Device and method for checking the flatness of a metal sheet

Abstract

The invention relates to a device for checking the flatness of a moving metal sheet, which comprises a light source that homogeneously illuminates the sheet, a camera that records the distorted image of the light source formed by the sheet, and a calculation device that continuously calculates a coefficient correlated to the flatness by comparing the distorted image of the light source with the reference image of the light source. The invention also relates to a method for checking the flatness by means of the device according to the invention using a predetermined model M for analysing the distorted image of the light source with respect to the reference image of the light source.

Description

DESCRIPTION TITLE: Device and method for checking the flatness of a metal sheet. TECHNICAL FIELD

[0001] The field of the invention is that of devices and methods for contactless flatness control of metal sheet or strip made in particular of aluminum or aluminum alloy. STATE OF PRIOR ART

[0002] A metal sheet may have flatness defects, such as non-developable defects (e.g. long edges, long centers or pockets, etc.) and developable defects (e.g. bend, tile and twist defects, etc.). Flatness is one of the many requirements that sheet users include in their specifications.

[0003] Patent EP1055905 discloses a method which involves using at least two radiation sources and a number of detectors to detect control values ​​from a number of control points arranged at intervals in a direction transverse to the direction of the moving sheet metal (2). The control points are detected by at least two detectors which each detect the radiation at different angles and evaluate the contours of the sheet metal. The patent also relates to a device for determining the flatness of a sheet of material and the use of a device for determining the flatness of a sheet of material.

[0004] Patent US6480802 discloses a method for determining the flatness of a sheet of material and a device for carrying out the method. The method and the device solve the technical problem of calculating the elongation of the sheet from the values ​​of the contour of the sheet and determining the flatness of the sheet of material. The method consists in calculating, from the changes in slope values ​​measured at a plurality of control points, the wavelength and the phase of these changes. From this is calculated the position of at least one extremum, at which the measured slope values ​​have only a transverse component. The slopes are added together to calculate a contour, from which the amplitude is calculated. The elongation of the sheet as a control of the flatness of the sheet of material is then determined from the wavelength and the amplitude.

[0005] Patent EP3487642 relates to a method and device for checking the flatness of the transported sheet metal material, comprising the following steps: moving the material of sheet metal through a sheet metal machining device, wherein a tensile force of at least 10 N / mm2 is applied to the sheet metal material; generating at least one projection pattern in the form of a projection grid on the surface of the sheet metal material by means of a projection-producing optical system, said projection pattern being projected onto the surface of the sheet metal material from a position which is laterally offset from the central plane of the sheet metal material, such that a projection angle formed between the surface of the sheet metal material and a projection beam is between 1° and 45°; and detecting the projection pattern by means of a camera, said camera being arranged on a transverse plane above the sheet metal material when viewed in the direction of movement.

[0006] Patent EP2910893 discloses a method for determining flatness variations during the processing of sheet-shaped material, in particular in the hot rolling of thin sheet metal, in which a TOF camera extending transversely to the direction of movement of the sheet-shaped material area and by means of an evaluation unit, is connected to the TOF camera, a flatness deviation can be determined.

[0007] Patent EP0864847 discloses a method of evaluating a line trace directed onto the surface of sheet metal by a projector. A CCD camera is used to directly examine the line trace. The image of the pattern provided by the camera can be compared with a reference pattern. The measured values ​​provided in this way can be used to adjust the production path of the sheet metal.

[0008] Patent EP1899086 discloses a method for producing a metal sheet, in which the metal sheet is guided over a number of rollers under such a sheet stress, and is moved in a conveying direction, that it is largely flat at least between two rollers. In order to be able to inspect the sheet in a simple and space-saving manner, the invention provides the internal tensile stresses acting in the metal sheet which is largely flat under the tensile stress between at least two rollers to be made optically visible and for the tensile stresses or tensile stress differences which are determined in this way to be used during the production of the metal sheet. Furthermore, the invention relates to an apparatus for producing a metal sheet.

[0009] Application EP1570236 discloses an apparatus and method for detecting twist in an object such as pieces of wood carried on a conveyor using a non-contact scanning technique in which a pair of transverse scanning line beams are directed onto a surface of the article in spaced relationship, and scans successive scans of simultaneously scanned pairs of spaced cross-sectional areas on the article are carried out repeatedly while the article is being transported, to generate profile data characterizing the position of each cross-sectional area in a reference system. The profile data characterizing the respective position of the two cross-sectional areas are compared with each other to generate indicative partial torsion data associated with each acquisition, followed by a summation of the indicative partial torsion data associated with all acquisitions to obtain an indication of the torsion in the considered part of the article.

[0010] Patent EP2931447 discloses flatness measurement and residual stress measurement in a metal flat product. A problem addressed by the invention is that of increasing the accuracy and reliability of existing flatness measuring devices and / or methods. This problem is solved by a method for measuring the flatness of a metal flat product which comprises the following method steps: - bending the flat product in a bending device such that a flat flat product would form an arc with a target bending radius rO after bending; - measuring the contour, in particular the actual bending radii, in the region of the arc of the bent flat product at several positions (y) in the width direction of the flat product; and - determining the flatness of the flat product taking into account the measured contour of the bent flat product.

