A method, device and storage medium for detecting scratches on steel plates

Abstract

The present invention provides a method for detecting scratches on a steel plate, a detection device and a computer-readable storage medium, and belongs to the field of visual measurement technology. The method comprises: obtaining a grid point set corresponding to the target defective surface of the steel plate to be measured, and obtaining the coordinate value of each grid point in the grid point set; performing curve fitting on the coordinate values ​​of all grid points in the grid point set to obtain a fitting curve; calculating the deviation between the coordinate value of each grid point in the grid point set and the fitting curve; and determining whether the target defective surface is a scratch according to the deviation. This method can at least solve the problems existing in the prior art of high labor intensity, easy to cause missed detection, inability to adapt to the production environment of high-speed units, and low detection accuracy of manual visual steel plate detection, and is suitable for scenes of visual measurement and steel plate detection.

Description

Technical Field

[0001] The present invention relates to the field of visual measurement technology, and in particular to a steel plate scratch detection method, a detection device and a computer-readable storage medium. Background Art

[0002] Steel plate surface inspection is mainly used to detect discontinuity defects on the surface of steel plates (such as pits, scratches, flaws, etc.). The detection methods include manual experience detection method, non-destructive testing technology based on electromagnetic induction and ultrasound.

[0003] At present, the detection method of large steel plates mainly uses manual visual methods. Due to the requirements of production rhythm, the manual visual method cannot detect the bottom surface of the steel plate, so it cannot fully cover the upper and lower surfaces of the tested steel plate; and the manual visual detection method cannot perform numerical measurement of the discontinuity of the steel plate surface. It is necessary to increase the process and use special instruments to measure the numerical value before performing steel plate grading operations. Therefore, the current detection of large steel plates has problems such as high labor intensity, easy to cause missed detection, inability to adapt to the production environment of high-speed units, and low detection accuracy. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a steel plate scratch detection method, a detection device and a computer-readable storage medium in response to the above-mentioned deficiencies in the prior art, so as to at least solve the problems existing in the related art such as high labor intensity, easy missed detection, inability to adapt to the production environment of high-speed units, and low detection accuracy of manual visual steel plate detection.

[0005] In a first aspect, the present invention provides a method for detecting scratches on a steel plate, comprising: obtaining a grid point set corresponding to a target defective surface of the steel plate to be tested, and obtaining a coordinate value of each grid point in the grid point set; performing curve fitting on the coordinate values ​​of all grid points in the grid point set to obtain a fitting curve; calculating the deviation between the coordinate values ​​of each grid point in the grid point set and the fitting curve; and determining whether the target defective surface is a scratch based on the deviation.

[0006] Preferably, after acquiring the coordinate value of each grid point in the grid point set and before performing curve fitting on the coordinate values ​​of all grid points in the grid point set, the method for detecting scratches on a steel plate further includes: arranging all grid points in the grid point set in ascending order of the horizontal coordinate values ​​to obtain a horizontal coordinate sequence and a vertical coordinate sequence.

[0007] Preferably, the curve fitting of the coordinate values ​​of all grid points in the grid point set to obtain the fitting curve specifically includes: using the polyfit function to curve fit the horizontal coordinate sequence and the vertical coordinate sequence to obtain the coefficients of the fitting curve; using the polval function to calculate the coefficients of the fitting curve and the horizontal coordinate sequence to obtain the fitting curve.

[0008] Preferably, the calculating the deviation between the coordinate value of each grid point in the grid point set and the fitting curve specifically comprises: calculating the deviation between the coordinate value of each grid point in the grid point set and the fitting curve according to the following formula:

[0009]

[0010] Among them, l(k) is the total number of grid points in the grid point set, y i is the ordinate value of each grid point, y_p i is the ordinate value of each point on the fitting curve.

[0011] Preferably, the determining whether the target defect surface is a scratch based on the deviation specifically includes: judging whether the deviation is less than or equal to the square of the step length of the measured steel plate; in response to the deviation being less than or equal to the square of the step length of the measured steel plate, determining that the target defect surface is a scratch, the step length being the step length of the steel plate in the longitudinal axis direction at each scanning moment.

[0012] Preferably, the coordinate values ​​of the grid points are recorded as Among them, T is the total number of scanning moments, A is the total number of sampling points on the upper surface or lower surface of the measured steel plate at each scanning moment, and the method of obtaining the grid point set corresponding to the target defect surface of the measured steel plate specifically includes: dividing the measured steel plate into T·A grids; obtaining a set of concave defect points of the measured steel plate, and obtaining the grids corresponding to each concave defect point in the set of concave defect points, wherein the grids are marked as (t, i); storing the grids corresponding to all the concave defect points in the first queue; extracting the grids in the first queue one by one as the central grid, and traversing other grids adjacent to the central grid until the grids with adjacent relationships in the first queue are connected in series to form a second queue, wherein one second queue corresponds to one defect surface, and the second queue is used to represent the defect surface to which a group of concave defect points with adjacent relationships belong; extracting all the grids in the second queue corresponding to the target defect surface as the grid point set corresponding to the target defect surface.

[0013] Preferably, the step of obtaining the set of concave defect points of the steel plate under test specifically includes: obtaining the coordinate data of the sampling points on the upper surface and the lower surface of the steel plate under test; Where p = u represents the upper surface of the steel plate under test, and p = d represents the lower surface of the steel plate under test. The steel plate thickness corresponding to all sampling points is calculated according to the following formula:

[0014]

[0015] in, is the Z-axis data in the coordinate data of the sampling point on the surface of the steel plate being tested, is the Z-axis data in the coordinate data of the sampling points on the lower surface of the tested steel plate; according to the difference between the steel plate thickness corresponding to the sampling points and the target thickness, determine whether each sampling point among all the sampling points is a concave defect point one by one to obtain the determination result of each sampling point; obtain the set of sampling points whose determination results are concave defect points to obtain the set of concave defect points of the tested steel plate.

[0016] Preferably, the step of obtaining the grid corresponding to each concave defect point in the concave defect point set specifically includes: The t value and i value of each concave defect point in the concave defect point set are respectively obtained; and the grid corresponding to each concave defect point is respectively determined according to the t value and i value of each concave defect point.

