Large-diameter steel pipe contour data acquisition method based on multi-line laser sensor

Abstract

The invention discloses a large-diameter steel pipe contour data acquisition method based on multiple line laser sensors. The method comprises the following steps: constructing a steel pipe contour data acquisition mechanism based on the multiple line laser sensors; calibrating a rotation translation matrix of each line laser sensor; the method comprises the following steps: collecting steel pipe contour segmentation data, and splicing the data to obtain initial contour data; and reconstructing the contour data to generate uniformly distributed final contour data. According to the method, the section contour data of the large-diameter steel pipe is acquired based on the cooperation of the plurality of line laser sensors, the section contour information of the large-diameter steel pipe can be quickly and accurately acquired, linear regression clustering is performed on the contour of the calibration block by using the FCRM algorithm, and the contour linear equation of the calibration block under the local coordinate system of the sensor can be accurately determined; therefore, accurate contour angular point coordinates are obtained, the splicing precision of steel pipe segment contour data is improved, the obtained steel pipe section contour data are more complete and accurate, and high-precision detection of the geometric dimension of the large-diameter steel pipe is ensured.

Description

Technical Field

[0001] The invention belongs to the technical field of laser measurement, and in particular relates to a method for collecting contour data of a large-diameter steel pipe based on a multi-line laser sensor. Background Art

[0002] In oil and gas pipeline engineering, the precise measurement of steel pipe dimensions is directly related to the safety and reliability of the pipeline system. As pipeline construction evolves toward higher steel grades, larger diameters, and higher pressures, more stringent requirements are placed on the accuracy and efficiency of measuring geometric dimensions such as pipe diameter and out-of-roundness.

[0003] The accuracy of steel pipe geometry is crucial for ensuring the progress and quality of pipeline construction. Substandard geometry not only increases construction difficulty but also impacts pipeline safety, potentially leading to leaks, explosions, and other accidents, causing environmental pollution and significant economic losses. Therefore, to effectively mitigate these potential risks, precise, rigorous, and meticulous testing of steel pipe geometry is crucial. To achieve high-precision measurements, the primary task is to accurately capture the profile data of the steel pipe cross section. After all, the efficiency of data acquisition directly determines the duration of the entire inspection process.

[0004] Currently, the industry mainly relies on a method based on circular traversal to obtain the profile data of steel pipe cross sections. However, this method requires the collaborative operation of a point laser sensor and a precision mechanism to obtain the profile information of a cross section by rotating around the steel pipe once. This is time-consuming and greatly restricts the detection efficiency. Even more problematic is that this method is incapable of collecting partial information on the steel pipe body, making it difficult to achieve all-round, no-dead-angle measurement of the geometric dimensions of all parts of the entire steel pipe. In view of this, how to improve the accuracy and efficiency of collecting steel pipe cross-sectional profile data has become a major problem that needs to be solved urgently in the field of oil and gas pipeline engineering. Summary of the Invention

[0005] The purpose of the present invention is to provide a large-diameter steel pipe profile data acquisition method based on a multi-line laser sensor to solve the problem of low efficiency of the existing steel pipe cross-section profile data acquisition method based on circumferential traversal.

[0006] The technical solution adopted by the present invention is a method for collecting contour data of large-diameter steel pipes based on a multi-line laser sensor, which specifically includes the following steps: Step 1: Construct a steel pipe profile data acquisition mechanism based on multiple line laser sensors; Step 2: Calibrate the rotation and translation matrix of each line laser sensor; Step 3: Collect the segmented data of the steel pipe contour and splice them to obtain the initial contour data; Step 4: Contour data reconstruction to generate final contour data with uniform distribution.

[0007] The present invention is also characterized in that: Step 1 The specific steps are as follows: Step 1.1: Mechanism Design: Design and construct a mechanical structure specifically for collecting steel pipe profile data. The core of this structure integrates N (N > 2) line laser sensors. These sensors are arranged diagonally to ensure full coverage of the steel pipe's cross-section, capturing the most complete profile information.