[0011] Application EP2265895 discloses a structured light sensor system for measuring the contour of a surface comprising an imaging lens system, an image capture device, a first set of microelectromechanical system (MEMS) mirrors, and a control module. The imaging lens system focuses light reflected from the surface, wherein the imaging lens system has a corresponding lens plane. The image capture device captures the focused light and generates data corresponding to the captured light, wherein the image capture device has a corresponding image plane that is not parallel to the lens plane. The first set of MEMS mirrors directs the focused light toward the image capture device.The control module receives the data, determines the focusing quality of the captured light based on the received data, and controls the first set of MEMS mirrors based on the focusing quality to maintain a Scheimpflug tilt condition between the objective plane and the image plane.

[0012] Patent EP0397672 discloses a method and system for high-speed, high-resolution 3D imaging of an object at a viewing station, including an anamorphic magnification and field lens system, for transmitting reflected light from the object to a small-area position detector having a position-detecting direction. Preferably, an acousto-optic deflector having no moving parts with associated lens elements scans a modulated laser light beam in a scanning direction across the object to be inspected to produce a flat-field telecentric scan. The deflector has a feedback loop to allow uniform illumination of the object (i.e., flat-field correction). Light scattered by the object is collected by a telecentric receiving lens.

[0013] Application WO2012 / 037186 ​​discloses a non-contact sensing system is provided for acquiring three-dimensional contour information of an object. The system is comprised of a light source subsystem operable to scan a spot of light within an illumination area; a first imaging device having a field of view disposed to intersect with the illumination area and operable to capture image data; and a second imaging device having a field of view disposed to intersect the illumination area and operable to capture image data. A first control module is in data communication with the first imaging device to determine contour information of an object within the field of view of the first imaging device and report the contour information for the object in a common coordinate system.A second control module is in data communication with the second imaging device to determine the contour information of the object in the field of view of the second imaging device and to report the contour information for the object in the common coordinate system. In addition, the light source subsystem is calibrated to indicate the position of the light point in the common coordinate system.

[0014] US6252659 discloses a three-dimensional measuring apparatus comprising an optical system for scanning a reference beam across a target object to be measured, a light sensor that receives light reflected from the target object, and a processor for calculating a dimensional shape of the target object from the received light. An image for calculating the three-dimensional shape of the target object and an image for displaying the target object are both captured by the same sensor. The displayed image is a grayscale image based on a centroid calculated from multiple data samples taken for each pixel of the image during target acquisition.

[0015] Patent application EP1346204 discloses a device for automatic surface inspection of an unwound strip to detect surface defects, comprising at least one camera whose optical axis is directed towards the surface to be inspected and forming on said surface a transverse line of sight and at least one lighting system whose incident rays are directed towards the transverse line of sight and distributed over the entire length of said line. At each point on the transverse line of sight, the direction of observation of the camera forms a constant angle with the specular ray reflected at said point.

[0016] Patent application US3590258 discloses an apparatus for controlling the shape of a rolled product moving in a longitudinal direction, comprising means for projecting a linear image onto the surface of the rolled product in its transverse direction, means for detecting the longitudinal deviation of a portion of said linear image at least at the central portion and the end of said image, and oil pressure means for regulating the pressure actions on the central portion and the output of said detecting means.

[0017] Application DE102014104338 discloses a simple and rapid method for determining small surface deformations on reflective or shiny moving parts. For this purpose, a method is provided for detecting surface deformations on reflective objects, in which a light source is provided which has at least one linear light-dark transition, in which the surface of the object is illuminated by the light source at an oblique angle. The light from the light source reflected by the surface of the object and the path of the light-dark transition are detected by means of an image detector, while the object is moved relative to the arrangement consisting of the image detector and the light source. The direction of movement of the object is transverse to the longitudinal direction of the light-dark transition.A surface deformation on the object surface is detected based on a deviation from the linear trajectory of the light-dark transition imaged with the imaging detector. STATEMENT OF THE INVENTION

[0018] The invention aims to propose a device and a method for non-contact flatness control of sheet metal or strip which is simpler than the prior art and compatible with the demanding environment of a heavy industry metalworking production workshop.

[0019] The invention is a device for checking the flatness of a metal sheet or strip 1 moving in a direction X, comprising: a) A light source 2 illuminating an observation field 4 of the sheet or strip 1 in a substantially homogeneous manner, b) A camera 3, which is a 2D digital camera and which is arranged to record in the field of observation 4 at least one photograph of a deformed image 22 of the light source 2 formed by the sheet or strip 1, c) A calculation device making it possible to compare at least one photograph of the deformed image 22 of the light source 2 with a photograph of a reference image 20 of the light source 2 and to continuously calculate at least one coefficient correlated with the flatness of the sheet or strip 1.

[0020] The invention is also a sheet metal or strip production machine comprising the device according to the invention.

[0021] The invention is also a method for controlling the flatness with the device according to the invention comprising the successive steps: a) The camera 3 transmits a photograph of the deformed image 22 of the light source 2 to the calculation device, b) a predetermined model M implemented in the calculation device determines a photograph of the reference image 20 of the light source 2 c) the predetermined model M compares the photograph of the deformed image 22 of the light source 2 with the photograph of the reference image 20 of the light source 2, d) the predetermined model M calculates at least one coefficient correlated with the flatness of the sheet or strip 1. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Other aspects, aims, advantages and characteristics of the invention will appear better on reading the following detailed description of preferred embodiments thereof, given by way of non-limiting example, and made with reference to the appended drawings in which:

[0023] Figure 1 is a schematic view of the device seen from above.

