A method and system for monitoring and identifying damage of light-weight small and medium span bridges

Abstract

The application discloses a kind of small and medium-sized bridge light weight monitoring and damage identification method and system, belong to bridge structure health monitoring technical field.The application determines targeted monitoring index and general monitoring index according to bridge structure type;Then online visual perception equipment with MCU intelligent gateway of integrated edge computing function is used to collect monitoring data, and local preliminary processing and storage are carried out;MCU intelligent gateway carries out edge computing according to data time delay demand, and exports preliminary monitoring result after real-time data, and historical data is transmitted to cloud for depth analysis;Based on all the stored monitoring data, combined with damage identification model, bridge disease identification and degree evaluation are completed;According to the evaluation result, trigger early warning and push to maintenance personnel terminal.The application realizes low-cost, low-power, high-precision whole-process monitoring and maintenance through the division of labor of edge computing, depth analysis and damage identification three kinds of data processing methods, and adapts to the actual application needs of small and medium-sized bridge.

Description

Technical Field

[0001] This invention relates to the field of bridge structural health monitoring technology, specifically to a lightweight monitoring and damage identification method and system for small and medium span bridges. Background Technology

[0002] With the improvement of transportation infrastructure, small and medium-span bridges with spans of less than 50m are numerous in highway and municipal road networks, forming an important part of the transportation system. These bridges often use hollow slab beams and single-column piers, and are subjected to vehicle loads and environmental factors over long periods of time, making them prone to typical defects such as hinge joint damage and pier tilting. If these defects are not detected and addressed in time, they will threaten structural safety.

[0003] Currently, bridge health monitoring technology is mainly concentrated on large bridges, generally employing industrial control computers paired with high-precision sensors. This approach suffers from problems such as large equipment size, high energy consumption, high cost, and low deployment flexibility, making it unsuitable for the dispersed distribution, limited maintenance funds, and poor power supply conditions of small and medium-span bridges. Existing monitoring of small and medium-span bridges still relies primarily on manual inspections, which are inefficient, subjective, and have a high rate of missed detections, failing to achieve real-time monitoring. Some simplified monitoring schemes suffer from limited monitoring indicators and insufficient accuracy, failing to meet the needs of precision maintenance. Therefore, there is an urgent need for a monitoring and damage identification solution that is adapted to the characteristics of small and medium-span bridges, combining lightweight design, low cost, high precision, and real-time capabilities. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a lightweight monitoring and damage identification solution for small and medium-span bridges throughout the entire process. It specifically addresses pain points such as high cost, difficult deployment, insufficient accuracy, and ambiguous damage location, thereby achieving refined and intelligent monitoring and maintenance of small and medium-span bridges.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: On the one hand, a lightweight monitoring and damage identification method for small and medium span bridges is provided, including: S1. Based on the structural type of small and medium span bridges, determine the target monitoring indicators and monitoring data types corresponding to their main defects, and determine the general monitoring indicators and monitoring data types; S2. An online visual sensing device with integrated edge computing function and an MCU smart gateway are used to collect monitoring data corresponding to the monitoring data types determined in S1. The online visual sensing device is used to collect monitoring data corresponding to the targeted monitoring indicators and perform image preprocessing. The MCU smart gateway receives the monitoring data after image preprocessing from the online visual sensing device, and at the same time collects monitoring data corresponding to general monitoring indicators through environmental sensors. All monitoring data are pre-processed and stored in the MCU smart gateway. Based on data latency requirements, the S3.MCU smart gateway performs edge computing on the real-time data in the stored monitoring data and outputs preliminary monitoring results. It also transmits the historical data in the stored monitoring data to the cloud data center for storage and in-depth analysis. S4. Based on all monitoring data stored in the MCU smart gateway, and combined with the preset damage identification model corresponding to the targeted monitoring indicators, bridge defects are identified and their severity is assessed. S5. Trigger an early warning based on the identification and evaluation results, and push the early warning information and the identification and evaluation results to the maintenance personnel's terminal.

[0006] Furthermore, in S1: When the small-span bridge is a hollow slab beam bridge, the targeted monitoring indicators include the relative displacement of the slab beams on both sides of the hinge joint, and the corresponding monitoring data type is the vertical deflection data of the slab beams on both sides of the hinge joint. When the small-span bridge is a single-column pier bridge, the targeted monitoring indicators include the pier rotation angle, and the corresponding monitoring data type is the pier rotation angle data. The general monitoring indicators include bridge deck displacement and ambient temperature and humidity, and the corresponding data types are bridge deck displacement data and ambient temperature and humidity data, respectively.

