Bridge stay cable monitoring method and device
By constructing a three-dimensional point cloud sequence using binocular cameras and feature matching technology, and combining it with multi-order natural frequencies and finite element models, the problem of insufficient spatial morphology perception in bridge cable-stayed monitoring was solved, and the accuracy of cable force monitoring was improved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- WUHAN UNIV OF TECH
- Filing Date
- 2025-12-30
- Publication Date
- 2026-04-21
AI Technical Summary
Existing bridge cable-stayed bridge monitoring schemes are unable to accurately perceive spatial topography, resulting in insufficient accuracy of cable force monitoring results. Contact monitoring schemes affect accuracy, while frequency methods rely on theoretical models and cannot fully consider boundary conditions.
Binocular cameras were used to acquire images of the cable-stayed bridge. A three-dimensional point cloud sequence was constructed through feature extraction and matching. Cable force inversion was performed by combining multi-order natural frequencies and a finite element model. Feature matching was performed using a feature extraction network and a graph neural network. The LK optical flow algorithm was used for time-series tracking to construct the cable force monitoring results.
This improves the accuracy of bridge cable-stayed bridge monitoring by combining measured data with theoretical inversion, ensuring the precision and reliability of monitoring results.
Smart Images
Figure CN121898288A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of bridge monitoring technology, and in particular to a method and device for monitoring bridge stay cables. Background Technology
[0002] The cable-stayed bridge is the core load-bearing component of the cable-stayed bridge, and accurate monitoring of its cable force is crucial for assessing the overall safety and service performance of the bridge. Currently, cable force monitoring technology has formed a pattern dominated by the frequency method, with various direct sensing technologies coexisting and continuously developing, and showing a trend of development from single-point static monitoring to distributed, dynamic, and intelligent monitoring.
[0003] Existing bridge cable-stayed bridge monitoring schemes are unable to perceive the spatial morphology of the cables, resulting in insufficient accuracy of monitoring results. Contact-based monitoring schemes also have a significant impact on the accuracy of monitoring results. The accuracy of the frequency method also depends on the theoretical model, but the theoretical model cannot fully consider the boundary conditions, which also leads to insufficient accuracy of the monitoring results of the frequency method.
[0004] Therefore, improving the accuracy of cable tension monitoring for bridge stay cables has become an urgent technical problem to be solved. Summary of the Invention
[0005] In view of this, it is necessary to provide a monitoring method and device for bridge stay cables to solve the problem of insufficient accuracy in existing bridge stay cable force monitoring.
[0006] To address the aforementioned problems, in a first aspect, the present invention provides a method for monitoring bridge stay cables, comprising: Acquire monitoring images of the cable stays of the target bridge within the target time period. The monitoring images of the cable stays of the target bridge are obtained by cameras located at two different positions. Feature extraction and feature matching are performed on the preprocessed monitoring images of the cable-stayed bridge within the target time period. Based on the feature matching results of the preprocessed monitoring images of the cable-stayed bridge within the target time period and the calibration parameters between two cameras at different locations, a three-dimensional point cloud sequence corresponding to the cable-stayed bridge within the target time period is constructed. The average cable force and multiple natural frequencies of the cable stays of the target bridge within the target time period are determined based on the three-dimensional point cloud sequence corresponding to the cable stays of the target bridge within the target time period. Based on the multi-order natural frequencies of the stay cables of the target bridge within the target time period and the corresponding three-dimensional point cloud sequence of the stay cables of the target bridge within the target time period, the cable force inversion is performed to determine the inverted cable force of the stay cables of the target bridge within the target time period. Based on the average cable force and the inverted cable force of the stay cables of the target bridge within the target time period, the monitoring results of the stay cables of the target bridge within the target time period are determined.
[0007] In one possible implementation, the feature extraction and feature matching of the preprocessed monitoring images of the cable-stayed bridge within the target time period includes: Feature extraction is performed on the monitoring images of the cable-stayed bridge of the target bridge within the target time period based on the feature extraction network. The feature extraction network is trained on the SuperPoint network using a general image dataset and a sample image dataset of cable-stayed bridges as samples. The encoder of the SuperPoint network is equipped with a convolutional block attention module at the end. The graph neural network matching algorithm is used to match feature points at the same sampling time in the monitoring images of the cable-stayed bridge of the target bridge within the target time period, and the LK optical flow algorithm is used to track the temporal sequence of the feature points.
[0008] In one possible implementation, the construction of a 3D point cloud sequence corresponding to the cable-stayed bridge within the target time period, based on the feature matching results of the preprocessed monitoring images of the cable-stayed bridge within the target time period and the calibration parameters between two cameras at different locations, includes: Based on the calibration parameters between two cameras at different locations, the coordinates of each feature point in the three-dimensional space of the feature matching result of the cable-stayed monitoring image of the target bridge in the preprocessed target time period are determined, and the three-dimensional point cloud sequence corresponding to the cable-stayed cable of the target bridge in the target time period is constructed. The calibration parameters between the two cameras at different locations include camera parameters, lens distortion coefficients, and rotation matrix and translation vector between the two cameras at different locations.
[0009] In one possible implementation, determining the average cable force and multiple natural frequencies of the stay cables of the target bridge within the target time period based on the three-dimensional point cloud sequence corresponding to the stay cables of the target bridge within the target time period includes: The static three-dimensional point cloud sequence is determined by averaging the three-dimensional point cloud sequence corresponding to the cable stays of the target bridge within the target time period in the time dimension. The static 3D point cloud sequence is projected onto a 2D plane defined by the two endpoints of the cable to determine the 2D cable shape point set, and the 2D cable shape point set is fitted to determine the cable shape curve; Based on the cable shape curve, determine the average cable force of the stay cables of the target bridge within the target time period; Modal analysis is performed based on the three-dimensional point cloud sequence corresponding to the stay cables of the target bridge within the target time period to determine the multiple natural frequencies of the stay cables of the target bridge within the target time period.
[0010] In one possible implementation, the process of inverting cable forces based on the multi-order natural frequencies of the stay cables of the target bridge within the target time period and the corresponding three-dimensional point cloud sequence of the stay cables of the target bridge within the target time period, to determine the inverted cable forces of the stay cables of the target bridge within the target time period, includes: A finite element model is constructed based on cable force, bending stiffness, and end boundary constraint stiffness. A multi-objective function is constructed that includes shape residuals and frequency residuals. The shape residuals are used to represent the root mean square error between the static three-dimensional point cloud sequence and the node positions of the finite element model. The frequency residuals are used to represent the sum of the relative errors between the multi-order natural frequencies of the cable stays of the target bridge and the multi-order natural frequencies of the cable stays of the target bridge determined by the finite element model during the target time period. The cable force that minimizes the multi-objective function corresponding to the finite element model is determined based on the multi-objective optimization algorithm, and is used as the inverse cable force of the cable-stayed bridge in the target time period.