[0017] In a second aspect, the present invention further provides a device for detecting scratches on a steel plate, comprising an acquisition module, a fitting module, a calculation module and a determination module.

[0018] The acquisition module is used to obtain a grid point set corresponding to the target defect surface of the steel plate under test, and to obtain the coordinate value of each grid point in the grid point set.

[0019] The fitting module is connected to the acquisition module and is used to perform curve fitting on the coordinate values ​​of all grid points in the grid point set to obtain a fitting curve.

[0020] The calculation module is connected to the acquisition module and the fitting module to calculate the deviation between the coordinate value of each grid point in the grid point set and the fitting curve.

[0021] The determination module is connected to the calculation module and is used to determine whether the target defect surface is a scratch according to the deviation.

[0022] Preferably, the determination module includes a determination unit and a determination unit. The determination unit is used to determine whether the deviation is less than or equal to the square of the step length of the steel plate being measured. The determination unit is connected to the determination unit and is used to determine that the target defect surface is a scratch in response to the deviation being less than or equal to the square of the step length of the steel plate being measured, and the step length is the step length of the steel plate in the longitudinal direction at each scanning moment.

[0023] In a third aspect, the present invention further provides a detection device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to implement the method for detecting steel plate scratches as described in the first aspect.

[0024] In a fourth aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for detecting scratches on a steel plate as described in the first aspect is implemented.

[0025] The steel plate scratch detection method, detection device and computer-readable storage medium provided by the present invention can use conventional visual equipment (such as a camera) to obtain the sampling point data of the steel plate to be tested, obtain a fitting curve according to the coordinate data of the grid points corresponding to the target defect surface in the XOY plane, and then determine whether the target defect surface is a steel plate scratch according to the deviation between the coordinate data of each grid point in the XOY plane and the fitting curve. Compared with the manual visual method, it has a higher degree of automation, comprehensive detection, and is suitable for the production environment of high-speed units, and the steel plate scratch determination based on high-precision grid point coordinates has the characteristics of high detection accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 A schematic flow chart of a method for detecting scratches on a steel plate according to Embodiment 1 of the present invention;

[0027] Figure 2 A schematic diagram of a grid of a steel plate according to Example 1 of the present invention;

[0028] Figure 3 A schematic diagram of a defect point grid adjacent to each other in embodiment 1 of the present invention;

[0029] Figure 4 This is a structural schematic diagram of a steel plate detection system according to Embodiment 1 of the present invention;

[0030] Figure 5 A schematic diagram of a world coordinate system according to Embodiment 1 of the present invention;

[0031] Figure 6 This is a schematic structural diagram of a steel plate scratch detection device according to Embodiment 2 of the present invention;

[0032] Figure 7 This is a schematic diagram of the structure of a detection device according to Example 3 of the present invention. DETAILED DESCRIPTION

[0033] In order to enable those skilled in the art to better understand the technical solution of the present invention, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0034] It should be understood that the specific embodiments and drawings described herein are only used to explain the present invention rather than to limit the present invention.

[0035] It can be understood that, in the absence of conflict, the various embodiments of the present invention and the various features in the embodiments can be combined with each other.

[0036] It can be understood that, for the convenience of description, the drawings of the present invention only show the parts related to the present invention, while the parts irrelevant to the present invention are not shown in the drawings.

[0037] It can be understood that each unit and module involved in the embodiments of the present invention may correspond to only one physical structure, or may be composed of multiple physical structures, or multiple units and modules may be integrated into one physical structure.

[0038] It can be understood that, without conflict, the functions and steps marked in the flowcharts and block diagrams of the present invention may occur in an order different from that marked in the drawings.

[0039] It is understood that the flowcharts and block diagrams of the present invention illustrate the possible architectures, functions, and operations of the systems, devices, equipment, and methods according to the various embodiments of the present invention. Each box in the flowchart or block diagram may represent a unit, module, program segment, or code, which contains executable instructions for implementing the specified functions. Moreover, each box or combination of boxes in the block diagram and flowchart may be implemented by a hardware-based system that implements the specified functions, or may be implemented by a combination of hardware and computer instructions.

[0040] It can be understood that the units and modules involved in the embodiments of the present invention can be implemented by software or hardware. For example, the units and modules can be located in a processor.

[0041] Embodiment 1:

[0042] like Figure 1 As shown, this embodiment provides a method for detecting scratches on a steel plate, comprising:

[0043] Step 101, obtaining a grid point set corresponding to a target defect surface of a steel plate under test, and obtaining a coordinate value of each grid point in the grid point set.

[0044] In this embodiment, the sampling point data of the steel plate to be tested is obtained based on conventional visual equipment (such as a camera), and the steel plate to be tested is divided into multiple grids. According to the correspondence between the sampling points and the grids and the adjacent relationship between the grids, the defect surface is obtained, and then it is determined whether the defect surface is a scratch on the steel plate.

[0045] Optionally, obtaining a grid point set corresponding to a target defective surface of the steel plate under test specifically includes steps S11 to S15:

[0046] S11, such as Figure 2 As shown, the steel plate to be tested is divided into T·A grids, where T is the total number of scanning moments, and A is the total number of sampling points on the upper or lower surface of the steel plate to be tested at each scanning moment.

[0047] S12, obtaining a set of concave defect points of the tested steel plate, and obtaining a grid corresponding to each concave defect point in the set of concave defect points, wherein the grid is marked as (t, i).

[0048] S13, storing the grids corresponding to all the concave defect points into a first queue (such as queue J).

[0049] S14, extracting the grids in the first queue one by one as the central grid, and traversing other grids adjacent to the central grid, until the grids with adjacent relationships in the first queue are connected in series to form a second queue (such as queue R(k)), wherein one second queue corresponds to one defect surface, and the second queue is used to represent the defect surface to which a group of concave defect points with adjacent relationships belong.