[0008] Step 1.2: Installation and Fixing: Firmly install N line laser sensors on the mechanical structure to ensure that they remain stable during the acquisition process to avoid vibration or displacement that may affect the accuracy of data acquisition. The specific steps are as follows: Step 1.2.1: Build the gantry, making sure it is perpendicular to the roller conveyor; Step 1.2.2: Install the sensors. Install N line laser sensors in corresponding positions using a customized fixture. Step 1.2.3: Adjust the sensor angles, including the angle of rotation around its horizontal axis (pitch angle) and the free angle of rotation around its vertical axis (azimuth angle), so that the laser beams of N sensors are located in the same plane and there is overlap between the N laser lines.

[0009] Step 2 is as follows: Step 2.1: Collect the calibration block contour data: The calibration block is a regular N-gon with a side length of S. Each line laser sensor collects the calibration block contour data within its field of view; Step 2.2: Determine the contour line equation and contour corner coordinates in the local coordinate system of each sensor: Use the FCRM algorithm to perform regression clustering on the calibration block contour data collected by each line laser sensor to determine the equation of the line on which the contour of each side of the calibration block lies, thereby determining the local coordinates of the calibration block contour corner points; Step 2.2 The method for determining the contour line equation and contour corner coordinates in the local coordinate system of each sensor is: The FCRM algorithm is used to perform regression clustering on the calibration block contour data collected by each line laser sensor. The specific process is as follows: Step 2.2.1: Set the number of cluster models c=2, the fuzziness value m=2, and the iteration termination threshold , set the number of iterations , and initialize the membership matrix , recorded as ,in: For line c A matrix of columns, For data pairs The number of for Matrix Rank The membership value of the column, , , the elements in the membership matrix Constraints must be met: .

[0010] Step 2.2.2: According to the formula Update model parameter matrix , . Among them: matrix is the output value matrix, is the augmented input vector The Vandermonde matrix is ​​formed, For the first ( ) iterations, the diagonal membership matrix, ; is the straight line intercept, is the slope.

[0011] Step 2.2.3: Based on the results from the previous step Update the membership matrix .

[0012]

[0013] in: Indicates the For data on The error of the model, if , then the model number Add to collection Otherwise, put it into the collection middle.

[0014] Step 2.2.4: Determine whether to terminate the iteration. If ,but , and return to step 2.2.2; otherwise, the iteration terminates, is the final model parameter matrix.

[0015] Step 2.2.5: Find the coordinates of the intersection of the two straight lines in the local coordinate system. These coordinates are the coordinates of the contour corner points. The equations of the straight lines of the two adjacent contours of the calibration block are determined by the model parameters obtained in step 2.2.4. and , then the coordinates of the intersection of these two sides are , which is the coordinate of the contour corner point in the local coordinate system of the sensor.

[0016] Step 2.3: Determine the vertex coordinates and straight line equations of the calibration block in the world coordinate system: Take the center of the calibration block as the origin, the horizontal line through the origin as the x-axis, and the y-axis through the origin and perpendicular to the x-axis as the world coordinate system. Since the calibration block is a regular N-gon with a side length of S, the coordinates of the vertices of the calibration block are ,in: , , is the vertex number (counterclockwise), is the angle between the first vertex and the line where the origin is and the horizontal axis. Finally, the linear equation of each side is determined by two adjacent vertices. ,in: , , , the side where the first and last points are located , .

[0017] Step 2.4: Solve the rotation and translation matrix of each sensor: First, determine the rotation matrix of each sensor's local coordinate system to the world coordinate system based on the angular deviation of the equations of each side in the local coordinate system and the world coordinate system; then use the rotation matrix to rotate the vertex coordinates of the local coordinate system; finally, determine the translation matrix of each sensor based on the coordinate deviation of each vertex after the rotation transformation and each vertex in the world coordinate system. The specific process is as follows: Step 2.4.1: Match the corresponding edges of the calibration block cross section in the world coordinate system according to the actual installation position of the sensor. The line laser sensor at the i-th position obtains the contours of the two edges of the calibration block where the i-th corner is located; Step 2.4.2: Calculate the rotation angle and rotation matrix: Solve the equation of the line of one side of the calibration block in the world coordinate system and the equation of the line of the line laser sensor in the local coordinate system. The direction vectors of the two corresponding lines are 、 , then the rotation angle of the local coordinate system relative to the world coordinate system is , the rotation matrix .

