Intelligent analysis method and system for bridge state

Abstract

The application provides a bridge state intelligent analysis method and system, and relates to the technical field of bridge state analysis.The method comprises the following steps: reading target bridge characteristics of a target bridge; obtaining a state analysis model; obtaining a target initial state index; preprocessing a target bridge video to obtain a target dynamic image; obtaining a real-time structure dynamic index of the target bridge; and dynamically adjusting the target initial state index based on the real-time structure dynamic index to obtain a target real-time state index of the target bridge.The method solves the technical problems that the prior art can only detect bridge state indexes periodically or rely on subjective experience of technical personnel to determine the bridge state, cannot monitor the bridge state information in real time, and cannot accurately analyze the bridge state, and thus affects the safe and stable operation of the bridge.The method achieves the technical effect of accurately and intelligently analyzing the bridge state and improves the accuracy of bridge state analysis.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of bridge state analysis, and particularly relates to an intelligent analysis method and system for bridge state. BACKGROUND

[0002] Bridge state analysis is to check and detect the whole and each component part of the existing bridge, then determine the bearing capacity and defects of the bridge through theoretical analysis to obtain the current health state of the bridge, so as to facilitate the staff to take relevant measures to maintain the bridge in time. Maintenance is the main means to ensure the bridge in normal working state. With the rapid development of bridge construction, bridge monitoring becomes particularly important. However, bridge monitoring has not been widely used in bridge maintenance work. The existing bridge state monitoring method mostly subjectively evaluates the bearing of the bridge by manual work, lacks dynamic change analysis of the bridge state, and thus leads to inaccurate bridge state analysis results, insufficient state analysis in time, and difficulty in assisting the staff to take measures to maintain the bridge in time, which affects the service life of the bridge.

[0003] In summary, the prior art only judges the bridge state by periodically detecting the bridge state index or the subjective experience of the technical personnel, cannot real-time monitor the bridge state information, and has the technical problems of inaccurate bridge state analysis, lack of generalization, and influence on safe and stable operation of the bridge. SUMMARY

[0004] The present application provides an intelligent analysis method and system for bridge state, which solves the technical problems that the prior art only judges the bridge state by periodically detecting the bridge state index or the subjective experience of the technical personnel, cannot real-time monitor the bridge state information, and has the technical problems of inaccurate bridge state analysis, lack of generalization, and influence on safe and stable operation of the bridge.

[0005] According to a first aspect of the present application, a method for intelligent analysis of bridge state is provided, comprising: reading target bridge features of a target bridge, wherein the target bridge features include target structure features and target material features; collecting sample structure features and sample material features of a plurality of bridge samples, matching a plurality of sample bridge state index identifiers, combining as construction data, and using the construction data for supervised training, verification and testing to obtain a state analysis model; analyzing the target structure features and target material features through the state analysis model to obtain target bridge state analysis results, and taking the target bridge state analysis results as target initial state indexes; real-time monitoring the target bridge through an intelligent monitoring device to obtain target bridge videos, and pre-processing the target bridge videos through a dynamic load identification model to obtain target dynamic load images, wherein the target dynamic load images include first images and second images; analyzing the first images and the second images to obtain target dynamic load parameters, and analyzing the target dynamic load parameters through a structure dynamic prediction model to obtain real-time structure dynamic indexes of the target bridge; and dynamically adjusting the target initial state indexes based on the real-time structure dynamic indexes to obtain target real-time state indexes of the target bridge.

[0006] According to a second aspect of the present application, an intelligent analysis system for bridge state is provided, comprising: a bridge feature reading module for reading target bridge features of a target bridge, wherein the target bridge features include target structure features and target material features; a state analysis model obtaining module for collecting sample structure features and sample material features of a plurality of bridge samples, matching a plurality of sample bridge state index identifiers, combining as construction data, and using the construction data for supervised training, verification and testing to obtain a state analysis model; a bridge state analysis module for analyzing the target structure features and target material features through the state analysis model to obtain target bridge state analysis results, and taking the target bridge state analysis results as target initial state indexes; a dynamic load identification module for real-time monitoring the target bridge through an intelligent monitoring device to obtain target bridge videos, and pre-processing the target bridge videos through a dynamic load identification model to obtain target dynamic load images, wherein the target dynamic load images include first images and second images; a dynamic prediction module for analyzing the first images and the second images to obtain target dynamic load parameters, and analyzing the target dynamic load parameters through a structure dynamic prediction model to obtain real-time structure dynamic indexes of the target bridge; and a state index adjusting module for dynamically adjusting the target initial state indexes based on the real-time structure dynamic indexes to obtain target real-time state indexes of the target bridge.

