Warpage measurement method of metal plate, warpage control method, manufacturing method, warpage measurement device, and program

Abstract

To provide a warpage measurement method of a metal plate for detecting warpage of a metal plate with high accuracy, and to provide a warpage measurement device, a program, a warpage control method for reducing the warpage of the metal plate, and a manufacturing method of the metal plate with reduced warpage.SOLUTION: A warpage measurement method of a metal plate includes: an imaging step of capturing an image including at least one of a leading end portion and a trailing end portion of the metal plate; an extraction step of detecting a side surface region corresponding to a side surface portion of the metal plate from the image and extracting a coordinate group representing a boundary position of the side surface region; and a specification step of specifying warpage of the metal plate based on the extracted coordinate group.SELECTED DRAWING: Figure 5

Description

[Technical Field]

[0001] The present disclosure relates to a method for measuring warpage of a metal plate, a method for controlling warpage, a manufacturing method, a warpage measuring device, and a program. [Background technology]

[0002] In a hot rolling line for producing hot-rolled steel sheets, for example, a continuously cast slab with a thickness of about 200 mm to 300 mm is heated to about 1100°C to 1300°C in a heating furnace. After the heat treatment, the slab is rough rolled into a rough bar (sheet bar) with a thickness of about 20 mm to 80 mm by multiple or a single roughing mill. The rolled material made of the sheet bar is then finish rolled to a thickness of about 1 mm to 25 mm by a multi-stand finishing mill.

[0003] During rough rolling or finish rolling, if there are factors that cause vertical asymmetry in the steel plate (rolled material) being rolled or the rolling mill, it is known that warpage occurs, in which the steel plate curves in the rolling direction. Warpage along the longitudinal direction of the steel plate is called upward warpage, and downward warpage is called downward warpage, and these forms can also be combined. Warpage in hot rolling is known to occur due to factors such as the temperature difference between the top and bottom surfaces of the rolled material, the difference in peripheral speed between the top and bottom work rolls, and the difference in diameter between the top and bottom work rolls. Warpage in steel plates is primarily noticeable at the leading and trailing ends of the steel plate. If the warpage becomes excessive, the steel plate may catch on the rolling mill's auxiliary equipment as it passes through the rolling mill, causing equipment failure. Therefore, there is a need to suppress warpage in steel plates. Therefore, control technologies for suppressing warpage in steel plates and warpage detection technologies for detecting the amount of warpage are important.

[0004] In response to this, Patent Document 1 discloses an imaging device installed to the side of a steel plate being transported so that the height position of the imaging center is within a range of 100 mm above the surface level of the minimum plate thickness of the steel plate being transported, and a warp detection device that detects warping at the leading end of the steel plate based on an image including the leading end of the steel plate captured by the imaging device.In Patent Document 1, the position of the leading end of the steel plate is determined on the image, and the edge of the steel plate is detected based on the brightness of the captured image.

[0005] Patent Document 2 discloses a plate warpage detection device that is provided between table rollers on the entry and exit sides of a rolling mill, scans a laser beam vertically on the side of the rolled material in the gap on one side between the table rollers, and detects the position of the reflected light to determine the thickness of the rolled material and the height position of the plate passing in the vertical direction.

[0006] Patent Document 3 discloses a method for measuring the amount of warpage of a rolled material before and after a rolling mill on a hot rolling line, which involves using a camera capable of measuring brightness in wavelength bands from visible light to near-infrared to capture an image of the rolled material from diagonally above the rolled material after rolling, detecting widthwise edges of the rolled material based on the brightness values ​​of the captured image of the rolled material, dividing the image of the rolled material from the leading end of the rolled material along the rolling direction at a pitch set according to the radius of the work rolls of the rolling mill, performing quadratic approximation on the shape of the widthwise edge in each divided image, and quantifying the amount of warpage of the rolled material as a curvature based on the quadratic-approximated shape of the widthwise edge. Patent Document 3 also describes a method for detecting the edge of a steel sheet by using an information processing device to read the brightness value of each pixel in the height direction, calculate the average brightness value of multiple pixels, and detect the position where the difference between the average brightness values ​​of multiple adjacent pixels is greatest as the edge of the steel sheet. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] Japanese Patent Application Laid-Open No. 2009-250723 [Patent Document 2] Japanese Patent Application Laid-Open No. 2006-7235 [Patent Document 3] Japanese Patent Application Publication No. 2019-181562 Summary of the Invention [Problem to be solved by the invention]

[0008] The technology described in Patent Document 1 uses an image of a steel sheet captured from the side using an imaging device (CCD camera), integrates brightness in the horizontal direction at each vertical position in the image, and then performs a differentiation process on the integrated brightness. Furthermore, the warpage of the steel sheet is detected by taking a moving average of the integrated brightness values ​​after the differentiation process. In this case, when distinguishing the boundary between the steel sheet and its surroundings based on the brightness of the captured image, a binarization process is first performed using a preset threshold value for brightness. However, if the captured image contains steam, water, fumes, or the like scattered around the steel sheet, the boundary between the steel sheet and its surroundings cannot be accurately distinguished in the initial binarization process. Therefore, when performing the binarization process, it is necessary to set the threshold value through trial and error, which results in a problem of time-consuming adjustment of the detection device.

[0009] The technology described in Patent Document 2 employs a method of scanning a laser beam vertically on the side of the rolled material, but there is a problem in that if steam, water, fumes, etc. are present around the steel plate, the laser beam may be absorbed or scattered, making it impossible to identify the rolled material being detected.

[0010] The technology described in Patent Document 3 detects the width edge of a rolled material based on the brightness values ​​of an image of the rolled material. The brightness value of each pixel in the height direction of the image is read, the average brightness value of multiple pixels is calculated, and the position where the difference between the average brightness values ​​of adjacent multiple pixels is greatest is identified as the edge of the steel plate. However, because the difference between the average brightness values ​​of multiple pixels is used as a criterion for identifying the edge of the steel plate, if steam, water, fumes, etc. are present within a certain range around the edge of the steel plate, the edge of the steel plate may not be identified accurately, which leaves room for improvement. Furthermore, Patent Document 3 approximates the shape of the width edge in the captured image by a quadratic equation and quantifies the amount of warpage of the rolled material by curvature based on the shape of the width edge that has been approximated by the quadratic equation. This leaves room for improvement, as it may be difficult to detect warpage in a rolled material with a complex curvature distribution.

[0011] The present disclosure has been made to solve the above-mentioned problems, and an object of the present disclosure is to provide a metal plate warpage measurement method, a warpage measurement device, and a program for detecting warpage of a metal plate with high accuracy. Another object of the present disclosure is to provide a warpage control method for reducing warpage of a metal plate. Another object of the present disclosure is to provide a metal plate manufacturing method with reduced warpage. [Means for solving the problem]

[0012] (1) A method for measuring warpage of a metal plate according to one embodiment of the present disclosure includes an imaging step of capturing an image including at least one of a leading end and a trailing end of a metal plate, an extraction step of detecting a side region corresponding to a side portion of the metal plate from the image and extracting a group of coordinates representing the boundary position of the side region, and an identification step of identifying the warpage of the metal plate based on the extracted group of coordinates.

[0013] (2) In the extraction step of the method for measuring warpage of a metal plate described in (1) above, an area in the image that includes a portion where the boundary between the side surface of the metal plate and the upper or lower surface of the metal plate is not captured may be detected as the side surface area.

[0014] (3) In the extraction step of the method for measuring warpage of a metal plate described in (1) or (2) above, a part of the boundary of the region detected as the side region may not appear in the image.

[0015] (4) In the extraction step of the method for measuring warpage of a metal plate described in any one of (1) to (3) above, the image may be input and the coordinate group may be extracted using a neural network trained to output a coordinate group representing the boundary position of the side region of the metal plate contained in the image.

[0016] (5) In the identification step of the method for measuring warpage of a metal plate described in any one of (1) to (4) above, a group of coordinates representing at least one of the upper surface and the lower surface of the metal plate may be selected from the group of coordinates representing the boundary positions of the side region, and the warpage of the metal plate may be identified based on the selected group of coordinates.

[0017] (6) In the imaging step of the method for measuring warpage of a metal plate described in any one of (1) to (5) above, an image including at least one of a leading end and a trailing end of the metal plate may be captured at the exit side of a rolling mill that rolls the metal plate.

[0018] (7) A method for controlling warpage of a metal plate according to an embodiment of the present disclosure includes a step of setting operating conditions for another rolling mill arranged downstream of the rolling mill based on the warpage of the metal plate measured using the method for measuring warpage of a metal plate described in (6) above.

[0019] (8) In the method for controlling warpage of a metal plate described in (7) above, the operating conditions of the other rolling mill may be at least one of a water-cooling condition of a water-cooling device arranged on the inlet side of the other rolling mill, a difference in peripheral speed between upper and lower work rolls of the other rolling mill, a pick-up amount of the other rolling mill, or a shape ratio of the other rolling mill.

[0020] (9) A method for controlling warpage of a metal plate according to one embodiment of the present disclosure includes a step of setting operating conditions of the rolling mill for other metal plates following the metal plate based on the warpage of the metal plate measured using the method for measuring warpage of a metal plate described in (6) above.

[0021] (10) In the method for controlling warpage of a metal plate described in (9) above, the operating conditions of the rolling mill for the other metal plate may be at least one of a water-cooling condition of a water-cooling device arranged on the inlet side of the rolling mill, a difference in peripheral speed between upper and lower work rolls of the rolling mill, a pickup amount in the rolling mill, or a shape ratio in the rolling mill.

[0022] (11) A method for manufacturing a metal plate according to one embodiment of the present disclosure includes a step of manufacturing a metal plate using the method for controlling warpage of a metal plate described in any one of (7) to (10) above.

[0023] (12) A metal plate warpage measuring device according to one embodiment of the present disclosure includes an acquisition unit that acquires an image including at least one of the leading end or trailing end of a metal plate, an extraction unit that detects a side region of the metal plate from the image acquired by the acquisition unit and extracts a set of coordinates representing the boundary positions of the detected side region, and an identification unit that identifies the warpage of the metal plate based on the set of coordinates extracted by the extraction unit.

