Multi-sensor strip steel width gauge and strip steel width measuring method

Abstract

The invention discloses a multi-sensor strip steel width gauge and a strip steel width measuring method.The width gauge comprises a light source fixedly arranged on a production line and two precision modules, the precision modules are located under a strip steel tension roller, and a line scanning camera used for collecting strip steel edge images is fixedly arranged on a sliding table of each precision module; the line-scan camera and the light source are respectively positioned on the front and rear sides of to-be-detected strip steel; a servo motor and a line scanning camera of the precision module are connected with a computer deployed with a sub-pixel-level edge detection algorithm based on interpolation, and the computer is connected with a strip steel production line control system through the Ethernet and used for obtaining the width and the transmission speed of strip steel. According to the width of the strip steel, the precision module is controlled to drive the line scanning camera to move to the edge position of the strip steel, the collection frame rate of the line scanning camera is adjusted according to the transmission speed of the strip steel, and the computer obtains the strip steel image collected by the line scanning camera and calculates the width of the strip steel. And the measurement precision is high.

Description

Technical Field

[0001] The invention belongs to the technical field of strip steel detection, and in particular relates to a multi-sensor strip steel width measuring instrument and a strip steel width measuring method. Background Art

[0002] Strip width is one of the main factors affecting the quality of steel pipes. Online width detection is an important part of strip production process control and finished product quality inspection. It is of great significance to improve the yield rate and is the key to improving the automation level of strip production. It is estimated that the yield rate can be increased by 0.1% for every 1mm reduction in width deviation. Therefore, welded pipe plants have increasingly stringent requirements on the width of strips. Improving the dynamic measurement accuracy of width gauges, thereby realizing width measurement throughout the entire process, is the development direction of online width measurement technology for hot-rolled strips.

[0003] The detection method of hot-rolled strip width has undergone a continuous evolution from manual sampling to offline and online detection. Among them, the earliest manual sampling was for operators to use micrometer calipers for contact measurement. This method is time-consuming and inaccurate, and cannot achieve real-time detection. In addition, the detection environment is harsh. Quality inspectors working in such an environment cannot be well guaranteed for safety and are prone to fatigue, which in turn affects the detection quality. At the same time, manual detection can only be performed after the strip has cooled down, and cannot detect steel plates with very high temperatures, and cannot play the role of real-time detection. With the increasing application of high-tech in engineering practice, CCD technology has realized non-contact dynamic width measurement of transmission steel plates, and the measurement results are directly displayed on the screen. , the strip width can be observed more intuitively and conveniently, which solves the labor intensity of workers to a greater extent. This non-contact measurement is mainly divided into two categories, namely offline measurement and online measurement, among which: laser width measurement is a more common offline measurement method. It uses the principle of symmetrical triangle reflection method, and the main device used is also CCD, which has good stability and high precision. However, when the calibration plate of laser width measurement is under the strip, the strip conveyor line must be stopped first, and then the laser is accurately calibrated by the calibration plate, and then the strip width is measured, so the online measurement of the strip width cannot be realized; online measurement mainly solves the problem of real-time online measurement of strip width, which is more convenient than offline measurement.

[0004] However, the existing online strip width measurement device has the following problems in the strip width measurement process: 1) The existing width gauge uses multiple line scan cameras to stitch images for measurement, and in the process of shooting the strip image, the line scan camera covers a larger field of view, resulting in lower detection accuracy; 2) In order to reduce the field of view covered by a single line scan camera, the number of line scan cameras must be increased, which not only leads to high equipment costs, but also increases the amount of software calculations; 3) When a single calibration plate is used on the production line, when measuring the width of strips with different width requirements, the line needs to be stopped to replace the calibration plate, which not only affects the efficiency of the production line, but also fails to achieve online measurement of the strip width, resulting in a decrease in the product qualification rate and an increase in unnecessary loss of raw materials. Summary of the invention

