System and method for measuring straight angle between continuous casting blank edge and two side edges of roller way
Through binocular cameras and multi-line lasers, they identify the edge feature points of the continuous casting blank and compare it with the roller reference axis. Combined with the laser rangefinder to detect the edge distance, the problem of the edge offset of the continuous casting blank is solved, and the edge of the continuous casting blank is realized. Efficient and accurate monitoring of the straight angle of the roller edge, timely detection of deviation and taking measures, reducing production costs and safety risks.
Patent Information
- Application Number
- CN202310592056.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-05-24
- Publication Date
- 2024-11-26
AI Technical Summary
During the conveying process, the edges are offset due to roller surface processing errors or uneven wear, which may cause the conveyor track to collide with the edges of the continuous casting billet, increasing production costs and reducing economic benefits. It is difficult to effectively monitor the existing technology. and deal with such offsets.
A system combining a binocular camera and a multi-line laser is used to identify the edge feature points of the continuous casting blank through projected structure light, fit the contour straight line equation, and compare it with the roller reference axis to detect the first included angle in real time; at the same time, laser measurement is performed The distance meter measures the distance between the edge and the roller track, detects the second angle, and optimizes the angle to determine if the straight angle exceeds the safety threshold.
It realizes efficient and accurate monitoring of the straight angle of the edge of the continuous casting billet and the roller edge, timely detection of deviation, and taking safety measures to minimize the losses caused by failure.
Smart Images

Figure CN119022827A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field related to visual ranging and laser ranging, and specifically relates to a system and method for measuring the straightness angle between the edge of a continuous casting billet and the edges on both sides of a roller. Background Art
[0002] Continuous casting billets are products obtained by casting molten steel from a steelmaking furnace through a continuous casting machine. Continuous casting billets need to be transported through the rollers of a conveyor. Due to machining errors, sticking or uneven wear of the roller surface of the conveyor, the diameters of the rollers in the rollers are different, or the position of the material at the transfer point is not correct, etc., the continuous casting billets transported on the rollers may have a certain offset in the horizontal direction from both sides of the rollers. If effective treatment measures are not taken in time, the conveyor track will collide with the edge of the continuous casting billet, causing serious safety accidents, increasing the production costs of the enterprise, and reducing the economic benefits of the enterprise. Summary of the invention
[0003] In view of the above-mentioned defects of the prior art, the present invention provides a system and method for measuring the straightness angle between the edge of the continuous casting billet and the edges on both sides of the roller. By measuring the straightness angle between the edge of the continuous casting billet and the edges on both sides of the roller, it can be determined whether the measured straightness angle is within the safety threshold.
[0004] The technical solution adopted by the present invention to solve its technical problem is:
[0005] A system for measuring the straightness angle of the edge of a continuous casting billet and the edges on both sides of a roller, comprising a roller, wherein cameras are arranged at the upper left and upper right of the roller, the fields of view of the cameras on the left and right sides overlap and form a binocular camera, a multi-line laser is arranged between the cameras on the left and right sides, and is used to project a plurality of parallel line structured lights to the continuous casting billet running on the roller, and the binocular camera is used to identify the edge feature points of the continuous casting billet provided by the multi-line laser in real time and fit the straight line equation of the contour of the continuous casting billet; point light source groups are arranged on the left and right sides of the roller, a center line connecting multiple point light sources in the point light source group on the left side is parallel to the left edge of the roller, and a center line connecting multiple point light sources in the point light source group on the right side is parallel to the right edge of the roller, and the point light source groups on the left and right sides are both in the field of view of the binocular camera, and the binocular camera cooperates with the point light source groups on the left and right sides to calibrate the reference axis of the corresponding edge of the roller; the binocular camera is communicatively connected to a signal input end of a control system, a signal output end of the control system is communicatively connected to an alarm, and the control system is used to calculate and obtain a first angle between the reference axis and the fitted straight line equation of the contour of the continuous casting billet in real time.
[0006] Furthermore, the control system is used to compare the first angle acquired in real time with a safety threshold stored in the control system, and when the first angle exceeds the safety threshold, the control system is used to control the alarm to send out an alarm signal.
[0007] Furthermore, laser rangefinders are provided at the left and right edges of the roller, and the laser rangefinders on the left and right sides are communicatively connected to the signal input terminals of the control system. The laser rangefinder on the left side is used to obtain the distance between the left edge of the continuous casting billet and the left edge of the roller in real time and transmit it to the control system, and the laser rangefinder on the right side is used to obtain the distance between the right edge of the continuous casting billet and the right edge of the roller in real time and transmit it to the control system. The control system is used to detect the second angle between the edge of the continuous casting billet and the corresponding edge of the roller in real time according to the received distance information.
[0008] Furthermore, the control system is used to compare the second angle acquired in real time with a safety threshold stored in the control system, and when the second angle exceeds the safety threshold, the control system is used to control the alarm to send out an alarm signal.
[0009] Furthermore, the control system is used to optimize the first angle and the second angle obtained in real time to obtain an optimized angle. The control system is used to compare the optimized angle with a safety threshold stored in the control system. If the optimized angle exceeds the safety threshold, the control system is used to control the alarm to send an alarm signal.
[0010] Furthermore, the camera on the left and the camera on the right are symmetrically arranged relative to the center line of the roller conveying direction, the point light source group on the left and the point light source group on the right are symmetrically arranged relative to the center line of the roller conveying direction, the laser rangefinder on the left and the laser rangefinder on the right are arranged left and right relative to the center line of the roller conveying direction, the multi-line laser is located directly above the roller, the alarm is located above the multi-line laser, and the cameras on the left and right sides, the point light source groups on the left and right sides, the laser rangefinders on the left and right sides, the multi-line lasers and the alarm are all installed through the mounting beam.
[0011] A method for measuring the straightness angle between the edge of a continuous casting billet and the edges on both sides of a roller, using the above-mentioned system for measuring the straightness angle between the edge of a continuous casting billet and the edges on both sides of a roller for measurement, comprises the following steps:
[0012] S1 includes two parallel sub-steps S1-1 and S1-2:
[0013] S1-1, the binocular camera cooperates with the point light source groups on the left and right sides to calibrate the reference axis of the corresponding edge of the roller, and transmits the calibrated reference axis data of the corresponding edge of the roller to the control system; the multi-line laser projects multiple parallel line structured lights to the continuous casting billet running on the roller, the binocular camera identifies the edge feature points of the continuous casting billet provided by the multi-line laser and fits the continuous casting billet contour straight line equation, and transmits the fitted continuous casting billet contour straight line equation data to the control system; the control system calculates and obtains the first angle between the reference axis and the fitted continuous casting billet contour straight line equation in real time;
[0014] S1-2, at the same time, the laser rangefinder on the left side obtains the distance between the left edge of the continuous casting billet and the left edge of the roller in real time and transmits it to the control system, and the laser rangefinder on the right side obtains the distance between the right edge of the continuous casting billet and the right edge of the roller in real time and transmits it to the control system, and the control system detects the second angle between the edge of the continuous casting billet and the corresponding edge of the roller in real time according to the received distance information;
[0015] S2. The control system optimizes the first angle and the second angle obtained in real time to obtain an optimized angle. The control system compares the optimized angle with a safety threshold stored in the control system. When the optimized angle exceeds the safety threshold, the control system controls the alarm to send an alarm signal.
[0016] Furthermore, in step S1-1: the binocular camera identifies the edge feature points of the continuous casting billet provided by the multi-line laser and fits the straight line equation of the contour of the continuous casting billet. Specifically, the binocular camera collects a binocular image of the continuous casting billet surface with line structured light, wherein the projection of each line structured light on the continuous casting billet intersects with the left edge and the right edge of the continuous casting billet respectively, to obtain a left contour point and a right contour point, and the three-dimensional coordinates of the left contour point and the right contour point in the camera coordinate system are calculated by the binocular camera, and then the three-dimensional coordinates of the left contour point and the right contour point are projected onto the plane where the reference axis is located, to fit the straight line equation of the contour of the continuous casting billet.
[0017] Furthermore, step S1-1 specifically includes the following sub-steps:
[0018] S1-1-1. Calibration of binocular camera
[0019] S1-1-11, automatic exposure processing in camera image acquisition: the direction of exposure adjustment is the direction to maximize the gradient information, and the exposure adjustment will eventually stabilize when the image gradient information is near the maximum value, specifically, extracting clear line structured light features through the automatic exposure method, wherein the automatic exposure method includes updating the camera parameters and updating the exposure value during the exposure adjustment;
[0020] S1-1-1101, Update camera parameters when adjusting exposure
[0021] Update the exposure by adjusting the exposure time and exposure gain, where the camera's exposure is measured by the exposure value EV:
[0022] EV=log(t*G) (1)
[0023] In formula (1), t represents the exposure time and G represents the exposure gain;
[0024] Increasing the exposure gain will bring noise to the image. Generally, the exposure gain is set to a smaller value, and the exposure value is updated by adjusting the exposure time first. When the exposure value update requirement cannot be met by adjusting the exposure time, the exposure gain is adjusted.
[0025] S1-1-1102, Update exposure value
[0026] The image gradient information is approximately a convex function with a maximum value with respect to the exposure value. The exposure value is updated using the golden section method. The image gradient value is f(x), x∈[a,b], and the image gradient value f(x) has a maximum value in the exposure value interval [a,b], and the exposure value corresponding to the maximum value of the image gradient value f(x) is x * ;
[0027] Before automatically adjusting the image exposure, the scene is exposed twice continuously. Specifically, two trial points x1 and x2 are taken on the interval [a, b], and a<x1<x2<b. Then, the image gradient value f(x1) corresponding to the exposure value x1 and the image gradient value f(x2) corresponding to the exposure value x2 are calculated.