[0050] In this embodiment, continuous concave defect points can constitute a concave defect surface. In order to improve the detection accuracy of the concave defect surface, the concave defect points are uniquely identified by grids. In addition, the grids corresponding to multiple groups of continuous concave defect points are identified one by one in combination with the adjacent relationship of the grids, and a group of continuous grids are formed into a second queue. One second queue corresponds to one concave defect surface, so that the steel plate defect surface can be detected more simply and accurately. In this embodiment, the first queue is queue J, the second queue is queue R(k), where k is a positive number, and Figure 3 , an example of a method for obtaining a defect surface to which each concave defect point in a concave defect point set belongs to, so as to obtain a defect surface set, includes: Figure 3As shown, the grid with background color represents the grid corresponding to the concave defect point, and the grid without background color represents the grid corresponding to the sampling point without concave defect. The (4,3) grid is extracted from queue J for the first time and put into queues Q(1) and R(1), and the (4,3) grid is deleted from queue J. The (4,3) grid is taken as the center grid, and a circle of grids adjacent to it is traversed. The number of grids in a circle is eight, and the (4,3) grid is deleted from queue Q(1). The (4,4) grid and the (5,4) grid are put into queues Q(1) and R(1) in turn, and the (4,4) grid and the (5,4) grid are deleted from queue J, so that the queue J stores the set of defect points that are not assigned to the defect surface. Among them, R(k) is used to store the grid set of the k-th defect surface, and Q(k) is used to store the defect points in the k-th defect surface that have not yet been traversed and analyzed for the surrounding grids; the second time, the first grid (4,4) in the queue Q(1) is used as the central grid, and the surrounding grids adjacent to it are traversed, and the (4,4) grid is deleted from the queue Q(1). Since the (4,3) grid and the (5,4) grid are already in the queue R(1), they are skipped and not processed. The (5,5) grid is put into the queues Q(1) and R(1), and the (5,5) grid is deleted from the queue J; the third time, the first grid (5,4) in the queue Q(1) is used as the central grid, and the surrounding grids adjacent to it are traversed, and the (5,4) grid is deleted from the queue Q(1). Since the (4,3) grid and the (5,4) grid are already in the queue R(1), they are skipped and not processed. The (5,5) grid is put into the queues Q(1) and R(1), and the (5,5) grid is deleted from the queue J. The (4,4) grid and the (5,4) grid are already in queue R(1), so they are skipped. The fourth time, the first grid (5,5) in queue Q(1) is taken as the center grid, and the grids around it are traversed. Grid (5,5) is deleted from queue Q(1). Since grids (4,4) and (5,4) are already in queue R(1), they are skipped. The fifth time, queue Q(1) is empty, and the next grid (8,8) is selected from queue J. The above steps are repeated to generate queue R(2). There are 4 grids in queue R(1), namely (4,3), (4,4), (5,4) and (5,5), representing the first defective surface. There are 3 grids in queue R(2), namely (8,8), (9,8) and (9,9), representing the second defective surface.

[0051] S15, extracting all grids in the second queue corresponding to the target defect surface as a grid point set corresponding to the target defect surface.

[0052] It should be noted that the grid points in step 101 refer to the grids in steps S11 to S15.

[0053] Optionally, obtaining the set of concave defect points of the tested steel plate in step S12 specifically includes steps S121 to S124:

[0054] S121, obtaining the coordinate data of the sampling points on the upper and lower surfaces of the steel plate being tested Among them, p=u represents the upper surface of the steel plate under test, and p=d represents the lower surface of the steel plate under test.

[0055] In this embodiment, the Figure 4The steel plate detection system shown obtains the coordinate data of the sampling points on the upper surface and the lower surface of the steel plate to be tested. The steel plate detection system includes an image acquisition device 11, a camera calibration software 12 and a data analysis software 13. The image acquisition device 11 is used to acquire image data of the upper surface and the lower surface of the steel plate to be tested. The camera calibration software 12 is connected to the image acquisition device 11 and is used to perform data preprocessing on the acquired image data. The data analysis software 13 is connected to the camera calibration software 12 and is used to calculate the thickness standard deviation of the steel plate to be tested based on the preprocessed image data, and evaluate the thickness uniformity of the steel plate to be tested based on the thickness standard deviation. Optionally, the image acquisition device includes an array camera and an acquisition signal controller. The array camera is connected to the acquisition signal controller and is used to acquire image data of the upper surface and the lower surface of the steel plate to be tested according to the acquisition signal sent by the acquisition signal controller, and output the first coordinate of the sampling point, wherein the first coordinate is in a coordinate system with the position of the array camera as the origin. The acquisition signal controller is provided with an acquisition signal cycle, and the acquisition signal controller is used to send an acquisition signal to the array camera according to the acquisition signal cycle, wherein the acquisition signal cycle is the ratio of the steel plate step length L (unit: meter) to the steel plate transmission rate S (unit: meter / second), that is, the time interval for the acquisition signal controller to send the acquisition signal is: τ=L / S (unit: second), that is, the array camera scans and collects data once every interval τ. Specifically, the array camera includes two groups of linear scanning cameras. The two groups of linear scanning cameras are used to be fixed at the upper and lower parts of the detection gate, respectively, and respectively maintain a preset shooting distance with the corresponding surface of the steel plate to be tested, wherein the multiple linear scanning cameras in the group jointly cover the width of the steel plate to be tested through a cascade of equipment. A group of three-dimensional (3D) linear scanning cameras are installed at the upper and lower parts of the detection gate (i.e., above and below the steel plate to be tested) to collect image data, and the upper (or lower) cameras are cascaded to cover the entire width of the steel plate to be tested. Each camera covers a straight line range of a certain length, and the six cameras cover the width of the steel plate to be tested through a cascaded device. The number of cameras is determined by the width of the steel plate to be tested (or other objects to be tested) and the scanning range of the camera (for example, when the image data acquisition coverage width of each 3D camera is 0.45 meters and the width of the steel plate to be tested is 4.2 meters, a total of 18 to 20 3D cameras are required to cover the width of the upper and lower surfaces of the entire steel plate to be tested through a cascaded device). When the resolutions of the cascaded cameras are different, according to the different resolutions of the cameras, each camera obtains a set of coordinates of the pixel points of the steel plate to be tested within the coverage range, and outputs the first coordinate of the steel plate to be tested as a coordinate system based on the position of each camera as the origin. The first coordinate is written as: Wherein p=u or p=d, u represents the camera above the upper surface of the steel plate under test, d represents the camera below the lower surface of the steel plate under test, t=0,1,2,...T represents the sampling time, j=0,1,2,...M represents the camera index ID, k=0,1,2,...N represents the index ID of each camera collecting N sampling points each time, all cameras are uniformly controlled by the acquisition signal controller, and when the acquisition signal controller sends an acquisition signal, all cameras start to shoot and collect image data of the steel plate under test at the same time. Optionally, the image acquisition device also includes a rate controller and a roller bed. The rate controller is connected to the roller bed, and the rate controller is provided with a steel plate transmission rate, which is used to control the rotation speed of the roller bed according to the steel plate transmission rate. The roller bed is used to place the steel plate under test and rotate according to the rotation speed to transmit the steel plate under test. In this embodiment, the rate controller and the acquisition signal controller are controlled and coordinated by the rate matching software. The camera calibration software includes a coordinate conversion module. The coordinate conversion module is used to convert the first coordinate of the sampling point output by each camera into the world coordinate system according to the relative position relationship of each camera in the array camera to obtain the converted acquisition data. Preferably, if Figure 5 As shown in the figure, the origin O of the world coordinate system is a corner of the steel plate to be measured (such as Figure 3 One corner of the object to be inspected is the unified coordinate origin O), its X-axis is parallel to the camera array composed of each camera, its Y-axis is parallel to the transmission direction of the steel plate to be inspected, and its Z-axis is perpendicular to the XOY plane. The converted coordinate data is: Where i = j·N+k, j is the camera index, i takes values ​​of 0, 1, 2, ... A, A = M·N+N M Indicates the total number of sampling points on any surface of the steel plate being tested at each scanning moment, N M ≤M represents the number of sampling points of the last camera that photographs any surface of the steel plate under test. The coordinate values ​​of the grid points are recorded as It should be noted that the origin of the world coordinate system is not limited to a corner of the steel plate being measured in this embodiment. By converting the data collected by each camera into the collected data in the same coordinate system, the subsequent calculation of the thickness of the steel plate being measured is facilitated and the accuracy of the calculation result is guaranteed.

