CN117076928A

CN117076928A - Bridge health state monitoring method, device and system and electronic equipment

Info

Publication number: CN117076928A
Application number: CN202311080993.3A
Authority: CN
Inventors: 孟熙; 付金璐; 崔春雷; 魏广宾; 韩红彩; 贾林萱; 刘永
Original assignee: China Road & Bridge Technology Co ltd
Current assignee: China Road & Bridge Technology Co ltd
Priority date: 2023-08-25
Filing date: 2023-08-25
Publication date: 2023-11-17

Abstract

The application discloses a bridge health state monitoring method, device and system and electronic equipment, and belongs to the technical field of structural safety monitoring. According to the embodiment of the application, various real-time monitoring data are acquired by arranging various real-time monitoring data acquisition equipment, and are processed, so that the health state of the bridge is judged, compared with manual monitoring, manual operation is not needed completely, the labor cost is saved completely, the health state of the bridge can be acquired in time through automatic equipment, the processing efficiency and timeliness are improved, in addition, the types of the real-time monitoring data can be automatically screened through one large model, the sub-model of the corresponding type is selected for data processing, various types of monitoring data can be fused, the health state of the bridge is comprehensively analyzed, the process is completely automated, human errors are avoided, and the processing efficiency and accuracy of the monitoring result are greatly improved.

Description

Bridge health state monitoring method, device and system and electronic equipment

Technical Field

The application relates to the technical field of structural safety monitoring, in particular to a bridge health state monitoring method, device and system and electronic equipment.

Background

The bridge is an important traffic infrastructure, provides great convenience for people living and provides strong support for social and economic development, so that the bridge has great significance in guaranteeing the safe operation of the bridge facility.

At present, the bridge health status monitoring method mainly adopts manual monitoring, namely, manual in-situ monitoring data is adopted, and the health status of the bridge is judged according to self experience by monitoring staff, so that the method cannot detect the status of the bridge comprehensively in real time, the labor cost is high, the method depends on the technical level of the monitoring staff, and the accuracy of finally judging the obtained health status is required to be checked.

Disclosure of Invention

The embodiment of the application provides a bridge health state monitoring method, device and system and electronic equipment, which can achieve the effects of improving monitoring efficiency and accuracy and reducing cost. The technical scheme is as follows:

in one aspect, a method for monitoring health status of a bridge is provided, the method comprising:

acquiring first real-time monitoring data acquired by a plurality of sensors, wherein the plurality of sensors are arranged at different structural positions on a bridge and are used for acquiring data of different types of the bridge;

Acquiring second real-time monitoring data of the microwave radar for deflection dynamic high-precision measurement monitoring;

acquiring third real-time monitoring data acquired by a monitoring device of the Beidou satellite navigation system;

acquiring fourth real-time monitoring data acquired by a bridge weighing system, wherein the bridge weighing system is arranged on the bridge;

inputting the first real-time monitoring data, the second real-time monitoring data, the third real-time monitoring data and the fourth real-time monitoring data into a state estimation model, judging the type of the input real-time monitoring data by the state estimation model, inputting each real-time monitoring data into a state estimation sub-model corresponding to the type, processing the real-time monitoring data of the corresponding type based on each sub-model to obtain a candidate judging result of the bridge health state, and determining and outputting the health state of the bridge based on the candidate judging results of a plurality of sub-models;

and sending early warning information in response to the abnormal health state of the bridge.

In some embodiments, the determining, by the state estimation model, the type of the input real-time monitoring data, and inputting each real-time monitoring data into a state estimation sub-model corresponding to the type, where the state estimation sub-model includes at least one of the following:

In response to the state estimation model determining that the input real-time monitoring data is image data, inputting the real-time monitoring data into an image state estimation sub-model;

in response to the state estimation model judging that the input real-time monitoring data is temperature data, inputting the real-time monitoring data into a temperature state estimation sub-model;

and in response to the state estimation model judging that the input real-time monitoring data is signal data, inputting the real-time monitoring data into a signal state estimation sub-model.

In some embodiments, the processing the real-time monitoring data of the corresponding type based on each sub-model to obtain the candidate discrimination result of the bridge health status includes at least one of the following:

responding to the real-time monitoring data as image data, wherein the submodel is an image state estimation submodel, preprocessing the image data by the image state estimation submodel, and comparing the extracted image characteristics with big data image characteristics of the bridge by extracting characteristics of the preprocessed image data to obtain candidate discrimination results of the health state of the bridge;

responding to the real-time monitoring data as temperature data, wherein a sub-model is a temperature state estimation sub-model, obtaining temperature load data based on the temperature data by the temperature state estimation sub-model, and carrying out finite element analysis on the structure of the bridge based on the temperature load data to obtain a candidate discrimination result of the health state of the bridge;

Responding to the real-time monitoring data as signal data, wherein the sub-model is a signal state estimation sub-model, if the signal data is a vibration signal, performing feature extraction on the preprocessed vibration signal by the signal state estimation sub-model to obtain depth features, and obtaining a candidate discrimination result of the bridge health state based on the depth feature recognition;

responding to the real-time monitoring data as signal data, wherein the submodel is a signal state estimation submodel, if the signal data is an electromagnetic wave signal, performing difference frequency differential processing on the phases of the transmitted and received electromagnetic wave signals by the signal state estimation submodel, eliminating noise and correcting the atmosphere to obtain a radar line-of-sight deformation graph, converting the radar line-of-sight deformation into bridge longitudinal displacement based on the radar line-of-sight deformation graph, determining a shape variable, and determining a candidate discrimination result of the bridge health state based on the deformation quantity;

responding to the real-time monitoring data as signal data, wherein the sub-model is a signal state estimation sub-model, if the signal data is a strain signal, calculating an initial strain stiffness representative value matrix under closed traffic, an initial strain stiffness representative value expected matrix under open traffic and a confidence interval by the signal state estimation sub-model according to the strain signal, comparing the current strain stiffness representative value expected matrix with the initial strain stiffness representative value expected matrix, if the current strain stiffness representative value expected matrix changes, performing joint investigation on the strain stiffness representative value confidence interval and the acquired strain signal, performing damage judgment after eliminating traffic modification interference, and determining candidate judging results of the bridge health state.

In some embodiments, the difference frequency differential processing is implemented by the following equation one:

；

wherein,for interference phase +.>For the phase component related to the actual displacement, < +.>Is the atmospheric phase component>Is a noise phase component>The whole-cycle ambiguities of the target point phase at two moments are respectively.

In some embodiments, after the noise cancellation and the atmospheric correction, determining the deformation value of the radar line of sight deformation map is implemented by the following formula two:

；

wherein,for deformation values of the target in the LOS direction, < >>For wavelength, < >>Is the phase component associated with the actual displacement.

In some embodiments, the determining and outputting the health status of the bridge based on the candidate discrimination results of the plurality of sub-models includes any one of the following:

determining and outputting the abnormal health state of the bridge in response to the candidate discrimination result of any one of the plurality of sub-models indicating the abnormal health state of the bridge; responding to candidate discrimination results of all sub-models in the plurality of sub-models to indicate that the health state of the bridge is normal, and determining and outputting that the health state of the bridge is normal;

determining and outputting the abnormal health state of the bridge in response to the candidate discrimination results in the plurality of sub-models indicating that the number of the abnormal health state of the bridge is greater than or equal to the target number; and determining and outputting that the health state of the bridge is normal in response to the candidate discrimination results in the plurality of sub-models indicating that the number of the health state anomalies of the bridge is smaller than the target number.

In some embodiments, the model parameters of the plurality of sub-models in the state estimation model are obtained by joint training based on a first error and a second error, wherein the first error is an error between candidate discrimination results of the first sample monitoring data, the second sample monitoring data, the third sample monitoring data and the fourth sample monitoring data and the target health state, and the second error is an error between the predicted health state and the target health state obtained based on the four sample monitoring data.