[0024] Figure 2 is a schematic view of the side of the device in the case of a perfectly flat sheet or strip.

[0025] Figure 3 is a schematic side view of the device showing the displacement of the image of the light source when the sheet or strip has a flatness defect.

[0026] Figure 4 is a schematic view from the camera location showing the image of the light source with a perfectly flat sheet or strip and with a sheet with a flatness defect.

[0027] Figure 5 shows an example of a photograph of a distorted image 22 of the light source 2 obtained with the device according to the invention.

[0028] Figure 6 shows a photograph of the distorted image 22 of a light source obtained with a device which is not according to the invention.

[0029] Figure 7 shows an example of analysis of the photograph of the distorted image 22 of the light source 2.

[0030] Figure 8 shows another example of analyzing the photograph of the distorted image 22 of the light source 2.

[0031] Figure 9 shows a counterexample of the device according to the invention.

[0032] Figure 10 shows a schematic example of a production line comprising the device according to the invention.

[0033] Figure 11 shows a schematic example of a production line comprising the device according to the invention. DETAILED DESCRIPTION OF SPECIFIC EMBODIMENTS

[0034] In the figures and in the remainder of the description, the same references represent identical or similar elements. In addition, the different elements are not shown to scale so as to enhance the clarity of the figures. Furthermore, the different embodiments and variants are not mutually exclusive and may be combined with each other. Unless otherwise indicated, the terms "substantially", "approximately", "of the order of" mean to within 10%, and preferably to within 5%. Furthermore, the terms "between ... and ..." and equivalent mean that the limits are included, unless otherwise indicated.

[0035] Sheet or strip 1 is a metal sheet or strip, preferably made of aluminum or aluminum alloy. Indeed, sheets or strips made of aluminum or aluminum alloy have a reflectance of light visible to humans. The terms and definitions of EN 12258-1 (2012) are applicable. The difference between the terms sheet and strip is defined by note 3 of paragraph 2.6.1 of EN 12258-1 (2012), in this case: in Europe, the term "sheet" is only used for rolled products supplied in a straight length, for sheets in coils, the term "strip" is used. Sheet is therefore a flat product. A strip can be transformed into sheet metal by any cutting or lengthening operation known to those skilled in the art. However, the term strip is extended to rolled products with a thickness of less than 0.20 mm. Standards EN 485-3 (2003) and EN 485-4 (1994) define the shape tolerances for sheets, strips and thick plates made of aluminium or aluminium alloy. These standards describe in particular certain flatness defects well known to those skilled in the art. The metallurgical conditions are defined by EN 515 (2017).

[0036] The sheet or strip 1 is in motion in the direction of movement X. The sheet or strip therefore has a translational movement. The movement can be relative, the device can therefore be moved relative to the sheet. In the case of a sheet, the preferred direction of movement is either the long rolling direction L, or the long cross direction TL, the one corresponding to the long side of the sheet being preferred to reduce the size of the machine. The directions L and TL are not shown. In the case of a strip, it is unwound and the direction of movement of the strip is the long direction or rolling direction. The perpendicular direction Y is the direction perpendicular to the direction of movement X in the plane of the sheet or strip 1.

[0037] Figure 1 shows the sheet or strip 1 and the device seen from above. The X direction is the direction of movement of the sheet or strip 1 and the Y direction is the direction perpendicular to the X direction in the plane of the sheet or strip 1. Since this is a top view, neither the reference image 20 of the light source 2 nor the deformed image 22 of the light source 2 are in the field of observation 4. Given the laws of geometric optics, these images are located under the light source 2. The reference image 20 of the light source 2 is masked by the presence of the light source 2. On the other hand, the deformed image 22 of the light source 2 may be visible in a top view depending on the deformation of the sheet or strip 1.

[0038] Figures 2 and 3 show a side view of the device. Figure 2 corresponds to a perfectly flat sheet or strip 1. The reference image 20 of the light source 2 is the symmetrical image of the light source 2 with respect to the perfectly flat sheet or strip, taking into account the laws of geometric optics. Figure 3 shows a flatness defect of the sheet or strip 1 crossing the observation field 4 (not shown). Taking into account the laws of geometric optics, the image of the light source 2 moves and deforms and becomes the deformed image 22 of the light source 2. This movement and this deformation depend on the progress of the flatness defect of the sheet or strip 1.

[0039] Figure 4 shows the previous device with camera 3 as observation point. Camera 3 sees and can photograph the image of light source 2 as reference 20 that distorted 22 in the field of observation 4. Figure 9 shows the device with an observation point that is inappropriate for the camera 3 because this observation point is too close to the light source 2 because the photograph of the distorted image 22 of the light source 2 is not in the field of observation 4. The analysis cannot therefore take place correctly. The camera 3, the light source 2 and the field of observation must therefore be arranged taking into account the laws of geometric optics known to those skilled in the art.