[0011] In one possible implementation, determining the monitoring results of the stay cables of the target bridge within the target time period based on the average cable force and inverted cable force of the stay cables within the target time period includes: If the relative error between the average cable force of the cable-stayed bridge in the target time period and the inverted cable force of the cable-stayed bridge in the target time period is less than or equal to the error threshold, the inverted cable force of the cable-stayed bridge in the target time period or the average of the average cable force and the inverted cable force of the cable-stayed bridge in the target time period shall be determined as the monitoring result of the cable-stayed bridge in the target time period. If the relative error between the average cable force of the stay cables of the target bridge within the target time period and the inverted cable force of the stay cables of the target bridge within the target time period is greater than the error threshold, the average cable force and inverted cable force of the stay cables of the target bridge within the target time period, as well as the uncertainty range corresponding to the average cable force and inverted cable force of the stay cables of the target bridge within the target time period, are determined as the monitoring results of the stay cables of the target bridge within the target time period.
[0012] In one possible implementation, the preprocessing of the cable-stayed bridge monitoring images within the target time period includes: The monitoring images of the cable-stayed bridge within the target time period are sequentially converted to grayscale, histogram equalization, and guided filtering.
[0013] On the other hand, the present invention also provides a monitoring device for bridge stay cables, comprising: The acquisition module is used to acquire monitoring images of the cable stays of the target bridge within the target time period. The monitoring images of the cable stays of the target bridge are captured by cameras at two different locations. The module is used to extract and match features from the preprocessed monitoring images of the cable-stayed bridge within the target time period, and to construct a three-dimensional point cloud sequence corresponding to the cable-stayed bridge within the target time period based on the feature matching results of the preprocessed monitoring images of the cable-stayed bridge within the target time period and the calibration parameters between two cameras at different locations. The first determining module is used to determine the average cable force and multiple natural frequencies of the cable-stayed bridge within the target time period based on the three-dimensional point cloud sequence corresponding to the cable-stayed bridge within the target time period. The second determining module is used to perform cable force inversion based on the multi-order natural frequencies of the cable stays of the target bridge within the target time period and the corresponding three-dimensional point cloud sequence of the cable stays of the target bridge within the target time period, to determine the inverted cable force of the cable stays of the target bridge within the target time period, and to determine the monitoring results of the cable stays of the target bridge within the target time period based on the average cable force and the inverted cable force of the cable stays of the target bridge within the target time period.
[0014] Secondly, the present invention also provides a monitoring device, including a memory and a processor, wherein, The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the bridge cable-stayed cable monitoring method described in any of the above implementations.
[0015] Thirdly, the present invention also provides a computer-readable storage medium for storing a computer-readable program or instruction, which, when executed by a processor, can implement the steps in the bridge cable monitoring method described in any of the above implementations.
[0016] The beneficial effects of this invention are as follows: The bridge cable-stayed cable monitoring method and device provided by this invention acquire monitoring images of the bridge cable-stayed cables through a binocular camera to improve the accuracy of monitoring data, thereby ensuring the accuracy of subsequent bridge cable-stayed cable monitoring. Then, a three-dimensional point cloud sequence corresponding to the cable-stayed cables of the target bridge within the target time period is constructed to obtain cable force monitoring values based on measured data. At the same time, cable force inversion is also performed to obtain cable force inversion values based on theory. Finally, the final bridge cable-stayed cable monitoring results are determined based on both, thus ensuring the accuracy of bridge cable-stayed cable monitoring. Attached Figure Description
[0017] Figure 1 A schematic flowchart of an embodiment of the monitoring method for bridge stay cables provided by the present invention; Figure 2 This is a schematic flowchart of an embodiment of the cable force monitoring process provided by the present invention; Figure 3 A schematic diagram of an embodiment of the stereo vision measurement system provided by the present invention; Figure 4 A schematic flowchart of an embodiment of the cable force inversion process provided by the present invention; Figure 5 A schematic diagram of an embodiment of the monitoring device for bridge stay cables provided by the present invention; Figure 6 A schematic diagram of an embodiment of the monitoring device provided by the present invention. Detailed Implementation
[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.
[0019] In the description of the embodiments of the present invention, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.
[0020] The terms "first," "second," etc., used in the embodiments of this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a technical feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature.
[0021] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.
[0022] This invention provides a method and device for monitoring bridge stay cables, which will be described below.
[0023] Figure 1 This is a schematic flowchart of an embodiment of the monitoring method for bridge stay cables provided by the present invention, as shown below. Figure 1 As shown, the monitoring methods for bridge stay cables include: S101. Acquire monitoring images of the cable stays of the target bridge within the target time period. The monitoring images of the cable stays of the target bridge are obtained by two cameras located at different positions.
[0024] It should be noted that the bridge cable monitoring method provided by this invention can be applied to bridge monitoring scenarios, especially the cable force monitoring scenario of bridge cables.
[0025] When monitoring bridge stay cables, the monitoring equipment (such as a portable or desktop computer) first acquires monitoring images of the target bridge's stay cables within the target time period. These images are captured by two cameras located at different positions (i.e., a binocular camera system). Using these images from two cameras at different locations effectively improves the accuracy of subsequent bridge stay cable monitoring.
[0026] S102. Perform feature extraction and feature matching on the preprocessed monitoring images of the cable-stayed bridge within the target time period. Based on the feature matching results of the preprocessed monitoring images of the cable-stayed bridge within the target time period and the calibration parameters between two cameras at different locations, construct a three-dimensional point cloud sequence corresponding to the cable-stayed bridge within the target time period.
[0027] It should be noted that after acquiring the cable-stayed bridge monitoring images of the target bridge within the target time period, the cable-stayed bridge monitoring images can be preprocessed first, and then feature extraction and feature matching can be performed on the preprocessed cable-stayed bridge monitoring images of the target bridge within the target time period. Then, based on the feature matching results and the calibration parameters between two cameras at different locations, a three-dimensional point cloud sequence corresponding to the cable-stayed bridge within the target time period can be constructed, providing data basis for bridge cable-stayed bridge monitoring.