[0007] Furthermore, in S4, the damage identification model includes: Hinge joint damage identification model: Based on the vertical deflection values ​​of the beams on both sides of the hinge joint in the monitoring data, the damage degree of each hinge joint is calculated using the damage degree calculation formula. The damage degree calculation formula is as follows: ; in, For the first i The degree of damage to the hinge joint. and The first i Vertical deflection of the beams on both sides of the hinge joint; Pier overturning risk identification model: Based on the pier rotation angle value in the monitoring data, combined with the preset rotation angle threshold, it is determined whether the pier has an overturning risk; Anomaly detection model: Combining bridge deck displacement data and environmental temperature and humidity data from the monitoring data, the 3σ criterion is used to detect anomalies and identify abnormalities in bridge deck settlement and component deformation.

[0008] Furthermore, in S2, the online visual perception device is equipped with an infrared array target and spot centroid tracking algorithm; image preprocessing includes image distortion correction, denoising, dehazing, binarization and area consistency constraint optimization, as well as centroid localization processing.

[0009] Furthermore, the centroid localization process uses a grayscale centroid localization algorithm to calculate the target centroid coordinates. The calculation formula for the grayscale centroid localization algorithm is as follows: ; in , The coordinates of the centroid of the target region within the region of interest. pixel coordinates ( i,j The grayscale value at ) m , n The pixel size of the region of interest.

[0010] Furthermore, the MCU smart gateway adopts a low-power design, including a working mode and a sleep mode. In sleep mode, only the interrupt mechanism of the wireless communication module is enabled. The MCU smart gateway also uses dynamic frequency adjustment technology to adjust the MCU main frequency.

[0011] Furthermore, in S5, tiered early warnings are issued based on the identification and assessment results, and the early warning information and identification and assessment results are pushed to the maintenance personnel's terminals.

[0012] On the other hand, a system for implementing the above method is provided, comprising: Targeted monitoring unit: includes online visual sensing device and environmental sensor. The online visual sensing device is used to collect monitoring data corresponding to the targeted monitoring indicators and perform image preprocessing; the environmental sensor is used to collect monitoring data corresponding to general monitoring indicators. Lightweight Gateway Unit: Employs an MCU smart gateway, integrating a data acquisition module, a data processing module, a wireless communication module, and a power management module. The data acquisition module receives monitoring data collected by the target monitoring unit; the data processing module performs preliminary processing and storage of all monitoring data; the wireless communication module receives pre-processed monitoring data from online visual sensing devices and transmits historical data from the stored monitoring data to the cloud data center; the power management module provides low-power, stable power to the system. Cloud-edge collaborative processing unit: includes an edge processing module integrated into the MCU smart gateway and a cloud data center. The edge processing module is used to perform edge computing on the real-time data in the stored monitoring data and output preliminary monitoring results. The cloud data center is used to receive historical data in the stored monitoring data and perform storage and in-depth analysis. Damage identification unit: Integrated into the cloud-edge collaborative processing unit, it is used to identify and assess the severity of bridge defects based on all monitoring data stored in the MCU smart gateway and a preset damage identification model that corresponds to the targeted monitoring indicators. Early warning and feedback unit: including an audible and visual alarm and a terminal push module. The audible and visual alarm is used to trigger an early warning based on the identification and evaluation results, and the terminal push module is used to push the early warning information and the identification and evaluation results to the maintenance personnel's terminal.

[0013] Furthermore, online visual sensing devices include industrial cameras, infrared array targets, and microprocessors; The industrial camera uses a global shutter CMOS industrial camera with a resolution of no less than 2448×2048 pixels and a target size of no less than 2 / 3". The infrared array target uses active infrared LEDs. Furthermore, the MCU smart gateway uses the STM32F407 controller chip, with a main frequency of up to 168MHz and power consumption reduced to below 10mW in sleep mode; The wireless communication module adopts a dual-mode design of LoRa and Wi-Fi, with a LoRa communication distance of no less than 1km; the power management module uses a combination of lithium battery and solar power.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention employs an online visual sensing device with integrated edge computing capabilities and an MCU smart gateway to replace the bulky, high-power, and expensive combination of industrial control computers and high-precision sensors in traditional bridge monitoring systems. The MCU smart gateway features low-power design with switching between working and sleep modes and dynamic frequency adjustment. In sleep mode, power consumption can be reduced to below 10mW. It supports a combination of lithium battery and solar power supply, eliminating the need for an external 220V power supply. The online visual sensing device is equipped with a microprocessor, embedding a spot centroid tracking algorithm and an image preprocessing algorithm, enabling it to independently complete image acquisition and processing without the need for external computing devices. The overall device is small in size, low in power consumption, and inexpensive, making it flexibly adaptable to real-world scenarios where small and medium-span bridges are scattered, maintenance funds are limited, and power supply conditions are poor, resulting in convenient and efficient deployment.

