A low-delay strip deviation amount detection method based on machine vision

Abstract

The application discloses a low-delay strip deviation detection method based on machine vision, which comprises the following steps: step 1, simultaneously collecting images of the strip at the inlet side and the outlet side of the roller during the rolling process; step 2, calculating the edge positions of the strip images collected at the inlet side and the outlet side by using an edge detection algorithm, and obtaining the edge coordinates of both sides of each strip image at the collection moment; step 3, calculating the position coordinates of each point on the center line of the strip at the inlet side and the outlet side of the roller by using the edge coordinates of both sides of the strip; step 4, performing circle fitting on the center line of the strip at the inlet side and the outlet side of the roller by using the position coordinates of each point on the center line of the strip; step 5, calculating the intersection positions of the fitting circles at the inlet side and the outlet side of the roller and the central axis of the roller, and recording the intersection positions as the deviation values at the inlet side and the outlet side; and step 6, calculating the average value of the deviation values at the inlet side and the outlet side as the strip deviation at the center of the roller.

Description

Technical Field

[0001] This invention belongs to the field of strip steel finishing control technology, and relates to a low-delay strip steel deviation detection method based on machine vision. Background Technology

[0002] Strip misalignment refers to the phenomenon where the centerline of the strip deviates from the centerline of the rolls during the finishing rolling process due to factors such as lateral bending and misalignment of the intermediate billet, as well as precision issues with the rolling equipment. Once strip misalignment occurs, it leads to significant deviations in the rolling force and roll gap width on both the operating and drive sides, thus interfering with the mill's shape control system. This results in quality problems such as strip waviness, and in severe cases, even "tailing" and "stacking" accidents, affecting production progress and safety. Therefore, real-time correction of strip misalignment is necessary. To achieve this, it is essential to obtain the amount of strip misalignment at the rolls. Currently, using machine vision to obtain the center position of strip steel is one of the commonly used methods. However, due to limitations such as installation location, the camera cannot directly detect the strip steel deviation at the roll position, but can only detect the position of the strip steel at the roll inlet and roll outlet. This results in a delay between the detection result and the actual controlled quantity (the deviation of the strip steel at the roll). This delay is related to the rolling speed, which needs to be adjusted constantly. Therefore, this increases the control difficulty of the strip steel deviation correction system and reduces the stability of the control system.

[0003] Numerous studies have been conducted by researchers both domestically and internationally regarding the detection and control of strip position during the rolling process. A Chinese journal article, "Research and Application of Asymmetric Measurement and Control System for Hot Strip Rolling Operation" (Metallurgical Automation, 2020, 44(01):48-54), designed and developed an asymmetric detection and control system for hot strip rolling operation based on machine vision technology. The system utilizes a binocular linear array camera to acquire images of the strip width and deviation, thereby calculating the roll gap inclination value to correct the strip's orientation. This system has been successfully applied in hot strip rolling production practice. Application results show that the detection device operates stably, can perform online detection and control of deviation in the finishing mill, and meets the required accuracy, effectively improving rolling stability.

[0004] The international conference paper "A vision-based system for strip tracking measurement in the finishing train of a hot strip mill" (International Conference on Mechatronics & Automation, IEEE, 2010) proposes a method for fitting and predicting the edge position of the strip using Bézier curves to overcome the interference of the environment, especially cooling water, on the strip position detection, which greatly improves the robustness of the detection results. The Chinese journal article "Strip Deviation Detection System Using Novel Sensors" (Welded Pipe, 2007, No. 151(05): 79-81+107-108) utilizes grating sensors to overcome the problems of vertical displacement of the strip during its movement and inconsistent strip specifications, realizing the detection of strip deviation during the production process of spiral submerged arc welded pipes.