[0028] S103. Determine the average cable force and multiple natural frequencies of the cable-stayed bridge within the target time period based on the three-dimensional point cloud sequence corresponding to the cable-stayed bridge within the target time period.
[0029] It should be noted that after constructing the three-dimensional point cloud sequence corresponding to the cable stays of the target bridge within the target time period, the average cable force and multiple natural frequencies of the cable stays of the target bridge within the target time period can be determined through the three-dimensional point cloud sequence corresponding to the cable stays of the target bridge within the target time period, thus completing the monitoring based on the measured data.
[0030] S104. Based on the multi-order natural frequencies of the stay cables of the target bridge within the target time period and the corresponding three-dimensional point cloud sequence of the stay cables of the target bridge within the target time period, perform cable force inversion to determine the inverted cable force of the stay cables of the target bridge within the target time period. Based on the average cable force and the inverted cable force of the stay cables of the target bridge within the target time period, determine the monitoring results of the stay cables of the target bridge within the target time period.
[0031] It should be noted that, in addition to monitoring based on measured data, cable force inversion can also be performed using the multi-order natural frequencies of the stay cables of the target bridge within the target time period and the corresponding three-dimensional point cloud sequence of the stay cables of the target bridge within the target time period. This allows for the determination of the inverted cable force of the stay cables of the target bridge within the target time period. Then, by using the average cable force and the inverted cable force of the stay cables of the target bridge within the target time period, the monitoring results of the stay cables of the target bridge within the target time period can be determined, thereby further improving the accuracy of bridge stay cable monitoring.
[0032] In summary, the bridge stay cable monitoring method provided by this invention acquires monitoring images of the bridge stay cables using a binocular camera to improve the accuracy of the monitoring data, thereby ensuring the accuracy of subsequent bridge stay cable monitoring. Then, a three-dimensional point cloud sequence corresponding to the stay cables of the target bridge within the target time period is constructed to obtain cable force monitoring values based on measured data. At the same time, cable force inversion is performed to obtain cable force inversion values based on theory. Finally, the final bridge stay cable monitoring result is determined based on both, thus ensuring the accuracy of bridge stay cable monitoring.
[0033] In some embodiments of the present invention, the step of extracting and matching features from the preprocessed cable-stayed monitoring images of the target bridge within the target time period includes: Feature extraction is performed on the monitoring images of the cable-stayed bridge of the target bridge within the target time period based on the feature extraction network. The feature extraction network is trained on the SuperPoint network using a general image dataset and a sample image dataset of cable-stayed bridges as samples. The encoder of the SuperPoint network is equipped with a convolutional block attention module at the end. The graph neural network matching algorithm is used to match feature points at the same sampling time in the monitoring images of the cable-stayed bridge of the target bridge within the target time period, and the LK optical flow algorithm is used to track the temporal sequence of the feature points.
[0034] It should be noted that when performing feature extraction and feature matching on the preprocessed monitoring images of the cable-stayed bridge within the target time period, image feature extraction can first be performed using an improved SuperPoint network. The improved SuperPoint network introduces a convolutional block attention module at the end of the encoder, which automatically learns and weights information-rich channels and spatial locations in the image. The improved SuperPoint network is first pre-trained on a general dataset (such as MS-COCO), and then transfer learning is performed using a self-built dataset containing a large number of low-texture images of the cable-stayed lines, specializing it in identifying stable features of the cable surface. During deployment, the preprocessed images are input into the network, and the network outputs the feature point locations (pixel coordinates) of each image and their corresponding high-dimensional feature descriptors.
[0035] During feature matching, graph neural network matching algorithms (such as SuperGlue) can be used to match feature points of two camera images at the same time to obtain initial matching point pairs. This algorithm utilizes the similarity of feature descriptors and the local geometric consistency between feature points for optimal matching, and has a strong ability to suppress false matches. Then, for the 3D points that have been successfully triangulated in frame t, the LK optical flow method is used to predict their approximate positions in the images around frames t+1, forming a small search window. Within this window, feature descriptors are used for precise matching.
[0036] In some embodiments of the present invention, the construction of a three-dimensional point cloud sequence corresponding to the cable-stayed bridge within the target time period based on the feature matching results of the preprocessed monitoring images of the cable-stayed bridge within the target time period and the calibration parameters between two cameras at different locations includes: Based on the calibration parameters between two cameras at different locations, the coordinates of each feature point in the three-dimensional space of the feature matching result of the cable-stayed monitoring image of the target bridge in the preprocessed target time period are determined, and the three-dimensional point cloud sequence corresponding to the cable-stayed cable of the target bridge in the target time period is constructed. The calibration parameters between the two cameras at different locations include camera parameters, lens distortion coefficients, and rotation matrix and translation vector between the two cameras at different locations.
[0037] It should be noted that when constructing the three-dimensional point cloud sequence corresponding to the cable stays of the target bridge within the target time period based on the feature matching results of the preprocessed monitoring images of the cable stays of the target bridge within the target time period and the calibration parameters between two cameras at different locations, the coordinates of each feature point in the three-dimensional space can be calculated frame by frame using the triangulation method based on the calibration parameters between the two cameras at different locations, thus obtaining the three-dimensional point cloud sequence corresponding to the cable stays of the target bridge within the target time period.
[0038] In some embodiments of the present invention, determining the average cable force and multiple natural frequencies of the stay cables of the target bridge within a target time period based on the three-dimensional point cloud sequence corresponding to the stay cables of the target bridge within a target time period includes: The static three-dimensional point cloud sequence is determined by averaging the three-dimensional point cloud sequence corresponding to the cable stays of the target bridge within the target time period in the time dimension. The static 3D point cloud sequence is projected onto a 2D plane defined by the two endpoints of the cable to determine the 2D cable shape point set, and the 2D cable shape point set is fitted to determine the cable shape curve; Based on the cable shape curve, determine the average cable force of the stay cables of the target bridge within the target time period; Modal analysis is performed based on the three-dimensional point cloud sequence corresponding to the stay cables of the target bridge within the target time period to determine the multiple natural frequencies of the stay cables of the target bridge within the target time period.