[0056] S122, calculate the steel plate thickness corresponding to all sampling points according to the following formula:

[0057]

[0058] in, is the Z-axis data in the coordinate data of the sampling point on the surface of the steel plate being tested, It is the Z-axis data in the coordinate data of the sampling point on the lower surface of the steel plate being tested.

[0059] S123, according to the difference between the steel plate thickness corresponding to the sampling point and the target thickness, determine whether each sampling point among all the sampling points is a concave defect point one by one to obtain the determination result of each sampling point.

[0060] In this embodiment, when the steel plate thickness corresponding to the sampling point is less than the target thickness, the sampling point can be judged as a concave defect point. However, since the steel plate under test may experience mechanical vibration during the detection process, in order to improve the detection accuracy, the first threshold is set to the tolerance thickness within the target thickness range. Specifically, it is determined whether the steel plate thickness corresponding to the sampling point is less than or equal to the difference between the target thickness and the first threshold; in response to the steel plate thickness corresponding to the sampling point being less than or equal to the difference between the target thickness and the first threshold, the sampling point is determined to be a concave defect point. For example, the first threshold is set to Th, and combined with the thickness requirements of steel plate production, when the following conditions are met, the sampling point is determined to be a concave defect point:

[0061]

[0062] If the value of H is 1 cm and the value of Th is 0.2 cm, when the thickness of the sampling point is less than or equal to 0.8 cm, it is judged as a concave defect point.

[0063] S124, obtaining a set of sampling points whose determination results are concave defect points, to obtain a set of concave defect points of the tested steel plate.

[0064] Optionally, the step S12 of obtaining the grid corresponding to each concave defect point in the concave defect point set specifically includes: The t value and i value of each concave defect point in the concave defect point set are respectively obtained; and the grid corresponding to each concave defect point is respectively determined according to the t value and i value of each concave defect point.

[0065] In this embodiment, according to When determining the concave defect point, the scanning time t value corresponding to the steel plate thickness and the index i value of the sampling point are determined. The t value and the i value correspond to the row value and column value of the grid respectively, so the grid position corresponding to the concave defect point is determined. The position of the concave defect point can be located by the grid position, which can improve the accuracy of defect point location.

[0066] Optionally, since the t value and i value are determined, the coordinate data of the sampling point The coordinate values ​​of each grid point in the grid point set can be obtained (i.e. ).

[0067] Optionally, when the cameras on the upper and lower surfaces of the steel plate being tested are Figure 4If the cameras on the upper and lower surfaces of the steel plate are aligned one by one, the coordinates of the grid points are obtained and then step 102 is performed. If the cameras on the upper and lower surfaces of the steel plate are not aligned one by one, the coordinates of the grid points are processed as follows to obtain the center of the steel plate at each defect point:

[0068]

[0069]

[0070] Step 102: Perform curve fitting on the coordinate values ​​of all grid points in the grid point set to obtain a fitting curve.

[0071] Specifically, curve fitting is performed on the coordinate values ​​of all grid points in the grid point set to obtain a fitting curve, including steps S1021 and S1022:

[0072] S1021, using the polyfit function in MATLAB to perform curve fitting on the abscissa sequence and the ordinate sequence to obtain coefficients of the fitting curve.

[0073] S1022, using the polval function in MATLAB to calculate the coefficients and the horizontal coordinate sequence of the fitting curve to obtain the fitting curve.

[0074] In this embodiment, the polyfit function is used to perform polynomial fitting in MATLAB, and its principle is to perform fitting based on the least squares method. When the coordinate values ​​of x are not arranged from small to large, as long as the coordinate points of (x, y) correspond one to one, the polyfit function can be used for fitting.

[0075] Optionally, after obtaining the coordinate value of each grid point in the grid point set and before performing curve fitting on the coordinate values ​​of all grid points in the grid point set, the method for detecting scratches on a steel plate also includes: arranging all grid points in the grid point set in ascending order of horizontal coordinate values ​​to obtain a horizontal coordinate sequence and a vertical coordinate sequence.