[0040] The light source 2 illuminates an observation field 4 on the sheet or strip 1 in a substantially homogeneous manner. The observation field 4 is fixed relative to the light source 2 and to the camera 3. The observation field 4 is a predetermined part of the plane which corresponds to the upper face of the sheet or strip 1 if it is perfectly flat. Preferably, the illumination of the observation field 4 does not vary by more than 10%, more preferably 5%, more preferably 2% relative to the average value of the illumination (luminous flux received per unit area) of the observation field 4. A heterogeneous illumination degrades the analysis of the photograph of the distorted image 22 of the light source 2, which degrades the result of the calculations of the calculation device. The illumination is the quantity of light per unit area. The light source 2 therefore does not project a particular pattern onto the sheet.

[0041] The field of observation 4 is chosen so that the photograph of the distorted image 22 of the light source 2 as well as the photograph of the reference image 20 of the light source 2 are in the field of observation 4 for the camera 3. The field of observation 4 is therefore predetermined according to the flatness defects of the sheet or strip 1, which the person skilled in the art knows from experience.

[0042] Preferably, the observation field 4 is a rectangle whose short sides are parallel to the edges of the sheet or strip 1. Preferably, two of the edges of said rectangle correspond to the edges of the sheet or strip. Preferably, the ratio between the short side of said rectangle and the width of the sheet or strip is greater than 30%. According to the experience of the inventors, this ratio of 30% is sufficient to observe or control the usual flatness defects of the sheet or strip 1 made of aluminum or aluminum alloy. If said ratio is less than 30%, a photograph of the deformed image 22 of the light source 22 may no longer be recorded completely by the camera 3, which will degrade the analysis of the photograph of the deformed image 22 of the light source 2, which will therefore degrade the result of the calculations of the calculation device.

[0043] The light source preferentially emits light that is visible to humans. Visible light is advantageous because it allows the use of many sources or cameras. commercially available. The light source 2 is not a coherent light source such as that produced by a laser, for example, but not limited to. Preferably, the color of the light source is white to have the best contrast on the photographs. Preferably, the light source 2 is made up of light-emitting diodes (LEDs). LEDs are less expensive than a laser. LEDs are also less dangerous for the eyes than a laser. LEDs do not cause an iridescence phenomenon. A light source such as a fluorescent tube causes iridescence which deteriorates the quality of the image for the control of flatness. The flicker frequency of the light source 2 is sufficiently higher than the acquisition frequency of the camera 3 to avoid flickering of the photograph of the distorted image 22 of the light source 2 recorded by the camera 3.Given the high flicker frequency of LEDs, LEDs are preferably used as light source 2 to check the flatness of a fast moving strip such as on an aluminum strip finishing shear whose speed can typically reach 900 m / min. The LEDs make it possible to obtain a photograph of the distorted image 22 of the light source that is very clear as shown in Figure 5. Figure 6 shows a photograph of the distorted image of a fluorescent tube whose quality is insufficient due to flicker. Preferably, the density of the LEDs of the light source 2 is sufficient for the light source 2 to appear continuous in the photograph of the reference image 20 or in the photograph of the distorted image 22 of the light source 2. This continuity facilitates the analysis of the image by the predetermined model M as a continuous object, in particular a line.

[0044] Preferably, the light source 2 is sufficiently extended in the perpendicular direction Y so that the images of the ends of the light source 2 in the perpendicular direction Y are not photographed in the field of observation 4, preferably so that the images of the ends of the light source 2 in the perpendicular direction Y are not photographed by the camera 3. This ensures control of the flatness of the sheet or strip 1 in the field of observation 4, preferably in the entire field of observation that the camera 3 can photograph.

[0045] Preferably, the light source 2 is of substantially linear and straight shape. More preferably, the light source 2 is substantially perpendicular to the direction of movement X. The light source 2 is of substantially linear shape if the photograph of the reference image 20 of the light source 2 if there is a line such that any point of the photograph of the reference image 20 of the light source 2 is at a distance of less than 10%, preferably 5%, of the width of the sheet or strip 1 at least a point on said line. The light source 2 is of substantially straight shape if the length of the long side of the smallest rectangle, small in the sense of the smallest possible surface area, which contains the photograph of the reference image 20 of the light source 2 is at least 10 times longer than the short side of said rectangle. The substantially linear and straight shape of the light source 2 makes it possible to achieve a preferred embodiment of the method according to the invention described later. This orientation and this shape make it possible to reduce the length along the direction of movement X of the flatness control device. Preferably, the light source 2 is in a plane substantially parallel to the plane of the sheet or strip 1. The plane of the sheet or strip 1 is the plane if the sheet or strip 1 were perfectly flat. If referencing means, as described below, are used, they define such a plane.This reduces the difference in length between the longest and shortest optical path between the light source 2 and the camera 3 reflected by the sheet or strip 1 according to the laws of geometric optics. This reduces the distortion of the photograph of the distorted image 22 of the light source 2 and improves the analysis of the photograph of the distorted image 22 of the light source 2.

[0046] Preferably, there is no other light source that could directly or indirectly create another image photographable by the camera 3 to avoid degrading the quality of the photographs for their analysis. The device is therefore advantageously protected from other light sources with protections or covers while allowing the movement of the sheet or strip 1. Preferably, the camera 3 and the light source 2 are arranged so that the light source 2 does not appear on the photographs transmitted by the camera 3 to avoid any confusion with the distorted image 22 of the light source 2 which could disturb the calculation of at least one coefficient correlated with the flatness.

