average length in the rolling direction, median length in the rolling direction, and other index information related to the size of the reflected light region in each partition, the information on the distribution in the width direction related to the strip wave shape in the rolling direction is sent as a signal.

[0066] Preferably, the aforementioned index information can be set as the average length in the rolling direction calculated for each pixel unit in the width direction of the plate, or the median length in the rolling direction, or the area in each segmented partition.

[0067] Regarding the visible strip-shaped reflected light zones along the width direction of the metal strip 1, if the waviness along the rolling direction (strip length direction) at all positions along the width direction of the metal strip 1 is uniform, then because the surface of the metal strip 1 is flat and the waviness is small, therefore... Figure 2 As shown, the distribution difference caused by the illumination method is small. Therefore, the boundary lines of the reflected light region 1A based on the illumination are approximately parallel on the upstream and downstream sides. In the case where the reflected light region 1A is divided into multiple sections along the width of the plate, the area value of each section, the average length in the rolling direction, and other parameters are approximately uniform across all sections.

[0068] In contrast, when the wavy shape in the rolling direction differs depending on its position in the width direction (e.g., edge wavy, center wavy), the illumination method differs due to the varying heights of the wavy areas, such as... Figure 3 As shown, the boundary line of the reflected light region 1B based on illumination will be wavy on one or both of the upstream and downstream sides instead of being parallel. Therefore, when the metal strip-shaped reflected light region 1B is divided into multiple sections along the width direction, the length and other parameters of each section in the rolling direction become non-uniform depending on its position in the width direction.

[0069] Therefore, in the image processing computer 80, various parameters (indicator information), including area values ​​related to the distance between the two boundary lines on the upstream and downstream sides of the reflected light region, are divided into multiple parts along the width direction and calculated. Preferably, these values ​​are decomposed into components (distribution of 0th-order, 1st-order, 2nd-order, and 4th-order components) using Chebyshev polynomials. The judgment results corresponding to each component value are output to the monitor 85 and the control device 82. Details are provided using... Figure 4 The following figures and descriptions will follow in detail. Preferably, the image processing computer 80 serves as the execution body for the image processing steps.

[0070] Return to Figure 1 The control device 82 is a device that controls the operation of each device in the rolling equipment 100. In this embodiment, it is a device that performs various controls in the image processing computer 80 corresponding to the determination of the shape of the metal strip 1.

[0071] These image processing computers 80 and control devices 82 can be configured as computers having a monitor 85 such as a liquid crystal display (described later), input devices, storage devices, CPUs, memory, etc., and can be configured as a single computer or as different computers, without particular limitation.

[0072] The control of the actions of each device based on the image processing computer 80 and the control device 82 is performed according to various programs recorded in the storage device. Furthermore, the control processing of actions performed by the image processing computer 80 and the control device 82 can be summarized in a single program, distributed among multiple programs, or a combination thereof. Additionally, some or all of the programs can be implemented in dedicated hardware, or they can be modularized.

[0073] The monitor 85 is a display device such as a monitor or an audio device such as an alarm. For example, it is a device used to communicate the operator's response when the image processing computer 80 determines that there is a problem with the shape of the board. Therefore, as such a monitor 85, a monitor is mostly used.

[0074] Here, the image processing computer 80 described above includes a display signal unit that sends signals related to the content displayed on the monitor 85 to the monitor 85.

[0075] The operator can confirm the status of the plate shape by visually observing the display screen of the monitor 85 and the individual racks and the racks in relation to each other during the operation.

[0076] Furthermore, it is not limited to the method of automatically improving the board shape by means of the control device 82 while communicating the board shape problem to the operator; it can be set to a method of displaying only on the monitor 85, or to a method of omitting the display to the monitor 85 and automatically improving the board shape problem by means of the control device 82.

[0077] Next, use Figure 4 The following figures illustrate specific examples of the plate shape determination device and method for the rolled metal strip 1 in this invention. First, using… Figures 4 to 7 This section details the method for calculating the state variables in the width direction based on the calculated values ​​of the boundary line of the reflected light region.

[0078] Figures 4 to 6 This is a diagram illustrating an example of the calculation method for index information in each segmented partition of the reflected light region in the metal strip shape determination device of the embodiment, and the distribution of components of each degree of the Chebyshev polynomial. Figure 7 This is a diagram illustrating an example of the distribution of the distance between the upstream and downstream boundary lines of each reflected light region when the area, including the reflected light region of the metal strip, is divided into seven parts along the width of the strip.

[0079] First, in the image processing computer 80, the selected rolling surface image range in the images captured by cameras 61, 62, 63, and 64 is binarized pixel by pixel to determine an appropriate threshold for brightness. Based on this, the width-direction boundary lines of the reflected light region are determined at two points along the rolling length direction, on the upstream and downstream sides. This processing is performed along the entire width direction to determine the reflected light regions 1A and 1B. Details can be set using known methods.

[0080] Next, the reflected light regions 1A and 1B are divided into N parts along the width of the plate (in Figures 4 to 6 The middle is divided into 5 parts, in Figure 7 The image is divided into 7 sections. When j is set as the number (No.) of each section (j = 1 to N), for each pixel in the image of section No. j, the distance between the boundary lines on the upstream and downstream sides of the reflected light region along the rolling direction (Y-axis) is calculated for each pixel position in the plate width direction (X-axis direction), and the index information is calculated. For example, in... Figure 4 In the process, the average length Laj in the rolling direction, calculated as an indicator, is plotted as a bar chart.

[0081] Furthermore, the data in the bar chart is not limited to those mentioned above. Figure 4 In addition to the average length Laj in that rolling direction, it can also be like Figure 5 As shown, as indicator information included within this segmentation, all area values ​​Sj of each segmented partition are plotted as bar charts. In addition, they can also be used as indicator information included within this segmentation, such as... Figure 6 As shown, the median Mj of the distance between the upstream and downstream boundary lines of the reflected light region in each segmented partition will be plotted as a bar graph when the frequency of occurrence of the distance is histogramized.

