Tunnel digital intelligent management and maintenance method and system

Abstract

The invention relates to the technical field of tunnel intelligent management and maintenance, in particular to a tunnel digital intelligent management and maintenance method and system, and the method comprises the steps: building a tunnel real-time model through the collection of tunnel geometric representation and tunnel monitoring information, carrying out the recognition of license plates and vehicle types of passing vehicles according to the collected historical traffic data, and obtaining a real-time model; the method comprises the steps of obtaining high-risk vehicle information, performing road risk detection on a tunnel road, obtaining high-risk road section information, integrating the high-risk vehicle information and the high-risk road section information, obtaining a traffic feature set, performing disease prediction based on the traffic feature set and a tunnel real-time model, and preventing tunnel diseases based on a disease prediction result. According to the invention, the problem that the traditional tunnel disease monitoring method cannot accurately identify high-risk vehicles and road sections in real time is effectively solved, the high-risk vehicles and road sections can be accurately identified by integrating the tunnel real-time model and the traffic feature set, and the accuracy of tunnel disease prediction is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent tunnel management and maintenance, and in particular to a digital intelligent tunnel management and maintenance method and system. Background Art

[0002] In the field of modern tunnel disease monitoring, the rapid response capability of tunnels is one of the key factors in improving the efficiency of tunnel health management and maintenance. With the changes in tunnel operations and traffic flow, the real-time response requirements for tunnel disease prediction models are increasing. Traditional monitoring methods often face the problem of response delays, which leads to the inability to effectively guarantee the accuracy and timeliness of disease prediction when facing complex traffic conditions and changes in tunnel environment.

[0003] Although existing technologies have made certain progress in tunnel disease monitoring and prediction, they are still insufficient in terms of rapid response capabilities. Many traditional methods rely on static data analysis and standardized parameters. This means that when faced with sudden traffic loads, transient loads or tunnel structure changes, the model's response speed cannot meet the needs of real-time monitoring and prediction, which in turn affects the accuracy and timeliness of disease warnings. In a highly dynamic tunnel operation environment, tunnel models need to respond to variable factors such as traffic flow, vehicle type, and road section risks in a very short time, and traditional static monitoring methods cannot meet this demand.

[0004] The information disclosed in this background technology section is only intended to deepen the understanding of the overall background technology of the present disclosure, and should not be regarded as acknowledging or suggesting in any form that the information constitutes the prior art known to those skilled in the art. Summary of the invention

[0005] The present invention provides a tunnel digital intelligent management and maintenance method and system, which can effectively solve the problems in the background technology.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is: A digital intelligent tunnel maintenance method, the method comprising: Collecting tunnel geometry representation and tunnel monitoring information, and building a tunnel real-time model according to the tunnel geometry representation and the tunnel monitoring information; Collect historical traffic data, identify the vehicle license plate and vehicle type of passing vehicles, obtain vehicle attribute information, and identify high-risk vehicles based on the vehicle attribute information and the historical traffic data; Perform risk monitoring on the tunnel section to obtain section risk information, identify high-risk sections based on the section risk information and the historical traffic data, integrate the vehicle attribute information and the section risk information, and obtain a traffic feature set; Tunnel diseases are predicted based on the tunnel real-time model and the traffic feature set, disease prediction results are obtained, and tunnel diseases are prevented according to the disease prediction results.

[0007] Furthermore, collecting tunnel structure information and tunnel monitoring information, and constructing a tunnel real-time model according to the tunnel structure information and the tunnel monitoring information, including: Scanning the tunnel using three-dimensional data acquisition technology to obtain point cloud data of the tunnel surface, generating a geometric representation of the tunnel based on the point cloud data, and monitoring the tunnel in real time to obtain tunnel monitoring information; Performing adaptive disease recognition on the tunnel monitoring information, identifying tunnel disease characteristics, and obtaining tunnel disease information; A real-time model of the tunnel is created based on the tunnel geometry representation, the tunnel monitoring information and the tunnel disease information.

