Segment beam prefabrication linear measurement method and system based on hybrid sign positioning

Abstract

The application discloses a segmental beam prefabrication linear measurement method and system based on mixed mark positioning, and the method comprises the following steps: arranging circular diagonal coding marks on the control points of a fixed end mold and a segmental beam top; the circular diagonal coding marks adopt two features of a white positioning circle and a diagonal mark; collecting multiple images including all the circular diagonal coding marks, and adopting an image processing method to identify and acquire the center coordinates of each coding mark in each image; matching the marks with the same coding value in different images, and based on the known three-dimensional world coordinates of the fixed end mold control points, initially solving the camera external parameters of each image and the three-dimensional world coordinates of the segmental beam top surface control points; and then globally jointly optimizing the camera external parameters of all the images and the three-dimensional world coordinates of each control point of the segmental beam top, so as to obtain the three-dimensional world coordinates of each control point of the segmental beam. The application can solve the problems of large manual measurement error and low efficiency, and realizes the automation of the linear measurement of the prefabricated beam field.

Description

Technical Field

[0001] This invention relates to the field of segmental beam measurement and positioning, and specifically to a method and system for measuring the precast alignment of segmental beams based on hybrid marker positioning. Background Technology

[0002] Currently, in the mass production and geometric alignment control of precast segmental beams, most domestic and international engineering projects rely on the "six-point control method," employing manual operation of a total station with a prism to obtain the three-dimensional spatial position of the segmental beams. Due to the large number of beam segments and control points on each segment in these projects, and the need to measure multiple sets of data for each segment, including the cast-in-place position, the matching position, and the final actual matching position, the measurement workload is enormous. Manual measurement is inefficient, resulting in high construction costs and long construction periods, failing to fully realize the advantages of prefabricated bridges, such as high production quality, rapid construction, and lower cost.

[0003] In recent years, with the rapid development of technologies such as computer vision and image recognition, visual measurement has become one of the important means of precise measurement and positioning due to its advantages of non-contact, high precision, and high efficiency. However, for the measurement needs of segmental beam prefabrication sites, the detection and positioning features of most targets are relatively simple. When there are complex conditions such as poor lighting, blurriness, or occlusion on site, their positioning accuracy is relatively low, and even image detection failure may occur. For example, when the camera shooting angle is large, the circular coded marker (CCT) is easily affected by the eccentricity difference caused by projection deformation, and the positioning accuracy decreases accordingly. The coded markers related to corner recognition are more sensitive to noise and blurriness, especially when the image edges are not clear, which can easily lead to false detections or missed detections. Summary of the Invention

[0004] This invention provides a method and system for measuring the precast alignment of segmental beams based on hybrid marker positioning, in order to solve the technical problems of numerous human interference factors, large errors, and low efficiency in the current short-line method for measuring and controlling the precast alignment of segmental beams.

[0005] To achieve the above technical objectives, the present invention adopts the following technical solution:

[0006] A method for measuring the precast alignment of segmental beams based on hybrid marker positioning, comprising:

[0007] Circular diagonal coding marks are placed at the control points on the fixed end formwork and the top surface of the segmental beam;

[0008] The circular diagonal coding mark includes a white positioning circle and an outer ring coding band with the same center. The area between the positioning circle and the outer ring coding band, and outside the outer ring coding band, is black. A black diagonal mark is set at the center of the positioning circle. The outer ring coding band is divided into n equal parts, and black and white are used to encode the given value corresponding to the outer ring coding band in binary.

[0009] Multiple images of all circular diagonal coded marks on the top surface of the fixed end formwork and segmental beam were collected, and image processing methods were used to detect and identify the center coordinates of each circular diagonal coded mark in each image.

[0010] Based on the encoding value of the outer ring encoding band, the circular diagonal encoding marks in different images are matched; and based on the known three-dimensional world coordinates of the fixed end module control points, the camera extrinsic parameters and the three-dimensional world coordinates of the control points on the top surface of the segment beam are initially calculated for each image; then, the camera extrinsic parameters and the three-dimensional world coordinates of each control point on the top surface of the segment beam are globally jointly optimized to finally obtain the three-dimensional world coordinates of each control point on the top surface of the segment beam.

[0011] Furthermore, the diagonal sign is composed of two black isosceles right triangles.

[0012] Furthermore, the step of using image processing methods to detect and identify the center coordinates of each circular diagonal coded marker in each image includes:

[0013] The center of the fitted ellipse is extracted by fitting an ellipse to the projection of the white positioning circle in the image.