[0024] (13) A program according to one embodiment of the present disclosure is a program for causing a computer to function as a metal plate warpage measurement device, and the program enables the computer to realize the following functions: inputting an image including at least one of the leading end and the trailing end of a metal plate; detecting a side region of the metal plate from the image and extracting a set of coordinates representing the boundary positions of the detected side region; and identifying the warpage of the metal plate based on the extracted set of coordinates. [Effects of the Invention]

[0025] According to the metal plate warpage measurement method, warpage measurement device, and program of the present disclosure, the warpage of the metal plate can be detected with high accuracy. Furthermore, according to the metal plate warpage control method of the present disclosure, the warpage of the metal plate can be reduced. Furthermore, according to the metal plate manufacturing method of the present disclosure, a metal plate with reduced warpage can be manufactured. [Brief explanation of the drawings]

[0026] [Figure 1] 1 is a block diagram illustrating a configuration example of an information processing system according to the present disclosure. [Figure 2] FIG. 1 is a diagram illustrating a configuration example of a hot rolling line. [Figure 3] FIG. 2 is a diagram illustrating a configuration example of a finishing rolling mill. [Figure 4] FIG. 1 is a diagram illustrating an example of the configuration of a four-high rolling mill. [Figure 5] 1 is a flowchart showing an example of a procedure for a method for measuring warpage of a metal plate according to the present disclosure. [Figure 6A] FIG. 2 is a side view showing an example of the arrangement of the imaging device. [Figure 6B] FIG. 2 is a front view showing an example of the arrangement of the imaging device. [Figure 7] 10 is an example of an image of a tip of a metal plate. [Figure 8] 8 is an image showing a region representing a side portion of a metal plate, which is sectioned as a side region in the image of FIG. 7. [Figure 9] 9 is an image showing the boundary positions of the side surface portions of the metal plate extracted by applying a binary mask to the side surface region of FIG. 8. [Figure 10] FIG. 6 is a block diagram illustrating an example of a model for performing the extraction step of the method of FIG. 5. [Figure 11] 8 is an image showing an example of a first feature map for identifying the position of a side surface portion from the image of FIG. 7 using anchor boxes. [Figure 12] 10 is an example of an image in which coordinate groups of the upper and lower surfaces of a metal plate are identified from coordinate groups representing boundary positions of side surfaces of the metal plate extracted in FIG. 9. [Figure 13] FIG. 1 is a diagram schematically illustrating two rolling mills arranged in series. [Figure 14] 1 is a diagram schematically illustrating a situation in which a second metal sheet enters a rolling mill and is rolled after a first metal sheet. [Figure 15] 10A and 10B are diagrams illustrating a method for measuring warpage of a metal plate according to a comparative example. [Figure 16A] FIG. 10 is a diagram showing an example of a measurement result obtained by a warpage measurement method according to a comparative example. [Figure 16B] FIG. 2 is a diagram showing an example of a measurement result obtained by the warpage measurement method according to the first embodiment of the present disclosure. [Figure 17] 10 is a graph showing an example of the results of controlling warpage based on the warpage measurement results. [Figure 18] 10 is an example of an image obtained in a warpage measurement method according to a second embodiment of the present disclosure. [Figure 19] 19 is an example of a side region extracted from the image of FIG. 18. DETAILED DESCRIPTION OF THE INVENTION

[0027] Hereinafter, embodiments of a metal plate warpage measurement method, warpage control method, manufacturing method, warpage measurement device, and warpage measurement program according to the present disclosure will be described with reference to the drawings. The drawings are schematic and may differ from the actual product. Furthermore, the following embodiments exemplify devices or methods for embodying the technical ideas of the present disclosure, and are not intended to limit the configuration to those described below. In other words, the technical ideas of the present disclosure can be modified in various ways within the technical scope described in the claims.

[0028] The present disclosure relates to a method for measuring warpage occurring at a leading end or a trailing end of a metal plate. The warpage measurement method of the present disclosure includes an imaging step of capturing an image including at least one of the leading end and the trailing end of the metal plate, an extraction step of detecting a side region of the metal plate from the image and extracting a set of coordinates representing the boundary positions of the detected side region, and an identification step of identifying the warpage of the metal plate based on the extracted set of coordinates. By configuring in this manner, the warpage of the metal plate can be measured with high accuracy even in an environment where steam or cooling water splashes.

[0029] <Metal plate warpage measuring device 30> An example of the configuration of an information processing system 1 according to the present disclosure will be described with reference to Fig. 1. The information processing system 1 includes a hot rolling line 10, a warpage measuring device 30, and a display device 44.

[0030] The hot rolling line 10 includes a control controller (PLC) 102 for controlling each piece of equipment that constitutes the hot rolling line 10, a control computer (process computer) 101 that gives control commands to the control controller, and a host computer 100 that gives manufacturing instructions to the hot rolling line 10.

[0031] Various controls such as the thickness or width of the metal plate, which is the material S to be rolled (see FIG. 2, etc.) in the hot rolling line 10, are performed by a control computer 101 setting control target values ​​for each device (see FIGS. 2 and 3) constituting the hot rolling line 10 based on a host computer 100 or a manufacturing instruction from the host computer 100, and setting operating conditions for each device. In addition, a control controller 102 has a function of collecting information obtained from various sensors (tracking sensors, thermometers, thickness gauges, etc.) installed in the hot rolling line 10 at a predetermined sampling period and outputting the information to the control computer 101.

[0032] The hot rolling line 10 further includes an imaging device 21. As will be described later, the imaging device 21 is arranged in the hot rolling line 10 so as to be able to capture images including the leading end or the trailing end of the rolled material in the hot rolling line 10. A camera that captures images in the visible light range, such as a normal CCD camera, may be used as the imaging device 21. Water, steam, etc. are present around the rolled material being transported in a high-temperature state, and these may act as disturbances that make it difficult for light in the near-infrared range or the infrared range to pass through. For this reason, it is preferable to use a camera that captures images in the visible light range as the imaging device 21, rather than a camera that captures images in the near-infrared range or the infrared range.

[0033] The warpage measuring device 30 includes a storage unit 32, an acquisition unit 33, an extraction unit 34, an identification unit 35, and an output unit 36. The warpage measuring device 30 acquires identification information such as the serial number or product number of the metal plate, which is the rolled material S to be measured for warpage, from the control computer 101. The warpage measuring device 30 can communicate with the control computer 101 via a network, and may acquire the identification information of the metal plate, which is the rolled material S, via another operation data server.

[0034] The warpage measuring device 30 inputs an image (see Figure 7) including at least one of the leading end and the trailing end of the metal plate, which is the rolled material S, extracts a coordinate group Pi (see Figure 9) representing the boundary position GB (see Figure 9) between the side region GS (see Figure 8) extracted by the extraction model and other regions, and further performs a process to identify the warpage of the metal plate, which is the rolled material S, based on the extracted coordinate group Pi.

[0035] The storage unit 32 is, for example, an information recording medium such as an updatable flash memory, a hard disk, or a memory card. The storage unit 32 stores, for example, programs or data for executing each function of the warp measuring device 30. The storage unit 32 may store an extraction model generated in advance for executing the extraction step S12 (see FIG. 5). The storage unit 32 may store a correspondence relationship between pre-specified dimensions on the image G and the actual dimensions of the metal plate, which is the rolled material S.

[0036] The acquisition unit 33 may include a communication interface that acquires an image including at least one of the leading end and the trailing end of the metal plate, which is the rolled material S, captured by the imaging device 21. The acquisition unit 33 acquires the image captured by the imaging device 21 via a network. The acquisition unit 33 may be configured to acquire the image from the control computer 101 or another data server in which the image captured by the imaging device 21 is stored.

[0037] The extraction unit 34 detects a side region GS of the metal plate, which is the material S to be rolled, from the image G including at least one of the leading end and the trailing end of the metal plate, which is the material S to be rolled, acquired by the acquisition unit 33, and extracts a coordinate group Pi representing a boundary position GB of the detected side region GS. The extraction unit 34 may extract the coordinate group Pi by inputting the image G into an extraction model stored in the memory unit 32.

[0038] The identification unit 35 identifies warpage of the metal plate, which is the rolled material S, based on the coordinate group Pi extracted by the extraction unit 34. The identification unit 35 identifies at least one of the coordinate group CV1 representing the shape on the top surface of the metal plate, which is the rolled material S, and the coordinate group CV2 representing the shape on the bottom surface from the coordinate group Pi, calculates the difference between the maximum and minimum Y coordinate values ​​of the coordinate group CV1, or the difference between the maximum and minimum Y coordinate values ​​of the coordinate group CV2, and identifies this as warpage.

[0039] The extraction unit 34 or the identification unit 35 may be configured to include at least one processor, such as a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit). The extraction unit 34 and the identification unit 35 may each be configured by separate processors. The extraction unit 34 and the identification unit 35 may be combined into one processor. The extraction unit 34 or the identification unit 35 may be configured by one processor or multiple processors. The processor configuring the extraction unit 34 or the identification unit 35 may implement the functions of the warpage measuring device 30 by reading and executing a program stored in the storage unit 32.

[0040] The output unit 36 ​​outputs the measured value of the warpage of the metal plate, which is the rolled material S, identified by the identification unit 35 to the display device 44. The output unit 36 ​​may be configured to output the measured value of the warpage of the metal plate, which is the rolled material S, to the control computer 101 or the control controller 102. The output unit 36 ​​is configured to output the measured value of the warpage of the metal plate, which is the rolled material S, to these via a network.

[0041] The display device 44 may be configured to include a liquid crystal display, an organic EL panel, or the like. The display device 44 may also be configured as a display of a terminal device such as a smartphone or a tablet. The output unit 36 ​​may be configured to output to the display device 44 the identification information of the metal plate that is the rolled material S, along with the measured value of the warp of the metal plate that is the rolled material S. This allows the operator to take action to reduce the warp of the metal plate that is the rolled material S, based on the measured value of the warp of the metal plate that is the rolled material S displayed on the display device 44. In other words, by displaying the measured value of the warp of the metal plate that is the rolled material S on the display device 44, the warp measuring device 30 can function as an operation guidance device.