[0005] In view of the deficiencies in the prior art, the object of the present invention is to provide a multi-sensor strip width gauge and a strip width measuring method, which can not only measure the width of strips of different specifications in real time, but also have high measurement accuracy.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] A multi-sensor strip width gauge, comprising a light source fixedly installed on a production line and two left-right symmetrical precision modules, the precision modules are located directly below the strip tension roller, a line scan camera for collecting strip edge images is fixedly installed on the slide table of the precision modules, the line scan camera and the light source are respectively located at the front and rear sides of the strip to be measured, and the light source and the line scan camera have corresponding heights;

[0008] The servo motor and line scan camera of the precision module are connected to a computer equipped with a sub-pixel edge detection algorithm based on interpolation. The computer is connected to the control system of the strip production line via Ethernet to obtain the width and transmission speed of the strip. The computer controls the precision module to drive the line scan camera to move to the edge of the strip according to the width of the strip, and adjusts the acquisition frame rate of the line scan camera according to the transmission speed of the strip. The line scan camera uploads the acquired strip image to the computer, and the computer uses the sub-pixel edge detection algorithm based on interpolation to identify the position of the strip edge in the image coordinate system and calculate the width of the strip.

[0009] Furthermore, it also includes a base plate fixedly installed on the production line, and two precision modules are fixedly installed on the base plate and are symmetrical about the center of the base plate.

[0010] Furthermore, a line scan camera protection box and a line scan camera bracket are fixedly installed on the slide, the line scan camera bracket is located inside the line scan camera protection box, and the line scan camera is fixedly installed on the line scan camera bracket.

[0011] Furthermore, a motor cover is fixedly mounted on the bottom plate, and two servo motors are located inside the motor cover.

[0012] A method for measuring the width of a strip steel comprises the following steps:

[0013] Step 1: The computer obtains the width and transmission speed of the strip from the strip production line control system through the TCP communication protocol;

[0014] Step 2: The computer starts the servo motor according to the width of the strip, and the precision module drives the two line scan cameras to move to the edges of both sides of the strip, so that the edges of both sides of the strip are located within the field of view of the two line scan cameras;

[0015] Step 3: The computer first adjusts the acquisition frame rate of the line scan camera according to the transmission speed of the strip, and then triggers the line scan camera to take pictures of the edges of both sides of the strip by soft triggering;

[0016] Step 4: The line scan camera uploads the edge images of both sides of the strip to the computer. The computer first uses a sub-pixel edge detection algorithm based on interpolation to identify the position of the edge of the strip in the image coordinate system, and then converts the strip edge coordinates to the same coordinate system based on the positional relationship between the two line scan cameras obtained in advance. The edge pixels collected by the two line scan cameras at the same time are taken as a group of edge pixels, and then the Euclidean distance between the coordinates of a group of edge pixels is calculated. The Euclidean distance is the width of the strip.

[0017] Furthermore, the calculation formula of the acquisition frame rate of the line scan camera in step 3 is:

[0018] F c =H c ×(V0 / L0)

[0019] In the formula, H C is the resolution of the line scan camera, L O is the target field of view, V O is the transmission speed of the strip.

[0020] Furthermore, the specific process of step 4 using the interpolation-based sub-pixel edge detection algorithm to identify the position of the strip edge in the image coordinate system is as follows:

[0021] Step 4.1.1: On the edge contour of the strip, select n pixel points in the grayscale transition zone according to the change of pixel grayscale value. The coordinates and grayscale values ​​of these pixel points are recorded as (x i ,y i ,g i ), where i = 1, 2, ..., n;

[0022] Step 4.1.2: Perform bilinear interpolation around the selected n pixel points to obtain the gray value distribution of the strip steel image; the specific process is:

[0023] Step 4.1.2.1, set the gray value of any pixel (x, y) to g(x, y), and the gray values ​​of its four nearest neighboring pixels (x1, y1), (x1, y2), (x2, y1) and (x2, y2) to g1(x1, y1), g2(x1, y2), g3(x2, y1) and g4(x2, y2) respectively;