[0082] At this point, in detail, using the Chebyshev polynomials described later...

[0083] E(x) = C0' + C1' × x + C2' × (2x) 2 -1)+C4'×(8x 4 -8x 2 +1)... (1)

[0085] When approximating the index information E(x), there are four unknown coefficients: C0', C1', C2', and C4'. To find the unknowns, more than four relationships are needed. Therefore, the number of segments in the width direction is set to more than four. In addition, since it is desired that the central part of the width direction belongs to a certain segmentation zone, it is desirable to set the number of segments to an odd number and to ensure that the central part of the width direction is not located at the boundary of each segmentation zone.

[0086] Furthermore, it is preferable that a single segment contains multiple pixels along the width of the plate. This is because it is less susceptible to the impact of pixel anomalies along the width of the plate. The desired number of segments is 4 to 11, especially an odd number. The lower limit of 4 for the number of segments is to be able to obtain the coefficients of the Chebyshev polynomial, and the upper limit of 11 is because 11 segments can obtain coefficients with sufficient accuracy, and even if the number of segments is further increased, the accuracy will not increase to the extent that the computational burden increases.

[0087] Next, when the image processing computer 80 sets the value of the position in the width direction of the bar chart as a variable (x), it converts the position coordinates of the region in the width direction of the image into a range of -1≤x≤1, where x=-1 represents the drive side (DS) end of the width, x=1 represents the working side (WS) end of the width, and x=0 represents the middle position of the width. It then uses E(x)=C0+C1×x+C2×x, which consists only of 0th, 1st, 2nd, and 4th degree terms, to set the index information in each partition as the distribution (E(x)) of each partition. 2 +C4×x 4 The function is represented by first calculating the coefficient vectors (C0, C1, C2, C4) of each component. Then, the Chebyshev polynomial and the above formula are set as equivalents to obtain the Chebyshev polynomial coefficient vectors (C0', C1', C2', C4') of each component, and then sent as a signal indicating the shape of the plate.

[0088] As a feature of approximating the index information corresponding to the plate wave distribution in the rolling direction relative to the plate width direction with Chebyshev polynomials, the following aspects can be listed: (1) normalizing the plate width range (X-axis) in the range from -1 to +1; (2) separating the first-order component (single-sided wave), the second-order component (medium-sided wave), and the fourth-order component (quarter-sided wave) so that the control operation amount for each component can be easily determined.

[0089] The following describes the method for calculating the Chebyshev polynomial coefficient vector (C0', C1', C2', C4') based on the measured values ​​of each partition. Here, as... Figure 7 As shown, with Figure 4 The following explanation will be based on the example of dividing the width of the metal strip plate 1 into 7 parts, for example.

[0090] like Figure 7 As shown, if the relationship between the distribution value vectors (E1, E2, E3, E4, E5, E6, E7) of the index information when performing a 7-part width division and the quartic coefficient vector (C0, C1, C2, C4) is set as x = x i Then it can be expressed as shown in the following formula (2). That is,

[0091] E(x i )=C0+C1×x i +C2×x i 2 +C4×x i 4 (i = 1 to 7) · · · (2)

[0092] If we express formula (2) in terms of vectors and matrices, then it becomes formula (3) as shown below.

[0093]

Number 1

[0094]

[0095] Here, in formula (3), the plate width position (x) is (-1≤x≤1), where 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). i (i = 1 to 7) represents the width-direction position of the detection value at each segmentation location (-1 ≤ x). i ≤1), can be referred to as the width-direction position coordinate (x) of the center position of each segmentation partition with the middle position of the plate width set to x=0. i ).

[0096] According to the formula (3), the unknown vector (C0, C1, C2, C4) is calculated by the least squares method, and the Chebyshev polynomial coefficient vector (C0', C1', C2', C4') is obtained. At this time, the least squares method can be set as shown below.

[0097] First, when the formula (3) is expressed as E=M×C in matrix and vector form, if the left and right sides are multiplied by the transpose matrix M from the left... T Then it is represented by the following formula (4).

[0098] M T ×E =(M T ×M)×C· · · (4)

[0099] Let E = [E1, E2, E3, E4, E5, E6, E7] and M = [[1, x1, x2]]. 2 x1 4 ],[1,x2,x2 2 x2 4 ],[1,x3,x3 2 x3 4 ],[1,x4,x4 2 x4 4 ],[1,x5,x5 2 x5 4 ],[1,x6,x6 2 x6 4 ],[1,x7,x7 2 x7 4 ]], C=[C0, C1, C2, C4].

[0100] Therefore, the coefficient vector C uses a matrix (M) T The inverse matrix of (M × M) T ×M)-1 The result is calculated using the following formula (5).

[0101] C = (M T ×M) -1 ×M T ×E· · · (5)

[0102] According to the formula (5), (C0, C1, C2, C4) are obtained. If the index information E(x) of the state of the reflected light region is represented by the Chebyshev polynomial coefficient vector (C0', C1', C2', C4'), then it becomes the above formula (1).

[0103] Therefore, (C0', C1', C2', C4') can be calculated in the following formula (6) using formulas (1) and (2).

[0104] C0' = C0 + C2' - C4'

[0105] C1'=C1

[0106] C2' = 1 / 2 × (C2 + C4)

[0107] C4' = 1 / 8 × C4··· (6)

[0108] Here, since the term C0' of the 0th order component needs to be included in the formula, it is desired to calculate it. The term C0' of the 0th order component represents the index information that forms the baseline (foundation) of the whole, and the other order components represent the component distribution of each order based on this baseline. In the absence of a wavy distribution corresponding to components of order 1 or higher, the distance between the boundary lines of the upstream and downstream sides of the reflected light region is as follows: Figure 2 As shown, it becomes a length that is approximately uniform along the rolling direction.