[0042] On the other hand, if the output unit 36 ​​is configured to output the measured value of the warpage of the metal plate, which is the rolled material S, to the control computer 101 or the control controller 102, the control computer 101 or the control controller 102 can be configured to set the operating conditions for controlling the warpage of the metal plate, which is the rolled material S, based on the measured value of the warpage. In other words, a metal plate warpage control device can be configured.

[0043] The warpage measuring device 30 may be realized by, for example, a computer. The computer may include, for example, a memory or a hard disk drive (storage device), and a CPU (processing device). The program may be stored in the hard disk drive and may be read from the hard disk drive to the memory when executed by the CPU. Data during processing is stored in the memory and, if necessary, stored in the hard disk drive. The storage unit 32 may be realized by, for example, a storage device. The functions of the acquisition unit 33, extraction unit 34, identification unit 35, and output unit 36 ​​may be realized by, for example, the CPU reading and executing the program.

[0044] <Hot rolling line 10> FIG. 2 shows the configuration of a hot rolling line 10 to which the metal sheet warpage measurement method according to the present disclosure is applied. The hot rolling line 10 shown in FIG. 2 includes a heating furnace 11, a descaling device 12, a width reduction device 13, a roughing mill 14, a finishing mill 15, a water cooling device 16, and a coiler 17. In the hot rolling line 10, a cast slab is charged into the heating furnace 11, heated to a predetermined set temperature, and extracted from the heating furnace 11 as a hot slab. The slab extracted from the heating furnace 11 has primary scale formed on its surface removed by the descaling device 12, and then its width is reduced to a predetermined set width by the width reduction device 13. The width-reduced slab is then rolled to a predetermined thickness in the roughing mills 14 (reversing mill 14a and non-reversing mill 14b) to be transported to the finishing mill 15 as a rough bar (sheet bar). In the finishing mill 15, the rough bar is rolled to the product thickness by a continuous rolling mill including any number of stands between five and seven. Downstream of the finishing mill 15, a water cooler 16 is provided in a facility called a run-out table, and the hot-rolled steel sheet is cooled to a predetermined temperature by the water cooler 16 and then wound into a coil by a coiler 17. Note that a rough bar cooler 18 for cooling the rough bar may be provided between the roughing mill 14 and the finishing mill 15. The hot rolling line 10 is provided with a transport mechanism for transporting the material S to be rolled (slab or steel sheet) until the slab extracted from the heating furnace 11 is wound into a coil by the coiler 17.

[0045] The material S to be rolled transported in the hot rolling line 10 is heated in the heating furnace 11 to, for example, 1100°C to 1300°C. The temperature of the material S to be rolled reaches 950°C to 1150°C when rough rolling by the roughing mill 14 is completed, and reaches 850°C to 1200°C when it is loaded into the finishing mill 15. The temperature of the material S to be rolled reaches 700°C to 1000°C when it is discharged from the finishing mill 15, and reaches 400°C to 800°C before it is coiled by the coiler 17.

[0046] In the hot rolling line 10 shown in Fig. 2, cameras 21a and 21b are arranged as the imaging device 21 for performing the metal plate warpage measuring method according to this embodiment. In the hot rolling line 10 shown in Fig. 2, the camera 21a is arranged downstream of the non-reversing rolling mill 14b. In addition, the camera 21b is arranged between the first stand and the second stand of the finishing rolling mill 15. The cameras 21a and 21b are arranged so as to be able to capture an image G (see Fig. 7, etc.) including the leading end or the trailing end of the rolled material S.

[0047] FIG. 3 is a schematic diagram showing an example of the configuration of the finishing mill 15 of the hot rolling line 10. After rough rolling by the roughing mill 14, the rough-rolled material (rough bar) is transported to the entrance of the finishing mill 15. A crop shear 23 is disposed at the entrance of the finishing mill. The crop shear 23 is a device that cuts and removes crops (irregularly shaped portions at the leading and trailing ends of the rough bar) formed at the leading and trailing ends of the rough bar. This shapes the material S to be rolled into a substantially rectangular planar shape that allows it to be easily fitted into the finishing mill 15.

[0048] The finishing mill 15 shown in Figure 3 is made up of seven stands F1 to F7, but the number of stands is not limited to seven. Generally, the number of stands in the finishing mill 15 is six to seven, and in some cases it is made up of five stands. The finishing mill 15 takes the form of a hot tandem finishing mill in which multiple stands simultaneously roll a rough bar that has been cooled to a temperature in the range of 800°C to 1100°C that is set depending on the steel type, etc., but is simply referred to as a "finishing mill" for short.

[0049] Strip cooling devices SC for cooling the material S to be rolled may be arranged between the stands of the finishing mill 15. The strip cooling devices SC are configured to spray cooling water onto the top and bottom sides of the material S to be rolled, and may be configured so that the amount of cooling water on the top side and the amount of cooling water on the bottom side can be set to different values. In the finishing mill 15 shown in FIG. 3, a strip cooling device SC0 is arranged on the inlet side of the first stand F1. Furthermore, strip cooling devices SC1 to SC6 are arranged on the outlet sides of each of the first stand F1 to the sixth stand of the finishing mill 15.

[0050] FIG. 4 is a schematic diagram of a rolling mill that constitutes the hot rolling line 10. The rolling mill 60 shown in FIG. 4 is a four-high rolling mill, and is equipped with a pair of work rolls 61a and 61b, one above the other and one below the pass line PL. The work rolls 61a and 61b are supported by backup rolls 62a and 62b, respectively. The pass line PL of the rolling mill 60 is set as a reference height for the threading position of the material S to be rolled, and is set, for example, to the height of the upper end of the lower work roll 61b. At this time, the height of the bottom surface of the material S to be rolled when the material S is charged into the rolling mill 60 can be set by an inlet guide or the like, and the relative height of the bottom surface of the material S with respect to the pass line PL is sometimes referred to as the pick-up amount. In addition, the rolling mill 60 may be configured so that the rotational speed of the upper work roll 61a and the rotational speed of the lower work roll 61b can be set to different speeds, in which case a difference in peripheral speed can be imparted between the rotation of the upper work roll 61a and the rotation of the lower work roll 61b.

[0051] <Warp measurement method> An embodiment of the method for measuring warpage of a metal plate according to the present disclosure will be described below.

[0052] 5 is a flowchart of a method for measuring warpage of a metal plate. The method for measuring warpage of a metal plate includes an imaging step S11 of capturing an image including a leading end or a trailing end of a metal plate that is a rolled material S, an extraction step S12 of detecting a side region of the metal plate that is the rolled material S from the image captured in the imaging step and extracting a set of coordinates that represent the boundary positions of the detected side region, and an identification step S13 of identifying the warpage of the metal plate that is the rolled material S based on the set of coordinates extracted in the extraction step S12.

[0053] The method for measuring warpage of a metal plate may be realized as a program for measuring warpage of a metal plate to be executed by a processor constituting the extraction unit 34 or the identification unit 35. The program for measuring warpage of a metal plate may be stored in a non-transitory computer-readable medium.

[0054] In the imaging step S11, the imaging device 21 captures an image including at least one of the leading and tailing ends of the metal plate, which is the material S to be rolled. Warpage that occurs in the metal plate, which is the material S to be rolled, is often greatest at the leading or tailing end, and if the warpage at the leading or tailing end of the metal plate, which is the material S to be rolled, is large, operational problems are likely to occur when the metal plate, which is the material S to be rolled, is transported or passed through a rolling mill. In order to prevent such operational problems, the warpage at at least one of the leading and tailing ends of the metal plate, which is the material S to be rolled, is measured.

[0055] The imaging device 21 for performing the imaging step S11 may be a still camera that captures still images or a video camera that captures moving images. When a still camera is used as the imaging device 21, the imaging device 21 may control the imaging timing in accordance with the movement of the metal plate that is the rolled material S being transported, and capture an image that includes the leading end or the tail end of the metal plate that is the rolled material S. When a video camera is used as the imaging device 21, the imaging device 21 may select an image that includes the leading end or the tail end of the metal plate that is the rolled material S from multiple captured images. For example, with regard to the leading end of the metal plate that is the rolled material S, the imaging device 21 may refer to a brightness value at a specific coordinate on the image that corresponds to the traveling direction of the metal plate that is the rolled material S, and when the brightness value at the set coordinate is equal to or greater than a predetermined threshold, determine that the leading end has passed, and select this image as an image that includes the leading end of the metal plate that is the rolled material S. In addition, for the tail end of the metal plate being the rolled material S, the imaging device 21 refers to the brightness value at a specific coordinate on the image corresponding to the direction of travel of the metal plate being the rolled material S, and when the brightness value at the set coordinate is below a predetermined threshold value, it determines that the tail end has passed, and selects this image as an image including the tail end of the metal plate being the rolled material S.

[0056] 2, the imaging step S11 can be performed using cameras 21a and 21b as the imaging device 21. That is, in the imaging step S11, an image including at least one of the leading end and the trailing end of the metal plate, which is the material S to be rolled, may be captured, for example, downstream of the non-reversing rolling mill 14b of the hot rolling line 10 where the camera 21a is installed, or between stands of the finishing rolling mill 15 where the camera 21b is installed.