[0024] Step 4.1.2.2, interpolate in the x direction, and calculate the interpolation results R1 and R2 on the two lines y=y1 and y=y2, expressed as:

[0025]

[0026] Step 4.1.2.3, interpolate R1 and R2 in the y direction to obtain the grayscale value of the pixel point (x, y), expressed as:

[0027]

[0028] Step 4.1.3: Use edge detection algorithm to detect edges on the interpolated strip image, mark the potential edge positions in the strip image, and obtain the edge pixel set of the strip image, denoted as (x i ,y i ), where i = 1, 2, ..., n;

[0029] Step 4.1.4: Use the centroid method to calculate the centroid of each edge pixel (x c ,y c ), as the position of the edge of the strip in the image coordinates, is expressed as:

[0030]

[0031] In the formula, I(x i ,y i ) is the point (x i ,y i )’s pixel value or weighted value.

[0032] Furthermore, the edge detection algorithm of step 4.1.3 is a Canny edge detector, a Sobel operator or a Prewitt operator.

[0033] Furthermore, in step 4, according to the positional relationship between the two line scan cameras obtained in advance by calibration, the specific process of converting the strip edge coordinates into the same coordinate system is as follows:

[0034] Step 4.2.1, set the coordinate systems of the two line scan cameras, the two precision modules and the base plate to be O and Cl , O Cr , OAl , O Ar and O B ;

[0035] Step 4.2.2: Determine the coordinate transformation relationship T between the two precision modules and the base plate through measurement. AlB and T ArB , and then determine the coordinate system transformation relationship T between the two line scan cameras and the corresponding precision modules according to the moving distance of the precision module. CrAr and T ClAl ;

[0036] Step 4.2.3: After image processing, two line scan camera coordinate systems O are obtained. Cl and O Cr The corresponding left edge point P of the strip Cl (x Cl ,y Cl ,z Cl ) and the right edge point P Cr (x Cr ,y Cr ,z Cr ), and then through coordinate transformation, we can obtain the coordinates of the corresponding precision module coordinate system O Al and O Ar The coordinates P in Al (x Al ,y Al ,z Al ) and P Ar (x Ar ,y Ar ,z Ar ), expressed as:

[0037] P Al (x Al ,y Al ,z Al )=P Cl (x Cl ,y Cl ,z Cl )·T ClAl

[0038] P Ar (x Ar ,y Ar ,z Ar )=P Cr (x Cr ,y Cr ,z Cr )·T CrAr

[0039] Then, through coordinate transformation, we can obtain the left edge point and the right edge point in the base coordinate O. BThe coordinates P in Bl (x Bl ,y Bl ,z Bl ) and P Br (x Br ,y Br ,z Br ), expressed as:

[0040] P Br (x Br ,y Br ,z Br )=P Ar (x Ar ,y Ar ,z Ar )·T ArB

[0041] P Bl (x Bl ,y Bl ,z Bl )=P Al (x Al ,y Al ,z Al )·T AlB .

[0042] Furthermore, the width of the strip in step 4 is expressed as:

[0043]

[0044] In the formula, x Bl and x Br Respectively expressed as base coordinates O B The coordinate values ​​of the left and right edge points in the x direction.

[0045] Compared with the prior art, the present invention has the following technical effects:

[0046] 1) According to the width of the steel strip, a set of precision modules are used to drive two line scan cameras to move toward or away from each other, thereby adjusting the distance between the two line scan cameras so that the edge of the strip is always within the field of view of the first scan camera. This method is suitable for measuring steel strips of different widths without the need for additional line scan cameras, which not only reduces equipment costs and software calculations, but also eliminates the need to stop the machine to replace the calibration plate, thus achieving online real-time measurement.