[0109] Furthermore, in this embodiment, the third component is omitted. This is because it is practically rare for a significant wavy pattern to appear in the third component. By eliminating the calculations and handling methods for the third component, it is easier to determine the condition of the rolled plate shape.

[0110] The image processing computer 80 can output control command signals to the control device 82 based on the polynomial approximation result in the width direction of the plate obtained by the formula (6), which corrects for leveling, bending force, and paired cross angles. Moreover, by outputting display command signals to the monitor 85 instead of or based thereon for guiding the display required for correcting leveling, bending force, and paired cross angles, the correction information for leveling, bending force, and paired cross angles can be conveyed to the operator.

[0111] Preferably, the image processing computer 80 is capable of displaying the 0th-order component (C0'), 1st-order component (C1'×x), and 2nd-order component (C2'×(2x)) of the function that displays the vectors of each order term (C0', C1', C2', C4') in the above-described E(x). 2- 1)), 4th component (C4'×(8x) 4 -8x 2 The signal is sent to monitor 85 in the form of a curve of each component of +1). The image displayed on monitor 85 becomes, for example... Figure 8 The scene shown.

[0112] Figure 8 This is an example diagram showing a monitor display. Figure 8 Although the zero-order component (C0') is also displayed, automatic control or operator support is not based on the zero-order component. Therefore, it is not necessarily necessary to send the zero-order component signal to or display it on the monitor 85. Furthermore, information about changes in line tension can be obtained from the zero-order component.

[0113] The operator can confirm this Figure 8 The screen shown is used to perform operations such as leveling, bending force, and paired cross angles (in the case of paired cross rolling mills) to make corrections. For example, if the primary component C1' shows a unilateral wave, then the leveling operation of the pressing cylinders 11, 21, 31, 41, and 51 needs to be performed.

[0114] Figure 9 This is a diagram representing the plate shape control block of the first-order component C1'. Since the first-order component C1' in the Chebyshev polynomial coefficients exhibits a unilateral wave, an operation command signal is output to the control device 82 to operate the leveling of the pressing cylinder 41 on the upstream drive side (DS) and working side (WS) of the corresponding camera 64, and / or the pressing cylinder 51 on the downstream drive side (DS) and working side (WS), so that the first-order component is normalized (within the target range).

[0115] In addition, Figures 9 to 11 The example shown is between F4 rack 40 and F5 rack 50, but the same structure can also be set between other racks such as F1 rack 10 and F2 rack 20, F2 rack 20 and F3 rack 30, or F3 rack 30 and F4 rack 40.

[0116] Figure 10This is a diagram representing the plate shape control block of the quadratic component C2'. The quadratic component C2' in the Chebyshev polynomial coefficients shows either double-sided waves or intermediate waves. Therefore, one or more of the following operations are performed: An operation command signal is output to the control device 82, causing the bending devices of the work rolls / intermediate rolls of the F4 stand 40 of the upstream mill (which corresponds to the camera 64) and / or the F5 stand 50 of the downstream mill to be operated. In the case of paired cross mills, the paired cross angle is operated. In the case of work roll shifting / intermediate roll shifting mills, since shifting is difficult during rolling, the intermediate waves / double-sided waves are predicted in advance to shift the work rolls / intermediate rolls, thereby normalizing the quadratic component (within the target range).

[0117] Figure 11 This diagram represents the plate shape control block of the fourth component C4', which shows a quarter wave. Therefore, one or more of the following operations are performed to correct the quarter wave: Bending operations are performed on the bending devices of the work rolls of the upstream mill F4 stand 40 (corresponding to camera 64) and / or the downstream mill F5 stand 50; in the case of paired cross mills, bending operations are performed simultaneously and / or the paired cross angles are operated separately. In the case of a 6-stage intermediate roll shifting mill, the quarter wave is predicted in advance, and the intermediate roll is shifted at the appropriate position. An operation command signal is output to the control device 82, causing bending operations and paired cross angle operations to be performed, thereby normalizing the fourth component showing the quarter wave (to be within the target range in a way that becomes the target plate shape). Furthermore, since the quarter wave is more easily generated in the region at the wide end of the roll by bending operations when the roll diameter is smaller relative to the roll length, the above operations can normalize it, even though the quarter wave is easy to generate.

[0118] Next, use Figures 12 to 17 This describes the details of the correction processing of the detected values ​​in the image processing computer 80 based on the angle between the metal strip plate 1 and the cameras 61, 62, 63, and 64. Figure 12 This is a diagram illustrating an example of the configuration relationship between a metal strip and a camera in three-dimensional space. Figure 13 It means to Figure 12 A diagram showing the state of the metal strip plate and the camera projected onto the XY two-dimensional plane. Figure 14 It means to Figure 12 A diagram showing the state of the metal strip plate and the camera projected onto the two-dimensional plane XZ. Figure 15 This diagram illustrates the correction status of the boundary line position of the reflected light region in the metal strip shape determination device of the embodiment. Figure 16 and Figure 17This is an example diagram showing the relationship between the distance (D) from the metal strip to the camera and the perspective correction ratio α (0).

[0119] As described above, it is desirable that cameras 61, 62, 63, and 64 be positioned on the outer side of the metal strip 1 in the width direction when viewed from above. Therefore, as Figure 12 The arrangement of the metal strip 1 and cameras 61, 62, 63, and 64 in the XYZ three-dimensional space, as shown, results in a configuration where cameras 61, 62, 63, and 64 view the metal strip 1 from a slightly upward angle. Furthermore, regarding... Figure 12 The XYZ three-dimensional coordinates are defined as follows: the Y coordinate represents the rolling direction, the X coordinate is set to the width direction of the metal strip 1, and the Z coordinate is set to the thickness direction of the metal strip 1, i.e., the vertical direction.