[0057] 6A and 6B are diagrams illustrating the placement of the imaging device 21 relative to the metal plate being the material S to be rolled. FIG. 6A is a side view of the rolling mill 60 and the downstream region in the conveyance direction of the metal plate being rolled, and FIG. 6B is a front view. The imaging device 21 is preferably located downstream of the rolling mill 60 and at a position where it can capture an image of the metal plate being rolled at an angle downward. Specifically, the imaging device 21 is located at a distance L1 of 1 m to 3 m from the housing 63 of the rolling mill 60. Furthermore, the imaging device 21 is preferably located slightly off to the DR side (drive side) of the work rolls 61 a or 61 b ​​or backup rolls 62 a or 62 b of the rolling mill 60, so as to capture an image of the metal plate being rolled at an angle downward. For example, the imaging device 21 may be located at a height H of 2 m from the pass line PL of the rolling mill 60 and at a distance L2 of 3 m from the center position (mill center) in the width direction of the rolling mill 60. In this case, if the imaging device 21 has an element size (length x width) of 7.09 mm x 8.5 mm and a lens focal length f of 16 mm, the distance between the metal plate being the rolled material S and the camera is approximately 3.61 m. In this case, the range imaged by the imaging device 21 is 2.04 m in the longitudinal direction of the metal plate being the rolled material S. By configuring the imaging device 21 in this manner, the imaging device 21 can capture an image that includes a range of approximately 2 m from the leading end of the metal plate being the rolled material S. Similarly, the imaging device 21 can capture an image that includes a range of approximately 2 m from the tail end of the metal plate being the rolled material S. However, the imaging range of the metal plate being the rolled material S by the imaging device 21 can be adjusted by appropriately selecting the installation position of the imaging device 21 or the focal length of the lens.

[0058] FIG. 7 shows an example of image G, captured using the imaging device 21, of the leading end of a metal plate, which is the material S to be rolled. Image G, captured in imaging step S11, captures the leading end LE, top surface US, and side surface SF of the metal plate, which is the material S to be rolled. The leading end LE, top surface US, and side surface SF of the metal plate, which is the material S to be rolled, have different surface temperatures, which can be visually distinguished by differences in color, lightness, or luminance in image G. That is, the leading end LE, top surface US, and side surface SF of the metal plate, which is the material S to be rolled, face in different directions, and therefore emit different amounts of radiant heat depending on the conditions of the surrounding equipment, resulting in differences in surface temperature between the regions. Furthermore, the temperature of the top surface US drops due to heat removal from the top surface, whereas the side surface SF is affected by heat removal from the sides as well as from the top and bottom surfaces, resulting in a difference in surface temperature between the top surface US and the side surface US. The front end portion LE is also affected by heat transfer from the front end surface as well as the top and bottom surfaces, resulting in a difference in surface temperature from the front end portion LE and the top surface portion US. The front end portion LE of the metal plate serving as the rolled material S can be identified because the boundary in the conveying direction of the metal plate serving as the rolled material S has a different brightness from the surrounding (background) BG. The top surface portion US of the metal plate serving as the rolled material S can be identified as a large area extending downstream in the conveying direction from the front end portion LE of the metal plate serving as the rolled material S. The side surface portion SF of the metal plate serving as the rolled material S can be identified as a region that is lower in temperature and lower in brightness than the top surface portion US, and can be visually identified because the underside of the metal plate serving as the rolled material S is positioned as the surrounding (background) BG. It is preferable that the image G captured in the imaging step S11 include three or more pixels in the thickness direction of the side surface portion SF of the metal plate serving as the rolled material S, so that the side surface region GS, described below, can be accurately extracted. 6A and 6B, it is preferable to apply the method for measuring warpage of a metal plate according to this embodiment when the thickness of the metal plate, which is the rolled material S, is 2 mm or more. More preferably, it is preferable to apply the method for measuring warpage of a metal plate according to this embodiment when the thickness of the metal plate is 6 mm or more.

[0059] In extraction step S12, the acquisition unit 33 of the warp measuring device 30 acquires the image G from the imaging device 21. As shown in Fig. 8, the extraction unit 34 of the warp measuring device 30 detects the side region GS of the metal plate, which is the rolled material S, from the image G captured in the imaging step S11, and extracts a coordinate group representing the boundary position of the detected side region GS. In this case, the extraction step S12 may be divided into a detection step of detecting the side region GS of the metal plate, which is the rolled material S, from the image G, and a coordinate group extraction step of extracting a coordinate group representing the boundary position GB of the detected side region GS, as shown in Fig. 9.

[0060] In the detection step, the extraction unit 34 identifies a side portion SF from the image G illustrated in Fig. 7 and detects a side region GS that is a region surrounding the boundary of the side portion SF. In the detection step, a deep learning model that performs object detection and segmentation from the image may be used.

[0061] Object detection is a method for detecting the bounding box of each object area in an input image and classifying the object. In other words, it detects a bounding box that encloses the region of interest (ROI) of an object contained in the image as a rectangular area, and performs class classification for the object in the ROI.

[0062] Specifically, in the detection step, a bounding box enclosing the "side portion SF" as a region of interest (ROI) is detected as an object contained in image G. In the detection step, a bounding box enclosing each of the "tip portion LE," "top portion US," "side portion SF," and "surrounding BG" as a region of interest (ROI) is detected as an object contained in image G, and the object class may be identified from these regions of interest. This allows the "side portion SF" to be reliably identified from image G.

[0063] Next, in the detection step, segmentation is performed on the region of interest identified as the object class "side portion SF." Segmentation is a technique for detecting objects at the pixel level within an image within a bounding box, which is the region of interest. Instance segmentation is a suitable segmentation technique. Instance segmentation distinguishes instances of an object class in an image from each instance. In other words, it distinguishes the "side portion SF" from other instances and detects them at the pixel level as an object contained in image G. Furthermore, it distinguishes and detects the "tip portion LE," "top portion US," "side portion SF," and "surrounding BG" as objects contained in image G. It also distinguishes and detects the boundaries between the "side portion SF" and the "top portion US," the boundaries between the "side portion SF" and the "tip portion LE," and the boundaries between the "side portion SF" and the "surrounding BG" at the pixel level in image G. The "side portion SF" identified by object detection and segmentation through the detection step in this manner is called the side region GS.

[0064] 8 shows an example of segmentation performed on a region of interest whose object class was identified as "side part SF" in the detection step. As shown in FIG. 8, the region representing the side part SF is partitioned in the image G as a side region GS (the region surrounded by a thick solid line).

[0065] The coordinate group extraction step, which extracts a coordinate group representing the boundary position GB in the side region GS detected in the detection step, can be performed by applying a binary mask to the side region GS. Binary masking refers to a process of encoding the spatial arrangement of the input object ("side region GS"), and generates a binarized image of the same size as the original image by, for example, assigning "1" to the side region GS and "0" to other regions including the surrounding BG. In this way, by applying the binary mask to the side region GS in the coordinate group extraction step, information on the boundary position (boundary position GB) of the side region GS is extracted.

[0066] 9 shows an example in which the boundary of the side surface portion SF is extracted as the boundary position GB by applying a binary mask to the side surface region GS. In this case, by setting a two-dimensional coordinate system on the image G, the points (e.g., pixels) that make up the boundary position GB can be extracted as a coordinate group Pi(Xi, Yi). In the detection step included in the extraction step S12, it is preferable to use Mask R-CNN (Region Convolutional Neural Network) as a deep learning model that performs object detection and segmentation from the image.

[0067] Fig. 10 shows an example of the configuration of a deep learning model using Mask R-CNN as an extraction model for executing extraction step S12. The extraction model M shown in Fig. 10 is configured with an input unit D1, a convolution unit D2, a region candidate setting unit D3, a cutout unit D4, a classification unit D5, a partitioning unit D6, and a mask unit D7.

[0068] The input unit D1 constitutes the input layer of the deep learning model, and receives as input the image G captured in the imaging step S11. The input unit D1 may receive as input a compressed version of the image G captured in the imaging step S11. In this case, the image G is compressed to, for example, 160 x 160 pixels and input.

[0069] The convolution unit D2 is configured with a convolutional neural network (CNN) and extracts features of the image G. The convolutional neural network used in the convolution unit D2 can be a neural network that performs classification tasks, such as VGG or ResNet. This generates a feature map in which the features of the image G are compressed.

[0070] The region candidate setting unit D3 uses the feature map generated by the convolution unit D2 to identify the position of the "side surface SF" (the detection target) on the image G using a rectangular anchor box (see FIG. 11) and sets it as a region proposal. The region candidate setting unit D3 is configured using an RPN (region proposal network) and sets candidates for the region where the "side surface SF" (the detection target) exists. For example, the RPN applied to Faster R-CNN may be used as the region candidate setting unit D3. By using the RPN, region candidates are identified on the image G so as to increase the probability that the "side surface SF" (the detection target) exists. Then, the feature map output from the convolution unit D2 and the region candidates set by the region candidate setting unit D3 are superimposed to generate a first feature map FM1, as illustrated in FIG. 11. The first feature map FM1 is generated as 512-channel information of 16 × 16 pixels.

[0071] The cropping unit D4 uses the information on the region candidates where the "side surface SF" identified by the region candidate setting unit D3 exists to convert (crop) the region candidates on the first feature map FM1 into fixed dimensions as preprocessing for segmentation, thereby generating a second feature map FM2. The cropping unit D4 is configured with an ROI Align layer. The ROI Align layer is provided to address the problem of misalignment of the region candidates due to the resolution of the region candidates identified by the region candidate setting unit D3 being lower than that of the original image G. Specifically, if the first feature map FM1 contains 512 channels of information in 16x16 pixels, and 512 channels of 10x8 pixels are identified as region candidates, this is assigned to 512 channels of information in a fixed size of 7x7 pixels. This generates a second feature map FM2, which contains information on the region candidates assigned to the fixed size.

[0072] The classification unit D5 uses the second feature map FM2 to identify the object class surrounded by the region candidate. That is, the classification unit D5 identifies whether the object surrounded by the region candidate is a "side surface SF." The classification unit D5 may be configured to identify whether the object surrounded by the region candidate is a "tip surface LE," a "top surface US," a "side surface SF," or a "surrounding BG." The classification unit D5 can be configured using a fully connected layer to compress the features of the input second feature map FM2 and output the number of object classes (dimensions). For example, the dimension of the object class is 4 when identifying whether the object is a "tip surface LE," a "top surface US," a "side surface SF," or a "surrounding BG."

[0073] The segmentation unit D6 uses the second feature map FM2 to identify the object class and position on a pixel-by-pixel basis. That is, it identifies whether each pixel in the second feature map FM2 belongs to the object class "side surface SF." The segmentation unit D6 is configured using a convolutional layer and a transposed convolutional layer, and interpolates the input data before performing convolution processing to enlarge it. That is, if the information in the second feature map FM2 is a fixed size of 7x7 pixels, it enlarges it to 14x14 pixels or 28x28 pixels and compresses the number of channels to the dimension of the object class to be identified.