[0047] 2) Use two line scan cameras to capture images within a small field of view. Moreover, adjust the frame rate of the line scan camera according to the transmission speed of the strip on the production line to ensure that the frame rate of the line scan camera matches the operating speed of the production line. The images captured by the two line scan cameras remain synchronized without causing distortion in the direction of strip movement, thereby improving the resolution of the image and improving the detection accuracy.

[0048] 3) The interpolation-based sub-pixel edge detection algorithm that combines sub-pixel and interpolation algorithms is used to identify the position of the strip edge in the image coordinate system, achieving pixel segmentation that exceeds the hardware limit, further improving the measurement accuracy, increasing the product qualification rate, and reducing raw material loss. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 : A schematic diagram of the three-dimensional structure of the multi-sensor strip width gauge of the present invention in use;

[0050] Figure 2 : The present invention Figure 1 Schematic diagram of the local enlarged structure in;

[0051] Figure 3 : Schematic diagram of the coordinate system of the line scan camera of the present invention;

[0052] Figure 4 : Schematic diagram of coordinate system conversion of the line scan camera of the present invention.

[0053] In the figure: 1. Base plate; 2. Precision module; 3. Slide; 4. Line scan camera bracket; 5. Line scan camera protection box; 6. Light source; 7. Line scan camera; 8. Strip steel; 9. Tension roller; 10. Servo motor; 11. Motor cover. DETAILED DESCRIPTION

[0054] The specific contents of the present invention are further explained in detail below in conjunction with embodiments.

[0055] like Figure 1 and Figure 2 As shown, a multi-sensor strip steel width gauge includes a bottom plate 1 and a light source 6 fixedly installed on a production line, two precision modules 2 symmetrically arranged about the center are fixedly installed on the bottom plate 1, the precision module 2 is located directly below the strip steel tension roller 9, a line scan camera protection box 5 and a line scan camera bracket 4 are fixedly installed on the slide 3 of the precision module 2 by screws, the line scan camera bracket 4 is located inside the line scan camera protection box 5, and a line scan camera 7 for collecting an edge image of the strip steel 8 is fixedly installed on the line scan camera bracket 4 by screws, the line scan camera 7 and the light source 6 are respectively located at the front and rear sides of the strip steel to be measured, and the light source 6 corresponds to the height of the line scan camera 7;

[0056] The precision module 2 is driven by a servo motor 10. The servo motor 10 and the line scan camera 7 are connected to a computer equipped with a sub-pixel edge detection algorithm based on interpolation. The computer is connected to the strip steel production line control system via Ethernet to obtain the width and transmission speed of the strip steel 8. The computer controls the precision module 2 to drive the line scan camera 7 to move to the edge of the strip steel according to the width of the strip steel 8, and adjusts the acquisition frame rate of the line scan camera 7 according to the transmission speed of the strip steel 8.

[0057] A motor cover 11 is also fixedly mounted on the base plate 1 by screws. Two servo motors 10 are located inside the motor cover 11. The motor cover 11 is used to protect the servo motors 10 to prevent accidental damage.

[0058] Preferably, this embodiment is suitable for measuring strip steel with a width of 1500 mm to 2000 mm.

[0059] Preferably, the angle between the moving direction of the steel strip 8 and the plane where the precision module 2 is located is 90°, and the steel strip 8 enters the measuring range of the line scan camera 7 by the rotation of the tension roller 9 .

[0060] In this embodiment, two line scan cameras 7 are used to collect edge contour images of the steel strip and upload them to a computer. The computer uses a sub-pixel edge detection algorithm based on interpolation to determine the edge contour of the steel strip. The coordinate relationship of the two line scan cameras 7 is obtained by calibration, and the edge coordinates are obtained according to the calibration relationship. The Euclidean distance between the two edges of the steel strip 8 is obtained by the two edge coordinates, which is the width of the steel strip. If the algorithm processing result shows that it is not the edge of the steel strip to be measured, the precision module 2 is used to drive the two line scan cameras 7 to continue to move left and right, thereby adjusting the scanning position of the line scan camera 7 to ensure that the edge of the steel strip to be measured is always within the field of view of the line scan camera 7.