[0120] First, such as Figure 13 As shown, the configuration relationship of the metal strip plate 1 and the cameras 61, 62, 63, and 64 in three-dimensional space is projected onto a two-dimensional plane (XY). In this case, regarding the relationship between the length d in the Y direction of the metal strip plate 1 that is near the front side (represented as the lower side in the image) and the length d' in the Y direction of the metal strip plate 1 that is near the back side (represented as the upper side in the image), although the lengths d and d' that are represented in the image are the same (the number of pixels constituting the line segments of d and d' is the same), the actual length can be expressed by the distance (D) from the metal strip plate to the camera and the width (W) of the plate as shown in the following formula (7).

[0121] d' / d=(D+W) / D· · · (7)

[0122] In other words, although the length of the metal strip 1 near the front in the Y direction shown in the image taken at a distance D from cameras 61, 62, 63, 64 in the Y direction is d, the length of the metal strip 1 on the inner side in the Y direction shown in the image taken at a distance D+W from cameras 61, 62, 63, 64 in the X direction is d'. In fact, a range longer than d is shown in the image.

[0123] That is, although the number of pixels visible in the image of the line segment d' on the inside and the line segment d near the front in the Y direction that constitute the image is the same, the actual length of d' is (D+W) / D times the actual length of d because the actual length corresponding to the spacing between adjacent pixels reflected in the image is different depending on the position in the image.

[0124] Regarding the ratio of the actual length of the interval between adjacent pixels at positions D and D+W in the X direction, when the actual length of the pixel interval at position D is set to 1, the actual length of the pixel interval at position D+W becomes (D+W) / D times in both the X and Y directions.

[0125] Therefore, it is also possible to correct the actual length ratio by continuously varying the ratio of the pixel interval from position D to position D+W in both the X and Y directions from 1 to (D+W) / D. That is, this is a correction that takes into account perspective, making the actual length correction appropriate.

[0126] In this way, when dividing the region in the image into multiple partitions along the width direction of the metal strip 1, and calculating one of the indicators related to the length in the rolling direction, such as the area value of each partition, the average length in the rolling direction, the median length in the rolling direction, or the total length in the rolling direction, the image processing computer 80 is expected to perform correction processing as described above, in a manner that assumes the distance between two points in the image is greater the farther the actual position corresponding to the position reflected in the image is from the camera 61, 62, 63, 64.

[0127] Therefore, when the index information is the average length, median length, or total length in the rolling direction between the upstream and downstream boundary lines of the reflected light area, the image processing computer 80 performs X-direction and Y-direction correction processing on the index information in each partition, assuming that the actual position corresponding to the position reflected in the image is farther away from the cameras 61, 62, 63, and 64, so that the positions between two pixels mapped on the image become longer as their actual positions are farther away from the cameras 61, 62, 63, and 64. More specifically, it is desirable to correct the pixel spacing reflected in the image in a manner proportional to the actual distance from the cameras 61, 62, 63, and 64 to each position in the image. When the index information is set as the area value of the reflected light area of ​​the partitioned area, it is also desirable to perform the same X-direction and Y-direction actual length correction processing as described above.

[0128] Next, as Figure 14 As shown, compared to the metal strip plate 1, cameras 61, 62, 63, and 64 are positioned at a height h above the Z-axis, i.e., the height direction.

[0129] In this case, if the configuration relationship between the metal strip plate 1 and the cameras 61, 62, 63, 64 in three-dimensional space is projected onto a two-dimensional plane (XZ), then the relationship between the length d in the Y direction of the metal strip plate 1 that appears in front of the captured image and the length d' in the Y direction of the metal strip plate 1 that appears in the captured image can be expressed as follows (8) since the cameras 61, 62, 63, 64 are higher than h in the Z direction.

[0130] d' / d=V' / V

[0131] V = (D 2 +h 2 ) 0.5

[0132] V' = {(D+W)} 2 +h 2} 0.5 ··· (8)

[0133] In addition, the distance from cameras 61, 62, 63, 64 to the front of the metal strip plate 1 is set as V, and the distance from cameras 61, 62, 63, 64 to the inside of the metal strip plate 1 is set as V'.

[0134] Regarding the above formula (8), if h = 0, it is the same as the relationship between d and d' in the XY plane (7).

[0135] Next, we will explain the method for correcting the position of the boundary lines of the reflected light regions 1A and 1B.

[0136] like Figure 15 As shown, the pixels on the upstream boundary line of the metal strip plate 1 constituting the reflected light regions 1A and 1B are defined as (U0, U1, U2, ..., U...). i U N The pixels constituting the boundary line of the other side (downstream of metal strip 1) of the reflected light regions 1A and 1B are defined as (R0, R1, R2, ..., R...). i , ···, R N ).

[0137] If U is placed in each of these pixels i Let the coordinates of the point in the image be (U ix U iy ), R i Let the coordinates of the point in the image be (R). ix R iy (i=0~N), then the spacing Li in the rolling direction of the reflected light regions 1A and 1B can be expressed as shown in the following formula (9).

[0138] Li(U ix R ix ) = R iy -U iy ··· (9)

[0139] Regarding the ratio by which the pixel spacing between positions D and D+W is corrected to the actual length, if considered in three-dimensional space (XYZ), it varies continuously from 1 to V' / V. Therefore, the perspective correction ratio α(i), which is the ratio of the actual length of the pixel spacing at position i, can be expressed as shown in the following formula (11).

[0140] When i = N, α(N) = 1

[0141] When i = 0, α(0) = V' / V = {(D+W)} 2 +h 2} 0.5 / (D 2 +h 2 ) 0.5 ···(10)

[0142] According to the formula (10), it can be expressed as the formula (11) below.

[0143] α(i)=-[{(D+W) 2 +h 2} 0.5 / (D 2 +h 2 ) 0.5 -1]×i / N+{(D+W) 2 +h 2} 0.5 / (D 2 +h 2 ) 0.5 ···(11)

[0144] Therefore, considering the perspective method of the spacing Li in the rolling direction of the reflected light region, a correction Lc corresponding to the actual length was made. i It can be expressed as shown in the following formula (12).