[0015] 2. This invention employs a targeted monitoring design for typical defects in small- and medium-span bridges: For hollow slab girder bridges, the relative displacement of the slabs on both sides of the hinge joint is used as the core monitoring indicator, and the degree of damage is calculated by measuring the vertical deflection data of the slabs on both sides of the hinge joint; for single-column pier bridges, the pier rotation angle is used as the core monitoring indicator, and the risk of overturning is determined by real-time monitoring of the rotation angle data. Simultaneously, general monitoring indicators such as bridge deck displacement and ambient temperature and humidity are set to consider the overall structural condition assessment. This targeted monitoring scheme avoids the problems of sensor redundancy and indicator generalization in traditional schemes, effectively reducing equipment costs and improving the targeting and accuracy of monitoring. Through the use of a spot centroid tracking algorithm and image preprocessing technology, the root mean square error of displacement monitoring does not exceed 0.24 mm, and the root mean square error of rotation angle monitoring does not exceed 0.0045°, fully meeting the engineering monitoring requirements.

[0016] 3. This invention adopts a cloud-edge collaborative processing architecture. The MCU smart gateway autonomously determines the data processing method based on data latency requirements: real-time data with high real-time requirements undergoes edge computing locally on the MCU smart gateway to quickly output preliminary monitoring results, ensuring timely detection of anomalies; historical data requiring long-term analysis is transmitted to the cloud data center for storage and in-depth analysis, achieving comprehensive, full-cycle monitoring and trend prediction of structural status. Simultaneously, based on all stored monitoring data and combined with damage identification models corresponding to targeted monitoring indicators, the identification and severity assessment of bridge defects are completed. Through the division of labor and collaboration among these three data processing methods—edge computing, in-depth analysis, and damage identification—the data transmission volume and network bandwidth pressure are reduced while ensuring the multiple requirements of real-time response, comprehensive monitoring, and accurate assessment are met.

[0017] 4. This invention relies on a targeted damage identification model to accurately locate and assess the severity of typical bridge defects. For hinge joint damage, the degree of damage is calculated using a formula, quantitatively reflecting the force transmission performance of the hinge joint; for pier overturning risk, a real-time assessment is made in conjunction with a preset rotation angle threshold; for abnormal conditions such as bridge deck settlement and component deformation, the 3σ criterion is used in conjunction with environmental temperature and humidity data for anomaly detection. The system automatically triggers early warnings based on the identification results and pushes the early warning information and severity assessment results to the maintenance personnel's terminals, forming a closed-loop management system of monitoring-identification-early warning-maintenance. This provides maintenance personnel with accurate and timely decision-making basis, effectively reducing the cost of manual inspections, preventing further development of defects, and ensuring bridge traffic safety.

[0018] 5. This invention's system supports the access of multiple types of sensors, and can flexibly adjust monitoring indicators and equipment deployment according to the monitoring needs of different types of small and medium-span bridges, adapting to various small and medium-span bridge structures such as hollow slab beam bridges and single-column pier bridges. The MCU smart gateway integrates a data acquisition module, a data processing module, a wireless communication module, and a power management module. The wireless communication module adopts a dual-mode design of LoRa and Wi-Fi, which not only meets the requirements of low-power long-distance transmission but also supports communication with cloud data centers, possessing good versatility and scalability, facilitating subsequent promotion and application. Attached Figure Description

[0019] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0021] like Figure 1 As shown, the present invention provides a lightweight monitoring and damage identification method for small and medium span bridges, comprising: S1. Based on the structural type of small and medium span bridges, determine the target monitoring indicators and monitoring data types corresponding to their main defects, and determine the general monitoring indicators and monitoring data types; S2. An online visual sensing device with integrated edge computing function and an MCU smart gateway are used to collect monitoring data corresponding to the monitoring data types determined in S1. The online visual sensing device is used to collect monitoring data corresponding to the targeted monitoring indicators and perform image preprocessing. The MCU smart gateway receives the monitoring data after image preprocessing from the online visual sensing device, and at the same time collects monitoring data corresponding to general monitoring indicators through environmental sensors. All monitoring data are pre-processed and stored in the MCU smart gateway. Based on data latency requirements, the S3.MCU smart gateway performs edge computing on the real-time data in the stored monitoring data and outputs preliminary monitoring results. It also transmits the historical data in the stored monitoring data to the cloud data center for storage and in-depth analysis. S4. Based on all monitoring data stored in the MCU smart gateway, and combined with the preset damage identification model corresponding to the targeted monitoring indicators, bridge defects are identified and their severity is assessed. S5. Trigger an early warning based on the identification and evaluation results, and push the early warning information and the identification and evaluation results to the maintenance personnel's terminal.