[0018] Step 2.4.3: Calculate the translation vector: by calibrating the coordinates of a vertex of the block in the world coordinate system The coordinates of the corner points in the local coordinate system of the sensor after rotation transformation Solve based on the translation vector .

[0019] Step 3 is as follows: Step 3.1: Collecting steel pipe segment profile data: Each line laser sensor can collect the profile data of part of the steel pipe section in its local coordinate system; Step 3.2: Splice the steel pipe segment profile data and unify them into the same world coordinate system to obtain complete profile data: Use the rotation and translation matrix obtained in step 2 to perform coordinate transformation on the partial steel pipe cross-section profile data collected by each line laser sensor. The data coordinates in the world coordinate system are obtained by the formula calculate.

[0020] The specific process of step 4 is: Step 4.1: Contour center positioning: Use the least squares method to perform ellipse fitting on the complete contour data obtained in step 3 to obtain the center coordinates of the steel pipe cross-section contour ( ); Step 4.2: Coordinate translation: Subtract the center coordinate from the complete contour data obtained in step 3 ( ), the contour data centered at the origin; Step 4.3: Coordinate transformation: convert the contour data obtained in step 3.1 into polar coordinates; Step 4.4: Linear interpolation: Use linear interpolation to make the coordinate points of each segment of data be spaced at intervals. ° ( The value is 0.1~1) and is evenly distributed; the angle of the first point of the contour data is taken The angle with the last point , and keep one decimal place to get and , we can determine that the contour of this segment has measuring points, calculate the angle of p measuring points °, where h is the index number of the measurement point, then in each contour data obtained in step 4.3, find the angle between the geometric position and the measurement point The two closest data points are located at the measurement point angles. The left and right sides of these two data points are used as the reference angle Perform linear interpolation to obtain The polar diameter of the angle.

[0021] Step 4.5: Process the overlapping parts of the contours; put the contour data obtained in step 4.4 together, if there is an angle The same two points, that is, find the average value of the polar radius of these two points as the angle The final polar diameter value of the measuring point; Step 4.6: Coordinate transformation: transform the coordinates obtained in step 4.5 By converting the standardized measurement data points into rectangular coordinates, the final standardized reconstructed steel pipe contour data can be obtained.

[0022] The beneficial effects of the present invention are: 1. The present invention is based on a large-diameter steel pipe profile data acquisition method using multiple line laser sensors. Multiple line laser sensors are used to collaboratively collect steel pipe cross-sectional profile data. Each sensor can collect partial profile data under its local coordinates. Multiple segments of profile data are then unified into the same world coordinate system to obtain complete steel pipe cross-sectional profile data. Overlapping profile data is processed through operations such as coordinate transformation and linear interpolation, ultimately obtaining uniformly distributed final profile data. This lays a good foundation for subsequent geometric parameter calculations. This method, based on the high efficiency of multiple line laser sensor fusion and data processing algorithms, can quickly and accurately collect steel pipe cross-sectional profile data, solving the problem of low efficiency in steel pipe profile data acquisition methods based on circular traversal.

