A method and system for monitoring the safety of a floating bridge at sea

Abstract

The present invention relates to the technical field of floating bridge safety monitoring, and discloses a method and system for monitoring the safety of a marine floating bridge. First, conventional simulation methods are used to determine stress thresholds, and this data is stored in a floating bridge response database; then, real-time images of the pontoon and the floating bridge are obtained, and by analyzing these images, the real-time displacement data between the pontoon and the floating bridge can be accurately obtained; when the real-time displacement data is greater than or equal to the displacement data pre-stored in the database, it indicates that a large relative displacement has occurred between the pontoon and the floating bridge. At this time, the system dynamically adjusts the original stress threshold according to various factors such as the current real-time displacement data, water flow velocity, and vehicle load, more accurately reflecting the stress condition in the actual situation, thereby improving the accuracy of safety assessment.

Description

Technical Field

[0001] The present invention relates to the technical field of floating bridge safety monitoring, and particularly to a method and system for monitoring the safety of a marine floating bridge. Background Art

[0002] When conducting a safety assessment of the floating bridge structure, it is particularly important to obtain the stress distribution of the pontoon and the floating bridge under different displacement conditions during the driving process of various vehicles through various simulation methods. Such a simulation method can not only help engineers understand the response behavior of the floating bridge structure under complex working conditions, but also provide a scientific basis for design optimization. Specifically, different types of vehicle loads and their dynamic effects on the floating bridge structure are considered during the simulation process, which helps to identify potential weak links, thereby taking corresponding reinforcement measures or adjusting the design scheme. In addition, through the quantitative analysis of stress thresholds under different working conditions, it can be ensured that the floating bridge can withstand the expected traffic flow and environmental load during actual use, and guarantee the safety of pedestrians and vehicle passage.

[0003] However, when the integrity of the floating bridge structure is challenged, such as a large relative displacement between the pontoon and the floating bridge due to the action of water flow, the stress thresholds obtained through the above simulation may no longer be applicable. This is because the pontoon, as one of the important support points of the floating bridge, the change in its position will directly affect the stability and bearing capacity of the entire structure. For example, when the river flow velocity increases, the pontoon may move due to the impact of the water flow, resulting in deviations in the distance and angle between the pontoon and the floating bridge. These changes will not only change the actual stress state of the floating bridge, but may also cause a significant change in the original stress distribution pattern, thereby affecting the effectiveness of the previously determined safety threshold.

[0004] Therefore, there is an urgent need for a method and system for monitoring the safety of a marine floating bridge, which can obtain stress thresholds with higher accuracy, be used to evaluate the safety of the floating bridge under large displacement conditions, and formulate more reasonable maintenance and management strategies. Summary of the Invention

[0005] In order to solve the above technical problems, on the one hand, the present invention provides a method for monitoring the safety of a marine floating bridge, including the following steps:

[0006] S1: Obtain the stress thresholds of different vehicles on different positions of the floating bridge when the pontoon and the floating bridge are under different displacement conditions through the finite element simulation method, and construct a floating bridge response database;

[0007] S2: Obtain the real-time images of the pontoon and the floating bridge, and analyze the real-time displacement data of the pontoon and the floating bridge from the real-time images, where the real-time displacement data includes the distance deviation and angle deviation between the pontoon and the floating bridge;

[0008] S3: Compare the real-time displacement data with the displacement data in the pontoon bridge response database to determine the final stress threshold:

[0009] If the real-time displacement data is less than the displacement data in the pontoon bridge response database, use the stress threshold in the pontoon bridge response database as the final stress threshold;

[0010] If the real-time displacement data is greater than or equal to the displacement data in the pontoon bridge response database, correct the stress threshold in the pontoon bridge response database according to the real-time displacement data, water flow velocity, and vehicle load to obtain the final stress threshold;

[0011] S4: Monitor the safety of the offshore pontoon bridge with the final stress threshold.

[0012] Preferably, correcting the stress threshold in the pontoon bridge response database according to the real-time displacement data, water flow velocity, and vehicle load to obtain the final stress threshold specifically includes:

[0013] S31: Construct a stress threshold correction model, use the structural characteristic parameters of the pontoon bridge as the basic variables, the real-time displacement data as the correction variable, and the vehicle load and water flow velocity as the reference variables to establish a stress threshold correction model; the input of the stress threshold correction model includes the structural characteristic parameters of the pontoon bridge, real-time displacement data, water flow velocity, and vehicle load, and the output is the final stress threshold; the structural characteristic parameters of the pontoon bridge are used to represent the structural characteristics of the pontoon bridge, including the pontoon bridge length;

[0014] S32: Determine the correction coefficient, perform fitting according to the finite element simulation data and actual test data to determine the correction coefficient in the stress threshold correction model;

[0015] S33: Correct the stress threshold in the pontoon bridge response database according to the structural characteristic parameters of the pontoon bridge, real-time displacement data, water flow velocity, and vehicle load through the stress threshold correction model to obtain the final stress threshold.