[0005] Chinese invention patent CN113828641A, entitled "A Method for Processing Strip Steel Belt Deviation Curves Between Stands Based on Machine Vision," discloses a method for processing strip steel belt deviation curves between stands. This method utilizes a binocular line-scan camera mounted at the top of the stand to capture the strip steel edge, calculates the strip steel width and deviation amount, and then calculates the difference between the actual strip steel width measured by the exit width measuring instrument and the detected strip steel width. Deviation data corresponding to width differences exceeding a threshold are filtered to remove noise from the deviation curve. Chinese invention patent CN102602681A, entitled "Online Detection Method for Conveyor Belt Deviation Faults Based on Machine Vision," discloses a machine vision-based online detection method for conveyor belt deviation faults. It uses a line-scan CCD camera to capture images of the conveyor belt in operation, performs binarization processing to obtain a binary image of the conveyor belt, maps the binary image to a one-dimensional fault feature function, and then extracts the deviation angle and offset from the one-dimensional fault feature function for deviation fault identification. Chinese utility model patent CN216889346U, entitled "A Visual Correction Device for Fabric Transport," discloses a visual correction device for fabric transport. Through the arrangement of a camera, control mechanism, servo motor, support, and swing wheel, real-time adjustment of the fabric during transport can be achieved, preventing fabric deviation and its impact on subsequent processing.

[0006] The shortcomings of the above research are that it did not take into account the time delay between the detection result and the actual controlled quantity, which makes the design of the control algorithm more difficult and the control effect worse. Summary of the Invention

[0007] To address the aforementioned technical problems, the present invention aims to provide a low-delay strip deviation detection method based on machine vision. This method involves using a circular fitting method to measure the centerline of the strip at the roll inlet and outlet sides, and calculating the average value of the intersection points of the fitted circles on both sides with the roll centerline to obtain the strip deviation at the roll.

[0008] This invention provides a low-latency strip deviation detection method based on machine vision, comprising the following steps:

[0009] Step 1: Simultaneously acquire images of the strip steel during the rolling process from both the inlet and outlet sides of the rolls;

[0010] Step 2: Calculate the edge positions of the strip images acquired from the inlet and outlet sides using an edge detection algorithm to obtain the edge coordinates on both sides of each strip image at the acquisition time;

[0011] Step 3: Calculate the position coordinates of each point on the center line of the strip on the roll inlet and outlet sides using the edge coordinates on both sides of the strip.

[0012] Step 4: Use the position coordinates of each point on the strip centerline to perform circle fitting on the centerlines of the strip on the roll inlet and outlet sides respectively;

[0013] Step 5: Calculate the intersection points of the fitted circles on the roll inlet side and roll outlet side with the roll centerline, and record them as the deviation values ​​on the inlet side and outlet side respectively;

[0014] Step 6: Calculate the average value of the deviation on the inlet and outlet sides as the strip deviation amount at the roll center.

[0015] Furthermore, in step 2, the edge positions of the strip images captured by the cameras on the mill entrance and exit sides are calculated based on the Canny edge detection algorithm, specifically as follows:

[0016] Step 2.1: Perform binary segmentation on the strip image captured by the camera, set the gray value of pixels with a value greater than the threshold to the maximum value of 255, and set the gray value of pixels with a value less than or equal to the threshold to the minimum value of 0, to obtain the binary segmented image M1;

[0017] Step 2.2: Perform dilation and erosion morphological operations on image M1 to fill the holes in image M1 caused by image noise, and obtain the morphologically processed image M2;

[0018] Step 2.3: Find the connected component N with the largest area in image M2. The region corresponding to connected component N is the region where the strip is located in image M2. Keep the gray value of connected component N unchanged, and set the gray value of other connected components in M2 to 0. At this time, we get image M3 with only connected component N as the foreground and background gray value of 0.