[0074] The masking unit D7 extracts information about the boundary position GB by applying a binary mask to the "side surface portion SF" identified by the segmentation unit D6 on a pixel-by-pixel basis. The binary mask generates a binarized image of the same size as the original image by assigning "1" to the side surface region GS and "0" to other regions, including the surrounding region BG. Therefore, a two-dimensional coordinate system defined for the image G can be used to extract a set of coordinates for the boundary position GB. For example, the coordinates (X, Y) of the boundary position GB are extracted by setting an arbitrary position in the image G as the origin, defining the horizontal direction of the image G as the x-axis, and the vertical direction as the y-axis. The coordinates (X, Y) of the boundary position GB may also be extracted as a set of coordinates converted to actual dimensions by identifying a correspondence between the dimensions in the image G and the actual dimensions of the metal plate being rolled (S) in advance. For example, a component of known dimensions may be imaged by the imaging device 21, and the correspondence between the dimensions in the image G and the actual dimensions of the component may be identified based on the actual dimensions of the component in the image.

[0075] The extraction model M shown in Figure 10 illustrates an example in which functional blocks for executing, as the extraction step S12, a detection step for detecting a side region GS of the metal plate, which is the rolled material S, from the image G and a coordinate group extraction step for extracting a coordinate group representing the boundary position GB of the detected side region GS are connected. However, the extraction model M may also be configured such that the input unit D1, convolution unit D2, region candidate setting unit D3, cut-out unit D4, classification unit D5, and partitioning unit D6 are configured as a single deep learning model, and software that performs the function of the masking unit D7 is separately provided, and the output of the partitioning unit D6 is input to the software that performs the function of the masking unit D7. This is because publicly available or commercially available software may be available for use as software that realizes the functions of the input unit D1, convolution unit D2, region candidate setting unit D3, cut-out unit D4, classification unit D5, and partitioning unit D6 in the extraction model M.

[0076] In the identification step S13, the warpage of the metal plate, which is the material S to be rolled, is identified based on the coordinate group extracted in the extraction step S12. A method for identifying the warpage of the metal plate, which is the material S to be rolled, will be described using Fig. 12. Fig. 12 shows a coordinate group representing the boundary position GB of the side region GS of the metal plate, which is the material S to be rolled, extracted from the image G in the extraction step S12. It should be noted that the coordinate group Pi(Xi, Yi) representing the boundary position GB is made up of N coordinate groups P1(X1, Y1) to PN(XN, YN) (N is an integer of 2 or more).

[0077] Here, the coordinate group Pi representing the boundary position GB is obtained by cutting out the side portion SF of the metal plate, which is the rolled material S, from the image G, and therefore the coordinate group Pi includes information representing the leading end or tail end of the metal plate, which is the rolled material S. Furthermore, the coordinate group Pi includes a boundary portion that does not correspond to the leading end or tail end of the metal plate, which is the rolled material S, but is connected to an area outside the range of the metal plate, which is the rolled material S, captured as image G. In the example shown in Figure 12, the coordinate group Pi includes a coordinate group representing an area R1 that corresponds to the leading end LE of the metal plate, which is the rolled material S, and an area R2 that is connected to a part of the metal plate, which is the rolled material S, that is not captured as image G.

[0078] Therefore, in the identification step S13, coordinate groups corresponding to regions R1 and R2 are excluded from the coordinate group Pi extracted in the extraction step S12 as follows. First, the coordinate group Pi representing the boundary position GB is ordered clockwise relative to the image, and the index of the coordinate group Pi is converted to a new one. That is, for N coordinate groups, the index is reassigned so that the coordinate Pi moves sequentially clockwise relative to the image as the index i increases from 1 to N. Next, since the coordinate group Pi includes coordinates representing the boundary positions GB on the top and bottom sides of the metal plate (the rolled material S), for example, starting from the origin of the x-coordinate of the image, a set of coordinate groups is extracted in which the difference (Xi+1-Xi) between adjacent coordinates is positive. Note that Xi+1 represents the x-coordinate with index i+1. For example, assume that the index i of the set extracted in this manner is i=j, j+1, ..., j+k+1. Next, the difference Yi+1-Yi (i=j, j+1, ..., j+k-1) between the y-coordinates of adjacent points in the extracted set is calculated, and if the value of Yi+1-Yi is negative, it is excluded as belonging to region R1. Note that Yi+1 represents the y-coordinate with index i+1. Also, the ratio (slope) of the difference Yi+1-Yi between the y-coordinates of adjacent points in the extracted set to the difference Xi+1-Xi between the x-coordinates is calculated, and if it is greater than a preset threshold, it is excluded as belonging to region R2. In this way, coordinate groups corresponding to regions R1 and R2 can be excluded from the coordinate group Pi extracted in extraction step S12 as follows.

[0079] In this manner, it is possible to identify a coordinate group CV1 representing the shape of the top surface of the metal plate that is the rolled material S and a coordinate group CV2 representing the shape of the bottom surface of the metal plate that is the rolled material S from the coordinate group Pi that represents the boundary position GB. However, because water or steam is present around the metal plate that is the rolled material S when the image G is captured, there is a possibility that the coordinate group Pi extracted in the extraction step S12 contains errors. In this case, it is advisable to select a coordinate group that represents at least one of the coordinate group CV1 representing the top surface of the metal plate that is the rolled material S and the coordinate group CV2 representing the bottom surface, and identify the warpage of the metal plate that is the rolled material S based on the selected coordinate group. With regard to whether to select the coordinate group CV1 representing the top surface of the metal plate that is the rolled material S or the coordinate group CV2 representing the bottom surface, structures associated with the rolling mill may be included in the captured image of the metal plate that is the rolled material S. In such a case, it is advisable to select the coordinate group that does not include disturbances caused by the structures. The warpage of the metal plate, which is the rolled material S, can be calculated by calculating the difference between the maximum and minimum Y coordinate values ​​from the coordinate group CV1 representing the top surface of the metal plate, which is the rolled material S. Alternatively, the warpage of the metal plate, which is the rolled material S, can be calculated by calculating the difference between the maximum and minimum Y coordinate values ​​from the coordinate group CV2 representing the bottom surface of the metal plate, which is the rolled material S.

[0080] In the above embodiment of the present disclosure, when measuring the warpage of the metal plate that is the rolled material S, an extraction step is performed to detect the side area of ​​the metal plate that is the rolled material S on the image, and to extract a group of coordinates that represent the boundary position of the detected side area. In other words, the side of the metal plate that is the rolled material S is detected as a "face," and the boundary position of the side area is extracted based on the detected "face."

[0081] In contrast, the method according to the comparative example extracts coordinates corresponding to the boundary position of the side region as a "point" or a "line" based on the brightness values ​​of an image of the metal plate. In this case, the method according to the comparative example detects changes in brightness values ​​near the boundary position of the side region and extracts coordinates corresponding to the boundary position of the side region based on these changes. That is, the method according to the comparative example focuses only on the area where brightness values ​​change within the entire image of the metal plate. As a result, the method according to the comparative example extracts coordinates corresponding to the boundary position of the side region using only a portion of the brightness value information contained in the image of the metal plate. Therefore, when attempting to extract coordinates corresponding to the boundary position of the side region as a "point" or a "line," a problem arises in that the boundary position may not be clearly extracted due to disturbances such as water or steam. In the above-described embodiment of the present disclosure, the side of the metal plate, which is the rolled material S, is detected as a "surface." Therefore, even if the image contains disturbances such as water or steam, the method is less susceptible to such influences, and the boundary position of the side region can be extracted with higher accuracy than the comparative example. That is, in the above-described embodiment of the present disclosure, coordinates corresponding to the boundary positions of the side surface regions are extracted using information on the overall brightness values ​​of the image obtained by capturing an image of the metal plate.

[0082] Furthermore, according to the above embodiment of the present disclosure, even if an image of the leading end or the tail end of a metal plate that is the rolled material S has a portion that is not captured at the boundary between the side surface and the top surface of the metal plate, or at the boundary between the side surface and the bottom surface of the metal plate, the warp measuring device 30 can detect the side surface of the metal plate that is the rolled material S as a "surface" and thereby detect the region including the portion where the boundary is not captured as a side surface region. In other words, the warp measuring device 30 can detect an image of the leading end or the tail end of a metal plate that is the rolled material S and a region including a portion where the boundary between the side surface and the top surface or the bottom surface of the metal plate is not captured as a side surface region.

[0083] Furthermore, according to the above embodiment of the present disclosure, even if part of the boundary of the area detected as a side area is not visible in the image showing the leading or trailing end of the metal plate being rolled material S, the coordinates of the boundary can be extracted from the detected side area by detecting the side area as a "surface."

[0084] From what has been said above, according to the above embodiment of the present disclosure, even if the image contains disturbances such as water or steam, it is less susceptible to these influences, and the boundary position of the side region can be extracted with higher accuracy than in the comparative example.

[0085] <How to generate an extraction model> A method for generating an extraction model for use in executing the extraction step S12 will now be described.

[0086] The learning data used to generate the extraction model is an image G in which a side region GS of a metal plate, which is the rolled material S, is partitioned. First, an image G including at least one of the leading end and the trailing end of the metal plate is captured, and an operator partitions the side region GS of the metal plate, which is the rolled material S, on the captured image. In this case, the operator has knowledge of the equipment that manufactures the metal plate, which is the rolled material S, and therefore understands the actual behavior of warping in the metal plate, which is the rolled material S, based on his or her experience, even if the image G contains disturbances such as water or steam, making it difficult to clearly see the side region GS, and can therefore estimate the side region GS on the image G within the visible range. In this way, images in which the side region GS of the metal plate, which is the rolled material S, is partitioned, can be collected as learning data.