[0016] Preferably, the final stress threshold is calculated using the following formula:

[0017]

[0018] where, σ corrected represents the corrected final stress threshold, σ original represents the stress threshold in the pontoon bridge response database, is the distance deviation in the real-time displacement data, is the angle deviation in the real-time displacement data, L represents the pontoon bridge length in the structural characteristic parameters, θ represents the angle value between the pontoon and the pontoon bridge, α, β, and γ are correction coefficients, P represents the vehicle load, P maxrepresents the maximum vehicle load, V represents the water flow velocity, and V max represents the maximum water flow velocity.

[0019] Preferably, the real-time displacement data of the pontoon and the floating bridge are analyzed from the real-time image, specifically including:

[0020] S21: For the real-time image, use the MOG2 algorithm for denoising;

[0021] S22: For the real-time image after denoising, use EfficientDet to determine the bounding box coordinates of the floating bridge and the pontoon in the real-time image; and intercept the floating bridge and pontoon area images according to the bounding box coordinates;

[0022] S23: Based on the floating bridge and pontoon area images, use the Mask R-CNN model to separate from the real-time image to obtain the pixel-level contour masks of the floating bridge and the pontoon;

[0023] S24: According to the pixel-level contour masks of the floating bridge and the pontoon, determine the final images of the floating bridge and the pontoon through a transfer learning model;

[0024] S25: Based on the selected reference object, determine the real-time displacement data of the pontoon and the floating bridge in the final images of the floating bridge and the pontoon.

[0025] Preferably, in S25, based on the selected reference object, determine the real-time displacement data of the pontoon and the floating bridge in the final images of the floating bridge and the pontoon, specifically:

[0026] S251: Obtain the actual size of the selected reference object and the pixel spacing in the final images of the floating bridge and the pontoon respectively;

[0027] S252: Calculate the scale factor based on the actual size and the pixel spacing;

[0028] S253: Obtain the coordinates of the pontoon and the floating bridge in the final images of the floating bridge and the pontoon respectively;

[0029] S254: Based on the scale factor, determine the real-time displacement data of the pontoon and the floating bridge according to the coordinates of the pontoon and the floating bridge in the final images of the floating bridge and the pontoon.

[0030] Preferably, the calculation formula for the real-time displacement data of the pontoon and the floating bridge is:

[0031] ;

[0032] ;

[0033] Wherein, is the distance deviation in the real-time displacement data, is the angle deviation in the real-time displacement data, k is the scale factor, x 趸船 is the x-axis coordinate of the barge in the final image of the pontoon bridge and the barge, y 趸船 is the y-axis coordinate of the barge in the final image of the pontoon bridge and the barge, x 浮桥 is the x-axis coordinate of the pontoon bridge in the final image of the pontoon bridge and the barge, y 浮桥 is the y-axis coordinate of the pontoon bridge in the final image of the pontoon bridge and the barge.

[0034] Preferably, the scale factor k is calculated using the following formula:

[0035] ;

[0036] wherein, L real represents the actual size of the selected reference object, L pixel represents the pixel pitch of the selected reference object in the final image of the pontoon bridge and the barge.

[0037] Preferably, after denoising processing of S21, it further includes performing distortion correction on the lens of the image acquisition device.

[0038] On the other hand, the present application also provides a safety monitoring system for a marine pontoon bridge, which implements a safety monitoring method for a marine pontoon bridge described in any one of the above, and the system includes:

[0039] A pontoon bridge response database construction module, which is used to obtain the stress thresholds at different positions of the pontoon bridge when different vehicles travel on the pontoon bridge under different displacement conditions of the barge and the pontoon bridge through the finite element simulation method, and construct a pontoon bridge response database;

[0040] A real-time displacement data acquisition module, which is used to acquire real-time images of the barge and the pontoon bridge, and analyze the real-time displacement data of the barge and the pontoon bridge from the real-time images, and the real-time displacement data includes the distance deviation and angle deviation of the barge and the pontoon bridge;

[0041] A final stress threshold acquisition module, which is used to compare the real-time displacement data with the displacement data in the pontoon bridge response database:

[0042] If the real-time displacement data is less than the displacement data in the pontoon bridge response database, the stress threshold in the pontoon bridge response database is used as the final stress threshold;

[0043] If the real-time displacement data is greater than or equal to the displacement data in the pontoon bridge response database, the stress threshold in the pontoon bridge response database is corrected according to the real-time displacement data, water flow velocity, and vehicle load to obtain the final stress threshold;

[0044] A safety monitoring module is used to monitor the safety of the offshore floating bridge with the final stress threshold.