[0019] Step 2.4: Use the Canny edge detection algorithm to perform edge detection on image M3, extracting the boundary lines L1 and L2 between the connected component N and the background. L1 and L2 are the two edges of the strip, where the position coordinates of the m pixels on L1 are (x... 11 ,y 11 ), (x 12 ,y 12 ), ..., (x 1m ,y 1m The coordinates of the n pixels on L2 are (x, y, y) 21 ,y 21 ), (x 22 ,y 22 )……(x 2n ,y 2n ).

[0020] Furthermore, step 3 specifically involves:

[0021] Step 3.1: Compare the x-coordinates of the m pixels on strip edge L1 with the x-coordinates of the n pixels on edge L2; extract pairs of pixels from the two strip edges that have the same x-coordinate, resulting in k pairs of pixels with the same x-coordinate but from different edges, denoted as (x1, y1, ..., y2). 11 ,y 21 (x2,y) 12 ,y 22 ), ..., (x k ,y 1k ,y 2k ), where x1 to x k Let y be the x-coordinate of these k pairs of pixels. 11 to y 1k Let y be the ordinate of these k pixels on the L1 edge. 21 to y 2k Let k be the ordinates of these k pixels on the L2 edge;

[0022] Step 3.2: Calculate the average of the ordinates of these k pairs of pixels, i.e., y1 = (y 11 +y 21 ) / 2, y2=(y 12 +y 22 ) / 2、……、y k =(y 1k +y 2k ) / 2;

[0023] Step 3.3: Obtain the coordinates of k pixels (x1, y1), (x2, y2), ..., (x...) on the center line of the strip. k ,y k).

[0024] Furthermore, in step 4, the fitted circles for the inlet and outlet sides are set to Φ1 and Φ2, respectively. A circle fitting is performed on the centerline of the strip. The specific process is as follows:

[0025] Step 4.1: Taking the fitted circle Φ1 on the entrance side as an example, the fitting process is described. The standard equation of the circle is known to be:

[0026] x 2 +y 2 +Ax+By+C=0 (1)

[0027] Where A, B, and C are the unknown parameters to be fitted; Equation (1) is transformed into:

[0028] Ax + By + C = -x 2 -y 2 (2)

[0029] The coordinates of k pixels (x1, y1), (x2, y2), ..., (x2, y2) on the center line of the strip on the entrance side obtained in step 3 are respectively... k ,y k Substituting into formula (2), we get:

[0030]

[0031] Equation (3) is converted into matrix form to obtain equation (4):

[0032]

[0033] According to the least squares method in matrix form, if:

[0034] X·R=Y (5)

[0035] Where X and Y are matrices with known parameters, and R is the matrix of unknowns to be fitted, then the least squares solution of R is:

[0036] R = (X T X) -1 X T Y (6)

[0037] Comparing equation (4) and equation (5), we can let:

[0038]

[0039]

[0040]

[0041] Substituting equations (7) to (9) into equation (6), we get:

[0042]

[0043] After obtaining A, B and C, substituting them into the standard equation of the circle in equation (1) will yield the equation of the fitted circle Φ1 on the entrance side.

[0044] Step 4.2: Similarly, based on step 4.1, obtain the equation of the fitted circle Φ2 on the exit side.

[0045] Furthermore, in step 5, the intersection points of the two fitted circles and the central axis of the roll are calculated, specifically as follows:

[0046] Step 5.1: Taking the fitted circle Φ1 on the inlet side as an example, calculate the intersection point of the fitted circle Φ1 and the central axis of the roll, and obtain the circle equation x. 2 +y 2 After +Ax+By+C=0, let x=0, and we get:

[0047] y 2 +By+C=0 (11)

[0048] The deviation value ΔY1 on the inlet side is obtained as follows:

[0049]

[0050] If the center of the fitted circle Φ1 is above the center line of the strip, then:

[0051]

[0052] If the center of the fitted circle Φ1 is below the centerline of the strip, then:

[0053]

[0054] Step 5.2: Similarly, obtain the deviation value ΔY2 on the exit side according to Step 5.1.