In some embodiments, the method further comprises:

generating a block based on the real-time monitoring data in response to acquiring the real-time monitoring data acquired at any moment;

adding the block to a blockchain of a blockchain system when the block is commonly passed in the blockchain system;

the acquiring process of the first real-time monitoring data, the second real-time monitoring data, the third real-time monitoring data and the fourth real-time monitoring data comprises the following steps:

and responding to a state estimation instruction of a certain period, and extracting real-time monitoring data in a plurality of blocks from the block chain based on block identifiers of the plurality of blocks corresponding to the period to obtain the first real-time monitoring data, the second real-time monitoring data, the third real-time monitoring data and the fourth real-time monitoring data.

In some embodiments, the method further comprises:

generating a first block based on the first real-time monitoring data, the second real-time monitoring data, the third real-time monitoring data, and the fourth real-time monitoring data;

adding the first block to a blockchain of a blockchain system when the first block is commonly known to pass in the blockchain system;

generating a second block based on the health state of the bridge and the association relation between the health state and the real-time monitoring data in the first block, which are output by the state estimation model;

adding the second block to a blockchain of a blockchain system when the second block is commonly known to pass in the blockchain system;

and correspondingly storing the block identifications of the first block and the second block in a block association relation table.

In one aspect, there is provided a bridge health monitoring device, the device comprising:

the acquisition module is used for acquiring first real-time monitoring data acquired by a plurality of sensors, wherein the plurality of sensors are arranged at different structural positions on the bridge and are used for acquiring data of different types of the bridge;

the acquisition module is used for acquiring second real-time monitoring data obtained by the high-precision measurement and monitoring of deflection dynamic of the microwave radar;

The acquisition module is used for acquiring third real-time monitoring data acquired by a monitoring device of the Beidou satellite navigation system;

the bridge weighing system is arranged on the bridge;

the state estimation module is used for inputting the first real-time monitoring data, the second real-time monitoring data, the third real-time monitoring data and the fourth real-time monitoring data into a state estimation model, judging the type of the input real-time monitoring data by the state estimation model, inputting each real-time monitoring data into a state estimation sub-model corresponding to the type, processing the real-time monitoring data of the corresponding type based on each sub-model to obtain a candidate judging result of the bridge health state, and determining and outputting the health state of the bridge based on the candidate judging results of a plurality of sub-models;

and the sending module is used for responding to the abnormal health state of the bridge and sending early warning information.

In some embodiments, the state estimation module is configured to perform at least one of:

；

In some embodiments, the state estimation module is configured to perform any one of:

In some embodiments, the apparatus further comprises:

the first generation module is used for responding to the acquired real-time monitoring data at any moment and generating a block based on the real-time monitoring data;

a first adding module for adding the block to a blockchain of a blockchain system when the block is commonly passed in the blockchain system;

the acquisition module is used for responding to a state estimation instruction of a certain period, extracting real-time monitoring data in a plurality of blocks from the block chain based on block identifiers of the plurality of blocks corresponding to the period, and obtaining the first real-time monitoring data, the second real-time monitoring data, the third real-time monitoring data and the fourth real-time monitoring data.

In some embodiments, the apparatus further comprises:

the second generation module is used for generating a first block based on the first real-time monitoring data, the second real-time monitoring data, the third real-time monitoring data and the fourth real-time monitoring data;

a second adding module for adding the first block to a blockchain of a blockchain system when the first block is commonly recognized in the blockchain system;

the second generation module is used for generating a second block based on the health state of the bridge and the association relation between the health state and the real-time monitoring data in the first block, which are output by the state estimation model;

the second adding module is used for adding the second block into a blockchain of the blockchain system when the second block is commonly recognized in the blockchain system;

and the storage module is used for correspondingly storing the block identifications of the first block and the second block in the block association relation table.

In one aspect, an electronic device is provided that includes one or more processors and one or more memories having at least one computer program stored therein, the at least one computer program loaded and executed by the one or more processors to implement various alternative implementations of the bridge health monitoring method described above.

In one aspect, a bridge health status monitoring system is provided, the bridge health status monitoring system includes a data acquisition device, a status estimation device and an alarm device; the data acquisition equipment is connected with the state estimation equipment through a network, and the state estimation equipment is connected with the alarm device through the network;

the data acquisition equipment is used for acquiring real-time monitoring data of the bridge, wherein the data acquisition equipment comprises a plurality of sensors, a microwave radar, a monitoring device of a Beidou satellite navigation system and a bridge weighing system, wherein the plurality of sensors are arranged at different structural positions on the bridge;

the state estimation device is loaded with a state estimation model and is used for processing and outputting the health state of the bridge based on real-time monitoring data;

the alarm device is used for responding to the abnormal health state output by the state estimation model to send alarm information.

In one aspect, a computer readable storage medium having at least one computer program stored therein is provided, the at least one computer program being loaded and executed by a processor to implement various alternative implementations of the bridge health monitoring method described above.

In one aspect, a computer program product or computer program is provided, the computer program product or computer program comprising one or more program codes, the one or more program codes being stored in a computer readable storage medium. One or more processors of the electronic device are capable of reading the one or more program codes from the computer readable storage medium, the one or more processors executing the one or more program codes such that the electronic device is capable of performing the bridge health monitoring method of any one of the possible embodiments described above.

According to the embodiment of the application, various real-time monitoring data are acquired by arranging various real-time monitoring data acquisition equipment, and are processed, so that the health state of the bridge is judged, compared with manual monitoring, manual operation is not needed completely, the labor cost is saved completely, the health state of the bridge can be acquired in time through automatic equipment, the processing efficiency and timeliness are improved, in addition, the types of the real-time monitoring data can be automatically screened through one large model, the sub-model of the corresponding type is selected for data processing, various types of monitoring data can be fused, the health state of the bridge is comprehensively analyzed, the process is completely automated, human errors are avoided, and the processing efficiency and accuracy of the monitoring result are greatly improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the description of the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic diagram of a data sharing system according to an embodiment of the present application;

FIG. 2 is a schematic diagram of a bridge health monitoring system according to an embodiment of the present application;

FIG. 3 is a flowchart of a method for monitoring bridge health according to an embodiment of the present application;

fig. 4 is a schematic structural diagram of a bridge health status monitoring device according to an embodiment of the present application;

fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the present application more apparent, the embodiments of the present application will be described in further detail with reference to the accompanying drawings.

The terms "first," "second," and the like in this disclosure are used for distinguishing between similar elements or items having substantially the same function and function, and it should be understood that there is no logical or chronological dependency between the terms "first," "second," and "n," and that there is no limitation on the amount and order of execution. It will be further understood that, although the following description uses the terms first, second, etc. to describe various elements, these elements should not be limited by the terms. These terms are only used to distinguish one element from another element. For example, a first image can be referred to as a second image, and similarly, a second image can be referred to as a first image, without departing from the scope of the various described examples. The first image and the second image can both be images, and in some cases, can be separate and distinct images.

The term "at least one" in the present application means one or more, and the term "plurality" in the present application means two or more, for example, a plurality of data packets means two or more data packets.

It is to be understood that the terminology used in the description of the various examples herein is for the purpose of describing particular examples only and is not intended to be limiting. As used in the description of various such examples and in the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

It should also be understood that, in the embodiments of the present application, the sequence number of each process does not mean that the execution sequence of each process should be determined by the function and the internal logic, and should not limit the implementation process of the embodiments of the present application.

It should also be understood that determining B from a does not mean determining B from a alone, but can also determine B from a and/or other information.

It will be further understood that the terms "Comprises" and/or "Comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.

It should also be understood that the term "if" may be interpreted to mean "when … …" ("white" or "upon") or "in response to a determination" or "in response to detection". Similarly, the phrase "if a determination … …" or "if a [ stated condition or event ] is detected" may be interpreted to mean "upon a determination … …" or "in response to a determination … …" or "upon a detection of a [ stated condition or event ]" or "in response to a detection of a [ stated condition or event ], depending on the context.