[0076] In this embodiment, in order to facilitate calculation and improve calculation efficiency, the grid points are arranged according to Sorting from small to large, the plane coordinate queues of the rearranged grid points are: (x'1, y'1), (x'2, y'2), ... (x' l(k) ,y' l(k) ), where l(k) represents the number of grid points in the grid point set (or the second queue R(k)) of the target defect surface. The sorted horizontal coordinate sequence is: x = [x'1, x'2, ..., x' l(k) ]; the sorted ordinate sequence is: y=[y'1,y'2,...,y' l(k)]; MATLAB function polyfit is used to fit the horizontal and vertical coordinate sequences: C = polyfit (x, y, n); where the matrix C contains the coefficients of each dimension of the fitting curve, and n is 2; since the scratches on the steel plate are usually straight lines or curved lines, n is 2, which means that the highest dimension is quadratic. MATLAB function polyval is used to obtain the vertical coordinate sequence (i.e., fitting curve) corresponding to each x coordinate on the fitting curve: y_p = polval (c, x); c is the coefficient of each dimension in matrix C.

[0077] Step 103, calculating the deviation between the coordinate value of each grid point in the grid point set and the fitting curve.

[0078] Optionally, calculating the deviation between the coordinate value of each grid point in the grid point set and the fitting curve specifically includes: calculating the deviation between the coordinate value of each grid point in the grid point set and the fitting curve according to the following formula:

[0079]

[0080] Among them, l(k) is the total number of grid points in the grid point set, y i is the ordinate value of each grid point, y_p i is the ordinate value of each point on the fitting curve.

[0081] Step 104: determine whether the target defect surface is a scratch based on the deviation.

[0082] Specifically, determining whether the target defect surface is a scratch based on the deviation includes: judging whether the deviation is less than or equal to the square of the step length of the measured steel plate; in response to the deviation being less than or equal to the square of the step length of the measured steel plate, determining that the target defect surface is a scratch, and the step length is the step length of the steel plate in the longitudinal axis direction at each scanning moment.

[0083] In this embodiment, the judgment principle is: a curve fitting is performed on a set of continuous defect points (i.e., a grid point set) of the target defect surface. If all defect points (grid points) are concentrated near the curve, the shape of the target defect surface is a linear scratch. Specifically, when the average deviation between the grid points and the fitting curve does not exceed one grid length (i.e., the step length of the steel plate, in meters), it means that the grid points are all near the fitting curve, and the target defect surface is judged to be a scratch.

[0084] The steel plate scratch detection method of this embodiment can use conventional visual equipment (such as a camera) to obtain the sampling point data of the steel plate to be tested, and obtain a fitting curve based on the coordinate data of the grid points corresponding to the target defect surface in the XOY plane, and then determine whether the target defect surface is a steel plate scratch based on the deviation between the coordinate data of each grid point in the XOY plane and the fitting curve. Compared with the manual visual method, it has a higher degree of automation, comprehensive detection, and is suitable for the production environment of high-speed units. Since continuous concave defect points can constitute a concave defect surface, in order to improve the detection accuracy of the concave defect surface, the concave defect points are uniquely identified by a grid. In addition, the grids corresponding to multiple groups of continuous concave defect points are identified one by one in combination with the adjacent relationship of the grids, and a group of continuous grids constitute a second queue, and one second queue corresponds to a concave defect surface, so that the steel plate defect surface can be detected more simply and accurately. Then, curve fitting is performed according to the grid corresponding to the target defect surface, and it is judged whether the deviation between each grid and the fitting curve is less than the length of a single grid to determine whether the target defect surface is a scratch. This method of judging steel plate scratches based on high-precision grid coordinates has the characteristics of high detection accuracy. And the calculation is based on MATLAB, which is simple and efficient.

[0085] Embodiment 2:

[0086] like Figure 6 As shown, this embodiment provides a steel plate scratch detection device, which includes an acquisition module 61, a fitting module 62, a calculation module 63 and a determination module 64.

[0087] The acquisition module 61 is used to acquire a grid point set corresponding to a target defect surface of the steel plate under test, and to acquire a coordinate value of each grid point in the grid point set.

[0088] The fitting module 62 is connected to the acquisition module 61 and is used to perform curve fitting on the coordinate values ​​of all grid points in the grid point set to obtain a fitting curve.

[0089] The calculation module 63 is connected to the acquisition module 61 and the fitting module 62, and is used to calculate the deviation between the coordinate value of each grid point in the grid point set and the fitting curve.

[0090] The determination module 64 is connected to the calculation module 63 and is used to determine whether the target defect surface is a scratch according to the deviation.

[0091] Optionally, the steel plate scratch detection device further comprises an arrangement module, which is connected to the acquisition module and the fitting module respectively.

[0092] The arrangement module is used to arrange all grid points in the grid point set in ascending order of the horizontal coordinate values ​​to obtain a horizontal coordinate sequence and a vertical coordinate sequence.

[0093] Optionally, the fitting module is used to use the polyfit function to perform curve fitting on the horizontal axis sequence and the vertical axis sequence to obtain coefficients of the fitting curve, and is used to use the polval function to calculate the coefficients of the fitting curve and the horizontal axis sequence to obtain the fitting curve.

[0094] Optionally, the calculation module is used to calculate the deviation between the coordinate value of each grid point in the grid point set and the fitting curve according to the following formula:

[0095]

[0096] Among them, l(k) is the total number of grid points in the grid point set, y i is the ordinate value of each grid point, y_p i is the ordinate value of each point on the fitting curve.

[0097] Optionally, the determination module includes a determination unit and a determination unit.

[0098] The judging unit is used to judge whether the deviation is less than or equal to the square of the step length of the steel plate being tested.

[0099] The determination unit is connected to the judgment unit and is used to determine that the target defect surface is a scratch in response to a deviation less than or equal to the square of the step length of the steel plate being tested, and the step length is the step length of the steel plate in the longitudinal direction at each scanning moment.

