Method for measuring the shape of a strip-like object, method for controlling the shape of a strip-like object, method for manufacturing a strip-like object, quality control method for a strip-like object, device for measuring the shape of a strip-like object, and manufacturing equipment for a strip-like object

Abstract

The shape measurement method for a strip-like object is a shape measurement method for measuring the shape of a strip-like object, and includes an imaging step of capturing an image of thermal radiation light from the strip-like object so that the angle θ between the plane α, which is the reference plane of the surface of the strip-like object, and the optical axis of the camera is not 90 degrees, and the angle φ between the orthogonal projection of the optical axis of the camera onto the plane α and the conveying direction p of the strip-like object is not 0 degrees, and an image processing step of calculating an index of the shape of the edge portion of the strip-like object by calculating a contour profile of the strip-like object from the obtained image.

Description

[Technical field]

[0001] The present invention relates to a method for measuring the shape of a strip-like object, a method for controlling the shape of a strip-like object, a method for manufacturing a strip-like object, a method for quality control of a strip-like object, an apparatus for measuring the shape of a strip-like object, and a manufacturing facility for a strip-like object. [Background technology]

[0002] In the materials industry, shape management of strip materials is important, and quantification of product shapes has been required. For example, there is a high demand for measuring the shape of steel materials in the steelmaking process from the viewpoints of operational stability and product quality assurance. In particular, shape measurement during rolling to produce a product into the desired shape is important because it leads to improved product quality and operational stability by providing feedback for initial setting of rolling conditions and for rolling control during rolling.

[0003] For example, in a hot-rolled steel sheet manufacturing line, a rectangular semi-finished product called a slab, which is in a high-temperature state after being extracted from a heating furnace, is processed into a sheet through a sizing mill, rough rolling, and finish rolling, and then wound up into a coil to be manufactured. During this process, depending on the rolling conditions, the reduction in the width direction of the steel sheet may become uneven, causing partial elongation and resulting in defective shape.

[0004] For example, when only the edge portion (widthwise end portion) of a steel sheet is stretched compared to the center portion (widthwise center portion), the edge portion will have a wavy shape. Conversely, when only the center portion is stretched, the center portion will have a wavy shape. Such shape defects not only result in a defective product, but also deteriorate the sheet passing properties in the next process such as pickling and cold rolling, causing problems, so there is a strong demand to prevent shape defects.

[0005] In order to improve shape defects, it is necessary to appropriately set the load in the width direction during rolling. However, due to various disturbances such as wear of the rolling rolls, temperature distribution of the steel plate, and variations in material property distribution, it has been very difficult to obtain optimal rolling conditions by calculation alone. Therefore, in order to set optimal rolling conditions, it is essential to derive optimal initial settings by comparing measurements of the steel plate shape with a rolling model, and perform preset control or feedback control of the rolling conditions in real time. Only by performing such control can it become possible to suppress shape defects.

[0006] However, it is difficult to measure the shape of a steel sheet after it has been made into a rolled product, and since feedback control requires measurement during rolling production, it is preferable to measure the shape immediately after the finish rolling exit, which is the final rolling process. For example, when the elongation of one of the edges of the steel sheet is significant relative to the center of the steel sheet, leveling control is performed to eliminate the bias of elongation to one of the edges by adjusting the load balance in the width direction of the rolling rolls. The shape described here mainly refers to the longitudinal sheet elongation that occurs locally in the width direction of the steel sheet during rolling, and specifically includes "elongation of the center" and "elongation of the edge".

[0007] As a technique for measuring the shape of a steel sheet immediately after the exit side of finish rolling, various techniques have been proposed in the past, such as a method using a rod-shaped light source or a magnetic sensor. Among these, Patent Documents 1 to 3, for example, disclose a particularly effective technique in which a point-like or linear light beam is irradiated onto a target surface using a laser, and the shape is measured by measuring the reflected light.

[0008] Furthermore, Patent Document 4 discloses a technology that uses three line lasers whose longitudinal direction is perpendicular to the conveying direction of the steel plate as a laser irradiation method. In the technology disclosed in Patent Document 4, a line laser is irradiated in parallel to the longitudinal direction of the steel plate so as to form three equally spaced lines, and the reflected images are acquired and the profiles of each laser are compared to remove the effects of vertical vibration of the steel plate.

[0009] Furthermore, Patent Documents 5 to 7 and Non-Patent Document 1 disclose a technology in which a powerful LED light source is used to project a stripe pattern consisting of multiple lines onto the target surface, thereby increasing the number of lines projected more cheaply than with a laser, thereby enabling stable shape measurement regardless of the specularity or tilt of the target.

[0010] The need for measuring the shape of steel sheets during manufacturing exists not only for hot-rolled steel sheets, but also for other strip-shaped material, regardless of whether it is in a red-hot state (e.g., 600°C or higher), a warm state (e.g., 300°C to 600°C), or a cold state (e.g., near room temperature). Note that the "strip" mentioned here refers to a long material. Strip-shaped objects include products that are ultimately wound into a roll, such as iron, paper, cloth, and non-ferrous metals such as aluminum, as well as products that are formed into rectangular plates, such as thick steel plates. [Prior art documents] [Patent documents]

[0011] [Patent Document 1] Japanese Patent Application Publication No. 56-124006 [Patent Document 2] Japanese Unexamined Patent Publication No. 55-40924 [Patent Document 3] Japanese Patent Application Publication No. 58-11708 [Patent Document 4] Japanese Patent Publication No. 61-40503 [Patent Document 5] JP 2008-58036 A [Patent Document 6] JP 2011-99821 A [Patent Document 7] JP 2016-65863 A [Non-patent literature]

[0012] [Non-Patent Document 1] Ise, et al., “Development of a hot-rolled steel sheet flatness meter using LED dot pattern projection method”, Iron and Steel, The Iron and Steel Institute of Japan, 2019, Vol. 105, No. 1, pp. 20-29 [Non-Patent Document 2] Kaneshige, “Heat Measurement by Radiation Thermometer”, Molding and Processing, Society of Plastics Processing, 2020, Vol. 32, No. 4, pp. 121-124 Summary of the Invention [Problem to be solved by the invention]

[0013] The techniques disclosed in Patent Documents 1 to 7 are all methods for measuring the shape of a target by irradiating light onto a hot-rolled steel sheet using a light source and capturing the reflected light with a camera. However, considering the effects of high temperatures due to radiant heat, and the adhesion of steam, dust, oil, and the like to the steel sheet, for example, advanced technology is required to perform stable measurements over a long period of time with a light source or sensor placed close to the steel sheet during transportation, and maintenance costs are also high.