The following description of the terms involved in the present application.

Blockchain system: and may also be referred to as a data sharing system, referring to the data sharing system shown in fig. 1, where the data sharing system refers to a system for performing data sharing between nodes, where the data sharing system may include multiple nodes 101, where the multiple nodes 101 may be respective clients in the data sharing system, and the data sharing system may configure one or more blockchains to store information and data. Each node 101 may receive input information while operating normally and maintain shared data within the data sharing system based on the received input information.

In order to ensure the information intercommunication in the data sharing system, information connection can exist between every two nodes in the data sharing system, the nodes can transmit information through the information connection, the nodes can communicate and achieve trust, a digital signature technology is needed to be relied on, and identity confirmation, information authenticity and integrity verification are mainly achieved.

Artificial intelligence (Artificial Intelligence, AI) is the theory, method, technique and application system that uses a digital computer or a machine controlled by a digital computer to simulate, extend and extend human intelligence, sense the environment, acquire knowledge and use the knowledge to obtain optimal results. In other words, artificial intelligence is an integrated technology of computer science that attempts to understand the essence of intelligence and to produce a new intelligent machine that can react in a similar way to human intelligence. Artificial intelligence, i.e. research on design principles and implementation methods of various intelligent machines, enables the machines to have functions of sensing, reasoning and decision.

The artificial intelligence technology is a comprehensive subject, and relates to the technology with wide fields, namely the technology with a hardware level and the technology with a software level. Artificial intelligence infrastructure technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation/interaction systems, mechatronics, and the like. The artificial intelligence software technology mainly comprises a computer vision technology, a voice processing technology, a natural language processing technology, machine learning/deep learning and other directions.

Computer Vision (CV) is a science of studying how to "look" a machine, and more specifically, to replace human eyes with a camera and a Computer to perform machine Vision such as recognition, tracking and measurement on a target, and further perform graphic processing to make the Computer process into an image more suitable for human eyes to observe or transmit to an instrument to detect. As a scientific discipline, computer vision research-related theory and technology has attempted to build artificial intelligence systems that can acquire information from images or multidimensional data. Computer vision techniques typically include image processing, image recognition, image semantic understanding, image retrieval, OCR, video processing, video semantic understanding, video content/behavior recognition, three-dimensional object reconstruction, 3D techniques, virtual reality, augmented reality, synchronous positioning, and map construction, among others, as well as common biometric recognition techniques such as face recognition, fingerprint recognition, and others.

Machine Learning (ML) is a multi-domain interdisciplinary, involving multiple disciplines such as probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory, etc. It is specially studied how a computer simulates or implements learning behavior of a human to acquire new knowledge or skills, and reorganizes existing knowledge structures to continuously improve own performance. Machine learning is the core of artificial intelligence, a fundamental approach to letting computers have intelligence, which is applied throughout various areas of artificial intelligence. Machine learning and deep learning typically include techniques such as artificial neural networks, confidence networks, reinforcement learning, transfer learning, induction learning, teaching learning, and the like.

With research and advancement of artificial intelligence technology, research and application of artificial intelligence technology is being developed in various fields, such as common smart home, smart wearable devices, virtual assistants, smart speakers, smart marketing, unmanned, automatic driving, unmanned aerial vehicles, robots, smart medical treatment, smart customer service, etc., and it is believed that with the development of technology, artificial intelligence technology will be applied in more fields and with increasing importance value.

The scheme provided by the embodiment of the application relates to the technologies of computer vision, machine learning and the like of artificial intelligence, and is specifically described by the following embodiments:

the environment in which the present application is implemented is described below.

Fig. 2 is a schematic diagram of a bridge health monitoring system according to an embodiment of the present application. The bridge health monitoring system 200 comprises a data acquisition device 201, a state estimation device 202 and an alarm 203. The data acquisition device 201 is connected to the state estimation device 202 via a wireless network or a wired network. The state estimation device 202 is connected to the alarm 203 via a wireless network or a wired network.

In the embodiment of the application, the data acquisition device 201 is used for acquiring real-time monitoring data of a bridge, wherein the data acquisition device 201 comprises a plurality of sensors, a microwave radar, a monitoring device of a Beidou satellite navigation system and a bridge weighing system, wherein the plurality of sensors are installed at different structural positions on the bridge.

The state estimation device 202 is loaded with a state estimation model, and the state estimation device 202 is used for outputting the health state of the bridge based on real-time monitoring data processing.

The alarm device 203 is configured to send alarm information in response to the abnormal health state output by the state estimation model.

The state estimation device 202 and the alarm device may be terminals or servers, which are not limited thereto.

The terminal can be at least one of a smart phone, a game console, a desktop computer, a tablet computer, an electronic book reader, an MP3 (Moving Picture Experts Group Audio Layer III, dynamic video expert compression standard audio plane 3) player or an MP4 (Moving Picture Experts Group Audio Layer IV, dynamic video expert compression standard audio plane 4) player, a laptop portable computer. And the terminal is provided with and runs an application program supporting bridge health status monitoring.

The server may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDNs, basic cloud computing services such as big data and artificial intelligence platforms. The terminal can be, but is not limited to, a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, etc.

If the state estimation device 202 is a server, the state estimation device 202 may include at least one of a server, a plurality of servers, a cloud computing platform, and a virtualization center for providing background services for bridge health monitoring.

Optionally, the state estimation device 202 includes at least one server and a database, where the database is used to store data, and in an embodiment of the present application, the database can store real-time monitoring data collected by the data collection device, and provide a data service for the at least one server.

Alternatively, the state estimation device 202 may be any one or more servers in a blockchain system that may perform bridge health monitoring work alone or in concert.

Those skilled in the art will appreciate that the number of terminals and servers can be greater or fewer. For example, the number of the terminals and the servers can be only one, or the number of the terminals and the servers can be tens or hundreds, or more, and the number and the device types of the terminals and the servers are not limited in the embodiment of the application.

Fig. 3 is a flowchart of a bridge health monitoring method according to an embodiment of the present application, where the method is applied to an electronic device, and the electronic device is a terminal or a server, and referring to fig. 3, the method includes the following steps.

In the present embodiment, the plurality of sensors may include a plurality of types of sensors, and for example, the plurality of sensors may include at least two of a vibration sensor, a displacement sensor, a temperature sensor, a stress strain sensor, an acceleration sensor, and a wind speed and direction sensor. The sensors can be arranged at different structural positions on the bridge, and accordingly, the sensors can acquire displacement, acceleration, wind speed, wind direction, stress, strain, temperature and other data of the corresponding structural positions. The present application is not limited by the specific type of sensor that the various sensors include, but is merely illustrative of.

For example, the various sensors may be mounted in structural locations such as the midspan of the bridge main span, 1/4 span, at the support, at the top of the bridge tower, and the like.

The bridge is monitored in all directions through the sensors of the plurality of types, and the health state of the bridge is evaluated and analyzed from the aspects of deformation, temperature, strain, acceleration and the like, so that the monitoring of the health state of the bridge is ensured to be more accurate, and the condition that the monitoring is not in place in a single aspect is effectively avoided.

The sensors are mounted at different structural locations of the bridge, and can monitor multiple structural locations of the bridge. Once an abnormality occurs in a structural location, the abnormality is detected by a sensor mounted in the structural location and detected by subsequent analysis, thereby alerting. Therefore, relevant staff can receive the alarm in time and take corresponding measures to check and repair the alarm so as to ensure bridge health and personnel safety.

In some embodiments, the plurality of sensors may move according to the movement instructions or may move periodically and by themselves according to a predetermined trajectory, so that the plurality of sensors may switch the location where the data is collected, thereby knowing that the bridge is more comprehensive.