[0100] Optionally, the acquisition module includes a first acquisition unit and a second acquisition unit. The coordinate values ​​of the grid points are recorded as Wherein, T is the total number of scanning moments, and A is the total number of sampling points on the upper or lower surface of the tested steel plate at each scanning moment. The first acquisition unit is used to obtain a set of grid points corresponding to the target defect surface of the tested steel plate. Specifically, the first acquisition unit is used to divide the tested steel plate into T·A grids, and obtain a set of concave defect points of the tested steel plate, and obtain grids corresponding to each concave defect point in the set of concave defect points, wherein the grids are marked as (t, i), and is also used to store all grids corresponding to the concave defect points in the first queue, and extract the grids in the first queue one by one as the central grid, and traverse other grids adjacent to the central grid until the grids with adjacent relationships in the first queue are connected in series to form a second queue, wherein one second queue corresponds to one defect surface, and the second queue is used to represent the defect surface to which a group of concave defect points with adjacent relationships belong, and is used to extract all grids in the second queue corresponding to the target defect surface as the set of grid points corresponding to the target defect surface.

[0101] In this embodiment, continuous concave defect points can constitute a concave defect surface. In order to improve the detection accuracy of the concave defect surface, the concave defect points are uniquely identified by grids. In addition, the grids corresponding to multiple groups of continuous concave defect points are identified one by one in combination with the adjacent relationship of the grids, and a group of continuous grids are formed into a second queue. One second queue corresponds to one concave defect surface, so that the steel plate defect surface can be detected more simply and accurately. In this embodiment, the first queue is queue J, the second queue is queue R(k), where k is a positive number, and Figure 3 , an example of a process in which the first acquisition unit is used to acquire the defect surface to which each concave defect point in the concave defect point set belongs, so as to obtain the defect surface set: Figure 3As shown in the figure, the grid with background color represents the grid corresponding to the concave defect point, and the grid without background color represents the grid corresponding to the sampling point without concave defect. The first time is used to extract the (4,3) grid from queue J and put it into queues Q(1) and R(1), and delete the (4,3) grid from queue J. With the (4,3) grid as the center grid, traverse a circle of grids adjacent to it, and the number of grids in a circle is eight. , Delete the (4,3) grid from queue Q(1), put the (4,4) grid and the (5,4) grid into queues Q(1) and R(1) in turn, and delete the (4,4) grid and the (5,4) grid from queue J, so that the queue J stores the defect point set that is not assigned to the defect surface. Among them, R(k) is used to store the grid set of the k-th defect surface, and Q(k) is used to store the defect points in the k-th defect surface that have not yet been traversed and analyzed for the surrounding grids; the second time is used to take the first grid (4,4) in the queue Q(1) as the central grid, traverse the surrounding grids adjacent to it, and delete the (4,4) grid from the queue Q(1). Since the (4,3) grid and the (5,4) grid are already in the queue R(1), they are skipped and not processed. The (5,5) grid is put into the queues Q(1) and R(1), and the (5,5) grid is deleted from the queue J; the third time is used to take the first grid (5,4) in the queue Q(1) as the central grid, traverse the surrounding grids adjacent to it, and delete the (5,4) grid from the queue Q(1). Since the (4,3) grid and the (5,4) grid are already in the queue R(1), they are skipped and not processed. The (5,5) grid is put into the queues Q(1) and R(1), and the (5,5) grid is deleted from the queue J. The (4,4) grid and the (5,4) grid are already in queue R(1), so they are skipped. The fourth time, the first grid (5,5) in queue Q(1) is used as the center grid, and the surrounding grids are traversed. The (5,5) grid is deleted from queue Q(1). Since the (4,4) grid and the (5,4) grid are already in queue R(1), they are skipped. The fifth time, queue Q(1) is empty, and the next (8,8) grid is selected from queue J. The above steps are repeated to generate queue R(2). There are 4 grids in queue R(1), namely (4,3), (4,4), (5,4) and (5,5), representing the first defective surface. There are 3 grids in queue R(2), namely (8,8), (9,8) and (9,9), representing the second defective surface.

[0102] Optionally, the first acquisition unit is used to acquire the coordinate data of the sampling points on the upper surface and the lower surface of the steel plate to be tested. Where p = u represents the upper surface of the steel plate under test, and p = d represents the lower surface of the steel plate under test. The steel plate thickness corresponding to all sampling points is calculated according to the following formula:

[0103]

[0104] in, is the Z-axis data in the coordinate data of the sampling point on the surface of the steel plate being tested, It is the Z-axis data in the coordinate data of the sampling points on the lower surface of the tested steel plate, and according to the difference between the steel plate thickness corresponding to the sampling points and the target thickness, it is determined one by one whether each sampling point in all the sampling points is a concave defect point, so as to obtain the determination result of each sampling point, and to obtain a set of sampling points whose determination results are concave defect points, so as to obtain the set of concave defect points of the tested steel plate.