[0087] Furthermore, the learning data used to generate the extraction model may be obtained by capturing an image G including at least one of the leading end and the trailing end of the metal plate, and extracting information about the boundary position GB by applying a binary mask to the captured image G, thereby partitioning the side region GS of the metal plate, which is the rolled material S, on the captured image G. In this case, if the captured image G contains disturbances such as water or steam, and the information about the boundary position GB cannot be extracted accurately, the side region GS of the metal plate, which is the rolled material S, may be partitioned by an operator adding information about the boundary position GB on the image G.

[0088] Because the thickness of the metal plate (rolled material S) is generally constant, the side region GS of the metal plate (rolled material S) does not have complex shapes such as "people" or "animals," as is the case with commonly used image recognition technologies. Therefore, the training data used to generate the extraction model may be relatively small, ranging from 500 to 2,000 pieces of training data. However, it is preferable to use multiple training data that satisfy the following conditions: the image G used as training data includes images with different brightness or color tone due to changes in the temperature of the metal plate (rolled material S), images captured with cooling water splashing, and images in which the side region SF is blurred due to steam present around the metal plate (rolled material S). Therefore, a new image G with a different color tone or brightness may be generated from the image G in which the side region GS of the metal plate (rolled material S) is defined, and this new image may be included in the training data. Furthermore, a new image G may be generated by blurring an image G in which the side region GS of the metal plate, which is the rolled material S, is partitioned, and this image G may be included in the learning data. Blurring can simulate disturbances such as water or steam, improving the accuracy of extracting the side region GS even for images that include these disturbances.

[0089] Regarding the learning method of the extraction model, there are no particular restrictions on the values ​​of its hyperparameters, and a general learning method may be used. For example, the maximum number of iterations when executing learning may be set to about 4000, momentum, which is a parameter for controlling the step size to minimize the loss function and means the inertia term of the adjustment parameters, may be set to about 0.9, the learning rate may be set to about 0.01, and weight decay, which is used to suppress overfitting, may be set to about 0.0005.

[0090] In this way, an extraction model that can be used to execute the extraction step S12 can be generated.

[0091] <Method for controlling warpage of metal plate and manufacturing method> In one embodiment of the method for controlling warpage of a metal plate according to the present disclosure, an image G including at least one of the leading end and the trailing end of the metal plate being rolled material S is captured at the exit side of a rolling mill that rolls the metal plate being rolled material S, the warpage of the metal plate being rolled material S is measured based on the image G captured using a warpage measuring device 30, and operating conditions for another rolling mill located downstream of the above rolling mill are set based on the measured warpage.

[0092] The above embodiment will be described with reference to Fig. 13. Fig. 13 schematically shows two rolling mills 60A and 60B arranged in series. In this case, the warpage measuring device 30 acquires an image G including at least one of the leading end and the trailing end of the metal plate, which is the material S to be rolled, captured by the imaging device 21 arranged between the rolling mills 60A and 60B. In Fig. 13, the imaging device 21 is arranged so as to acquire the image G including the leading end LE of the metal plate, which is the material S to be rolled. When the warpage measuring device 30 acquires an image including the leading end LE of the metal plate, which is the material S to be rolled, such as the image G shown in Fig. 7, the extraction unit 34 and the identification unit 35 perform processing to identify the warpage of the metal plate, which is the material S to be rolled.

[0093] In this embodiment, the system is configured so that the measured value of the warp of the metal plate, which is the identified material S to be rolled, is output to the control computer 101. Then, the control computer 101 compares the measured value of the warp of the metal plate, which is the material S to be rolled, with a preset threshold value, and determines whether the measured value of the warp is equal to or less than the threshold value or exceeds the threshold value. If the result of the determination is that the measured value of the warp of the metal plate, which is the material S to be rolled, is equal to or less than the threshold value, the operating conditions of another rolling mill 60B arranged downstream of the rolling mill 60A are maintained as those set in advance by the control computer 101, and the metal plate, which is the material S to be rolled, is rolled in the rolling mill 60B. On the other hand, if the measured value of warpage exceeds the threshold, the operating conditions of another rolling mill 60B arranged downstream of the rolling mill 60A are changed from those previously set by the control computer 101, and new operating conditions are set so as to reduce the warpage of the metal plate, which is the material S, at the delivery side of the rolling mill 60B, and the metal plate, which is the material S, is rolled in the rolling mill 60B. Here, the operating conditions of the rolling mill 60B refer to conditions that affect the warpage of the metal plate, which is the material S, observed at the delivery side of the rolling mill 60B when the metal plate, which is the material S, is rolled by the rolling mill 60B. This makes it possible to prevent excessive warpage of the metal plate, which is the material S, at the delivery side of the rolling mill 60B, and to prevent operational problems due to warpage of the metal plate. The preset threshold may be determined based on data from past operational performance and the frequency of operational problems caused by warpage when producing the metal plate, which is the material S.

[0094] The method for controlling warpage of a metal plate according to this embodiment is effective when warpage that has occurred in the preceding metal plate, which is the material S to be rolled, is mainly due to factors that induce warpage in the metal plate, which is the material S to be rolled, such as asymmetry in the temperature distribution of the metal plate, which is the material S to be rolled, or a difference in the coefficient of friction between the front and back sides of the metal plate, which is the material S to be rolled. That is, based on the measured value of warpage when the metal plate, which is the material S to be rolled, is rolled by the upstream rolling mill 60A, warpage can be reduced when the metal plate, which is the material S to be rolled, is rolled by the downstream rolling mill 60B.

[0095] In this case, it is preferable to set at least one of the operating conditions of the other rolling mill 60B, such as the water-cooling conditions of a cooling device, such as a strip cooler SC, arranged on the inlet side of the rolling mill 60B, the difference in peripheral speed between the upper and lower work rolls of the rolling mill 60B, the pick-up amount of the rolling mill 60B, or the shape ratio of the rolling mill 60B. By setting these, it is possible to reduce warpage of the metal plate, which is the material S to be rolled, at the outlet side of the rolling mill 60B. As the water-cooling conditions of the water-cooling device arranged between the rolling mills 60A and 60B, the ratio of the flow rate of the cooling water sprays installed above the metal plate, which is the material S, to the flow rate of the cooling water sprays installed below the metal plate, which is the material S, is changed. This causes a temperature difference between the top and bottom surfaces of the metal plate, which is the material S, to be rolled. Therefore, rolling in the rolling mill 60B changes the warpage of the metal plate, which is the material S, at the outlet side of the rolling mill 60B due to the temperature difference between the top and bottom surfaces of the metal plate, which is the material S. This makes it possible to reduce warpage of the metal plate, which is the material S to be rolled, at the exit side of the rolling mill 60B. By setting a difference in peripheral speed between the upper and lower work rolls of the rolling mill 60B, warpage of the metal plate, which is the material S to be rolled, is more likely to occur on the side of the rolling mill 60B where the peripheral speed of the work rolls is slower, and this can be used to reduce warpage of the metal plate, which is the material S to be rolled, at the exit side of the rolling mill 60B. By setting the pickup amount in the rolling mill 60B, the angle of incidence of the metal plate, which is the material S to be rolled, loaded between the work rolls of the rolling mill 60B changes. This changes the position where the upper and lower surfaces of the metal plate, which is the material S to be rolled, come into contact with the work rolls, and changes the warpage of the metal plate, which is the material S to be rolled, at the exit side of the rolling mill 60B. This makes it possible to reduce warpage of the metal plate, which is the material S to be rolled, at the exit side of the rolling mill 60B. The shape ratio in the rolling mill 60B means the ratio of the contact arc length to the average plate thickness in the roll bite when the metal plate, which is the material S to be rolled, is rolled by the rolling mill 60B. If the shape ratio changes, the warpage of the metal plate, which is the material S to be rolled, at the exit side of the rolling mill 60B changes. In this case, the shape ratio can be changed by changing the reduction rate of the metal plate, which is the material S to be rolled in the rolling mill 60B, and thereby the warpage of the metal plate, which is the material S to be rolled, at the exit side of the rolling mill 60B can be changed.

[0096] In another embodiment of the method for controlling warpage of a metal plate according to the present disclosure, an image G including at least one of the leading end and the trailing end of the metal plate being the rolled material S is captured at the exit side of a rolling mill that rolls the metal plate being the rolled material S, the warpage of the metal plate being the rolled material S is measured based on the image G captured using a warpage measuring device 30, and the operating conditions of the rolling mill for other metal plates that follow the metal plate being the rolled material S are set based on the measured warpage.

[0097] The above embodiment will be described with reference to FIG. 14. FIG. 14 schematically illustrates a situation in which a metal plate, which is a rolled material S2, is about to be rolled by one rolling mill 60A following a metal plate, which is a rolled material S1. In this case, the warp measuring device 30 acquires an image including at least one of the leading end and the trailing end of the metal plate, which is the rolled material S1, captured by an imaging device 21 disposed downstream of the rolling mill 60A. In FIG. 14, the imaging device 21 is disposed so as to acquire an image G including the leading end portion LE of the metal plate, which is the rolled material S1, such as the image G exemplified in FIG. 7. After acquiring the image G including the leading end portion LE of the metal plate, which is the rolled material S1, the warp measuring device 30 identifies the warp of the metal plate, which is the rolled material S1, through processing by the extraction unit 34 and the identification unit 35.

[0098] In this embodiment, the system is configured so that the measured value of the warp of the metal plate that is the identified rolled material S1 is output to the control computer 101. Then, the control computer 101 compares the measured value of the warp of the metal plate that is the rolled material S1 with a preset threshold value and determines whether the measured value of the warp is equal to or less than the threshold value or exceeds the threshold value. If the result of the determination is that the measured value of the warp of the metal plate that is the rolled material S1 is equal to or less than the threshold value, the operating conditions of the rolling mill 60A for the metal plate that is the rolled material S2 to be rolled next in the rolling mill 60A are maintained as those previously set by the control computer 101, and the metal plate that is the rolled material S2 is rolled in the rolling mill 60A. On the other hand, if the measured value of warpage exceeds the threshold value, the operating conditions of the rolling mill 60A for the metal plate, which is the material S2 to be rolled next in the rolling mill 60A, are changed from those previously set by the control computer 101, and new operating conditions are set so as to reduce the warpage of the metal plate, which is the material S2 to be rolled, at the delivery side of the rolling mill 60A, and the metal plate, which is the material S2 to be rolled, is rolled in the rolling mill 60A. Here, the operating conditions of the rolling mill 60A refer to conditions that affect the warpage of the metal plate, which is the material S2 to be rolled, observed at the delivery side of the rolling mill 60A when the metal plate, which is the material S2 to be rolled, is rolled by the rolling mill 60A. This makes it possible to prevent excessive warpage of the metal plate, which is the material S2 to be rolled, at the delivery side of the rolling mill 60A, and to prevent operational problems due to warpage of the metal plate. The preset threshold value may be determined based on the frequency of occurrence of operational troubles caused by warpage when manufacturing the metal plate, which is the material S to be rolled, based on data of past operational results.

