An intelligent analysis system and device for road, bridge and tunnel disasters based on deep learning and edge-cloud collaboration

Abstract

The present invention provides an intelligent analysis system and device for road, bridge and tunnel disasters based on deep learning and edge-cloud collaboration, including collecting historical monitoring data of road, bridge and tunnel structures to form a structural health database, the data including stress and strain, ambient temperature, humidity and wind speed, and image and video data. The system divides the road, bridge and tunnel structures into regions, and identifies factors affecting disaster risks in different regions, such as geographical location, terrain, environment, etc. Based on these factors, the system predicts the disaster risks of road, bridge and tunnel structures, obtains reference targets, and determines the disaster risk prediction level by analyzing historical structural information of the same type. The present invention can accurately predict the disaster risks of road, bridge and tunnel structures, improve the efficiency and accuracy of disaster warnings, and ensure the safety and reliability of transportation infrastructure.

Description

Technical Field

[0001] The present invention relates to the field of road, bridge and tunnel disaster monitoring and analysis. More specifically, the present invention relates to a road, bridge and tunnel disaster intelligent analysis system and device based on deep learning and edge-cloud collaboration. Background Art

[0002] Existing monitoring and analysis of road, bridge and tunnel disasters mainly rely on traditional monitoring equipment and manual analysis methods. These methods can usually only provide limited data types, such as stress-strain data and some environmental data, and lack the ability to comprehensively analyze multi-dimensional data such as images and videos. In addition, traditional disaster risk assessments mostly use empirical formulas and simple models, which are difficult to accurately capture the interaction between complex environmental factors and structural characteristics. In terms of disaster warning, existing systems often lack real-time and intelligent processing capabilities, resulting in slow warning response and inability to respond to sudden disaster events in a timely manner.

[0003] In the process of implementing the embodiments of the present invention, the inventors found that there are at least the following problems or defects in the prior art: data collection is not comprehensive, and there is a lack of comprehensive analysis of multi-dimensional data; the disaster risk assessment model is not accurate enough and has difficulty adapting to complex and changing environmental conditions; the early warning system has a slow response speed and cannot meet the needs of real-time monitoring and rapid warning. Summary of the invention

[0004] The present invention provides a road, bridge and tunnel disaster intelligent analysis system and device based on deep learning and edge-cloud collaboration.

[0005] In a first aspect of the present invention, a road, bridge and tunnel disaster intelligent analysis system based on deep learning and edge-cloud collaboration is provided, comprising:

[0006] Collect historical monitoring data of road, bridge and tunnel structures to form a structural health database, wherein the historical monitoring data includes stress and strain data, environmental temperature, humidity and wind speed data, and image and video data;

[0007] According to the monitoring characteristics of the road-bridge-tunnel structure, the road-bridge-tunnel structure is divided into regions, and the influencing factors of disaster risks are identified for different regions of the road-bridge-tunnel structure;

[0008] Based on the influencing factors of the disaster risk, the disaster risk of the road, bridge and tunnel structure is predicted to obtain a disaster risk reference target;

[0009] Setting key influencing factors for structures of the same type based on the influencing factors of the disaster risk, identifying historical structural information of the same type based on the key influencing factors, and then obtaining historical reference levels of disaster risks for historical structures of the same type through data analysis;

[0010] And determine the disaster risk prediction level of the road, bridge and tunnel structure based on the disaster risk reference target and the disaster risk historical reference level.

[0011] Furthermore, after determining the disaster risk prediction level of the road, bridge and tunnel structure, it further includes:

[0012] Combined with the regional division, the disaster risk design level of different regions is compared with the disaster risk prediction level to determine their deviation and impact ratio, and the areas with greater monitoring flexibility are adjusted first to control the deviation within the preset error range;

[0013] And use the system to predict the disaster risk design levels of different types of road, bridge and tunnel structures.

[0014] Furthermore, the factors affecting the disaster risk in different areas of the road-bridge-tunnel structure are identified, including:

[0015] Determine the monitoring characteristics of the structure based on the image and video data of the road, bridge and tunnel structure;

[0016] Decomposing the monitoring part of the road-bridge-tunnel structure into key parts, secondary parts and environmental impact areas according to the monitoring characteristics of the road-bridge-tunnel structure;

[0017] And identify the influencing factors of the disaster risk for the key parts, secondary parts and environmental impact areas of the road, bridge and tunnel structures, which include geographical location, topography, environment, wind speed, temperature, geology, structural material properties, structural design parameters, traffic loads, environmental corrosion and historical disaster records.