[0028] By calculating the image gradient value obtained after exposure, the exposure value range is narrowed. If f(x1) < f(x2), then x * ∈[x1,b], keep [x1,b] and remove [a,x1]; if f(x1)>f(x2), then x * ∈[a,x2], keep [a,x2] and remove [x2,b]; if f(x1)=f(x2), then x * ∈[x1,x2], keep [x1,x2] and remove [a,x1] and [x2,b];
[0029] Then, the left endpoint value of the interval is assigned to a, and the right endpoint value is assigned to b. The exposure value is updated repeatedly using the golden section method, and the exposure value interval of the image is continuously narrowed to obtain the best exposure value.
[0030] Let L k is the length of the interval left after comparing the image gradient value after the image is exposed for the kth time. When the interval is shortened, the length of the interval left each time is ω times the length of the interval left in the previous time. Then
[0031] L k+1 =ωL k (2)
[0032] In actual exposure adjustment, x1 and x2 are placed in symmetrical positions in the interval [a, b], specifically:
[0033] x1-a=b-x2 (3)
[0034] According to the principle of equal-proportional contraction, we can get
[0035] L1=ωL0 (4)
[0036] In formula (4), L0 = ba,
[0037] Then L1=x2-a=b-x1=ω(ba), we can get x1=b-ω(ba), and find x1=a+(1-ω)(ba), x2=a+ω(ba);
[0038] When the interval [x1, b] is retained and two trial points are needed on the retained interval [x1, b] in order to further shorten the interval, the existing point x2 in the interval [x1, b] is used as one of the trial points to reduce the amount of calculation;
[0039] And w 2 +w-1=0, then and then
[0040] EV1=x1=a+0.382(ba) (5)
[0041] EV2=x2=a+0.618(ba) (6)
[0042] In formula (5), EV1 represents the exposure value corresponding to the left test point, and in formula (6), EV2 represents the exposure value corresponding to the right test point;
[0043] S1-1-12. Calibrate the binocular camera to obtain the internal and external parameters of the camera and the transfer matrix between the left and right cameras
[0044] S1-1-1201. Calibrate the binocular camera to obtain the internal and external parameters of the camera
[0045] The calibration plate is freely moved within the field of view of the binocular camera, and k calibration plate images are obtained from different angles, where k ≥ 3. The sub-pixel coordinates of each feature point in the calibration plate image are detected, and the correspondence between the feature points is used to solve the world coordinate system OX in each perspective. W Y W Z W With pixel coordinate system o u -The homography matrix H between uv, taking the plane where the calibration plate is located as the world coordinate system Z W = 0, then the corresponding relationship between the world coordinate system and the pixel coordinate system of the point on the calibration plate plane is:
[0046]
[0047] In formula (7), is the camera's internal parameter matrix, fx represents the scale factor of the x-axis, f y represents the scale factor of the y-axis, [u0,v0] represents the coordinates of the intersection point O1 of the camera optical axis and the image plane in the pixel coordinate system, R is the rotation matrix, which is a 3×3 orthogonal unit matrix, r i represents the i-th column vector of the rotation matrix R, t is the translation matrix, and s is the scale factor;
[0048] make Then formula (7) can be expressed as:
[0049]
[0050] In formula (8),
[0051] By knowing the world coordinates and sub-pixel coordinates of the feature points on the calibration plate, the overdetermined equations are solved using the least squares method and further optimized using the Levenberg-Marquarat algorithm to obtain the H matrix and then obtain the initial solution of the camera's internal and external parameters.
[0052] S1-1-1202. Calibrate the binocular camera to obtain the transfer matrix between the left and right cameras
[0053] The three-dimensional coordinates of the same point in the known space in the camera coordinate system on the left are expressed as P l , the three-dimensional coordinates in the camera coordinate system on the right are expressed as P r , the three-dimensional coordinates in the world coordinate system are expressed as P W , then P l , P r With P W Has the following relationship:
[0054]
[0055] In formula (9), R l is the rotation matrix of the left camera, T l is the translation matrix of the left camera, R r is the rotation matrix of the right camera, T r is the translation matrix of the right camera, R 12 is the rotation matrix from the right camera coordinate system to the left camera coordinate system, t 12 is the translation matrix from the right camera coordinate system to the left camera coordinate system,
[0056] From formula (9), we can get
[0057]
[0058] S1-1-13. Calculate the three-dimensional coordinates of the characteristic points on the calibration plate
[0059] For the pixel coordinates p(u,v) of the intersection of the line structured light in the obtained image and the edge of one side of the calibration plate, it is assumed that the three-dimensional coordinates of the spatial measured point represented by it in the reference world coordinate system are P(X W ,Y W ,Z W ),but
[0060] Z C p=MP (11)
[0061] In formula (11), Z C represents the scale factor, which is related to the distance between the optical center of the camera and the measured point in space, and M represents the projection matrix containing the internal and external parameters of the camera, where
[0062] For the left and right cameras, respectively,
[0063]
[0064] In formula (12), R and t in the extrinsic parameter matrix included in the projection matrix M1 of the left camera are composed of the unit matrix and the zero matrix, and R and t in the extrinsic parameter matrix included in the projection matrix M2 of the right camera are composed of R 12 and t 12 The three-dimensional coordinates P of the corresponding intersection point in the reference world coordinate system are obtained by using the least squares method according to the projection matrix M1 of the left camera and the projection matrix M2 of the right camera. W1 ;
[0065] S1-1-2. Calibrate the reference axis and obtain the roller plane equation where the standard lines on the left and right sides of the roller are located
[0066] The binocular camera is used to capture the point light source groups on the left and right sides, and the pixel coordinates of the geometric center points of each point light source in the corresponding point light source group in the image are solved. According to step S1-1-12, the conversion relationship from the pixel coordinates of the point on the image to the world coordinate system can be obtained. The three-dimensional coordinates (x i ,y i ,z i ), i = 1, 2, 3, ..., N, where the center line of each point light source in the corresponding side point light source group forms the standard line of the corresponding edge of the roller, i.e., the reference axis. The general expression of the roller plane equation that does not pass through the origin in the left camera coordinate system is: Ax + By + Cz + 1 = 0. Let A, B, and C be the optimal parameters of the plane. Substitute the three-dimensional coordinates of each point light source into the expression of the roller plane equation. There is an error e i =Ax i +By i +Czi +1, summing up the squares of all errors gives
[0067]
[0068] Formula (13) is a convex hyperplane in four-dimensional space, and has a unique minimum value, which is also the global minimum value. Solving this set of equations yields the equation of the roller plane where the reference axis is located;
[0069] S1-1-3. Calculate the first angle between the edge of the continuous casting billet and the corresponding side edge of the roller
[0070] S1-1-31, continuous casting slab edge contour point extraction
[0071] By controlling the laser projection angle of the multi-line laser to add artificial features to the continuous casting billet image, the line structured light is broken at the edge of the continuous casting billet, which is convenient for edge point positioning, and the interval positioning method is used to distinguish the line structured light position to separate a single complete line structured light stripe, specifically:
[0072] Select an interval value dis and traverse the image along the y direction of the ROI image: if the gray value I of a pixel point of each column of line structured light is the same as the previous pixel point in the x direction (i-1,n·dis) is less than the laser line low threshold l_thre, and the gray value of the next pixel is I (i+1,n·dis) If it is greater than the laser line high threshold h_thre, the pixel point (i,n·dis) is determined to be the left edge point of the line structured light, and the right edge point is the opposite. The continuous left edge points and right edge points in the same row locate a line structured light stripe, and each row can locate multiple line structured light stripes. The row where the number of located line structured light stripes is equal to the number of image projection line structured light stripes is selected as the reference row, and the left edge point of each line structured light stripe is used to perform linear regression fitting of the line structured light straight line using the least squares method;
[0073] According to the upper edge point and lower edge point of a line structured light, the upper edge point of the line structured light corresponds to the right edge of the continuous casting billet, and the lower edge point of the line structured light corresponds to the left edge of the continuous casting billet. An upper edge distance and a lower edge distance are selected for ROI extraction and splicing to realize the extraction of a single line structured light ROI area. Then, the center point of the line structured light stripe obtained by the grayscale centroid method is optimized. The initial sub-pixel coordinates of each point are solved by the grayscale centroid method in the image coordinate system o-xy. p (x p ,y p ) is one of the points, with C p Centered on r p is the distance range, and the selected range is (2r p +1) points to recalculate the centroid As optimized sub-pixel coordinates:
[0074]
[0075] Finally, the center point of the line structured light stripe with high precision and smoothness is extracted, and then the endpoints on both sides of the line structured light are extracted, and the binocular image is corrected by the epipolar line. The edge points obtained by the continuous casting billet image taken by the left camera and the continuous casting billet image taken by the right camera are matched to ensure that the edge points obtained by the left and right images correspond to the same three-dimensional coordinates, so as to realize the extraction of the edge contour points of the continuous casting billet;
[0076] S1-1-32, Calculation of three-dimensional coordinates of the edge contour points of continuous casting billet
[0077] For the edge contour points extracted from the continuously collected continuous casting billet images, further splicing is required to fit the contour of the entire continuous casting billet. Specifically, the real position of the edge contour point is determined according to the world coordinates of the edge contour point, and the pixel coordinates in the camera image are extracted at the same time. The corresponding contour points in the continuous casting billet edge contour are assigned values to obtain the three-dimensional coordinates of the continuous casting billet edge contour points.
[0078] S1-1-33. Solution for the first angle between the straight line equation of the continuous casting billet contour and the reference axis
[0079] By matching a single set of left and right image pairs, the three-dimensional coordinates of the edge contour points of the continuous casting billet at a certain moment in the camera coordinate system are obtained, and the acquisition interval time is combined with the continuous casting billet travel speed to obtain the complete contour of the continuous casting billet. The three-dimensional coordinates of the edge contour points of the continuous casting billet obtained by a single measurement are all established in the same left camera coordinate system. According to the principle of relative displacement, the camera coordinate systems at different moments are displaced accordingly according to the travel distance, so as to unify the three-dimensional coordinates of the edge contour points of the continuous casting billet to the left camera coordinate system O1-X1Y1Z1 at time t1;
[0080] Assume that at time t, the continuous casting billet travel speed is v t mm / s, the roller is affected by the outside world, causing the speed of the roller to be disturbed to varying degrees, resulting in the continuous casting billet speed not being able to maintain a uniform speed. t Integrate to find the distance d that the continuous casting billet moves on the roller in the direction of the standard line.