[0022] This invention achieves lightweight, precise, and intelligent monitoring of small and medium-span bridges through the aforementioned methods. The system employs an MCU intelligent gateway and online visual sensing equipment, featuring low power consumption, low cost, and rapid deployment, making it suitable for outdoor scenarios without power supply. Through targeted monitoring design, it accurately locates hinge joint damage and pier overturning risks, with monitoring accuracy meeting engineering requirements. Relying on a cloud-edge collaborative architecture, real-time data edge processing ensures rapid response, while historical data cloud analysis supports trend prediction. Combining a damage identification model and a hierarchical early warning mechanism, a closed loop of monitoring-identification-early warning-maintenance is formed, effectively reducing manual inspection costs, improving maintenance efficiency, and ensuring bridge traffic safety.

[0023] The targeted monitoring indicators are the core physical quantities set by this invention for typical defects of small and medium span bridges. They are selected based on the structural type of the bridge and its main defect risks, and the key indicators that can most directly reflect the structural health status are monitored to avoid the problems of sensor redundancy and indicator generalization in traditional monitoring schemes.

[0024] In practical implementation, this invention determines targeted monitoring indicators corresponding to the main defects of small and medium-span bridges based on their structural types. For hollow slab girder bridges, the main defect is hinge joint damage, manifested as a decrease in the force transmission performance between adjacent slab girders. Therefore, the relative displacement of the slab girders on both sides of the hinge joint is used as the core monitoring indicator. By measuring the vertical deflection data (displacement data) at the lower ends of the slab girders on both sides of the hinge joint, the relative displacement is calculated, thereby quantitatively assessing the degree of hinge joint damage. For single-column pier bridges, the main risk is overturning, manifested as excessive rotation at the top of the pier. Therefore, the pier rotation angle is used as the core monitoring indicator. By monitoring the tilt angle of the pier in real time (rotation angle data) and combining it with a preset rotation angle threshold, the risk of overturning is judged. In addition, this invention also sets general monitoring indicators, including bridge deck displacement and ambient temperature and humidity, to monitor changes in the overall alignment of the bridge and provide environmental references for damage identification.

[0025] By designing the above-mentioned targeted monitoring indicators, this invention focuses monitoring resources on the most critical and weakest structural parts, using the fewest sensors to obtain the most critical structural status information. This reduces equipment costs and deployment complexity while improving the targeting and accuracy of monitoring, thus achieving lightweight and precise monitoring of small and medium-span bridges.

[0026] This invention employs an online visual sensing device with integrated edge computing capabilities and an MCU smart gateway to collect monitoring data corresponding to the various monitoring data types determined in S1. The online visual sensing device is the core hardware for achieving lightweight data acquisition in this invention. It is equipped with an infrared array target and a spot centroid tracking algorithm, and has a built-in microprocessor, enabling it to independently complete image acquisition and preprocessing without the need for external computing devices.

[0027] In practical implementation, the online visual perception device includes an industrial camera, an infrared array target, and a microprocessor. The industrial camera is a global shutter CMOS industrial camera with a resolution of no less than 2448×2048 pixels and a target size of no less than 2 / 3". The infrared array target uses active infrared LEDs, which can be adapted to outdoor rain and fog environments. The microprocessor has embedded spot centroid tracking algorithm and camera disturbance elimination algorithm.

[0028] Online visual sensing equipment is used to collect monitoring data (such as displacement data and rotation angle data) corresponding to target monitoring indicators and perform image preprocessing. Image preprocessing includes image distortion correction, denoising, dehazing, binarization, area consistency constraint optimization, and centroid localization. Specifically, image distortion correction uses Zhang's calibration method to determine camera intrinsic parameters, eliminating the impact of camera distortion on data accuracy; the corrected image is then denoised and dehazed to improve image clarity and ensure the accuracy of subsequent data calculations; the processed image is binarized to determine the region of interest (ROI) and calculate the target area; an adaptive threshold is fine-tuned based on area consistency constraints to make the target area before and after deformation as close as possible; centroid localization uses a grayscale centroid localization algorithm to calculate the target centroid coordinates. The calculation formula for this algorithm is... ,in , The coordinates of the centroid of the target region within the region of interest. pixel coordinates ( i,j The grayscale value at ) m , n The pixel size of the region of interest.

[0029] The MCU smart gateway is another core hardware component of this invention for achieving lightweight data processing. It adopts a low-power design, including a working mode and a sleep mode. In sleep mode, only the interrupt mechanism of the wireless communication module is enabled. When an external wake-up command is received, it immediately switches to the working mode and uses dynamic frequency adjustment technology to adjust the MCU main frequency according to the data processing requirements. In sleep mode, the power consumption can be reduced to below 10mW.