[0061] A strip steel width measuring method based on a multi-sensor strip steel width measuring instrument comprises the following steps:

[0062] Step 1, the computer obtains the width and transmission speed of the strip steel 8 from the strip steel production line control system through the TCP communication protocol;

[0063] When the strip steel width gauge is started, or when the type of product on the production line is changed, the computer is connected to the control system of the strip steel production line using a network cable with the help of Ethernet communication. Using the TCP communication protocol, the computer obtains the product parameters of the currently produced strip steel 8 from the control system of the production line, mainly the width of the strip steel 8 and the transmission speed of the strip steel 8 on the production line;

[0064] Step 2: The computer starts the servo motor 10 according to the width of the steel strip 8, controls the precision module 2 to drive the two line scan cameras 7 to move to the edge positions on both sides of the steel strip, so that the edges on both sides of the steel strip are respectively located within the field of view of the two line scan cameras 7;

[0065] Step 3: The computer first adjusts the acquisition frame rate of the line scan camera 7 according to the transmission speed of the steel strip 8, so that the acquisition frame rate of the line scan camera 7 matches the running speed of the production line, that is, the line frequency matches, so that the images acquired by the two line scan cameras 7 remain synchronized, and no distortion is generated in the moving direction of the steel strip 8. After the software is started, the line scan camera 7 will acquire images according to the running speed of the production line, that is, the line scan camera 7 is triggered by soft triggering to take pictures of the edges of both sides of the steel strip. The calculation formula of the acquisition frame rate of the line scan camera 7 is:

[0066] F c =H c ×(V0 / L0)

[0067] In the formula, H C is the resolution of the line scan camera, L O is the target field of view, V O is the transmission speed of the strip;

[0068] Since the line scan camera 7 continuously collects images, a fixed-length image cache is set for the line scan camera 7. The image cache timing is determined by setting the number of image lines to ensure that the image size is consistent each time it is processed.

[0069] Step 4: The line scan camera 7 uploads the edge images of the strip 8 to the computer. The computer identifies the edge position of the image and measures the width of the strip. The process is as follows:

[0070] Step 4.1: The computer first uses a sub-pixel edge detection algorithm based on interpolation to identify the position of the edge of the steel strip 8 in the image coordinate system. The specific process is as follows:

[0071] Step 4.1.1, on the edge contour of the steel strip 8, select n pixel points in the grayscale transition zone according to the change of the pixel grayscale value, and the coordinates and grayscale values ​​of these pixel points are recorded as (x i ,y i ,g i ), where i=1, 2, ..., n, and the selected n pixels represent the area where the gray value changes sharply from the background to the steel strip 8 or from the steel strip 8 to the background;

[0072] A small search area is set near the detected pixel-level edge of the strip, usually with a width and height of several pixels. This search area is used for sub-pixel positioning. For each sub-pixel point in the search area, that is, a pixel point with non-integer coordinates, in order to calculate the position of the sub-pixel edge, bilinear interpolation is used to calculate its grayscale value. The basic idea of ​​bilinear interpolation is to perform linear interpolation in two directions respectively.

[0073] Step 4.1.2: Perform bilinear interpolation around the selected n pixel points to obtain the gray value distribution of the strip steel image; the specific process is:

[0074] Step 4.1.2.1, set the gray value of any pixel (x, y) to g(x, y), and the gray values ​​of its four nearest neighboring pixels (x1, y1), (x1, y2), (x2, y1) and (x2, y2) to g1(x1, y1), g2(x1, y2), g3(x2, y1) and g4(x2, y2) respectively;

[0075] Step 4.1.2.2, interpolate in the x direction, and calculate the interpolation results R1 and R2 on the two lines y=y1 and y=y2, expressed as:

[0076]

[0077] Step 4.1.2.3, interpolate R1 and R2 in the y direction to obtain the grayscale value of the pixel point (x, y), expressed as:

[0078]

[0079] Step 4.1.3: Use edge detection algorithm to detect edges on the interpolated strip image, mark the potential edge positions in the strip image, and obtain the edge pixel set of the strip image, denoted as (x i ,y i ), where i = 1, 2, ..., n;

[0080] The edge detection algorithm may be a Canny edge detector, a Sobel operator or a Prewitt operator;

[0081] Step 4.1.4: Use the centroid method to calculate the centroid of each edge pixel (x c ,y c ), as the position of the edge of the strip steel 8 in the image coordinates, is expressed as:

[0082]

[0083] In the formula, I(x i ,y i ) is the point (x i ,y i )’s pixel value or weighted value;

[0084] The centroid method can further refine the edge position of the strip 8 to the sub-pixel level, and the centroid of the edge pixel (x c ,y c), as a more accurate edge position of the edge of the strip 8 in the image coordinates, which is then converted into the base coordinate system O after coordinate transformation. B The points in the middle are sampled at equal intervals and the width is calculated to obtain the width information of the entire strip steel 8, and the measurement data is saved;

[0085] Step 4.2: The computer converts the edge coordinates of the strip steel 8 into the same coordinate system according to the pre-calibrated positional relationship between the two line scan cameras 7. The specific process is as follows:

[0086] Step 4.2.1: Figure 3 As shown, the coordinate systems of the two line scan cameras 7, the two precision modules 2 and the base plate 1 are respectively set to be Cl , O Cr , O Al , O Ar and O B Specifically, for the convenience of distinction, according to Figure 1 The direction of O Cl is the coordinate system of the left line scan camera 7, O Cr is the coordinate system of the right line scan camera 7, O Al is the coordinate system of the left precision module, O Ar is the coordinate system of the right precision module;

[0087] Step 4.2.2: Determine the coordinate transformation relationship T between the two precision modules 2 and the base plate 1 through measurement. AlB and T ArB , and then determine the coordinate system transformation relationship T between the two line scan cameras 7 and the corresponding precision module 2 according to the moving distance of the precision module 2 CrAr and T ClAl ;

[0088] Step 4.2.3, if Figure 4 As shown, after image processing, two line scan camera 7 coordinate systems O are obtained respectively. Cl and O Cr The corresponding left edge point P of the strip Cl (x Cl ,y Cl ,z Cl ) and the right edge point P Cr (x Cr ,y Cr ,z Cr ), and then through coordinate conversion, we can obtain the coordinate system O of the corresponding precision module 2. Al and O Ar The coordinates P in Al (x Al ,y Al ,z Al ) and PAr (x Ar ,y Ar ,z Ar ), expressed as:

[0089] P Al (x Al ,y Al ,z Al )=P Cl (x Cl ,y Cl ,z Cl )·T ClAl

[0090] P Ar (x Ar ,y Ar ,z Ar )=P Cr (x Cr ,y Cr ,z Cr )·T CrAr

[0091] Then, through coordinate transformation, we can obtain the left edge point and the right edge point in the base coordinate O. B The coordinates P in Bl (x Bl ,y Bl ,z Bl ) and P Br (x Br ,y Br ,z Br ), expressed as:

[0092] P Br (x Br ,y Br ,z Br )=P Ar (x Ar ,y Ar ,z Ar )·T ArB

[0093] P Bl (x Bl ,y Bl ,z Bl )=P Al (x Al ,y Al ,z Al )·T AlB

[0094] Step 4.3: The edge pixels collected by the two line scan cameras 7 at the same time are taken as a group of edge pixels, and then the Euclidean distance between the coordinates of a group of edge pixels is calculated. The Euclidean distance is the width of the steel strip 8, which is expressed as:

[0095]

[0096] In the formula, x Bl and x Br Respectively expressed as base coordinates O B The coordinate values ​​of the left and right edge points in the x direction.