[0081]
[0082] In formula (15), t1 is the start time and t2 is the end time;
[0083] Record the acquisition time and interval of two consecutive images, so as to splice the edge contour point set of the continuous casting slab in each image in time sequence to obtain the edge contour point set in a unified coordinate system
[0084] After obtaining the three-dimensional coordinates of the intersection of the edge of the continuous casting billet and the line structure light, project the three-dimensional coordinates onto the plane where the roller is located, and set the coordinates of a point on the edge contour of the continuous casting billet as P(x p ,y p ,z p ), the roller plane equation Ax+By+Cz+1=0 obtained in step S1-1-2, and the corresponding normal vector Get point P(x p ,y p ,z p ) to the plane where the roller is located is:
[0085]
[0086] The intersection of the projection line and the plane where the roller is located is the projection point p(x, y, z) from the edge contour point of the continuous casting billet to the plane where the roller is located. The combined equation Ax+By+Cz+1=0 and equation (16) yield:
[0087]
[0088] The intersection points of all projection lines and the plane where the rollers are located are projected according to the above method to obtain the projection point set {p'(x i ,y i ,z i ),i=1,2,...,N}, and obtain the straight line equation of the continuous casting billet contour by straight line fitting based on the projection point set:
[0089]
[0090] In formula (18), A(x0,y0,z0) is a point on the straight line of the continuous casting billet contour, and the vector is a non-zero vector and is parallel to the continuous casting billet contour line, B(x,y,z) is an arbitrary point on the continuous casting billet contour line, and the vector With vector The components of the vector are proportional, and combined with the roller motion direction vector, the first angle between the straight line equation of the continuous casting billet profile and the reference axis is calculated as
[0091] Further,
[0092] Step S1-2 is specifically as follows:
[0093] Phase laser detection technology modulates the intensity of the emitted laser light wave by modulating the signal, and calculates the distance between the measured target and the laser rangefinder based on the measured phase difference, specifically:
[0094] Assuming that the distance between the laser emission point and the reflection point is x, the speed of the laser in the air is c, and the round-trip time between the laser emission point and the reflection point is t, then:
[0095] t=2x / c (19)
[0096] The period of the modulation signal is T, the frequency is f, and the phase difference between the laser emission point and the received signal at the reflection point is φ, then φ is expressed as:
[0097]
[0098] In formula (20), N is the number of complete cycle waves obtained, and Δφ is the phase of the wave that is less than a complete cycle;
[0099] In summary, the distance between the measured target and the laser rangefinder is:
[0100]
[0101] When the continuous casting billet passes through the laser rangefinder, the laser rangefinder continuously emits laser light waves to measure the distance between the edge of the continuous casting billet and the corresponding side edge of the roller, and transmits it to the control system in real time. The distance values measured by the laser rangefinders on the left and right sides gradually increase or decrease over time. Finally, the second angle between the edge of the continuous casting billet and the corresponding edge of the roller is calculated by the fitting straight line y=k1x+b1 and the reference line y=k2x+b2. The fitting straight line y=k1x+b1 is the relationship between the distance value measured by the laser rangefinder on the corresponding side and the time, and has an approximate straight line relationship with the direction of the corresponding edge of the continuous casting billet, and the reference line y=k2x+b2 is the straight line where the corresponding edge of the roller is located;
[0102] In step S2, the control system optimizes the first angle and the second angle obtained in real time to obtain an optimized angle, and the optimized angle is expressed as
[0103] Compared with the prior art, the beneficial effects of the present invention are:
[0104] In the present invention, a binocular camera cooperates with point light source groups on the left and right sides to calibrate the reference axis of the corresponding edge of the roller, and transmits the calibrated reference axis data of the corresponding edge of the roller to the control system, a multi-line laser projects multiple parallel line structured lights to the continuous casting billet running on the roller, the binocular camera identifies the edge feature points of the continuous casting billet provided by the multi-line laser and fits the straight line equation of the continuous casting billet contour, and transmits the fitted straight line equation data of the continuous casting billet contour to the control system, the control system calculates and obtains the first angle between the reference axis and the fitted straight line equation of the continuous casting billet contour in real time, the control system compares the first angle obtained in real time with the safety threshold stored in the control system, and when the first angle exceeds the safety threshold, the control system controls the alarm to send an alarm signal; in this way, the binocular camera and the multi-line laser can measure the angle between the edge of the continuous casting billet and the corresponding edge of the roller, and can prompt the equipment operator to promptly discover the phenomenon of the continuous casting billet deviation, take reasonable safety measures, and minimize the losses caused by the failure.
[0105] In addition, the left laser rangefinder obtains the distance between the left edge of the continuous casting billet and the left edge of the roller in real time and transmits it to the control system, and the right laser rangefinder obtains the distance between the right edge of the continuous casting billet and the right edge of the roller in real time and transmits it to the control system. The control system detects the second angle between the edge of the continuous casting billet and the corresponding edge of the roller in real time according to the received distance information, and the control system compares the second angle obtained in real time with the safety threshold stored in the control system. When the second angle exceeds the safety threshold, the control system controls the alarm to send an alarm signal; in this way, the angle between the edge of the continuous casting billet and the corresponding edge of the roller can also be measured by the laser rangefinder.
[0106] In the present invention, the control system optimizes the first angle and the second angle obtained in real time to obtain an optimized angle. The control system compares the optimized angle with a safety threshold stored in the control system. When the optimized angle exceeds the safety threshold, the control system controls the alarm to send an alarm signal. In this way, by optimizing the first angle and the second angle, the straightness angle between the edge of the continuous casting billet and the edges on both sides of the roller can be measured more accurately.
[0107] In summary, the measuring method of the present invention is efficient, accurate, and low-cost, and can monitor in real time whether the straight angle between the edge of the continuous casting billet and the edges on both sides of the roller table exceeds the safety threshold. BRIEF DESCRIPTION OF THE DRAWINGS
[0108] Figure 1 It is a schematic diagram of the three-dimensional structure of the system for measuring the straightness angle between the edge of the continuous casting billet and the edges on both sides of the roller table of the present invention;
[0109] Figure 2 This is the control principle diagram of the system for measuring the straightness angle between the edge of the continuous casting billet and the edges on both sides of the roller table;
[0110] Figure 3It is a working flow chart of the system for measuring the straightness angle between the edge of the continuous casting billet and the edges on both sides of the roller table of the present invention;
[0111] Figure 4 Target image renderings taken for different exposure values;
[0112] Figure 5 Schematic diagram of the coordinate system related to the binocular camera;
[0113] Figure 6 It is a schematic diagram of the sub-pixel coordinates of feature points on the calibration plate image;
[0114] Figure 7 Schematic diagram of the spatial relationship of the binocular camera coordinate system;
[0115] Figure 8 Schematic diagram of the line structured light stripes breaking at the edge of the continuous casting billet;
[0116] Fig. 9 It is the ROI image of the line structured light stripe;
[0117] Fig.10 It is a schematic diagram for collecting the edge contour points of the continuous casting billet;
[0118] Fig.11 It is a schematic diagram of projecting the three-dimensional coordinates of the intersection of the edge of the continuous casting billet and the line structured light onto the plane where the roller is located;
[0119] Fig.12 Optimize the schematic diagram for the grayscale centroid method;
[0120] Fig.13 This is a comparison chart of the center point extraction effect of structured light stripes with different brightness;
[0121] Fig.14 A schematic diagram is shown for the laser ranging distance point set.
[0122] Explanation of the reference numerals in the figure: 1. roller table, 2. camera, 3. multi-line laser, 4. point light source, 5. laser rangefinder, 6. mounting beam, 7. continuous casting billet, 8. camera field of view, 9. alarm. DETAILED DESCRIPTION
[0123] The specific embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings. These embodiments are only used to illustrate the present invention, but not to limit the present invention.
[0124] In the description of the present invention, it should be noted that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance.
[0125] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0126] Furthermore, in the description of the present invention, unless otherwise specified, “plurality” means two or more.
[0127] like Figure 1 and 2 As shown, a system for measuring the straightness angle between the edge of a continuous casting billet and the edges on both sides of a roller table comprises a roller table 1, cameras 2 are arranged on the upper left and upper right of the roller table 1, the fields of view of the cameras 2 on the left and right sides overlap and form a binocular camera, a multi-line laser 3 is arranged between the cameras 2 on the left and right sides, and is used to project a plurality of parallel line structured lights to the continuous casting billet 7 running on the roller table 1, and the binocular camera is used to identify the edge feature points of the continuous casting billet 7 provided by the multi-line laser 3 in real time and fit the contour straight line equation of the continuous casting billet 7; point light source groups are arranged on the left and right sides of the roller table 1, the center line of the plurality of point light sources 4 in the left point light source group is parallel to the left edge of the roller table 1, and the center line of the plurality of point light sources 4 in the right point light source group is parallel to the left edge of the roller table 1, and the center line of the plurality of point light sources 4 in the right point light source group is parallel to the left edge of the roller table 1. The center line of the light source 4 is parallel to the right edge of the roller 1, and the point light source groups on the left and right sides are both in the field of view of the binocular camera. The binocular camera and the point light source groups on the left and right sides are used to calibrate the reference axis of the corresponding edge of the roller 1; the binocular camera is communicatively connected to the signal input end of the control system, and the signal output end of the control system is communicatively connected to the alarm 9. The control system is used to calculate and obtain the first angle between the reference axis and the fitted straight line equation of the contour of the continuous casting billet 7 in real time, and the control system is used to compare the first angle obtained in real time with the safety threshold stored in the control system. If the first angle exceeds the safety threshold, the control system is used to control the alarm 9 to send an alarm signal.