[0099] The method for controlling warpage of a metal plate according to this embodiment is effective for reducing warpage of a metal plate that is a preceding material to be rolled S1, which is mainly caused by equipment characteristics such as asymmetry of the rolling mill 60A. That is, based on the measured value of warpage when the metal plate that is the material to be rolled S1 is rolled by the rolling mill 60A, it is possible to reduce warpage when the metal plate that is the succeeding material to be rolled S2 is rolled by the rolling mill 60A.

[0100] In this case, for the same reasons as above, it is preferable to set at least one of the following operating conditions for the rolling mill 60A for the metal plate to be rolled S2: the water cooling conditions of a water cooling device such as a strip cooling device SC arranged on the inlet side of the rolling mill 60A, the difference in peripheral speed between the upper and lower work rolls of the rolling mill 60A, the pick-up amount in the rolling mill 60A, or the shape ratio in the rolling mill 60A, and roll the metal plate to be rolled S2.

[0101] The metal plate manufacturing method according to the present disclosure manufactures a metal plate using the above-described method for controlling warpage of a metal plate. As a result, when a metal plate serving as a rolled material S is rolled using successive rolling mills 60A and 60B, warpage that occurs in the downstream rolling mill 60B can be reduced. Furthermore, when a metal plate serving as a rolled material S2 is rolled following a metal plate serving as a rolled material S1, warpage of the metal plate serving as the rolled material S2 can be reduced when the subsequent metal plate serving as the rolled material S2 is rolled by the rolling mill 60A. [Example]

[0102] Example 1 As Example 1, an example will be described in which a camera for measuring the warpage of a metal plate is placed between the first stand F1 and the second stand F2 in the finishing rolling mill 15 (see Figure 3) of the hot rolling line 10, and the warpage of the metal plate passing through the finishing rolling mill is measured and controlled.

[0103] In Example 1, an image G including the leading end portion LE of the metal plate, which is the material S to be rolled, was captured on the exit side of the first stand F1. That is, the leading end portion LE of the metal plate, which is the material S to be rolled, was located between the first stand F1 and the second stand F2, and image G was captured before it was bitten into the second stand F2. The camera that captured image G was an area camera that captures images in the visible light range.

[0104] In Example 1, an extraction model was generated in advance. Detectron, an open-source platform for object detection and segmentation developed by Meta, was used as the extraction model. In this case, 1,200 images including the leading edge of the metal plate were acquired at the exit side of the first stand F1, and images in which the side region was demarcated on the acquired images were prepared as learning data. In this case, the side region of the metal plate (rolled material S) is the region captured on the front side of the camera. The influence of surrounding water, steam, and fumes is relatively reduced in the front region, making it easier to demarcate the side region on the image. The extraction model was then configured to detect the side region GS of the metal plate (rolled material S) and extract a coordinate set Pi representing the boundary position GB of the detected side region GS.

[0105] In the first embodiment, in addition to the extraction model configured as a machine learning model, an identification unit configured as a mathematical model is provided. The identification unit is a program that identifies warpage of the metal plate, which is the rolled material S, based on the coordinate group Pi that represents the boundary position GB extracted by the extraction model. The identification unit rearranges indexes so that the extracted coordinate group Pi (i = 1 to N) is arranged clockwise on the image G. The identification unit is configured to identify the coordinate group CV2 that represents the shape of the bottom surface of the metal plate, which is the rolled material S, calculate the difference between the maximum and minimum Y coordinate values ​​of the coordinate group CV2, and identify this as warpage.

[0106] Next, the extraction model generated by machine learning and the identifying section configured by the mathematical model were installed in the warp measuring device 30 so that they could be executed as a program by a computer. Furthermore, the warp measuring device 30 was connected to a control computer 101, and the measured value of the warp of the metal plate, which is the rolled material S, was output from the output section 36 of the warp measuring device 30 to the control computer 101. The control computer 101 then acquired the measured value W (mm) of the warp at the tip end of the metal plate, which is the rolled material S, measured at the outlet side of the first stand F1, and compared it with a preset threshold value T (mm) to determine the warp of the metal plate, which is the rolled material S. If the measured value W exceeds the threshold value T as a result of the determination, the control computer 101 is configured to set the operating conditions of the first stand F1 for the metal plate, which is the rolled material S2, which is the subsequent metal plate, which is the rolled material S. Specifically, the water cooling conditions of the strip cooling device SC0 installed as a water cooling device on the inlet side of the first stand F1 were set as the operating conditions of the first stand F1.

[0107] The water cooling conditions of the water cooling device 16 installed on the inlet side of the first stand F1 are as follows: the amount of cooling water for cooling the top and bottom surfaces of the metal plate, which is the material S to be rolled, at the inlet side of the metal plate, which is the material S to be rolled and has a specified warpage measurement value W, is α(m 3 / s), β(m 3 / s), the new set values ​​α' and β' of the cooling water amounts for cooling the top and bottom surfaces of the metal plate, which is the subsequent rolled material S2, are set according to the following equations (1) and (2). α'=α×{1-γ1(WT)} (1) β'=min(β×{1+γ2(WT)},βs) (2) where min(·) is a function that outputs the minimum value among multiple arguments. βs is the upper limit for the set value β' of the cooling water flow rate. γ1 (1 / mm) and γ2 (1 / mm) are parameters for adjustment, and are set in advance as γ1 = 0.2 × 10 -2 (1 / mm), γ2=2.4×10 -2(1 / mm). Normally, even if the same amount of cooling water is supplied to the top and bottom surfaces of the metal plate being rolled (S), the temperature of the top surface of the metal plate being rolled (S) is more likely to drop than the bottom surface due to the effect of water riding. As a result, when the metal plate being rolled (S) is rolled, the temperature of the bottom surface becomes relatively higher, making it more likely that the metal plate being rolled (S) will camber upward. Therefore, by adjusting the amount of cooling water cooling the top surface of the metal plate being rolled (S) so that it is lower than the amount of cooling water cooling the bottom surface, the effect of reducing the camber of the metal plate can be increased.

[0108] On the other hand, as a first comparative example to the present disclosure, an example in which the warpage of a metal plate is measured based on the brightness of an image captured by a camera will be described. FIG. 15 is a diagram illustrating a method for identifying warpage when measuring the warpage of a metal plate using the first comparative example. In the first comparative example, a set of edge coordinates of the metal plate is first calculated from the brightness of the image captured by the camera. For example, an arbitrary coordinate X in the x-coordinate is selected on the image, and the brightness of the image is scanned from the bottom to the top of the y-coordinate, i.e., in the positive direction of the y-axis, on that x-coordinate. The first y-coordinate where the brightness is equal to or greater than a preset threshold is determined to be the Y coordinate of the boundary position. This identifies the X and Y coordinates of the boundary position. Then, the same process is performed by changing the coordinate X. Note that if the difference in y-coordinate values ​​between adjacent x-coordinate positions is greater than a preset threshold, it is determined to be an edge, such as the tip of the metal plate, and is excluded from the data for identifying warpage. From the coordinates (X, Y) extracted in this way, the difference between the maximum and minimum values ​​of the Y coordinate is identified as the warpage of the metal plate.

[0109] In this way, the warpage of the metal plate was measured by Example 1 and Comparative Example 1. Figures 16A and 16B show the results of verifying the difference in accuracy between warpage detection by Comparative Example 1 and Example 1, using the same image G.

[0110] In the first comparative example, it can be seen that the extraction results of the boundary position on the lower surface side of the side portion SF of the metal plate, which is the material S to be rolled, represented by the thick solid line in Fig. 16A, and the boundary position on the upper surface US side of the side portion SF of the metal plate, which is the material S to be rolled, represented by the thick dashed line in Fig. 16A, have a large error with respect to the side portion of the metal plate that can be recognized visually. This is because the area around the rolling stand is an environment where water, steam, etc. cause disturbances, and the first comparative example uses a method that directly detects the coordinates of the plate end, so the extraction accuracy of the boundary position of the metal plate is degraded.

[0111] In contrast to this, in Example 1, as shown as the side area GS in Fig. 16B, the side area GS of the metal plate, which is the rolled material S, is detected as a "surface," so that the side area GS corresponding to the side portion SF of the metal plate can be stably extracted even in an environment where water, steam, etc. are disturbances. By being able to stably extract the side area GS, the boundary position of the side area GS can also be stably extracted.

[0112] FIG. 17 shows the results of controlling the warpage of the metal plate according to the first embodiment.

[0113] "No flow rate adjustment" on the horizontal axis of Figure 17 corresponds to a second comparative example in which, when the rolled material S is continuously processed in the finishing rolling mill 15 illustrated in Figure 3, the flow rate of the strip cooling device SC0 located on the inlet side of the first stand F1 is not adjusted based on the warp of the previously processed metal plate as an operating condition of the first stand F1 for the subsequent rolled material S, which is another metal plate.

[0114] On the other hand, "with flow rate adjustment" on the horizontal axis of the graph in Figure 17 corresponds to Example 1, in which flow rate adjustment was performed on the strip cooling device SC0 located on the inlet side of the first stand F1 for other subsequent metal plates based on the warpage of the metal plate measured by the above-mentioned warpage measurement method.