For example, the movement instructions may be generated by a remote operation by an associated worker and sent to the various sensors, which, upon receiving the movement instructions, may move in accordance with the movement instructions.

The various sensors can monitor the related conditions of the bridge in real time, and can send the real-time monitoring data to the electronic equipment after the real-time monitoring data are collected, so that the electronic equipment obtains the first real-time monitoring data collected by the various sensors. For example, the vibration sensor may collect the vibration velocity of the bridge. The displacement sensor can collect the displacement of a certain structural position of the bridge. The temperature sensor can collect the temperature of a certain structural position of the bridge. The stress strain sensor can collect strain of a certain structural position of the bridge. The acceleration sensor can collect acceleration of a certain structural position of the bridge. The wind speed and direction sensor can collect wind speed and direction at a certain structural position of the bridge, namely the environmental data of the bridge.

In some embodiments, electromagnetic wave signals can be transmitted from the radar to the bridge surface by frequency modulated continuous wave technology based on synthetic aperture radar (Synthetic Aperture Radar, SAR), and the signals are reflected back by the bridge surface and received by the microwave radar, so that the transmitted signals and the received signals can be determined at the microwave radar, and the transmitted signals and the received signals are the second real-time monitoring data.

Among them, synthetic Aperture Radar (SAR) is a high-resolution imaging radar that can obtain a high-resolution radar image like an optical photograph under meteorological conditions where visibility is extremely low. The radar with a larger equivalent antenna aperture is synthesized by a data processing method by utilizing the relative motion of the radar and the target, and the radar is also called as a synthetic aperture radar. The synthetic aperture radar has the characteristics of high resolution, all-weather operation and effective recognition of camouflage and penetration of a mask. The resulting high azimuth resolution is equivalent to that provided by a large aperture antenna. By combining the advantages of the microwave radar, the health state of the bridge can be monitored in real time with high precision through the microwave radar.

In some embodiments, the microwave radar may also monitor the second real-time monitoring data of the different locations by transmitting and receiving electromagnetic wave signals to and from the different locations of the bridge.

Based on the transmitting signal and the receiving signal, the deformation condition of the bridge can be obtained by carrying out subsequent processing on the signals, and the health state of the bridge can be further judged. How these signals are processed may be referred to in step 305 below, and will not be described in detail here.

The Beidou satellite navigation system (BeiDou Navigation Satellite System, BDS) is provided with a monitoring device, so that the bridge on the ground can be monitored. The monitoring device can adopt an RTK (Real-time kinematic) carrier phase difference technology to position the bridge with high precision, and can monitor the position change of the bridge with high precision by positioning the position of the bridge.

The RTK carrier phase difference technology is a difference method for processing the observed quantity of the carrier phases of two measuring stations in real time, and the carrier phases acquired by the reference station are sent to a user receiver to calculate the difference and calculate the coordinates. The satellite positioning measurement method is characterized in that the previous static, quick static and dynamic measurement needs to be solved afterwards to obtain the centimeter-level precision, the RTK is a measurement method capable of obtaining the centimeter-level positioning precision in the field in real time, the carrier phase dynamic real-time differential method is adopted, and the method is a great milestone for GPS application.

Specifically, a frequency modulation continuous wave signal is sent and received through a microwave frequency modulation technology, the frequency of the sent and received frequency modulation continuous wave is extracted, calculated and unwrapped through a microwave interferometry technology and a signal analysis and processing technology, dynamic deflection is monitored through a phase ranging technology, particularly fine position changes of a bridge can be obtained through monitoring through a monitoring device of a Beidou satellite navigation system, and therefore whether the bridge is healthy or not is determined.

The bridge weighing system can be one or more industrial cameras and lenses, has very high acquisition precision, and is suitable for monitoring the tiny abnormality of the bridge. The bridge weighing system may also be other cameras, for example, an infrared camera, etc., and the type of the bridge weighing system is not limited in the embodiment of the present application.

The bridge weighing system is arranged on a bridge to collect images, and can be used for collecting image information of the bridge and monitoring traffic flow information of the bridge in real time. For example, the traffic flow information may include axle weights, vehicle speeds, axle numbers, etc. of all lane vehicles, and specifically, those parameters to be monitored may be set by the relevant technicians, which is not limited by the embodiment of the present application.

For example, the bridge weighing system may be mounted in a guy cable, bridge arch, bridge body, etc. In particular, how to analyze the data of the bridge weighing system can be seen in step 305 described below, which is not described in detail herein.

In some embodiments, a laser device may also be mounted around the bridge weighing system for providing illumination for the bridge weighing system to capture images of the bridge to improve imaging quality.

In some embodiments, the bridge weighing system may determine whether a certain structure of the bridge is normal by periodically (periodically) acquiring images, comparing the images or analyzing the position of the certain structure of the bridge in the images. If abnormal, the subsequent alarm operation can be performed. In other embodiments, the bridge weighing system may periodically (periodically) collect traffic information from which the load of the bridge is determined to determine the health of the bridge. If the health state of the bridge is abnormal, the subsequent alarm operation can be performed.

After various real-time monitoring data are obtained, comprehensive analysis can be performed on the various real-time monitoring data to determine the health state of the bridge. In the embodiment of the application, the state estimation process can be realized through a state estimation model.

The state estimation model is briefly described below.

In some embodiments, the state estimation model may be trained in advance and stored in the electronic device, and may be recalled from local storage when the electronic device needs to evaluate bridge health. In other embodiments, the state estimation model may be trained in advance and stored in a database of a server, and when the electronic device needs to evaluate the health state of the bridge, the state estimation model may be called from the database of the server, or the obtained various real-time monitoring data may be sent to the server, and the server performs the state estimation step in step 305, and then returns the output health state of the bridge to the electronic device. The foregoing provides two possible embodiments, and the embodiment of the present application does not limit the storage address of the state estimation model.

In the embodiment of the application, the state estimation model comprises a plurality of sub-models, and one sub-model can process one type of real-time monitoring data to judge the health state of the bridge.

It will be appreciated that different types of real-time monitoring data are handled differently. For example, for image data, feature extraction is required for pixels of an image. For the vibration signal, the difference frequency differential processing is needed to be carried out on the vibration signal, and the processing of the two data by adopting the same sub-model is not possible to realize. The embodiment of the application provides a model fused by a plurality of sub-models, which can process various types of real-time monitoring data at the same time.

For the training process of the state estimation model, in some embodiments, model parameters of the plurality of sub-models in the state estimation model are obtained by performing joint training based on a first error and a second error, where the first error is an error between candidate discrimination results of the first sample monitoring data, the second sample monitoring data, the third sample monitoring data and the fourth sample monitoring data and the target health state, and the second error is an error between the predicted health state obtained based on the four sample monitoring data and the target health state.

The first sample monitoring data are monitoring data acquired by various sensors, the second sample monitoring data are monitoring data acquired by microwave radar measurement, the third sample monitoring data are monitoring data acquired by a monitoring device of the Beidou satellite navigation system, and the fourth sample monitoring data are monitoring data acquired by a bridge weighing system. The above four sample monitoring data are sample data, and the monitoring data collected before can be used as samples, or the sample monitoring data can be extracted from an open source database to perform model training.

Specifically, multiple types of sample monitoring data (including first sample monitoring data, second sample monitoring data, third sample monitoring data and fourth sample monitoring data) may be input into a state estimation model, where each type of sample monitoring data is input into a respective type of sub-model, the respective type of sample monitoring data is processed by the respective type of sub-model, and each sub-model outputs a candidate discrimination result of the respective type of sample monitoring data, where the candidate discrimination result is a prediction result obtained by performing state estimation on the sample monitoring data by the sub-model. Then determining a first error between the candidate discrimination results output by each sub-model and the target health state, and determining the predicted health state by the state estimation model according to the candidate discrimination results output by the four sub-models, so as to determine a second error between the predicted health state and the target health state. And then, based on the first error and the second error, optimizing and adjusting model parameters of the state estimation model until the model parameters meet the model training ending condition, stopping the training process, and obtaining the jointly trained state estimation model.