[0128] In this way, the angle between the edge of the continuous casting billet 7 and the corresponding edge of the roller 1 can be measured by the binocular camera and the multi-line laser 3, and the equipment operator can be prompted to promptly discover the deviation of the continuous casting billet 7 and take reasonable safety measures to minimize the losses caused by the failure.
[0129] In one embodiment, Figure 1 and 2 As shown, laser rangefinders 5 are provided at the left and right edges of the roller 1. The laser rangefinders 5 on the left and right sides are connected to the signal input end of the control system. The laser rangefinder 5 on the left side is used to obtain the distance between the left edge of the continuous casting billet 7 and the left edge of the roller 1 in real time and transmit it to the control system. The laser rangefinder 5 on the right side is used to obtain the distance between the right edge of the continuous casting billet 7 and the right edge of the roller 1 in real time and transmit it to the control system. The control system is used to detect the second angle between the edge of the continuous casting billet 7 and the corresponding edge of the roller 1 in real time according to the received distance information. The control system is used to compare the second angle obtained in real time with the safety threshold stored in the control system. If the second angle exceeds the safety threshold, the control system is used to control the alarm 9 to send an alarm signal. In this way, the angle between the edge of the continuous casting billet 7 and the corresponding edge of the roller 1 can also be measured by the laser rangefinder 5.
[0130] In one embodiment, the control system is used to optimize the first angle and the second angle obtained in real time to obtain an optimized angle, and the control system is used to compare the optimized angle with the safety threshold stored in the control system. If the optimized angle exceeds the safety threshold, the control system is used to control the alarm 9 to send an alarm signal. In this way, by optimizing the first angle and the second angle, the straight angle between the edge of the continuous casting billet 7 and the edges on both sides of the roller table 1 can be measured more accurately.
[0131] In one embodiment, the left camera 2 and the right camera 2 are symmetrically arranged relative to the center line of the conveying direction of the roller 1, the left point light source group and the right point light source group are symmetrically arranged relative to the center line of the conveying direction of the roller 1, the left laser rangefinder 5 and the right laser rangefinder 5 are arranged left and right relative to the center line of the conveying direction of the roller 1, the multi-line laser 3 is located directly above the roller 1, the alarm 9 is located above the multi-line laser 3, and the left and right cameras 2, the left and right point light source groups, the left and right laser rangefinders 5, the multi-line laser 3 and the alarm 9 are all installed through the mounting beam 6.
[0132] A method for measuring the straightness angle between the edge of a continuous casting billet and the edges on both sides of a roller table, see Figure 3 , the above-mentioned continuous casting billet edge and roller edge straightness angle measurement system is used for measurement, including the following steps:
[0133] S1 includes two parallel sub-steps S1-1 and S1-2:
[0134] S1-1, the binocular camera cooperates with the point light source groups on the left and right sides to calibrate the reference axis of the corresponding edge of the roller 1, and transmits the calibrated reference axis data of the corresponding edge of the roller 1 to the control system; the multi-line laser 3 projects a plurality of parallel line structured lights onto the roller 1, and the binocular camera collects the image of the line structured light on the roller 1 in real time and judges whether the continuous casting billet 7 is transmitted to the field of view 8 of the binocular camera through the height information fed back by the line structured light. When the continuous casting billet 7 is transmitted to the field of view 8 of the binocular camera, the multi-line laser 3 projects a plurality of parallel line structured lights onto the continuous casting billet 7 running on the roller 1, and the binocular camera collects the image of the line structured light on the roller 1 in real time and judges whether the continuous casting billet 7 is transmitted to the field of view 8 of the binocular camera through the height information fed back by the line structured light. A binocular image of the surface of the continuous casting billet 7 of line structured light, wherein the projection of each line structured light on the continuous casting billet 7 intersects with the left edge and the right edge of the continuous casting billet 7 respectively, to obtain a left contour point and a right contour point, and the three-dimensional coordinates of the left contour point and the right contour point in the coordinate system of the camera 2 are calculated by the binocular camera, and then the three-dimensional coordinates of the left contour point and the right contour point are projected to the plane where the reference axis is located, and the contour straight line equation of the continuous casting billet 7 is fitted, and the fitted contour straight line equation data of the continuous casting billet 7 is transmitted to the control system; the control system calculates and obtains the first angle between the reference axis and the fitted contour straight line equation of the continuous casting billet 7 in real time;
[0135] S1-2, at the same time, the left laser rangefinder 5 obtains the distance between the left edge of the continuous casting billet 7 and the left edge of the roller 1 in real time and transmits it to the control system, and the right laser rangefinder 5 obtains the distance between the right edge of the continuous casting billet 7 and the right edge of the roller 1 in real time and transmits it to the control system, and the control system detects the second angle between the edge of the continuous casting billet 7 and the corresponding edge of the roller 1 in real time according to the received distance information;
[0136] S2. The control system optimizes the first angle and the second angle obtained in real time to obtain an optimized angle. The control system compares the optimized angle with a safety threshold stored in the control system. When the optimized angle exceeds the safety threshold, the control system controls the alarm 9 to send out an alarm signal.
[0137] In one embodiment, step S1-1 specifically includes the following sub-steps:
[0138] S1-1-1. Calibration of binocular camera
[0139] S1-1-11. Automatic exposure processing in camera image acquisition: In the actual production environment of the factory, the lighting conditions will continue to change. When the actual lighting conditions are not compatible with the built-in exposure parameters of camera 2, the quality of the image collected by camera 2 will be reduced. In severe cases, overexposure or underexposure will occur, which will have a serious impact on the measured straightness angle between the edge of the continuous casting billet 7 and the edges on both sides of the roller 1. In order to ensure the accuracy of the measured straightness angle between the edge of the continuous casting billet 7 and the edges on both sides of the roller 1, exposure adjustment is required, wherein the direction of exposure adjustment is the direction of maximizing the gradient information. The exposure adjustment will eventually stabilize when the image gradient information is near the maximum value. Specifically, clear line structured light features are extracted through the automatic exposure method. The automatic exposure method finally achieves the following effect. Figure 4 As shown, Figure 4 (a)-(f) show the effect of shooting the target image with different exposure values, where the exposure value increases gradually from small to large. Figure 4 The images in (a)-(b) are too dark, making it difficult to extract the line structured light features. Figure 4 In (f), overexposure occurs, which will seriously affect the measured straight angle between the edge of the continuous casting billet 7 and the edges on both sides of the roller 1, wherein the automatic exposure method includes updating the camera parameters and updating the exposure value during exposure adjustment;
[0140] S1-1-1101, Update camera parameters when adjusting exposure
[0141] The exposure of camera 2 is affected by four parameters: ambient brightness, aperture size, exposure time and exposure gain, and is proportional to the four parameters. Automatic exposure means that when the ambient brightness remains unchanged, the exposure is adjusted by changing the other three parameters so that the image acquired by camera 2 can achieve the best effect. In practical applications, the aperture size of camera 2 is generally difficult to change. The exposure is updated by adjusting the exposure time and exposure gain, where the exposure value EV is used to measure the exposure of camera 2:
[0142] EV=log(t*G) (1)
[0143] In formula (1), t represents the exposure time and G represents the exposure gain;
[0144] Increasing the exposure gain will bring noise to the image. Generally, the exposure gain is set to a smaller value, and the exposure value is updated by adjusting the exposure time first. When the exposure value update requirement cannot be met by adjusting the exposure time, the exposure gain is adjusted.
[0145] S1-1-1102, Update exposure value
[0146] The image gradient information is approximately a convex function with a maximum value with respect to the exposure value. The exposure value is updated using the golden section method. The image gradient value is f(x), x∈[a,b], and the image gradient value f(x) has a maximum value in the exposure value interval [a,b], and the exposure value corresponding to the maximum value of the image gradient value f(x) is x * ;
[0147] Before automatically adjusting the image exposure, the scene is exposed twice continuously. Specifically, two trial points x1 and x2 are taken on the interval [a, b], and a<x1<x2<b. Then, the image gradient value f(x1) corresponding to the exposure value x1 and the image gradient value f(x2) corresponding to the exposure value x2 are calculated.
[0148] By calculating the image gradient value obtained after exposure, the exposure value range is narrowed. If f(x1) < f(x2), then x * ∈[x1,b], keep [x1,b] and remove [a,x1]; if f(x1)>f(x2), then x * ∈[a,x2], keep [a,x2] and remove [x2,b]; if f(x1)=f(x2), then x * ∈[x1,x2], keep [x1,x2] and remove [a,x1] and [x2,b];
[0149] Then, the left endpoint value of the interval is assigned to a, and the right endpoint value is assigned to b. The exposure value is updated repeatedly using the golden section method, and the exposure value interval of the image is continuously narrowed to obtain the best exposure value.