[0018] Furthermore, based on the influencing factors of disaster risk, the disaster risk of road, bridge and tunnel structures is predicted, and the disaster risk reference targets include:

[0019] The disaster risks of the key parts, secondary parts and environmental impact areas of the road, bridge and tunnel structures are predicted separately to obtain the disaster risk levels of different areas;

[0020] And the disaster risk levels of the different areas are integrated to obtain the total disaster risk level of the road, bridge and tunnel structure as the disaster risk reference target.

[0021] Further, identifying the same type of historical structure information based on the key influencing factors includes:

[0022] Based on the key influencing factors, all sample data are searched, and a deep learning algorithm is applied to find the number of clusters with the smallest error square sum J value and the corresponding initial cluster center;

[0023] After the initial clustering center is determined, the categories of sample data and the centers of each category are finally determined according to the improved deep learning clustering algorithm, wherein the key influencing factors of the key parts include structural material properties, structural design parameters and traffic loads;

[0024] The key influencing factors of the secondary parts include environmental corrosion and historical disaster records;

[0025] The key influencing factors of the environmental impact area include geographical location, topography and environment.

[0026] Furthermore, the objective function of the deep learning clustering algorithm is:

[0027]

[0028] in, is the fuzzy matrix, Representation sample Belongs to the cluster center The membership degree of is the set of cluster centers, is the fuzzy index, is the sample size, is the number of clusters. By optimizing the objective function, the weighted sum of the distances between the sample and the center of the cluster to which it belongs is minimized, thereby achieving effective clustering of the sample data.

[0029] Furthermore, the historical reference levels of disaster risk for the same type of historical structures obtained through data analysis include:

[0030] Calculate the mean, standard deviation, maximum, minimum, and coefficient of variation of sample data;

[0031] The Z-score standardization method is used to standardize the sample data to eliminate the influence of different dimensions and magnitudes;

[0032] Based on the standardized data, the disaster risk score of each sample is calculated. The calculation formula is:

[0033]

[0034] in, is the sample data, is the sample mean, is the sample standard deviation, and are the minimum and maximum values ​​of the samples, respectively. and is the weight coefficient;

[0035] According to the disaster risk score, the sample data are divided into different risk level intervals to obtain the historical reference level of disaster risk for the same type of historical structures.

[0036] Furthermore, according to the disaster risk reference target and the disaster risk historical reference level, determining the disaster risk prediction level of the road, bridge and tunnel structure includes:

[0037] Calculate the difference between the disaster risk reference target and the disaster risk historical reference level , the calculation formula of difference is:

[0038]

[0039] in, For disaster risk reference targets, is the historical reference level of disaster risk; With the preset threshold For comparison, if , the forecast level is considered to be consistent with the historical reference level, otherwise the forecast level is adjusted according to the difference. The calculation formula is:

[0040]

[0041] in, is the initial prediction level, is the adjustment factor.

[0042] Furthermore, based on the influencing factors of the disaster risk, the disaster risk of the road, bridge and tunnel structure is predicted, and the disaster risk reference targets obtained include:

[0043] Based on the influencing factors of the disaster risk, a disaster risk assessment model based on deep learning optimization is used to predict the disaster risk of the road, bridge and tunnel structure to obtain a disaster risk reference target, wherein the disaster risk assessment model maps the input vector to a high-dimensional feature space through nonlinear mapping, and constructs an optimal classification hyperplane in the high-dimensional feature space. The classification function of the disaster risk assessment model is similar to a neural network, and the output is a linear combination of intermediate nodes, wherein each intermediate node corresponds to a support vector.

[0044] In a second aspect of the present invention, a road, bridge and tunnel disaster intelligent analysis device based on deep learning and edge-cloud collaboration is provided, comprising:

[0045] A data acquisition module is used to collect historical monitoring data of road, bridge and tunnel structures to form a structural health database, wherein the historical monitoring data includes stress and strain data, environmental temperature, humidity and wind speed data, and image and video data;

[0046] An influencing factor identification module is used to divide the road, bridge and tunnel structure into regions according to the monitoring characteristics of the road, bridge and tunnel structure, and identify the influencing factors of disaster risks for different regions of the road, bridge and tunnel structure;

[0047] A risk prediction module predicts the disaster risk of the road, bridge and tunnel structure based on the influencing factors of the disaster risk to obtain a disaster risk reference target;

[0048] A historical risk acquisition module is used to set key influencing factors of the same type of road, bridge and tunnel structures based on the influencing factors of the disaster risk, identify the same type of historical structure information based on the key influencing factors, and then obtain the historical reference level of disaster risk of the same type of historical structure through data analysis;

[0049] And a prediction target acquisition module is used to determine the disaster risk prediction level of the road, bridge and tunnel structure according to the disaster risk reference target and the disaster risk historical reference level.

