Visual geometry-based arch bridge cable crane carriage monitoring method

Through the visual geometric method based on the PNP algorithm, real-time and efficient positioning of the position of the steel pipe concrete arch bridge cable crane sports car is achieved, solving the problem of accurate positioning difficulties in construction, improving construction safety and efficiency, and reducing equipment costs and operating risks.

CN118968414AInactive Publication Date: 2024-11-15CHONGQING JIAOTONG UNIV

Patent Information

Application Number
CN202411057096.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-02
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

During the construction of steel pipe concrete arch bridges, it is difficult to accurately locate the cable cranes, the existing technology has limited accuracy or takes a long time, and cannot meet the real-time requirements, and the equipment costs are high and the operation risks are high.

Method used

The visual geometric method based on PNP algorithm combined with target tracking technology is adopted, through camera calibration, image acquisition and distortion correction, the camera position is solved by using the PNP algorithm to perform perspective transformation, real-time and efficient positioning of the position of the cable crane sports car.

Benefits of technology

It improves construction safety and efficiency, is easy to operate, has high accuracy, and has fast reaction speed, reduces equipment costs and operating risks, and adapts to different environmental conditions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118968414A_ABST
    Figure CN118968414A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of cable crane trolley construction, in particular to an arch bridge cable crane trolley monitoring method based on visual geometry. According to the method, the position and the motion state of the sports car in cable hoisting operation are monitored and positioned in real time by utilizing a vision measurement technology, so that an operator can conveniently control the operation of the cable crane. The method specifically comprises the steps that a PNP algorithm is used for solving an external parameter matrix from a camera coordinate system to a world coordinate system, then a pose transformation matrix is constructed, an inclined image is converted into an orthoimage through perspective transformation, at the moment, an arch bridge plane is parallel to a camera plane, and the arch bridge plane has the same scale in the orthoimage; and the real distance of the sports car can be measured and calculated through the scale, so that the sports car can be detected and positioned in real time. According to the method, accurate positioning of the position of the sports car is achieved through attitude estimation and orthographic projection images, real-time monitoring and cable rope position adjustment can be achieved under the condition that personnel do not need to work in the air, and the safety of the construction process is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of cable-suspended carriage construction, and in particular to a method for monitoring an arch bridge cable-suspended carriage based on visual geometry. Background Art

[0002] Steel tube concrete arch bridges are widely used in the mountainous areas of western my country due to their reasonable force, strong durability, high structural rigidity and good economy. This type of bridge is often used to cross valleys and rivers, has good spanning capacity, and can adapt to complex terrain and environmental conditions. The construction of steel tube concrete arch bridges not only meets the needs of transportation, but also significantly improves the economic and social benefits of the project.

[0003] During the construction of steel tube concrete arch bridges, it is crucial to accurately locate the position of the cable crane. Since these bridges usually have large spans and complex environmental conditions, operators may face obstructed vision during the lifting process. In this case, accurate positioning methods can ensure the smooth progress of the construction process and improve construction efficiency and safety. At present, the commonly used cable crane positioning methods in engineering include the use of high-precision measuring instruments such as total stations and GPS, combined with control points arranged on site for positioning. These measuring instruments can provide accurate three-dimensional coordinates to ensure the accuracy of the cable lifting position, but the corresponding time-consuming is long and cannot meet the real-time requirements. In addition, by setting conspicuous signs or markings on the bridge structure or temporary facilities, the operator confirms the position of the cable crane by observing these signs and markings. This method is simple and intuitive, but the accuracy is limited and it is easily affected by environmental factors. Modern cable lifting equipment is usually equipped with a remote control system, and the operator can monitor and adjust the position of the cable in real time through a computer or console. This method combines video monitoring and sensor data to provide higher accuracy and safety, but the equipment cost is high. In some special cases, operators may rely on visual inspection and walkie-talkies for coordination and adjustment. This method is often used for fine adjustments or when high-precision instruments are not available, but has lower accuracy and reliability. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In view of the limitations of the prior art, the present invention proposes a method for quickly monitoring and feeding back the position of the sports car in the working plane of the cable crane based on the PNP algorithm combined with the target tracking technology. The method can locate the position of the sports car in the cable hoisting system in real time and efficiently, so that the operator can intuitively identify the precise position of the sports car. In addition, the monitoring method also has the advantages of simple operation, high accuracy and fast response speed, and effectively solves the related problems in the prior art. Through the present invention, the safety and efficiency of arch bridge construction and maintenance can be significantly improved.