[0115] The vertical axis of the bar graph in Fig. 17 shows the average of the measured values ​​of the amount of warpage of the 10 continuously rolled metal plates, which are the rolled material S, normalized so that the average value of the amount of warpage at the outlet side of the first stand F1 obtained in the second comparative example is 100%. As can be seen from Fig. 17, it was confirmed that the warpage of the metal plate in Example 1 (with flow rate adjustment) was reduced by an average of 20% compared to the second comparative example (without flow rate adjustment).

[0116] Example 2 As Example 2, an example will be described in which a camera for measuring the warpage of a metal plate is placed between the fourth stand F4 and the fifth stand F5 in the finishing rolling mill 15 (see Figure 3) of the hot rolling line 10, and the warpage of a metal plate passing through the finishing rolling mill is measured.

[0117] In Example 2, image G including the leading end portion LE of the metal plate, which is the material S to be rolled, was captured on the delivery side of the fourth stand F4. That is, image G was captured when the leading end portion LE of the metal plate, which is the material S to be rolled, was between the fourth stand F4 and the fifth stand F5, and before it was bitten into the fifth stand F5. The camera that captured image G was an area camera that captures images in the visible light range.

[0118] In Example 2, the extraction model generated in Example 1 was used, and an identifying unit having the same configuration as in Example 1 was used. That is, the warpage measuring device 30 in Example 1 was configured to input an image G including the leading end portion LE of the metal plate, which is the rolled material S, captured on the delivery side of the fourth stand F4.

[0119] On the other hand, to confirm the measurement accuracy of the warpage of the metal plate in Example 2, a laser rangefinder was installed between the fourth stand F4 and the fifth stand F5. The laser rangefinder irradiated a laser beam from above onto the upper surface of the rolled material S to measure the distance to the rolled material S. Distance measurements using the laser rangefinder were continuously performed over an area including the leading end LE of the metal plate, which was the rolled material S, and the measured value of the warpage of the metal plate (warpage measured by the rangefinder) was calculated from the measured values. The reason for measuring the warpage of the metal plate between the fourth stand F4 and the fifth stand F5 is as follows. First, the area between the upstream stands (for example, between the first stand F1 and the second stand F2) is in an environment where a large amount of steam or cooling water splashes, so the accuracy of distance measurements using the laser rangefinder between the upstream stands is reduced. However, in the downstream stands (stands downstream of the fourth stand F4), operating conditions with relatively little splashing of steam or cooling water can often be achieved, and therefore, relatively reliable distance measurement values ​​can be obtained from the rangefinder.

[0120] In Example 2, a thick steel plate for line pipe was processed as the rolled material S by the finishing mill 15 of the hot rolling line 10. Then, the warpage of the leading end LE of the metal plate, which was the rolled material S, was measured by the warpage measuring device 30, and the warpage measurement value was calculated by a distance meter using a laser distance meter. In manufacturing a thick steel plate for line pipe, controlled rolling is performed upstream of the finishing mill 15 to cool the rolled material S to a predetermined temperature, and finish rolling is performed by the finishing mill 15 in a predetermined temperature range. For this reason, the injection of cooling water from the strip cooling device SC is often stopped in the rear stand. By stopping the injection of cooling water in the rear stand, operating conditions are realized in which there is little scattering of steam or cooling water in the rear stand, and conditions are realized in which the error in the warpage measurement value measured by the distance meter is small. In Example 2, the thickness of the rolled material S at the outlet of the final stand F7 of the finishing rolling mill 15 was 15 mm, the width was 1270 mm, and the thickness between the fourth stand F4 and the fifth stand F5 was 33.1 mm.

[0121] Figure 18 shows an example of image G including the leading end LE of the metal plate, which is the rolled material S, captured using an area camera at the exit side of the fourth stand F4. In image G of Figure 18, hatched, dashed, and rectangular areas have been added to make the side portions of the metal plate easier to recognize. The hatched areas are areas where hatched lines have been added to the top surface of the metal plate to make it easier to distinguish the top surface. The dashed areas indicate the side portions of the metal plate surrounded by dashed lines. The rectangular areas indicate areas where the boundary between the side and top surfaces is unclear. In other words, image G shown in Figure 18 includes areas where the boundary between the side and top surfaces of the metal plate is not clearly visible.

[0122] Fig. 19 shows the side area GS of the metal plate extracted by the extraction model of the specific part in the warpage measurement device 30 when the image G shown in Fig. 18 is input to the warpage measurement device 30. According to Fig. 19, even when the image G includes a portion in which the boundary between the side surface and the top surface of the metal plate is not clearly visible, the side area GS of the metal plate, which is the rolled material S, is detected as a "surface," and it can be seen that the side area GS corresponding to the side surface portion SF of the metal plate can be extracted with high accuracy.

[0123] Next, 50 coils of thick steel plate for line pipe, which has the same size as the thick steel plate for line pulp described above, were manufactured by the finishing mill 15 of the hot rolling line 10, and the measured value CI of the warpage of the metal plate measured by Example 2 was compared with the measured value CL of the warpage measured by using a laser distance meter for each rolled material S. As a result, the average value of the error rate ((CI-CL) / CL×100[%]) was 0.5%, and it was confirmed that the measurement result of the warpage of the metal plate according to the example and the measured value of the warpage measured by using a laser distance meter under operating conditions with little scattering of steam or cooling water were in good agreement.

[0124] Although the embodiments of the present disclosure have been described based on the drawings and examples, it should be noted that those skilled in the art could make various modifications or alterations based on the present disclosure. Therefore, it should be noted that these modifications and alterations are included within the scope of the present disclosure. For example, the functions included in each component or step can be rearranged so as not to cause logical inconsistencies, and multiple components or steps can be combined or divided into one. The embodiments of the present disclosure can also be realized as a program executed by a processor included in an apparatus or a storage medium on which a program is recorded. It should be understood that these are also included within the scope of the present disclosure. [Explanation of symbols]

[0125] 1. Information Processing Systems 10 Hot rolling line (11: heating furnace, 12: descaling device, 13: width reduction device, 14: roughing mill, 14a: reversing rolling mill, 14b: non-reversing rolling mill, 15: finishing rolling mill, 16: water cooling device, 17: coiler, 18: bar cooling device, 23: crop shear) 21 Imaging device (21a, 21b: camera) 30 Warpage measuring device (32: memory unit, 33: acquisition unit, 34: extraction unit, 35: identification unit, 36: output unit) 44 Display device 60, 60A, 60B Rolling mill (61a, 61b: work rolls, 62a, 62b: backup rolls, 63: housing) 100 top calculators 101 Control computer 102 Control Controller D1 Input section D2 Convolution section D3 Area candidate setting section D4 Cutout D5 classification section D6 Segmentation part D7 Mask section G image (GB: border position, GS: side area, LE: tip, US: top surface) M Extraction Model S(S1, S2) Rolled material

Claims

1. an imaging step of capturing an image including at least one of a leading end and a trailing end of the metal plate; an extraction step of detecting a side area corresponding to a side portion of the metal plate from the image and extracting a group of coordinates representing a boundary position of the side area; a step of identifying the warp of the metal plate based on the extracted coordinate group; A method for measuring warpage of a metal plate, comprising:

2. 2. The method for measuring warpage of a metal plate according to claim 1, wherein in the extraction step, an area including a portion in the image where a boundary between a side portion of the metal plate and an upper surface or a lower surface of the metal plate is not captured is detected as the side surface area.

3. The method for measuring warpage of a metal plate according to claim 1 , wherein a part of a boundary of the region detected as the side region in the extraction step is not captured in the image.

4. 2. The method for measuring warpage of a metal plate according to claim 1, wherein in the extraction step, the image is input and the group of coordinates is extracted using a neural network trained to output a group of coordinates representing boundary positions of the side region of the metal plate included in the image.

5. 2. The method for measuring warpage of a metal plate according to claim 1, wherein in the identifying step, a group of coordinates representing at least one of an upper surface and a lower surface of the metal plate is selected from a group of coordinates representing boundary positions of the side surface region, and the warpage of the metal plate is identified based on the selected group of coordinates.

6. 2. The method for measuring warpage of a metal plate according to claim 1, wherein in the imaging step, an image including at least one of a leading end and a trailing end of the metal plate is captured at the delivery side of a rolling mill that rolls the metal plate.

7. 7. A method for controlling warpage of a metal plate, comprising a step of setting operating conditions of another rolling mill arranged downstream of the rolling mill based on the warpage of the metal plate measured using the method for measuring warpage of a metal plate according to claim 6.

8. 8. The method for controlling warpage of a metal plate according to claim 7, wherein the operating conditions of the other rolling mill are at least one of a water-cooling condition of a water-cooling device arranged on an inlet side of the other rolling mill, a difference in peripheral speed between upper and lower work rolls of the other rolling mill, a pick-up amount of the other rolling mill, or a shape ratio of the other rolling mill.

9. A method for controlling warpage of a metal plate, comprising a step of setting operating conditions of the rolling mill for other metal plates following the metal plate measured using the method for measuring warpage of a metal plate according to claim 6.

10. The operating conditions of the rolling mill for the other metal plate are at least one of a water-cooling condition of a water-cooling device arranged on the inlet side of the rolling mill, a peripheral speed difference between upper and lower work rolls of the rolling mill, a pickup amount in the rolling mill, or a shape ratio in the rolling mill. The method for controlling warpage of a metal plate according to claim 9.

11. A method for manufacturing a metal plate, comprising the step of manufacturing a metal plate using the method for controlling warpage of a metal plate according to any one of claims 7 to 10.

12. an acquisition unit that acquires an image including at least one of the leading end and the trailing end of the metal plate; an extraction unit that detects a side area of ​​the metal plate from the image acquired by the acquisition unit and extracts a group of coordinates that represent boundary positions of the detected side area; an identification unit that identifies the warp of the metal plate based on the coordinate group extracted by the extraction unit; A metal plate warpage measuring device comprising:

13. A program for causing a computer to function as a metal plate warpage measuring device, On the computer, a function of inputting an image including at least one of the leading end and the trailing end of the metal plate; a function of detecting a side area of ​​the metal plate from the image and extracting a group of coordinates representing boundary positions of the detected side area; a function of identifying the warp of the metal plate based on the extracted coordinate group; A program to achieve this.

Images