[0007] According to the intelligent analysis method of bridge state adopted by the present application, based on the above analysis, the present application provides an intelligent analysis method of bridge state. In the present embodiment, first, the target bridge is analyzed according to the target structure characteristics and the target material characteristics to obtain a target initial state index, so as to achieve the effect of static analysis of the bridge state and preliminary acquisition of the bridge state. Further, a dynamic load identification model is constructed based on the SLOWFAST network, the target bridge video is analyzed to obtain a first image and a second image, and then the target dynamic load parameter is obtained by analyzing the first image and the second image, and the real-time structure dynamic index of the target bridge is obtained by analyzing the target dynamic load parameter through the structure dynamic prediction model, so as to realize real-time analysis of the dynamic performance of the bridge, thereby improving the accuracy and effectiveness of the bridge state analysis. Finally, the target initial state index is dynamically adjusted based on the real-time structure dynamic index, so as to realize dynamic correction of the bridge state analysis result, thereby achieving the technical effect of improving the accuracy and effectiveness of the bridge state analysis.

[0008] It should be understood that the content described in this part is not intended to identify key or important features of the embodiments of the present application, nor is it used to limit the scope of the present application. Other features of the present application will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only exemplary, and other drawings can be obtained by those skilled in the art without creative labor on the basis of the provided drawings.

[0010] Figure 1 A flowchart of an intelligent analysis method of bridge state provided by the embodiment of the present application is shown in the figure.

[0011] Figure 2 A flowchart of obtaining target structure characteristics in the embodiment of the present application is shown in the figure.

[0012] Figure 3 A flowchart of obtaining target dynamic load image in the embodiment of the present application is shown in the figure.

[0013] Figure 4 A structural diagram of an intelligent analysis system of bridge state provided by the embodiment of the present application is shown in the figure.

[0014] Explanation of reference numerals in the attached diagram: Bridge feature reading module 11, State analysis model acquisition module 12, Bridge state analysis module 13, Dynamic load identification module 14, Dynamic prediction module 15, State index adjustment module 16. Detailed Implementation

[0015] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of the present invention, including various details to aid understanding. These details should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0016] To address the technical problems of existing technologies that rely solely on periodic inspections of bridge condition indicators or subjective experience of technicians to determine bridge condition, which result in the inability to obtain real-time bridge condition information and inaccurate and non-specific analysis of bridge condition, thus affecting the safe and stable operation of bridges, the inventors of this invention have, through creative work, developed an intelligent bridge condition analysis method and system.

[0017] Example 1

[0018] Figure 1 This invention provides an intelligent analysis method for bridge status, comprising:

[0019] Step S100: Read the target bridge features of the target bridge, wherein the target bridge features include target structural features and target material features;

[0020] Among them, such as Figure 2 As shown, step S100 in this embodiment of the invention further includes:

[0021] Step S110: Read the target superstructure of the target bridge, wherein the target superstructure includes a main beam or a main arch rib;

[0022] Step S120: Read the target substructure of the target bridge, wherein the target substructure includes piers, abutments, and piles;

[0023] Step S130: Read the target ancillary structures of the target bridge, wherein the target ancillary structures include bridge deck pavement, sidewalk, curbstone, railing, and expansion joint;

[0024] Step S140: Combine the upper structure of the target, the lower structure of the target, and the auxiliary structure of the target to obtain the target structural features.

[0025] In this embodiment of the invention, step S150 includes:

[0026] Step S151: extracting a first structure in the target structure feature;

[0027] Step S152: reading a first structure material of the first structure;

[0028] Step S153: performing performance measurement on the first structure material based on a preset material detection scheme, to obtain a first structure material performance;

[0029] Step S154: obtaining the target material feature based on the first structure material performance.

[0030] Specifically, the target bridge is any type of bridge to be analyzed, such as a typical bridge on a highway. The target superstructure of the target bridge is read, wherein the target superstructure includes a main beam or a main arch rib. The main beam is the main load-bearing structure of a beam bridge, and the arch rib is the main structure of an arch bridge. Since current bridges mainly include beam bridges and arch bridges, of course, if the target bridge is of other types, the target superstructure can be replaced or adjusted according to the main load-bearing structure, which is not limited herein. The target substructure of the target bridge is read, wherein the target substructure includes a pier, an abutment, and a pile. Similarly, the target substructure is the structure that contacts the ground and plays a load-bearing role, including a pier, an abutment, and a pile. The target accessory structure of the target bridge is read, which is an evaluation index of the bridge load and quality except for the target superstructure and the target substructure, including bridge deck pavement, sidewalk, curbstone, railing, and expansion joint. It should be noted that the target superstructure, the target substructure, and the target accessory structure can be adjusted according to actual conditions, which is not limited herein. Finally, the target superstructure, the target substructure, and the target accessory structure are combined to obtain the target structure feature, which provides basic data for subsequent bridge state analysis.