[0150] Let L k is the length of the interval left after comparing the image gradient value after the image is exposed for the kth time. When the interval is shortened, the length of the interval left each time is ω times the length of the interval left in the previous time. Then
[0151] L k+1 =ωL k (2)
[0152] In actual exposure adjustment, in order to better determine which interval of the interval [a, x1] and [x2, b] is discarded first and speed up the iteration speed, the lengths of the intervals [a, x1] and [x2, b] are equal, so that x1 and x2 are in symmetrical positions in the interval [a, b], specifically:
[0153] x1-a=b-x2 (3)
[0154] According to the principle of equal-proportional contraction, we can get
[0155] L1=ωL0 (4)
[0156] In formula (4), L0 = ba,
[0157] Then L1=x2-a=b-x1=ω(ba), we can get x1=b-ω(ba), and find x1=a+(1-ω)(ba), x2=a+ω(ba);
[0158] When the interval [x1, b] is retained and two trial points are needed on the retained interval [x1, b] in order to further shorten the interval, the existing point x2 in the interval [x1, b] is used as one of the trial points to reduce the amount of calculation;
[0159] And w 2 +w-1=0, then and then
[0160] EV1=x1=a+0.382(ba) (5)
[0161] EV2=x2=a+0.618(ba) (6)
[0162] In formula (5), EV1 represents the exposure value corresponding to the left test point, and in formula (6), EV2 represents the exposure value corresponding to the right test point;
[0163] S1-1-12. Calibrate the binocular camera to obtain the internal and external parameters of the camera and the transfer matrix between the left and right cameras
[0164] S1-1-1201. Calibrate the binocular camera to obtain the internal and external parameters of the camera
[0165] Establish the binocular camera related coordinate system, such as Figure 5 As shown, Figure 5 A represents the running direction of the roller 1. The calibration plate is freely moved within the field of view 8 of the binocular camera. K calibration plate images are obtained from different angles, where k ≥ 3. The sub-pixel coordinates of each feature point in the calibration plate image are detected. Figure 6 , using the correspondence between feature points to solve the world coordinate system OX in each perspective W Y W Z W With pixel coordinate system o u -The homography matrix H between uv, taking the plane where the calibration plate is located as the world coordinate system Z W = 0, then the corresponding relationship between the world coordinate system and the pixel coordinate system of the point on the calibration plate plane is:
[0166]
[0167] In formula (7), is the internal parameter matrix of camera 2, f xrepresents the scale factor of the x-axis, f y represents the scale factor of the y-axis, [u0,v0] represents the coordinates of the intersection point O1 of the optical axis of camera 2 and the image plane in the pixel coordinate system, which is the coordinates of the principal point of the image, R is the rotation matrix, which is a 3×3 orthogonal unit matrix, r i represents the i-th column vector of the rotation matrix R, t is the translation matrix, and s is the scale factor;
[0168] make Then Equation 7 can be expressed as:
[0169]
[0170] In formula (8),
[0171] By knowing the world coordinates and sub-pixel coordinates of the feature points on the calibration plate, the overdetermined equations are solved using the least squares method and further optimized using the Levenberg-Marquarat algorithm to obtain the H matrix, and then obtain the initial solution of the internal and external parameters of camera 2;
[0172] Due to the errors in the manufacturing process of camera 2, the final image of camera 2 often has radial distortion and tangential distortion. Generally, there are 5 distortion coefficients of camera 2, namely k1, k2, k3, p1, p2, among which k1, k2, k3 are radial distortion coefficients, and p1, p2 are tangential distortion coefficients. If the distortion coefficients of camera 2 are known, the distorted image can be restored.
[0173] If you need to remove the radial distortion of the image, you can use the following two formulas for correction:
[0174] x'=x(1+k1r 2 +k2r 4 +k3r 6 ) (9)
[0175] y'=y(1+k1r 2 +k2r 4 +k3r 6 ) (10)
[0176] Combining equations (9) and (10), [x, y] T is the distorted point, [x',y'] T is the point after radial distortion, where r 2 =x 2 +y 2 ;
[0177] If you need to remove the tangential distortion of the image, you can use the following two formulas for correction:
[0178] x'=x+2p1xy+p2(r2 +2x 2 ) (11)
[0179] y'=y+p1(r 2 +2y 2 )+2p2xy (12)
[0180] Combining equations (11) and (12), [x, y] T is the distorted point, [x',y'] T is the point after tangential distortion;
[0181] If it is necessary to remove radial distortion and tangential distortion of the image at the same time, then by combining equations (9), (10), (11) and (12), we can get:
[0182] x'=x(1+k1r 2 +k2r 4 +k3r 6 )+2p1xy+p2(r 2 +2x 2 ) (13)
[0183] y'=y(1+k1r 2 +k2r 4 +k3r 6 )+p1(r 2 +2y 2 )+2p2xy (14)
[0184] Combining equations (13) and (14), [x, y] T is the distorted point, [x',y'] T is the point after removing radial distortion and tangential distortion;
[0185] When considering the radial distortion and tangential distortion of camera 2, the lens distortion parameters can be set to zero during initialization to simplify the calculation and allow it to participate in the next step of nonlinear optimization. The nonlinear optimization objective function is as follows:
[0186]
[0187] In formula (15), m ij Point M j The pixel coordinates in image i, m(A,R i ,t i ,M j ,k1,k2,k3,p1,p2) is point M j The projection in image i, R i represents the rotation matrix in image i, t iRepresents the translation matrix in image i. The global optimal solution is obtained through multiple iterations and convergence, and then the internal and external parameters and distortion coefficients of camera 2 are obtained.
[0188] S1-1-1202. Calibrate the binocular camera to obtain the transfer matrix between the left and right cameras
[0189] After the relevant parameters of the two cameras 2 are obtained, the positional relationship between the two cameras 2 needs to be calibrated to obtain the transformation relationship from the right camera 2 to the left camera 2. The schematic diagram of the coordinate system transformation relationship between the left camera 2 and the right camera 2 is as follows: Figure 7 As shown, the three-dimensional coordinates of the same point in the known space in the coordinate system of the left camera 2 are expressed as P l , the three-dimensional coordinates in the coordinate system of camera 2 on the right are expressed as P r , the three-dimensional coordinates in the world coordinate system are expressed as P W , then P l , P r With P W Has the following relationship:
[0190]
[0191] In formula (16), R l is the rotation matrix of the left camera 2, T l is the translation matrix of the left camera 2, R r is the rotation matrix of the right camera 2, T r is the translation matrix of the right camera 2, R 12 is the rotation matrix from the right camera 2 coordinate system to the left camera 2 coordinate system, t 12 is the translation matrix from the right camera 2 coordinate system to the left camera 2 coordinate system,
[0192] From formula (16), we can get
[0193]
[0194] S1-1-13. Calculate the three-dimensional coordinates of the characteristic points on the calibration plate
[0195] For the pixel coordinates p(u,v) of the intersection of the line structured light in the obtained image and the edge of one side of the calibration plate, it is assumed that the three-dimensional coordinates of the spatial measured point represented by it in the reference world coordinate system are P(X W ,Y W ,Z W ),but
[0196] Z C p=MP (18)
[0197] In formula (18), Z Crepresents the scale factor, which is related to the distance between the optical center of camera 2 and the measured point in space. M represents the projection matrix containing the internal and external parameters of camera 2, where
[0198] For the left and right cameras 2, we have
[0199]
[0200] In formula (19), R and t in the external parameter matrix included in the projection matrix M1 of the left camera 2 are composed of the unit matrix and the zero matrix, and R and t in the external parameter matrix included in the projection matrix M2 of the right camera 2 are composed of R 12 and t 12 The three-dimensional coordinates P of the corresponding intersection point in the reference world coordinate system are obtained by using the least squares method according to the projection matrix M1 of the left camera 2 and the projection matrix M2 of the right camera 2. W1 ;
[0201] S1-1-2. Calibrate the reference axis and obtain the roller plane equation where the standard lines on the left and right sides of the roller are located
[0202] The binocular camera is used to capture the point light source groups on the left and right sides, and the pixel coordinates of the geometric center points of each point light source 4 in the corresponding point light source groups on the image are solved. According to step S1-1-12, the conversion relationship from the pixel coordinates of the point on the image to the world coordinate system can be obtained. The three-dimensional coordinates (x i ,y i ,z i ), i = 1, 2, 3, ..., N, where the center line of each point light source 4 in the corresponding side point light source group forms the standard line of the corresponding edge of the roller 1, i.e., the reference axis. The general expression of the plane equation of the roller 1 that does not pass through the origin in the coordinate system of the left camera 2 is: Ax + By + Cz + 1 = 0. Let A, B, and C be the optimal parameters of the plane. Substitute the three-dimensional coordinates of each point light source 4 into the expression of the plane equation of the roller 1. There is an error e i =Ax i +By i +Cz i +1, summing up the squares of all errors gives
[0203]
[0204] Formula (20) is a downward convex hyperplane in four-dimensional space, and has a unique minimum value, which is also the global minimum value. Solving this set of equations yields the plane equation of roller 1 where the reference axis is located;
[0205] S1-1-3. Calculate the first angle between the edge of the continuous casting billet and the corresponding side edge of the roller
[0206] S1-1-31, continuous casting slab edge contour point extraction
[0207] By controlling the laser projection angle of the multi-line laser 3 to add artificial features to the image of the continuous casting billet 7, the line structured light appears at the edge of the continuous casting billet 7 as shown in FIG. Figure 8 The fracture situation shown is convenient for locating edge points, and the interval positioning method is used to distinguish the line structured light position to separate a single complete line structured light stripe, specifically:
[0208] like Fig. 9 As shown, an interval value dis is selected and the image is traversed along the y direction of the ROI image: if the gray value I of a pixel point of each column of line structured light in the previous pixel point in the x direction is (i-1,n·dis) is less than the laser line low threshold l_thre, and the gray value of the next pixel is I (i+1,n·dis) If it is greater than the laser line high threshold h_thre, the pixel point (i,n·dis) is determined to be the left edge point of the line structured light, and the right edge point is the opposite. The continuous left edge points and right edge points in the same row locate a line structured light stripe, and each row can locate multiple line structured light stripes. The row where the number of located line structured light stripes is equal to the number of image projection line structured light stripes is selected as the reference row, and the left edge point of each line structured light stripe is used to perform linear regression fitting of the line structured light straight line using the least squares method;
[0209] According to the determined upper edge point and lower edge point of a certain line structured light straight line, the position of the upper edge point of the line structured light corresponds to the right edge of the continuous casting billet 7, and the position of the lower edge point of the line structured light corresponds to the left edge of the continuous casting billet 7. An upper edge distance and a lower edge distance are selected for ROI extraction and splicing to realize the single line structured light ROI area extraction, see Fig.13 ,Then the center point of the line structured light stripe obtained by the grayscale centroid method is optimized. The optimization principle of the grayscale centroid method is as follows Fig.12 As shown, the initial sub-pixel coordinates of each point are solved by the grayscale centroid method in the image coordinate system o-xy, C p (x p ,y p ) is one of the points, with C p Centered on r p is the distance range, and the selected range is (2r p +1) points to recalculate the centroid As optimized sub-pixel coordinates:
[0210]
[0211] Finally, the center point of the line structured light stripe with high precision and smoothness is extracted, and then the endpoints on both sides of the line structured light are extracted, and the binocular image is subjected to epipolar correction. The edge points obtained by the image of the continuous casting billet 7 taken by the left camera 2 and the image of the continuous casting billet 7 taken by the right camera 2 are matched to ensure that the edge points obtained by the left and right images correspond to the same three-dimensional coordinates, so as to realize the extraction of the edge contour points of the continuous casting billet 7;
[0212] S1-1-32, Calculation of three-dimensional coordinates of the edge contour points of continuous casting billet
[0213] For the edge contour points extracted from the continuously collected images of the continuous casting billet 7, further splicing is required to fit the contour of the entire continuous casting billet 7, specifically: determine the real position of the edge contour point according to the world coordinates of the edge contour point, extract the pixel coordinates in the camera 2 image at the same time, and assign values to the corresponding contour points in the edge contour of the continuous casting billet 7 to obtain the three-dimensional coordinates of the edge contour point of the continuous casting billet 7;
[0214] S1-1-33. Solution for the first angle between the straight line equation of the continuous casting billet contour and the reference axis
[0215] By matching a single set of left and right image pairs, the three-dimensional coordinates of the edge contour point of the continuous casting billet 7 at a certain moment in the coordinate system of the camera 2 are obtained, and the acquisition interval time is combined with the travel speed of the continuous casting billet 7 to calculate, see Fig.10 To obtain the complete outline of the continuous casting billet 7, the three-dimensional coordinates of the edge contour points of the continuous casting billet 7 obtained by a single measurement are all established in the same left camera 2 coordinate system. According to the principle of relative displacement, the camera 2 coordinate system at different times is displaced accordingly according to the travel distance, so as to unify the three-dimensional coordinates of the edge contour points of the continuous casting billet 7 to the left camera 2 coordinate system O1-X1Y1Z1 at time t1;
[0216] Assume that at time t, the continuous casting billet 7 travels at a speed of v t mm / s, the roller 1 is affected by the outside world, causing the speed of the roller 1 to be disturbed to varying degrees, resulting in the continuous casting billet 7 being unable to maintain a uniform speed. t Integrate to find the distance d that the continuous casting billet 7 moves on the roller table 1 in the direction of the standard line.