[0008] Furthermore, the tunnel detection information is adaptively identified to identify tunnel disease information, including: Establishing a tunnel disease database, extracting disease features from the tunnel disease database using image processing technology, and obtaining tunnel disease features; A disease recognition model is constructed using machine learning technology as a framework, and deep learning of tunnel disease characteristics is performed according to the disease recognition model to identify the tunnel disease characteristics; The disease recognition model is used to identify the disease characteristics of the tunnel detection information to obtain tunnel disease information.

[0009] Further identifying the vehicle license plate and vehicle type of the passing vehicle, obtaining vehicle attribute information, and identifying high-risk vehicles based on the vehicle attribute information and the historical traffic data, including: Traversing the public transportation database, collecting vehicle image data according to the imaging device, identifying the vehicle license plate based on the vehicle image data, and identifying the vehicle type according to the public transportation database and the image recognition algorithm to obtain vehicle attribute information; According to the historical traffic data, the historical vehicle type judgment standard is collected, and according to the vehicle attribute information and the historical vehicle type judgment standard, it is judged whether the vehicle type is a high-risk vehicle. If so, the vehicle is marked and the vehicle is tracked and monitored in a focused manner.

[0010] Further, risk monitoring is performed on the tunnel section to obtain section risk information, and high-risk sections are identified based on the section risk information and the historical traffic data, including: Dividing the tunnel into sections according to the real-time model of the tunnel, and equally dividing the tunnel into a number of sections; Retrieving historical traffic data based on the historical traffic data, and setting a road section vehicle passage threshold and a vehicle speed threshold for the plurality of road sections respectively according to the historical traffic data; monitoring the number of vehicles and the speed of vehicles on each of the road sections in real time, and making a one-to-one correspondence between the number of vehicles and the speed of vehicles on each of the road sections and a plurality of the road sections; The number of vehicles and the speed of vehicles on each road section are compared with the vehicle traffic threshold and the vehicle speed threshold respectively. If the number of vehicles or the speed of vehicles on a road section exceeds the corresponding threshold, it is marked as a high-risk road section.

[0011] Furthermore, a tunnel disease database is established, including: Collecting historical tunnel disease information, classifying and managing tunnel disease types according to the historical tunnel disease information, obtaining classification management results, and constructing a disease index according to the classification management results; The data collection time sequence length is set, the tunnel disease database is regularly updated based on the time sequence length, and tunnel disease information is continuously imported and exported.

[0012] Further, predicting tunnel defects based on the tunnel real-time model and the traffic feature set to obtain defect prediction results includes: Dividing the traffic feature set into a disease training set, a disease test set and a disease verification set; The tunnel real-time model is trained using deep learning technology according to the disease training set, and the trained tunnel real-time model is used to predict the disease of the disease test set to obtain a preliminary prediction result; The preliminary prediction results are analyzed for accuracy using the disease verification set to obtain accuracy analysis results, and the tunnel real-time model is optimized based on the accuracy analysis results to obtain final prediction results.

[0013] Further, the preliminary prediction result is evaluated for accuracy using the disease verification set to obtain an accuracy evaluation result, and the tunnel real-time model is optimized based on the accuracy evaluation result to obtain a final prediction result, including: Recording the actual error between the preliminary prediction result and the actual disease, analyzing the actual error, extracting relevant disease features that produce the actual error, and marking the relevant disease features as high-error prediction features; The high error prediction feature is used as a training incremental feature to perform specialized training on the tunnel real-time model to obtain a specialized model, and a final prediction result is generated based on the specialized model.

[0014] A digital intelligent tunnel management and maintenance system, the system comprising: Collecting tunnel geometry representation and tunnel monitoring information, and building a tunnel real-time model according to the tunnel geometry representation and the tunnel monitoring information; Collect historical traffic data, identify the vehicle license plate and vehicle type of passing vehicles, obtain vehicle attribute information, and identify high-risk vehicles based on the vehicle attribute information and the historical traffic data; Perform risk monitoring on the tunnel section to obtain section risk information, identify high-risk sections based on the section risk information and the historical traffic data, integrate the vehicle attribute information and the section risk information, and obtain a traffic feature set; Tunnel diseases are predicted based on the tunnel real-time model and the traffic feature set, disease prediction results are obtained, and tunnel diseases are prevented according to the disease prediction results.