[0030] The MCU smart gateway receives pre-processed monitoring data (such as displacement and rotation data) from online visual sensing devices. Simultaneously, the MCU smart gateway collects monitoring data corresponding to general monitoring indicators (such as ambient temperature and humidity) through environmental sensors. All monitoring data undergoes preliminary processing and storage locally on the MCU smart gateway. The environmental sensors are electrically connected to the MCU smart gateway, transmitting the collected ambient temperature and humidity data to the MCU smart gateway, thereby achieving lightweight data acquisition and reducing data transmission burden.

[0031] In a specific embodiment of the present invention, the preliminary processing specifically includes the following: First, data reception and integration. The system receives pre-processed displacement and rotation data from online visual sensing devices, and simultaneously collects environmental temperature and humidity data from environmental sensors. Data from different sources is then time-aligned and format-unified to form a structured monitoring dataset.

[0032] Second, invalid data removal. The received monitoring data is validated, and data points that clearly exceed the measurement range, contain communication errors, or have abnormal formats are removed to ensure the accuracy of subsequent processing. For example, displacement data exceeding a preset physical limit value is considered invalid and removed.

[0033] Third, local caching and storage. All monitoring data is cached in local storage using a circular storage strategy. Recent data is retained for edge computing, while historical data is prepared for uploading to the cloud data center. Local caching effectively prevents data loss due to network interruptions.

[0034] Fourth, edge computing preparation. Monitoring data is categorized based on latency requirements. Displacement and rotation data with high real-time requirements are marked as real-time data, ready for edge computing; historical data requiring long-term analysis is marked as data to be uploaded, ready for transmission to the cloud data center.

[0035] Through the above preliminary processing, the MCU smart gateway realizes the localized organization, filtering, storage and classification of monitoring data, providing a high-quality data foundation for subsequent edge computing and cloud analysis.

[0036] This invention achieves the technical effects of small device size, low power consumption, and low cost by combining an online visual perception device with integrated edge computing function and an MCU smart gateway in a lightweight hardware architecture. It can be flexibly adapted to actual scenarios where small and medium-sized bridges are distributed in a dispersed manner, have limited maintenance funds, and have poor power supply conditions.

[0037] Based on this, the present invention employs three data processing methods to achieve intelligent processing throughout the entire process from data acquisition to disease assessment. The MCU intelligent gateway, according to data latency requirements, performs edge computing on real-time data from the stored monitoring data and outputs preliminary monitoring results. It also transmits historical data from the stored monitoring data to a cloud data center for storage and in-depth analysis. Simultaneously, it completes bridge disease identification and severity assessment based on all stored monitoring data. These three data processing methods have clear divisions of labor and work together to form a complete monitoring and damage identification method.

[0038] The first method is edge computing. For real-time data with high real-time requirements, edge computing is performed locally on the MCU smart gateway to quickly output preliminary monitoring results and ensure timely detection of anomalies.

[0039] In a specific embodiment of the present invention, edge computing specifically includes: For hollow slab girder bridges, the vertical deflection values ​​of the slab girders on both sides of the hinge joint are used as the basis for determining the vertical deflection values. and Real-time calculation of the relative displacement of the beams on both sides of the hinge joint. and relative displacement relative displacement threshold Compare them.

[0040] For single-column pier bridges, the pier rotation angle value is based on the pier rotation angle data. θ Compare it with the preset corner threshold (e.g., 0.01°) are compared.

[0041] The comparison results of the above edge computing can ensure that anomalies are detected in a timely manner. When or When an anomaly is detected, an alert is issued. The total delay from data acquisition to alert triggering does not exceed 100 milliseconds, ensuring that anomalies can be detected within milliseconds.

[0042] The second method is in-depth analysis. Historical data requiring long-term analysis is transmitted to a cloud data center for storage and in-depth analysis, enabling comprehensive, full-cycle monitoring and trend prediction of structural status.

[0043] In a specific embodiment of the present invention, in-depth analysis specifically includes the following: Hinge joint damage trend analysis: Based on the vertical deflection data of the beams on both sides of the hinge joint, the vertical deflection values ​​of the beams on both sides of the hinge joint are analyzed. and Analyze the relative displacement By analyzing historical data and calculating statistical characteristics such as mean, standard deviation, and rate of change, damage development trends can be predicted. Early warnings can be issued when damage is accelerating.

[0044] Pier overturning risk trend analysis: based on pier rotation angle values ​​in the pier rotation angle data θ By combining environmental temperature and humidity data to eliminate the influence of temperature, the historical mean and standard deviation of the turning angle are calculated, and a temperature compensation model is established. When the turning angle continues to deviate from the mean or the fluctuation increases, it is determined that there is a long-term risk of overturning.