[0039] It should be noted that when determining the average cable force and multiple natural frequencies of the cable-stayed bridge within a target time period based on the 3D point cloud sequence corresponding to the cable-stayed bridge within that target time period, the average value (i.e., static equilibrium position) of each point over the entire time series can be subtracted from the 3D coordinates of each point in the point cloud sequence to obtain the displacement time series data of each point on the three axes, forming the 3D dynamic displacement field of the cable. Then, the displacement field is averaged over the time dimension to obtain the static 3D point cloud sequence of the cable. Next, the 3D topographic point cloud is projected onto a 2D plane defined by the two ends of the cable to obtain a 2D cable shape point set. Using the nonlinear least squares method, these points are iteratively fitted with the catenary equation or parabola equation to obtain the optimal curve parameters. Finally, the average cable force is calculated based on the mechanical relationship between "cable shape and cable force" (considering the sag effect). The multiple natural frequencies of the cable-stayed bridge can be obtained by modal analysis of the 3D point cloud sequence corresponding to the cable-stayed bridge within the target time period.
[0040] In some embodiments of the present invention, the process of determining the inverted cable force of the stay cables of the target bridge within the target time period by performing cable force inversion based on the multi-order natural frequencies of the stay cables of the target bridge within the target time period and the corresponding three-dimensional point cloud sequence of the stay cables of the target bridge within the target time period includes: A finite element model is constructed based on cable force, bending stiffness, and end boundary constraint stiffness. A multi-objective function is constructed that includes shape residuals and frequency residuals. The shape residuals are used to represent the root mean square error between the static three-dimensional point cloud sequence and the node positions of the finite element model. The frequency residuals are used to represent the sum of the relative errors between the multi-order natural frequencies of the cable stays of the target bridge and the multi-order natural frequencies of the cable stays of the target bridge determined by the finite element model during the target time period. The cable force that minimizes the multi-objective function corresponding to the finite element model is determined based on the multi-objective optimization algorithm, and is used as the inverse cable force of the cable-stayed bridge in the target time period.
[0041] It should be noted that when determining the inverted cable forces of the target bridge's stay cables within the target time period by performing cable force inversion based on the multi-order natural frequencies of the target bridge's stay cables and the corresponding 3D point cloud sequence of the stay cables within the target time period, a finite element model can be constructed first based on the cable forces, bending stiffness, and end boundary constraint stiffness. A multi-objective function incorporating topographic and frequency residuals can then be built. Finally, a multi-objective optimization algorithm is used to determine the cable forces that minimize the multi-objective function corresponding to the finite element model, which are then used as the inverted cable forces of the target bridge's stay cables within the target time period.
[0042] In some embodiments of the present invention, determining the monitoring results of the stay cables of the target bridge within the target time period based on the average cable force and inverted cable force of the stay cables of the target bridge within the target time period includes: If the relative error between the average cable force of the cable-stayed bridge in the target time period and the inverted cable force of the cable-stayed bridge in the target time period is less than or equal to the error threshold, the inverted cable force of the cable-stayed bridge in the target time period or the average of the average cable force and the inverted cable force of the cable-stayed bridge in the target time period shall be determined as the monitoring result of the cable-stayed bridge in the target time period. If the relative error between the average cable force of the stay cables of the target bridge within the target time period and the inverted cable force of the stay cables of the target bridge within the target time period is greater than the error threshold, the average cable force and inverted cable force of the stay cables of the target bridge within the target time period, as well as the uncertainty range corresponding to the average cable force and inverted cable force of the stay cables of the target bridge within the target time period, are determined as the monitoring results of the stay cables of the target bridge within the target time period.
[0043] It should be noted that when determining the monitoring results of the stay cables of the target bridge within the target time period based on the average cable force and the inverted cable force, the relative error between the average cable force and the inverted cable force of the stay cables of the target bridge within the target time period can be calculated first. When the relative error is less than or equal to the error threshold (e.g., 5%), the inverted cable force of the stay cables of the target bridge within the target time period or the average of the average cable force and the inverted cable force of the stay cables of the target bridge within the target time period can be determined as the monitoring results of the stay cables of the target bridge within the target time period. When the relative error is greater than the error threshold, the average cable force and the inverted cable force of the stay cables of the target bridge within the target time period, as well as the uncertainty range corresponding to the average cable force and the inverted cable force of the stay cables of the target bridge within the target time period, can be determined as the monitoring results of the stay cables of the target bridge within the target time period. The uncertainty range corresponding to the average cable force and the inverted cable force of the target bridge during the target time period can be determined based on the relative error between the average cable force of the target bridge during the target time period and the inverted cable force of the target bridge during the target time period.
[0044] In some embodiments of the present invention, the preprocessing of the cable-stayed bridge monitoring images within the target time period includes: The monitoring images of the cable-stayed bridge within the target time period are sequentially converted to grayscale, histogram equalization, and guided filtering.
[0045] It should be noted that when preprocessing the monitoring images of the cable stays of the target bridge within the target time period, the monitoring images can be converted to grayscale, histogram equalization and guided filtering in sequence to enhance contrast and suppress noise.
[0046] Combination Figure 2 The bridge cable-stayed cable monitoring process provided by this invention specifically includes the following steps: 1. Deployment of stereo vision system and high-precision spatiotemporal calibration.
[0047] A stereo vision system consisting of at least two synchronously triggered high-speed, high-resolution industrial cameras is deployed at stable reference points such as bridge towers or bridge decks. In addition, a low-cost IMU is installed at the stereo vision pedestal to measure the vibration and attitude of the stereo vision system, providing a high-frequency reference for the motion of the reference points to address the problem of "dynamic tracking instability" in the vision system when the bridge towers or bridge decks vibrate violently. The motion state of the entire system can be decomposed into two parts: "motion of the reference points" and "flexible deformation of the stay cables relative to the reference points." The IMU's task is to accurately estimate the former, allowing the vision system to focus more on monitoring the latter.
[0048] An Extended Kalman Filter (EKF) or Graph Optimization (SLAM backend) framework is employed to fuse multi-source monitoring information. The state vector of this framework includes not only the 3D coordinates of the cable-stayed bridge feature points but also the overall motion state of the cable-stayed bridge. Its prediction-correction mechanism is as follows: a) Prediction step: Based on the cable-stayed bridge motion model and IMU data, predict the position of the cable-stayed bridge feature points at the next moment; b) Correction step: When visual measurement data arrives, compare the predicted value with the actual observed value, optimally correct the state estimate through a filtering algorithm, and output a smooth, continuous 3D trajectory; c) Short-term prediction: When the cable-stayed bridge loses visual perception due to other reasons, the system can rely on the motion model for short-term prediction to maintain data continuity, and then correct again after visual perception is restored.