[0014] By fitting the projections of the two perpendicular lines in the diagonal mark onto the image, and extracting the intersection point of the two fitted lines as the diagonal detection center;

[0015] The coordinates of the center of the fitted ellipse and the diagonal detection center are integrated based on the mean square error of the ellipse fitting and the straight line fitting, and used as the center coordinates of the circular diagonal coding mark.

[0016] Furthermore, ellipse fitting and extraction of the fitted ellipse center includes: first, detecting ellipses in the image and calculating the perimeter L and area S of each ellipse; then, calculating the roundness factor. Then, ellipses whose roundness factors meet the threshold conditions are fitted to obtain ellipses. Finally, the coordinates of the ellipse center are extracted. for:

[0017] ;

[0018] in, These are the parameters of the ellipse.

[0019] Furthermore, the extraction of diagonal detection centers through line fitting includes:

[0020] Extract the contours of connected components in the image and filter out the triangle contours;

[0021] Calculate the centroid coordinates of each triangle and the straight-line distance between the two centroids. Filter out a pair of triangles whose distance meets the threshold condition, which are the two triangles corresponding to the diagonal marker.

[0022] Perform a Hough transform within the area containing the diagonal markers to extract all line segments, and then select two mutually perpendicular lines from them. , The condition for determining whether lines L1 and L2 are perpendicular is: , They are straight lines and straight line polar angle They are straight lines and straight line The polar distance;

[0023] The least squares method is used to fit lines L1 and L2, and the pixel coordinates of the intersection point of the two lines are obtained. This refers to the diagonal detection center.

[0024] Furthermore, based on the mean square error of the ellipse fitting and the line fitting, the coordinate results obtained from the center of the fitted ellipse and the diagonal detection center are weighted and allocated, and the center coordinates of the circular diagonal coding mark are integrated and calculated, including:

[0025] Calculate the mean square error of ellipse fitting and line fitting respectively. , :

[0026] ;

[0027] ;

[0028] Where N represents the number of sample points used for ellipse fitting; N1 and N2 represent the number of sample points used for fitting lines L1 and L2, respectively. For ellipse parameters, It is a straight line The parameters, It is a straight line Parameters; For ellipse fitting or line fitting, the first Coordinates of each sample point;

[0029] Calculate the distance between the center of the fitted ellipse and the diagonal detection center. ;

[0030] When distance Less than or equal to the threshold When using mean square error , For the center coordinates of the fitted ellipse and the coordinates of the diagonal detection center By performing a weighted calculation, the center coordinates of the circular diagonal coded mark are obtained:

[0031] ;

[0032] like ,when At that time, the ellipse detection results were considered reliable, and the following method was adopted. As the center coordinates of the circular diagonal coding mark; when At that time, the diagonal detection results were considered reliable, and the following method was adopted. The center coordinates of the circular diagonal coding mark.

[0033] Furthermore, the decoding method for the outer ring coding band is as follows: starting from any starting point of the outer ring coding band, decoding is performed in a counterclockwise order. Each flag can obtain n n-bit binary sequences. Each binary sequence is converted into a decimal value, and the minimum value among the decimal values ​​is selected as the encoding value of the outer ring coding band.

[0034] Furthermore, the three-dimensional world coordinates of the control points on the top surface of the initial solution segment beam are as follows:

[0035] The center coordinates of each circular diagonal coded mark detected and identified by the image processing method are converted into the three-dimensional camera coordinates of the corresponding control points;

[0036] Using the known 3D world coordinates and 3D camera coordinates of the fixed end-mode control points, and using the known camera intrinsic parameters, the initial extrinsic parameters of the camera are solved.

[0037] Furthermore, for the control points of the segmental beam, their three-dimensional world coordinates are calculated using the three-dimensional camera coordinates and the camera's intrinsic and initial extrinsic parameters.

[0038] Furthermore, a global joint optimization is performed on the camera extrinsic parameters and the 3D world coordinates of each control point on the top surface of the segmental beam for all images. Specifically, this includes: using the camera extrinsic parameters and the 3D world coordinates of each control point on the segmental beam obtained from the initial solution of each image as initial values, and with the minimization of reprojection error as the optimization objective, a nonlinear global joint optimization is performed on the camera extrinsic parameters and the 3D world coordinates of each control point on the top surface of the segmental beam using bundle adjustment, to obtain the 3D world coordinates of each control point on the top surface of the segmental beam; where the optimization objective function is:

[0039] ;

[0040] In the formula, This represents the measured two-dimensional pixel coordinates of control point j in the k-th image, which is the center coordinates of the corresponding circular diagonal coded mark detected and identified by the image processing method. The three-dimensional world coordinates of control point j are reprojected onto the k-th image to obtain the two-dimensional pixel coordinates; K is the number of images including the segmental beam, and J is the number of control points on the segmental beam; This is the reprojection error.