[0023] 2. The original contour data segmentation method automatically locates the segmentation points based on the changing characteristics of the data. By calculating the slope or curvature between adjacent data points, the inflection point with the largest change can be found, and then the segmentation point can be determined. This method is quite effective when the data changes are relatively stable, and can better reflect the changing trend of the data. However, when collecting data, the line laser sensor is susceptible to external interference, and the surface contour data of the calibration block obtained is often mixed with noise points. In addition, due to the limitation of the sensor's own accuracy, the data itself will also fluctuate, resulting in the slope or curvature at the segmentation point not necessarily being the largest. Therefore, the present invention introduces clustering analysis technology, and uses the FCRM clustering algorithm to perform regression clustering on the cross-sectional contour data of the calibration block, which can accurately determine the fitting straight line and corner points of the contour segment, lay a solid foundation for the calculation of subsequent transformation information, make the solution of the system calibration parameters more accurate, and thus improve the accuracy of the steel pipe segment contour data splicing. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a flow chart of the method for collecting contour data of large-diameter steel pipes based on a multi-line laser sensor of the present invention; Figure 2 This is a schematic diagram of a large-diameter steel pipe profile data acquisition mechanism based on four line laser sensors constructed by the present invention; Figure 3 is a schematic diagram of a calibration block used in the present invention; DETAILED DESCRIPTION The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0025] Example 1 The present invention utilizes a steel pipe profile data acquisition mechanism based on multiple line laser sensors to acquire steel pipe cross-sectional profile data in segments; utilizes a calibration block and the FCRM algorithm to accurately acquire the rotation and translation matrices for transforming the sensor's local coordinate system into the world coordinate system; utilizes these rotation and translation matrices to perform coordinate system transformation on the steel pipe segmented profile data to obtain a complete cross-sectional profile; finally, utilizes a linear interpolation method to standardize the profile data, eliminate overlapping profiles, and reconstruct standard profile data.

[0026] Example 2 The large-diameter steel pipe profile data acquisition method based on the multi-line laser sensor of the present invention is as follows: Figure 1 As shown, please follow the steps below: Step 1: Construct a steel pipe profile data acquisition mechanism based on multiple line laser sensors; Step 1.1: Mechanism Design: Design and construct a mechanical structure specifically for collecting steel pipe profile data. The core of this structure integrates N (N > 2) line laser sensors. These sensors are arranged diagonally to ensure full coverage of the steel pipe's cross-section, capturing the most complete profile information.

[0027] Step 1.2: Installation and Fixing: Firmly install N line laser sensors on the mechanical structure to ensure that they remain stable during the acquisition process to avoid vibration or displacement that may affect the accuracy of data acquisition. The specific steps are as follows: Step 1.2.1: Build the gantry, making sure it is perpendicular to the roller conveyor; Step 1.2.2: Install the sensors. Install N line laser sensors in corresponding positions using a customized fixture. Step 1.2.3: Adjust the sensor angles, including the angle of rotation around its horizontal axis (pitch angle) and the free angle of rotation around its vertical axis (azimuth angle), so that the laser beams of N sensors are located in the same plane and there is overlap between the N laser lines.

[0028] Step 2: Calibrate the rotation and translation matrix of each line laser sensor; Step 2.1: Collect the calibration block contour data: The calibration block is a regular N-gon with a side length of S. Each line laser sensor collects the calibration block contour data within its field of view; Step 2.2: Determine the contour line equation and contour corner coordinates in the local coordinate system of each sensor: Use the FCRM algorithm to perform regression clustering on the calibration block contour data collected by each line laser sensor to determine the equation of the line on which the contour of each side of the calibration block lies, thereby determining the local coordinates of the calibration block contour corner points; Step 2.3: Determine the vertex coordinates and straight line equations of the calibration block in the world coordinate system: Take the center of the calibration block as the origin, the horizontal line through the origin as the x-axis, and the y-axis through the origin and perpendicular to the x-axis as the world coordinate system. Since the calibration block is a regular N-gon with a side length of S, the coordinates of the vertices of the calibration block are ,in: , , is the vertex number (counterclockwise), is the angle between the first vertex and the line where the origin is and the horizontal axis. Finally, the linear equation of each side is determined by two adjacent vertices. ,in: , , , the side where the first and last points are located , .

[0029] Step 2.4: Solve the rotation and translation matrices of each sensor: First, determine the rotation matrix of each sensor's local coordinate system to the world coordinate system based on the angular deviation between the equations of each side in the local coordinate system and the world coordinate system. Then, use the rotation matrix to rotate the vertex coordinates of the local coordinate system. Finally, determine the translation matrix of each sensor based on the coordinate deviation between each vertex after the rotation transformation and each vertex in the world coordinate system.

[0030] Step 3: Collect the segmented data of the steel pipe contour and splice them to obtain the initial contour data; Step 4: Contour data reconstruction to generate final contour data with uniform distribution.