[0049] The system calibration includes high-precision spatial calibration and high-synchronization temporal calibration. Spatial calibration employs the Zhang Zhengyou calibration method combined with nonlinear optimization to accurately acquire the camera's intrinsic and extrinsic parameters and distortion coefficients. Temporal calibration ensures that the synchronization error between the two cameras' exposure times is much smaller than the sampling interval (e.g., less than 0.1 ms), which is a prerequisite for the subsequent accurate extraction of high-frequency vibration signals.
[0050] 2. Dynamic feature tracking based on anti-interference deep feature network.
[0051] To achieve long-term robust monitoring, this invention employs an anti-interference deep feature network specifically designed for low-texture targets. This network architecture includes: a feature extraction module: based on a pre-trained SuperPoint network, channel and spatial attention mechanisms are introduced, enabling the network to focus on stable but subtle texture features on the cable surface (such as strand edges and anti-corrosion coating particles), suppressing interference from lighting changes and rain reflections. A sequence tracking module: abandoning traditional frame-by-frame matching, a sequence tracking algorithm based on optical flow guidance and graph matching fusion is adopted. This algorithm not only utilizes the appearance features of a single frame image but also combines the motion continuity between consecutive frames to achieve long-term, stable tracking of feature points in the video sequence. Even with brief occlusion, recapture can be achieved, ensuring the continuity of displacement data.
[0052] 3. High spatiotemporal resolution three-dimensional dynamic displacement field reconstruction.
[0053] Triangulation is performed on each frame of synchronized image to generate a time-varying 3D point cloud sequence. By tracking the trajectories of tens of thousands of feature points, the 3D displacement time-series data of densely packed measurement points on the cable surface is directly obtained, thereby constructing a high spatiotemporal resolution 3D dynamic displacement field for the cable.
[0054] Static morphology extraction and average cable force identification: By averaging the dynamic displacement field over time, the static three-dimensional equilibrium position of the cable with sub-pixel precision can be obtained, i.e., the initial spatial morphology. The obtained three-dimensional point cloud of the cable is projected onto a two-dimensional cable shape plane, thereby obtaining a cable shape point set that can be used to calculate the cable shape. Based on the cable's morphology and mechanical differential equations in three-dimensional space, the two-dimensional cable configuration (i.e., the initial cable shape) under its own weight and constant load is reconstructed. Based on the cable shape point set, the cable curve equation (such as parabola or catenary) is obtained by fitting using a nonlinear regression iterative algorithm, and the average cable force is calculated according to the mechanical relationship between "cable shape and cable force" (considering the sag effect).
[0055] Dynamic vibration information extraction: Modal analysis is performed on the displacement time series data of all measuring points, which can simultaneously identify the multiple natural frequencies of the cable and their corresponding three-dimensional full-field vibration modes. Compared with traditional accelerometers, which can only obtain single-direction vibrations from a few measuring points, this invention can obtain complete vibration mode information from tens of thousands of measuring points in three directions.
[0056] 4. Parametric finite element model update and cable force identification that integrates morphology and vibration information.
[0057] Parametric model construction: Establish a parametric finite element model of the cable, and set key mechanical parameters such as cable force, bending stiffness, end boundary constraint stiffness and cable density as variables to be identified.
[0058] Multi-objective collaborative inversion: The static morphology, multiple frequencies, and three-dimensional mode shapes obtained in step 3 are used as the joint objective function. A multi-objective optimization algorithm (such as NSGA-II) is employed to find a set of optimal mechanical parameters that achieve a globally optimal fit between the simulation results of the finite element model and the measured data under the least squares principle.
[0059] Cable force distribution identification: Based on the constructed finite element model, cable forces are identified and compared with the cable forces identified by the spatial cable force identification method based on cable sag effect in step 3, and the identification results are corrected.
[0060] Combination Figure 3 The core hardware of this invention is a stereo vision measurement unit, the specific structure of which is as follows: Image acquisition equipment: At least two high-speed, high-resolution industrial cameras should be deployed on a stable structure (such as the main beam or bridge tower) near the cable to be measured. The cameras must be mounted on a stable base and equipped with a protective shield to protect against wind and rain. A low-cost IMU should be installed at the base location. Global shutter speeds should be used for the cameras to avoid rolling shutter effects when capturing moving targets. The camera baseline length B and focal length f need to be optimized based on the measurement distance D, the cable diameter, and the required measurement accuracy to ensure that the cable occupies sufficient pixels in the image.
[0061] Synchronization and Control Unit: The two cameras achieve precise synchronization of exposure times via a synchronization trigger, with a synchronization error of less than 0.1 milliseconds, which is crucial for ensuring dynamic measurement accuracy. All equipment is controlled by an industrial control computer and can communicate with a remote monitoring center via wired or wireless networks.
[0062] Lighting Auxiliary Unit: To ensure monitoring at night or under low light conditions, a uniform lighting source, such as an LED supplementary light, can be added to ensure uniform illumination on the cable surface and reduce shadow interference.
[0063] This hardware deployment scheme enables non-contact, long-distance, and synchronous observation of the cables, providing high-quality, synchronous binocular video stream data for subsequent processing.
[0064] Before monitoring can begin, the stereo vision system must be precisely calibrated.
[0065] Spatial calibration: A high-precision checkerboard calibration plate is used. Multiple images (typically 15-25) of the calibration plate are taken from different angles and positions within the measurement area. Using Zhang Zhengyou's calibration method combined with nonlinear least squares optimization, the intrinsic parameter matrices (including focal length and principal point) and lens distortion coefficients of the two cameras, as well as the rotation matrix R and translation vector T (i.e., extrinsic parameters) between the two cameras, are accurately solved.
[0066] Function: Precise calibration parameters are the foundation for all subsequent three-dimensional coordinate calculations, and their accuracy directly determines the accuracy of the final morphology and displacement measurements.
[0067] The following provides a further explanation of the specific process of bridge monitoring: 1. Dynamic feature tracking based on anti-interference deep feature network.
[0068] 1) Image preprocessing and enhancement: The acquired raw images are preprocessed, including grayscale conversion and histogram equalization to enhance contrast, and guided filtering and other methods are used to suppress noise.
[0069] 2) Feature point self-identification: Network Structure: An improved SuperPoint network is used as the backbone. Specifically, a convolutional block attention module is introduced at the end of the encoder, which can automatically learn and weight information-rich channels and spatial locations in the image.
[0070] Training and Deployment: The network is first pre-trained on a general dataset (such as MS-COCO), and then transferred to a self-built dataset containing a large number of low-texture images of cable-stayed bridges to specialize it in recognizing stable features on the cable surface. During deployment, the pre-processed left and right images are input into the network, and the network outputs the feature point locations (pixel coordinates) of each image and their corresponding high-dimensional feature descriptors.