[0041] A segmental beam precast alignment measurement system based on hybrid marker positioning, comprising:

[0042] (1) Several circular diagonal coding marks: respectively arranged at the control points on the fixed end form and the top surface of the segment beam;

[0043] The circular diagonal coding mark includes a white positioning circle and an outer ring coding band with the same center. The area between the positioning circle and the outer ring coding band, and outside the outer ring coding band, is black. A black diagonal mark is set at the center of the positioning circle. The outer ring coding band is divided into n equal parts, and black and white are used to encode the given value corresponding to the outer ring coding band in binary.

[0044] (2) Image acquisition module: including a slide rail system and a camera base and camera fixedly installed on the slide rail system; the slide rail system drives the camera to move to the image acquisition position preset by each of the different platforms; the camera is used to acquire multiple images including all the circular diagonal coded marks on the top surface of the fixed end mold and the segment beam;

[0045] (3) Central control module: used to send instructions to the image acquisition module and the camera motion control module; used to detect and identify the center coordinates of each circular diagonal coded mark in each image using image processing methods; used to match circular diagonal coded marks with the same coded value in different images according to the coded value of the outer ring coded band; and used to perform initial calculations on the camera extrinsic parameters of each image and the three-dimensional world coordinates of the control points on the top surface of the segment beam based on the known three-dimensional world coordinates of the fixed end module control points; and then to perform global joint optimization on the camera extrinsic parameters of all images and the three-dimensional world coordinates of each control point on the top surface of the segment beam, and finally obtain the three-dimensional world coordinates of each control point on the top surface of the segment beam.

[0046] Compared with existing technologies, the technical advantages of the segmental beam precast alignment measurement method and system based on hybrid marker positioning in this invention are as follows: Combining the dual features of circular and diagonal markers, when image noise is high, circular marker positioning can compensate for the insufficient accuracy of diagonal marker positioning; when the shooting angle is large, diagonal marker positioning can compensate for the eccentricity problem caused by the deformation of circular markers, effectively reducing the situation where the failure of a single feature leads to the failure of the overall marker detection. At the same time, it helps to overcome the limitations of traditional manual operation of total stations in precast beam yards, such as low efficiency, high human error, and high construction costs. Furthermore, the intelligent measurement method facilitates integration with the on-site segmental beam attitude adjustment formwork trolley and alignment calculation system, achieving efficient and high-precision construction in precast beam yards with reduced manpower, automation, and intelligence. Attached Figure Description

[0047] Figure 1 This is a schematic diagram of the method steps in an embodiment of the present invention.

[0048] Figure 2 This is a schematic diagram of a circular diagonal coding mark according to an embodiment of the present invention, wherein the reference numerals are: 1-outer ring coding band, 2-white positioning circle, 3-black diagonal mark.

[0049] Figure 3 This is a schematic diagram showing the arrangement of the circular diagonal coding marks on the fixed end mold and segmental beam according to an embodiment of the present invention.

[0050] Figure 4 This is a schematic diagram of the spatial layout of each platform, industrial camera, and slide rail system in the precast beam yard according to an embodiment of the present invention, wherein the reference numerals are: 10-slide rail system, 20-camera base, 30-camera. Detailed Implementation

[0051] The embodiments of the present invention will be described in detail below. These embodiments are based on the technical solutions of the present invention and provide detailed implementation methods and specific operation processes to further explain the technical solutions of the present invention.

[0052] Example 1

[0053] This embodiment provides a method for measuring the precast alignment of segmental beams based on hybrid marker positioning, referencing... Figure 1 As shown, it includes:

[0054] Step 1: Arrange circular diagonal coding marks at the control points on the fixed end formwork and the top surface of the segmental beam.

[0055] In this invention, such as Figure 2As shown, the circular diagonal coding mark mainly consists of two parts: first, the central positioning part, which adds a black diagonal mark at the center of the white positioning circle, composed of two black isosceles right triangles; second, the outer ring coding band (a concentric ring of the white positioning circle), used to store coding information. The area between the positioning circle and the outer ring coding band, and the area outside the outer ring coding band, is black.

[0056] The encoding and decoding rules for the outer ring encoding band are as follows: Divide the ring encoding band into n equal parts, each part into one binary bit, and assign different binary numbers according to the two different colors of black and white on the binary bit (white is 1, black is 0); Decode in a counterclockwise order starting from any starting point, and each flag can obtain n n-bit binary sequences. Convert the binary sequences into decimal values, and select the minimum value among the decimal values ​​as the encoding value of the flag.