[0031] Specifically, after obtaining the target structure feature, a first structure in the target structure feature is extracted, which is any one of the target superstructure, the target substructure, and the target accessory structure, such as a main beam, a pier, and the like. The first structure material, such as steel bars, cement, and concrete, is read. Further, the performance of the first structure material is measured based on a preset material detection scheme, to obtain a first structure material performance. The preset material detection scheme refers to a scheme for detecting the performance of the first structure material, which is preset by the staff, such as strength detection of concrete and the like, and corrosion evaluation and measurement of steel bars and the like. Finally, the first structure material performance is taken as the target material feature.

[0032] The target bridge feature is composed of the target structure feature and the target material feature.

[0033] Step S200: Collecting a plurality of sample structure features and a plurality of sample material features of a plurality of bridge samples, and matching a plurality of sample bridge state index labels, combining as construction data, and using the construction data for supervised training, verification and testing to obtain a state analysis model;

[0034] In the embodiment of the present application, step S200 further comprises:

[0035] Step S210: Extracting a first bridge sample in the plurality of bridge samples, and matching a first sample structure feature and a first sample material feature of the first bridge sample;

[0036] The first sample structure feature and the first sample material feature have a first corresponding relationship;

[0037] Step S220: Extracting a first sample structure in the first sample structure feature, and obtaining a first sample structure material and a first sample structure material performance of the first sample structure in combination with the first corresponding relationship;

[0038] Step S230: Standardizing the first sample structure material performance, and weighting to obtain a first sample bridge state index;

[0039] Step S240: Obtaining the plurality of sample bridge state index labels based on the first sample bridge state index label.

[0040] Specifically, based on big data collection of multiple sample bridge structure characteristics and multiple sample material characteristics, and matching multiple sample bridge state index identifiers, the multiple sample bridge state index identifiers are the state indexes of the multiple bridge samples, and the multiple sample bridge structure characteristics and the multiple sample material characteristics and the multiple sample bridge state index identifiers have a one-to-one correspondence relationship, which are combined as construction data, and the construction data is used for supervised training, verification and testing to obtain a state analysis model. Specifically, the construction data is divided into training data, verification data and test data according to a certain proportion (which can be set by yourself), the training data includes multiple sample structure characteristics and multiple sample material characteristics, the verification data includes the state indexes of multiple bridge samples corresponding to the multiple sample structure characteristics and the multiple sample material characteristics in the training data, and the test data includes multiple groups of one-to-one corresponding sample structure characteristics, sample material characteristics and sample bridge state index identifiers. Based on the neural network model, the network structure of the state analysis model is built, each group of sample structure characteristics and sample material characteristics in the training data is input into the state analysis model, the state indexes of multiple bridge samples in the verification data are used to supervise and adjust the output of the state analysis model, so that the output of the state analysis model is consistent with the state indexes of the multiple bridge samples. After all the training data is trained, the state analysis model is tested for accuracy through the test data, so as to obtain the state analysis model with required accuracy, which provides a basis for subsequent target bridge state analysis.

[0041] Specifically, the process of matching multiple sample bridge state index identifiers is as follows: extracting a first bridge sample from the multiple bridge samples, the first bridge sample being any bridge sample in the multiple bridge samples, matching to obtain first sample structure characteristics and first sample material characteristics of the first bridge sample, wherein the first sample structure characteristics and the first sample material characteristics have a first corresponding relationship, extracting a first sample structure from the first sample structure characteristics, and combining the first corresponding relationship to obtain a first sample structure material and a first sample structure material performance of the first sample structure. The first sample structure material performance refers to the material performance of any structure of any bridge sample, such as the strength of the steel bars on the main beam, the hardness of the bridge pier made of cement, the corrosion degree, etc. The first sample structure material performance is standardized, the main function of standardization is to eliminate the dimensional relationship between data, so that the data has comparability. The first sample structure material performance after standardization is weighted to obtain a first sample bridge state index, and the first sample bridge state index is marked on the corresponding bridge sample to obtain the multiple sample bridge state index identifiers, which provide data support for the training and testing of the state analysis model.