[0006] (II) Technical solution

[0007] In order to achieve the above-mentioned purpose of real-time and efficient positioning of the cable-suspended sports car, so that the operator can intuitively identify the precise position of the sports car, the present invention provides the following technical solution: a method for monitoring a cable-suspended sports car on an arch bridge based on visual geometry, comprising the following steps:

[0008] S1: Use the calibration board to calibrate the camera to obtain the camera's intrinsic parameter matrix K and distortion parameter D;

[0009] S2: Before the cable hoisting system works, set up the camera and adjust the camera shooting range so that the camera shooting range covers the entire working plane of the cable hoisting system, including two cable towers on the left and right, and covers all cables and the maximum longitudinal activity range of the cable crane from top to bottom, and then adjust the focal length and related parameters to make the picture clear and stable;

[0010] S3: After the cable hoisting system operation starts, the camera is turned on to collect images, and distortion correction is performed on the collected images;

[0011] S4: Import the coordinates of the feature points of the arch bridge, use the PNP algorithm to solve the camera pose at this time, obtain the camera pose matrix [R|T], and transform the world coordinates of the feature points into the camera coordinate system;

[0012] S5: Invert the [R|T] matrix to obtain the matrix [C|t], and construct the pose transformation matrix [W1] based on the rotation matrix C to the camera coordinate system O c Perform a rotation transformation to obtain O′ c , and then construct the pose transformation matrix [W2] from the translation vector at this time, and transform the camera coordinate system O c ′ Perform translation transformation to translate the camera coordinate system to the front of the world coordinate system;

[0013] S6: Projecting the feature points onto the camera plane by orthographic projection, thereby obtaining 2D orthographic projection coordinates of the feature points;

[0014] S7: Set the default pixel coordinate system O p The pixel coordinates of the feature points under the condition are transformed into the pixel coordinate system O with the world coordinate system as the origin. p ', thereby translating the origin of the original image coordinate system to the position (u0, v0), and inverting the vertical coordinate system to match the orthographic projection coordinates, so that the pixel coordinates are aligned with the world coordinate system;

[0015] S8: Through 2D orthographic projection coordinates and O p‘ The one-to-one correspondence between 2D pixel coordinates in the coordinate system gives the perspective transformation matrix H;

[0016] S9: Perform perspective transformation on the arch bridge image of the current frame to obtain an orthographic projection of the arch bridge image;

[0017] S10: Calculate the image pixel scale S=L / d, where L is the distance between two points in the real coordinate system, and d is the pixel distance between two points in the orthoimage;

[0018] S11: Detect the position of the sports car on the orthographically projected cable image through a target detection algorithm, and return the default pixel coordinates of the sports car;

[0019] S12: transforming the default pixel coordinates of the sports car into the real coordinate system of the sports car with the arch foot of the arch bridge as the origin, and calculating the mileage x and elevation y of the sports car at this time by pixel scale, thereby completing the positioning.

[0020] Furthermore, the specific processing method of step S1 is:

[0021] A1: Place the calibration plate in the camera imaging area, ensure that the feature points on the calibration plate fully occupy a certain proportion in the camera field of view, and adjust the posture and distance of the calibration plate to ensure that the calibration plate can be photographed from different positions and angles;

[0022] A2: Use a camera to shoot the calibration plate and obtain at least 5 images of the calibration plate at different positions and angles. Ensure that the feature points on the calibration plate are clearly visible in each image and cover the entire imaging area.