[0047] Camera 3 is a simpler 2D digital camera than a 3D camera. Camera 3 is not a TOF (time of flight) camera, which is an expensive piece of equipment. Camera 3 is arranged to record at least one photograph of the distorted image 22 of the light source 2 formed by the metal sheet 1 in the field of observation 4. If the distorted image 22 of the light source 2 is not formed in the field of observation 4 as seen by camera 3, this will degrade the quality of the analysis of the photograph. This degradation is caused by a distorted image 22 of the light source 2 photographed by camera 3 incompletely in the field of observation 4.

[0048] Preferably, the camera 3 records photographs with a predetermined period. The recording period is predetermined by the speed of movement of the sheet or strip 1 and by the flatness defects that the person skilled in the art knows from experience. The predetermined period must be sufficient to correctly sample the flatness defects taking into account the speed of movement of the sheet or strip 1. The choice of camera 3 must be adapted according to the size of the defects to be checked and the speed of movement of the sheet or strip 1. The resolution of camera 3 must be sufficient to correctly sample the observation field 4 and observe the desired flatness defects.

[0049] Preferably, the camera 3 is substantially centered relative to the field of observation 4. The position of the camera 3 is centered if this position corresponds to the minimum of the difference in length between the longest and shortest optical path between the light source 2 and the camera 3 reflected by the sheet metal relative to the other positions of the camera 3 obtained by a translation in the perpendicular direction Y. If the field of observation 4 is a rectangle whose short sides are parallel to the edges of the sheet metal or strip 1, the camera 3 is substantially arranged in a plane perpendicular to the field of observation 4 and containing the mediator of the long side of the field of observation 4. This position reduces the distortion of the photograph of the distorted image 22 of the light source 2 and improves the analysis of the photograph of the distorted image 22 of the light source 2.The camera 3 being arranged to record the photograph of the distorted image 22 of the light source 2, it is not focused to record photographs of the surface of the sheet or strip. 1. Camera 3 cannot therefore photograph and analyze surface defects or patterns projected onto the surface of sheet metal or strip 1.

[0050] The relative position of camera 3, light source 2 and sheet or strip 1 is a compromise. The minimum optical path is the shortest light path from light source 2 to camera 3 reflected by sheet or strip 1 according to the laws of geometric optics. The longer the minimum optical path, the more sensitive the device will be to check for a flatness defect. On the other hand, the longer the minimum optical path, the more the sensitivity and resolution of camera 3 and / or the light power of the light source must be increased. 2. Preferably, the camera 3 and the light source 2 are at the same height to facilitate the design of the device. Preferably, the camera 3 and the light source 2 are at least one width of the strip or sheet 1 above the sheet or strip 1. Such an arrangement makes it possible to limit the depth of field of the camera 3 so that the photograph of the deformed image 22 of the light source 2 is clear between the closest and most distant point of the deformed image 22 of the light source 2 with the camera 3. This makes it possible to avoid the use of a camera 3 with a complex lens. Preferably, in the case of a strip stretched between two referencing means described below, the camera 3 is substantially above one of the referencing means, and the light source 2 is substantially above the other referencing means. This arrangement makes it possible to make the best use of the surface area of ​​the device without interfering with other equipment. In one embodiment, the camera 3 and the light source 2 are at least 2 m above the strip to facilitate access to the interior of the device when servicing the device in an industrial context.

[0051] Preferably, the device according to the invention comprises referencing means which are arranged to ensure that the sheet or strip 1 is positioned to avoid movements of the sheet or strip 1 in directions other than the direction of movement X. These movements in other directions are likely to cause displacements of the deformed image 22 of the light source 2 and therefore to degrade the quality of the coefficient correlated to the flatness of the strip or sheet 1. These referencing means are, for example, but not limited to, a laying surface or rollers in the case of a sheet. The means for referencing a strip may be, for example, but not limited to, rollers, pinch rollers, S-blocks, unwinders or winders, which are well known to those skilled in the art of strip production machines.Preferably, the observation field 4 for a strip is arranged between two referencing means and without any other referencing means between the two said referencing means. Preferably, these two referencing means are spaced apart by at least the width of the strip 1 so that flatness defects are not attenuated by the proximity of the two referencing means and to ensure good control of the flatness. Preferably, the area of ​​the strip 1 on which the light rays emitted by the light source 2 are reflected towards the camera 3 is substantially centered between the two referencing means so that flatness defects are not attenuated by the proximity of the two referencing means and to ensure good control of the flatness. The referencing means tension the strip.The band must not be too tight so as not to degrade the quality of the coefficient correlated with flatness as explained later.

[0052] The device according to the invention is preferably part of a sheet or strip production machine. Preferably, these machines are finishing or finishing machines such as, for example, but not limited to, shears, saws, planers, varnishers, degreasers or continuous heat treatment furnaces. The environment and the level of tension of the sheet during rolling make it more difficult to install the device according to the invention on a rolling mill.

[0053] Figure 10 shows a non-limiting example of a strip production machine 1 comprising an unwinder 11, two rollers 12 and a winder 13 which are means of referenced to the strip 1 and which are spaced at least by the width of the strip. The machine of Figure 10 can also have other functions not shown such as those of the aforementioned machines. Figure 11 shows a non-limiting example of a strip or sheet production machine. The strip is unwound on an unwinder 11. The strip then passes into a tensioned leveler framed by two S-shaped blocks 14, the leveling cage not being shown but known to those skilled in the art. The strip 1 then passes into the device according to the invention and then into pinch rollers 15. The device according to the invention is positioned in the space between the two referencing means which are the S-shaped block 14 and the pinch rollers 15. Then the strip 1 is cut into sheet 1', the cutting and stacking device, known to those skilled in the art, is not shown.