[0014] One way to prevent this is to move the light source away from the object, but this would result in a decrease in the amount of light due to diffusion, making it difficult to design an optical system that can stably collect light in order to ensure the amount of light. Another possibility is to use a magnetic sensor, but this requires that the sensor itself be brought close to the object being measured, which also makes installation and maintaining performance difficult.

[0015] The present invention has been made in consideration of the above, and aims to provide a method for measuring the shape of a strip-like object, a method for controlling the shape of a strip-like object, a method for manufacturing a strip-like object, a quality control method for a strip-like object, an apparatus for measuring the shape of a strip-like object, and manufacturing equipment for a strip-like object, which can be easily and stably operated for targeting the edge portion of the strip-like object without bringing a light source or sensor close to the object to be measured, and which also reduces maintenance costs. [Means for solving the problem]

[0016] (1) A shape measuring method of a belt-like object according to the present invention is a shape measuring method for measuring the shape of a belt-like object, comprising the steps of: The angle θ between the plane α, which is the reference plane of the surface of the belt-like object, and the optical axis of the camera is not 90 degrees. In addition, the angle φ between the orthogonal projection of the optical axis of the camera onto the plane α and the conveying direction p of the belt-like object is not 0 degrees. an imaging step of capturing an image of thermal radiation light from the belt-shaped object; an image processing step of calculating a contour profile of the belt-like object from the obtained image, thereby calculating an index of the shape of an edge portion of the belt-like object; It includes.

[0017] (2) The present invention also provides a method for measuring the shape of a belt-like object, comprising the steps of: The image processing step includes: Extracting the region of the belt-like object from the obtained image; In the extracted region, the positions of the edge portions are calculated to calculate a contour profile of the belt-like object.

[0018] (3) A method for measuring the shape of a belt-like object according to the present invention is the method for measuring the shape of a belt-like object according to the above (1) or (2), further comprising the steps of: The image processing step calculates one or more of the following indicators from the obtained contour profile of the band-like object: steepness of the edge portion of the band-like object, wave height, wave pitch, elongation amount, and elongation rate.

[0019] (4) A method for measuring the shape of a belt-like object according to the present invention is the method for measuring the shape of a belt-like object according to any one of (1) to (3) above, further comprising: The image processing step calculates the resolution in the wave height direction and the wave pitch direction from the positional relationship between the camera and the band-like object, and converts the resolution into actual dimensions.

[0020] (5) Furthermore, the method for controlling the shape of a strip-like object according to the present invention measures the shape of the strip-like object using a method for measuring the shape of a strip-like object described in any one of (1) to (4) above, and controls the shape of the strip-like object to a desired shape based on the measurement results.

[0021] (6) Furthermore, the method for manufacturing a strip-shaped object of the present invention involves measuring the shape of the strip-shaped object using a method for measuring the shape of a strip-shaped object described in any one of (1) to (4) above, and manufacturing the strip-shaped object based on the measurement results.

[0022] (7) Furthermore, the quality control method for a strip-like object of the present invention measures the shape of the strip-like object using the shape measurement method for a strip-like object described in any one of (1) to (4) above, and controls the quality of the strip-like object based on the measurement results.

[0023] (8) A shape measuring device for a belt-like object according to the present invention is a shape measuring device for measuring the shape of a belt-like object, comprising: The angle θ between the plane α, which is the reference plane of the surface of the belt-like object, and the optical axis of the camera is not 90 degrees. In addition, the angle φ between the orthogonal projection of the optical axis of the camera onto the plane α and the conveying direction p of the belt-like object is not 0 degrees. an imaging means for capturing an image of thermal radiation light from the belt-shaped object; an image processing means for calculating a contour profile of the belt-like object from the obtained image, thereby calculating an index of the shape of an edge portion of the belt-like object; It is equipped with the following.

[0024] (9) Furthermore, a manufacturing facility for a belt-like object according to the present invention includes the shape measuring device for a belt-like object described above in (8). Effect of the Invention

[0025] The shape measurement method for a strip-like object, the shape control method for a strip-like object, the manufacturing method for a strip-like object, the quality control method for a strip-like object, the shape measurement device for a strip-like object, and the manufacturing equipment for a strip-like object according to the present invention can be easily and stably operated without bringing a light source or sensor close to the object to be measured, and maintenance costs can also be reduced. [Brief description of the drawings]

[0026] [Figure 1] FIG. 1 is a diagram showing a schematic configuration of an apparatus for measuring the shape of a belt-shaped object according to an embodiment of the present invention. [Diagram 2] FIG. 2 is a diagram showing an example of the positional relationship between the conveying direction of the steel plate and the camera, where (a) is a diagram showing the positional relationship viewed from an oblique angle, (b) is a diagram showing the positional relationship viewed from above, and (c) is a diagram showing the positional relationship viewed from the direction δ of (b). [Diagram 3] FIG. 3 is a diagram showing an example of the positional relationship between the conveying direction of the steel plate and the camera, in which the optical axis of the camera is disposed perpendicular to the conveying direction of the steel plate. [Figure 4] FIG. 4 is a diagram showing an example of the positional relationship between the conveying direction of the steel plate and the camera, in which the optical axis of the camera is disposed obliquely with respect to the conveying direction of the steel plate. [Diagram 5] FIG. 5 shows examples of images of a steel sheet when it is in a normal state and when it has a defective shape. [Figure 6] FIG. 6 is a flow chart showing a specific process flow of the image processing steps performed by the image processing device of the apparatus for measuring the shape of a belt-shaped object according to the embodiment of the present invention. [Figure 7] FIG. 7 is a diagram for explaining the binarization process in the image processing step of the method for measuring the shape of a band-shaped object according to the embodiment of the present invention. [Figure 8] FIG. 8 is a diagram for explaining a contour profile calculation process in the image processing step of the method for measuring the shape of a band-shaped object according to an embodiment of the present invention. [Figure 9] Figure 9 shows an example of a photograph of waving that has occurred on a steel plate. [Figure 10] Figure 10 shows an example of a stationary wave generated in a steel plate. [Figure 11] FIG. 11 is a diagram showing an example of a case where it is difficult to distinguish between the plate surface of a steel plate and a structure. [Figure 12] FIG. 12 is a diagram showing an example of processing using a low-pass filter that is performed when it is difficult to distinguish between the plate surface of a steel plate and a structure in the method for measuring the shape of a belt-shaped object according to an embodiment of the present invention. [Figure 13] FIG. 13 shows an example in which the apparatus for measuring the shape of a belt-like object according to the embodiment of the present invention is applied to determining whether or not the shape is acceptable. [Figure 14] FIG. 14 shows an example in which the apparatus for measuring the shape of a strip-like object according to the embodiment of the present invention is applied to rolling feedback control. [Figure 15] FIG. 15 shows an example in which the shape measuring device for a strip-shaped object according to the embodiment of the present invention is applied to rolling control using machine learning. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0027] A method for measuring the shape of a strip-like object, a method for controlling the shape of a strip-like object, a method for manufacturing a strip-like object, a quality control method for a strip-like object, a shape measuring device for a strip-like object, and a manufacturing facility for a strip-like object according to embodiments of the present invention will be described with reference to the drawings. Note that the components in the following embodiments include those that are replaceable and easy for a person skilled in the art, or those that are substantially the same.