Parameters of multiple sub-models of the state estimation model are jointly trained, rather than simply placed in a large model after training of a single sub-model. The single sub-model training can only improve the accuracy of the single sub-model, a plurality of sub-models have no relation, and the accuracy of the final result cannot be ensured if the plurality of sub-models are used together. After the combined training, the multiple sub-models can show the effect of overlapping the training results of the multiple sub-models, and the overall accuracy of the multiple sub-models can be improved in the training process.

Wherein the first error is used to indicate the accuracy of the input real-time monitoring data processing by the single sub-model in the training process. The second error is used for indicating the accuracy of the input real-time monitoring data processing by the comprehensive multiple sub-models in the training process. And the first error and the second error are integrated, and parameters of all sub-models in the state estimation model are optimized, so that the performance of a single sub-model is improved, and the performance of an integral model after a plurality of sub-models are combined is also improved.

In some embodiments, the first error and the second error may be weighted and summed to obtain a third error, such that the parameter optimization is performed based on the third error. The weights of the first error and the second error may be set by a related technician according to requirements, or may be optimized together with model parameters in a model training process, which is not limited in the embodiment of the present application.

The automatic selection type function of the state estimation model is intelligent, which real-time monitoring data is input into which sub-model is not required to be set manually, and the state estimation model is used for carrying out type discrimination on the input real-time monitoring data so as to realize automatic selection, thereby realizing full automation.

Specifically, one field in the relevant fields of each real-time monitoring data is used for indicating the type of the data, which is referred to as a target field herein, and the relevant fields can be specifically set by relevant technicians according to requirements. In the embodiment of the application, the state estimation model can read the target field of the input real-time monitoring data to identify the type of the real-time monitoring data, and then input the real-time monitoring data into the sub-model of the corresponding type.

In some embodiments, step one may include at least one of, i.e., step one may include any one or more of, e.g., the following.

In the first case, the real-time monitoring data is input into the image state estimation sub-model in response to the state estimation model judging that the input real-time monitoring data is image data.

And in the second case, the real-time monitoring data is input into the temperature state estimation sub-model in response to the fact that the state estimation model judges that the input real-time monitoring data is temperature data.

And in the third case, the real-time monitoring data is input into the signal state estimation sub-model in response to the state estimation model judging that the input real-time monitoring data is signal data.

The foregoing provides several exemplary descriptions, and of course, the input real-time monitoring data may be other types of data, and further, the state estimation model may also correspond to a corresponding type of sub-model, where the sub-model in the state estimation model of the present application may be arbitrarily expanded into other types of data processing, for example, in response to determining, by the state estimation model, that the input real-time monitoring data is acceleration data, the real-time monitoring data is input into the acceleration state estimation sub-model, where embodiments of the present application are not limited herein.

In the first step, different types of real-time monitoring data are input into corresponding sub-models, and each sub-model can process the input real-time monitoring data to obtain candidate discrimination results. The sub-model is called as the candidate discrimination results because the sub-model only gives one discrimination, and finally the state estimation model can synthesize the candidate discrimination results of a plurality of sub-models to carry out overall discrimination.

Corresponding to the first step, when the types of the real-time monitoring data are different, the processing measures taken by the sub-model on the real-time monitoring data can be different. Specifically, the second step may include at least one of the following cases, that is, the first step may include any one or more of the following.

And 1, responding to the real-time monitoring data as image data, wherein the submodel is an image state estimation submodel, preprocessing the image data by the image state estimation submodel, and comparing the extracted image features with the big data image features of the bridge by extracting features of the preprocessed image data to obtain candidate discrimination results of the health state of the bridge.

In this case 1, preprocessing the image data may include processing steps such as image enhancement, image denoising, and image alignment.

The image enhancement is an image processing method which can make the original unclear image clear or emphasize some interesting features, inhibit the uninteresting features, improve the image quality, enrich the information quantity and strengthen the image interpretation and recognition effects.

The image enhancement process may be implemented by any image enhancement method, for example, a CLAHE (Contrast Limited Adaptive Histogram Equalization, limiting contrast adaptive histogram equalization) algorithm may be adopted, and the specific steps may be: dividing an image into a plurality of sub-blocks, counting pixels of each sub-block to obtain a histogram, setting an upper amplitude limit for the histogram, marking the upper amplitude limit as ClipLimit, summing parts higher than the upper amplitude limit in the histogram, and marking the result as total Excess. The total appearance is then averaged over all gray levels in the histogram and the height l=total appearance/N where N is the number of gray levels for this operation to cause the histogram to rise overall is found. Then, the difference upper=cliplimit-L between the Upper amplitude limit and the height L is obtained, and then, if the amplitude is higher than the ClipLimit, the difference upper=cliplimit-L is directly set as the ClipLimit by taking the ClipLimit and the Upper as boundaries. If the amplitude is between Upper and ClipLimit, filling the amplitude to ClipLimit. If the amplitude is lower than Upper, L pixel points are directly filled. The embodiment of the application does not limit the image enhancement process.

For image denoising, image denoising refers to the process of reducing noise in a digital image. In reality, digital images are often affected by interference of imaging equipment and external environment noise and the like in the processes of digitizing and transmitting, and are called noisy images or noise images.

The image denoising process can also be realized by adopting any image denoising algorithm, such as Gaussian filtering, median filtering, P-M equation denoising, TV method denoising and the like, for example, an image is divided into a plurality of sub-blocks, one of the sub-blocks is randomly selected, pixels of the sub-block are compared with pixels of other areas in the whole image, the similarity of the two is calculated, and the higher the similarity is, the larger the obtained weight is; if the similarity is below a certain threshold, the weight is zero; and weighting and summing the similar pixel values according to the normalized weights to obtain a denoised image. The embodiment of the application does not limit the image denoising process.

For image alignment, any image alignment algorithm may be used for image alignment, which is not limited by the embodiment of the present application. For example, a global image alignment method may be employed: extracting features of the image, extracting feature edges of the image by using an edge operator, and rotating the original feature image to form a series of feature images; and generating a plurality of positive and negative rotation characteristic images according to the preset rotation angle and precision configuration. And then, dimension reduction is carried out on the characteristic subarrays, 1/4 of the original image is taken by length and width, 1/16 of the original image is obtained, downsampling is carried out by a mean value method, and normalized cross correlation coefficients are obtained for all the characteristics of the characteristic subarrays, so that a coarse-granularity matching image is obtained. Then, adjacent characteristic images are found in a high-precision space to match the high-precision characteristic images, and then rotation translation and shearing of the images are carried out according to image matching results, so that aligned images are obtained.

For the preprocessed image, feature extraction can be performed to extract image features of the image, wherein the image features are extracted based on pixels and can represent the content of the image. And then, the extracted image features and the large data image features can be compared, and candidate discrimination results based on the large data of the image can be obtained.

The image features mainly comprise color features, texture features, shape features and space features of the image. The conventional feature extraction method is divided into two categories, namely feature extraction based on structural morphology and feature extraction based on geometric distribution, and of course, other feature extraction modes are also included, and the embodiment of the application is not particularly limited thereto, and can be specifically selected by relevant technicians according to requirements.

The comparison process is a matching process, and the condition that the extracted image features can be matched with the image features in the big data is the same or similar, the health state of the bridge is the same or similar.

And 2, responding to the real-time monitoring data as temperature data, wherein the sub-model is a temperature state estimation sub-model, obtaining temperature load data based on the temperature data by the temperature state estimation sub-model, and carrying out finite element analysis on the structure of the bridge based on the temperature load data to obtain a candidate discrimination result of the health state of the bridge.