[0071] Function: The attention mechanism enables the network to focus on stable features such as the edges of twisted wires and minor imperfections in the paint surface, rather than volatile lighting, which greatly improves the robustness of feature extraction under complex lighting conditions.
[0072] 3) Feature matching and sequence tracking: Initial matching: A graph neural network matching algorithm (such as SuperGlue) is used to match feature points of the left and right images at the same time to obtain initial matching point pairs. This algorithm utilizes the similarity of feature descriptors and the local geometric consistency between feature points to perform optimal matching, and has a strong ability to suppress mismatches.
[0073] Sequence Tracking: To achieve long-term temporal tracking, this invention does not employ simple inter-frame matching, but instead uses an optical flow-guided tracking strategy. Specifically, for a successfully triangulated 3D point in frame t, the LK optical flow method is used to predict its approximate position in the image around frame t+1, forming a small search window. Then, within this window, feature descriptors are used for precise matching. This method combines the efficiency of optical flow with the accuracy of feature matching and can effectively handle re-tracking after brief occlusion.
[0074] Function: This strategy ensures that thousands of feature points are stably and continuously tracked over long time sequences, generating reliable three-dimensional displacement time series data.
[0075] 2. High spatiotemporal resolution three-dimensional dynamic displacement field reconstruction.
[0076] 1) 3D coordinate calculation: For each successfully matched feature point pair in a synchronized image frame, using the camera parameters obtained from step 2, the coordinates of each feature point in 3D space are calculated frame by frame using triangulation. For N continuously tracked feature points, after T frames, a 3D dynamic point cloud sequence with dimension N×T can be obtained.
[0077] 2) Dynamic displacement field extraction: Subtract the average value (i.e., static equilibrium position) of each point in the point cloud sequence from the three-dimensional coordinates of the point in the entire time series to obtain the displacement time series data of each point in three directions, which together constitute the three-dimensional dynamic displacement field of the cable.
[0078] Average cable force identification based on static morphology.
[0079] Static morphology: The static three-dimensional spatial configuration of the cable can be obtained by averaging the displacement field over the time dimension.
[0080] Cable curve fitting and cable force calculation: The three-dimensional topographic point cloud is projected onto a two-dimensional plane defined by the two ends of the cable to obtain a two-dimensional cable shape point set; the nonlinear least squares method is used to iteratively fit these points using the catenary equation or parabola equation to obtain the optimal curve parameters; finally, the average cable force S1 of the cable is calculated according to the mechanical relationship between "cable shape and cable force" (considering the sag effect).
[0081] Dynamic vibration information extraction.
[0082] Frequency and mode shape identification: By performing modal analysis on the displacement time series data of each point, the multiple natural frequencies of the cable and the corresponding three-dimensional full-field mode shape can be identified.
[0083] 3. Parametric model updating and cable force identification integrating morphology and vibration information. Combined with... Figure 4 Specifically, it includes: Parametric finite element model establishment: Establish a refined finite element model of the cable. Parameters that are difficult to predict precisely, such as cable force, bending stiffness, and end boundary constraint stiffness, are set as variables to be inverted.
[0084] 1) Multi-objective optimization inversion: Objective function construction: Construct a multi-objective function that includes shape residuals and frequency residuals.
[0085] Shape residual R_shape: Calculates the root mean square error between the measured static shape point set and the node positions of the finite element model.
[0086] Frequency residual R_freq: The sum of the relative errors between the measured first n frequencies and the first n frequencies calculated by the finite element model.
[0087] 2) Optimization Solution: A multi-objective optimization algorithm (such as NSGA-II) is used to search within the given parameter space. The optimization objective is to find an optimal combination of parameters that minimizes the objective function F = w1 × R_shape + w2 × R_freq (where w1 and w2 are weighting coefficients). After the optimization process converges, the T value in the optimal solution is the high-precision cable force obtained through inversion. Furthermore, this parameterized model itself, due to its high consistency with measured data, can serve as a digital twin model of the cable for subsequent health assessment.
[0088] 3) Result comparison, analysis and correction: Calculate the relative error δ between the cable force values S1 and S2 obtained from the two paths, and set a threshold ε (e.g., 5%).
[0089] If δ≤ε, it indicates that the results obtained by the two methods based on different physical principles are highly consistent, and the reliability of the results is extremely high. In this case, the theoretically more accurate result S2 from path two can be output first, or the weighted average of S1 and S2 can be used as the final cable force.
[0090] If δ > ε, it indicates a significant difference in the results. The system will issue a warning and initiate a diagnostic process. For example, the cable method is sensitive to boundary condition assumptions; if the boundaries do not match reality, it will lead to a large S1 error. The model update method may also be distorted if optimization is insufficient. In this case, the system will output the results of both paths and their uncertainty ranges, and suggest checking the boundary conditions or model parameters, providing clear guidance for manual intervention. This dual-path comparison mechanism greatly enhances the reliability and credibility of the monitoring system, upgrading it from "single result output" to "result credibility assessment."
[0091] This invention, through stereo vision and dynamic tracking, can synchronously and homogeneously extract high-precision static spatial morphology and complete three-dimensional dynamic vibration modes from the same data source, completely changing the situation where traditional methods can only obtain single information, and providing a complete data foundation for comprehensively evaluating the state of cables.
[0092] This invention abandons the reliance on a single simplified model and proposes a parametric finite element model update method. This method treats model parameters as unknowns, utilizing massive amounts of full-field measured data (morphology + mode shapes) to "constrain" and "train" them, significantly reducing the errors caused by uncertain boundary conditions and model simplification in the traditional frequency method. Theoretically, it can achieve higher cable force identification accuracy. Simultaneously, this method has the potential to identify the distribution trend of cable force along the cable length, providing a technical means for diagnosing local damage such as anchorage slippage and saddle wear.
[0093] This invention replaces manual labeling with deep learning feature self-identification technology and combines it with a dual-path force inversion and decision-making mechanism to construct a highly automated and intelligent system. This overcomes the poor reliability of traditional methods in complex outdoor environments, achieving truly long-term unattended automated monitoring. The dual-path verification mechanism enables the system to self-verify, allowing for the evaluation of credibility while outputting results, greatly enhancing the system's practical value and reliability.