[0054] When the person skilled in the art measures the various flatness characteristics of a sheet or strip, he places the sheet or strip, in the case of a strip generally a sample, on a control surface to carry out the flatness measurements with a ruler. Flatness defects such as long edges or pocket lines result from a difference in length in the width of the sheet or strip 1. The stretched strip lengthens, which reduces the difference in length and which modifies the flatness of the moving strip 1. If the strip is stretched by too much traction, this can cancel out the difference in length and make it impossible to check the flatness. Too much traction can also permanently deform the strip if the traction exceeds the elastic limit of the strip. The level of traction applied to the strip is predetermined by the various equipment of the machine on which the device according to the invention is used.For example, if one of the referencing means comprises a winder which winds the strip into a reel, the traction must be sufficient to ensure the cohesion of the reel and to avoid surface defects due to friction of the strip on itself. The traction must not be too strong to avoid collapse of the strip wound into a reel on its internal diameter. Preferably, the traction is at least 10 MPa, preferably 15 MPa. Preferably, the traction is at most 30 MPa.

[0055] The reference image 20 of the light source 2 is the image of the light source 2 for a perfectly flat sheet, that is to say that the reference image 20 is symmetrical with respect to the sheet or strip 1 of the light source according to the laws of geometric optics. When a flatness defect of the moving sheet or strip 1 crosses the observation field 4, the reference image 20 of the light source 2 moves and deforms according to the laws of geometric optics to become the deformed image 22 of the light source 2. For those skilled in the art of geometric optics, the light source 2 and the reference image 20 of the light source 2 are conjugated by the plane of the sheet or strip 1, which is the plane if the sheet or strip 1 were perfectly flat. For those skilled in the art of geometric optics, the light source 2 and the deformed image 22 of the light source 2 are conjugated by the sheet or strip 1. The invention therefore consists of analyzing the geometric aberration caused by the sheet or strip 1 which behaves like a deforming mirror. When the curvature of the deformation is small, the deformed image 22 of the light source 2 moves while deforming very little. Taking into account the laws of geometric optics, there may be doublings. Figure 5 shows an example where the deformed image of a linear and straight light source is locally doubled with a fork or branch. When the curvature of the deformation is significant, taking into account the laws of geometric optics, there is an enlargement effect of the deformed image 22 of the light source 2.This results in a distorted image 22 of the light source 2 which becomes larger and less bright. The displacement and distortion of the image of the light source 2 are therefore a consequence of the flatness defects of the sheet or strip 1 and are also a consequence of the displacement of the sheet or strip 1 and its flatness defects. The laws of geometric optics have been known for a long time but solving them for the numerous flatness defects of a sheet is complex.

[0056] The method for controlling the moving sheet or strip comprises the following steps: a) The camera transmits a photograph of the deformed image 22 of the light source 2 to the calculation device, b) a predetermined model M implemented in the calculation device determines the photograph of the reference image 20 of the light source 2, c) the predetermined model M compares the photograph of the deformed image 22 of the light source 2 with the photograph of the reference image 20 of the light source 2, d) the predetermined model M calculates at least one coefficient correlated with the flatness of the sheet or strip 1.

[0057] The predetermined model M uses the characteristic data of the sheet or strip such as the composition of the alloy constituting it, its dimensions, its mechanical properties, its metallurgical state, its speed and the level of traction of the sheet. The predetermined model M can be a database (abacus) obtained previously, for example experimentally and / or numerically. The experimental method consists of comparing the maximum deflection or the local maximum deflections of a sheet or a sample of the strip motionless measured on a marble with the photograph of the deformed image 22 of the light source 2 and with the photograph of the reference image 20 of the light source 2. The predetermined model M can also be obtained by any learning method. The predetermined model can use image analysis tools known to those skilled in the art such as, for example, non-limiting pixelation, skeletonization.

[0058] The comparison made by the predetermined model M consists of analyzing the geometry of the photograph of the distorted image 22 of the distorted light source 2. For non-limiting example, the distortion of a straight line can be a curve. For non-limiting example, the distortion of a shape such as a circle can be an ellipse.

[0059] The photograph can also be analyzed by the predetermined grayscale model for each pixel of the photograph transmitted by the camera 3. Analyzing in grayscale is advantageous because it allows the identification of flatness defects that cannot be geometrically quantified. For example, if the light source is a straight line, if the image distorted by the flatness defect of the sheet or strip is an arc of a parabola lying in the plane defined by the camera 3 and the reference image 20 of the light source 2, then the distorted image 22 of the light source 2 is photographed by the camera 3 only as a line that does not appear geometrically distorted.However, as each point of the distorted image 22 of the light source 2 has moved relative to the reference image 20 of the light source 2 by approaching or moving away from the camera 3, the pixels of the photograph transmitted by the camera 3 will be more or less illuminated and will appear more or less gray. The gray level analysis of the image makes it possible to check for flatness defects in the strip or sheet 1. The term gray does not limit this method to an analysis of a photograph obtained with white light. The term gray must be interpreted as the variation in the illumination of the pixels of the camera 3.