[0031] Example 3 Based on Example 2, step 2.2 is as follows: The FCRM algorithm is used to perform regression clustering on the calibration block contour data collected by each line laser sensor. The specific process is as follows: Step 2.2.1: Set the number of cluster models c=2, the fuzziness value m=2, and the iteration termination threshold , set the number of iterations , and initialize the membership matrix , recorded as ,in: For line c A matrix of columns, For data pairs The number of for Matrix Rank The membership value of the column, , , the elements in the membership matrix Constraints must be met: .

[0032] Step 2.2.2: According to the formula Update model parameter matrix , . Among them: matrix is the output value matrix, is the augmented input vector The Vandermonde matrix is ​​formed, For the first ( ) iterations, the diagonal membership matrix, ; is the straight line intercept, is the slope; Step 2.2.3: Based on the results from the previous step Update the membership matrix .

[0033]

[0034] in: Indicates the For data on The error of the model, if , then the model number Add to collection Otherwise, put it into the collection middle.

[0035] Step 2.2.4: Determine whether to terminate the iteration. If ,but , and return to step 2.2.2; otherwise, the iteration terminates, is the final model parameter matrix.

[0036] Step 2.2.5: Find the coordinates of the intersection of the two straight lines in the local coordinate system. These coordinates are the coordinates of the contour corner points. The equations of the straight lines of the two adjacent contours of the calibration block are determined by the model parameters obtained in step 2.2.4. and , then the coordinates of the intersection of these two sides are , which is the coordinate of the contour corner point in the local coordinate system of the sensor.

[0037] Example 4 Based on Example 3, step 2.4 is as follows: Step 2.4.1: Match the corresponding edges of the calibration block cross section in the world coordinate system according to the actual installation position of the sensor. The line laser sensor at the i-th position obtains the contours of the two edges of the calibration block where the i-th corner is located; Step 2.4.2: Calculate the rotation angle and rotation matrix: Solve the equation of the line of one side of the calibration block in the world coordinate system and the equation of the line of the line laser sensor in the local coordinate system. The direction vectors of the two corresponding lines are 、 , then the rotation angle of the local coordinate system relative to the world coordinate system is , the rotation matrix .

[0038] Step 2.4.3: Calculate the translation vector: by calibrating the coordinates of a vertex of the block in the world coordinate system The coordinates of the corner points in the local coordinate system of the sensor after rotation transformation Solve based on the translation vector .

[0039] Example 5 Based on Example 4, Step 3: Collect the segmented data of the steel pipe contour and splice them to obtain the initial contour data; Step 3.1: Collecting steel pipe segment profile data: Each line laser sensor can collect the profile data of part of the steel pipe section in its local coordinate system; Step 3.2: Splice the steel pipe segment profile data and unify them into the same world coordinate system to obtain complete profile data: Use the rotation and translation matrix obtained in step 2 to perform coordinate transformation on the partial steel pipe cross-section profile data collected by each line laser sensor. The data coordinates in the world coordinate system are obtained by the formula calculate.

[0040] Step 4: Contour data reconstruction to generate final contour data with uniform distribution.

[0041] Step 4.1: Contour center positioning: Use the least squares method to perform ellipse fitting on the complete contour data obtained in step 3 to obtain the center coordinates of the steel pipe cross-section contour ( ); Step 4.2: Coordinate translation: Subtract the center coordinate from the complete contour data obtained in step 3 ( ), the contour data centered at the origin; Step 4.3: Coordinate transformation: convert the contour data obtained in step 3.1 into polar coordinates; Step 4.4: Linear interpolation: Use linear interpolation to make the coordinate points of each segment of data be spaced at intervals. ° is evenly distributed; take the angle of the first point of the contour data The angle with the last point , and keep one decimal place to get and , we can determine that the contour of this segment has measuring points, calculate the angle of p measuring points °, where h is the index number of the measurement point, then in each contour data obtained in step 4.3, find the angle between the geometric position and the measurement point The two closest data points are located at the measurement point angles. The left and right sides of these two data points are used as the reference angle Perform linear interpolation to obtain The polar diameter of the angle.