[0145] Lc i =α(i)×Li=α(i)×(R) iy- U iy )···(12)

[0146] Typically, when dealing with pixel coordinates on an image, the top-left corner is set as 0 (origin), the bottom is set as positive, and the right is set as positive.

[0147] Pixel coordinates U on the boundary lines of the upstream and downstream sides of the rolling direction (Y-axis direction) for the reflected light region. ix and pixel coordinates R ix Considering perspective, points that have been corrected to correspond to actual distances are set as pixel coordinate points U. Cix and pixel coordinates R Cix In the case of the upstream boundary line, the X coordinate of point U is transformed as shown in the following formula (13).

[0148] U C0x =U 0x (where i = 0)

[0149] U C1x = (1-α(0))×U 0x +α(0)×U 1x (where i = 1)

[0150] U C2x = (1-α(0))×U 0x +(α(0)-α(1))×U 1x +α(1)×U 2x (where i = 2)

[0151] U C3x = (1-α(0))×U 0x +(α(0)-α(1))×U 1x +(α(1)-α(2))×U 2x +α(2)×U 3x (where i = 3)

[0152] When we rearrange the above recursive relation using i, if we consider i=2 and beyond (i=2 to N), it is expressed as follows.

[0153] U C0x =U 0x (where i = 0)

[0154] U C1x = (1-α(0))×U 0x +α(0)×U 1x (where i = 1)

[0155] U Cix = (1-α(0))×U 0x +Σ i i＝2 {α(i-2)-α(i-1)}×U (i-1)x +α(i-1)

[0156] ×U ix (where i = 2 to N)

[0157] ···(13)

[0158] The X coordinates of point R on the downstream boundary line are also expressed in the same way, as in the following formula (14).

[0159] R C0x =R 0x (where i = 0)

[0160] R C1x = (1-α(0))×R 0x +α(0)×R 1x (where i = 1)

[0161] R Cix = (1-α(0))×R 0x +Σ i i＝2 {α(i-2)-α(i-1)}×R (i-1)x +α(i-1)

[0162] ×R ix (where i = 2 ~ N) · · · (14)

[0163] This correction incorporates perspective adjustment, ensuring that the distance between two points projected onto the image increases proportionally to their actual distance as the image moves further away from the camera. When segmenting along the width of the image, the pixel coordinates U are based on this correction. Cix and pixel coordinates R Cix Divide the plate into equal parts along its width (e.g., 7 parts).

[0164] Conversely, the spacing of the scale applied to the image can also be changed. When such a modification is made to divide the reflective area of ​​the image along the width of the plate, when a scale of fixed actual length is taken on the image, the scale spacing becomes narrower towards the top of the image and wider towards the bottom of the image.

[0165] The pixels constituting the image are arranged at equal intervals in the top, bottom, left, and right directions. When the reflected light area projected onto the image is divided at equal intervals along the width of the plate, the number of pixels arranged along the width of the plate is the same in each segment. When the reflected light area is divided equally along the width of the plate based on a scale of actual length that has been corrected for perspective, the number of pixels of the original image arranged along the width of the plate in each segment decreases as it moves further inward (top of the image) and increases as it moves closer to the front (bottom of the image).

[0166] The length (specifically, the number of pixels with a brightness of a fixed brightness or higher in the Y-axis direction) corresponding to the pixels arranged along the width direction (X-axis) within each segment of the reflective region is measured. The average length is obtained by dividing the sum of these lengths by the number of pixels arranged along the width direction. The median length is the length in the middle of these lengths arranged in ascending order. Furthermore, while the calculation method for the area of ​​the segmented reflective region is complex, it can be obtained by integrating the actual length within the segmented reflective region based on the actual length after considering perspective correction.

[0167] In light of these circumstances, when the indicator information is an area value, the image processing computer 80 processes the image in such a way that the area value of the reflected light region in each partition increases as the actual position corresponding to the position reflected in the image moves further away from the cameras 61, 62, 63, 64. As a correction process to increase the area value, it is expected that the perspective correction ratio α(i) of the pixel interval at position i at the actual distance from the cameras 61, 62, 63, 64 to each position in the image will be accumulated in both the X and Y directions to process the area in the image.

[0168] Next, we will explain the appropriate setting range for cameras 61, 62, 63, and 64.

[0169] In this embodiment, the appropriate setting range of cameras 61, 62, 63, and 64 is determined based on the above formula (10) and the camera setting position range is determined from the point of view of taking into account the necessity of perspective correction when applying control.

[0170] For example, assuming that the correction amount within the width of the plate during the correction is less than 5%, it will not have a significant impact on control. Figure 16 As shown, after investigating the change in the perspective correction ratio α(0) corresponding to the deviation from the actual length, it is found that when the width W of the metal strip 1 is 2m, as long as the distance D from the metal strip 1 is greater than 40m, even if the setting height h of cameras 61, 62, 63, and 64 is changed to 1m, 3m, and 5m respectively, no deviation from the actual length exceeding 5% will be observed. Even at distances greater than 40m, which can be considered to have a minimal impact, there is a possibility that the ease of processing may be prioritized over actual correction. Furthermore, it is known that if the distance D from the metal strip 1 is less than 5m, the correction amount becomes larger and is considered to have a significant impact; therefore, it is desirable to avoid this as much as possible.