[0055] Furthermore, in step 5.1, the coordinates of the center of the fitted circle are calculated according to the following formula:

[0056] Let the fitted circle on the entrance side be Φ1(a1,b1,r1), where a1, b1, and r1 are the x-coordinate of the center, y-coordinate of the center, and radius of Φ1, respectively. According to formula (1), we can obtain:

[0057]

[0058] Therefore, we get:

[0059]

[0060] Compare the ordinate b1 of the center of the fitted circle with the ordinate of the point on the strip centerline obtained in step 3. If the ordinate of the center is greater than the ordinate of the point on the strip centerline, the center of the fitted circle is above the strip centerline; otherwise, the center of the fitted circle is below the strip centerline.

[0061] Furthermore, in step 6, the strip deviation at the roll center is calculated according to the following formula.

[0062]

[0063] This invention discloses a low-latency strip deviation detection method based on machine vision. The method involves capturing images of the strip at the roll inlet and outlet sides; obtaining the coordinates of edge pixels on both sides of the strip using an edge detection algorithm; calculating the coordinates of points along the strip's centerline; approximating the strip's centerline as an arc and obtaining a fitted circle using the coordinates of points along the centerline; and finally, calculating the strip deviation at the roll inlet and outlet sides by taking the average of the intersections of the fitted circles with the roll's centerline. This eliminates the delay between the detection result and the actual controlled variable, reducing the control difficulty of the strip deviation correction system and improving the stability of the control system. Attached Figure Description

[0064] Figure 1 This is a flowchart of a low-latency strip deviation detection method based on machine vision according to the present invention.

[0065] Figure 2 This is a schematic diagram of the camera's installation position in this invention;

[0066] Figure 3 This is a schematic diagram illustrating the calculation principle of the low-latency strip deviation detection method based on machine vision of the present invention. Detailed Implementation

[0067] like Figure 1 The present invention provides a low-latency strip deviation detection method based on machine vision, comprising the following steps:

[0068] Step 1: Simultaneously acquire images of the strip steel during the rolling process from both the inlet and outlet sides of the rolls, such as... Figure 2 As shown;

[0069] Step 2: Calculate the edge positions of the strip images acquired from the inlet and outlet sides using an edge detection algorithm to obtain the edge coordinates on both sides of each strip image at the acquisition time;

[0070] In practical implementation, edge detection algorithms can be implemented using the Canny edge detection algorithm or deep learning-based edge detection methods, etc. In this embodiment, the edge positions of the strip images captured by cameras on the mill inlet and outlet sides are calculated based on the Canny edge detection algorithm, specifically as follows:

[0071] Step 2.1: Perform binary segmentation on the strip image captured by the camera, set the gray value of pixels with a value greater than the threshold to the maximum value of 255, and set the gray value of pixels with a value less than or equal to the threshold to the minimum value of 0, to obtain the binary segmented image M1;

[0072] Step 2.2: Perform dilation and erosion morphological operations on image M1 to fill the holes in image M1 caused by image noise, and obtain the morphologically processed image M2;

[0073] Step 2.3: Find the connected component N with the largest area in image M2. The region corresponding to connected component N is the region where the strip is located in image M2. Keep the gray value of connected component N unchanged, and set the gray value of other connected components in M2 to 0. At this time, we get image M3 with only connected component N as the foreground and background gray value of 0.

[0074] Step 2.4: Use the Canny edge detection algorithm to perform edge detection on image M3, extracting the boundary lines L1 and L2 between the connected component N and the background. L1 and L2 are the two edges of the strip, where the position coordinates of the m pixels on L1 are (x... 11 ,y 11 ), (x 12 ,y 12 ), ..., (x 1m ,y 1m The coordinates of the n pixels on L2 are (x, y, y) 21 ,y 21 ), (x 22 ,y 22 )……(x 2n ,y 2n ).