[0015] Furthermore, the road section risk identification module includes: A tunnel section division unit, which divides the tunnel into sections according to the real-time tunnel model, and divides the tunnel into a plurality of sections; A road section threshold setting unit, which retrieves historical traffic data based on the historical traffic data, and sets road section vehicle passage thresholds and vehicle speed thresholds for the plurality of road sections respectively according to the historical traffic data; A road section flow monitoring unit monitors the number of vehicles and the speed of vehicles on each road section in real time, and associates the number of vehicles and the speed of vehicles on each road section with a plurality of road sections respectively; The risk section identification unit compares the number of vehicles and the vehicle speed of each section with the vehicle passage threshold and the vehicle speed threshold. If the number of vehicles or the vehicle speed of a section exceeds the corresponding threshold, it is marked as a high-risk section.

[0016] The technical solution of the present invention can achieve the following technical effects: It effectively solves the problem that traditional tunnel disease monitoring methods cannot accurately identify high-risk vehicles and road sections in real time. By integrating the tunnel real-time model and traffic feature set, it can accurately identify high-risk vehicles and road sections, effectively improving the accuracy of tunnel disease prediction.

[0017] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 This is a flowchart of the digital intelligent tunnel maintenance method; Figure 2 This is a schematic diagram of the road section risk identification process; Figure 3 This is the structural diagram of the tunnel digital intelligent management and maintenance system. DETAILED DESCRIPTION

[0020] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0022] Embodiment 1 like Figure 1 As shown, the present application provides a tunnel digital intelligent management and maintenance method, the method comprising: S10: Collecting tunnel geometry representation and tunnel monitoring information, and building a tunnel real-time model based on the tunnel geometry representation and tunnel monitoring information; S20: Collect historical traffic data, identify the vehicle license plate and vehicle type of the passing vehicle, obtain vehicle attribute information, and identify high-risk vehicles based on the vehicle attribute information and historical traffic data; S30: Perform risk monitoring on the tunnel section, obtain section risk information, identify high-risk sections based on the section risk information and historical traffic data, integrate vehicle attribute information and section risk information, and obtain a traffic feature set; S40: Predict tunnel diseases based on the tunnel real-time model and the traffic feature set, obtain disease prediction results, and prevent tunnel diseases according to the disease prediction results.

[0023] Specifically, first, various sensors and measuring devices are used to collect geometric surface feature information of the tunnel, including the length, width, curvature and structural materials of the tunnel. At the same time, tunnel monitoring information is collected in real time, such as tunnel internal environmental parameters, structural stress data, crack monitoring data, etc. A real-time tunnel model is constructed using the collected tunnel geometry representation and tunnel monitoring information. The model can reflect the physical state and possible structural changes of the tunnel in real time. The next step is to monitor passing vehicles in real time, identify vehicle license plates and vehicle types, thereby obtaining vehicle attribute information, analyzing the collected vehicle attribute information and historical traffic data, and identifying High-risk vehicles, such as heavily loaded vehicles and vehicles that frequently enter and exit, are identified. Then, risk monitoring is carried out on each section of the tunnel to evaluate the safety status of each section. Section risk information, such as road wear and structural deformation data, is collected. Historical traffic data is used to further analyze possible risks. The collected vehicle attribute information and section risk information are integrated into a traffic feature set. Based on the constructed real-time tunnel model and the traffic feature set, machine learning and data analysis techniques can be used to predict possible tunnel defects. According to the defect prediction results, corresponding tunnel maintenance and reinforcement measures can be formulated to prevent or mitigate the development of defects.

[0024] It effectively solves the problem that traditional tunnel disease monitoring methods cannot accurately identify high-risk vehicles and road sections in real time. By integrating the tunnel real-time model and traffic feature set, it can accurately identify high-risk vehicles and road sections, and effectively improve the accuracy of tunnel disease prediction.

[0025] Specifically, the tunnel structure information and tunnel monitoring information are collected, and a tunnel real-time model is constructed based on the tunnel structure information and tunnel monitoring information, including: Use 3D data acquisition technology to scan the tunnel, obtain point cloud data of the tunnel surface, generate tunnel geometry representation based on the point cloud data, and monitor the tunnel in real time to obtain tunnel monitoring information; Adaptively identify tunnel damage information, identify tunnel damage characteristics, and obtain tunnel damage information; A real-time tunnel model is created based on tunnel geometry representation, tunnel monitoring information, and tunnel disease information.