[0171] In addition, such as Figure 17As shown, after investigating the change in the perspective correction ratio α(0) corresponding to the deviation from the actual length, it is found that when the width W of the metal strip plate 1 is 1m, if the distance D from the metal strip plate 1 is greater than 25m, the correction amount for the actual length within the plate width becomes less than 5%, and the influence of the installation height h of cameras 61, 62, 63, and 64 is not very noticeable. Furthermore, it is found that if the distance D from the metal strip plate 1 is less than 5m, the correction amount is also large, so it is desirable to avoid it as much as possible. Regarding the installation height h, it is the same as when the plate width is 2m. In this case, even when it is farther than 25m, there may be a situation where ease of processing is prioritized over actual correction. In addition, there may be a situation where the installation positions of cameras 61, 62, 63, and 64 are far from the metal strip plate in order to avoid correction and complete the process easily.

[0172] Conversely, from the perspective of needing correction, it can be seen that when the plate width W is 2m, if the horizontal distance from the metal strip plate 1 to the cameras 61, 62, 63, and 64 is less than 40m, then a correction taking into account perspective should be made. When the plate width W is 1m, if the distance from the metal strip plate 1 to the cameras 61, 62, 63, and 64 is less than 25m, then a correction taking into account perspective is desired.

[0173] Based on the above, regarding the positional relationship between the metal strip plate 1 and cameras 61, 62, 63, and 64, which should be corrected for perspective, it is preferable that cameras 61, 62, 63, and 64, according to... Figure 16 and Figure 17 The distance between the metal strip plate 1 and the end of the strip is between 5m and 40m.

[0174] Regarding the height h, it can be known that if it is in Figure 16 and Figure 17 The range studied is considered suitable, therefore it is expected to be set at a height of more than 1m and less than 5m from the metal strip plate 1.

[0175] Next, refer to Figure 18 The flowchart of the metal strip shape determination method of this embodiment, which is suitable for execution by the metal strip shape determination device described above, is explained. Figure 18 This is a diagram illustrating the detection and control process of the plate shape in the plate shape determination device for the metal strip of the embodiment.

[0176] First, the image processing computer 80 acquires images captured by cameras 61, 62, 63, and 64 between racks of racks F1 10, F2 20, F3 30, F4 40, and F5 50 (step S101).

[0177] Next, in the image processing computer 80, the boundary lines of the upstream and downstream sides of the reflected light regions 1A and 1B in the rolling direction are calculated by image processing (step S102).

[0178] Next, in the image processing computer 80, the correction processing (Ucix) and (Rcix) in the X direction of the boundary lines of the reflected light regions 1A and 1B calculated in step S102, and the correction processing (Lci) of the distance between the boundary lines in the Y direction of the reflected light regions are executed (step S103). The correction processing in step S103 is preferably set to be based on the above formulas (12), (13), and (14).

[0179] Next, in the image processing computer 80, the number of divisions in the width direction of the reflective light area of ​​the metal strip plate 1 is set (step S104). It is desired that the number of divisions be selected from an odd number of 4 or higher, as described above.

[0180] Next, in the image processing computer 80, the distance or area between the boundary lines of the upstream and downstream sides of the reflected light region in the rolling direction is calculated, and the index information (average or median distance, or area, etc.) of each segmented partition is calculated based on the distance or area (step S105).

[0181] Next, in the image processing computer 80, the width distribution of the metal strip plate 1 of the index information is calculated, and the coefficients of the Chebyshev polynomial (0th-order component, 1st-order component, 2nd-order component, 4th-order component) are obtained (step S106).

[0182] Next, in the image processing computer 80, the difference between the target value and the detected value of each component of the coefficients of the Chebyshev polynomial is calculated (step S107).

[0183] Next, in the image processing computer 80, it is determined whether all components of the Chebyshev polynomial coefficients are within the control target value tolerance (step S108). If it is determined that all components are within the control target value tolerance, the processing is completed, and the detection and control process is repeated from the beginning. Conversely, if it is determined that one or more components of the Chebyshev polynomial coefficients are greater than the control target value tolerance, the processing proceeds to steps S109, S110, and S111.

[0184] Next, in the image processing computer 80, when it is determined in step S108 that the first-order component is greater than the allowable value, control command signals for the operation of the upstream and downstream racks of the cameras 61, 62, 63, and 64 that have captured the corresponding images are generated and output to the control device 82, or / and display command signals for the monitor 85 are generated and output (step S109). Furthermore, signals of the zero-order component do not necessarily need to be output. This is because control or operation is not based on the zero-order component.

[0185] Additionally, in the image processing computer 80, when it is determined in step S108 that the second component is greater than the allowable value, a control command signal for the operation amount of the rack upstream of the cameras 61, 62, 63, and 64 that have captured the corresponding images is generated and output to the control device 82, or / and a display command signal for the monitor 85 is generated and output (step S110).

[0186] Next, in the image processing computer 80, when it is determined in step S108 that the fourth component is greater than the allowable value, a control command signal for the operation amount of the rack on the upstream side of the cameras 61, 62, 63, 64 that have captured the corresponding images is generated and output to the control device 82, or / and a display command signal for the monitor 85 is generated and output (step S111).

[0187] Next, intervention control is automatically executed via control device 82, or the content of intervention control is displayed on monitor 85 (intervention control guidance) or Figure 8 The system displays the image shown and requests the operator to make a decision on whether manual intervention is appropriate (step S112). Then, the process returns to step S101 and the camera image is captured again. Adjustments are performed in such a way that all components of the coefficients of the Chebyshev polynomial are within the allowable value of the control target value.

[0188] Next, the effects of this embodiment will be explained.

[0189] The plate shape determination device for the rolled metal strip 1 described in this embodiment includes: cameras 61, 62, 63, and 64, which are configured to capture images of the rolled metal strip 1, including a region reflecting a strip of reflected light transversely along the width direction; and an image processing computer 80, which determines the plate shape of the metal strip 1 based on the images captured by the cameras 61, 62, 63, and 64. The image processing computer 80 divides the region in the image into multiple partitions along the width direction of the metal strip 1, and sends information corresponding to the distribution of the plate waviness in the rolling direction along the width direction as a signal based on the index information representing the size related to the region in each partition.