[0075] Step 3: Calculate the position coordinates of each point on the center line of the strip on both the roll inlet and outlet sides using the edge coordinates of the strip. Specifically, Step 3 involves:

[0076] Step 3.1: Compare the x-coordinates of the m pixels on strip edge L1 with the x-coordinates of the n pixels on edge L2; extract pairs of pixels from the two strip edges that have the same x-coordinate, resulting in k pairs of pixels with the same x-coordinate but from different edges, denoted as (x1, y1, ..., y2). 11 ,y 21 (x2,y)12 ,y 22 ), ..., (x k ,y 1k ,y 2k ), where x1 to x k Let y be the x-coordinate of these k pairs of pixels. 11 to y 1k Let y be the ordinate of these k pixels on the L1 edge. 21 to y 2k Let k be the ordinates of these k pixels on the L2 edge;

[0077] Step 3.2: Calculate the average of the ordinates of these k pairs of pixels, i.e., y1 = (y 11 +y 21 ) / 2, y2=(y 12 +y 22 ) / 2、……、y k =(y 1k +y 2k ) / 2;

[0078] Step 3.3: Obtain the coordinates of k pixels (x1, y1), (x2, y2), ..., (x...) on the center line of the strip. k ,y k ).

[0079] Step 4: Use the position coordinates of each point on the strip centerline to perform circle fitting on the centerlines of the strip on the roll inlet and outlet sides respectively;

[0080] In practice, in step 4, the fitted circles for the inlet and outlet sides are set as Φ1 and Φ2, respectively. A circle fitting is performed on the centerline of the strip. The specific process is as follows:

[0081] Step 4.1: Taking the fitted circle Φ1 on the entrance side as an example, the fitting process is described. The standard equation of the circle is known to be:

[0082] x 2 +y 2 +Ax+By+C=0 (1)

[0083] Where A, B, and C are the unknown parameters to be fitted; Equation (1) is transformed into:

[0084] Ax + By + C = -x 2 -y 2 (2)

[0085] The coordinates of k pixels (x1, y1), (x2, y2), ..., (x2, y2) on the center line of the strip on the entrance side obtained in step 3 are respectively... k ,y k Substituting into formula (2), we get:

[0086]

[0087] Equation (3) is converted into matrix form to obtain equation (4):

[0088]

[0089] According to the least squares method in matrix form, if:

[0090] X·R=Y (5)

[0091] Where X and Y are matrices with known parameters, and R is the matrix of unknowns to be fitted, then the least squares solution of R is:

[0092] R = (X T X) -1 X T Y (6)

[0093] Comparing equation (4) and equation (5), we can let:

[0094]

[0095]

[0096]

[0097] Substituting equations (7) to (9) into equation (6), we get:

[0098]

[0099] After obtaining A, B and C, substituting them into the standard equation of the circle in equation (1) will yield the equation of the fitted circle Φ1 on the entrance side.

[0100] Step 4.2: Similarly, based on step 4.1, obtain the equation of the fitted circle Φ2 on the exit side.

[0101] Step 5: Calculation principle diagram of the strip deviation detection method, as shown below. Figure 3 As shown. The positions of the intersection points of the fitted circles on the roll inlet side and roll outlet side with the roll centerline are calculated respectively, and denoted as the deviation values ​​on the inlet and outlet sides, respectively. Specifically:

[0102] Step 5.1: Taking the fitted circle Φ1 on the inlet side as an example, calculate the intersection point of the fitted circle Φ1 and the central axis of the roll, and obtain the circle equation x. 2 +y 2 After +Ax+By+C=0, let x=0, and we get:

[0103] y 2 +By+C=0 (11)

[0104] The deviation value ΔY1 on the inlet side is obtained as follows:

[0105]

[0106] If the center of the fitted circle Φ1 is above the center line of the strip, then:

[0107]

[0108] If the center of the fitted circle Φ1 is below the centerline of the strip, then:

[0109]