[0028] (shape measuring device) A shape measuring device for a strip-like object according to an embodiment will be described with reference to Figs. 1 to 12. The shape measuring device is a device for measuring the shape of a strip-like object. In the following, a case where the shape measuring device is applied to hot finish rolling will be described. Also, in the following, a case where the strip-like object to be measured is a steel plate will be described. Also, in the following, a case where the shape measured by the shape measuring device is the elongation of an edge portion of a steel plate will be described.

[0029] 1, the shape measuring device according to the embodiment includes a camera 2 and an image processing device 3. First, the camera 2 will be described in detail.

[0030] As described later, the camera 2 is positioned so that the angle θ between the "optical axis of the camera 2" and the "plane (plane α) that is the reference plane of the strip-like object (steel sheet S)" is not 90 degrees, and the angle φ between the "orthogonal projection of the optical axis of the camera 2 onto the plane α" and the "conveying direction p of the strip-like object (steel sheet S)" is not 0 degrees (see FIG. 2). Then, using the camera 2 positioned in this way, an image of thermal radiation light of the steel sheet S on the hot finish rolling exit side that has been rolled by the rolling rolls 1 is captured.

[0031] If a commercially available digital camera or the like is used as the camera 2, problems such as blurring of the image or graininess occur, making it impossible to clearly capture the shape of the steel sheet S. Therefore, technical points (1) to (4) for accurately capturing the shape of the steel sheet S with the camera 2 will be described below.

[0032] (1) Ensuring sufficient light for imaging First, it is necessary to ensure the amount of light necessary to obtain a clear image. At the hot finishing rolling exit, the steel sheet S may pass through at a speed of 20 m / s or more, which is extremely fast. To capture a fast-moving object clearly and without blur, it is preferable to shorten the exposure time. For example, when capturing an image at 20 m / s with a resolution of about 2 mm and pixel blur of 1 pixel, i.e., within 2 mm, the allowable exposure time is only 0.1 ms, and the amount of light received, which is proportional to the exposure time, is also very small.

[0033] The temperature of the steel sheet at the hot finish rolling exit side is about 900°C. As shown in Fig. 5 of the above Non-Patent Document 2, the peak wavelength of thermal radiation light at about 900°C (1200K) is 2.5μm, and the sensitivity is low in the visible region (0.4 to 0.7μm). Therefore, it is preferable to use a camera 2 equipped with an imaging element such as InGaAs, PbS, or PbSe.

[0034] On the other hand, cameras using imaging elements such as InGaAs, PbS, PbSe, etc. are expensive, and the price increases depending on the resolution, so the introduction cost is high when trying to capture images with high resolution. Therefore, it is preferable to use a camera 2 equipped with an inexpensive Si imaging element and capable of using the near-infrared sensitivity region of 0.8 to 1.0 μm. By using such a camera 2, it is possible to obtain a sufficient amount of thermal radiation light at a low cost.

[0035] In addition, when applying this method to a lower temperature object, it is preferable to use a camera 2 equipped with an imaging element having sensitivity on the longer wavelength side. For example, when the measurement object is about 400°C, if the introduction cost is not considered, a camera 2 equipped with an InGaAs imaging element having sensitivity at 1.2 to 1.7 μm can be used. In addition, when the measurement object is about 200°C, a camera 2 equipped with an imaging element such as PbS or PbSe having sensitivity at 3 to 5 μm can be used. This makes it possible to ensure a sufficient amount of light. In this way, it is preferable to select a camera 2 equipped with an imaging element having sensitivity in an appropriate wavelength band depending on the transport speed, resolution, temperature, and depth of field described later of the measurement object.

[0036] (2) Positional relationship between the steel plate and the camera Secondly, in order to clearly capture the shape of the steel sheet S, the positional relationship between the steel sheet S to be measured and the camera 2 is considered. Here, attention is focused on the edge elongation of the steel sheet S, and consideration is given to capturing the change in shape of the edge portion as the outline of the steel sheet S. An example of the positional relationship between the conveying direction p of the steel sheet S and the camera 2 is shown in Fig. 2. In Fig. 2, (a) is a diagram of the positional relationship seen from an oblique angle, (b) is a diagram of the positional relationship seen from above, and (c) is a diagram of the positional relationship seen from the direction δ in (b). n Here, the view is taken from the direction δ n is the direction in which the optical axis of the camera 2 can be seen from the side. n By defining the above equation, the magnitude of the light-receiving angle θ, which will be described later, can be seen accurately.

[0037] 2(a), (b), and (c), a plane parallel to the conveying table for the steel sheet S is designated as α, and a plane including the conveying direction p of the steel sheet S and the normal n of the conveying table is designated as β. Furthermore, the angle between the optical axis of the camera 2 and the plane α is designated as θ (light-receiving angle θ), and the angle between the orthogonal projection of the optical axis of the camera 2 onto the plane α and the conveying direction p is designated as φ. The plane α includes the surface of the belt-like object (steel sheet S) when the belt-like object (steel sheet S) is stably conveyed in a substantially flat shape. Therefore, the plane α is also called the reference plane of the surface of the belt-like object (steel sheet S).