[0023] A3: Extract feature points from each calibration plate image and calculate the coordinates of the feature points on the calibration plate in the image;

[0024] A4: Use the RANSAC algorithm to match the feature points extracted from different images to eliminate incorrect matching points, and calculate the world coordinates of the matching feature points in 3D space based on the known geometric structure and relative position of the calibration plate;

[0025] A5: Use the calibration algorithm to calculate the camera's intrinsic parameter matrix K and distortion parameter D using the obtained 3D feature point coordinates and the feature point coordinates in the 2D image.

[0026] Furthermore, the specific processing method of step S4 is as follows:

[0027] B1: Determine the 3D world coordinates of n (n≥3) feature points on the arch bridge plane as P1, P2, ..., Pn and the corresponding 2D pixel coordinates as p1, p2, ..., pn;

[0028] B2: Use the camera intrinsic parameter matrix K and eliminate the scale factor Z c To solve the pose transformation matrix [R|T] of the arch bridge world coordinate system to the camera coordinate system, the specific solution formula is as follows:

[0029]

[0030] Among them, R is a 3×3 rotation matrix and T is a 1×3 translation vector.

[0031] Furthermore, in step S4, the world coordinate system O of the arch bridge is w The point under the pose transformation is transformed to transform the world coordinates of the 3D point into the camera coordinate system. The expression is as follows:

[0032]

[0033] Furthermore, in step S5, the matrix [R|T] is inverted to obtain the matrix [C|t], and the camera coordinate system O c Perform rotation transformation to obtain the world coordinate system O of the arch bridge w X w , Y w , Z w The camera coordinate system O with three coordinate axes in the same direction c’ , and its calculation expression is as follows:

[0034]

[0035] Furthermore, in step S5, the camera coordinate system O c’ Perform translation transformation to obtain the final camera coordinate system O c '', translate the rotated camera coordinate system to the front of the world coordinate system, so that when the camera is shooting the arch bridge, n 3D points on the arch bridge are located in the camera coordinate system O c The coordinates under '' are calculated as follows:

[0036]

[0037] Furthermore, in the step S7, the default pixel coordinates of the origin of the known world coordinate system and the arch bridge feature point are (u0, v0) and (u i ,v i ), the image pixel coordinate system O p Transform to O p ', so that O p The 2D pixel coordinates in the coordinate system correspond one-to-one to the 2D coordinates of the orthographic projection, and the calculation relationship expression is as follows:

[0038]

[0039] Furthermore, in the steps S8 and S9, the perspective transformation matrix H is obtained according to the one-to-one correspondence between the 2D pixel coordinates before and after the camera posture transformation, and the image captured by the camera in real time is perspective transformed to convert the originally tilted image into an orthographic projection of the arch bridge image.

[0040] Furthermore, in the step S12, the pixel coordinate system O p Transform to the sports car coordinate system O with the arch foot as the origin p‘ Then, the pixel coordinates (u′, v′) of the sports car after coordinate transformation are multiplied by the image pixel scale S, so that the pixel coordinates of the sports car are returned to the real world coordinates, and the mileage x and elevation y of the sports car are fed back. The calculation expression is as follows:

[0041]

[0042] Where u and v are the image coordinates of the sports car in the orthophoto image, and u0 and v0 are the default image coordinate origins.

[0043] The present invention provides a method for monitoring a cable car on an arch bridge based on visual geometry, which has the following features:

[0044] Beneficial effects:

[0045] The present invention constructs a posture transformation matrix through the PNP algorithm, obtains the posture of the camera orthophoto of the arch bridge steel pipe, and then converts the oblique photography image into an orthophoto image through perspective transformation, so that the scale of each pixel point is the same, so that the real distance between two points in the image can be accurately measured. The PNP algorithm has a simple structure and fast operation speed, which significantly improves the measurement efficiency. Compared with the traditional sensor positioning and visual observation method, the present invention adopts visual measurement means, does not need to directly contact the sports car, monitors and tracks the sports car through camera equipment, avoids physical interference with the sports car structure, and significantly reduces the operation risk. The operator adjusts the operation of the cable-cabled sports car through the returned coordinates, and can flexibly adjust the layout and parameters of the camera equipment according to actual needs, adapt to the sports car monitoring needs under different environments and working conditions, increase the applicability and flexibility of the system, and enable it to be widely used in various monitoring scenarios. By changing the camera posture and the orthophoto projection image, the accurate positioning of the cable-cabled sports car is achieved. Compared with the traditional visual observation method, it does not require personnel to work in the air, improves safety, and can more quickly and efficiently feedback the position of the sports car, and the accurate judgment of machine vision solves the problem of large errors in traditional measurement methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 This is a flow chart of the method for monitoring a cable car on an arch bridge based on visual geometry according to Example 1 of the present invention.