[0110] In specific implementation, step 5.1 calculates the coordinates of the center of the fitted circle according to the following formula:

[0111] Let the fitted circle on the entrance side be Φ1(a1,b1,r1), where a1, b1, and r1 are the x-coordinate of the center, y-coordinate of the center, and radius of Φ1, respectively. According to formula (1), we can obtain:

[0112]

[0113] Therefore, we get:

[0114]

[0115] Compare the ordinate b1 of the center of the fitted circle with the ordinate of the point on the strip centerline obtained in step 3. If the ordinate of the center is greater than the ordinate of the point on the strip centerline, the center of the fitted circle is above the strip centerline; otherwise, the center of the fitted circle is below the strip centerline.

[0116] Step 5.2: Similarly, obtain the deviation value ΔY2 on the exit side according to Step 5.1.

[0117] Step 6: Calculate the average value of the deviation on the inlet and outlet sides as the strip deviation amount at the roll center.

[0118] In practice, step 6 involves calculating the strip deviation at the roll center using the following formula.

[0119]

[0120] The above description is only a preferred embodiment of the present invention and is not intended to limit the ideas of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A low-latency strip steel deviation detection method based on machine vision, characterized in that, Includes the following steps: Step 1: Simultaneously acquire images of the strip steel during the rolling process from both the inlet and outlet sides of the rolls; Step 2: Calculate the edge positions of the strip images acquired from the inlet and outlet sides using an edge detection algorithm to obtain the edge coordinates on both sides of each strip image at the acquisition time; Step 3: Calculate the position coordinates of each point on the center line of the strip on the roll inlet and outlet sides using the edge coordinates on both sides of the strip. Step 4: Use the position coordinates of each point on the strip centerline to perform circle fitting on the centerlines of the strip on the roll inlet and outlet sides respectively; Step 5: Calculate the intersection points of the fitted circles on the roll inlet side and roll outlet side with the roll centerline, and record them as the deviation values ​​on the inlet side and outlet side respectively; Step 6: Calculate the average value of the deviation on the inlet and outlet sides as the strip deviation amount at the roll center.

2. The low-latency strip misalignment detection method based on machine vision as described in claim 1, characterized in that, In step 2, the edge positions of the strip images captured by the cameras on the mill entrance and exit sides are calculated based on the Canny edge detection algorithm, specifically as follows: Step 2.1: Perform binary segmentation on the strip image captured by the camera, set the gray value of pixels with a value greater than the threshold to the maximum value of 255, and set the gray value of pixels with a value less than or equal to the threshold to the minimum value of 0, to obtain the binary segmented image M1; Step 2.2: Perform dilation and erosion morphological operations on image M1 to fill the holes in image M1 caused by image noise, and obtain the morphologically processed image M2; Step 2.3: Find the connected component N with the largest area in image M2. The region corresponding to connected component N is the region where the strip is located in image M2. Keep the gray value of connected component N unchanged, and set the gray value of other connected components in M2 to 0. At this time, we get image M3 with only connected component N as the foreground and background gray value of 0. Step 2.4: Use the Canny edge detection algorithm to perform edge detection on image M3, extracting the boundary lines L1 and L2 between the connected component N and the background. L1 and L2 are the two edges of the strip, where the position coordinates of the m pixels on L1 are (x... 11 ,y 11 ), (x 12 ,y 12 ), ..., (x 1m ,y 1m The coordinates of the n pixels on L2 are (x, y, y) 21 ,y 21 ), (x 22 ,y 22 )……(x 2n ,y 2n ).