[0026] As a preferred embodiment of the above embodiment, first, a three-dimensional data acquisition technology such as laser scanning or structured light scanning technology is used to comprehensively scan the tunnel. During this process, the scanning device will move in the tunnel and continuously capture point cloud data on the tunnel surface. These point cloud data can accurately reflect the current geometric shape of the tunnel and form a detailed geometric representation of the tunnel. According to the acquired point cloud data, these point clouds are converted into a detailed three-dimensional tunnel model through point cloud processing software such as GeoModeler. This model not only includes the geometric features of the tunnel, but can also be updated in real time to reflect the latest status of the tunnel structure. At the same time, tunnel monitoring information is collected based on various sensors such as strain gauges, temperature sensors and vibration sensors. Using image processing and machine learning technology, the collected monitoring data is deeply analyzed to identify the characteristics of tunnel diseases, such as cracks, water seepage and leakage, deformation and corrosion, etc., so as to obtain tunnel disease information, and then the tunnel geometric representation, tunnel monitoring information and tunnel disease information are integrated into a comprehensive tunnel real-time model to obtain a tunnel real-time model.

[0027] Furthermore, adaptive recognition is performed on tunnel detection information to identify tunnel disease information, including: Establish a tunnel disease database, use image processing technology to extract disease features from the tunnel disease database, and obtain tunnel disease features; Machine learning technology is used as a framework to build a disease recognition model. Deep learning of tunnel disease characteristics is performed based on the disease recognition model to identify tunnel disease characteristics. The disease recognition model is used to identify the disease characteristics of tunnel detection information and obtain tunnel disease information.

[0028] In this embodiment, first, a database containing a large number of tunnel disease images is created, each image is marked with a detailed disease description and classification, such as cracks, water seepage, corrosion, etc., and image processing technology such as edge detection is used to extract key disease features from the tunnel disease database. This may include edge detection, texture analysis, and color segmentation methods to accurately capture the visual identification of each disease. A disease recognition model is constructed using a machine learning framework such as deep learning. The model uses the disease features as input and can recognize new tunnel disease images after training. The model training process includes multiple iterations to continuously optimize the recognition accuracy and reduce misidentification. The trained disease recognition model is used to identify new tunnel detection information such as real-time images and videos. The model can automatically identify and label disease features, thereby quickly and accurately obtaining tunnel disease information, which is helpful for tunnel maintenance decision support and ensures safe tunnel operation.

[0029] Specifically, the license plate and vehicle type of the passing vehicle are identified to obtain vehicle attribute information. Based on the vehicle attribute information and historical traffic data, high-risk vehicles are identified, including: Traverse the public transportation database, collect vehicle image data according to the imaging equipment, identify the vehicle license plate based on the vehicle image data, and identify the vehicle type according to the public transportation database and image recognition algorithm to obtain vehicle attribute information; According to the historical traffic data, the historical vehicle type judgment standard is collected, and according to the vehicle attribute information and the historical vehicle type judgment standard, it is judged whether the vehicle type is a high-risk vehicle. If so, the vehicle is marked and tracked and monitored in a focused manner.

[0030] Specifically, smart cameras deployed in the tunnel are used to collect continuous images of passing vehicles. These imaging devices are used to capture panoramic images of vehicles, and through image processing software and image recognition algorithms, the vehicle license plate and model are identified from the collected vehicle image data. License plate recognition can be performed on the license plate through optical character recognition technology to accurately extract license plate information. The vehicle type can be determined by building a deep learning model and training the deep learning model using the public transportation database. The vehicle type is determined by matching the vehicle's appearance features such as size, shape, and roof structure with the data in the trained deep learning model, including 2-axle buses, 2-axle trucks, 3-axle trucks, 4-axle trucks, 5-axle trucks, and 6-axle and above trucks. The identified license plate and vehicle model information is then integrated to form vehicle attribute information. The public transportation database is traversed to collect historical vehicle type judgment standards related to the vehicle type. These standards include risk ratings of vehicle types. For example, heavy vehicles and vehicles transporting flammable materials are classified as higher risk levels. Combining the vehicle attribute information and historical vehicle type judgment standards, it is determined whether the currently identified vehicle is a high-risk vehicle. This judgment is based on the vehicle type and its potential risk contribution. For example, certain vehicle models have shown a higher accident rate in past accidents. Once a vehicle is marked as a high-risk vehicle, a key monitoring and tracking procedure will be initiated, which may include enhanced real-time monitoring, regular inspections, and preventive measures in specific circumstances, such as restricting its passage at specific times or sections of the road.