Claims

1. A multi-sensor strip width gauge, characterized in that: The invention comprises a light source (6) fixedly installed on a production line and two left-right symmetrical precision modules (2), wherein the precision module (2) is located directly below a steel strip tension roller (9), and a line scan camera (7) for collecting edge images of a steel strip (8) is fixedly installed on a slide table (3) of the precision module (2), wherein the line scan camera (7) and the light source (6) are respectively located on the front and rear sides of the steel strip to be tested, and the light source (6) and the line scan camera (7) are at corresponding heights; The servo motor (10) and the line scan camera (7) of the precision module (2) are both connected to a computer equipped with a sub-pixel edge detection algorithm based on interpolation. The computer is connected to a strip steel production line control system via Ethernet and is used to obtain the width and transmission speed of the strip steel (8). The computer controls the precision module (2) to drive the line scan camera (7) to move to the edge position of the strip steel according to the width of the strip steel (8), and adjusts the acquisition frame rate of the line scan camera (7) according to the transmission speed of the strip steel (8). The line scan camera (7) uploads the acquired strip steel image to the computer. The computer uses the sub-pixel edge detection algorithm based on interpolation to identify the position of the strip steel edge in the image coordinate system and calculates the width of the strip steel (8).

2. The multi-sensor strip width gauge according to claim 1, characterized in that: It also includes a base plate (1) fixedly mounted on the production line, and two precision modules (2) fixedly mounted on the base plate (1) and symmetrically disposed about the center of the base plate (1).

3. The multi-sensor strip width gauge according to claim 1 or 2, characterized in that: A line scan camera protection box (5) and a line scan camera bracket (4) are fixedly mounted on the slide (3); the line scan camera bracket (4) is located inside the line scan camera protection box (5); and the line scan camera (7) is fixedly mounted on the line scan camera bracket (4).

4. The multi-sensor strip width gauge according to claim 2, characterized in that: A motor cover (11) is fixedly mounted on the bottom plate (1), and two servo motors (10) are located inside the motor cover (11).

5. A strip width measuring method based on the multi-sensor strip width gauge according to claim 2, characterized in that: The steps include: Step 1, the computer obtains the width and transmission speed of the strip steel (8) from the strip steel production line control system through the TCP communication protocol; Step 2: The computer starts the servo motor (10) according to the width of the steel strip (8), and the precision module (2) drives the two line scan cameras (7) to move to the edge positions on both sides of the steel strip, so that the edges on both sides of the steel strip are respectively located within the field of view of the two line scan cameras (7); Step 3, the computer first adjusts the acquisition frame rate of the line scan camera (7) according to the transmission speed of the steel strip (8), and then triggers the line scan camera (7) to take pictures of the edges of both sides of the steel strip by soft triggering; Step 4: The line scan camera (7) uploads the edge images of both sides of the steel strip (8) collected to a computer. The computer first uses a sub-pixel edge detection algorithm based on interpolation to identify the position of the edge of the steel strip (8) in the image coordinate system, and then converts the edge coordinates of the steel strip (8) into the same coordinate system based on the positional relationship between the two line scan cameras (7) obtained in advance. The edge pixels collected by the two line scan cameras (7) at the same time are taken as a group of edge pixels, and then the Euclidean distance between the coordinates of a group of edge pixels is calculated. The Euclidean distance is the width of the steel strip (8).

6. The strip width measuring method of the multi-sensor strip width gauge according to claim 5, characterized in that: The calculation formula of the acquisition frame rate of the line scan camera (7) in step 3 is: F c =H c ×(V0 / L0) In the formula, H C is the resolution of the line scan camera, L O is the target field of view, V O is the transmission speed of the strip.