[0045] Damage trend analysis of bridge deck and other components: Combining bridge deck displacement and environmental temperature and humidity data, the 3σ criterion is used for anomaly detection. The historical mean and standard deviation of displacement are calculated. When the displacement exceeds the mean ± 3 times the standard deviation, anomalies are identified. After eliminating the influence of temperature, bridge deck settlement and component deformation are identified, and the changing trend of anomaly locations is analyzed.

[0046] The total delay from data upload to analysis completion is no more than 10 minutes, enabling rapid and in-depth mining of historical monitoring data while meeting the needs of real-time monitoring and long-term trend analysis.

[0047] The third method: Damage identification. Based on all stored monitoring data, and combined with the damage identification model corresponding to the targeted monitoring indicators, bridge defects are identified and their severity is assessed.

[0048] In a specific embodiment of the present invention, the damage recognition model includes: Hinge joint damage identification model: Based on the vertical deflection values ​​of the beams on both sides of the hinge joint in the monitoring data, the damage degree of each hinge joint is calculated using the damage degree calculation formula. The damage degree calculation formula is as follows: ; in, For the first i The degree of damage to the hinge joint. and The first i Vertical deflection of the beams on both sides of the hinge joint.

[0049] Specifically, when When the hinge is undamaged and its force transmission performance is normal; when This indicates that the hinge joint has varying degrees of damage and its force transmission capacity has decreased; when This indicates complete failure of the hinge joint and loss of force transmission connection between the two side beams. Furthermore, combining this with historical variation curves output from depth analysis can help determine damage development trends and provide a reference for maintenance decisions.

[0050] Pier overturning risk identification model: Based on the pier rotation angle value in the monitoring data, combined with the preset rotation angle threshold, it determines whether the pier has an overturning risk.

[0051] Specifically, an early warning is immediately triggered when the pier's rotation angle exceeds a preset threshold. Simultaneously, the long-term overturning risk level can be assessed by combining historical statistical characteristics and temperature influence coefficients output from in-depth analysis.

[0052] Anomaly detection model: Combining bridge deck displacement data and environmental temperature and humidity data from the monitoring data, the 3σ criterion is used to detect anomalies and identify abnormalities in bridge deck settlement and component deformation.

[0053] Specifically, when the bridge deck displacement data exceeds the historical average by ±3 times the standard deviation, it is judged as abnormal. The influence of temperature is eliminated by combining the ambient temperature and humidity data, so as to achieve accurate location of damage.

[0054] An early warning is triggered based on the above identification and severity assessment results, and the warning information, along with the identification and assessment results, is pushed to the maintenance personnel's terminal.

[0055] In a specific embodiment of the present invention, the warning is divided into three levels based on the degree of damage and abnormal monitoring data in the identification and severity assessment results: a minor warning corresponds to minor damage to the hinge joint and displacement approaching the threshold; a moderate warning corresponds to moderate damage to the hinge joint and rotation angle approaching the threshold; and a severe warning corresponds to severe damage to the hinge joint, rotation angle exceeding the threshold, and abnormal bridge deck settlement.

[0056] Furthermore, different warning levels correspond to different audio-visual prompts, facilitating on-site personnel to quickly identify the warning level and take appropriate measures. When a warning is triggered, warning information including the warning level, as well as identification and severity assessment results including damage location, damage extent, and monitoring data, are simultaneously pushed to the maintenance personnel's terminal (handheld terminal or on-site monitoring equipment). At the same time, a damage report is automatically generated, providing clear maintenance recommendations and offering maintenance personnel accurate maintenance basis to ensure timely and efficient maintenance work.

[0057] The present invention also provides a system for implementing the above method, including a target monitoring unit, a lightweight gateway unit, a cloud-edge collaborative processing unit, a damage identification unit, and an early warning and feedback unit.

[0058] The targeted monitoring unit includes an online visual sensing device and an environmental sensor. The online visual sensing device is used to collect displacement and rotation data corresponding to the targeted monitoring indicators and to perform image preprocessing; the environmental sensor is used to collect ambient temperature and humidity data corresponding to general monitoring indicators.

[0059] The lightweight gateway unit, employing an MCU smart gateway, integrates a data acquisition module, a data processing module, a wireless communication module, and a power management module. The data acquisition module receives monitoring data collected by the target monitoring unit; the data processing module performs preliminary local processing and storage of all monitoring data; the wireless communication module receives pre-processed monitoring data from the online visual sensing device and transmits historical data from the stored monitoring data to the cloud data center; the power management module combines lithium battery and solar power to provide a stable power supply for the MCU smart gateway and its connected online visual sensing device, environmental sensors, and wireless communication module, achieving low-power, stable power supply through switching between operating and sleep modes and dynamic frequency adjustment technology.