[0045] The embodiments of the present invention have the following technical effects:

[0046] In a method for monitoring the safety of an offshore floating bridge in this application, first, conventional simulation methods are used to determine the stress threshold, and this data is stored in the floating bridge response database; then, real-time images of the pontoon and the floating bridge are obtained, and by analyzing these images, the real-time displacement data between the pontoon and the floating bridge can be accurately obtained; when the real-time displacement data is greater than or equal to the displacement data pre-stored in the database, it indicates that a large relative displacement has occurred between the pontoon and the floating bridge. At this time, the system will dynamically adjust the original stress threshold according to various factors such as the current real-time displacement data, water flow velocity, and vehicle load, more accurately reflecting the stress condition in the actual situation, thereby improving the accuracy of safety assessment.

[0047] Through the above process, especially the dynamic correction of the stress threshold, this technical solution effectively reduces the errors that may be caused by the large relative displacement between the pontoon and the floating bridge. This method not only considers the changes in the structure itself but also comprehensively takes into account the influence of the external environment (such as water flow velocity) and usage conditions (such as vehicle load), thereby ensuring that the finally obtained stress threshold is closer to the actual working conditions and greatly improving the safety of the offshore floating bridge. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0049] Figure 1 is a flowchart of the steps of a method for monitoring the safety of an offshore floating bridge provided by an embodiment of the present invention;

[0050] Figure 2 is a flowchart of the calculation steps of the final stress threshold provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0051] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

[0052] Traditional floating bridge safety monitoring methods rely on static simulation data and are difficult to cope with dynamic interferences such as barge displacement and water flow impact in the actual environment. The present invention captures structural changes through real-time image analysis, combines external factors such as water flow velocity and vehicle load, dynamically corrects the stress threshold, and realizes the transformation from "static simulation" to "dynamic adaptation", taking into account both the structural mechanics characteristics and the dynamic environmental impacts. By constructing a "simulation - perception - correction" closed-loop system, the offline simulation data and the online perception data are fused to ensure that the stress threshold always fits the actual working conditions.

[0053] Figure 1 It is a flowchart of the steps of a method for monitoring the safety of a sea floating bridge provided by an embodiment of the present invention. As Figure 1 shown, a method for monitoring the safety of a sea floating bridge provided by the present application includes the following steps:

[0054] S1: Obtain the stress thresholds of different vehicles traveling on the floating bridge at different positions of the floating bridge under different displacement conditions of the barge and the floating bridge through the finite element simulation method, and construct a floating bridge response database;

[0055] S2: Obtain the real-time images of the barge and the floating bridge, and analyze the real-time displacement data of the barge and the floating bridge from the real-time images. The real-time displacement data includes the distance deviation and the angle deviation between the barge and the floating bridge, which are used to reflect the structural dynamic changes of the barge and the floating bridge;

[0056] In some embodiments, analyzing the real-time displacement data of the barge and the floating bridge from the real-time images specifically includes:

[0057] S21: For the real-time images, perform denoising processing using the MOG2 algorithm;

[0058] The MOG2 algorithm, that is, the Gaussian Mixture Model Separation Algorithm (Mixture of Gaussians 2), is a background segmentation algorithm based on the Gaussian Mixture Model (Gaussian Mixture Model, GMM). In special application scenarios involving barges and floating bridges, this algorithm can accurately separate the target object from the displacement image. By using the Gaussian Mixture Model to model the probability distribution of pixel points, the MOG2 algorithm can effectively distinguish the pixel changes caused by background changes (such as water surface fluctuations, light changes, etc.) from the real displacements of the target object. With its strong adaptability and robustness, it demonstrates excellent performance in processing image separation tasks in complex dynamic scenarios, providing strong technical support for the safety monitoring of sea floating bridges.

[0059] In some embodiments, after the denoising process in S21, it further includes performing distortion correction on the lens of the image acquisition device. Lens distortion can cause the captured image to be deformed, affecting the actual shape, size, and relative position relationship between objects, which is particularly disadvantageous for applications relying on image analysis. Therefore, performing distortion correction on the lens of the image acquisition device can improve the quality and accuracy of the image, ensuring that the information extracted from the image is true and reliable.

[0060] S22: For the real-time image after denoising processing, use EfficientDet to determine the bounding box coordinates of the floating bridge and the pontoon in the real-time image; and intercept the floating bridge and pontoon area images according to the bounding box coordinates.

[0061] Exemplarily, the EfficientDet model is an efficient object detection model based on deep learning, integrating the powerful feature extraction ability of EfficientNet and various object detection strategies. When locating the bounding boxes of the floating bridge and the pontoon in the real-time image based on this model, the following specific steps can be included:

[0062] S221: Model input and preprocessing

[0063] Input image: Obtain the real-time image of the pontoon and the floating bridge, with a resolution recommended to be ≥1920×1080 to ensure clear details.