[0042] Step 4.5: Process the overlapping parts of the contours; put the contour data obtained in step 4.4 together, if there is an angle The same two points, that is, find the average value of the polar radius of these two points as the angle The final polar diameter value of the measuring point; Step 4.6: Coordinate transformation: transform the coordinates obtained in step 4.5 By converting the standardized measurement data points into rectangular coordinates, the final standardized reconstructed steel pipe contour data can be obtained.

[0043] Example 6 The present invention proposes a large-diameter steel pipe contour data acquisition method based on multi-line laser sensors. First, multiple line laser sensors are used to segmentally collect the steel pipe cross-sectional contour data. Then, a calibration block and the FCRM algorithm are used to obtain the rotation and translation matrix for transforming the sensor's local coordinate system into the world coordinate system. The coordinate system is then transformed and integrated with the segmented data through the matrix to obtain the complete cross-sectional contour. Finally, the standard contour data is reconstructed using linear interpolation.

[0044] Example 7 To better illustrate the beneficial effects of the present invention, taking N=4 as an example, a large-diameter steel pipe profile data acquisition mechanism based on four line laser sensors is built according to step 1, as shown in FIG. Figure 2 As shown, it is necessary to ensure that the lasers of the four line laser sensors are coplanar, and the laser plane is perpendicular to the axis of the steel pipe, and the field of view of the four sensors can fully cover the cross section of the steel pipe.

[0045] According to step 2, the four line laser sensors are calibrated using the calibration block. The calibration block used is as follows: Figure 3As shown, its cross-section is a square with a side length of 270 mm, a thickness of 5 mm, and a verticality of ≤0.05 mm on each side. With the center of the calibration block as the origin, the horizontal line passing through the origin is the x-axis, and the y-axis is established through the origin and perpendicular to the x-axis. The angle between the first vertex of the calibration block and the line containing the origin and the horizontal axis is 45°. The coordinates of the four vertices of the calibration block are A(-135, 135), B(-135, -135), C(135, -135), and D(135, 135), respectively. From this, the direction vectors of the four sides can be obtained.

[0046] First, the contour data of the calibration block is collected. Each line laser sensor can collect contour data within its field of view. Then, the FCRM algorithm is used to perform regression clustering on the contour data of each calibration block collected by the line laser to determine the equation of the straight line on which each small contour segment lies, thereby determining the coordinates of the contour corner points. Finally, the translation matrix is ​​solved according to the corresponding relationship. For example, the laser sensor on line 1 obtains the contours of the two edges of the vertex A of the calibration block, then the fitting straight line equations of the two edges of the cross-section contour are 、 , the corner coordinates are (5.2, -171.14); the direction vector of one edge is , , then the rotation angle of the local coordinate system relative to the world coordinate system is , the rotation matrix , then the translation vector Similarly, the rotation matrices of the other three sensor local coordinate systems can be obtained as follows: 、 、

[0047] The translation vectors are: 、 、

[0048] According to step 3, the steel pipe profile data is collected. First, four line laser sensors are used to collect the cross-sectional profile data of the 813mm diameter steel pipe. Then, the rotation and translation matrix obtained in step 2 is used to transform the coordinate system and unify it to the world coordinate system to obtain the complete profile data. The data coordinates are obtained by the formula calculate.

[0049] According to step 4, the steel pipe cross-section profile data is reconstructed. 1) The least squares method is used to perform elliptical fitting on the complete profile data to obtain the center coordinates of the steel pipe cross-section profile (14.30, -327.41); 2) The center coordinates are subtracted from the coordinate points of each segment of the profile data to obtain the profile data with the center as the origin; 3) The profile data is converted to polar coordinates and the interval angle is determined to be 0.1°; 4) Linear interpolation is performed on the polar coordinate data of the four segments of the profile. For example, for the first segment of the profile data, the angle of the first point of the segment of the profile data is taken. The angle with the last point , and keep one decimal place to get and , we can determine that the contour of this segment has measuring points, angles of p measuring points °, where h is the index number of the measuring point, and among the polar coordinate data points of the contour segment, find the angle with the measuring point in the geometric position. The two closest data points are located at the measurement point angles. The left and right sides of these two data points are used as the reference angle Perform linear interpolation to obtain 5) Put the four segments of contour polar coordinate data together, if there is an angle The same two points, that is, find the average value of the polar radius of these two points as the angle The final polar diameter value of the measurement point is obtained, and a total of 3600 data points are obtained; 6) these 3600 polar coordinate data points are converted into rectangular coordinates to obtain the final standardized reconstructed steel pipe profile data.