[0047] Figure 2Schematic diagram of camera posture transformation in Example 1 of the present invention.

[0048] Figure 3 This is a rendering of a sports car that is orthogonally projected in Example 1 of the present invention.

[0049] Figure 4 This is a diagram showing the coordinate feedback effect of a sports car in Example 2 of the present invention. DETAILED DESCRIPTION

[0050] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] Example 1

[0052] refer to Figure 1-3 As shown, the first embodiment of the present invention provides a method for monitoring a cable car on an arch bridge based on visual geometry, comprising the following steps:

[0053] S1: Use the calibration board to calibrate the camera to obtain the camera's intrinsic parameter matrix K and distortion parameter D.

[0054] The specific processing method of this step is as follows:

[0055] A1: Place the calibration plate in the camera imaging area, ensure that the feature points on the calibration plate fully occupy a certain proportion in the camera field of view, and adjust the posture and distance of the calibration plate to ensure that the calibration plate can be photographed from different positions and angles;

[0056] A2: Use a camera to shoot the calibration plate and obtain at least 5 images of the calibration plate at different positions and angles. Ensure that the feature points on the calibration plate are clearly visible in each image and cover the entire imaging area.

[0057] A3: Extract feature points from each calibration plate image and calculate the coordinates of the feature points on the calibration plate in the image;

[0058] A4: Use the RANSAC algorithm to match the feature points extracted from different images to eliminate incorrect matching points, and calculate the world coordinates of the matching feature points in 3D space based on the known geometric structure and relative position of the calibration plate;

[0059] A5: Use the calibration algorithm to calculate the camera's intrinsic parameter matrix K and distortion parameter D using the obtained 3D feature point coordinates and the feature point coordinates in the 2D image.

[0060] S2: Before the cable hoisting system starts working, set up the camera and adjust the camera shooting range so that the camera shooting range covers the entire working plane of the cable hoisting system. The left and right sides of the picture include two cable towers, and the upper and lower parts cover the maximum longitudinal activity range of all cables and cable cranes. Then adjust the focal length and related parameters to make the picture clear and stable.

[0061] S3: After the cable lifting system operation starts, the camera is turned on to collect images, and distortion correction is performed on the collected images.

[0062] S4: Import the coordinates of the feature points of the arch bridge, use the PNP algorithm to solve the camera pose at this time, obtain the camera pose matrix [R|T], and transform the world coordinates of the feature points to the camera coordinate system.

[0063] The specific processing method of this step is as follows:

[0064] B1: Determine the 3D world coordinates of n (n≥3) feature points on the arch bridge plane as P1, P2, ..., Pn and the corresponding 2D pixel coordinates as p1, p2, ..., pn;

[0065] B2: Use the camera intrinsic parameter matrix K and eliminate the scale factor Z c To solve the pose transformation matrix [R|T] of the arch bridge world coordinate system to the camera coordinate system, the specific solution formula is as follows:

[0066]

[0067] Among them, R is a 3×3 rotation matrix and T is a 1×3 translation vector.

[0068] World coordinate system O for arch bridge w The point under the pose transformation is transformed to transform the world coordinates of the 3D point into the camera coordinate system. The expression is as follows:

[0069]

[0070] S5: Invert the [R|T] matrix to obtain the matrix [C|t], and construct the pose transformation matrix [W1] based on the rotation matrix C to the camera coordinate system O c Perform a rotation transformation to obtain O′ c , and then construct the pose transformation matrix [W2] from the translation vector at this time, and transform the camera coordinate system O c ′ Make a translation transformation to move the camera coordinate system to the front of the world coordinate system.