[0064] Preprocessing process:

[0065] Optical flow method assistance: Use the optical flow method to estimate the motion trend of objects in the image sequence, initially screen out the areas that may contain the floating bridge and the pontoon, narrow the detection range, and reduce the computational load of the EfficientDet model.

[0066] Image enhancement: Use data enhancement techniques such as random cropping, color jittering, and rotation to simulate various lighting and angle changes, and improve the robustness of the model.

[0067] Size adjustment: Uniformly adjust the image size to the input size of the EfficientDet model (such as 512×512 or 640×640), and maintain the aspect ratio to avoid object deformation.

[0068] S222: Model architecture optimization

[0069] Feature fusion: Introduce a bidirectional Feature Pyramid Network (BiFPN) to fuse information between feature maps of different scales multiple times, and strengthen the capture of different-size details of the floating bridge and the pontoon.

[0070] Anchor box design: Customize the size of the anchor box according to the common length-width ratios of floating bridges and pontoons to improve the matching degree between the anchor box and the target. For example, for a long-strip floating bridge, set anchor boxes with length-width ratios of 3:1 or 5:1; for approximately square objects such as pontoons, use 1:1 anchor boxes.

[0071] Loss function adjustment: Combine the Intersection over Union (IoU) loss and the confidence loss to guide the model to accurately locate the bounding box. Introduce a dynamic trade-off parameter to automatically adjust the proportion of the classification loss and the regression loss to address the problem of uneven target categories in complex backgrounds.

[0072] S223: Post-processing strategy

[0073] Improvement of NMS: Use Soft-NMS or Diou-NMS to replace the traditional NMS, allow bounding boxes with appropriate overlaps to exist, reduce the probability of misdeletion, and smoothly filter the final boxes according to the confidence.

[0074] False alarm suppression: Set the bounding box confidence threshold ≥0.8, and at the same time combine context information (such as whether the relative positions of the floating bridge and the pontoon conform to physical logic) to eliminate misdetected targets.

[0075] S23: Based on the images of the floating bridge and pontoon areas, use the Mask R-CNN model to separate from the real-time image to obtain the pixel-level contour masks of the floating bridge and the pontoon;

[0076] Exemplarily, S23 is specifically as follows:

[0077] S231: Model input and preprocessing

[0078] Input image: A real-time image with the bounding boxes of the floating bridge and the pontoon located.

[0079] Preprocessing:

[0080] Semantic segmentation mask: First, use a pre-trained semantic segmentation model (such as DeepLab) to generate a rough mask of the floating bridge and the pontoon, providing prior information for Mask R-CNN.

[0081] Edge enhancement: Apply the Canny operator or the Hough transform to detect the image edges, enhance the contours of the floating bridge and the pontoon, and assist Mask R-CNN in extracting the pixel-level boundaries.

[0082] S232: Model architecture optimization

[0083] Expansion of the feature pyramid: Add deeper-level feature pyramid layers to improve the description ability for small-size details (such as the connection parts of the floating bridge and the edge gaps of the pontoon).

[0084] Attention mechanism: Incorporate SENet or CBAM attention modules to focus on the key areas of the floating bridge and pontoon, weaken background interference, and improve the rationality of feature weight distribution.

[0085] Mask generation: Adopt a hierarchical mask generation strategy. First, generate the overall mask of the target, and then generate refined masks for key areas (such as the main beam of the floating bridge and the mooring point of the pontoon) to enhance the contour accuracy.

[0086] S233: Post-processing strategy

[0087] Contour refinement: Use mathematical morphology operations (such as opening and closing operations, skeletonization) to denoise and smooth the extracted pixel-level contours, and repair contour defects caused by light and shadow changes or image noise.

[0088] Multi-view correction: Introduce multi-view geometry methods. According to the masks generated from images of different camera perspectives, correct the contour distortion caused by perspective differences to ensure contour consistency in three-dimensional space.

[0089] S24: According to the pixel-level contour masks of the floating bridge and pontoon, determine the final images of the floating bridge and pontoon through a transfer learning model;

[0090] S25: Based on the selected reference objects, determine the real-time displacement data of the pontoon and the floating bridge in the final images of the floating bridge and pontoon.

[0091] Exemplarily, select the fixed markers at both ends of the floating bridge (such as mooring bitts, bridge piers) as reference objects to ensure their stable size, easy recognition, and clear visibility in the image. Traditional calibration requires additional installation of markers, which is costly and easily damaged. The present invention uses the inherent structure of the floating bridge as a reference object, which not only reduces costs but also avoids human intervention.