[0042] Step S300: analyzing the target structure characteristics and target material characteristics through the state analysis model to obtain a target bridge state analysis result, and taking the target bridge state analysis result as a target initial state index;

[0043] Specifically, the target structure characteristics and target material characteristics are input into the constructed state analysis model, and a target bridge state analysis result is output, which is taken as a target initial state index.

[0044] Step S400: real-time monitoring of the target bridge through an intelligent monitoring device to obtain a target bridge video, and pre-processing of the target bridge video through a dynamic load identification model to obtain a target dynamic load image, wherein the target dynamic load image includes a first image and a second image;

[0045] As shown in FIG. 4, the step S400 includes: Figure 3

[0046] Step S410: the dynamic load identification model includes a dynamic load identification fast branch and a dynamic load identification slow branch;

[0047] Step S420: the dynamic load identification fast branch extracts the target bridge video based on a first predetermined frequency to obtain a first bridge image time sequence;

[0048] Step S430: the dynamic load identification slow branch extracts the target bridge video based on a second predetermined frequency to obtain a second bridge image time sequence;

[0049] The first predetermined frequency is less than the second predetermined frequency.

[0050] Step S440: based on the first bridge image time sequence and the second bridge image time sequence, the target dynamic load image is obtained.

[0051] Specifically, the above-mentioned intelligent monitoring device is a device for real-time monitoring of the target bridge, such as an intelligent camera, which has the functions of automatic acquisition and automatic transmission. Through the intelligent monitoring device, the target bridge video is obtained by real-time monitoring of the target bridge, and the target dynamic load image is obtained by pre-processing of the target bridge video through the dynamic load identification model, wherein the target dynamic load image includes a first image and a second image. The dynamic load identification model is a convolutional neural network model based on a SLOWFAST network, which includes a fast (Fast) channel and a slow (Slow) channel, and the first image and the second image are obtained based on the fast (Fast) channel and the slow (Slow) channel.

[0052] ​Specifically, the dynamic load identification model comprises a dynamic load identification fast branch and a dynamic load identification slow branch, which are two parallel convolutional neural networks. The dynamic load identification fast branch analyzes the static content in the target bridge video through a slow high-resolution convolutional neural network Slow channel, and the dynamic load identification slow branch analyzes the dynamic content in the target bridge video using a fast low-resolution convolutional neural network (Fast channel). Specifically, the dynamic load identification fast branch extracts the target bridge video based on a first predetermined frequency to obtain a first bridge image time sequence, and the dynamic load identification slow branch extracts the target bridge video based on a second predetermined frequency to obtain a second bridge image time sequence. The first predetermined frequency and the second predetermined frequency can be set according to actual conditions. The first predetermined frequency is less than the second predetermined frequency. For example, the first predetermined frequency can be set to 2 frames of images collected per second, and the second predetermined frequency is 15 frames of images collected per second. Based on the first bridge image time sequence and the second bridge image time sequence, the first image is obtained by image extraction of the target bridge video according to the first predetermined frequency. The second image is obtained by image extraction processing of the target bridge video according to the second predetermined frequency, and the image extraction result is down-sampled to reduce the image resolution to generate a thumbnail corresponding to the image. The operation method can be realized by using existing technologies. The first image and the second image are obtained by constructing a dynamic load identification model based on a SLOWFAST network, so as to improve the accuracy of bridge state analysis.

[0053] Step S500: analyzing the first image and the second image to obtain a target dynamic load parameter, and analyzing the target dynamic load parameter through a structural dynamic prediction model to obtain a real-time structural dynamic index of the target bridge;

[0054] In the embodiment of the present application, step S500 further comprises:

[0055] Step S510: the first image has a first time identifier, and the second image has a second time identifier;

[0056] Step S520: comparing the first time identifier and the second time identifier to obtain an interval duration;

[0057] Step S530: identifying a first vehicle in the first image based on image processing technology, and obtaining a position of the first vehicle in the first image, denoted as a first vehicle position;

[0058] Step S540: obtaining a position of the first vehicle in the second image, denoted as a second vehicle position;

[0059] Step S550: obtaining a position interval by comparing the first vehicle position and the second vehicle position;

[0060] Step S560: calculating a first speed of the first vehicle by combining the interval duration and the position interval;

[0061] Step S570: constructing a vehicle database based on big data and obtaining a first vehicle body weight of the first vehicle by traversal;

[0062] Step S580: taking the first speed and the first vehicle body weight as the target dynamic load parameter.