7. The strip width measuring method of the multi-sensor strip width gauge according to claim 5, characterized in that: The specific process of step 4 using the interpolation-based sub-pixel edge detection algorithm to identify the position of the edge of the steel strip (8) in the image coordinate system is as follows: Step 4.1.1, on the edge contour of the steel strip (8), select n pixel points in the grayscale transition zone according to the change of the pixel grayscale value, and the coordinates and grayscale values ​​of these pixel points are recorded as (x i ,y i ,g i ), where i = 1, 2, ..., n; Step 4.1.2: Perform bilinear interpolation around the selected n pixel points to obtain the gray value distribution of the strip steel image; the specific process is: Step 4.1.2.1, set the gray value of any pixel (x, y) to g(x, y), and the gray values ​​of its four nearest neighboring pixels (x1, y1), (x1, y2), (x2, y1) and (x2, y2) to g1(x1, y1), g2(x1, y2), g3(x2, y1) and g4(x2, y2) respectively; Step 4.1.2.2, interpolate in the x direction, and calculate the interpolation results R1 and R2 on the two lines y=y1 and y=y2, expressed as: Step 4.1.2.3, interpolate R1 and R2 in the y direction to obtain the grayscale value of the pixel point (x, y), expressed as: Step 4.1.3: Use edge detection algorithm to detect edges on the interpolated strip image, mark the potential edge positions in the strip image, and obtain the edge pixel set of the strip image, denoted as (x i ,y i ), where i = 1, 2, ..., n; Step 4.1.4: Use the centroid method to calculate the centroid of each edge pixel (x c ,y c ), as the position of the edge of the strip (8) in the image coordinates, is expressed as: In the formula, I(x i ,y i ) is the point (x i ,y i )’s pixel value or weighted value.

8. The strip width measuring method of the multi-sensor strip width gauge according to claim 7, characterized in that: The edge detection algorithm of step 4.1.3 is a Canny edge detector, a Sobel operator or a Prewitt operator.

9. The strip width measuring method of the multi-sensor strip width gauge according to claim 5, characterized in that: In step 4, the specific process of converting the edge coordinates of the steel strip (8) into the same coordinate system according to the positional relationship between the two line scan cameras (7) obtained by pre-calibration is as follows: Step 4.2.1, respectively set the coordinate systems of the two line scan cameras (7), the two precision modules (2) and the base plate (1) to be O Cl , O Cr , O Al , O Ar and O B ; Step 4.2.2: Determine the coordinate transformation relationship T between the two precision modules (2) and the base plate (1) through measurement. AlB and T ArB , and then determine the coordinate system conversion relationship T between the two line scan cameras (7) and the corresponding precision module (2) according to the moving distance of the precision module (2). CrAr and T ClAl ; Step 4.2.3: After image processing, the coordinate systems O of the two line scan cameras (7) are obtained respectively. Cl and O Cr The corresponding left edge point P of the strip (8) Cl (x Cl ,y Cl ,z Cl ) and the right edge point P Cr (x Cr ,y Cr ,z Cr ), and then through coordinate conversion, we can obtain the coordinate system O of the corresponding precision module (2) Al and O Ar The coordinates P in Al (x Al ,y Al ,z Al ) and P Ar (x Ar ,y Ar ,z Ar ), expressed as: P Al (x Al ,y Al ,z Al )=P Cl (x Cl ,y Cl ,z Cl )·T ClAl P Ar (x Ar ,y Ar ,z Ar )=P Cr (x Cr ,y Cr ,z Cr )·T CrAr Then, through coordinate transformation, we can obtain the left edge point and the right edge point in the base coordinate O. B The coordinates P in Bl (x Bl ,y Bl ,z Bl ) and P Br (x Br ,y Br ,z Br ), expressed as: P Br (x Br ,y Br ,z Br )=P Ar (x Ar ,y Ar ,z Ar )·T ArB P Bl (x Bl ,y Bl ,z Bl )=P Al (x Al ,y Al ,z Al )·T AlB 。 10. The strip width measuring method of the multi-sensor strip width gauge according to claim 9, characterized in that: The width of the strip steel (8) in step 4 is expressed as: In the formula, x Bl and x Br Respectively expressed as the base coordinate system O B The coordinate values ​​of the left and right edge points in the x direction.

Images