[0092] Displacement calculation needs to solve the problem of image scale distortion. Establish a pixel-physical mapping relationship through reference object calibration to eliminate lens distortion and perspective error and ensure data reliability. Therefore, in some embodiments, in S25, based on the selected reference objects, determine the real-time displacement data of the pontoon and the floating bridge in the final images of the floating bridge and pontoon, specifically:

[0093] S251: Respectively obtain the actual size of the selected reference object and the pixel spacing in the final images of the floating bridge and pontoon;

[0094] S252: Calculate the scale factor based on the actual size and the pixel spacing;

[0095] S253: Respectively obtain the coordinates of the pontoon and the floating bridge in the final images of the floating bridge and pontoon;

[0096] S254: Based on the scale factor, determine the real-time displacement data of the pontoon and the floating bridge according to the coordinates of the pontoon and the floating bridge in the final image of the floating bridge and the pontoon.

[0097] In some embodiments, the calculation formula for the real-time displacement data of the pontoon and the floating bridge is:

[0098] ;

[0099] ;

[0100] Wherein, is the distance deviation in the real-time displacement data, is the angle deviation in the real-time displacement data, k is the scale factor, x 趸船 is the x-axis coordinate of the pontoon in the final image of the floating bridge and the pontoon, y 趸船 is the y-axis coordinate of the pontoon in the final image of the floating bridge and the pontoon, x 浮桥 is the x-axis coordinate of the floating bridge in the final image of the floating bridge and the pontoon, y 浮桥 is the y-axis coordinate of the floating bridge in the final image of the floating bridge and the pontoon. The positions of the centers of the outlines of the floating bridge and the pontoon in the image, the above x and y coordinates are extracted by the segmentation mask. Take the geometric center of the contour mask to reduce the error caused by local deformation.

[0101] In some embodiments, the scale factor k is calculated using the following formula:

[0102] ;

[0103] Wherein, L real represents the actual size of the selected reference object, obtained through engineering drawings or on-site measurements; L pixel represents the pixel spacing of the selected reference object in the final image of the floating bridge and the pontoon, extracted by image analysis. The scale factor k is the conversion coefficient between the pixel distance and the actual distance, calibrated by the reference object;

[0104] MOG2 algorithm: A background separation algorithm based on a mixture Gaussian model, used to eliminate dynamic interferences such as water surface reflection and waves. EfficientDet: An efficient object detection model that balances computing resources and detection accuracy through compound scaling. MaskR-CNN: An instance segmentation model that can output the object pixel-level contour to improve the displacement measurement accuracy.

[0105] The extraction of displacement data needs to solve the problem of image interference in a complex water environment. Through a four-level pipeline of "denoising - positioning - segmentation - calibration", the data accuracy is gradually improved:

[0106] Image preprocessing: Use a polarization filter to eliminate the water surface reflection and reduce image noise; apply the MOG2 algorithm to separate the dynamic foreground (floating bridge, floating dock) from the static background (water surface, sky). The dynamic background separation filters out the interference of water surface fluctuations.

[0107] Target localization: Input the preprocessed image into the EfficientDet model to output the bounding box coordinates of the floating bridge and the floating dock; optimize the anchor box design (aspect ratio of the floating bridge is 3:1, and that of the floating dock is 1:1) to improve the localization efficiency. Target detection locates the areas of the floating bridge and the floating dock.

[0108] Contour segmentation: Crop the target area based on the bounding box and input it into the Mask R-CNN model to generate a pixel-level mask; repair contour breaks or noise through morphological operations (dilation, erosion). Instance segmentation extracts the precise contour.

[0109] Displacement calculation: Select fixed reference objects (such as mooring bollards) at both ends of the floating bridge to calculate the scale factor between pixels and the actual distance; calculate the distance deviation (Δd) and the angle deviation (Δθ) based on the center coordinates of the mask. Geometric calibration maps the pixel displacement to physical quantities.

[0110] By combining computer vision and deep learning, automated and non-contact displacement monitoring is achieved, which is especially suitable for harsh water environments.

[0111] S3: Compare the real-time displacement data with the displacement data in the floating bridge response database to determine the final stress threshold:

[0112] If the real-time displacement data is less than the displacement data in the floating bridge response database, use the stress threshold in the floating bridge response database as the final stress threshold;

[0113] If the real-time displacement data is greater than or equal to the displacement data in the floating bridge response database, correct the stress threshold in the floating bridge response database according to the real-time displacement data, water flow velocity, and vehicle load to obtain the final stress threshold;

[0114] In some embodiments, the stress threshold in the floating bridge response database is corrected according to the real-time displacement data, water flow velocity, and vehicle load to obtain the final stress threshold. See Figure 2 and specifically includes:

[0115] S31: Build a stress threshold correction model. Based on the structural characteristic parameters of the floating bridge as the basic variables, the real-time displacement data as the correction variable, and the vehicle load and water flow velocity as the reference variables, establish a stress threshold correction model; the input of the stress threshold correction model includes the structural characteristic parameters of the floating bridge, real-time displacement data, water flow velocity, and vehicle load, and the output is the final stress threshold; the structural characteristic parameters of the floating bridge are used to represent the structural characteristics of the floating bridge, including the floating bridge length.