[0063] In the embodiment of the present application, step S590 comprises:

[0064] Step S591: constructing a historical operation data set based on historical operation record data of similar bridges in operation, wherein the historical operation data set comprises a historical speed, a historical vehicle body weight and a historical structure dynamic index;

[0065] Step S592: constructing the structure dynamic prediction model according to the historical operation data set.

[0066] In the embodiment of the present application, step S591 comprises:

[0067] Step S5911: constructing a structure dynamic index set, wherein the structure dynamic index set comprises a dynamic characteristic model parameter and a dynamic response index parameter;

[0068] The dynamic characteristic model parameter comprises a bridge frequency, a bridge vibration shape and a bridge damping ratio.

[0069] The dynamic response index parameter comprises a bridge dynamic deflection, a bridge dynamic stress and an impact coefficient.

[0070] Specifically, the target initial state index obtained in step S300 is a static state parameter of the target bridge, that is, the initial state of the target bridge. However, as time changes, the target bridge is subjected to wind and rain, vehicle passing and other wear and tear, and its state may change. Therefore, in step S400, the target bridge is video collected, and a first image and a second image are obtained. The target dynamic load parameter is obtained by analyzing the first image and the second image, and the real-time structure dynamic index of the target bridge is obtained by analyzing the target dynamic load parameter through the structure dynamic prediction model. The real-time structure dynamic index represents the structure dynamic performance of the target bridge, and the structure dynamic performance is an important index for judging the operation condition and carrying capacity of the bridge.

[0071] The process of analyzing the first image and the second image to obtain the target dynamic load parameter is as follows: the collection time is marked when the video is collected, so that the first image has a first time identifier, the second image has a second time identifier, and the interval duration is obtained by comparing the first time identifier and the second time identifier. Based on image processing technology, a first vehicle in the first image is identified, and if the first image contains multiple vehicles, the first vehicle refers to any vehicle in the first image. The position of the first vehicle in the first image is obtained, denoted as the first vehicle position, which is the position of the first vehicle on the target bridge at the first time identifier. The collection times of the first image and the second image are different, and the position of the first vehicle has changed. The position of the first vehicle in the second image is obtained, denoted as the second vehicle position, which is the position of the first vehicle on the target bridge at the second time identifier. The position interval is obtained by comparing the first vehicle position and the second vehicle position. The interval duration and the position interval are combined, and the first speed of the first vehicle is obtained by dividing the position interval by the interval duration. Based on big data, a vehicle database is constructed, which contains the vehicle body weight data of different types of vehicles, and the vehicle type and vehicle weight data have a corresponding relationship. The first vehicle type is traversed in the vehicle database to obtain the first vehicle body weight of the first vehicle. The first speed and the first vehicle body weight are taken as the target dynamic load parameter, which provides data basis for subsequent structural dynamic performance analysis.

[0072] Before analyzing the target dynamic load parameter by the structural dynamic prediction model to obtain the real-time structural dynamic index of the target bridge, a structural dynamic prediction model needs to be constructed, and the construction process is as follows: based on the historical operation record data of the same type of bridge as the target bridge during operation, a historical operation data set is established, wherein the historical operation data set includes historical speed, historical vehicle body weight and historical structural dynamic index, and the historical speed, historical vehicle body weight and historical structural dynamic index have a one-to-one correspondence. The structural dynamic prediction model is constructed according to the historical operation data set. The input of the structural dynamic prediction model is vehicle speed and vehicle body weight, and the output is structural dynamic index. The structural dynamic prediction model is a neural network model in machine learning. The historical speed and the historical vehicle body weight are input into the structural dynamic prediction model, and the output result of the structural dynamic prediction model is supervised and adjusted through the historical structural dynamic index. The structural dynamic prediction model is trained to convergence, and then the target dynamic load parameter (i.e. the first speed and the first vehicle body weight) is input into the structural dynamic prediction model, and the real-time structural dynamic index is output.

[0073] Before the historical operation data set is established, a structural dynamic index set needs to be established, the structural dynamic index set includes multiple evaluation indexes for evaluating the structural dynamic performance of the bridge, and can be divided into dynamic characteristic model parameters and dynamic response index parameters, wherein the dynamic characteristic model parameters include bridge frequency, bridge vibration shape, bridge damping ratio, and the dynamic response index parameters include bridge dynamic deflection, bridge dynamic stress and impact coefficient.