[0031] Furthermore, if Figure 2 As shown, risk monitoring is performed on tunnel sections to obtain section risk information. Based on the section risk information and historical traffic data, high-risk sections are identified, including: S31: dividing the tunnel into sections according to the real-time model of the tunnel, and equally dividing the tunnel into a number of sections; S32: Retrieving historical traffic data based on historical traffic data, and setting a road section vehicle passage threshold and a vehicle speed threshold for a plurality of road sections respectively according to the historical traffic data; S33: monitoring the number of vehicles and the speed of vehicles on each road section in real time, and corresponding the number of vehicles and the speed of vehicles on each road section to a number of road sections respectively; S34: Compare the number of vehicles and the speed of each road section with the vehicle traffic threshold and the vehicle speed threshold. If the number of vehicles or the speed of a road section exceeds the corresponding threshold, it is marked as a high-risk road section.

[0032] As a preferred embodiment of the above, first, according to the total length and structural characteristics of the tunnel, the tunnel is equally divided into several sections, for example, every 1km is divided into an identification section, and then based on the stored historical traffic data, a specific vehicle passage threshold and vehicle speed threshold are set for each section. These thresholds are set based on past traffic flow, accident records and passage speed to identify potential high-risk conditions, monitor the number of vehicles and vehicle speed of each section in real time, and use sensors and cameras installed at the entrance and exit of each section to collect the number and speed data of vehicles passing in real time. The vehicle data of each section corresponds to the previously set section one by one to ensure accurate matching of the data, and the collected real-time vehicle number and speed data are compared with the preset thresholds. When the number of vehicles or the speed of any section exceeds the corresponding threshold, the section is determined to be a high-risk section. Once the section is marked as high-risk, corresponding traffic management measures will be implemented, including but not limited to traffic speed restrictions, adding traffic signs and warnings, or guiding traffic to detours when necessary.

[0033] Furthermore, a tunnel disease database is established, including: Collect historical tunnel disease information, classify and manage tunnel disease types according to the historical tunnel disease information, obtain classification management results, and build a disease index based on the classification management results; Set the data collection time series length, regularly update the tunnel disease database based on the time series length, and continuously import and export tunnel disease information.

[0034] In this embodiment, historical tunnel disease data is first collected from existing tunnel maintenance records, inspection reports and monitoring systems, including various types of disease information such as cracks, leakage, structural deformation, damaged lining, waterproof layer damage, etc., and the collected historical disease information is classified according to the type and characteristics of the tunnel disease. For example, the diseases can be divided into several categories such as structural diseases, functional diseases and aesthetic diseases. Each type of disease is further classified in more detail according to its specific characteristics such as location, severity and scope of influence. According to the results of classification management, an index is established for each disease type so that the disease information can be quickly retrieved. The index includes disease Type, frequency of occurrence, severity assessment, record of relevant repair measures, and evaluation of historical treatment effects, etc., determine an appropriate time series length for regular updates of database information. The time series length can be set based on the speed of disease development and the use of the tunnel, such as a comprehensive data update every six months or every year. Based on the set time series length, the tunnel disease database is regularly updated, including adding newly discovered tunnel disease information to the database, and updating the latest status and treatment results of existing diseases to achieve continuous operation of the database, including continuously importing the latest tunnel disease data and exporting data for analysis and report preparation. Ensure data consistency and integrity, as well as efficient data access and use.