[0050] The above shows only some implementation schemes of the present invention. For those skilled in the art, several improvements and modifications made to the present invention without departing from the principles of the present invention should also be regarded as within the scope of protection of the present invention.

Claims

1. A method for collecting contour data of large-diameter steel pipes based on a multi-line laser sensor, characterized in that: The specific operations are as follows: Step 1: Construct a steel pipe profile data acquisition mechanism based on multiple line laser sensors; Step 2: Calibrate the rotation and translation matrix of each line laser sensor; Step 3: Collect the segmented data of the steel pipe contour and splice the segmented data to obtain the initial contour data; Step 4: Contour data reconstruction to generate final contour data with uniform distribution.

2. The method for collecting contour data of large-diameter steel pipes based on a multi-line laser sensor according to claim 1, characterized in that: Step 2: The specific steps are as follows: Step 2.1: Collect the calibration block contour data: The calibration block is a regular N-gon with a side length of S. Each line laser sensor collects the calibration block contour data within its field of view; Step 2.2: Determine the contour line equation and contour corner coordinates in the local coordinate system of each laser sensor: Use the FCRM algorithm to perform regression clustering on the calibration block contour data collected by each line laser sensor to determine the equation of the line on which the contour of each side of the calibration block lies, thereby determining the local coordinates of the contour corner points of the calibration block; Step 2.3: Determine the vertex coordinates and straight line equations of the calibration block in the world coordinate system: Take the center of the calibration block as the origin, the horizontal line through the origin as the x-axis, and the y-axis through the origin and perpendicular to the x-axis as the world coordinate system. Since the calibration block is a regular N-gon with a side length of S, the coordinates of the vertices of the calibration block are ,in: , , is the vertex number, counterclockwise; is the angle between the first vertex and the line where the origin is and the horizontal axis. Finally, the linear equation of each side is determined by two adjacent vertices. ,in: , , , the side where the first and last points are located , ; Step 2.4: Solve the rotation and translation matrix of each laser sensor: First, determine the rotation matrix of each laser sensor from the local coordinate system to the world coordinate system based on the angular deviation of the line equations of each side in the local coordinate system and the world coordinate system; then use the rotation matrix to rotate the vertex coordinates of the local coordinate system; finally, determine the translation matrix of each sensor based on the coordinate deviation of each vertex after the rotation transformation and each vertex in the world coordinate system.

3. The method for collecting contour data of large-diameter steel pipes based on a multi-line laser sensor according to claim 2, characterized in that: Step 2.2 The method for determining the contour line equation and contour corner coordinates in the local coordinate system of each sensor is: The FCRM algorithm is used to perform regression clustering on the calibration block contour data collected by each line laser sensor. The specific process is as follows: Step 2.2.1: Set the number of cluster models c=2 and the fuzziness value m =2, iteration termination threshold , set the number of iterations , and initialize the membership matrix , recorded as ,in: For line c A matrix of columns, For data pairs The number of for Matrix Rank The membership value of the column, , , the elements in the membership matrix Constraints must be met: ; Step 2.2.2: According to the formula Update model parameter matrix , ; Among them: matrix is the output value matrix, is the augmented input vector The Vandermonde matrix is ​​formed, For the The diagonal membership matrix of the iteration, ; is the straight line intercept, is the slope; Step 2.2.3: Based on the results from the previous step Update the membership matrix in: Indicates the For data on The error of the model, if , then the model number Add to collection Otherwise, put it into the collection middle; Step 2.2.4: Determine whether to terminate the iteration. If ,but , and return to step 2.2.2; otherwise, the iteration terminates, is the final model parameter matrix; Step 2.2.5: Find the coordinates of the intersection of the two straight lines in the local coordinate system. These coordinates are the coordinates of the contour corner points. The equations of the straight lines of the two adjacent contours of the calibration block are determined by the model parameters obtained in step 2.2.