[0071] In this step, the matrix [R|T] is inverted to obtain the matrix [C|t], and the camera coordinate system O c Perform rotation transformation to obtain the world coordinate system O of the arch bridge w X w , Yw , Z w The camera coordinate system O with three coordinate axes in the same direction c’ , and its calculation expression is as follows:

[0072]

[0073] For the camera coordinate system O c’ Perform translation transformation to obtain the final camera coordinate system O c '', translate the rotated camera coordinate system to the front of the world coordinate system, so that when the camera is shooting the arch bridge, n 3D points on the arch bridge are located in the camera coordinate system O c The coordinates under '' are calculated as follows:

[0074]

[0075] S6: Project the feature points onto the camera plane by orthographic projection, thereby obtaining 2D orthographic projection coordinates of the feature points.

[0076] S7: Set the default pixel coordinate system O p The pixel coordinates of the feature points under the condition are transformed into the pixel coordinate system O with the world coordinate system as the origin. p ', thereby translating the origin of the original image coordinate system to the position (u0, v0) and inverting the vertical coordinate system to match the orthographic projection coordinates, so that the pixel coordinates are aligned with the world coordinate system.

[0077] It is known that the default pixel coordinates of the origin of the world coordinate system and the arch bridge feature point are (u0,v0) and (u i ,v i ), the image pixel coordinate system O p Transform to O p ', so that O p The 2D pixel coordinates in the coordinate system correspond one-to-one to the 2D coordinates of the orthographic projection, and the calculation relationship expression is as follows:

[0078]

[0079] S8: Through 2D orthographic projection coordinates and O p‘ The one-to-one correspondence between 2D pixel coordinates in the coordinate system gives the perspective transformation matrix H.

[0080] S9: Perform perspective transformation on the arch bridge image of the current frame to obtain an orthographic projection of the arch bridge image.

[0081] The perspective transformation matrix H is obtained according to the one-to-one correspondence between the 2D pixel coordinates before and after the camera posture transformation. The image captured by the camera in real time is transformed into an orthographic projection image of the arch bridge.

[0082] S10: Calculate the image pixel scale S=L / d, where L is the distance between two points in the real coordinate system, and d is the pixel distance between the two points in the orthoimage.

[0083] S11: Detect the position of the sports car on the orthographically projected cable image through a target detection algorithm, and return the default pixel coordinates of the sports car.

[0084] S12: transforming the default pixel coordinates of the sports car into the real coordinate system of the sports car with the arch foot of the arch bridge as the origin, and calculating the mileage x and elevation y of the sports car at this time by pixel scale, thereby completing the positioning.

[0085] Specifically, the pixel coordinate system O p Transform to the sports car coordinate system O with the arch foot as the origin p‘ Then, the pixel coordinates (u′, v′) of the sports car after coordinate transformation are multiplied by the image pixel scale S, so that the pixel coordinates of the sports car are returned to the real world coordinates, and the mileage x and elevation y of the sports car are fed back. The calculation expression is as follows:

[0086]

[0087] Where u and v are the image coordinates of the sports car in the orthophoto image, and u0 and v0 are the default image coordinate origins.

[0088] Example 2

[0089] On the arch bridge image after orthographic projection, feature points P1 and P2 are selected to calculate the scale of the image after perspective transformation, where the world coordinates and default pixel coordinates of P1 are and The world coordinates and pixel coordinates of P2 are and The image scale is Use pixel scale to locate the sports car. The default pixel coordinates of the origin in the orthophoto image are: The default pixel coordinates of the sports car are

[0090] Then the mileage of the sports car is x = (u-u0)S = (428-59) × 0.2584 ≈ 95.46m;

[0091] Height y = (v0-v)S = (418-360) × 0.2584 ≈ 15.04m, Figure 4 The diagram shows the feedback effect of the sports car coordinates after the sports car is positioned using the above calculation method.