[0035] Furthermore, tunnel defects are predicted based on the tunnel real-time model and traffic feature set to obtain defect prediction results, including: Divide the traffic feature set into a disease training set, a disease test set and a disease verification set; The tunnel real-time model is trained using deep learning technology based on the disease training set, and the trained tunnel real-time model is used to predict the disease on the disease test set to obtain preliminary prediction results; Use the disease verification set to perform accuracy analysis on the preliminary prediction results to obtain accuracy analysis results. Based on the accuracy analysis results, optimize the tunnel real-time model to obtain the final prediction results.

[0036] Specifically, the collected traffic feature set is first divided into two parts: a disease training set and a disease test set. The training set is used to establish and train the tunnel disease prediction model, while the test set is used to evaluate the prediction performance of the model. A deep learning model suitable for processing time series data and spatial data is selected, such as a convolutional neural network for image data analysis and a recurrent neural network or a long short-term memory network for processing time series data. The tunnel real-time model is trained according to the disease training set. This training process involves feature extraction, pattern recognition, and relationship learning to identify potential disease patterns and influencing factors in the tunnel. The data is preprocessed as necessary, including feature standardization, normalization, and Data enhancement can be performed to improve the efficiency and effect of model training. Appropriate model parameters such as learning rate, batch size, number of iterations, etc. are set. The training set data is used for model training. After the training is completed, the trained tunnel real-time model is used to predict the disease test set to obtain preliminary disease prediction results. The preliminary disease prediction results reflect the performance of the model on unknown data and provide a preliminary evaluation. The disease validation set is then used to perform precision analysis on the preliminary prediction results. According to the precision analysis results, the deficiencies of the model are identified, and the model is adjusted and optimized as necessary. The optimized tunnel real-time model is used to predict the test set again to obtain the final disease prediction results.

[0037] Furthermore, the preliminary prediction results are evaluated for accuracy using the disease validation set to obtain the accuracy evaluation results. Based on the accuracy evaluation results, the tunnel real-time model is optimized to obtain the final prediction results, including: Record the actual error between the preliminary prediction result and the actual disease, analyze the actual error, extract the relevant disease features that produce the actual error, and mark the relevant disease features as high error prediction features; The high-error prediction features are used as training incremental features to perform specialized training on the tunnel real-time model, obtain the specialized model, and generate the final prediction results based on the specialized model.

[0038] As a preferred embodiment of the above embodiment, the prediction result is first compared with the actual situation, and the difference between each predicted disease and the actual disease, that is, the actual error, is recorded. The recorded errors are analyzed to understand the source of the errors, including the size and distribution of the statistical errors, and the relationship between the errors and specific predictive variables. Through error analysis, it is determined which features are associated with larger errors, and these features associated with larger errors are marked as high-error prediction features. These features will be focused on in subsequent training, in order to improve the overall prediction performance of the model by adjusting the model's processing of these features. The above-marked high-error prediction features are included in the incremental feature set of model training, and incremental training is performed. Features other than high-error prediction features are frozen, and only high-error prediction features are trained separately. On the one hand, this is because features other than high-error prediction features have been trained on a large amount of data and usually perform well, and do not need to be trained again. On the other hand, freezing existing parameters can prevent too many parameters from being adjusted on a small data set, thereby reducing the risk of overfitting.

[0039] Embodiment 2 Based on the same inventive concept as the tunnel digital intelligent management and maintenance method in the aforementioned embodiment, the present invention also provides a tunnel digital intelligent management and maintenance system, such as Figure 3 As shown, the system includes: The tunnel model building module collects tunnel geometry representation and tunnel monitoring information, and builds a tunnel real-time model based on the tunnel geometry representation and tunnel monitoring information; The vehicle risk identification module collects historical traffic data, identifies the vehicle license plate and vehicle type of passing vehicles, obtains vehicle attribute information, and identifies high-risk vehicles based on the vehicle attribute information and historical traffic data; The road section risk identification module monitors the risks of tunnel sections, obtains road section risk information, identifies high-risk sections based on the road section risk information and historical traffic data, integrates vehicle attribute information and road section risk information, and obtains a traffic feature set; The disease prediction module predicts tunnel diseases based on the tunnel real-time model and traffic feature set, obtains disease prediction results, and prevents tunnel diseases according to the disease prediction results.

[0040] The above-mentioned adjustment system in the present invention can effectively realize a digital intelligent management and maintenance method for tunnels, and the technical effects that can be achieved are as described in the above-mentioned embodiments and will not be repeated here.