[0116] S32: Determine the correction coefficient. Based on the finite element simulation data and the actual test data, perform fitting to determine the correction coefficient in the stress threshold correction model.

[0117] Exemplarily, generate stress thresholds under different working conditions based on the finite element simulation data; obtain the real stress data through actual tests (such as hydraulic laboratory loading tests); use the least squares method for fitting to minimize the error between the model prediction value and the measured value.

[0118] S33: According to the structural characteristic parameters of the floating bridge, the real-time displacement data, the water flow velocity, and the vehicle load, correct the stress threshold in the floating bridge response database through the stress threshold correction model to obtain the final stress threshold.

[0119] In some embodiments, the final stress threshold is calculated using the following formula:

[0120]

[0121] where, σ corrected represents the corrected final stress threshold, which is the safety stress value dynamically adjusted according to the real-time displacement, water flow velocity, and vehicle load; σ original represents the stress threshold in the floating bridge response database, which is the static stress value pre-determined through finite element simulation; is the distance deviation in the real-time displacement data, reflecting the horizontal displacement of the pontoon relative to the floating bridge (unit: meter); L represents the length of the floating bridge in the structural characteristic parameters (unit: meter), characterizing the influence of the overall size of the floating bridge on the stress distribution; is the angle deviation in the real-time displacement data (unit: radian or degree), reflecting the inclination degree of the pontoon and the floating bridge; θ represents the angle value between the pontoon and the floating bridge, usually the angle in the design or initial state (unit: radian or degree); α, β, and γ are correction coefficients obtained by fitting the finite element simulation data and the actual test data, used to adjust the weights of different parameters on the final stress threshold; P represents the vehicle load (unit: kilonewton or ton), representing the real-time load of the vehicle on the floating bridge; P max represents the maximum vehicle load (unit: kilonewton or ton), determined according to the design bearing capacity of the floating bridge; V represents the water flow velocity (unit: meter per second), reflecting the influence of the water flow dynamics in the water area where the floating bridge is located; V max represents the maximum water flow velocity (unit: meter per second), which is the threshold set based on the anti-flow ability of the floating bridge structure.

[0122] The length L of the floating bridge is a key structural parameter. A longer floating bridge may generate a greater bending moment under the same displacement, thereby affecting the stress distribution. When the real-time displacement data (Δd and Δθ) exceeds the preset threshold, it means that there is a large change in the connection between the floating bridge and the pontoon, which may lead to a redistribution of the stress inside the structure. For example, a large distance deviation Δd may cause tension or compression of the floating bridge, while an angular deviation Δθ may lead to an increase in torsional or shear stress.

[0123] Next, the water flow velocity V affects the hydrodynamic load of the floating bridge. A higher flow velocity will increase the lateral force on the floating bridge, which may cause structural vibration or displacement, and thus affect the stress. The vehicle load P is a dynamic load, and different vehicle weights and distributions will directly change the local stress of the floating bridge. When the vehicle load approaches the maximum vehicle load P max the original stress threshold may be insufficient to cope with the actual working conditions and needs to be dynamically adjusted.

[0124] The correction terms in the formula take these factors into account: α combines the relative effects of the distance deviation and the vehicle load, β correlates the angular deviation and the water flow velocity, and γ deals with the cross effects of the two. The correction coefficients α, β, and γ are determined by fitting finite element simulations and actual data to ensure the theoretical reliability and practicality of the model.

[0125] Therefore, by integrating structural parameters, real-time displacement, environmental load, and dynamic load, the correction model can dynamically reflect the complex stress conditions under actual working conditions. When the real-time displacement data is greater than or equal to the displacement data in the floating bridge response database, the system automatically adjusts the stress threshold, avoiding the deficiencies of traditional static models, ensuring the accuracy and timeliness of safety assessment, and thus significantly improving the safety of the floating bridge.

[0126] This formula dynamically corrects the original stress threshold (σ max ) by combining the real-time displacement deviation (Δd, Δθ), the vehicle load (P / PmaxP), and the water flow velocity (V / V original ), so that the corrected stress threshold (σ corrected ) is more in line with the actual working conditions, thereby improving the accuracy of the safety monitoring of the floating bridge.

[0127] Correction coefficient α: Quantify the coupling effect of the distance deviation and the vehicle load on the stress. Correction coefficient β: Quantify the coupling effect of the angular deviation and the water flow velocity on the stress. Correction coefficient γ: Synthesize the interaction between the distance deviation and the angular deviation to capture the non-linear effect under complex displacements.