[0092] In summary, the method proposed in the present invention can be used to quickly monitor the position of the sports car in real time, utilize the properties of the orthophoto, use the global scale factor to calculate the true distance between two points in the image, and complete the detection of the cable-suspended sports car system through the method of multi-view geometry. The accurate positioning of the sports car position is achieved through posture calculation and orthographic projection images. Compared with the traditional sports car positioning method, the advantage is that it reduces the use and maintenance of expensive equipment, does not require personnel to work in the air, and the layout of the visual system is more flexible, and can be relatively easily moved and adjusted to adapt to different monitoring scenarios and needs. The feedback on the sports car position can be faster and more efficient, effectively improving the safety and efficiency of the sports car lifting process.

[0093] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0094] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for monitoring a cable car on an arch bridge based on visual geometry, characterized in that: The following steps are involved: S1: Use the calibration board to calibrate the camera to obtain the camera's intrinsic parameter matrix K and distortion parameter D; S2: Before the cable hoisting system works, set up the camera and adjust the camera shooting range so that the camera shooting range covers the entire working plane of the cable hoisting system, including two cable towers on the left and right, and covers all cables and the maximum longitudinal activity range of the cable crane from top to bottom, and then adjust the focal length and related parameters to make the picture clear and stable; S3: After the cable hoisting system operation starts, the camera is turned on to collect images, and distortion correction is performed on the collected images; S4: Import the coordinates of the feature points of the arch bridge, use the PNP algorithm to solve the camera pose at this time, obtain the camera pose matrix [R|T], and transform the world coordinates of the feature points into the camera coordinate system; S5: Invert the [R|T] matrix to obtain the matrix [C|t], and construct the pose transformation matrix [W1] based on the rotation matrix C to the camera coordinate system O c Perform a rotation transformation to obtain O′ c , and then construct the pose transformation matrix [W2] from the translation vector at this time, and transform the camera coordinate system O′ c Make a translation transformation to move the camera coordinate system to the front of the world coordinate system; S6: Projecting the feature points onto the camera plane by orthographic projection, thereby obtaining 2D orthographic projection coordinates of the feature points; S7: Set the default pixel coordinate system O p The pixel coordinates of the feature points under the pixel coordinate system O are transformed to the pixel coordinate system O with the world coordinate system as the origin. p ', thereby translating the origin of the original image coordinate system to the position (u0, v0), and inverting the vertical coordinate system to match the orthographic projection coordinates, so that the pixel coordinates are aligned with the world coordinate system; S8: Through 2D orthographic projection coordinates and O p‘ The one-to-one correspondence between 2D pixel coordinates in the coordinate system gives the perspective transformation matrix H; S9: Perform perspective transformation on the arch bridge image of the current frame to obtain an orthographic projection of the arch bridge image; S10: Calculate the image pixel scale S=L / d, where L is the distance between two points in the real coordinate system, and d is the pixel distance between two points in the orthoimage; S11: Detect the position of the sports car on the orthographically projected cable image through a target detection algorithm, and return the default pixel coordinates of the sports car; S12: transforming the default pixel coordinates of the sports car into the real coordinate system of the sports car with the arch foot of the arch bridge as the origin, and calculating the mileage x and elevation y of the sports car at this time by pixel scale, thereby completing the positioning.

2. The method for monitoring a cable car on an arch bridge based on visual geometry according to claim 1, characterized in that: The specific processing method of step S1 is: A1: Place the calibration plate in the camera imaging area, ensure that the feature points on the calibration plate fully occupy a certain proportion in the camera field of view, and adjust the posture and distance of the calibration plate to ensure that the calibration plate can be photographed from different positions and angles; A2: Use a camera to shoot the calibration plate and obtain at least 5 images of the calibration plate at different positions and angles. Ensure that the feature points on the calibration plate are clearly visible in each image and cover the entire imaging area. A3: Extract feature points from each calibration plate image and calculate the coordinates of the feature points on the calibration plate in the image; A4: Use the RANSAC algorithm to match the feature points extracted from different images to eliminate incorrect matching points, and calculate the world coordinates of the matching feature points in 3D space based on the known geometric structure and relative position of the calibration plate; A5: Use the calibration algorithm to calculate the camera's intrinsic parameter matrix K and distortion parameter D using the obtained 3D feature point coordinates and the feature point coordinates in the 2D image.