[0038] The direction of the normal n of the conveying table is the same as the normal direction of the surface of the steel sheet S when the steel sheet S is stably conveyed in a substantially flat shape. Therefore, the plane α includes the conveying direction p of the belt-like object (steel sheet S) and is perpendicular to the normal direction of the surface of the belt-like object (steel sheet S). The plane β also includes the conveying direction p of the belt-like object (steel sheet S) and the normal direction of the surface of the belt-like object (steel sheet S). In FIG. 2, the intersection of the plane α and the optical axis is defined as point 0. The plane perpendicular to the conveying direction of the steel sheet is defined as plane γ. In other words, as the angle φ approaches 90 degrees, the plane γ and the optical axis become closer to being parallel.

[0039] In Fig. 2, if the resolution of the camera 2 at the position of the steel plate S to be measured is "r (mm / pixel)", the resolution rn (mm / pixel) in the wave height direction and the resolution rp (mm / pixel) in the wave pitch direction can be expressed by the following formulas (1) and (2). Note that the smaller the numerical value of the resolution, the higher the resolution.

[0040] rn=r / cosθ (1) rp = r / sinφ (2)

[0041] When the receiving angle θ is close to 90 degrees, i.e., when the optical axis of the camera 2 is nearly perpendicular to the plane α parallel to the conveying table of the steel sheet S, the resolution rn in the wave height direction decreases, making it difficult to capture the change in the contour of the edge portion. Therefore, the closer the receiving angle θ is to 0 degrees, the better. Similarly, when the angle φ is close to 0 degrees, i.e., when the optical axis of the camera 2 is nearly parallel to the conveying direction p of the steel sheet S, the resolution rp in the wave pitch direction decreases. Therefore, the closer the angle φ is to 90 degrees, the better. In particular, since the wave height to be measured is very small compared to the wave pitch, in order to capture the wave height with high accuracy, it is more preferable to make the receiving angle θ as small as possible and to image the steel sheet S at a low angle.

[0042] Furthermore, by capturing an image of the steel sheet S at a low angle relative to the light receiving angle θ as shown in Figure 2, the distance between both ends of the steel sheet S in the image becomes smaller, making it possible to capture images of both edges in a smaller field of view. In addition, a smaller field of view also results in a smaller image size, which is advantageous when the transport speed of the steel sheet S to be measured is high and high-speed imaging and image processing are required.

[0043] In addition, since the wave pitch is sufficiently larger than the wave height in a hot-rolled steel sheet, the resolution rp in the wave pitch direction may be smaller than the resolution rn in the wave height direction. Therefore, for example, as shown in Fig. 3, it is preferable to set the angle φ to 90 degrees, that is, to place the camera 2 perpendicular to the conveying direction of the steel sheet S and look in from the side. However, if the angle φ is not extremely small, the camera 2 may be placed at an angle to the conveying direction of the steel sheet S, for example, as shown in Fig. 4.

[0044] (3) Distance between the steel plate and the camera Thirdly, the distance between the steel sheet S to be measured and the camera 2 is increased. The reason for this is that by placing the camera 2 as far away as possible from the production line, it is possible to install it in a good environment and to reduce the difference in optical conditions such as the light receiving angle θ and the angle φ in the width direction of the steel sheet S, i.e., at both edges. If there is a difference in optical conditions in the width direction of the steel sheet S, it will appear as a difference in appearance. Therefore, when judging by visual inspection, there is a risk of misjudging the degree of shape defect, and when quantifying the degree of shape defect by image processing, a larger correction is required. To increase the distance, it is preferable to use a telephoto lens when capturing the image. By using a telephoto lens, the difference in optical conditions can be reduced.

[0045] (4) Setting the correct aperture Fourth, to ensure the depth of field, the aperture value of the camera 2 is appropriately set. For example, assume that an image is to be captured of the entire surface of the steel plate S, including both the front (lower) and rear (upper) edge portions. In this case, if the width of the steel plate S is d (mm), the difference ΔL (mm) between the distance from the camera 2 to the front edge portion and the distance from the camera 2 to the rear edge portion can be expressed by the following formula (3).

[0046] ΔL = dcosθ / sinφ (3)

[0047] To obtain an image in focus on both edges of the steel sheet S, a depth of field (range in focus) of at least ΔLd (mm) is required. Since the greater the aperture value, the greater the depth of field, it is preferable to set the aperture to ΔLd (mm) or more. However, if the aperture value is increased, the amount of light will be insufficient, and there is a concern that blurring will occur. Therefore, it is preferable to determine the aperture value and focal length that can achieve both blurring and depth of field depending on the conveying speed of the steel sheet S to be measured and the required resolution. Next, the image processing device 3 will be described.

[0048] The image processing device 3 is realized by, for example, a general-purpose computer such as a workstation or a personal computer. The image processing device 3 may be installed near the camera 2, or may be installed on the cloud if high speed is not required.

[0049] The image processing device 3, as described later, calculates a contour profile of the steel sheet S from an image captured by the camera 2, thereby calculating an index of the shape of the edge portion of the steel sheet S. Also, as described later, the image processing device 3 extracts an area of ​​the steel sheet S from the image (step S1 in FIG. 6), and calculates the position of the edge portion in the extracted area, thereby calculating a contour profile of the strip-shaped object of the steel sheet S (step S4 in FIG. 6). Also, as described later, the image processing device 3 appropriately selects one or more of the steepness, wave height, wave pitch, elongation amount, and elongation rate of the edge portion of the steel sheet S from the contour profile of the steel sheet S, and calculates them as the above indexes. Below, a process of calculating an index of the shape of the edge portion of the steel sheet S from an image captured by the camera 2 using the image processing device 3 will be described.

[0050] Fig. 5(a) shows an image of the steel sheet S under normal conditions, and Fig. 5(b) shows an image of the steel sheet S when a shape defect has occurred due to sheet elongation. As shown in Fig. 5, the state of sheet elongation (particularly elongation of the edge portion of the strip-shaped object) can be determined visually. The image processing device 3 calculates an index of the edge portion of the steel sheet S from the acquired image of the steel sheet S, for example, by the procedure shown in Fig. 6. Note that, for ease of explanation, in the following, with regard to the vertical and horizontal axes of the image, the axis closer to being parallel to the conveying direction of the steel sheet S is taken as the horizontal axis.