[0057] Furthermore, the actual size of the circular diagonal coding mark can be determined according to specific measurement requirements. In this embodiment, the circular diagonal coding mark is a square with a side length of 6cm, the diameter of the white positioning circle is 3cm, the inner and outer diameters of the annular coding strip are 4cm and 5cm respectively, and the diagonal mark is composed of two isosceles right triangles with a side length of 0.75cm.

[0058] In this embodiment, the control point distribution is as follows: A three-dimensional local coordinate system is established in the space of the fixed end model: the origin I0 is the intersection of the fixed end model and the beam body interface and the centerline of the beam top surface along the bridge direction. The positive x-axis points along the centerline of the beam top surface towards the matching end, the positive z-axis points upward, and the y-axis direction is determined according to the right-hand rule. Then, with the origin I0(0,0,0) of the local coordinate system as the central control point, control points are generated from point I0 towards... Points are taken at equal intervals (distance a) on both sides of the positive and negative axes as control points L0(0,a,0) and R0(0,-a,0) respectively. The distance a is determined based on the actual site conditions, generally approximately the distance from the web line to the beam centerline. Throughout the construction process, since the fixed end formwork is a rigid body, its position remains fixed and it does not deform. Therefore, the three control points with known three-dimensional coordinates, I0, L0, and R0, are used as the reference points for measurement. Then, based on the "six-point control method" for controlling the geometric alignment of segmental beams using the short-line method, horizontal control points FH and BH are set on the centerline of the beam top surface along the bridge direction in each segment, and elevation control points FL, BL, FR, and BR are set on the web lines on the left and right sides of the beam top surface, resulting in a total of six control points per segment.

[0059] The present invention is a method for measuring the precast alignment of segmental beams based on hybrid marker positioning. This method uses the known three-dimensional world coordinates of the fixed end mold, employs visual detection technology to extract the center coordinates of the circular diagonally coded markers from the image, and uses multi-image optimization to obtain the three-dimensional world coordinates of each control point on the top surface of the segmental beam.

[0060] Step 2: Collect multiple images including all circular diagonal coded marks on the top surface of the fixed end formwork and segmental beam, and use image processing methods to detect and identify the center coordinates of each circular diagonal coded mark in each image.

[0061] It is important to note that the pre-set image acquisition position and camera acquisition angle must ensure that the acquired image includes all the circular diagonal coded marks on the top surface of both the fixed end mold and the segmental beam.

[0062] Before identifying the center coordinates of the circular diagonal coded marker, this embodiment first stores the acquired raw image data and performs image preprocessing. The image preprocessing steps mainly include grayscale processing, illumination correction, Gaussian filtering for noise reduction, Otsu binarization, and Canny edge detection to optimize image quality and improve subsequent detection results. If image data loss or corruption is detected during this process, the image is reacquired.

[0063] The method of detecting and identifying the center coordinates of each circular diagonal coded marker in each image using image processing includes:

[0064] (1) The center of the fitted ellipse is extracted by fitting the projection of the white positioning circle in the image to an ellipse.

[0065] Since a circle deforms into an ellipse during projection, this embodiment first uses the Canny operator to detect and extract image edges, and then uses the cv2.findContours function from the OpenCV library to detect and extract the closed contour of the ellipse. To quantify the degree of deformation of the circle, a shape factor is introduced based on shape parameters (circumference L, area S). The value of C ranges from (0,1), and the theoretical shape factor of a standard circle is 1. Generally, the larger the shooting angle, the greater the deformation of the circle, and the smaller the value of C. The threshold of C should be set reasonably according to the actual situation to initially filter out contours that do not meet the threshold conditions. When the shooting angle is small, the threshold of the roundness factor can be appropriately increased to improve the screening efficiency; when the shooting angle is large, the threshold should be appropriately decreased.

[0066] The least squares method was used to fit the elliptical contours that met the threshold conditions after preliminary screening, based on the general equation of an ellipse. Construct the least squares fitting optimization objective function The ellipse parameters, including the pixel coordinates of the ellipse center O, are thus calculated and obtained. And the major and minor axis pixel radii a and b, as shown below:

[0067] ;

[0068] in, These are the parameters of the ellipse.

[0069] (2) By fitting the projections of the two perpendicular lines in the diagonal mark onto the image, the intersection of the two fitted lines is extracted as the diagonal detection center.

[0070] A diagonal sign consists of two black isosceles right triangles with sides of length b. The area of ​​a single isosceles right triangle is 0.5b. 2 The straight-line distance between the centroids of the two small triangles is The `findContours` function from the OpenCV library is used to detect and extract connected component contours in a binary image. The extracted contours are initially screened based on the geometric features of the diagonal markers, meeting the following two conditions: (a1) Each connected component contour is approximated using the `approxPolyDP` function, retaining contours that approximate as triangles (i.e., containing three corner points); (a2) Since the two isosceles right triangles of the diagonal markers form a diagonal distribution, the pixel coordinates of the centroids C1 and C2 of the two triangles are calculated respectively. , , ( (This is a threshold value, set according to specific circumstances).