[0105] In this embodiment, the method used is as follows: Figure 4The steel plate detection system shown obtains the coordinate data of the sampling points on the upper surface and the lower surface of the steel plate to be tested. The steel plate detection system includes an image acquisition device 11, a camera calibration software 12 and a data analysis software 13. The image acquisition device 11 is used to acquire image data of the upper surface and the lower surface of the steel plate to be tested. The camera calibration software 12 is connected to the image acquisition device 11 and is used to perform data preprocessing on the acquired image data. The data analysis software 13 is connected to the camera calibration software 12 and is used to calculate the thickness standard deviation of the steel plate to be tested based on the preprocessed image data, and evaluate the thickness uniformity of the steel plate to be tested based on the thickness standard deviation. Optionally, the image acquisition device includes an array camera and an acquisition signal controller. The array camera is connected to the acquisition signal controller and is used to acquire image data of the upper surface and the lower surface of the steel plate to be tested according to the acquisition signal sent by the acquisition signal controller, and output the first coordinate of the sampling point, wherein the first coordinate is in a coordinate system with the position of the array camera as the origin. The acquisition signal controller is provided with an acquisition signal cycle, and the acquisition signal controller is used to send an acquisition signal to the array camera according to the acquisition signal cycle, wherein the acquisition signal cycle is the ratio of the steel plate step length L (unit: meter) to the steel plate transmission rate S (unit: meter / second), that is, the time interval for the acquisition signal controller to send the acquisition signal is: τ=L / S (unit: second), that is, the array camera scans and collects data once every interval τ. Specifically, the array camera includes two groups of linear scanning cameras. The two groups of linear scanning cameras are used to be fixed at the upper and lower parts of the detection gate, respectively, and respectively maintain a preset shooting distance with the corresponding surface of the steel plate to be tested, wherein the multiple linear scanning cameras in the group jointly cover the width of the steel plate to be tested through a cascade of equipment. A group of three-dimensional (3D) linear scanning cameras are installed at the upper and lower parts of the detection gate (i.e., above and below the steel plate to be tested) to collect image data, and the upper (or lower) cameras are cascaded to cover the entire width of the steel plate to be tested. Each camera covers a straight line range of a certain length, and the six cameras cover the width of the steel plate to be tested through a cascaded device. The number of cameras is determined by the width of the steel plate to be tested (or other objects to be tested) and the scanning range of the camera (for example, when the image data acquisition coverage width of each 3D camera is 0.45 meters and the width of the steel plate to be tested is 4.2 meters, a total of 18 to 20 3D cameras are required to cover the width of the upper and lower surfaces of the entire steel plate to be tested through a cascaded device). When the resolutions of the cascaded cameras are different, according to the different resolutions of the cameras, each camera obtains a set of coordinates of the pixel points of the steel plate to be tested within the coverage range, and outputs the first coordinate of the steel plate to be tested as a coordinate system based on the position of each camera as the origin. The first coordinate is written as: Wherein p=u or p=d, u represents the camera above the upper surface of the steel plate under test, d represents the camera below the lower surface of the steel plate under test, t=0,1,2,...T represents the sampling time, j=0,1,2,...M represents the camera index ID, k=0,1,2,...N represents the index ID of each camera collecting N sampling points each time, all cameras are uniformly controlled by the acquisition signal controller, and when the acquisition signal controller sends an acquisition signal, all cameras start to shoot and collect image data of the steel plate under test at the same time. Optionally, the image acquisition device also includes a rate controller and a roller bed. The rate controller is connected to the roller bed, and the rate controller is provided with a steel plate transmission rate, which is used to control the rotation speed of the roller bed according to the steel plate transmission rate. The roller bed is used to place the steel plate under test and rotate according to the rotation speed to transmit the steel plate under test. In this embodiment, the rate controller and the acquisition signal controller are controlled and coordinated by the rate matching software. The camera calibration software includes a coordinate conversion module. The coordinate conversion module is used to convert the first coordinate of the sampling point output by each camera into the world coordinate system according to the relative position relationship of each camera in the array camera to obtain the converted acquisition data. Preferably, if Figure 5 As shown in the figure, the origin O of the world coordinate system is a corner of the steel plate to be measured (such as Figure 3 One corner of the object to be inspected is the unified coordinate origin O), its X-axis is parallel to the camera array composed of each camera, its Y-axis is parallel to the transmission direction of the steel plate to be inspected, and its Z-axis is perpendicular to the XOY plane. The converted coordinate data is: Where i = j·N+k, j is the camera index, i takes values ​​of 0, 1, 2, ... A, A = M·N+N M Indicates the total number of sampling points on any surface of the steel plate being tested at each scanning moment, N M ≤M represents the number of sampling points of the last camera that photographs any surface of the steel plate under test. It should be noted that the origin of the world coordinate system is not limited to a corner of the steel plate under test in this embodiment. By converting the data collected by each camera into the collected data in the same coordinate system, the subsequent calculation of the thickness of the steel plate under test is facilitated and the accuracy of the calculation result is guaranteed.

[0106] Optionally, the first acquisition unit is further configured to obtain the thickness of the steel plate corresponding to the concave defect point. The t value and i value of each concave defect point in the concave defect point set are respectively obtained, and the grid corresponding to each concave defect point is determined according to the t value and i value of each concave defect point.

[0107] Embodiment 3:

[0108] like Figure 7As shown, this embodiment provides a detection device, including a memory 71 and a processor 72, wherein the memory 71 stores a computer program, and the processor 72 is configured to run the computer program to implement the steel plate scratch detection method as described in Example 1.

[0109] Embodiment 4:

[0110] This embodiment provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the method for detecting scratches on a steel plate as described in Example 1 is implemented.

[0111] The detection device for steel plate scratches of Example 2, the detection device of Example 3 and the computer-readable storage medium of Example 4 are used to obtain the sampling point data of the steel plate to be tested by using conventional visual equipment (such as a camera), and are also used to obtain a fitting curve based on the coordinate data of the grid points corresponding to the target defect surface in the XOY plane, and to determine whether the target defect surface is a steel plate scratch based on the deviation between the coordinate data of each grid point in the XOY plane and the fitting curve. Compared with the manual visual method, it has a higher degree of automation, comprehensive detection, and is suitable for the production environment of high-speed units. Since continuous concave defect points can constitute a concave defect surface, in order to improve the detection accuracy of the concave defect surface, the concave defect points are uniquely identified by a grid. In addition, the grids corresponding to multiple groups of continuous concave defect points are identified one by one in combination with the adjacent relationship of the grids, and a group of continuous grids are formed into a second queue, and one second queue corresponds to a concave defect surface, so that the steel plate defect surface can be detected more simply and accurately. It is used to fit the curve according to the grid corresponding to the target defect surface, and judge whether the deviation between each grid and the fitting curve is less than the length of a single grid, so as to judge whether the target defect surface is a scratch. Therefore, the high-precision grid coordinates used to judge the steel plate scratch have the characteristics of high detection accuracy. And the calculation is based on MATLAB, which is simple and efficient.

[0112] It is to be understood that the above embodiments are merely exemplary embodiments used to illustrate the principles of the present invention, but the present invention is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also considered to be within the scope of protection of the present invention.

Claims

1. A method for detecting scratches on a steel plate, characterized in that: include: Obtaining a grid point set corresponding to the target defect surface of the steel plate under test, and obtaining the coordinate value of each grid point in the grid point set; The coordinate values ​​of the grid points are recorded as Wherein, T is the total number of scanning moments, and A is the total number of sampling points on the upper or lower surface of the steel plate under test at each scanning moment; Perform curve fitting on the coordinate values ​​of all grid points in the grid point set to obtain a fitting curve; Calculate the deviation between the coordinate value of each grid point in the grid point set and the fitting curve; Determine whether the target defect surface is a scratch based on the deviation; The step of obtaining a grid point set corresponding to a target defect surface of the steel plate under test specifically includes: Divide the steel plate to be tested into T·A grids; Obtain a set of concave defect points of the tested steel plate, and obtain a grid corresponding to each concave defect point in the set of concave defect points, wherein the grid is marked as (t, i); The grids corresponding to all the concave defect points are stored in the first queue; Extract the grids in the first queue one by one as the central grid, and traverse other grids adjacent to the central grid until the grids in the first queue with adjacent relationships are connected in series to form a second queue, wherein one second queue corresponds to one defect surface, and the second queue is used to represent the defect surface to which a group of concave defect points with adjacent relationships belong; All grids in the second queue corresponding to the target defect surface are extracted as a grid point set corresponding to the target defect surface.