[0051] First, as shown in Fig. 7, only the plate region of the steel plate S is extracted by binarization processing (step S1 in Fig. 6). Fig. 7(a) is an image before binarization processing, and Fig. 7(b) is an image after binarization processing.

[0052] At this time, disturbances may occur, such as water riding on the plate surface or splashes of cooling water causing localized reductions in brightness on the plate surface, or splashes of water in the space above the plate scattering the thermal radiation light from the steel plate S and causing it to shine brightly. Since these disturbances are often very small and in the high frequency range relative to the pitch of the shape to be calculated, it is desirable to carry out expansion / contraction processing or removal of connected and isolated points by a median filter, etc. (step S2 in Fig. 6) on the plate region extracted in the binarization processing step.

[0053] Furthermore, after removing connected and isolated points, multiple blobs (lumps recognized by connecting surrounding pixels during binarization) that are candidates for plate regions may be generated. In this case, the plate region may be determined by extracting the blobs (step S3 in FIG. 6) or by determining whether or not it is a plate region based on the size, orientation, etc. of the steel plate S.

[0054] Next, the contour profiles of both edge portions are calculated from the plate region of the steel plate S obtained as described above (step S4 in FIG. 6). There are various methods for calculating the contour profile of the edge portion, but as an example, a method for extracting the contour of the edge portion by searching is shown in FIG. 8. FIG. 8(a) is a diagram showing how the contour of the edge portion is searched, and FIG. 8(b) is a diagram showing the contour of the extracted edge portion.

[0055] As shown in (a) of Fig. 8, a search may be performed from outside the plate region in the image toward the plate region and the coordinates of the plate region are recorded, or conversely, a search may be performed from inside the plate region to outside the plate region. In this embodiment, the contour profiles of the upper edge portion and the lower edge portion are calculated as one-dimensional vectors by searching the contour in the vertical direction at each point on the horizontal axis on the image.

[0056] Here, since the image obtained in the imaging step is imaged obliquely with respect to the conveyance direction of the steel sheet S, the resolution differs between the wave height direction and the wave pitch direction. Therefore, the resolution of the contour profile on the obtained image is corrected separately in the wave height direction and the wave pitch direction (step S5 in FIG. 6).

[0057] Specifically, in step S5, first, the inclination of the optical axis of the camera 2 is corrected by a rotation process so that the longitudinal direction and the horizontal axis of the steel plate S coincide. Since the resolution in the wave height direction, i.e., the vertical axis, is "rn (mm / pixel)" and the resolution in the wave pitch direction, i.e., the horizontal axis, is "rp (mm / pixel)", it can be converted into a contour profile of the steel plate S.

[0058] Furthermore, the resolution actually changes because the distance from the camera 2 differs between the front edge and the rear edge. Therefore, the resolution may be calculated and corrected separately for the position of the front edge and the position of the rear edge according to the distance from the camera 2. That is, the resolution in the wave height direction and the wave pitch direction may be calculated from the positional relationship between the camera 2 and the steel plate S, and converted into actual dimensions.

[0059] Alternatively, instead of carrying out the above-mentioned processing on the obtained contour profile of the edge portion, the contour profile of the steel sheet S may be calculated by performing coordinate transformation using the attitude parameters of the camera 2 and obtaining an orthogonal projection onto the plane β (see FIG. 2). This makes it possible to perform strict geometric correction in the entire field of view of the camera 2.

[0060] In addition, since the wave pitch and period of the sheet elongation of the steel sheet S are almost fixed, noise that does not contribute to the shape can be removed from the contour profile of the edge portion by applying a low-pass filter or band-pass filter that removes frequency components other than these.

[0061] Next, an index of the shape of the edge portion is calculated from the contour profile of the edge portion of the steel sheet S obtained as described above (step S6 in FIG. 6). The index of the shape of the edge portion is often discussed as a parameter called steepness, which is the ratio of wave height to wave pitch, but the index may be one or more appropriately selected from wave height, wave pitch, elongation amount, elongation rate, etc. There are various methods for calculating the index of the shape of the edge portion of the steel sheet S, but the following (i) to (iii) can be mentioned as representative methods.

[0062] (i) The maximum and minimum points are calculated, and the vertical distance of the image is taken as the wave height, and the horizontal distance is taken as the wave pitch. The steepness is calculated as the ratio of the wave height to the wave pitch. (ii) A sine curve is fitted, and the wave height and wave pitch are calculated from the amplitude and period. The steepness is calculated as the ratio of the wave height to the wave pitch. (iii) The amount and rate of sheet elongation are calculated from the contour length, and the steepness is directly calculated by the method of Non-Patent Document 2 mentioned above.

[0063] By using the thus obtained index of the shape of the edge portion of the steel sheet S, it becomes possible to carry out parameter control and feedback control during rolling, and further to judge the pass / fail of the sheet shape of the coil. Note that instead of calculating the index of the shape from the contour profile of the steel sheet S, for example, it is also possible to calculate the wave height, wave period, steepness, elongation amount, elongation rate, etc. from the contour profile on an image, and then correct them later using the resolution rn (mm / pixel) in the wave height direction and the resolution rp (mm / pixel) in the wave pitch direction.

[0064] Furthermore, as a behavior of the steel sheet S during transportation, there are cases where a different curve is detected even if sheet elongation (especially elongation of the edge portion of a strip-like object) does not occur. For example, there is a phenomenon in which the steel sheet S momentarily rises up and forms a large wave (called waving), and a phenomenon in which the steel sheet S continues to be transported without changing the shape of the wave within the field of view (called fixed wave). Since these phenomena are different from sheet elongation, it is preferable to exclude them from the detection results. Therefore, below, a method for distinguishing sheet elongation from waving and fixed wave will be explained.

[0065] Waving is a phenomenon that occurs when the conveying speed of steel plate S accelerates during transportation, and when the speed on the downstream side of a certain plate point becomes faster than that on the upstream side, the plate with nowhere to go bounces upward and forms large waves. This phenomenon occurs even if the shape of the plate edge is flat and no plate elongation has occurred, so it must be distinguished from shapes caused by plate elongation.

[0066] Waving is characterized by a very large wave pitch, a sudden and very large wave height, and a sudden increase in wave pitch or wave height compared to the preceding and following waves. Therefore, if a method can directly calculate the wave pitch or wave height, it can be determined that a sudden increase in these is waving rather than plate elongation. Then, a process such as masking the data for the period determined to be waving or filling in the gaps with data from before and after the period is performed. This makes it possible to exclude waving from the detection results of plate elongation, and to reduce the impact of the waving on the measurement results.