[0060] The cloud-edge collaborative processing unit includes an edge processing module integrated into the MCU smart gateway and a cloud data center. The edge processing module performs edge computing on real-time data from the stored monitoring data and outputs preliminary monitoring results; the cloud data center receives historical data from the stored monitoring data and performs storage and in-depth analysis.

[0061] The damage identification unit, integrated into the cloud-edge collaborative processing unit, is used to identify and assess the severity of bridge defects based on all stored monitoring data and the damage identification model corresponding to the targeted monitoring indicators.

[0062] The early warning and feedback unit includes an audible and visual alarm and a terminal push module. The audible and visual alarm is used to trigger an early warning based on the identification and severity assessment results; the terminal push module is used to push the early warning information and the identification and severity assessment results to the maintenance personnel's terminal.

[0063] In a specific embodiment of the present invention, the online visual sensing device includes an industrial camera, an infrared array target, and a microprocessor. The industrial camera is a global shutter CMOS industrial camera with a resolution of no less than 2448×2048 pixels and a target surface size of no less than 2 / 3". The infrared array target uses active infrared LEDs, adaptable to outdoor rain and fog environments. The microprocessor embeds a spot centroid tracking algorithm and a camera disturbance cancellation algorithm, preferably using an STM32F407, capable of independently completing image acquisition and preprocessing without the need for external computing devices. An environmental sensor is used to collect environmental temperature and humidity data corresponding to general monitoring indicators, providing environmental references for damage identification. The environmental sensor is electrically connected to the MCU smart gateway, transmitting the collected environmental temperature and humidity data to the MCU smart gateway for local processing and storage.

[0064] In a specific embodiment of the present invention, the MCU smart gateway uses an STM32F407 controller chip with a main frequency of up to 168MHz and power consumption reduced to below 10mW in sleep mode.

[0065] In a specific embodiment of this invention, the wireless communication module adopts a dual-mode design of LoRa and Wi-Fi. The LoRa module is used for long-distance, low-power data transmission with the cloud data center, with a communication distance of not less than 1km. It is responsible for uploading historical data from the monitoring data to the cloud and receiving parameter configuration and wake-up commands from the cloud. The Wi-Fi module is used for near-field, high-speed communication with maintenance personnel's terminals. At the bridge site, early warning information can be directly pushed to the handheld terminals of maintenance personnel or on-site monitoring equipment. Through the cloud-edge collaborative processing architecture, both the data transmission volume and network bandwidth pressure are reduced, while ensuring the dual requirements of real-time response and in-depth analysis are met.

[0066] Finally, it should be noted that the above embodiments are merely preferred embodiments of the present invention used to illustrate the technical solutions of the present invention, and are not intended to limit the invention, nor are they intended to limit the patent scope of the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. These modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention. That is to say, any changes or refinements made to the main design concept and spirit of the present invention that are not of substantial significance, but whose technical problems are still consistent with the present invention, should be included within the protection scope of the present invention. In addition, the direct or indirect application of the technical solutions of the present invention to other related technical fields are similarly included within the patent protection scope of the present invention.

Claims

1. A method for lightweight monitoring and damage identification of small-to-medium span bridges, characterized in that, include: S1. Based on the structural type of small and medium span bridges, determine the target monitoring indicators and monitoring data types corresponding to their main defects, and determine the general monitoring indicators and monitoring data types; S2. An online visual sensing device with integrated edge computing function and an MCU smart gateway are used to collect monitoring data corresponding to the monitoring data types determined in S1. The online visual sensing device is used to collect monitoring data corresponding to the targeted monitoring indicators and perform image preprocessing. The MCU smart gateway receives the monitoring data after image preprocessing from the online visual sensing device, and at the same time collects monitoring data corresponding to general monitoring indicators through environmental sensors. All monitoring data are pre-processed and stored in the MCU smart gateway. Based on data latency requirements, the S3.MCU smart gateway performs edge computing on the real-time data in the stored monitoring data and outputs preliminary monitoring results. It also transmits the historical data in the stored monitoring data to the cloud data center for storage and in-depth analysis. S4. Based on all monitoring data stored in the MCU smart gateway, and combined with the preset damage identification model corresponding to the targeted monitoring indicators, bridge defects are identified and their severity is assessed. S5. Trigger an early warning based on the identification and evaluation results, and push the early warning information and the identification and evaluation results to the maintenance personnel's terminal.