[0074] Step S600: dynamically adjusting the target initial state index based on the real-time structural dynamic index to obtain a target real-time state index of the target bridge.

[0075] Specifically, the real-time structural dynamic index includes real-time bridge frequency, real-time bridge vibration shape, real-time bridge damping ratio, real-time bridge dynamic deflection, real-time bridge dynamic stress and real-time impact coefficient, and the real-time structural dynamic index reflects the real-time use performance and bearing capacity of the target bridge. With the change of use time, the real-time structural dynamic index is also dynamically changed, and the target initial state index needs to be dynamically corrected by the real-time structural dynamic index, so as to ensure the accuracy of the bridge state.

[0076] Based on the above analysis, the present application provides an intelligent analysis method for bridge state. In the embodiment, first, the target bridge is analyzed according to the target structural characteristics and the target material characteristics to obtain a target initial state index, so as to achieve the effect of static analysis of the bridge state and preliminary acquisition of the bridge state. Further, a dynamic load identification model is constructed based on the SLOWFAST network, the target bridge video is analyzed to obtain a first image and a second image, and then the target dynamic load parameter is obtained by analyzing the first image and the second image, and the real-time structural dynamic index of the target bridge is obtained by analyzing the target dynamic load parameter through the structural dynamic prediction model, so as to realize real-time analysis of the dynamic performance of the bridge, thereby improving the accuracy and effectiveness of the bridge state analysis. Finally, the target initial state index is dynamically adjusted based on the real-time structural dynamic index, so as to realize dynamic correction of the bridge state analysis result, thereby achieving the technical effect of improving the accuracy and effectiveness of the bridge state analysis.

[0077] Embodiment Two

[0078] Based on the same inventive concept as the intelligent analysis method for bridge state in the foregoing embodiments, as Figure 4As shown, the present application also provides an intelligent analysis system for bridge state, which comprises:

[0079] a bridge feature reading module 11 for reading target bridge features of a target bridge, wherein the target bridge features comprise target structural features and target material features;

[0080] a state analysis model acquisition module 12 for collecting sample structural features and sample material features of a plurality of bridge samples, and matching a plurality of sample bridge state index identifiers, combining as construction data, and using the construction data for supervised training, verification and testing to obtain a state analysis model;

[0081] a bridge state analysis module 13 for analyzing the target structural features and target material features through the state analysis model to obtain a target bridge state analysis result, and taking the target bridge state analysis result as a target initial state index;

[0082] a dynamic load identification module 14 for real-time monitoring of the target bridge through intelligent monitoring equipment to obtain a target bridge video, and pre-processing the target bridge video through a dynamic load identification model to obtain a target dynamic load image, wherein the target dynamic load image comprises a first image and a second image;

[0083] a dynamic force prediction module 15 for analyzing the first image and the second image to obtain a target dynamic load parameter, and analyzing the target dynamic load parameter through a structural dynamic force prediction model to obtain a real-time structural dynamic force index of the target bridge;

[0084] a state index adjustment module 16 for dynamically adjusting the target initial state index based on the real-time structural dynamic force index to obtain a target real-time state index of the target bridge.

[0085] Further, the bridge feature reading module 11 is also used for:

[0086] reading a target superstructure of the target bridge, wherein the target superstructure comprises a main beam or a main arch rib;

[0087] reading a target substructure of the target bridge, wherein the target substructure comprises a pier, an abutment, and a pile;

[0088] reading a target accessory structure of the target bridge, wherein the target accessory structure comprises bridge deck pavement, sidewalk, curb, railing, and expansion joint;

[0089] Combining the target superstructure, the target substructure and the target accessory structure to obtain the target structure feature.

[0090] Further, the bridge feature reading module 11 is further used for:

[0091] Extracting a first structure in the target structure feature;

[0092] Reading a first structure material of the first structure;

[0093] Performing performance measurement on the first structure material based on a preset material detection scheme to obtain a first structure material performance;

[0094] Obtaining the target material feature based on the first structure material performance.

[0095] Further, the state analysis model obtaining module 12 is further used for:

[0096] Extracting a first bridge sample in the plurality of bridge samples, and matching a first sample structure feature and a first sample material feature of the first bridge sample;

[0097] Wherein, the first sample structure feature and the first sample material feature have a first corresponding relationship;

[0098] Extracting a first sample structure in the first sample structure feature, and obtaining a first sample structure material and a first sample structure material performance of the first sample structure in combination with the first corresponding relationship;

[0099] Standardizing the first sample structure material performance, and weighting to obtain a first sample bridge state index;

[0100] Marking the plurality of sample bridge state indexes based on the first sample bridge state index.