[0060] The maximum deflection of a sheet is defined by EN 485-3 or EN485-4. Considering said standards, the maximum deflection is correlated to the flatness. The local maximum deflections are defined by dividing the sheet strip into a plurality of sectors substantially parallel to the direction of movement X. The local maximum deflection on a sector is the maximum of the deflection on said sector. Similar to the local flatness, a local flatness for each sector can also be calculated. Considering said standards, the maximum deflection is correlated to the flatness of sheet or strip 1 and the local maximum deflections are correlated to the local flatnesses of sheet or strip 1. Monitoring the local flatness of the sheet or strip is advantageous because it allows for a more detailed analysis of flatness defects in sheet or strip 1.

[0061] In a preferred embodiment, the light source 2 is substantially linear and straight in shape. The predetermined model M will therefore analyze the photograph of the deformed image 22 of the light source 2 as a deformed line 23 by image analysis methods known to those skilled in the art, such as, for example, non-limiting pixelation, skeletonization. The predetermined model M will therefore analyze the photograph of the reference image 20 of the light source 2 as a reference line 21 by image analysis methods known to those skilled in the art, such as, for example, non-limiting pixelation, skeletonization. The deformed line 23 can be split taking into account the rules of geometric optics as the example in FIG. 5 shows. In this preferred embodiment,steps b to d of the method according to the invention comprise the following successive steps: i) the predetermined model M analyzes the photograph of the deformed image 22 of the light source 2 as a deformed line 23, ii) a reference line 21 is preferentially estimated by the linear regression of a deformed line 23, iii) a distance to the reference line 21 is obtained by calculating the average of the Euclidean distance from each point of the deformed line 23 to the reference line 21, and / or the following successive steps: iv) the predetermined model M analyzes the photograph of the deformed image 22 of the light source 2 as a deformed line 23, iv) the deformed line 23 is segmented into a predetermined plurality of elements, v) vi) a segmented reference line 24 is calculated by the linear regression of each element of the segmented deformed line 23, vi) vii) for at least one element of the deformed line 23 segmented,a distance to the segmented reference line 24 is obtained by calculating the average of the Euclidean distance from each point of the element of the deformed line 23 to the segmented reference line 24.,

[0062] In this preferred embodiment, the estimation of the reference line 21 by the linear regression of the deformed line 23 is advantageous because this simplifies the determination of the reference line 21. Figure 7 illustrates the principle of step iii of calculating the average of the Euclidean distance between the reference line 21 and the deformed line 23. This average is a coefficient correlated with the flatness of the sheet or strip 1. The predetermined model M can also calculate the flatness of the sheet or strip 1. It uses the characteristic data of the sheet or strip 1 such as its composition, dimensions, mechanical properties, metallurgical state, speed and the level of traction of the sheet. The predetermined model M can be a database (abacus) obtained previously, for example experimentally and / or numerically. The experimental method consists of comparing the maximum deflection or local maximum deflections of a stationary sheet or strip 1 measured on a marble with the average of the Euclidean distance between the reference line 21 and the deformed line 23. The predetermined model M can also be obtained by any learning method.

[0063] Figure 8 illustrates the principle of steps iv to vii. The deformed line 23 is segmented into a predetermined plurality of elements. This segmentation is predetermined according to the sheet or strip 1 and the usual defects expected by the experience of a person skilled in the art. This segmentation can be obtained by segmenting the deformed line 23 into segments of predetermined relative length. In the case where there is a fork or a branch, for example that visible in Figure 5, this fork or branch is treated as a cusp point on a curve to determine the relative lengths of the segments of the segmented deformed line 23. Another equivalent method consists of using the reference line 21 and segmenting it into a predetermined plurality of segments. Then the deformed line 23 is segmented into a plurality of elements by the projection along the direction of movement X of the segmented reference line 21.Then a segmented reference line 24 is calculated by the linear regression of each element of the segmented deformed line 23. It should be noted that the segmented reference line 24 is not necessarily continuous. In Figure 8, the direction of movement X is represented several times to represent the segmentation of the reference line 21 which is projected onto the deformed line 23 to obtain the segmented deformed line 23. The average of the Euclidean distance of all the points of a segment of the segmented deformed line 23 to the segmented reference line 24 is a coefficient correlated with the local flatness of the sheet or strip 1. These methods of calculating the segmented reference line 24 are advantageous because they allow focusing on the local flatness defect. Similar to the above, the predetermined model M can also calculate the local flatnesses of the sheet or strip 1.

[0064] Preferably, the camera 3 transmits, according to a predetermined period, a photograph of the deformed image 22 of the light source 2 to the calculation device to control the flatness along the length of the sheet or strip 1. Preferably, at least one coefficient correlated with the flatness of the sheet or strip 1 is an average over a plurality predetermined sliding photographs of the coefficients correlated to the flatness calculated for each photograph. Calculating at least one coefficient correlated to the flatness on an average over a plurality of sliding images consists of calculating the average of the at least one coefficient correlated to the flatness on each of the plurality of photographs of the distorted image 22 of the light source 2 previously transmitted by the camera 3. Calculating these averages is advantageous because it makes it possible to filter the noise linked to the predetermined period of transmission by the camera 3.