[0041] Specifically, the road section risk identification module includes: The tunnel section division unit divides the tunnel into sections according to the real-time tunnel model, and divides the tunnel into several sections; A road section threshold setting unit retrieves historical traffic data based on historical traffic data, and sets road section vehicle passage thresholds and vehicle speed thresholds for a number of road sections according to the historical traffic data; The road section flow monitoring unit monitors the number of vehicles and vehicle speeds on each road section in real time, and associates the number of vehicles and vehicle speeds on each road section with a number of road sections; The risk section identification unit compares the number of vehicles and the speed of each section with the vehicle passage threshold and the vehicle speed threshold. If the number of vehicles or the speed of a section exceeds the corresponding threshold, it is marked as a high-risk section.

[0042] Similarly, the above-mentioned optimization schemes for the system can also respectively achieve the optimization effects corresponding to the method in Example 1, which will not be repeated here.

[0043] Although the present application has been described in conjunction with specific features and embodiments thereof, it is obvious that various modifications and combinations may be made thereto without departing from the spirit and scope of the present application. Accordingly, this specification and the accompanying drawings are merely exemplary illustrations of the present application as defined therein, and are deemed to have covered any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. A digital intelligent tunnel maintenance method, characterized in that: The method comprises: Collecting tunnel geometry representation and tunnel monitoring information, and building a tunnel real-time model according to the tunnel geometry representation and the tunnel monitoring information; Collect historical traffic data, identify the vehicle license plate and vehicle type of passing vehicles, obtain vehicle attribute information, and identify high-risk vehicles based on the vehicle attribute information and the historical traffic data; Perform risk monitoring on the tunnel section to obtain section risk information, identify high-risk sections based on the section risk information and the historical traffic data, integrate the vehicle attribute information and the section risk information, and obtain a traffic feature set; Tunnel diseases are predicted based on the tunnel real-time model and the traffic feature set, disease prediction results are obtained, and tunnel diseases are prevented according to the disease prediction results.

2. The digital intelligent tunnel maintenance method according to claim 1 is characterized in that: Collecting tunnel structure information and tunnel monitoring information, and building a tunnel real-time model according to the tunnel structure information and the tunnel monitoring information, including: Scanning the tunnel using three-dimensional data acquisition technology to obtain point cloud data of the tunnel surface, generating a geometric representation of the tunnel based on the point cloud data, and monitoring the tunnel in real time to obtain tunnel monitoring information; Performing adaptive disease recognition on the tunnel monitoring information, identifying tunnel disease characteristics, and obtaining tunnel disease information; A real-time model of the tunnel is created based on the tunnel geometry representation, the tunnel monitoring information and the tunnel disease information.

3. The digital intelligent tunnel maintenance method according to claim 2 is characterized in that: Adaptively identifying the tunnel detection information to identify tunnel disease information includes: Establishing a tunnel disease database, extracting disease features from the tunnel disease database using image processing technology, and obtaining tunnel disease features; A disease recognition model is constructed using machine learning technology as a framework, and deep learning of tunnel disease characteristics is performed according to the disease recognition model to identify the tunnel disease characteristics; The disease recognition model is used to identify the disease characteristics of the tunnel detection information to obtain tunnel disease information.

4. The digital intelligent tunnel maintenance method according to claim 1 is characterized in that: Identify the vehicle license plate and vehicle type of the passing vehicle, obtain vehicle attribute information, and identify high-risk vehicles based on the vehicle attribute information and the historical traffic data, including: Traversing the public transportation database, collecting vehicle image data according to the imaging device, identifying the vehicle license plate based on the vehicle image data, and identifying the vehicle type according to the public transportation database and the image recognition algorithm to obtain vehicle attribute information; According to the historical traffic data, the historical vehicle type judgment standard is collected, and according to the vehicle attribute information and the historical vehicle type judgment standard, it is judged whether the vehicle type is a high-risk vehicle. If so, the vehicle is marked and the vehicle is tracked and monitored in a focused manner.