[0217]
[0218] In formula (22), t1 is the start time and t2 is the end time;
[0219] Record the acquisition time and acquisition interval of two consecutive images, so as to splice the edge contour point set of the continuous casting billet 7 in each image in time sequence to obtain the edge contour point set in a unified coordinate system
[0220] After obtaining the three-dimensional coordinates of the intersection of the edge of the continuous casting billet 7 and the line structure light, the three-dimensional coordinates are projected onto the plane where the roller 1 is located. Fig.11 , let the coordinates of a point on the edge contour of the continuous casting billet 7 be P(x p ,y p ,z p ), the plane equation of roller 1 obtained in step S1-1-2 is Ax+By+Cz+1=0, and the corresponding normal vector Get point P(x p ,y p ,z p ) to the plane where roller 1 is located is:
[0221]
[0222] The intersection of the projection line and the plane where the roller 1 is located is the projection point p(x, y, z) from the edge contour point of the continuous casting billet 7 to the plane where the roller 1 is located. The simultaneous equation Ax+By+Cz+1=0 and equation (23) yield:
[0223]
[0224] The intersection points of all projection lines and the plane where the roller 1 is located are projected according to the above method to obtain the projection point set {p'(x i ,y i ,z i ),i=1,2,...,N}, and obtain the straight line equation of the continuous casting billet 7 by straight line fitting according to the projection point set:
[0225]
[0226] In formula (25), A(x0,y0,z0) is a point on the straight line of the continuous casting billet 7, and the vector is a non-zero vector and is parallel to the straight line of the continuous casting billet 7, B(x, y, z) is an arbitrary point on the straight line of the continuous casting billet 7, and the vector With vector The components of the vector are proportional, and combined with the motion direction vector of the roller 1, the first angle between the straight line equation of the continuous casting billet 7 and the reference axis is calculated as
[0227] Further,
[0228] Step S1-2 is specifically as follows:
[0229] Phase laser detection technology modulates the intensity of the emitted laser light wave by modulating the signal, and calculates the distance between the measured target and the laser rangefinder 5 according to the measured phase difference, specifically:
[0230] Assuming that the distance between the laser emission point and the reflection point is x, the speed of the laser in the air is c, and the round-trip time between the laser emission point and the reflection point is t, then:
[0231] t=2x / c (26)
[0232] The period of the modulation signal is T, the frequency is f, and the phase difference between the laser emission point and the received signal at the reflection point is φ, then φ is expressed as:
[0233]
[0234] In formula (27), N is the number of complete cycle waves obtained, and Δφ is the phase of the wave that is less than a complete cycle;
[0235] In summary, the distance between the measured target and the laser rangefinder 5 is:
[0236]
[0237] When the continuous casting billet 7 passes through the laser distance meter 5, the laser distance meter 5 continuously emits laser light waves to measure the distance between the edge of the continuous casting billet 7 and the corresponding side edge of the roller table 1, and transmits it to the control system in real time. The distance values measured by the laser distance meters 5 on the left and right sides gradually increase or decrease over time, as shown in FIG. Fig.14 Finally, the second angle between the edge of the continuous casting billet 7 and the corresponding edge of the roller 1 is calculated by the fitting straight line y=k1x+b1 and the reference line y=k2x+b2 The fitting straight line y=k1x+b1 is the relationship between the distance value measured by the laser rangefinder 5 on the corresponding side and the time, and has an approximate straight line relationship with the direction of the corresponding edge of the continuous casting billet 7, and the reference line y=k2x+b2 is the straight line where the corresponding edge of the roller 1 is located;
[0238] In step S2, the control system optimizes the first angle and the second angle obtained in real time to obtain an optimized angle, and the optimized angle is expressed as This can improve the accuracy and reliability of the measured straightness angle between the edge of the continuous casting billet 7 and the edges on both sides of the roller table 1.
[0239] Among them, the alarm 9 is an audible and visual warning light, the control system is an industrial computer, and the industrial computer end uses a serial port relay to control the audible and visual warning light. The workflow of the industrial computer controlling the audible and visual warning light is as follows: Figure 2 As shown, if the measured optimized angle exceeds the safety threshold, the industrial computer sends a command to the serial port relay, the serial port relay energizes the sound and light alarm light, and the sound and light alarm light gives an alarm prompt to remind the staff to perform subsequent operations.
[0240] The system for measuring the straightness angle between the edge of the continuous casting billet and the edges on both sides of the roller table of the present invention also includes a display, the signal output end of the industrial computer is communicatively connected to the display, and the optimized angle measured by the industrial computer is displayed on the display.
[0241] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and substitutions can be made without departing from the technical principles of the present invention. These improvements and substitutions should also be regarded as the scope of protection of the present invention.
Claims
1. A system for measuring the straightness angle between the edge of a continuous casting billet and the edges on both sides of a roller, characterized in that: The invention comprises a roller conveyor (1), wherein cameras (2) are arranged on the upper left and upper right of the roller conveyor (1), the fields of view of the cameras (2) on the left and right sides overlap and form a binocular camera, a multi-line laser (3) is arranged between the cameras (2) on the left and right sides, and is used to project a plurality of parallel line structured lights onto a continuous casting billet (7) running on the roller conveyor (1), and the binocular camera is used to identify edge feature points of the continuous casting billet (7) provided by the multi-line laser (3) in real time and fit the contour straight line equation of the continuous casting billet (7); a point light source group is arranged on the left and right sides of the roller conveyor (1) and is within the field of view of the binocular camera, and the left and right sides of the roller conveyor (1) are provided with a point light source group within the field of view of the binocular camera, and the binocular camera on the left and right sides of the roller conveyor (1) is used to project a plurality of parallel line structured lights onto a continuous casting billet (7) running on the roller conveyor (1). The center line of the multiple point light sources (4) in the point light source group on the side is parallel to the left edge of the roller (1), and the center line of the multiple point light sources (4) in the point light source group on the right is parallel to the right edge of the roller (1). The binocular camera cooperates with the point light source groups on the left and right sides to calibrate the reference axis of the corresponding edge of the roller (1); the binocular camera is communicatively connected to the signal input end of the control system, and the signal output end of the control system is communicatively connected to the alarm (9). The control system is used to calculate and obtain in real time the first angle between the reference axis and the fitted continuous casting billet (7) contour straight line equation.
2. A system for measuring the straightness angle between the edge of a continuous casting billet and the edges on both sides of a roller table according to claim 1, characterized in that: The control system is used to compare the first angle acquired in real time with a safety threshold stored in the control system. When the first angle exceeds the safety threshold, the control system is used to control the alarm (9) to send out an alarm signal.
3. The system for measuring the straightness angle between the edge of a continuous casting billet and the edges on both sides of a roller table according to claim 1, characterized in that: Laser rangefinders (5) are provided at the left and right edges of the roller (1). The laser rangefinders (5) on the left and right sides are communicatively connected to the signal input end of the control system. The laser rangefinder (5) on the left side is used to obtain the distance between the left edge of the continuous casting billet (7) and the left edge of the roller (1) in real time and transmit it to the control system. The laser rangefinder (5) on the right side is used to obtain the distance between the right edge of the continuous casting billet (7) and the right edge of the roller (1) in real time and transmit it to the control system. The control system is used to detect the second angle between the edge of the continuous casting billet (7) and the corresponding edge of the roller (1) in real time according to the received distance information.