[0065] In a preferred embodiment, the method according to the invention further comprises step e: the predetermined model M compares the photograph of the distorted image 22 of the light source 2 with the photographs of the distorted images 22 of the light source 2 which were previously transmitted by the camera 3. This comparison is advantageous because it makes it possible to identify repetitive flatness defects of the sheet or strip 1. Two flatness defects are repetitive if the photographs of their distorted image 22 of the light source 2 can be superimposed with a margin of error of less than 5%, preferably 2% of the width of the sheet or strip 1. Identifying repetitive defects in the length of the sheet or strip is advantageous in particular for rolled products because it makes it possible to efficiently search for the rolling mill rolls or cylinders or machines which may be the cause of said defect.Considering the length of a strip, this is much more advantageous on a strip whose length can be greater than 100 m while the longest known sheets, sheets for airliner wings, are not more than 36 m long. Advantageously, this comparison and identification can be made with the deformed line.

[0066] The advantage of the device and method is to allow real-time control of the flatness of the sheet or strip 1 without stopping production to measure the flatness on a marble, which also makes it possible to avoid cutting the strip to take a sample to bring it to a marble. The advantage of the device and method is that they can be installed in a heavy industry metallurgical product workshop whose temperature and humidity are not regulated and depend on the climate and whose atmosphere may contain dust. The device and method do not require operating conditions that can only be obtained in a laboratory. The other advantage of the device and method is that it is simpler than the prior art, in particular because the predetermined model M can be implemented by the ibavision and Halcon library. The device uses less equipment or less expensive equipment.

Claims

CLAIMS 1) Device for checking the flatness of a metal sheet or strip (1) moving in a direction X, comprising: a) A light source (2) illuminating a field of observation (4) of the sheet or strip (1) in a substantially homogeneous manner, b) A camera (3), which is a 2D digital camera and which is arranged to record in the field of observation (4) at least one photograph of a deformed image (22) of the light source (2) formed by the sheet or strip (1), c) A calculation device for comparing at least one photograph of the deformed image (22) of the light source (2) with a photograph of a reference image (20) of the light source (2) and for continuously calculating at least one coefficient correlated with the flatness of the sheet or strip (1) 2) Flatness control device according to claim 1 characterized in that the light source (2) is of substantially linear and straight shape, preferably substantially perpendicular to the direction of movement X. 3) Flatness control device according to claim 1 or 2 characterized in that the light source (2) is in a plane substantially parallel to the plane of the sheet or strip (1). 4) Device according to one of claims 1 to 3 characterized in that the light source (2) consists of light-emitting diodes (LEDs). 5) Flatness control device according to one of claims 1 to 4, characterized in that the camera (3) is substantially centered relative to the field of observation (4). 6) Flatness control device according to one of claims 1 to 5 characterized in that the images of the ends of the light source (2) in the perpendicular direction Y are not in the field of observation (4), and preferably in that the images of the ends of the light source (2) in the perpendicular direction Y are not visible by the camera 3. 7) Device for checking the flatness of a strip according to one of claims 1 to 6, characterized in that the device comprises referencing means (12, 13, 14, 15) which are preferably spaced apart by at least the width of the sheet or strip (1). ) Sheet metal or strip production machine comprising the device according to one of claims 1 to 7. ) Method for checking flatness with the device according to any one of claims 1 to 7 comprising the following successive steps: a) The camera (3) transmits a photograph of the deformed image (22) of the light source (2) to the calculation device, b) a predetermined model M implemented in the calculation device determines the photograph of the reference image (20) of the light source (2), c) the predetermined model M compares the photograph of the deformed image (22) of the light source (2) with the photograph of the reference image (20) of the light source (2),d) the predetermined model M calculates at least one coefficient correlated with the flatness of the sheet or strip (1). 0) Method according to claim 9 with the device whose light source (2) is of substantially linear and straight shape characterized in that steps b to d comprise the following successive steps: i) the predetermined model M analyzes the photograph of the deformed image (22) of the light source (2) as a deformed line (23), ii) a reference line (21) is preferentially estimated by the linear regression of the deformed line (23), iii) a distance to the reference line (21) is obtained by calculating the average of the Euclidean distance of each point of the deformed line (23) to the reference line (21), and / or the following successive steps iv) the predetermined model M analyzes the photograph of the deformed image (22) of the light source (2) as a deformed line (23),v) the deformed line (23) is segmented into a plurality of elements, vi) a segmented reference line (24) is calculated by the linear regression of each segment of the deformed line (23), vii) for at least one element of the segmented deformed line (23), a distance to the segmented reference line (24) is obtained by calculating the average of the Euclidean distance from each point of the element of the deformed line (23) to the segmented reference line (24). 11) Continuous flatness control method according to one of claims 9 or 10, characterized in that the camera (3) transmits, over a predetermined period, a photograph of the deformed image (22) of the light source (2) to the calculation device. 12) Continuous control method according to claim 11 characterized in that at least one coefficient correlated with the flatness of the sheet or strip (1) is an average over a predetermined plurality of sliding photographs. 13) Method according to one of claims 11 or 12 further comprising step e) the predetermined model M compares the photograph of the distorted image (22) of the light source (2) with the photographs of the distorted image (22) of the light source (2) which were previously transmitted by the camera 3. 14) Method according to one of claims 12 characterized in that the predetermined model M identifies the repetitive flatness defects.

Images