[0050] The above-mentioned embodiments of the present invention have at least the following beneficial effects: the system can effectively integrate the historical monitoring data of road, bridge and tunnel structures, including stress and strain, environmental temperature and humidity, wind speed, and image and video data, to form a comprehensive structural health database. Through deep learning and edge-cloud collaborative technology, the system can accurately divide road, bridge and tunnel structures into different regions and identify the factors affecting disaster risks in different regions, such as geographical location, terrain, environment, etc. This enables the system to more accurately predict the disaster risks of road, bridge and tunnel structures and improve the accuracy and reliability of disaster warnings.

[0051] In addition, the system can also determine the disaster risk prediction level by analyzing the same type of historical structural information, thereby providing a scientific basis for the maintenance and management of road, bridge and tunnel structures. This data-driven intelligent analysis method can effectively improve the safety and reliability of transportation infrastructure, reduce the economic losses and social impacts caused by disasters, and provide strong technical support and decision-making basis for transportation management departments. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood by reading the detailed description below with reference to the accompanying drawings. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, in which:

[0053] Figure 1 A schematic diagram of the process of an intelligent analysis system for road, bridge and tunnel disasters based on deep learning and edge-cloud collaboration provided by an embodiment of the present invention;

[0054] Figure 2A schematic diagram of the structure of an intelligent analysis device for road, bridge and tunnel disasters based on deep learning and edge-cloud collaboration provided by an embodiment of the present invention;

[0055] Figure 3 The schematic diagram schematically shows the structure of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0056] The principles and spirit of the present invention will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided only to enable those skilled in the art to better understand and implement the present invention, and are not intended to limit the scope of the present invention in any way. On the contrary, these embodiments are provided to make the present invention more thorough and complete, and to fully convey the scope of the present invention to those skilled in the art.

[0057] Those skilled in the art know that the embodiments of the present invention can be implemented as a system, device, apparatus, method or computer program product. Therefore, the present invention can be specifically implemented in the following forms, namely: complete hardware, complete software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0058] It should be noted that any number of elements in the drawings is for illustration rather than limitation, and any naming is only for distinction and does not have any limiting meaning.

[0059] Reference below Figure 1 , Figure 1 A schematic diagram of the process of an intelligent analysis system for road, bridge and tunnel disasters based on deep learning and edge-cloud collaboration provided by an embodiment of the present invention. Figure 1 As shown, a road, bridge and tunnel disaster intelligent analysis system 100 based on deep learning and edge-cloud collaboration includes:

[0060] Step 101, collecting historical monitoring data of road, bridge and tunnel structures to form a structural health database, wherein the historical monitoring data includes stress and strain data, environmental temperature, humidity and wind speed data, and image and video data;

[0061] Step 102, dividing the road, bridge and tunnel structure into regions according to the monitoring characteristics of the road, bridge and tunnel structure, and identifying the influencing factors of disaster risks for different regions of the road, bridge and tunnel structure;

[0062] Step 103, predicting the disaster risk of the road, bridge and tunnel structure based on the influencing factors of the disaster risk to obtain a disaster risk reference target;

[0063] Step 104, setting key influencing factors for structures of the same type based on the influencing factors of the disaster risk, identifying historical structural information of the same type based on the key influencing factors, and then obtaining historical reference levels of disaster risk for historical structures of the same type through data analysis;

[0064] Step 105, and determining the disaster risk prediction level of the road, bridge and tunnel structure according to the disaster risk reference target and the disaster risk historical reference level.

[0065] It should be noted that the system forms a structural health database by collecting historical monitoring data of road, bridge and tunnel structures. The historical monitoring data includes stress and strain data, environmental temperature, humidity and wind speed data, and image and video data. These data are an important basis for assessing the health status of road, bridge and tunnel structures and predicting disaster risks. Stress and strain data reflect the deformation of the structure under various loads, environmental temperature, humidity and wind speed data provide the physical conditions of the environment in which the structure is located, and image and video data can be used to observe the appearance changes and damage of the structure.