In this case 2, after the temperature sensor collects the temperature data, the temperature sub-model performs preprocessing on the temperature data, and the preprocessing process may include denoising, where noise and obvious abnormal data in the bridge temperature data are removed in the denoising process, so as to ensure accuracy of subsequent calculation results. And then finishing the temperature data to obtain temperature load data, wherein the temperature load data comprises bridge integral temperature rise and fall load data and section temperature gradient load data.

The calculation mode of the whole bridge temperature rise and fall load data is as follows: setting a reference temperature, and calculating a temperature difference according to the bridge temperature data and the reference temperature to be used as the whole bridge temperature rise and fall load data. The calculation mode of bridge section temperature gradient load data is as follows: and converting the bridge temperature data measured by the temperature sensors at different positions of the bridge structure section into the bridge section temperature gradient load.

Finite element analysis (FEA, finite Element Analysis) uses a mathematical approximation method to simulate real physical systems (geometry and load conditions), among others. With simple and interactive elements, i.e. units, it is possible to approximate an infinite number of real systems with a finite number of unknowns. The finite element analysis can carry out real simulation on the structure of the bridge based on the temperature load data, so as to determine the health state of the bridge and realize accurate prediction.

And 3, responding to the real-time monitoring data as signal data, wherein the sub-model is a signal state estimation sub-model, and if the signal data is a vibration signal, performing feature extraction on the preprocessed vibration signal by the signal state estimation sub-model to obtain depth features, and identifying based on the depth features to obtain candidate discrimination results of the bridge health state.

In this case 3, the bridge health status feature information acquired by the vibration sensor is subjected to data preprocessing, the depth feature can be extracted by a depth feature extraction network, the depth feature extraction network can extract the depth feature information in the input vibration signal by a one-dimensional displacement convolution network, finally the depth feature is identified by an identification network, and then the status identification result of the bridge is output.

The depth feature extraction network may be comprised of a plurality of one-dimensional displacement convolution networks. In each one-dimensional displacement convolution module, an input signal is processed in two paths, one path uses displacement operation to carry out multichannel displacement processing on the input data so as to redistribute space information, then carries out point-by-point convolution processing to realize cross-channel mixed information, and batch normalization and nonlinear activation processing are needed on the data before point-by-point convolution. The other path fuses the characteristics of the first path output by carrying out average pooling and convolution operation on the input and by a characteristic addition connection mode, and the fused characteristics are used as the output of the one-dimensional convolution module.

The above case 3 is merely an exemplary illustration, and the processing procedure of the vibration signal may be implemented in other manners, which is not limited in the embodiment of the present application. For example, the vibration signal is intercepted by adopting two adjacent moving time windows to obtain a first window signal and a second window signal, the first window signal is input into a preset first feature extraction model to obtain a first feature vector, and the second window signal is input into a preset second feature extraction model to obtain a second feature vector. And then analyzing according to the first characteristic vector and the second characteristic vector to obtain a bridge safety state evaluation result, wherein the bridge safety state evaluation result is used for representing the safety state of the bridge. The preset first feature extraction model and the preset second feature extraction model are two feature extraction models with shared parameters, and are obtained based on the vibration signal training of the bridge.

And 4, responding to the real-time monitoring data as signal data, wherein the submodel is a signal state estimation submodel, if the signal data is an electromagnetic wave signal, performing difference frequency differential processing on the phases of the transmitted and received electromagnetic wave signals by the signal state estimation submodel, eliminating noise and correcting the atmosphere to obtain a radar line-of-sight deformation graph, converting the radar line-of-sight deformation into bridge longitudinal displacement based on the radar line-of-sight deformation graph, determining a shape variable, and determining a candidate discrimination result of the bridge health state based on the deformation quantity.

In this case 4, the microwave radar can record only phase information of less than the whole circumference in the target echo phase, i.e., the obtained phase is entangled. The interference phase of the two-phase radar images of the same target at different moments can be obtained by differential interference, as shown in a formula I:

；

wherein,for interference phase +.>For the phase component related to the actual displacement, < +.>Is the atmospheric phase component>Is a noise phase component>The whole-cycle ambiguities of the target point phase at two moments are respectively. After noise elimination and atmospheric correction, the deformation value deltar of the target in the LOS direction can be solved by the phase difference after unwrapping by referring to the following formula II:

；

wherein,is the wavelength.

In the identification and detection of the deformation of the ground SAR cracks, the influences of atmospheric refraction, shielding and the like usually exist, and the process can cause the measurement error of the radar vision line, so that the formula I is obtained、/>And the ki value is wrong, and a deformation calculation error is introduced. For->Correcting the influence of the error by adopting an error model; the influence of the ki value error is solved by adopting a PS method to process and a corresponding phase unwrapping method; for noise phase component->And (3) carrying out filtering treatment by adopting a model smooth spline function. For one input signal f (xi), a smooth spline can be used to realize low-pass filtering of the deformation time sequence of the foundation SAR point location in the time dimension, and the optimal estimation of deformation information is obtained by calculating the minimum value of the cost function J (lambda).

Case 5: responding to the real-time monitoring data as signal data, wherein the sub-model is a signal state estimation sub-model, if the signal data is a strain signal, calculating an initial strain stiffness representative value matrix under closed traffic, an initial strain stiffness representative value expected matrix under open traffic and a confidence interval by the signal state estimation sub-model according to the strain signal, comparing the current strain stiffness representative value expected matrix with the initial strain stiffness representative value expected matrix, if the current strain stiffness representative value expected matrix changes, performing joint investigation on the strain stiffness representative value confidence interval and the acquired strain signal, performing damage judgment after eliminating traffic modification interference, and determining candidate judging results of the bridge health state.

In this case 5, if the acquired signal is a strain signal, a candidate determination result may be obtained by processing the strain signal.

The calculation process of the initial strain stiffness representative value matrix under the closed traffic is as follows:

a. the standard vehicles sequentially pass through each lane at the same speed, the vehicles keep straight running at a constant speed, and the lanes of the vehicles cannot be changed in the test process; and recording the dynamic response conditions of each strain sensor of each key section and each reference section, wherein the time for collecting signals is at least 300 seconds(s).

b. And processing the strain time interval signal in the acquisition period by adopting a Fourier transform technology and an inverse Fourier transform technology to obtain a vehicle load static strain signal.

c. And calculating to obtain an initial strain stiffness representative value matrix.

Then, the acquired strain time-interval signal can be subjected to fast Fourier transformation to obtain a frequency domain signal, the amplitude of high frequency and low amplitude in the frequency domain signal is set to be zero, the vehicle load static strain signal is obtained by adopting inverse Fourier transformation, and then the ratio of the strain peak value of the reference section to the strain peak value corresponding to the monitoring section is taken as a rigidity representative value to form a strain rigidity representative value matrix of each section.

The calculation process of the initial strain stiffness representative value expected matrix and the confidence interval under open traffic is as follows:

1. and under the condition of open traffic, collecting monitoring data.

2. And carrying out zero point correction on the strain signal by adopting a moving average filtering technology, and eliminating environmental influence to obtain a vehicle load strain signal.

3. And further processing the vehicle load strain signal by adopting a Fourier transform technology and an inverse Fourier transform technology to obtain a vehicle load static strain signal.

4. And screening the obtained vehicle load static strain signals by using an automatic program to obtain strain rigidity representative value samples of all the measuring points.

5. And obtaining a strain stiffness representative value expected matrix and a strain stiffness representative value confidence interval by utilizing non-parameter estimation.

The processing method for the strain signal under the open traffic specifically comprises the following steps:

1. under open traffic, the data collected by the health monitoring system for a long time comprise environmental effect data and vehicle load effect data, the collected strain signals are processed by adopting a moving average filtering technology, and the window width is set to 900s, so that the strain signals of the environmental effect are obtained. Wherein, calculate the vehicle load strain signal CY: cy=zy-HY. Wherein ZY is the acquired strain signal, i.e. the total strain; HY is a strain signal of the environmental effect; the vehicle load strain signal CY is the difference between the total strain ZY and the environmental effect strain HY.