[0071] Based on the above conditions, the connected region contour is clipped as the region of interest (ROI) centered on the midpoint of the line connecting points C1 and C2 to reduce background interference and improve the efficiency and accuracy of Hough transform detection of lines. Within the ROI, Hough transform detection is performed to extract all line segments, and two mutually perpendicular lines are selected from them. , (The condition for determining whether they are perpendicular is:) The least squares method is used to fit lines L1 and L2 to obtain the pixel coordinates of the intersection point of the two lines. This refers to the diagonal detection center.

[0072] If the diagonal markers in the image are severely affected by noise (stains, shadows, etc.), severely occluded, or have poor local image quality, these situations may lead to contour extraction errors or inaccurate triangle fitting. In such cases, the two preliminary screening conditions mentioned above cannot be met, and the system automatically determines that the diagonal marker feature detection has failed or is unreliable, and considers it as not detecting the diagonal detection center. However, the system can still rely on the center coordinates of the white positioning circle obtained by ellipse fitting as the image coordinates of the control point. This ensures that usable measurement data can still be provided under complex working conditions, greatly improving the robustness of control point detection.

[0073] This method obtains the center by detecting and fitting the inherent two vertical straight line features of the diagonal mark, utilizing more image edge information and fusing information from multiple pixels. It has stronger robustness to local contour deformation and partial occlusion, and can achieve higher accuracy and more stable center positioning. On this basis, combined with the ellipse fitting center, a complementary dual positioning mechanism is formed. (3) The coordinates of the fitted ellipse center and the diagonal detection center are integrated according to the mean square error of the ellipse fitting and the straight line fitting, and used as the center coordinates of the circular diagonal coded mark. Including:

[0074] First, calculate the mean square error of the ellipse fitting and the line fitting respectively. , The overall fitting accuracy between the actual detected edge points and the theoretical equation is quantified, and this result is used as the confidence level for the two feature detection localization methods. Generally, the smaller the mean square error, the more accurate the fit; ideally, the mean square error is 0. The formula for calculating the mean square error is as follows:

[0075] ;

[0076] ;

[0077] Where N represents the number of sample points used for ellipse fitting; N1 and N2 represent the number of sample points used for fitting lines L1 and L2, respectively. For ellipse parameters, It is a straight line The parameters, It is a straight line Parameters; For ellipse fitting or line fitting, the first Coordinates of each sample point.

[0078] Secondly, calculate the distance between the center of the fitted ellipse and the diagonal detection center. .

[0079] Based on the design of the circular diagonal coding mark, the center of the white positioning circle should theoretically coincide with the center of the diagonal mark. This embodiment utilizes an ellipse to fit the center O. With diagonal detection center C The distance between them is used to perform a geometric verification of the two center coordinates.

[0080] Then based on distance To determine the center coordinates of the circular diagonal coding mark:

[0081] When distance Less than or equal to the threshold Then, based on the positioning results of the positioning circle and the diagonal markers (i.e., the ellipse fitting center O), the positioning is determined. With diagonal detection center C All are reliable, using mean square error. , We calculate the center coordinates of the circular diagonal coded marker by weighting the coordinates of the fitted ellipse center and the diagonal detection center:

[0082] ;

[0083] like ,when At that time, the ellipse detection results were considered reliable, and the following method was adopted. As the center coordinates of the circular diagonal coding mark; when At that time, the diagonal detection results were considered reliable, and the following method was adopted. The center coordinates of the circular diagonal coding mark.

[0084] Step 3: Match the circular diagonal coding marks in different images according to the coding value of the outer ring coding band; and perform initial calculations on the camera extrinsic parameters and the three-dimensional world coordinates of the control points on the top surface of the segment beam for each image based on the known three-dimensional world coordinates of the fixed end module control points; then perform global joint optimization on the camera extrinsic parameters and the three-dimensional world coordinates of each control point on the top surface of the segment beam for all images, and finally obtain the three-dimensional world coordinates of each control point on the top surface of the segment beam.

[0085] The decoding method for the outer ring coding band is as follows: starting from any starting point of the outer ring coding band, decoding is performed in a counterclockwise order. Each flag can obtain n n-bit binary sequences. Each binary sequence is converted into a decimal value, and the minimum value among the decimal values ​​is selected as the coding value of the outer ring coding band.

[0086] In step 3, the three-dimensional world coordinates of the control points on the top surface of the initial solution segment beam are as follows:

[0087] (1) The center coordinates of each circular diagonal coded mark detected and identified by the image processing method are converted into the three-dimensional camera coordinates of the corresponding control points.