[0101] Further, the dynamic load identification module 14 is further used for:

[0102] The dynamic load identification model includes a dynamic load identification fast branch and a dynamic load identification slow branch;

[0103] The dynamic load identification fast branch extracts the target bridge video based on a first predetermined frequency to obtain a first bridge image time sequence;

[0104] The dynamic load identification slow branch extracts the target bridge video based on a second predetermined frequency to obtain a second bridge image time sequence;

[0105] Wherein, the first predetermined frequency is less than the second predetermined frequency;

[0106] Based on the first bridge image time sequence and the second bridge image time sequence, the target dynamic load image is obtained.

[0107] Further, the power prediction module 15 is further used for:

[0108] The first image has a first time identifier, and the second image has a second time identifier;

[0109] The interval duration is obtained by comparing the first time identifier and the second time identifier;

[0110] A first vehicle in the first image is identified based on image processing technology, and a position of the first vehicle in the first image is obtained, denoted as a first vehicle position;

[0111] A position of the first vehicle in the second image is obtained, denoted as a second vehicle position;

[0112] The position interval is obtained by comparing the first vehicle position and the second vehicle position;

[0113] The first speed of the first vehicle is calculated based on the interval duration and the position interval;

[0114] A vehicle database is constructed based on big data, and a first vehicle body weight of the first vehicle is obtained by traversal;

[0115] The first speed and the first vehicle body weight are taken as the target dynamic load parameters.

[0116] Further, the power prediction module 15 is further used for:

[0117] Based on historical operation record data of the same type of bridge as the target bridge during operation, a historical operation data set is established, wherein the historical operation data set includes historical speed, historical vehicle body weight, and historical structure power index;

[0118] The structure power prediction model is constructed according to the historical operation data set.

[0119] Further, the power prediction module 15 is further used for:

[0120] A structure power index set is established, wherein the structure power index set includes power characteristic model parameters and power response index parameters;

[0121] The power characteristic model parameters include bridge frequency, bridge vibration shape, and bridge damping ratio;

[0122] The power response index parameters include bridge dynamic deflection, bridge dynamic stress, and impact coefficient.

[0123] The intelligent analysis method of the bridge state in the first preceding embodiment is also applicable to the intelligent analysis system of the bridge state in the present embodiment. Those skilled in the art can clearly understand the intelligent analysis system of the bridge state in the present embodiment through the detailed description of the intelligent analysis method of the bridge state. Therefore, for the sake of brevity of the description, no further detailed description is given herein.

[0124] It should be understood that the various forms of flow shown above can be reordered, steps added or deleted. For example, the steps described in the present application can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in the present application can be achieved, which is not limited herein.

[0125] The above detailed description does not constitute a limitation on the scope of protection of the present application. Those skilled in the art should understand that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A method of intelligent analysis of the state of a bridge, characterized in that, The method comprises the following steps: reading target bridge features of a target bridge, wherein the target bridge features comprise target structure features and target material features; collecting a plurality of sample structure features and a plurality of sample material features of a plurality of bridge samples, and matching a plurality of sample bridge state index identifiers, combining them as construction data, and using the construction data for supervised training, verification and testing to obtain a state analysis model; analyzing the target structure features and the target material features through the state analysis model to obtain target bridge state analysis results, and taking the target bridge state analysis results as a target initial state index; real-time monitoring the target bridge through an intelligent monitoring device to obtain a target bridge video, and pre-processing the target bridge video through a dynamic load identification model to obtain a target dynamic load image, wherein the target dynamic load image comprises a first image and a second image; analyzing the first image and the second image to obtain target dynamic load parameters, and analyzing the target dynamic load parameters through a structure dynamic prediction model to obtain a real-time structure dynamic index of the target bridge; dynamically adjusting the target initial state index based on the real-time structure dynamic index to obtain a target real-time state index of the target bridge; the pre-processing of the target bridge video through the dynamic load identification model to obtain the target dynamic load image comprises: the dynamic load identification model comprises a dynamic load identification fast branch and a dynamic load identification slow branch; the dynamic load identification fast branch extracts the target bridge video based on a first predetermined frequency to obtain a first bridge image time sequence; the dynamic load identification slow branch extracts the target bridge video based on a second predetermined frequency to obtain a second bridge image time sequence; wherein the first predetermined frequency is less than the second predetermined frequency; the target dynamic load image is obtained based on the first bridge image time sequence and the second bridge image time sequence.