4. A system for measuring the straightness angle between the edge of a continuous casting billet and the edges on both sides of a roller table according to claim 3, characterized in that: The control system is used to compare the second angle acquired in real time with a safety threshold stored in the control system, and when the second angle exceeds the safety threshold, the control system is used to control the alarm (9) to send out an alarm signal.
5. The system for measuring the straightness angle between the edge of a continuous casting billet and the edges on both sides of a roller table according to claim 3, characterized in that: The control system is used to optimize the first angle and the second angle obtained in real time to obtain an optimized angle. The control system is used to compare the optimized angle with a safety threshold stored in the control system. When the optimized angle exceeds the safety threshold, the control system is used to control the alarm (9) to send an alarm signal.
6. A system for measuring the straightness angle between the edge of a continuous casting billet and the edges on both sides of a roller table according to claim 5, characterized in that: The left camera (2) and the right camera (2) are symmetrically arranged relative to the center line of the conveying direction of the roller (1); the left point light source group and the right point light source group are symmetrically arranged relative to the center line of the conveying direction of the roller (1); the left laser rangefinder (5) and the right laser rangefinder (5) are arranged left and right relative to the center line of the conveying direction of the roller (1); the multi-line laser (3) is located directly above the roller (1); the alarm (9) is located above the multi-line laser (3); the left and right cameras (2), the left and right point light source groups, the left and right laser rangefinders (5), the multi-line laser (3) and the alarm (9) are all installed via a mounting beam (6).
7. A method for measuring the straightness angle between the edge of a continuous casting billet and the edges on both sides of a roller, using the system for measuring the straightness angle between the edge of a continuous casting billet and the edges on both sides of a roller as claimed in any one of claims 3 to 6, characterized in that: The following steps are involved: S1 includes two parallel sub-steps S1-1 and S1-2: S1-1, the binocular camera cooperates with the point light source groups on the left and right sides to calibrate the reference axis of the corresponding edge of the roller (1), and transmits the calibrated reference axis data of the corresponding edge of the roller (1) to the control system; the multi-line laser (3) projects a plurality of parallel line structured lights to the continuous casting billet (7) running on the roller (1), the binocular camera identifies the edge feature points of the continuous casting billet (7) provided by the multi-line laser (3) and fits the contour straight line equation of the continuous casting billet (7), and transmits the fitted contour straight line equation data of the continuous casting billet (7) to the control system; the control system calculates and obtains the first angle between the reference axis and the fitted contour straight line equation of the continuous casting billet (7) in real time; S1-2, at the same time, the laser rangefinder (5) on the left side acquires the distance between the left edge of the continuous casting billet (7) and the left edge of the roller (1) in real time and transmits it to the control system, and the laser rangefinder (5) on the right side acquires the distance between the right edge of the continuous casting billet (7) and the right edge of the roller (1) in real time and transmits it to the control system, and the control system detects the second angle between the edge of the continuous casting billet (7) and the corresponding edge of the roller (1) in real time according to the received distance information; S2, the control system optimizes the first angle and the second angle obtained in real time to obtain an optimized angle, and the control system compares the optimized angle with a safety threshold stored in the control system. When the optimized angle exceeds the safety threshold, the control system controls the alarm (9) to send an alarm signal.
8. A method for measuring the straightness angle between the edge of a continuous casting billet and the edges on both sides of a roller table according to claim 7, characterized in that: In step S1-1: the binocular camera identifies the edge feature points of the continuous casting billet (7) provided by the multi-line laser (3) and fits the contour straight line equation of the continuous casting billet (7). Specifically, the binocular camera collects a binocular image of the surface of the continuous casting billet (7) with line structured light, wherein the projection of each line structured light on the continuous casting billet (7) intersects with the left edge and the right edge of the continuous casting billet (7) respectively, to obtain a left contour point and a right contour point, and calculates the three-dimensional coordinates of the left contour point and the right contour point in the camera (2) coordinate system through the binocular camera, and then projects the three-dimensional coordinates of the left contour point and the right contour point to the plane where the reference axis is located, to fit the contour straight line equation of the continuous casting billet (7).
9. A method for measuring the straightness angle between the edge of a continuous casting billet and the edges on both sides of a roller table according to claim 8, characterized in that: Step S1-1 specifically includes the following sub-steps: S1-1-1. Calibration of binocular camera S1-1-11, automatic exposure processing in camera image acquisition: the direction of exposure adjustment is the direction to maximize the gradient information, and the exposure adjustment will eventually stabilize when the image gradient information is near the maximum value, specifically, extracting clear line structured light features through the automatic exposure method, wherein the automatic exposure method includes updating the camera parameters and updating the exposure value during the exposure adjustment; S1-1-1101, Update camera parameters when adjusting exposure The exposure is updated by adjusting the exposure time and exposure gain, where the exposure of the camera is measured by the exposure value EV (2): EV=log(t*G) (1) In formula (1), t represents the exposure time and G represents the exposure gain; Increasing the exposure gain will bring noise to the image. Generally, the exposure gain is set to a smaller value, and the exposure value is updated by adjusting the exposure time first. When the exposure value update requirement cannot be met by adjusting the exposure time, the exposure gain is adjusted. S1-1-1102, Update exposure value The image gradient information is approximately a convex function with a maximum value with respect to the exposure value. The exposure value is updated using the golden section method. The image gradient value is f(x), x∈[a,b], and the image gradient value f(x) has a maximum value in the exposure value interval [a,b], and the exposure value corresponding to the maximum value of the image gradient value f(x) is x * ; Before automatically adjusting the image exposure, the scene is exposed twice continuously. Specifically, two trial points x1 and x2 are taken on the interval [a, b], and a<x1<x2<b. Then, the image gradient value f(x1) corresponding to the exposure value x1 and the image gradient value f(x2) corresponding to the exposure value x2 are calculated. By calculating the image gradient value obtained after exposure, the exposure value range is narrowed. If f(x1) < f(x2), then x * ∈[x1,b], keep [x1,b] and remove [a,x1]; if f(x1)>f(x2), then x * ∈[a,x2], keep [a,x2] and remove [x2,b]; if f(x1)=f(x2), then x * ∈[x1,x2], keep [x1,x2] and remove [a,x1] and [x2,b]; Then, the left endpoint value of the interval is assigned to a, and the right endpoint value is assigned to b. The exposure value is updated repeatedly using the golden section method, and the exposure value interval of the image is continuously narrowed to obtain the best exposure value. Let L k is the length of the interval left after comparing the image gradient value after the image is exposed for the kth time. When the interval is shortened, the length of the interval left each time is ω times the length of the interval left in the previous time. Then L k+1 =ωL k (2) In actual exposure adjustment, x1 and x2 are placed in symmetrical positions in the interval [a, b], specifically: x1-a=b-x2 (3) According to the principle of equal-proportional contraction, we can get L1=ωL0 (4) In formula (4), L0 = ba, Then L1=x2-a=b-x1=ω(ba), we can get x1=b-ω(ba), and find x1=a+(1-ω)(ba), x2=a+ω(ba); When the interval [x1, b] is retained and two trial points are needed on the retained interval [x1, b] in order to further shorten the interval, the existing point x2 in the interval [x1, b] is used as one of the trial points to reduce the amount of calculation; And w 2 +w-1=0, then and then EV1=x1=a+0.382(ba) (5) EV2=x2=a+0.618(ba) (6) In formula (5), EV1 represents the exposure value corresponding to the left test point, and in formula (6), EV2 represents the exposure value corresponding to the right test point; S1-1-12. Calibrate the binocular camera to obtain the internal and external parameters of the camera and the transfer matrix between the left and right cameras S1-1-1201. Calibrate the binocular camera to obtain the internal and external parameters of the camera The calibration plate is freely moved within the field of view (8) of the binocular camera, and k images of the calibration plate are obtained from different angles, where k ≥ 3. The sub-pixel coordinates of each feature point in the calibration plate image are detected, and the correspondence between the feature points is used to solve the world coordinate system OX in each viewing angle. W Y W Z W With pixel coordinate system o u -The homography matrix H between uv, taking the plane where the calibration plate is located as the world coordinate system Z W = 0, then the corresponding relationship between the world coordinate system and the pixel coordinate system of the point on the calibration plate plane is: In formula (7), is the internal parameter matrix of the camera (2), f x Indicates the scale factor of the x-axis, f y represents the scale factor of the y-axis, [u0,v0] represents the coordinates of the intersection point O1 of the optical axis of the camera (2) and the image plane in the pixel coordinate system, R is the rotation matrix, which is a 3×3 orthogonal unit matrix, r i represents the i-th column vector of the rotation matrix R, t is the translation matrix, and s is the scale factor; make Then formula (7) can be expressed as: In formula (8), By knowing the world coordinates and sub-pixel coordinates of the feature points on the calibration plate, the overdetermined equations are solved using the least squares method and further optimized using the Levenberg-Marquarat algorithm to obtain the H matrix, and then obtain the initial solution of the internal and external parameters of the camera (2); S1-1-1202. Calibrate the binocular camera to obtain the transfer matrix between the left and right cameras The three-dimensional coordinates of the same point in the known space in the coordinate system of the left camera (2) are expressed as P l , the three-dimensional coordinates in the coordinate system of the right camera (2) are expressed as P r , the three-dimensional coordinates in the world coordinate system are expressed as P W , then P l , P r With P W Has the following relationship: In formula (9), R l is the rotation matrix of the left camera (2), T l is the translation matrix of the left camera (2), R r is the rotation matrix of the right camera (2), T r is the translation matrix of the right camera (2), R 12 is the rotation matrix from the right camera (2) coordinate system to the left camera (2) coordinate system, t 12 is the translation matrix from the right camera (2) coordinate system to the left camera (2) coordinate system, From formula (9), we can get S1-1-13. Calculate the three-dimensional coordinates of the characteristic points on the calibration plate For the pixel coordinates p(u,v) of the intersection of the line structured light in the obtained image and the edge of one side of the calibration plate, it is assumed that the three-dimensional coordinates of the spatial measured point represented by it in the reference world coordinate system are P(X W ,Y W ,Z W ),but WITH C p=MP (11) In formula (11), Z C represents the scale factor, which is related to the distance between