[0190] Therefore, instead of using parameters with low information content such as linear reflected light in the form of metal strips, parameters with high information content such as strip reflected light are used. Furthermore, instead of relying solely on data from a portion of the reflected light regions 1A and 1B, the overall information of the reflected light portion is used to determine the plate wavy shape. Thus, even if the reflected light is affected by interference, high-precision detection of the plate wavy shape can still be performed.

[0191] In addition, the index information is the average length in the rolling direction, the median length in the rolling direction, or the area value in the segmented area of ​​the reflected light region, calculated for each pixel unit in the width direction of the plate. Therefore, the index information can obtain accuracy within the practical range.

[0192] When the image processing computer 80 sets the value of the center position in the width direction of each partition of the reflected light region as a variable (x), it uses E(x) = C0 + C1×x + C2×x, which is used to convert the position of the region in the width direction of the image into the range of -1≤x≤1 and set the index information in each partition as the distribution (E(x)) of each partition, and consists only of terms with x as the fourth, second, first and 0th order. 2 +C4×x 4 The function is represented by E(x), which first calculates the coefficient vector of each component (C0, C1, C2, C4), and then uses E(x) = C0' + C1' × x + C2' × (2x) 2 -1)+C4'×(8x 4 -8x 2 +1) represents the Chebyshev polynomial and the above formula as equivalents. The Chebyshev polynomial coefficient vectors (C0', C1', C2', C4') as each component are obtained and sent as signals indicating the plate shape. Thus, the plate shape, which reflects the area value, the average length in the rolling direction, or the median length in the rolling direction of each partition divided into multiple sections along the width direction of the metal strip 1, can be decomposed into the first-order, second-order, and fourth-order components of the Chebyshev polynomial. Therefore, it is easier to normalize the plate shape (within the target range).

[0193] In addition, the image processing computer 80 sends a signal to the monitor 85 to display a graph of the function of each degree term in the formula (1), which is then displayed on the screen that the operator can see. In this case, the operator can determine the operation to normalize the board shape to the target range for each degree component and implement it, making it easier to respond appropriately compared to the past.

[0194] Furthermore, when viewed from above, cameras 61, 62, 63, and 64 are positioned on the outer side of the metal strip 1 in the width direction. The image processing computer 80 processes the image in such a way that the further away the actual position of the metal strip 1, corresponding to the position reflected in the image, is from the cameras 61, 62, 63, and 64, the longer the distance between the positions in the image. Thus, by setting up cameras 61, 62, 63, and 64 at a certain distance from the rolling mill, the effect of easy maintenance and inspection can be achieved due to the need to ensure the working space and the proximity to the monitoring room. Furthermore, in this case, the farther the strip-shaped reflected light regions 1A and 1B of the metal strip plate 1 in the image are from the cameras 61, 62, 63, and 64, the smaller they appear in the image. The closer the plate end is to the end that is farther from the camera 61, 62, 63, and 64, the smaller the reflected light regions will appear in the image, even if objects are the same distance or size. If the distance and size are based on the original image, there is a concern that the detection of the distance and size of the reflected light regions may introduce non-negligible errors. However, for image information, the reflected light regions 1A and 1B appearing in the image... The further away the images of A and 1B are from cameras 61, 62, 63, and 64, the greater the distance between each point. Therefore, when viewed from the upper surface of the metal strip plate 1, even if cameras 61, 62, 63, and 64 are positioned away from the plate end, the difference in the position information of the reflected light from the metal strip plate 1 in the image can be reduced regardless of the position. Thus, the Chebyshev polynomial coefficient vectors (C0', C1', C2', C4') as components of each degree term can be obtained with high precision, thereby enabling high-precision plate shape determination.

[0195] Furthermore, the image processing computer 80 can obtain more accurate index information of the reflected light regions 1A and 1B through the following correction processing, thereby further improving the detection accuracy of the plate shape. The correction processing is performed in such a way that the area value in the index information of each partition increases as the actual position corresponding to the position reflected in the image moves further away from the cameras 61, 62, 63, and 64; or in such a way that the average length or median length in the rolling direction in the index information of each partition increases as the actual position corresponding to the position reflected in the image moves further away from the cameras 61, 62, 63, and 64.

[0196] Furthermore, the image processing computer 80 processes the area in the image by increasing the area value from each position in the image from the cameras 61, 62, 63, and 64, in both the rolling direction and the plate width direction, so that the area increases as the intended position moves further away from the cameras. The image processing computer 80 also processes the average length or the median length in the rolling direction by extending the average length or the median length in the rolling direction from the cameras 61, 62, 63, and 64, so that the length increases as the distance from the cameras increases. Thus, through calculations based on clear principles, the above-mentioned effects can be reliably achieved.

[0197] Furthermore, since the deviation relative to the actual length is determined based on the relationship with the horizontal distance D from the metal strip plate 1 to the cameras 61, 62, 63, 64, α(0) of formula (10) is obtained. If the deviation relative to the actual length is set as a threshold of 5%, the actual length is corrected if it exceeds 5%. This allows the Chebyshev polynomial coefficient vectors (C0', C1', C2', C4') as components of each degree term to be obtained with high accuracy. This allows the cameras 61, 62, 63, 64 to be set at a position away from the metal strip plate 1. This is practical from the viewpoint of ensuring space for maintenance and inspection and easy access. It also allows the mutual distance between the positions in the image to be corrected within the practical range.

[0198] <Other>

[0199] It should be noted that the present invention is not limited to the above embodiments and can be modified and applied in various ways. The above embodiments have been described in detail for ease of understanding of the present invention and are not limited to having all the structures described.