2. The intelligent analysis method of claim 1, wherein, The reading of the target bridge features of the target bridge comprises: reading a target superstructure of the target bridge, wherein the target superstructure comprises a main beam or a main arch rib; reading a target substructure of the target bridge, wherein the target substructure comprises a pier, an abutment, a pile; reading a target accessory structure of the target bridge, wherein the target accessory structure comprises a bridge deck pavement, a sidewalk, a curb, a railing, a expansion joint; combining the target superstructure, the target substructure and the target accessory structure to obtain the target structure features.

3. The method of claim 2, wherein the step of analyzing the data comprises the step of: After obtaining the target structure features, the method comprises the following steps: extracting a first structure in the target structure features; reading a first structure material of the first structure; performing performance determination on the first structure material based on a preset material detection scheme to obtain a first structure material performance; obtaining the target material features based on the first structure material performance.

4. The intelligent analysis method of claim 1, wherein, The matching of the plurality of sample bridge state index identifiers comprises: extracting a first bridge sample in the plurality of bridge samples, and matching a first sample structure feature and a first sample material feature of the first bridge sample; wherein the first sample structure feature and the first sample material feature have a first corresponding relationship; extracting a first sample structure in the first sample structure feature, and obtaining a first sample structure material and a first sample structure material performance of the first sample structure in combination with the first correspondence relationship; standardizing the first sample structure material performance, and weighting to obtain a first sample bridge state index; labeling the first sample bridge state index based on the first sample bridge state index to obtain a plurality of sample bridge state index labels.

5. The intelligent analysis method of claim 1, wherein, After the target dynamic load image is obtained, the method further comprises: The first image has a first time identifier, and the second image has a second time identifier; comparing the first time identifier with the second time identifier to obtain an interval duration; identifying a first vehicle in the first image based on image processing technology, and obtaining a position of the first vehicle in the first image, denoted as a first vehicle position; obtaining a position of the first vehicle in the second image, denoted as a second vehicle position; comparing the first vehicle position with the second vehicle position to obtain a position interval; combining the interval duration and the position interval to calculate a first speed of the first vehicle; constructing a vehicle database based on big data, and traversing to obtain a first vehicle body weight of the first vehicle; taking the first speed and the first vehicle body weight as the target dynamic load parameter.

6. The intelligent analysis method of claim 5, wherein, Before the target dynamic load parameter is analyzed by the structure dynamic prediction model to obtain the real-time structure dynamic index of the target bridge, the method further comprises: based on historical operation record data of the same type of bridges as the target bridge during operation, a historical operation data set is established, wherein the historical operation data set includes historical speed, historical vehicle body weight and historical structure dynamic index; constructing the structure dynamic prediction model according to the historical operation data set.

7. The intelligent analysis method of claim 6, wherein, Before the historical operation data set is established, the method further comprises: establishing a structure dynamic index set, wherein the structure dynamic index set includes dynamic characteristic model parameters and dynamic response index parameters; wherein the dynamic characteristic model parameters include bridge frequency, bridge vibration shape and bridge damping ratio; wherein the dynamic response index parameters include bridge dynamic deflection, bridge dynamic stress and impact coefficient.

8. An intelligent analysis system for bridge condition, characterized by, The system is used to execute the method of any one of claims 1-7, and the system comprises: a bridge feature reading module, which is used to read target bridge features of a target bridge, wherein the target bridge features include target structure features and target material features; a state analysis model acquisition module, which is used to collect a plurality of sample structure features and a plurality of sample material features of a plurality of bridge samples, and match a plurality of sample bridge state index labels, combine as construction data, and use the construction data for supervised training, verification and testing to obtain a state analysis model; a bridge state analysis module, which is used to analyze the target structure features and the target material features by the state analysis model to obtain a target bridge state analysis result, and take the target bridge state analysis result as a target initial state index; The dynamic load identification module is configured to monitor the target bridge in real time by an intelligent monitoring device, obtain a target bridge video, and preprocess the target bridge video by a dynamic load identification model to obtain a target dynamic load image, wherein the target dynamic load image includes a first image and a second image; The dynamic force prediction module is configured to analyze the first image and the second image to obtain a target dynamic load parameter, and analyze the target dynamic load parameter by a structural dynamic force prediction model to obtain a real-time structural dynamic force index of the target bridge; The state index adjustment module is configured to dynamically adjust the target initial state index based on the real-time structural dynamic force index to obtain a target real-time state index of the target bridge.

Images