[0088] Assume the center coordinates of the circular diagonal coding mark are The 3D camera coordinates of the corresponding control points are then expressed as: .

[0089] (2) Using the known three-dimensional world coordinates and three-dimensional camera coordinates of the fixed end control point, and using the known camera intrinsic parameters, solve for the initial extrinsic parameters of the camera.

[0090] For a single image k, the three-dimensional world coordinates of three coded marker points I0, L0, and R0 on the fixed end model in the image are known. Then, combined with the transformed three-dimensional camera coordinates and using the coordinate system mapping relationship, the PnP (Perspective-n-Point) algorithm is used to iteratively solve the initial extrinsic parameters corresponding to image k.

[0091] Among them, the three-dimensional camera coordinates of each control point and three-dimensional world coordinates The following mapping relationship exists:

[0092] ;

[0093] In the formula, s is the scale factor; K is the camera intrinsic parameter matrix, which can be obtained by camera calibration before the system is put into operation; R is the camera extrinsic parameter matrix, representing the transformation relationship between 3D spatial coordinates and 3D camera coordinates, where R is a 3×3 rotation matrix and T is a 3×1 translation vector.

[0094] (3) For the control points of the segmental beam, the three-dimensional world coordinates are calculated using the three-dimensional camera coordinates and the camera intrinsic parameters and initial extrinsic parameters, according to the mapping relationship between the two coordinate systems mentioned above.

[0095] Based on the camera extrinsic parameters of image k Using the principle of triangulation, the three-dimensional world coordinates of control point j on the top surface of the segment beam corresponding to image k were initially calculated. .

[0096] In step 3, the camera extrinsic parameters and the 3D world coordinates of each control point on the top surface of the segmental beam of all images are globally jointly optimized. Specifically, this includes: using the camera extrinsic parameters and the 3D world coordinates of each control point of the segmental beam obtained from the initial solution of each image as initial values, and taking the minimization of reprojection error as the optimization objective, the bundle adjustment (BA) method is used to optimize the camera extrinsic parameters of each image. The three-dimensional world coordinates of each control point of the segmental beam Nonlinear global joint optimization is performed to obtain the three-dimensional world coordinates of each control point on the top surface of the segmental beam. ;

[0097] The objective function is:

[0098] ;

[0099] In the formula, This represents the measured two-dimensional pixel coordinates of control point j in the k-th image, which is the center coordinates of the corresponding circular diagonal coded mark detected and identified by the image processing method. The three-dimensional world coordinates of control point j are reprojected onto the k-th image to obtain the two-dimensional pixel coordinates; K is the number of images including the segmental beam, and J is the number of control points on the segmental beam; This is the reprojection error.

[0100] Example 2

[0101] This embodiment provides a segmental beam precast alignment measurement system based on hybrid marker positioning, such as... Figure 3 , Figure 4 As shown, it includes:

[0102] (1) Several circular diagonal coding marks: respectively arranged at the control points on the fixed end form and the top surface of the segment beam;

[0103] The circular diagonal coding mark includes a white positioning circle and an outer ring coding band with the same center. The area between the positioning circle and the outer ring coding band, and outside the outer ring coding band, is black. A black diagonal mark is set at the center of the positioning circle. The outer ring coding band is divided into n equal parts, and black and white are used to encode the given value corresponding to the outer ring coding band in binary.

[0104] (2) Image acquisition module: including a slide rail system 10 above the platform and a camera base 20 and a camera 30 fixedly installed on the slide rail system; the slide rail system drives the camera to move to the image acquisition position preset by each platform; the camera acquires multiple images including the fixed end mold and all the circular diagonal coded marks on the top surface of the segment beam.

[0105] It is important to note that the pre-set image acquisition position and camera acquisition angle must ensure that the acquired image includes all the circular diagonal coded marks on the top surface of the fixed end mold and segmental beam. With the support of the sliding rail system, a single industrial camera can handle the image acquisition tasks of multiple platforms in the precast beam yard.

[0106] (3) Central control module: used to send instructions to the image acquisition module and the camera motion control module, used to detect and identify the center coordinates of each circular diagonal coded mark in each image using image processing methods; and used to: match the circular diagonal coded marks in different images according to the coding value of the outer ring coding band; and to perform initial calculations on the camera extrinsic parameters and the three-dimensional world coordinates of the control points on the top surface of the segment beam for each image based on the known three-dimensional world coordinates of the fixed end module control points; and then to perform global joint optimization on the camera extrinsic parameters and the three-dimensional world coordinates of each control point on the top surface of the segment beam for all images, and finally obtain the three-dimensional world coordinates of each control point on the top surface of the segment beam.