Claims

1. A sheet shape judging device for a metal strip, comprising: A camera configured to capture an image of an area on the surface of a rolled metal strip that reflects a strip of light transversely along the width of the strip; and The image processing unit determines the shape of the metal strip based on the image captured by the camera. The feature of the metal strip shape determination device is that... When the image processing unit divides the region in the image into multiple partitions along the width direction of the metal strip and sets the value showing the width direction position of each partition as a variable (x), it applies the following formula (1) of the distribution E(x) of each partition, which is composed only of the 0th, 1st, 2nd and 4th degree terms of x, to convert the width direction position of the region in the image into the range of -1≤x≤1, and sets the index information representing the size of the region in the partition as the Chebyshev polynomial of the following formula (1) of the distribution E(x) of each partition, and calculates the coefficients (C0', C1', C2', C4') of the Chebyshev polynomial as information corresponding to the distribution of the plate waviness in the width direction in the rolling direction. It sends one or more of the coefficients of the 0th degree coefficient (C0'), the 1st degree coefficient (C1'), the 2nd degree coefficient (C2') and the 4th degree coefficient (C4') as the judgment result signal of the plate waviness distribution in the width direction. E(x)=C0'+C1'×x+C2'×(2x 2 -1)+C4'×(8x 4 -8x 2 +1) Where -1≤x≤1···(1).

2. The metal strip shape determination device as described in claim 1, characterized in that, The indicator information is the average length in the rolling direction, the median length in the rolling direction, or the area in the partition, calculated for each pixel unit in the width direction of the plate.

3. The metal strip shape determination device as described in claim 2, characterized in that, The image processing unit sends a display signal to the display device, causing the zero-order component (C0'), the first-order component (C1'×x), and the second-order component (C2'×(2x)) in the formula (1) to be displayed. 2 -1)) and 4th component (C4'×(8x 4 -8x 2 The graph of the function of the term +1)) is a graph of one or more of the degree components related to the coefficient sent as the judgment result signal.

4. The sheet shape determination device for metal strips as described in any one of claims 1 to 3, characterized in that, The camera is positioned on the outer side of the metal strip in the width direction when viewed from above. The image processing unit performs correction processing in a manner that assumes the distance between two points in the image is greater the further away the actual position corresponding to the position reflected in the image is from the camera.

5. The metal strip shape determination device as described in claim 4, characterized in that, As a correction process, the image processing unit reduces the scale interval of the two-dimensional axis applied to the image as the position of the actual position, which is imagined to correspond to the position reflected in the image, is further away from the camera.

6. The sheet shape determination device for metal strips as described in any one of claims 1 to 3, characterized in that, The camera is positioned on the outer side of the metal strip in the width direction when viewed from above. When calculating the average length or the median length in the rolling direction of the index information in each of the partitions, the image processing unit processes the data in such a way that the length in the rolling direction of the region increases proportionally to the distance from the camera relative to the actual position that is imagined to correspond to the position reflected in the image.

7. The sheet shape determination device for metal strips as described in any one of claims 4 to 6, characterized in that, The camera is positioned at a height of more than 1m and less than 5m above the metal strip, and at a distance of more than 5m and less than 40m from the end of the metal strip.

8. A rolling mill, characterized in that, have: The sheet shape determination device for the metal strip as described in any one of claims 1 to 7; and Control device, Based on the judgment result signal, the control device sends one or more operation signals related to the leveling amount, bending force, or paired cross angle of the rolling mill.

9. A method for determining the shape of a metal strip, specifically a method for determining the shape of a rolled metal strip, comprising: The shooting steps involve capturing an image using a camera of the area containing the rolled metal strip that reflects a strip of light transversely along the width of the strip; and The image processing step, based on the image captured in the shooting step, determines the shape of the metal strip plate. The method for determining the shape of the metal strip is characterized by the following: In the image processing step, when the region in the image is divided into multiple partitions along the width direction of the metal strip and the value showing the center position of each partition in the width direction is set as variable (x), the width direction position of the region in the image is converted into the range of -1≤x≤1, and the index information representing the size of the region in each partition is set as the Chebyshev polynomial of the distribution E(x) of each partition, consisting only of 0th, 1st, 2nd and 4th degree terms, and the Chebyshev polynomial coefficient vector (C0', C1', C2', C4') is obtained as information corresponding to the distribution of the plate waviness in the width direction in the rolling direction. One or more of the 0th degree coefficient (C0'), 1st degree coefficient (C1'), 2nd degree coefficient (C2') and 4th degree coefficient (C4') are sent as the judgment result signal of the plate waviness distribution in the width direction. E(x)=C0'+C1'×x+C2'×(2x 2 -1)+C4'×(8x 4 -8x 2 +1) Where -1≤x≤1···(1).

10. The method for determining the shape of a metal strip as described in claim 9, characterized in that, As the indicator information, for each pixel unit in the width direction of the plate in the partition, the average length in the rolling direction, the median length in the rolling direction, or the area in the partition is calculated.

11. The method for determining the shape of a metal strip as described in claim 9, characterized in that, In the image processing step, a display signal is sent to the display device to display the 0th-order component (C0'), the 1st-order component (C1'×x), and the 2nd-order component (C2'×(2x)) in the formula (1). 2 -1)) and 4th component (C4'×(8x 4 -8x 2 The graph of the function of the term +1)) is a graph of one or more of the degree components related to the coefficient sent as the judgment result signal.

12. The method for determining the shape of a metal strip as described in any one of claims 9 to 11, characterized in that, The camera is a camera positioned on the outer side of the metal strip in the width direction when viewed from above. In the image processing step, a correction process is performed in such a way that the further away the actual position corresponding to the position reflected in the image is from the camera, the longer the distance between the positions of two points in the image becomes.

13. The method for determining the shape of a metal strip as described in any one of claims 9 to 11, characterized in that, The camera is positioned on the outer side of the metal strip in the width direction when viewed from above. In the image processing step, when calculating the average length or the median length in the rolling direction in the index information, the processing is performed such that the length in the rolling direction of the region increases proportionally with respect to the actual position that is conceived to correspond to the position reflected in the image from the distance from the camera.

Images