2. The method for detecting steel plate scratches according to claim 1, characterized in that: After acquiring the coordinate value of each grid point in the grid point set, and before performing curve fitting on the coordinate values ​​of all grid points in the grid point set, the method further includes: Arrange all grid points in the grid point set in ascending order of the horizontal coordinate values ​​to obtain a horizontal coordinate sequence and a vertical coordinate sequence.

3. The method for detecting steel plate scratches according to claim 2, characterized in that: The step of performing curve fitting on the coordinate values ​​of all grid points in the grid point set to obtain a fitting curve specifically includes: The polyfit function is used to perform curve fitting on the abscissa sequence and the ordinate sequence to obtain the coefficients of the fitting curve; The polyval function is used to calculate the coefficients and the abscissa sequence of the fitting curve to obtain the fitting curve.

4. The method for detecting steel plate scratches according to claim 1, characterized in that: The calculating of the deviation between the coordinate value of each grid point in the grid point set and the fitting curve specifically includes: The deviation between the coordinate value of each grid point in the grid point set and the fitting curve is calculated according to the following formula: Among them, l(k) is the total number of grid points in the grid point set, y i is the ordinate value of each grid point, y_p i is the ordinate value of each point on the fitting curve.

5. The method for detecting steel plate scratches according to claim 1, characterized in that: The step of determining whether the target defective surface is a scratch according to the deviation specifically includes: Determine whether the deviation is less than or equal to the square of the step length of the steel plate being tested. In response to the deviation being less than or equal to the square of the step length of the steel plate being tested, the target defect surface is determined to be a scratch, and the step length is the step length of the steel plate in the longitudinal direction at each scanning moment.

6. The method for detecting steel plate scratches according to claim 1, characterized in that: The step of obtaining a set of concave defect points of the tested steel plate specifically includes: Obtain the coordinate data of the sampling points on the upper and lower surfaces of the steel plate being tested Wherein, t=0,1,2,…,T, i=0,1,2,…,A, where T is the total number of scanning moments, A is the total number of sampling points on the upper surface or lower surface of the tested steel plate at each scanning moment, p=u represents the upper surface of the tested steel plate, p=d represents the lower surface of the tested steel plate, two groups of linear scanning cameras are used to collect image data of the upper surface and lower surface of the tested steel plate respectively, and u represents the camera above the upper surface of the tested steel plate, d represents the camera below the lower surface of the tested steel plate, i=j·N+k, j is the camera index, and k represents the index of N sampling points collected by each camera each time; Calculate the steel plate thickness corresponding to all sampling points according to the following formula: in, is the Z-axis data in the coordinate data of the sampling point on the surface of the steel plate being tested, is the Z-axis data in the coordinate data of the sampling point on the lower surface of the steel plate being tested; According to the difference between the steel plate thickness corresponding to the sampling point and the target thickness, determine whether each sampling point among all the sampling points is a concave defect point one by one to obtain the determination result of each sampling point; A set of sampling points whose determination results are concave defect points is obtained to obtain a set of concave defect points of the tested steel plate.

7. The method for detecting steel plate scratches according to claim 1, characterized in that: The obtaining of the grid corresponding to each concave defect point in the concave defect point set specifically includes: According to the thickness of the steel plate corresponding to the concave defect point Respectively obtain the t value and i value of each concave defect point in the concave defect point set; The grid corresponding to each concave defect point is determined according to the t value and i value of each concave defect point.

8. A device for detecting scratches on a steel plate, characterized in that: It includes acquisition module, fitting module, calculation module and judgment module. The acquisition module is used to obtain the grid point set corresponding to the target defect surface of the steel plate under test, and to obtain the coordinate value of each grid point in the grid point set. The coordinate value of the grid point is recorded as Wherein, T is the total number of scanning moments, A is the total number of sampling points on the upper or lower surface of the steel plate under test at each scanning moment, The fitting module is connected to the acquisition module and is used to perform curve fitting on the coordinate values ​​of all grid points in the grid point set to obtain a fitting curve. The calculation module is connected with the acquisition module and the fitting module, and is used to calculate the deviation between the coordinate value of each grid point in the grid point set and the fitting curve. A determination module, connected to the calculation module, for determining whether the target defect surface is a scratch according to the deviation; The acquisition module is specifically used to divide the steel plate to be tested into T·A grids, and obtain a set of concave defect points of the steel plate to be tested, and obtain the grids corresponding to each concave defect point in the set of concave defect points, wherein the grid is marked as (t, i), and the acquisition module is also used to store the grids corresponding to all the concave defect points in a first queue, extract the grids in the first queue one by one as the central grid, and traverse other grids adjacent to the central grid until the grids with adjacent relationships in the first queue are connected in series to form a second queue, wherein one second queue corresponds to one defect surface, and the second queue is used to represent the defect surface to which a group of concave defect points with adjacent relationships belong, and the acquisition module is also used to extract all grids in the second queue corresponding to the target defect surface as a set of grid points corresponding to the target defect surface.

9. The steel plate scratch detection device according to claim 8, characterized in that: The determination module includes a determination unit and a determination unit, The judging unit is used to judge whether the deviation is less than or equal to the square of the step length of the steel plate being tested. The determination unit is connected to the judgment unit and is used to determine that the target defect surface is a scratch in response to a deviation less than or equal to the square of the step length of the steel plate being tested, and the step length is the step length of the steel plate in the longitudinal direction at each scanning moment.

10. A detection device, characterized in that: It comprises a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to implement the method for detecting steel plate scratches as described in any one of claims 1-7.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for detecting steel plate scratches as described in any one of claims 1 to 7 is implemented.

Images