[0107] The specific implementation method of the central control module in this embodiment is the same as that described in Embodiment 1.

[0108] The above embodiments are preferred embodiments of this application. Those skilled in the art can make various changes or improvements based on them. Without departing from the overall concept of this application, these changes or improvements should fall within the scope of protection claimed in this application.

Claims

1. A segmental girder precast linear measurement method based on hybrid landmark positioning, characterized in that, The method comprises the following steps: A circular diagonal coding mark is arranged at each control point on the top surface of the fixed end mold and the segmental beam; The circular diagonal coding mark comprises a white positioning circle and an outer ring-shaped coding band with the same center, and the space between the white positioning circle and the outer ring-shaped coding band and the space outside the outer ring-shaped coding band are black; a black diagonal mark is arranged at the center of the white positioning circle; the outer ring-shaped coding band is equally divided into n parts, and the given value corresponding to the outer ring-shaped coding band is coded by using black and white colors in binary form; A plurality of images including all circular diagonal coding marks on the top surface of the fixed end mold and the segmental beam are collected, and the center coordinates of each circular diagonal coding mark in each image are detected and recognized by using an image processing method; According to the coding value of the outer ring-shaped coding band, the circular diagonal coding marks with the same coding value in different images are matched, and the three-dimensional world coordinates of the control points on the top surface of the segmental beam are initially solved based on the known three-dimensional world coordinates of the control points of the fixed end mold; then the camera external parameters of all images and the three-dimensional world coordinates of the control points on the top surface of the segmental beam are globally optimized, and finally the three-dimensional world coordinates of the control points on the top surface of the segmental beam are obtained. The diagonal mark is composed of two black isosceles right triangles; the center coordinates of each circular diagonal coding mark in each image are detected and recognized by using an image processing method, which comprises the following steps: The center of the fitted ellipse is extracted by fitting the projection of the white positioning circle in the image as an ellipse; The intersection point of the two fitted straight lines is taken as the diagonal detection center by fitting the projections of the two mutually perpendicular straight lines in the diagonal mark in the image respectively; The coordinate results obtained by the fitting ellipse center and the diagonal detection center are weighted and allocated according to the mean square errors of the ellipse fitting and the straight line fitting, and the center coordinates of the circular diagonal coding mark are integrated and calculated.

2. The segmental beam precast linear alignment method of claim 1, wherein, The ellipse fitting and extraction of the fitted ellipse center comprises: firstly detecting the ellipse in the image and counting the circumference L and area S of each ellipse, then calculating the roundness factor , and then screening the ellipse whose roundness factor meets the threshold condition for fitting to obtain the ellipse , and finally extracting the ellipse center coordinates : ； wherein are the ellipse parameters.

3. The segmental beam precast linear alignment method of claim 1, wherein, The straight line fitting to extract the diagonal detection center comprises the following steps: The triangular contours in the connected domain contours in the image are screened out; The barycentric coordinates of each triangle and the straight line distance between the two barycenters are calculated, and a pair of triangles whose distance satisfies the threshold condition are screened out, that is, the two triangles corresponding to the diagonal mark; Hough transform detection is performed in the region where the diagonal sign is located to extract all straight line segments, and two mutually perpendicular straight lines are selected from the straight line segments , ; wherein the determination condition that the straight lines L1 and L2 are perpendicular to each other is: , are the polar angles of the straight lines and the straight line , respectively, are the polar distances of the straight lines and the straight line , respectively. The least square method is used to perform straight line fitting on the straight lines L1 and L2 to obtain the pixel coordinates of the intersection point of the two straight lines That is, the diagonal detection center.

4. The segmental beam precast linear alignment method of claim 1, wherein, The coordinate results obtained by the fitting ellipse center and the diagonal detection center are weighted and allocated according to the mean square errors of the ellipse fitting and the straight line fitting, and the center coordinates of the circular diagonal coding mark are integrated and calculated, which comprises the following steps: Calculate the mean square error of the ellipse fit and the straight line fit, respectively , : ； ； wherein N represents the number of sample points for the ellipse fitting; N1, N2 represent the number of sample points for the straight line L1, L2 fitting respectively; are parameters of the ellipse, are parameters of the straight line are parameters of the straight line are parameters of the straight line are parameters of the straight line is the coordinate of the Nth sample point for the ellipse fitting or the straight line fitting; is the coordinate of the Nth sample point for the ellipse fitting or the straight line fitting; calculating a distance between the center of the fitted ellipse and the diagonal detection center ; When distance Less than or equal to the threshold When using mean square error , For the center coordinates of the fitted ellipse and the coordinates of the diagonal detection center By performing a weighted calculation, the center coordinates of the circular diagonal coded mark are obtained: ； like ,when At that time, the ellipse detection results were considered reliable, and the following method was adopted. As the center coordinates of the circular diagonal coding mark; when At that time, the diagonal detection results were considered reliable, and the following method was adopted. The center coordinates of the circular diagonal coding mark.