3. The method for monitoring a cable car on an arch bridge based on visual geometry according to claim 2 is characterized in that: The specific processing method of step S4 is: B1: Determine the 3D world coordinates of n (n≥3) feature points on the arch bridge plane as P1, P2, ..., Pn and the corresponding 2D pixel coordinates as p1, p2, ..., pn; B2: Use the camera intrinsic parameter matrix K and eliminate the scale factor Z c To solve the pose transformation matrix [R|T] of the arch bridge world coordinate system to the camera coordinate system, the specific solution formula is as follows: Among them, R is a 3×3 rotation matrix and T is a 1×3 translation vector.

4. The method for monitoring a cable car on an arch bridge based on visual geometry according to claim 3 is characterized by: In step S4, the world coordinate system O of the arch bridge is w The point under the pose transformation is transformed to transform the world coordinates of the 3D point into the camera coordinate system. The expression is as follows:

5. The method for monitoring a cable car on an arch bridge based on visual geometry according to claim 4, characterized in that: In step S5, the matrix [R|T] is inverted to obtain the matrix [C|t], and the camera coordinate system O c Perform rotation transformation to obtain the world coordinate system O of the arch bridge w X w , Y w , Z w The camera coordinate system O with three coordinate axes in the same direction c ', its calculation expression is as follows:

6. The method for monitoring a cable car on an arch bridge based on visual geometry according to claim 5, characterized in that: In step S5, the camera coordinate system O c 'Perform translation transformation to obtain the final camera coordinate system O c '', translate the rotated camera coordinate system to the front of the world coordinate system, so that when the camera is shooting the arch bridge, n 3D points on the arch bridge are located in the camera coordinate system O c The coordinates under '' are calculated as follows:

7. The method for monitoring a cable car on an arch bridge based on visual geometry according to claim 6, characterized in that: In the step S7, it is known that the default pixel coordinates of the origin of the world coordinate system and the arch bridge feature point are (u0, v0) and (u i ,v i ), the image pixel coordinate system O p Transform to O p ', so that at this time O p The 2D pixel coordinates in the coordinate system correspond one-to-one to the 2D coordinates of the orthographic projection, and the calculation relationship expression is as follows:

8. The method for monitoring a cable car on an arch bridge based on visual geometry according to claim 7, characterized in that: In the steps S8 and S9, the perspective transformation matrix H is obtained according to the one-to-one correspondence between the 2D pixel coordinates before and after the camera posture transformation, and the image captured by the camera in real time is perspective transformed to convert the originally tilted image into an orthographic projection of the arch bridge image.

9. The method for monitoring a cable car on an arch bridge based on visual geometry according to claim 8, characterized in that: In the step S12, the pixel coordinate system O p Transform to the sports car coordinate system O with the arch foot as the origin p‘ Then, the pixel coordinates (u′, v′) of the sports car after coordinate transformation are multiplied by the image pixel scale S, so that the pixel coordinates of the sports car are returned to the real world coordinates, and the mileage x and elevation y of the sports car are fed back. The calculation expression is as follows: Where u and v are the image coordinates of the sports car in the orthophoto image, and u0 and v0 are the default image coordinate origins.

Citation Information

Patent Citations

  • Single-lens three-dimensional image reconstruction method based on laser radar point cloud data assistance

    CN112102458A

  • Vehicle spatio-temporal information identification method based on computer vision and pose estimation

    CN115205397A

  • Method for detecting vertical splicing line shape of arch rib of steel tube arch bridge

    CN115540765A

  • Construction method for installing steel cover beam on concrete-filled steel tube arch bridge

    CN115976966A

  • Infrared vision positioning method for concrete pumping liquid level in steel tube arch bridge tube

    CN117968796A

Cited By

  • Arch bridge cable hoisting dynamic measurement and early warning method based on machine vision

    CN120117536A

  • Image correction method and system based on camera pose, equipment and storage medium

    CN120510230A

  • Bridge hoisting pose real-time measurement method and system based on monocular vision

    CN120635211A

  • A monocular vision-based bridge hoisting position real-time measurement method and system

    CN120635211B