[0094] To better implement the bridge stay cable monitoring method in this embodiment of the invention, based on the bridge stay cable monitoring method, correspondingly, as follows: Figure 5 As shown, this embodiment of the invention also provides a monitoring device for bridge stay cables. The monitoring device 500 for bridge stay cables includes: The acquisition module 501 is used to acquire monitoring images of the cable stays of the target bridge within the target time period. The monitoring images of the cable stays of the target bridge are captured by two cameras at different locations. The construction module 502 is used to extract and match features from the preprocessed monitoring images of the cable-stayed bridge in the target time period, and to construct a three-dimensional point cloud sequence corresponding to the cable-stayed bridge in the target time period based on the feature matching results of the preprocessed monitoring images of the cable-stayed bridge in the target time period and the calibration parameters between two cameras at different positions. The first determining module 503 is used to determine the average cable force and multiple natural frequencies of the cable of the target bridge within the target time period based on the three-dimensional point cloud sequence corresponding to the cable of the target bridge within the target time period. The second determining module 504 is used to perform cable force inversion based on the multi-order natural frequencies of the cable-stayed cables of the target bridge within the target time period and the corresponding three-dimensional point cloud sequence of the cable-stayed cables of the target bridge within the target time period, to determine the inverted cable force of the cable-stayed cables of the target bridge within the target time period, and to determine the monitoring results of the cable-stayed cables of the target bridge within the target time period based on the average cable force and the inverted cable force of the cable-stayed cables of the target bridge within the target time period.
[0095] The bridge cable monitoring device 500 provided in the above embodiments can realize the technical solutions described in the above bridge cable monitoring method embodiments. The specific implementation principles of each module or unit can be found in the corresponding content in the above bridge cable monitoring method embodiments, and will not be repeated here.
[0096] like Figure 6 As shown, the present invention also provides a monitoring device 600. The monitoring device 600 includes a processor 601, a memory 602, and a display 603. Figure 6 Only some components of the monitoring device 600 are shown; however, it should be understood that implementation of all shown components is not required, and more or fewer components may be implemented instead.
[0097] In some embodiments, processor 601 may be a central processing unit (CPU), microprocessor, or other data processing chip, used to run program code stored in memory 602 or process data, such as the bridge cable-stayed cable monitoring method of the present invention.
[0098] In some embodiments, processor 601 may be a single server or a group of servers. The server group may be centralized or distributed. In some embodiments, processor 601 may be local or remote. In some embodiments, processor 601 may be implemented on a cloud platform. In one embodiment, the cloud platform may include a private cloud, public cloud, hybrid cloud, community cloud, distributed cloud, internal cloud, multi-cloud, etc., or any combination thereof.
[0099] In some embodiments, memory 602 may be an internal storage unit of monitoring device 600, such as a hard disk or memory of monitoring device 600. In other embodiments, memory 602 may also be an external storage device of monitoring device 600, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., provided on monitoring device 600.
[0100] Furthermore, the memory 602 may include both internal storage units of the monitoring device 600 and external storage devices. The memory 602 is used to store the application software and various types of data installed on the monitoring device 600.
[0101] In some embodiments, display 603 may be an LED display, a liquid crystal display, a touch-screen liquid crystal display, or an organic light-emitting diode (OLED) touchscreen. Display 603 is used to display information from monitoring device 600 and to display a user interface for visualization. Components 601-603 of monitoring device 600 communicate with each other via a system bus.
[0102] In one embodiment, when processor 601 executes the bridge cable-stayed bridge monitoring program in memory 602, the following steps can be implemented: Acquire monitoring images of the cable stays of the target bridge within the target time period. The monitoring images of the cable stays of the target bridge are obtained by cameras located at two different positions. Feature extraction and feature matching are performed on the preprocessed monitoring images of the cable-stayed bridge within the target time period. Based on the feature matching results of the preprocessed monitoring images of the cable-stayed bridge within the target time period and the calibration parameters between two cameras at different locations, a three-dimensional point cloud sequence corresponding to the cable-stayed bridge within the target time period is constructed. The average cable force and multiple natural frequencies of the cable stays of the target bridge within the target time period are determined based on the three-dimensional point cloud sequence corresponding to the cable stays of the target bridge within the target time period. Based on the multi-order natural frequencies of the stay cables of the target bridge within the target time period and the corresponding three-dimensional point cloud sequence of the stay cables of the target bridge within the target time period, the cable force inversion is performed to determine the inverted cable force of the stay cables of the target bridge within the target time period. Based on the average cable force and the inverted cable force of the stay cables of the target bridge within the target time period, the monitoring results of the stay cables of the target bridge within the target time period are determined.
[0103] It should be understood that when the processor 601 executes the bridge cable-stayed bridge monitoring program in the memory 602, in addition to the functions mentioned above, it can also perform other functions, as detailed in the description of the corresponding method embodiments above.
[0104] Furthermore, this embodiment of the invention does not specifically limit the type of monitoring device 600 mentioned. The monitoring device 600 can be a portable electronic device such as a mobile phone, tablet computer, personal digital assistant (PDA), wearable device, or laptop computer. Exemplary embodiments of portable electronic devices include, but are not limited to, portable electronic devices running iOS, Android, Microsoft, or other operating systems. The aforementioned portable electronic device can also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the invention, the monitoring device 600 may not be a portable electronic device, but rather a desktop computer with a touch-sensitive surface (e.g., a touch panel).
[0105] Accordingly, this application also provides a computer-readable storage medium for storing computer-readable programs or instructions. When the programs or instructions are executed by a processor, they can implement the steps or functions of the bridge cable monitoring methods provided in the above-described method embodiments.
[0106] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.), and the computer program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.
[0107] The monitoring method and device for bridge stay cables provided by the present invention have been described in detail above. Specific examples have been used to illustrate the principle and implementation of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core idea of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation and application scope based on the idea of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.
Claims
1. A method for monitoring bridge stay cables, characterized in that, include: Acquire monitoring images of the cable stays of the target bridge within the target time period. The monitoring images of the cable stays of the target bridge are obtained by cameras located at two different positions. Feature extraction and feature matching are performed on the preprocessed monitoring images of the cable-stayed bridge within the target time period. Based on the feature matching results of the preprocessed monitoring images of the cable-stayed bridge within the target time period and the calibration parameters between two cameras at different locations, a three-dimensional point cloud sequence corresponding to the cable-stayed bridge within the target time period is constructed. The average cable force and multiple natural frequencies of the cable stays of the target bridge within the target time period are determined based on the three-dimensional point cloud sequence corresponding to the cable stays of the target bridge within the target time period. Based on the multi-order natural frequencies of the stay cables of the target bridge within the target time period and the corresponding three-dimensional point cloud sequence of the stay cables of the target bridge within the target time period, the cable force inversion is performed to determine the inverted cable force of the stay cables of the target bridge within the target time period. Based on the average cable force and the inverted cable force of the stay cables of the target bridge within the target time period, the monitoring results of the stay cables of the target bridge within the target time period are determined.