[0066] Specifically, stress-strain data can be collected in real time through sensors installed on road, bridge and tunnel structures. The sensors can be strain gauges or fiber optic sensors, etc., which can accurately measure the strain changes of the structure under different loads. Ambient temperature, humidity and wind speed data can be obtained through meteorological stations or environmental monitoring equipment. These devices can be installed near or on the structure to monitor changes in environmental conditions in real time. Image and video data can be obtained through cameras installed on the structure or drones to observe the appearance and damage of the structure, such as cracks and deformations.

[0067] Preferably, the system can divide the area according to the monitoring characteristics of the road, bridge and tunnel structure, and divide the structure into key parts, secondary parts and environmental impact areas. Key parts refer to the parts of the structure that bear the main load or are crucial to the safety of the structure, such as the main beam and support of the bridge; secondary parts refer to the relatively less important parts of the structure, such as the bridge deck pavement; environmental impact areas refer to the areas where the environmental conditions of the structure have a greater impact on it, such as the entrance and exit areas of the tunnel. By analyzing the monitoring data of different areas, the influencing factors of disaster risks can be identified more accurately, thereby improving the accuracy of disaster risk prediction.

[0068] In some embodiments, after determining the disaster risk prediction level of the road, bridge and tunnel structure, the method further includes:

[0069] Combined with the regional division, the disaster risk design level of different regions is compared with the disaster risk prediction level to determine their deviation and impact ratio, and the areas with greater monitoring flexibility are adjusted first to control the deviation within the preset error range;

[0070] And use the system to predict the disaster risk design levels of different types of road, bridge and tunnel structures.

[0071] It should be noted that after determining the disaster risk prediction level of the road, bridge and tunnel structure, the system will further conduct risk assessment in combination with regional division. Regional division refers to dividing the road, bridge and tunnel structure into different functional areas or structural units in order to analyze the disaster risk of each area in more detail. The disaster risk design level refers to the risk level determined during the design phase based on factors such as the importance of the structure and the use environment, while the disaster risk prediction level is the risk assessment result obtained through monitoring data and intelligent analysis. By comparing these two levels, the deviation between the actual risk and the design expectation can be judged, and the monitoring strategy and maintenance plan can be adjusted accordingly.

[0072] Specifically, the area division can be carried out according to the physical characteristics and functional requirements of the structure. For example, in a bridge structure, the piers, main beams, bridge decks, etc. can be divided into different areas. The disaster risk design level is usually determined by engineers based on the design specifications of the structure and historical experience, while the disaster risk prediction level is calculated through the system's intelligent analysis module, combined with real-time monitoring data and historical data.

[0073] More specifically, when comparing the two levels, a pre-determined error range, such as 5% or 10%, can be set. If the deviation between the actual predicted level and the design level exceeds this range, more frequent monitoring of the relevant area or reinforcement measures will be required.

[0074] Preferably, the system can give priority to adjusting areas with greater monitoring elasticity. Areas with greater monitoring elasticity are those areas that can quickly reflect changes in structural status when monitoring data changes. For example, in a bridge, the settlement monitoring data of the pier changes quickly, and its monitoring frequency and accuracy can be adjusted first.

[0075] Furthermore, the system can also use deep learning algorithms to predict the disaster risk design levels of different types of road, bridge and tunnel structures, and identify the risk characteristics of different structural types through training models, thereby improving the accuracy and reliability of the prediction.

[0076] In some embodiments, identifying disaster risk factors in different areas of a road, bridge, or tunnel structure includes:

[0077] Determine the monitoring characteristics of the structure based on the image and video data of the road, bridge and tunnel structure;

[0078] Decomposing the monitoring part of the road-bridge-tunnel structure into key parts, secondary parts and environmental impact areas according to the monitoring characteristics of the road-bridge-tunnel structure;

[0079] And identify the influencing factors of the disaster risk for the key parts, secondary parts and environmental impact areas of the road, bridge and tunnel structures, which include geographical location, topography, environment, wind speed, temperature, geology, structural material properties, structural design parameters, traffic loads, environmental corrosion and historical disaster records.

[0080] It should be noted that the system determines the monitoring characteristics of road, bridge and tunnel structures by analyzing image and video data, and divides the structures into key parts, secondary parts and environmental impact areas accordingly. Image and video data refers to structural images and videos taken by devices such as cameras or drones, which can be used to observe changes in the appearance and damage of the structure. Key parts refer to the parts of the structure that bear the main load or are crucial to the safety of the structure, such as the main beams and supports of the bridge; secondary parts refer to the relatively less important parts of the structure, such as the bridge deck pavement; environmental impact areas refer to areas where the environmental conditions of the structure have a greater impact on it, such as the entrance and exit areas of the tunnel.