3. The low-latency strip misalignment detection method based on machine vision as described in claim 1, characterized in that, Step 3 specifically involves: Step 3.1: Compare the x-coordinates of the m pixels on strip edge L1 with the x-coordinates of the n pixels on edge L2; extract pairs of pixels from the two strip edges that have the same x-coordinate, resulting in k pairs of pixels with the same x-coordinate but from different edges, denoted as (x1, y1, ..., y2). 11 ,y 21 (x2,y) 12 ,y 22 ), ..., (x k ,y 1k ,y 2k ), where x1 to x k Let y be the x-coordinate of these k pairs of pixels. 11 to y 1k Let y be the ordinate of these k pixels on the L1 edge. 21 to y 2k Let k be the ordinates of these k pixels on the L2 edge; Step 3.2: Calculate the average of the ordinates of these k pairs of pixels, i.e., y1 = (y 11 +y 21 ) / 2, y2=(y 12 +y 22 ) / 2、……、y k =(y 1k +y 2k ) / 2; Step 3.3: Obtain the coordinates of k pixels (x1, y1), (x2, y2), ..., (x...) on the center line of the strip. k ,y k ).

4. The low-latency strip deviation detection method based on machine vision as described in claim 3, characterized in that, In step 4, the fitted circles for the inlet and outlet sides are set as Φ1 and Φ2, respectively. A circle fitting is performed on the centerline of the strip steel. The specific process is as follows: Step 4.1: Taking the fitted circle Φ1 on the entrance side as an example, the fitting process is described. The standard equation of the circle is known to be: x 2 +y 2 +Ax+By+C=0 (1) Where A, B, and C are the unknown parameters to be fitted; Equation (1) is transformed into: Ax+By+C=-x 2 -y 2 (2) The coordinates of k pixels (x1, y1), (x2, y2), ..., (x2, y2) on the center line of the strip on the entrance side obtained in step 3 are respectively... k ,y k Substituting into formula (2), we get: Equation (3) is converted into matrix form to obtain equation (4): According to the least squares method in matrix form, if: X·R=Y (5) Where X and Y are matrices with known parameters, and R is the matrix of unknowns to be fitted, then the least squares solution of R is: R=(X T X) -1 X T Y (6) Comparing equation (4) and equation (5), we can let: Substituting equations (7) to (9) into equation (6), we get: After obtaining A, B and C, substituting them into the standard equation of the circle in equation (1) will yield the equation of the fitted circle Φ1 on the entrance side. Step 4.2: Similarly, based on step 4.1, obtain the equation of the fitted circle Φ2 on the exit side.

5. The low-latency strip misalignment detection method based on machine vision as described in claim 3, characterized in that, Step 5 calculates the intersection points of the two fitted circles with the central axis of the roll, specifically as follows: Step 5.1: Taking the fitted circle Φ1 on the inlet side as an example, calculate the intersection point of the fitted circle Φ1 and the central axis of the roll, and obtain the circle equation x. 2 +y 2 After +Ax+By+C=0, let x=0, and we get: y 2 +By+C=0 (11) The deviation value ΔY1 on the inlet side is obtained as follows: If the center of the fitted circle Φ1 is above the center line of the strip, then: If the center of the fitted circle Φ1 is below the centerline of the strip, then: Step 5.2: Similarly, obtain the deviation value ΔY2 on the exit side according to Step 5.

1.

6. The low-latency strip misalignment detection method based on machine vision as described in claim 5, characterized in that, In step 5.1, the coordinates of the center of the fitted circle are calculated according to the following formula: Let the fitted circle on the entrance side be Φ1(a1,b1,r1), where a1, b1, and r1 are the x-coordinate of the center, y-coordinate of the center, and radius of Φ1, respectively. According to formula (1), we can obtain: Therefore, we get: Compare the ordinate b1 of the center of the fitted circle with the ordinate of the point on the strip centerline obtained in step 3. If the ordinate of the center is greater than the ordinate of the point on the strip centerline, the center of the fitted circle is above the strip centerline; otherwise, the center of the fitted circle is below the strip centerline.

7. The low-latency strip misalignment detection method based on machine vision as described in claim 5, characterized in that, In step 6, the strip deviation at the roll center is calculated according to the following formula.

Images