[0067] In addition, it is possible to use the waving characteristics to distinguish indices that do not directly calculate wave height or wave pitch, such as calculating sheet elongation from the contour of the steel sheet S. For example, even without using the entire field of view in the image, it is possible to monitor the contour position of one point in the longitudinal direction of the steel sheet S within the field of view, as shown in Figure 9, and determine that waving occurs when the variation in the contour position is extremely large.

[0068] A method for determining waving using the variation in contour position will be described with reference to Fig. 9. (a) to (e) of Fig. 9 show examples of photographs of waving on a steel sheet S. Moreover, in Fig. 9, the images are arranged in chronological order in the order of (a), (b), (c), (d), and (e). Specifically, the determination of waving can be performed in the following steps (1) to (3).

[0069] (1) A position Xt in the image width direction is determined. (2) Calculate the lower edge position Yt of each image with respect to the image width direction position Xt determined in (1). This lower edge position Yt is the lower edge position at the center of the image in the horizontal direction. (3) Fluctuations in the calculated lower contour position Yt are monitored, and locations where there has been a large fluctuation are detected.

[0070] The diagram (f) at the bottom of Fig. 9 shows a schematic diagram in which the lower contour position Yt is plotted in chronological order. Using such a schematic diagram, the presence or absence of waving can be determined. As a method of determining whether waving occurs, the number of maximum and minimum values ​​in a certain section of the transition of the lower contour position Yt may be monitored, and if the number of these values ​​suddenly decreases, it may be determined that waving occurs.

[0071] In the example of Fig. 9, the presence or absence of waving is determined using the lower contour position Yt, but the presence or absence of waving may be determined using the upper contour position. In addition, waving may occur independently on both edges of the steel sheet S. Therefore, it is preferable to monitor each of the upper contour position and the lower contour position to determine the presence or absence of waving.

[0072] Next, a method for determining fixed waves will be described. Fixed waves are a phenomenon in which the steel sheet S continues to be transported without changing the wave form. Within the field of view of the camera 2, the position and shape of the outline of the steel sheet S are observed to remain unchanged for a certain period of time. Like waving, this is a phenomenon in which the sheet speed at the front of the transport line steadily decreases, and the excess steel sheet S bends upward, resulting in a constantly curved outline. Like waving, fixed waves occur even when sheet elongation does not occur, so they must be distinguished from shapes caused by sheet elongation.

[0073] Since a fixed wave literally generates a stationary contour (contour that does not change) in the time direction, it is possible to detect the fixed wave by comparing the contour shapes before and after a certain time and calculating the amount of change. For example, the absolute value difference between contour profiles is taken, and if the sum or square sum of the difference falls below a threshold, it is determined to be a fixed wave. Then, processing such as masking the data for the period determined to be a fixed wave or filling in the gaps with data before and after is performed. This makes it possible to exclude the fixed wave from the detection results of plate elongation, and to reduce the impact of the fixed wave on the measurement results.

[0074] Also, similarly to waving, it is not necessary to use the entire field of view in the image. For example, as shown in Figure 10, the contour position of one point in the longitudinal direction of the steel plate S within the field of view can be monitored, and a fixed wave can be determined when the fluctuation in the contour position does not change extremely.

[0075] A method for determining whether a fixed wave exists using the variation in the contour position will be described with reference to Fig. 10. (a) to (e) of Fig. 10 show examples of images of fixed waves in a steel sheet S. In Fig. 10, the images are arranged in chronological order in the order of (a), (b), (c), (d), and (e). Specifically, the determination of whether a fixed wave exists can be performed by the following steps (1) to (3).

[0076] (1) A position X2t in the image width direction is determined. (2) Calculate the upper contour position Y2t of each image with respect to the image width direction position X2t determined in (1). This upper contour position Y2t is the upper contour position at the center in the horizontal direction of the image. (3) Fluctuations in the calculated upper contour position Y2t are monitored, and a portion where there is almost no fluctuation is detected.

[0077] The diagram (f) at the bottom of Fig. 10 shows a schematic diagram in which the upper contour position Y2t is plotted in chronological order. Using such a schematic diagram, the presence or absence of a fixed wave can be determined. As a method for determining whether or not a fixed wave exists, the standard deviation in a certain section of the transition of the upper contour position Y2t may be monitored, and a portion where the standard deviation falls below a threshold value may be determined as a fixed wave. Furthermore, the difference between the maximum value and the minimum value in a certain section may be monitored, and a portion where the standard deviation falls below a threshold value may be determined as a fixed wave.

[0078] In the example of Fig. 10, the presence or absence of a fixed wave is determined using the upper contour position Y2t, but the presence or absence of a fixed wave may also be determined using the lower contour position. In addition, a fixed wave may occur independently for both edges. Therefore, it is preferable to monitor the upper contour position and the lower contour position respectively to determine the presence or absence of a fixed wave.

[0079] On the other hand, even when the steel sheet S is in a flat state with no waves, the contour shape does not change for a certain period of time. At this time, the steel sheet S is not a fixed wave, and the shape and running condition of the steel sheet S are good. Fixed waves are a type of running defect caused by jamming of the steel sheet, and if the degree is severe, the steel sheet S may come into contact with equipment and cause trouble, so it is preferable to detect them automatically.

[0080] Therefore, by adding information on the occurrence of plate elongation to the evaluation using the change in the contour shape, it is possible to determine whether a fixed wave has occurred. For example, even if the change in the contour shape is large, if the index of plate elongation is below a threshold, it is not determined to be a fixed wave, but only if it is above the threshold. In addition, a method of designing an evaluation function and determining based on a threshold, or a method of visually determining whether a fixed wave is present as correct answer data and creating a determiner for determining whether a fixed wave is present using machine learning may be used.

[0081] In the binarization process in step S1, the brightness of the thermal radiation light varies depending on the temperature of the object to be measured. Therefore, if the binarization process is performed with a fixed value, when the steel sheet S is bright, part of the background will be detected as the sheet area, and conversely, when the steel sheet S is dark, the background cannot be detected.