2. The vehicle load static strain signal adopts a rolling time window of 300s and utilizes the Fourier transform and inverse Fourier transform technology to obtain an accurate static strain signal.

3. Only one vehicle is ensured to pass through the bridge within a fixed time period, and an automatic program is adopted to screen data.

4. Setting the uplink direction and the downlink direction of the bridge as two nominal lanes, and referring to the calculation mode of the rigidity representative value under the closed traffic for the screened data meeting the conditions, calculating all strain rigidity representative value samples R of each measuring point in the period.

5. And calculating an expected value of R and a 95% confidence interval to obtain an expected matrix of the strain stiffness representative value and a confidence interval of the strain stiffness representative value.

The judging process for the bridge damage is as follows:

comparing the current strain stiffness representative value expected matrix with the initial strain stiffness representative value expected matrix, if the current strain stiffness representative value expected matrix changes, performing joint investigation on a confidence interval of the strain stiffness representative value and a video image, performing damage judgment after traffic modification interference is eliminated, if damage occurs, taking the damaged strain stiffness representative value expected matrix as an initial matrix, and performing follow-up monitoring by taking the current state as a reference; if the bridge structure is unchanged, the bridge structure is not damaged, and the bridge structure is directly monitored in the next period.

And if the candidate discrimination results are obtained in each sub-model, executing the third step, integrating a plurality of candidate discrimination results, and determining the health state of the bridge from a global angle.

In some embodiments, this third step may be implemented in any of the following two manners, which is not limited in the embodiment of the present application.

In the first mode, the health state abnormality of the bridge is determined and output in response to the candidate discrimination result of any one of the plurality of sub-models indicating the health state abnormality of the bridge; and responding to candidate discrimination results of all the sub-models in the plurality of sub-models to indicate that the health state of the bridge is normal, and determining and outputting that the health state of the bridge is normal.

In the first mode, the multiple sub-models analyze the health state of the bridge from different angles, and any sub-model determines that the bridge state is abnormal, which indicates that the bridge may have potential safety hazards, and the potential safety hazards need to be alarmed, so that relevant staff can check the potential safety hazards to ensure safety. And if the state of the bridge is normal, the plurality of sub-models are required to judge that the health state of the bridge is normal.

A second mode is that the number of the abnormal health states of the bridge is larger than or equal to the target number in response to the candidate discrimination results in the plurality of sub-models, and the abnormal health states of the bridge are determined and output; and determining and outputting that the health state of the bridge is normal in response to the candidate discrimination results in the plurality of sub-models indicating that the number of the health state anomalies of the bridge is smaller than the target number.

In the second mode, the plurality of sub-models analyze the health state of the bridge from different angles, and the target number can be set in consideration of the fault tolerance rate, and the health state of the bridge is determined through the target number.

The target number may be set by a person skilled in the relevant art as required, and the embodiment of the present application is not limited thereto.

The candidate discrimination results in the plurality of sub-models indicate that the number of the abnormal health states of the bridge is larger than or equal to the target number, and the fact that most of the sub-models consider the abnormal health states of the bridge is indicated, and then investigation and maintenance can be conducted through alarming. If the candidate discrimination results in the plurality of sub-models indicate that the number of the abnormal bridge health states is smaller than the target number, the condition that only the individual sub-models consider the abnormal bridge health states possibly has some abnormal values or environmental factors to cause erroneous discrimination is indicated, so that the condition can not be alarmed.

Through the steps 301 to 305, whether the health state of the bridge is normal or abnormal is estimated through the state estimation model, and if the health state of the bridge is abnormal, early warning information can be sent to remind related staff to conduct checking and repairing. If the health state of the bridge is normal, the health state can be ignored, and no early warning is performed.

In some embodiments, the obtained real-time monitoring data has time sequence, and the data volume is large, so that the real-time monitoring data can be stored through the blockchain system and then extracted from the blockchain when needed. The data reliability of the block chain and the fact that the data cannot be modified are considered, the fact that the data cannot be modified on a data storage block chain system is monitored in real time, and the accuracy and the reliability of the data can be guaranteed.

Specifically, in response to acquiring real-time monitoring data acquired at any moment, the electronic device generates a block based on the real-time monitoring data, and adds the block to a blockchain of the blockchain system when the block is commonly passed in the blockchain system. In this way, in the above steps 301 to 304, the process of acquiring the real-time monitoring data may be: in response to detecting a state estimation instruction of a certain period, the electronic device extracts real-time monitoring data in a plurality of blocks from the block chain based on block identifiers of the plurality of blocks corresponding to the period to obtain the first real-time monitoring data, the second real-time monitoring data, the third real-time monitoring data and the fourth real-time monitoring data.

The block can comprise the four real-time monitoring data, the data are collected at each moment, the block is generated based on the real-time monitoring data at the moment, and the block is added to the blockchain after the consensus passes.

The electronic device may be a node device in a blockchain system, and after the electronic device generates a block, the electronic device may broadcast the block to other node devices, so that the blockchain system may perform consensus on the first block. Of course, in another possible implementation, one or more node devices in the blockchain system may generate a block and broadcast the block to other node devices for consensus, which the embodiments of the present application do not limit.

The block generation process may be: the last block in the current blockchain is called the last block, that is, the last block is the block with the highest block height in the current blockchain, and the block being generated by the electronic device is called the next block. The electronic device can acquire all information of the previous block from the block chain, so that a block head characteristic value of the previous block can be generated based on all information of the previous block, and characteristic value calculation is carried out on real-time monitoring data to be stored in the next block to obtain a block main body characteristic value of the next block, and further, the electronic device can store the block head characteristic value of the previous block and the block main body characteristic value of the next block to the block head of the next block, store the real-time monitoring data (including version numbers, difficulty, time stamps and the like) to the block main body of the next block, so that the next block is generated, and the previous block and the next block are related through the block head characteristic value of the previous block, so that the purpose of block concatenation in the block chain can be realized. The foregoing is merely illustrative, and the generation process of the block may also be implemented in other manners, which are not limited by the embodiments of the present application.

In other embodiments, when using a blockchain system, the obtained real-time monitoring data may also be stored in the blockchain along with the results output by the state estimation model. Specifically, the electronic device may generate a first block based on the first real-time monitoring data, the second real-time monitoring data, the third real-time monitoring data, and the fourth real-time monitoring data, and add the first block to a blockchain of a blockchain system when the first block passes through in common in the blockchain system. The electronic device may generate a second block based on the health status of the bridge and the association between the health status and the real-time monitoring data in the first block output by the status estimation model, and add the second block to a blockchain of the blockchain system when the second block is commonly passed in the blockchain system. And finally, the electronic equipment correspondingly stores the block identifications of the first block and the second block in the block association relation table. The generation of the first block and the second block may refer to the generation of the next block in the previous embodiment, which is just an exemplary illustration, and the embodiment of the present application is not limited thereto.

In other embodiments, when using a blockchain system, the obtained real-time monitoring data may also be stored in the blockchain along with the results output by the state estimation model. Specifically, the electronic device may generate a third block based on the first real-time monitoring data, the second real-time monitoring data, the third real-time monitoring data, the fourth real-time monitoring data, and the health status of the bridge output by the status estimation model, and add the first block to a blockchain of the blockchain system when the third block is commonly passed in the blockchain system. The generation of the third block may be referred to as the generation of the next block in the previous embodiment, which is only an exemplary description, and the embodiment of the present application is not limited thereto.

All the above optional solutions can be combined to form an optional embodiment of the present application, and will not be described in detail herein.