5. The segmental beam precast linear alignment method of claim 1, wherein, The decoding method of the outer ring-shaped coding band is as follows: decoding is started from any starting point of the outer ring-shaped coding band in counterclockwise order, each mark can obtain n n-bit binary sequences, each binary sequence is converted into a decimal value, and the minimum value in the decimal values is selected as the coding value of the outer ring-shaped coding band.

6. The segmental beam precast linear measurement method of claim 1, wherein, The initial solving of the three-dimensional world coordinates of the control points on the top surface of the segmental beam comprises the following steps: The center coordinates of each circular diagonal coding mark detected and recognized by the image processing method are converted into the three-dimensional camera coordinates of the corresponding control points; The initial external parameters of the camera are solved by using the known three-dimensional world coordinates and three-dimensional camera coordinates of the control points of the fixed end mold and the known camera internal parameters; Then, the three-dimensional world coordinates of the control points of the segmental beam are solved by using the three-dimensional camera coordinates, the camera internal parameters and the initial external parameters.

7. The segmental beam precast linear measurement method of claim 1, wherein, The camera extrinsic parameters of all images and the three-dimensional world coordinates of the control points on the top surface of the segmental beam are globally and jointly optimized, specifically including: taking the camera extrinsic parameters and the three-dimensional world coordinates of the control points on the segmental beam obtained from the initial calculation of each image as initial values, minimizing the re-projection error as the optimization objective, using the bundle adjustment method to perform nonlinear global and joint optimization on the camera extrinsic parameters and the three-dimensional world coordinates of the control points on the segmental beam of each image, and obtaining the three-dimensional world coordinates of the control points on the top surface of the segmental beam; wherein the optimization objective function is: ； In the formula, represents the measured two-dimensional pixel point coordinates of control point j in the kth image, that is, the center coordinates of the corresponding circular corner coding mark recognized by the image processing method; represents the two-dimensional pixel point coordinates obtained by reprojecting the three-dimensional world coordinates of control point j to the kth image; K is the number of images including the segmental beam, and J is the number of control points on the segmental beam; is the re-projection error.

8. A segmental girder precast linear measurement system based on hybrid landmark positioning, characterized by, including: (1) A plurality of circular diagonal coding marks: arranged on the control points on the top surface of the segmental beam and the fixed end mold respectively; Wherein the circular diagonal coding mark includes a white positioning circle and an outer ring-shaped coding band with the same center, and the space between the positioning circle and the outer ring-shaped coding band and outside the outer ring-shaped coding band is black; a black diagonal mark is arranged at the center of the positioning circle; the outer ring-shaped coding band is equally divided into n parts, and the given value corresponding to the outer ring-shaped coding band is binary coded using different colors of black and white; (2) An image acquisition module: including a slide rail system, a camera base and a camera fixedly installed on the slide rail system; the slide rail system drives the camera to move to different pedestal image acquisition positions preset respectively; the camera is used for acquiring a plurality of images including all circular diagonal coding marks on the fixed end mold and the top surface of the segmental beam; (3) A central control module: used for sending instructions to the image acquisition module and the camera motion control module, used for detecting and identifying the center coordinates of each circular diagonal coding mark in each image by using an image processing method; and used for matching the circular diagonal coding marks with the same coding value in different images based on the coding value of the outer ring-shaped coding band; and based on the known three-dimensional world coordinates of the control points of the fixed end mold, the camera extrinsic parameters of each image and the three-dimensional world coordinates of the control points on the top surface of the segmental beam are initially calculated; then the camera extrinsic parameters of all images and the three-dimensional world coordinates of the control points on the top surface of the segmental beam are globally and jointly optimized, and finally the three-dimensional world coordinates of the control points on the top surface of the segmental beam are obtained; The diagonal mark is composed of two black isosceles right triangles; the center coordinates of each circular diagonal coding mark in each image are detected and identified by using an image processing method, including: The center of the fitted ellipse is extracted by fitting the projection of the white positioning circle in the image as an ellipse; The intersection point of the two fitted straight lines is extracted as the diagonal detection center by fitting the projections of the two mutually perpendicular straight lines in the diagonal mark in the image respectively; The coordinate results obtained by the fitting ellipse center and the diagonal detection center are weighted and allocated according to the mean square error of the ellipse fitting and the straight line fitting, and the center coordinates of the circular diagonal coding mark are integrated and calculated.

Images