2. The monitoring method for bridge stay cables according to claim 1, characterized in that, The process of extracting and matching features from the preprocessed monitoring images of the cable-stayed bridge within the target time period includes: Feature extraction is performed on the monitoring images of the cable-stayed bridge of the target bridge within the target time period based on the feature extraction network. The feature extraction network is trained on the SuperPoint network using a general image dataset and a sample image dataset of cable-stayed bridges as samples. The encoder of the SuperPoint network is equipped with a convolutional block attention module at the end. The graph neural network matching algorithm is used to match feature points at the same sampling time in the monitoring images of the cable-stayed bridge of the target bridge within the target time period, and the LK optical flow algorithm is used to track the temporal sequence of the feature points.
3. The monitoring method for bridge stay cables according to claim 1, characterized in that, Based on the feature matching results of the preprocessed monitoring images of the cable-stayed bridge within the target time period and the calibration parameters between two cameras at different locations, a three-dimensional point cloud sequence corresponding to the cable-stayed bridge within the target time period is constructed, including: Based on the calibration parameters between two cameras at different locations, the coordinates of each feature point in the three-dimensional space of the feature matching result of the cable-stayed monitoring image of the target bridge in the preprocessed target time period are determined, and the three-dimensional point cloud sequence corresponding to the cable-stayed cable of the target bridge in the target time period is constructed. The calibration parameters between the two cameras at different locations include camera parameters, lens distortion coefficients, and rotation matrix and translation vector between the two cameras at different locations.
4. The monitoring method for bridge stay cables according to claim 1, characterized in that, The determination of the average cable force and multiple natural frequencies of the stay cables of the target bridge within the target time period based on the three-dimensional point cloud sequence corresponding to the stay cables of the target bridge within the target time period includes: The static three-dimensional point cloud sequence is determined by averaging the three-dimensional point cloud sequence corresponding to the cable stays of the target bridge within the target time period in the time dimension. The static 3D point cloud sequence is projected onto a 2D plane defined by the two endpoints of the cable to determine the 2D cable shape point set, and the 2D cable shape point set is fitted to determine the cable shape curve; Based on the cable shape curve, determine the average cable force of the stay cables of the target bridge within the target time period; Modal analysis is performed based on the three-dimensional point cloud sequence corresponding to the stay cables of the target bridge within the target time period to determine the multiple natural frequencies of the stay cables of the target bridge within the target time period.
5. The monitoring method for bridge stay cables according to claim 4, characterized in that, The method involves performing cable force inversion based on the multi-order natural frequencies of the stay cables of the target bridge within the target time period and the corresponding three-dimensional point cloud sequence of the stay cables within the target time period, to determine the inverted cable forces of the stay cables of the target bridge within the target time period, including: A finite element model is constructed based on cable force, bending stiffness, and end boundary constraint stiffness. A multi-objective function is constructed that includes shape residuals and frequency residuals. The shape residuals are used to represent the root mean square error between the static three-dimensional point cloud sequence and the node positions of the finite element model. The frequency residuals are used to represent the sum of the relative errors between the multi-order natural frequencies of the cable stays of the target bridge and the multi-order natural frequencies of the cable stays of the target bridge determined by the finite element model during the target time period. The cable force that minimizes the multi-objective function corresponding to the finite element model is determined based on the multi-objective optimization algorithm, and is used as the inverse cable force of the cable-stayed cable of the target bridge within the target time period.
6. The monitoring method for bridge stay cables according to claim 1, characterized in that, The method for determining the monitoring results of the stay cables of the target bridge within the target time period, based on the average cable force and inverted cable force of the stay cables within the target time period, includes: If the relative error between the average cable force of the cable-stayed bridge in the target time period and the inverted cable force of the cable-stayed bridge in the target time period is less than or equal to the error threshold, the inverted cable force of the cable-stayed bridge in the target time period or the average of the average cable force and the inverted cable force of the cable-stayed bridge in the target time period shall be determined as the monitoring result of the cable-stayed bridge in the target time period. If the relative error between the average cable force of the stay cables of the target bridge within the target time period and the inverted cable force of the stay cables of the target bridge within the target time period is greater than the error threshold, the average cable force and inverted cable force of the stay cables of the target bridge within the target time period, as well as the uncertainty range corresponding to the average cable force and inverted cable force of the stay cables of the target bridge within the target time period, are determined as the monitoring results of the stay cables of the target bridge within the target time period.
7. The monitoring method for bridge stay cables according to any one of claims 1 to 6, characterized in that, The preprocessing of the cable-stayed bridge monitoring images within the target time period includes: The monitoring images of the cable-stayed bridge within the target time period are sequentially converted to grayscale, histogram equalization, and guided filtering.
8. A monitoring device for bridge stay cables, characterized in that, include: The acquisition module is used to acquire monitoring images of the cable stays of the target bridge within the target time period. The monitoring images of the cable stays of the target bridge are captured by two cameras located at different positions. The module is used to extract and match features from the preprocessed monitoring images of the cable-stayed bridge within the target time period, and to construct a three-dimensional point cloud sequence corresponding to the cable-stayed bridge within the target time period based on the feature matching results of the preprocessed monitoring images of the cable-stayed bridge within the target time period and the calibration parameters between two cameras at different locations. The first determining module is used to determine the average cable force and multiple natural frequencies of the cable-stayed bridge within the target time period based on the three-dimensional point cloud sequence corresponding to the cable-stayed bridge within the target time period. The second determining module is used to perform cable force inversion based on the multi-order natural frequencies of the cable stays of the target bridge within the target time period and the corresponding three-dimensional point cloud sequence of the cable stays of the target bridge within the target time period, to determine the inverted cable force of the cable stays of the target bridge within the target time period, and to determine the monitoring results of the cable stays of the target bridge within the target time period based on the average cable force and the inverted cable force of the cable stays of the target bridge within the target time period.
9. A monitoring device, characterized in that, Including memory and processor, among which, The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the bridge cable-stayed cable monitoring method according to any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that, Used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps in the bridge cable-stayed cable monitoring method according to any one of claims 1 to 7.