[0082] To address such issues, it is desirable to carry out luminance correction. For example, a method of luminance correction is to multiply the luminance of the entire image by a fixed value so that a representative value such as the maximum value, average value, median value, percentile value, etc. of the luminance of the entire image becomes a target value. Dark current correction may also be performed before the multiplication.

[0083] Also, as shown in Figure 11, for example, there are cases where the thermal radiation light emitted from the steel sheet S is irradiated onto structures such as transport rolls and tables, causing the structures to appear bright in the image, making it difficult to distinguish them from the sheet surface. In this case, the sheet surface appears bright in the image because it directly receives the thermal radiation light, but the structures are often darker than the sheet surface because they are reflections of the thermal radiation light emitted from the sheet surface.

[0084] Therefore, the binarization threshold value that can separate the structure from the plate by performing the above-mentioned brightness correction may be determined as a fixed value, but more stable detection is possible by automatically detecting the valley portion from the shape of the histogram of the entire image. When detecting the valley portion, for example, as shown in FIG. 12, by applying a low-pass filter to the brightness histogram itself, it becomes possible to stably calculate the minimum value by differentiation processing, search processing, etc. FIG. 12(a) is an example of a brightness histogram, and is a graph with brightness on the horizontal axis and the number of pixels N on the vertical axis. FIG. 12(b) is an example of applying a low-pass filter to the brightness direction of the brightness histogram. Note that there are two valley portions in the histogram, one at the boundary between the plate surface and the structure, and the other at the boundary between the structure and the background, so it is preferable to select the larger of the two valleys. In FIG. 12(b), the portion divided by the dotted line corresponds to the binarization threshold value.

[0085] In addition, in this embodiment, an example has been described in which the image from the camera 2 is processed by the image processing device 3 to calculate an index of the shape of the edge portion of the steel sheet S, but the obtained camera image may simply be presented to the worker as it is and fed back to the rolling control.

[0086] Also, a machine learning discriminator may be used to estimate an index of the shape of the edge portion of the steel sheet S from an image. Specifically, learning data is created by linking an index of the shape of the edge portion measured visually or by some other method as a correct answer to an obtained image of the steel sheet S. Then, using the created learning data and a machine learning method, a discriminator is generated that takes an image as input and outputs an index of the shape of the edge portion, and this discriminator is used to calculate an index of the shape of the edge portion of the target steel sheet S. Also, input data to the machine learning and discriminator may be a feature calculated from an image, in addition to an image. Also, there is no restriction on the machine learning method, but if real-time performance is not required, a convolutional neural network or the like may be used.

[0087] (Example of application to shape pass / fail judgment) An example in which the shape measuring device for a belt-shaped object according to the embodiment is applied to a shape pass / fail judgment will be described with reference to Fig. 13. The pass / fail judgment system in Fig. 13 includes a camera 2, an image processing device 3, and a pass / fail judgment device 6.

[0088] First, the image processing device 3 calculates an index (shape data) of the shape of the edge portion of the steel sheet S from the image captured by the camera 2, and transmits it to the pass / fail determination device 6. The pass / fail determination device 6 determines whether the shape is pass / fail based on the steel sheet information (e.g., sheet thickness, sheet width, steel type, temperature, etc.) acquired from the upper system and the index of the shape of the edge portion of the steel sheet S. The pass / fail determination result is transmitted to the upper system. Based on the pass / fail determination result, the upper system determines whether correction is necessary, whether the defective shape portion should be cut off, whether rolling can be performed in the next process, etc. In this way, by utilizing the index of the shape of the edge portion of the steel sheet S in the action of the next process, it is possible to prevent trouble and improve product quality.

[0089] (Example of application to rolling feedback control) An example in which the shape measuring device for a strip-like object according to the embodiment is applied to rolling feedback control will be described with reference to Fig. 14. The rolling control system in Fig. 14 includes a camera 2, an image processing device 3, and a rolling control device 7.

[0090] First, the image processing device 3 calculates an index of the shape of the edge portion of the steel sheet S from the image captured by the camera 2, and transmits the calculated index to the rolling control device 7. The rolling control device 7 calculates control parameters using the index of the shape of the edge portion based on the steel sheet information (e.g., sheet thickness, sheet width, steel type, temperature, etc.) acquired from the upper system, other measurement data such as the sheet passing position, and the index of the shape of the edge portion of the steel sheet S. In addition, the rolling control device 7 performs feedback control such as leveling by transmitting a control signal to the rolling rolls 1. By performing such feedback control, the shape of the product can be stabilized, and shape defects themselves can be reduced.

[0091] (Application of machine learning to rolling control) An example in which the shape measuring device for a strip-like object according to the embodiment is applied to rolling control using machine learning will be described with reference to Fig. 15. The rolling control system in Fig. 15 includes a camera 2, an image processing device 3, a rolling control device 7, a data server 8, a machine learning device 9, and a control parameter estimation device 10.

[0092] First, the image processing device 3 calculates an index of the shape of the edge portion of the steel sheet S from the image captured by the camera 2, and transmits the calculated index to the data server 8. The data server 8 accumulates the index of the shape of the edge portion in a state linked to the steel sheet information (e.g., sheet thickness, sheet width, steel type, temperature, etc.) acquired from a higher-level system and the control parameters during rolling.

[0093] These accumulated data are transmitted to a machine learning device 9. The machine learning device 9 constructs a model by machine learning to estimate control parameters for rolling while suppressing shape defects. The estimated model is transmitted to a control parameter estimation device 10, which estimates control parameters based on the steel plate information acquired from a higher-level system. The control parameters are transmitted to a rolling control device 7. The rolling control device 7 performs preset control such as leveling by transmitting a control signal to the rolling rolls 1. By performing such preset control, the shape of the product can be stabilized and shape defects themselves can be reduced.

[0094] In this application example, the control parameters in the preset control have been described, but machine learning may also be applied to calculate the control parameters in the feedback control. In addition, by combining the preset control and the feedback control, it is possible to further improve the shape quality of the product.

[0095] In addition to the above application examples, the device for measuring the shape of a belt-like object according to the embodiment can be provided as part of a manufacturing facility for belt-like objects.

[0096] The method for measuring the shape of a strip-like object according to the embodiment may also be applied to a method for controlling the shape of a strip-like object, in which the shape of the strip-like object is measured by the method for measuring the shape of a strip-like object, and the shape of the strip-like object is controlled to a desired shape based on the measurement result.