[0128] S4: Monitor the safety of the offshore floating bridge with the final stress threshold.

[0129] On the other hand, this application also provides a safety monitoring system for a marine floating bridge, which implements the safety monitoring method for a marine floating bridge described in any one of the above. The system includes:

[0130] A floating bridge response database construction module, which is used to obtain the stress thresholds at different positions of the floating bridge when different vehicles travel on the floating bridge under different displacement conditions of the pontoon and the floating bridge through the finite element simulation method, and construct a floating bridge response database;

[0131] A real-time displacement data acquisition module, which is used to acquire real-time images of the pontoon and the floating bridge, and analyze the real-time displacement data of the pontoon and the floating bridge from the real-time images. The real-time displacement data includes the distance deviation and the angle deviation between the pontoon and the floating bridge;

[0132] A final stress threshold acquisition module, which is used to compare the real-time displacement data with the displacement data in the floating bridge response database:

[0133] If the real-time displacement data is less than the displacement data in the floating bridge response database, the stress threshold in the floating bridge response database is used as the final stress threshold;

[0134] If the real-time displacement data is greater than or equal to the displacement data in the floating bridge response database, the stress threshold in the floating bridge response database is corrected according to the real-time displacement data, the water flow velocity, and the vehicle load to obtain the final stress threshold;

[0135] A safety monitoring module, which is used to monitor the safety of the marine floating bridge with the final stress threshold.

[0136] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the technical solutions of the embodiments of the present invention.

Claims

1. A method for monitoring the safety of a floating bridge at sea, characterized in that, It includes the following steps: S1: Obtain the stress thresholds at different positions of the pontoon bridge when different vehicles are traveling on the pontoon bridge under different displacement conditions of the barge and the pontoon bridge through the finite element simulation method, and construct a pontoon bridge response database; S2: Obtain the real-time images of the barge and the pontoon bridge, and analyze the real-time displacement data of the barge and the pontoon bridge from the real-time images. The real-time displacement data includes the distance deviation and the angle deviation between the barge and the pontoon bridge; S3: Compare the real-time displacement data with the displacement data in the pontoon bridge response database to determine the final stress threshold: If the real-time displacement data is less than the displacement data in the pontoon bridge response database, use the stress threshold in the pontoon bridge response database as the final stress threshold; If the real-time displacement data is greater than or equal to the displacement data in the pontoon bridge response database, correct the stress threshold in the pontoon bridge response database according to the real-time displacement data, the water flow velocity, and the vehicle load to obtain the final stress threshold; Specifically, it includes: S31: Construct a stress threshold correction model. Based on the structural characteristic parameters of the pontoon bridge as the basic variables, the real-time displacement data as the correction variable, and the vehicle load and the water flow velocity as the reference variables, establish a stress threshold correction model; The input of the stress threshold correction model includes the structural characteristic parameters of the pontoon bridge, the real-time displacement data, the water flow velocity, and the vehicle load, and the output is the final stress threshold; The structural characteristic parameters of the pontoon bridge are used to represent the structural characteristics of the pontoon bridge, including the pontoon bridge length; S32: Determine the correction coefficient. According to the finite element simulation data and the actual test data, perform fitting to determine the correction coefficient in the stress threshold correction model; S33: According to the structural characteristic parameters of the pontoon bridge, the real-time displacement data, the water flow velocity, and the vehicle load, correct the stress threshold in the pontoon bridge response database through the stress threshold correction model to obtain the final stress threshold; The final stress threshold is calculated using the following formula: ； Among them, σ corrected represents the final stress threshold after correction, and σ original represents the stress threshold in the pontoon bridge response database. is the distance deviation in the real-time displacement data, is the angle deviation in the real-time displacement data. L represents the pontoon bridge length in the structural characteristic parameters, θ represents the angle value between the barge and the pontoon bridge, α, β, and γ are correction coefficients, P represents the vehicle load, and P max represents the maximum vehicle load, V represents the water flow velocity, and V max represents the maximum water flow velocity. S4: Use the final stress threshold to monitor the safety of the offshore pontoon bridge.

2. The safety monitoring method of an offshore floating bridge according to claim 1, characterized in that Analyze the real-time displacement data of the barge and the pontoon bridge from the real-time images. Specifically, it includes: S21: For the real-time images, use the MOG2 algorithm for denoising; S22: For the denoised real-time images, use EfficientDet to determine the bounding box coordinates of the pontoon bridge and the barge in the real-time images; and intercept the pontoon bridge and the barge area images according to the bounding box coordinates; S23: Based on the pontoon bridge and the barge area images, use the Mask R-CNN model to separate from the real-time images to obtain the pixel-level contour masks of the pontoon bridge and the barge; S24: According to the pixel-level contour masks of the pontoon bridge and the barge, determine the final images of the pontoon bridge and the barge through the transfer learning model; S25: Based on the selected reference object, determine the real-time displacement data between the barge and the pontoon bridge in the final images of the pontoon bridge and the barge.