the optical center of the camera (2) and the measured point in space, and M represents the projection matrix containing the internal and external parameters of the camera (2), where For the left and right cameras (2), we have In formula (12), R and t in the external parameter matrix included in the projection matrix M1 of the left camera (2) are composed of the unit matrix and the zero matrix, and R and t in the external parameter matrix included in the projection matrix M2 of the right camera (2) are composed of R 12 and t 12 The three-dimensional coordinates P of the corresponding intersection point in the reference world coordinate system are obtained by using the least square method according to the projection matrix M1 of the left camera (2) and the projection matrix M2 of the right camera (2). W1 ; S1-1-2. Calibrate the reference axis and obtain the roller plane equation where the standard lines on the left and right sides of the roller are located The binocular camera is used to photograph the point light source groups on the left and right sides, and the pixel coordinates of the geometric center points of each point light source (4) in the corresponding point light source groups on the image are solved. According to step S1-1-12, the conversion relationship from the pixel coordinates of the point on the image to the world coordinate system can be obtained. The three-dimensional coordinates (x ) of each point light source (4) in the point light source groups on the left and right sides of the image can be measured through the conversion relationship. i ,y i ,z i ), i = 1, 2, 3, ..., N, wherein the center line of each point light source (4) in the corresponding side point light source group forms the standard line of the corresponding edge of the roller (1), i.e., the reference axis. The general expression of the plane equation of the roller (1) that does not pass through the origin in the coordinate system of the left camera (2) is: Ax + By + Cz + 1 = 0. Let A, B, and C be the optimal parameters of the plane. Substitute the three-dimensional coordinates of each point light source (4) into the expression of the plane equation of the roller (1). There is an error e i =Ax i +By i +Cz i +1, summing up the squares of all errors gives Formula (13) is a convex hyperplane in four-dimensional space, and has a unique minimum value, which is also the global minimum value. Solving this set of equations yields the plane equation of the roller table (1) where the reference axis is located; S1-1-3. Calculate the first angle between the edge of the continuous casting billet and the corresponding side edge of the roller S1-1-31, continuous casting slab edge contour point extraction By controlling the laser projection angle of the multi-line laser (3) to add artificial features to the image of the continuous casting billet (7), the line structured light is broken at the edge of the continuous casting billet (7), which is convenient for edge point positioning, and the interval positioning method is used to distinguish the position of the line structured light to separate a single complete line structured light stripe, specifically: Select an interval value dis and traverse the image along the y direction of the ROI image: if the gray value I of a pixel point of each column of line structured light is the same as the previous pixel point in the x direction (i-1,ndis) is less than the laser line low threshold l_thre, and the next pixel gray value I (i+1,ndis) If it is greater than the laser line high threshold h_thre, the pixel point (i,n·dis) is determined to be the left edge point of the line structured light, and the right edge point is the opposite. The continuous left edge points and right edge points in the same row locate a line structured light stripe, and each row can locate multiple line structured light stripes. The row where the number of located line structured light stripes is equal to the number of image projection line structured light stripes is selected as the reference row, and the left edge point of each line structured light stripe is used to perform linear regression fitting of the line structured light straight line using the least squares method; According to the determined upper edge point and lower edge point of a certain line structured light straight line, wherein the position of the upper edge point of the line structured light corresponds to the right edge of the continuous casting billet (7), and the position of the lower edge point of the line structured light corresponds to the left edge of the continuous casting billet (7), an upper edge distance and a lower edge distance are selected for ROI extraction and splicing, so as to realize the single line structured light ROI area extraction, and then the center point of the line structured light stripe obtained by the grayscale centroid method is optimized, wherein the initial sub-pixel coordinates of each point are solved by the grayscale centroid method in the image coordinate system o-xy, C p (x p ,y p ) is one of the points, with C p Centered on r p is the distance range, and the selected range is (2r p +1) points to recalculate the centroid As optimized sub-pixel coordinates: Finally, the center point of the line structured light stripe with high precision and smoothness is extracted, and then the endpoints on both sides of the line structured light are extracted, and the binocular image is subjected to epipolar correction, and the edge points obtained from the image of the continuous casting billet (7) taken by the left camera (2) and the image of the continuous casting billet (7) taken by the right camera (2) are matched to ensure that the edge points obtained from the left and right images correspond to the same three-dimensional coordinates, thereby realizing the extraction of the edge contour points of the continuous casting billet (7); S1-1-32, Calculation of three-dimensional coordinates of the edge contour points of continuous casting billet For the edge contour points extracted from the continuously collected continuous casting billet (7) images, further splicing is required to fit the contour of the entire continuous casting billet (7), specifically: the real position of the edge contour point is determined according to the world coordinates of the edge contour point, and the pixel coordinates in the camera (2) image are extracted at the same time, and the corresponding contour points in the edge contour of the continuous casting billet (7) are assigned values to obtain the three-dimensional coordinates of the edge contour points of the continuous casting billet (7); S1-1-33. Solution for the first angle between the straight line equation of the continuous casting billet contour and the reference axis By matching a single set of left and right image pairs, the three-dimensional coordinates of the edge contour points of the continuous casting billet (7) at a certain moment in the camera (2) coordinate system are obtained, and the acquisition interval time is combined with the travel speed of the continuous casting billet (7) to calculate, so as to obtain the complete contour of the continuous casting billet (7). The three-dimensional coordinates of the edge contour points of the continuous casting billet (7) obtained by a single measurement are all established in the same left camera (2) coordinate system. According to the principle of relative displacement, the camera (2) coordinate system at different moments is displaced accordingly according to the travel distance, so as to unify the three-dimensional coordinates of the edge contour points of the continuous casting billet (7) to the left camera (2) coordinate system O1-X1Y1Z1 at moment t1; Assume that at time t, the continuous casting billet (7) travels at a speed of v t mm / s, the roller (1) is affected by the outside world, causing the traveling speed to be disturbed to varying degrees, resulting in the continuous casting billet (7) being unable to maintain a uniform speed. t Integrate to find the distance d that the continuous casting billet (7) moves on the roller table (1) in the direction of the standard line. In formula (15), t1 is the start time and t2 is the end time; Record the acquisition time and acquisition interval of two consecutive images, so as to splice the edge contour point set of the continuous casting billet (7) in each image in time sequence to obtain the edge contour point set in a unified coordinate system. After obtaining the three-dimensional coordinates of the intersection of the edge of the continuous casting billet (7) and the line structured light, the three-dimensional coordinates are projected onto the plane where the roller (1) is located, and the coordinates of a point on the edge contour of the continuous casting billet (7) are set as P(x p ,y p ,z p ), the plane equation of the roller table (1) obtained in step S1-1-2 is Ax+By+Cz+1=0, and the corresponding normal vector Get point P(x p ,y p ,z p ) to the plane where the roller (1) is located is: The intersection of the projection line and the plane where the roller (1) is located is the projection point p (x, y, z) from the edge contour point of the continuous casting billet (7) to the plane where the roller (1) is located. The combined equation Ax+By+Cz+1=0 and equation (16) yield: The intersection points of all projection lines and the plane where the roller (1) is located are projected according to the above method to obtain the projection point set {p'(x i ,y i ,z i ),i=1,2,...,N}, and obtain the contour straight line equation of the continuous casting billet (7) by linear fitting based on the projection point set: In formula (18), A(x0,y0,z0) is a point on the contour line of the continuous casting billet (7), and the vector is a non-zero vector and is parallel to the straight line of the continuous casting billet (7), B(x, y, z) is an arbitrary point on the straight line of the continuous casting billet (7), and the vector With vector The components of the vector are proportional, and combined with the moving direction vector of the roller (1), the first angle between the straight line equation of the continuous casting billet (7) and the reference axis is calculated as follows:
10. A method for measuring the straightness angle between the edge of a continuous casting billet and the edges on both sides of a roller table according to claim 8, characterized in that: Step S1-2 is specifically as follows: Phase laser detection technology modulates the intensity of the emitted laser light wave by means of a modulation signal, and calculates the distance between the measured target and the laser rangefinder (5) according to the measured phase difference, specifically: Assuming that the distance between the laser emission point and the reflection point is x, the speed of the laser in the air is c, and the round-trip time between the laser emission point and the reflection point is t, then: t=2x / c (19) The period of the modulation signal is T, the frequency is f, and the phase difference between the laser emission point and the received signal at the reflection point is φ, then φ is expressed as: In formula (20), N is the number of complete cycle waves obtained, and Δφ is the phase of the wave that is less than a complete cycle; In summary, the distance between the measured target and the laser rangefinder (5) is: When the continuous casting billet (7) passes through the laser distance meter (5), the laser distance meter (5) continuously emits laser light waves to measure the distance between the edge of the continuous casting billet (7) and the corresponding side edge of the roller (1), and transmits the distance to the control system in real time. The distance values measured by the laser distance meters (5) on the left and right sides gradually increase or decrease over time. Finally, the second angle between the edge of the continuous casting billet (7) and the corresponding edge of the roller (1) is calculated by the fitting straight line y=k1x+b1 and the reference line y=k2x+b2. The fitting straight line y=k1x+b1 is the relationship between the distance value measured by the laser rangefinder (5) on the corresponding side and the time, and has an approximate straight line relationship with the direction of the corresponding edge of the continuous casting billet (7), and the reference line y=k2x+b2 is the straight line where the corresponding edge of the roller (1) is located; In step S2, the control system optimizes the first angle and the second angle obtained in real time to obtain an optimized angle, and the optimized angle is expressed as
Citation Information
Cited By
High-precision right angle detection device and detection method
CN119509418A
A high-precision right-angle detection device and detection method
CN119509418B
Angle detection method and device for continuous casting billet, electronic equipment and storage medium
CN120707479A