Fig. 4 is a schematic structural diagram of a bridge health status monitoring device according to an embodiment of the present application, referring to fig. 4, the device includes:

the acquisition module 401 is configured to acquire first real-time monitoring data acquired by a plurality of sensors, where the plurality of sensors are installed at different structural positions on the bridge, and are configured to acquire different types of data of the bridge;

the acquisition module 401 is configured to acquire second real-time monitoring data that is monitored by performing dynamic high-precision measurement on deflection by using the microwave radar;

the acquisition module 401 is configured to acquire third real-time monitoring data acquired by a monitoring device of the beidou satellite navigation system;

the acquiring module 401 is configured to acquire fourth real-time monitoring data acquired by a bridge weighing system, where the bridge weighing system is installed on the bridge;

the state estimation module 402 is configured to input the first real-time monitoring data, the second real-time monitoring data, the third real-time monitoring data, and the fourth real-time monitoring data into a state estimation model, determine a type of the input real-time monitoring data by the state estimation model, input each real-time monitoring data into a state estimation sub-model corresponding to the type, process the real-time monitoring data of the corresponding type based on each sub-model, obtain a candidate discrimination result of the health state of the bridge, and determine and output the health state of the bridge based on the candidate discrimination results of the plurality of sub-models;

And the sending module 403 is configured to send early warning information in response to the abnormal health status of the bridge.

In some embodiments, the state estimation module 402 is configured to perform at least one of:

in response to determining by the state estimation model that the input real-time monitoring data is image data, inputting the real-time monitoring data into an image state estimation sub-model;

responding to the real-time monitoring data as image data, wherein the submodel is an image state estimation submodel, preprocessing the image data by the image state estimation submodel, and comparing the extracted image characteristics with the big data image characteristics of the bridge by extracting the characteristics of the preprocessed image data to obtain candidate discrimination results of the health state of the bridge;

Responding to the real-time monitoring data as temperature data, wherein the sub-model is a temperature state estimation sub-model, obtaining temperature load data based on the temperature data by the temperature state estimation sub-model, and carrying out finite element analysis on the structure of the bridge based on the temperature load data to obtain a candidate discrimination result of the health state of the bridge;

responding to the real-time monitoring data as signal data, wherein the sub-model is a signal state estimation sub-model, if the signal data is a vibration signal, performing feature extraction on the preprocessed vibration signal by the signal state estimation sub-model to obtain depth features, and identifying based on the depth features to obtain candidate discrimination results of the bridge health state;

responding to the real-time monitoring data as signal data, wherein the submodel is a signal state estimation submodel, if the signal data is an electromagnetic wave signal, performing difference frequency differential processing, noise elimination and atmosphere correction on the phase of the transmitted and received electromagnetic wave signal by the signal state estimation submodel to obtain a radar line-of-sight deformation graph, converting the radar line-of-sight deformation into bridge longitudinal displacement based on the radar line-of-sight deformation graph, determining a shape variable, and determining a candidate discrimination result of the bridge health state based on the deformation quantity;

；

In some embodiments, the state estimation module 402 is configured to perform any of the following:

In some embodiments, the model parameters of the plurality of sub-models in the state estimation model are obtained by joint training based on a first error and a second error, the first error is an error between candidate discrimination results of the first sample monitoring data, the second sample monitoring data, the third sample monitoring data and the fourth sample monitoring data and the target health state, and the second error is an error between the predicted health state and the target health state obtained based on the four sample monitoring data.

In some embodiments, the apparatus further comprises:

a first adding module for adding the block to a blockchain of the blockchain system when the block is commonly passed in the blockchain system;

the obtaining module 401 is configured to extract real-time monitoring data in a plurality of blocks from the blockchain based on block identifiers of the plurality of blocks corresponding to a certain period in response to detecting a state estimation instruction of the certain period, and obtain the first real-time monitoring data, the second real-time monitoring data, the third real-time monitoring data and the fourth real-time monitoring data.

In some embodiments, the apparatus further comprises:

a second adding module for adding the first block to a blockchain of the blockchain system when the first block is commonly passed in the blockchain system;

The second adding module is used for adding the second block into a blockchain of the blockchain system when the second block is commonly passed in the blockchain system;

According to the device provided by the embodiment of the application, various real-time monitoring data are acquired by arranging various real-time monitoring data acquisition equipment, the real-time monitoring data are processed, so that the health state of the bridge is judged, compared with manual monitoring, the manual operation is not needed, the labor cost is completely saved, the health state of the bridge can be timely acquired through automatic equipment, the processing efficiency and timeliness are improved, in addition, the types of the real-time monitoring data can be automatically screened through one large model, the sub-model of the corresponding type is selected for data processing, various types of monitoring data can be fused, the health state of the bridge is comprehensively analyzed, the process is fully automated, human errors are avoided, and the processing efficiency and accuracy of the monitoring result are greatly improved.

It should be noted that: in the bridge health monitoring device provided in the above embodiment, only the division of the above functional modules is used for illustration, and in practical application, the above functional allocation can be completed by different functional modules according to needs, that is, the internal structure of the bridge health monitoring device is divided into different functional modules, so as to complete all or part of the functions described above. In addition, the bridge health status monitoring device provided in the above embodiment and the bridge health status monitoring method embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which is not described herein again.

Fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present application, where the electronic device 500 may have a relatively large difference due to different configurations or performances, and may include one or more processors (Central Processing Units, CPU) 501 and one or more memories 502, where at least one computer program is stored in the memories 502, and the at least one computer program is loaded and executed by the processors 501 to implement the bridge health monitoring method provided in the above method embodiments. The electronic device can also include other components for implementing device functions, for example, the electronic device can also have wired or wireless network interfaces, input-output interfaces, and the like for input-output. The embodiments of the present application are not described herein.

In an exemplary embodiment, a computer program product or a computer program is also provided, the computer program product or the computer program comprising one or more program codes, the one or more program codes being stored in a computer readable storage medium. The one or more processors of the electronic device are capable of reading the one or more program codes from the computer-readable storage medium, the one or more processors executing the one or more program codes such that the electronic device is capable of performing the bridge health monitoring method described above.

Those of ordinary skill in the art will appreciate that all or a portion of the steps implementing the above-described embodiments can be implemented by hardware, or can be implemented by a program instructing the relevant hardware, and the program can be stored in a computer readable storage medium, and the above-mentioned storage medium can be a read-only memory, a magnetic disk or an optical disk, etc.

Claims

1. A method for monitoring the health status of a bridge, the method comprising:

2. The method according to claim 1, wherein the determining, by the state estimation model, the type of the input real-time monitoring data, and inputting each real-time monitoring data into the state estimation sub-model corresponding to the type, includes at least one of:

3. The method according to claim 1 or 2, wherein the processing the real-time monitoring data of the corresponding type based on each sub-model to obtain the candidate discrimination result of the bridge health status comprises at least one of the following:

4. A method according to claim 3, wherein the difference frequency differential processing is achieved by the following formula one:

；

5. A method according to claim 3, wherein, after the noise cancellation and the atmospheric correction, the determining of the deformation value of the radar line of sight deformation map is achieved by the following formula two:

；

6. The method according to claim 1 or 2, wherein the determining and outputting the health status of the bridge based on the candidate discrimination results of the plurality of sub-models includes any one of the following:

7. The method of claim 1, wherein model parameters of the plurality of sub-models in the state estimation model are jointly trained based on a first error and a second error, the first error being an error between candidate discrimination results of first sample monitoring data, second sample monitoring data, third sample monitoring data and fourth sample monitoring data and a target health state, the second error being an error between a predicted health state obtained based on the four sample monitoring data and the target health state.

8. A bridge health monitoring device, the device comprising:

the bridge weighing system is arranged at different positions of the bridge;

9. The bridge health state monitoring system is characterized by comprising data acquisition equipment, state estimation equipment and an alarm device; the data acquisition equipment is connected with the state estimation equipment through a network, and the state estimation equipment is connected with the alarm device through the network;

10. An electronic device comprising one or more processors and one or more memories, the one or more memories having stored therein at least one computer program loaded and executed by the one or more processors to implement the bridge health monitoring method of any of claims 1-7.