3. The safety monitoring method of an offshore floating bridge according to claim 2, characterized in that, In S25, based on the selected reference object, determine the real-time displacement data between the barge and the pontoon bridge in the final images of the pontoon bridge and the barge. Specifically, it is: S251: Respectively obtain the actual size of the selected reference object and the pixel spacing in the final images of the pontoon bridge and the barge; S252: Calculate a scale factor based on the actual size and the pixel pitch; S253: Obtain the coordinates of the pontoon and the floating bridge in the final images of the floating bridge and the pontoon respectively; S254: Based on the scale factor, determine the real-time displacement data of the pontoon and the floating bridge according to the coordinates of the pontoon and the floating bridge in the final images of the floating bridge and the pontoon.

4. The safety monitoring method of an offshore floating bridge according to claim 3, characterized in that, The calculation formula for the real-time displacement data of the pontoon and the floating bridge is: ； ； Among them, is the distance deviation in the real-time displacement data, is the angle deviation in the real-time displacement data, k is the scale factor, x 趸船 is the x-axis coordinate of the barge in the final image of the floating bridge and the barge, y 趸船 is the y-axis coordinate of the barge in the final image of the floating bridge and the barge, x 浮桥 is the x-axis coordinate of the floating bridge in the final image of the floating bridge and the barge, y 浮桥 is the y-axis coordinate of the floating bridge in the final image of the floating bridge and the barge.

5. The method for monitoring the safety of an offshore floating bridge according to claim 3, characterized in that, The scale factor k is calculated using the following formula: ； Among them, L real represents the actual size of the selected reference object, and L pixel represents the pixel pitch of the selected reference object in the final images of the pontoon bridge and the floating dock.

6. The safety monitoring method of an offshore floating bridge according to claim 2, characterized in that, After the denoising process in S21, it further includes performing distortion correction on the lens of the image acquisition device.

7. A safety monitoring system for a floating bridge at sea, which implements the safety monitoring method for a floating bridge at sea according to any one of claims 1-6, characterized in that, The system includes: A floating bridge response database construction module, which is used to obtain the stress thresholds at different positions of the floating bridge when different vehicles travel on the floating bridge under different displacement conditions of the pontoon and the floating bridge through the finite element simulation method, and construct a floating bridge response database; A real-time displacement data acquisition module, which is used to acquire the real-time images of the pontoon and the floating bridge, and analyze the real-time displacement data of the pontoon and the floating bridge from the real-time images. The real-time displacement data includes the distance deviation and the angle deviation between the pontoon and the floating bridge; A final stress threshold acquisition module, which is used to compare the real-time displacement data with the displacement data in the floating bridge response database: If the real-time displacement data is less than the displacement data in the floating bridge response database, use the stress threshold in the floating bridge response database as the final stress threshold; If the real-time displacement data is greater than or equal to the displacement data in the floating bridge response database, correct the stress threshold in the floating bridge response database according to the real-time displacement data, the water flow velocity, and the vehicle load to obtain the final stress threshold; Specifically, it includes: S31: Construct a stress threshold correction model. Based on the structural characteristic parameters of the floating bridge as the basic variables, the real-time displacement data as the correction variables, and the vehicle load and the water flow velocity as the reference variables, establish a stress threshold correction model; the input of the stress threshold correction model includes the structural characteristic parameters of the floating bridge, the real-time displacement data, the water flow velocity, and the vehicle load, and the output is the final stress threshold; the structural characteristic parameters of the floating bridge are used to represent the structural characteristics of the floating bridge, including the length of the floating bridge; S32: Determine the correction coefficient. Through fitting based on the finite element simulation data and the actual test data, determine the correction coefficient in the stress threshold correction model; S33: According to the structural characteristic parameters of the floating bridge, the real-time displacement data, the water flow velocity, and the vehicle load, correct the stress threshold in the floating bridge response database through the stress threshold correction model to obtain the final stress threshold; The final stress threshold is calculated using the following formula: ； Among them, σ corrected represents the final stress threshold after correction, and σ original represents the stress threshold in the pontoon bridge response database. is the distance deviation in the real-time displacement data, is the angle deviation in the real-time displacement data. L represents the pontoon bridge length in the structural characteristic parameters, θ represents the angle value between the pontoon and the pontoon bridge, α, β, and γ are correction coefficients, P represents the vehicle load, and P max represents the maximum vehicle load, V represents the water flow velocity, and V max represents the maximum water flow velocity. A safety monitoring module, which is used to monitor the safety of the offshore floating bridge with the final stress threshold.

Images