2. The monitoring and damage identification method according to claim 1, characterized in that, In S1: When the small-span bridge is a hollow slab beam bridge, the targeted monitoring indicators include the relative displacement of the slab beams on both sides of the hinge joint, and the corresponding monitoring data type is the vertical deflection data of the slab beams on both sides of the hinge joint. When the small-span bridge is a single-column pier bridge, the targeted monitoring indicators include the pier rotation angle, and the corresponding monitoring data type is the pier rotation angle data. The general monitoring indicators include bridge deck displacement and ambient temperature and humidity, and the corresponding data types are bridge deck displacement data and ambient temperature and humidity data, respectively.

3. The monitoring and damage identification method according to claim 2, characterized in that, In S4, the damage identification model includes: Hinge joint damage identification model: Based on the vertical deflection values ​​of the beams on both sides of the hinge joint in the monitoring data, the damage degree of each hinge joint is calculated using the damage degree calculation formula. The damage degree calculation formula is as follows: ; in, For the first i The degree of damage to the hinge joint. and The first i Vertical deflection of the beams on both sides of the hinge joint; Pier overturning risk identification model: Based on the pier rotation angle value in the monitoring data, combined with the preset rotation angle threshold, it is determined whether the pier has an overturning risk; Anomaly detection model: Combining bridge deck displacement data and environmental temperature and humidity data from the monitoring data, the 3σ criterion is used to detect anomalies and identify abnormalities in bridge deck settlement and component deformation.

4. The monitoring and damage identification method according to claim 1, characterized in that, In S2, the online visual perception device is equipped with an infrared array target and spot centroid tracking algorithm; image preprocessing includes image distortion correction, denoising, dehazing, binarization and area consistency constraint optimization, as well as centroid localization processing.

5. The monitoring and damage identification method according to claim 4, characterized in that, The centroid localization process uses a grayscale centroid localization algorithm to calculate the target centroid coordinates. The calculation formula for the grayscale centroid localization algorithm is as follows: ; in , The coordinates of the centroid of the target region within the region of interest. pixel coordinates ( i,j The grayscale value at ) m , n The pixel size of the region of interest.

6. The monitoring and damage identification method according to claim 1, characterized in that, The MCU smart gateway adopts a low-power design, including a working mode and a sleep mode. In sleep mode, only the interrupt mechanism of the wireless communication module is enabled. The MCU smart gateway also uses dynamic frequency adjustment technology to adjust the MCU main frequency.

7. The monitoring and damage identification method according to claim 1, characterized in that, In S5, tiered early warnings are issued based on the identification and assessment results, and the early warning information and identification and assessment results are pushed to the maintenance personnel's terminals.

8. A system for implementing the method according to any one of claims 1 to 7, characterized in that, include: Targeted monitoring unit: includes online visual sensing device and environmental sensor. The online visual sensing device is used to collect monitoring data corresponding to the targeted monitoring indicators and perform image preprocessing. Environmental sensors are used to collect monitoring data corresponding to general monitoring indicators; Lightweight Gateway Unit: Employs an MCU smart gateway, integrating a data acquisition module, a data processing module, a wireless communication module, and a power management module. The data acquisition module receives monitoring data collected by the target monitoring unit; the data processing module performs preliminary processing and storage of all monitoring data; the wireless communication module receives pre-processed monitoring data from online visual sensing devices and transmits historical data from the stored monitoring data to the cloud data center; the power management module provides low-power, stable power to the system. Cloud-edge collaborative processing unit: includes an edge processing module integrated into the MCU smart gateway and a cloud data center. The edge processing module is used to perform edge computing on the real-time data in the stored monitoring data and output preliminary monitoring results. The cloud data center is used to receive historical data in the stored monitoring data and perform storage and in-depth analysis. Damage identification unit: Integrated into the cloud-edge collaborative processing unit, it is used to identify and assess the severity of bridge defects based on all monitoring data stored in the MCU smart gateway and a preset damage identification model that corresponds to the targeted monitoring indicators. Early warning and feedback unit: including an audible and visual alarm and a terminal push module. The audible and visual alarm is used to trigger an early warning based on the identification and evaluation results, and the terminal push module is used to push the early warning information and the identification and evaluation results to the maintenance personnel's terminal.

9. The system according to claim 8, characterized in that, Online visual sensing equipment includes industrial cameras, infrared array targets, and microprocessors; The industrial camera uses a global shutter CMOS industrial camera with a resolution of no less than 2448×2048 pixels and a target size of no less than 2 / 3". The infrared array target uses active infrared LEDs.

10. The system according to claim 8, characterized in that, The MCU smart gateway uses the STM32F407 controller chip, with a main frequency of up to 168MHz and power consumption reduced to below 10mW in sleep mode; The wireless communication module adopts a dual-mode design of LoRa and Wi-Fi, with a LoRa communication distance of no less than 1km; the power management module uses a combination of lithium battery and solar power.

Images