5. The digital intelligent tunnel maintenance method according to claim 2 is characterized in that: Perform risk monitoring on the tunnel section, obtain section risk information, and identify high-risk sections based on the section risk information and the historical traffic data, including: Dividing the tunnel into sections according to the real-time model of the tunnel, and equally dividing the tunnel into a number of sections; Retrieving historical traffic data based on the historical traffic data, and setting a road section vehicle passage threshold and a vehicle speed threshold for the plurality of road sections respectively according to the historical traffic data; monitoring the number of vehicles and the speed of vehicles on each of the road sections in real time, and making a one-to-one correspondence between the number of vehicles and the speed of vehicles on each of the road sections and a plurality of the road sections; The number of vehicles and the speed of vehicles on each road section are compared with the vehicle traffic threshold and the vehicle speed threshold respectively. If the number of vehicles or the speed of vehicles on a road section exceeds the corresponding threshold, it is marked as a high-risk road section.

6. The digital intelligent tunnel maintenance method according to claim 3 is characterized in that: Establish a tunnel disease database, including: Collecting historical tunnel disease information, classifying and managing tunnel disease types according to the historical tunnel disease information, obtaining classification management results, and constructing a disease index according to the classification management results; The data collection time sequence length is set, the tunnel disease database is regularly updated based on the time sequence length, and tunnel disease information is continuously imported and exported.

7. The digital intelligent tunnel maintenance method according to claim 1 is characterized in that: Predicting tunnel defects based on the tunnel real-time model and the traffic feature set to obtain defect prediction results includes: Dividing the traffic feature set into a disease training set, a disease test set and a disease verification set; The tunnel real-time model is trained using deep learning technology according to the disease training set, and the trained tunnel real-time model is used to predict the disease of the disease test set to obtain a preliminary prediction result; The preliminary prediction results are analyzed for accuracy using the disease verification set to obtain accuracy analysis results, and the tunnel real-time model is optimized based on the accuracy analysis results to obtain final prediction results.

8. The digital intelligent tunnel maintenance method according to claim 7 is characterized in that: Using the disease verification set to perform accuracy evaluation on the preliminary prediction result to obtain an accuracy evaluation result, and optimizing the tunnel real-time model based on the accuracy evaluation result to obtain a final prediction result, including: Recording the actual error between the preliminary prediction result and the actual disease, analyzing the actual error, extracting relevant disease features that produce the actual error, and marking the relevant disease features as high-error prediction features; The high error prediction feature is used as a training incremental feature to perform specialized training on the tunnel real-time model to obtain a specialized model, and a final prediction result is generated based on the specialized model.

9. A digital intelligent tunnel maintenance system, characterized in that: The system comprises: A tunnel model building module collects tunnel geometry representation and tunnel monitoring information, and builds a tunnel real-time model according to the tunnel geometry representation and the tunnel monitoring information; A vehicle risk identification module collects historical traffic data, identifies the vehicle license plate and vehicle type of passing vehicles, obtains vehicle attribute information, and identifies high-risk vehicles based on the vehicle attribute information and the historical traffic data; A road section risk identification module performs risk monitoring on the tunnel section, obtains road section risk information, identifies high-risk road sections based on the road section risk information and the historical traffic data, integrates the vehicle attribute information and the road section risk information, and obtains a traffic feature set; The disease prediction module predicts tunnel diseases based on the tunnel real-time model and the traffic feature set, obtains disease prediction results, and prevents the tunnel diseases according to the disease prediction results.

10. The digital intelligent tunnel maintenance system according to claim 9, characterized in that: The road section risk identification module includes: A tunnel section division unit, which divides the tunnel into sections according to the real-time tunnel model, and divides the tunnel into a plurality of sections; A road section threshold setting unit, which retrieves historical traffic data based on the historical traffic data, and sets road section vehicle passage thresholds and vehicle speed thresholds for the plurality of road sections respectively according to the historical traffic data; A road section flow monitoring unit monitors the number of vehicles and the speed of vehicles on each road section in real time, and associates the number of vehicles and the speed of vehicles on each road section with a plurality of road sections respectively; The risk section identification unit compares the number of vehicles and the vehicle speed of each section with the vehicle passage threshold and the vehicle speed threshold. If the number of vehicles or the vehicle speed of a section exceeds the corresponding threshold, it is marked as a high-risk section.

Images