4. and , then the coordinates of the intersection of these two sides are , which is the coordinate of the contour corner point in the local coordinate system of the sensor.

4. The method for collecting contour data of large-diameter steel pipes based on a multi-line laser sensor according to claim 2, characterized in that: Step 2.4: The method for solving the rotation and translation matrices of each laser sensor is: Step 2.4.1: Match the corresponding edge of the calibration block section in the world coordinate system according to the actual installation position of the laser sensor. i The line laser sensor at the position gets the calibration block i The outline of the two sides where the corner is located; Step 2.4.2: Calculate the rotation angle and rotation matrix: Solve the equation of the line of one side of the calibration block in the world coordinate system and the equation of the line of the line laser sensor in the local coordinate system. The direction vectors of the two corresponding lines are 、 , then the rotation angle of the local coordinate system relative to the world coordinate system is , the rotation matrix ; Step 2.4.3: Calculate the translation vector: by calibrating the coordinates of a vertex of the block in the world coordinate system The coordinates of the corner points in the local coordinate system of the sensor after rotation transformation Solve based on the translation vector .

5. The method for collecting contour data of large-diameter steel pipes based on a multi-line laser sensor according to claim 4, characterized in that: Step 3: Collect the segmented data of the steel pipe contour and splice them to obtain the initial contour data. The process is as follows: Step 3.1: Collecting steel pipe segment profile data: Each line laser sensor can collect the profile data of part of the steel pipe section in its local coordinate system; Step 3.2: Splice the steel pipe segmented contour data and unify them into the same world coordinate system to obtain the complete contour data.

6. The method for collecting contour data of large-diameter steel pipes based on a multi-line laser sensor according to claim 5, characterized in that: In step 3.2, the rotation matrix and translation vector obtained in step 2 are used to transform the partial steel pipe cross-section profile data collected by each line laser sensor. The data coordinates in the world coordinate system are calculated by the formula calculate.

7. The method for collecting contour data of large-diameter steel pipes based on a multi-line laser sensor according to claim 6, characterized in that: The specific process of reconstructing the contour data in step 4 is: Step 4.1: Contour center positioning: Use the least squares method to perform ellipse fitting on the complete contour data obtained in step 3 to obtain the center coordinates of the steel pipe cross-section contour ( ); Step 4.2: Coordinate translation: Subtract the center coordinate from the complete contour data obtained in step 3 ( ), the contour data centered at the origin; Step 4.3: Coordinate transformation: convert the contour data obtained in step 3.1 into polar coordinate representation; Step 4.4: Linear interpolation: Use linear interpolation to make the coordinate points of each segment of data be spaced at intervals. The contour data is evenly distributed in the form of °; Step 4.5: Process the overlapping parts of the contours; put the contour data obtained in step 4.4 together, if there is an angle The same two points, that is, find the average value of the polar radius of these two points as the angle The final polar diameter value of the measuring point; Step 4.6: Coordinate transformation: transform the coordinates obtained in step 4.5 By converting the standardized measurement data points into rectangular coordinates, the final standardized reconstructed steel pipe contour data can be obtained.

8. The method for collecting contour data of large-diameter steel pipes based on a multi-line laser sensor according to claim 7, characterized in that: Step 4.4 is as follows: Get the angle of the first point of the contour data segment The angle with the last point , and keep one decimal place to get and , that is, to determine the contour of this segment measuring points, calculate the angle of p measuring points °, where h is the index number of the measurement point, and then in each contour data obtained in step 4.3, find the angle between the geometric position and the measurement point The two closest data points are located at the measurement point angles. The left and right sides of these two data points are used as the reference angle Perform linear interpolation to obtain The polar diameter of the angle.

Images