[0097] The method for measuring the shape of a web-like object according to the embodiment may also be applied to a method for manufacturing a web-like object, in which the shape of the web-like object is measured by the method for measuring the shape of a web-like object, and the web-like object is manufactured based on the measurement results.

[0098] The method for measuring the shape of a strip-like object according to the embodiment can also be applied to a method for controlling the quality of a strip-like object, in which the shape of the strip-like object is measured by the method for measuring the shape of a strip-like object, and the quality of the strip-like object is controlled based on the measurement result.

[0099] The method for measuring the shape of a strip-like object, the method for controlling the shape of a strip-like object, the method for manufacturing a strip-like object, the quality control method for a strip-like object, the device for measuring the shape of a strip-like object, and the manufacturing equipment for a strip-like object according to the above-described embodiments provide the following effects.

[0100] First, by passively capturing images of the thermal radiation light from the steel sheet S being measured, it is possible to capture images from a good installation environment at a distance without having to install light sources or sensors around the steel sheet S while it is being passed, and the shape of the edge of the steel sheet S can be clearly captured.

[0101] In addition, by selecting an imaging element suitable for the temperature of the target as the imaging element of the camera 2, it is possible to efficiently receive thermal radiation light. For example, for a measurement target of about 900°C, sensitivity can be increased by using the near-infrared components of a Si imaging element, and clear, blur-free images can be obtained by shortening the exposure time, even for steel plates S transported at high speed.

[0102] Moreover, by capturing an image so that the angle θ between the plane α, which is the reference plane of the surface of the strip-like object (steel sheet S), and the optical axis of the camera 2 is not 90 degrees, and the angle φ between the orthogonal projection of the optical axis of the camera 2 onto the plane α and the conveying direction p of the strip-like object (steel sheet S) is not 0 degrees, the contours of both edges of the steel sheet S can be clearly captured and an index of the shape of the edge can be quantitatively calculated, so that the sheet elongation can be clearly captured. Furthermore, a higher effect can be obtained by capturing an image of the strip-like object (steel sheet S) from the width direction (i.e., the direction in which the angle φ is close to 90 degrees). Alternatively, a higher effect can be obtained by capturing an image at a low angle (i.e., the direction in which the angle θ is close to 0 degrees). Of course, an even higher effect can be obtained by capturing an image from the direction in which the angle φ is close to 90 degrees and the angle θ is close to 0 degrees.

[0103] In addition, the present invention has been described with respect to the steel plate S at the finish rolling exit side of the hot rolling process, but it goes without saying that the present invention can also be applied to other high-temperature plate-shaped steel materials such as thick steel plates and slabs. Furthermore, it goes without saying that the present invention can be applied to various strip-shaped objects made of different materials, not just strip-shaped objects in the steelmaking process, as long as the thermal radiation light of the object to be measured can be obtained. Furthermore, the present invention is preferable because it provides greater effects when the elongation of the edge of a strip-shaped object is the object of shape measurement.

[0104] The method for measuring the shape of a strip-like object, the method for controlling the shape of a strip-like object, the method for manufacturing a strip-like object, the method for quality control of a strip-like object, the shape measuring device for a strip-like object, and the manufacturing equipment for a strip-like object according to the present invention have been specifically described above using the embodiments and examples for carrying out the invention, but the gist of the present invention is not limited to these descriptions and should be broadly interpreted based on the claims. Furthermore, it goes without saying that various changes and modifications based on these descriptions are also included in the gist of the present invention. [Explanation of symbols]

[0105] 1. Rolling mill 2 Camera 3. Image Processing Device 6. Pass / fail judgment device 7. Rolling control device 8 Data Server 9 Machine Learning Devices 10 Control parameter estimation device S steel plate

Claims

1. A shape measuring method for measuring a shape of a band-like object, comprising the steps of: The angle θ between the plane α, which is the reference plane of the surface of the belt-like object, and the optical axis of the camera is not 90 degrees. In addition, the angle φ between the orthogonal projection of the optical axis of the camera onto the plane α and the conveying direction p of the belt-like object is not 0 degrees. an imaging step of capturing an image of thermal radiation light from the belt-shaped object; an image processing step of calculating a contour profile of the belt-like object from the obtained image, thereby calculating an index of the shape of an edge portion of the belt-like object; A method for measuring the shape of a strip-shaped object, comprising:

2. The image processing step includes: Extracting the region of the belt-like object from the obtained image; calculating a contour profile of the belt-like object by calculating the position of the edge portion in the extracted region; The method for measuring the shape of a belt-like object according to claim 1.

3. The shape measuring method of a strip-like object as described in claim 1, wherein the image processing step calculates one or more of the steepness, wave height, wave pitch, elongation amount, and elongation rate of the edge portion of the strip-like object from the obtained contour profile of the strip-like object as the index.

4. 2. The method for measuring the shape of a band-like object according to claim 1, wherein said image processing step calculates resolutions in the wave height direction and wave pitch direction from a positional relationship between said camera and said band-like object, and converts the resolutions into actual dimensions.

5. A method for controlling the shape of a strip-like object, comprising: measuring the shape of the strip-like object using a method for measuring the shape of the strip-like object according to any one of claims 1 to 4; and controlling the shape of the strip-like object to a desired shape based on the measurement results.

6. A method for manufacturing a belt-like object, comprising the steps of: measuring a shape of the belt-like object by the method for measuring a shape of the belt-like object according to any one of claims 1 to 4; and manufacturing the belt-like object based on the measurement result.

7. 5. A quality control method for a web-like object, comprising the steps of: measuring a shape of the web-like object by the method for measuring a shape of the web-like object according to claim 1; and controlling the quality of the web-like object based on a measurement result.

8. A shape measuring device for measuring a shape of a belt-like object, comprising: The angle θ between the plane α, which is the reference plane of the surface of the belt-like object, and the optical axis of the camera is not 90 degrees. In addition, the angle φ between the orthogonal projection of the optical axis of the camera onto the plane α and the conveying direction p of the belt-like object is not 0 degrees. an imaging means for capturing an image of thermal radiation light from the belt-shaped object; an image processing means for calculating a contour profile of the belt-like object from the obtained image, thereby calculating an index of the shape of an edge portion of the belt-like object; An apparatus for measuring the shape of a strip-shaped object comprising:

9. A manufacturing facility for a web-like object, comprising the device for measuring the shape of a web-like object according to claim 8.

Images