[0081] Specifically, image and video data can be obtained through high-definition cameras installed on the structure or by using drones for regular filming. This data can be analyzed by image processing algorithms to identify damage and deformation of the structure. For example, cracks, rust and other damage on the surface of a bridge can be detected through image recognition technology to determine its health status.

[0082] More specifically, in terms of regional division, the specific location and range of key and secondary parts can be determined based on the structural design drawings and actual monitoring data. For example, for a bridge, the main beam and support can be classified as key parts, while the bridge deck pavement and guardrail can be classified as secondary parts.

[0083] Preferably, the system can further refine the process of identifying factors affecting disaster risk. For example, in key locations, in addition to structural material properties and structural design parameters, changes in traffic loads, such as vehicle weight, speed, and flow, can also be considered. In secondary locations, in addition to environmental corrosion, factors such as the age and maintenance history of the structure can also be considered. In environmental impact areas, in addition to geographical location and topography, factors such as climate conditions and geological stability can also be considered.

[0084] Furthermore, the system can also use multi-source data fusion technology to conduct comprehensive analysis of image and video data with stress and strain data, environmental monitoring data, etc., in order to more comprehensively identify the factors affecting disaster risks.

[0085] In some embodiments, disaster risk of road, bridge and tunnel structures is predicted based on the influencing factors of disaster risk, and the disaster risk reference target is obtained, including:

[0086] The disaster risks of the key parts, secondary parts and environmental impact areas of the road, bridge and tunnel structures are predicted separately to obtain the disaster risk levels of different areas;

[0087] And the disaster risk levels of the different areas are integrated to obtain the total disaster risk level of the road, bridge and tunnel structure as the disaster risk reference target.

[0088] It should be noted that the system predicts the disaster risks of key parts, secondary parts and environmental impact areas of road, bridge and tunnel structures separately to obtain the disaster risk levels of different areas, and combines these levels to determine the total disaster risk level of the entire structure. The disaster risk level refers to the result of a quantitative assessment of the disaster risks that the structure may face in different areas based on monitoring data and analysis results. The total disaster risk level is a comprehensive assessment of the disaster risk of the entire structure after comprehensively considering the risk levels of each area, which is used to guide the maintenance and management decisions of the structure.

[0089] Specifically, disaster risk prediction for key parts can be based on factors such as stress-strain data, structural material properties, structural design parameters, and traffic loads. For example, by analyzing the stress-strain data of the main beam of a bridge, combined with its material properties and design parameters, its fatigue damage risk under different traffic loads can be predicted. Disaster risk prediction for secondary parts can take into account factors such as environmental corrosion and historical disaster records. For example, for secondary parts of a tunnel, its corrosion risk under different environmental conditions can be predicted by analyzing its surface corrosion conditions and historical water damage records. Disaster risk prediction for environmental impact areas can take into account factors such as geographical location, topography, and environment. For example, for tunnels located in mountainous areas, their risks in geological disasters can be predicted by analyzing their geographical location and topographical features.

[0090] Preferably, the system can use a deep learning model to predict disaster risks in different areas. For example, a convolutional neural network (CNN) can be used to analyze image and video data to identify damage and deformation of the structure, thereby predicting its disaster risk level.

[0091] Furthermore, the system can also improve the accuracy of disaster risk prediction by building a multi-factor coupling model and comprehensively considering the interaction between multiple influencing factors. For example, when predicting the total disaster risk level of a bridge, the fatigue damage risk of key parts, the corrosion risk of secondary parts, and the geological disaster risk of the environmental impact area can be coupled and analyzed to obtain a more accurate total disaster risk level.

[0092] In some embodiments, identifying the same type of historical structural information based on the key influencing factors includes:

[0093] Based on the key influencing factors, all sample data are searched, and a deep learning algorithm is applied to find the number of clusters with the smallest error square sum J value and the corresponding initial cluster center;

[0094] After the initial clustering center is determined, the categories of sample data and the centers of each category are finally determined according to the improved deep learning clustering algorithm, wherein the key influencing factors of the key parts include structural material properties, structural design parameters and traffic loads;

[0095] The key influencing factors of the secondary parts include environmental corrosion and historical disaster records;

[0096] The key influencing factors of the environmental impact area include geographical location, topography and environment.

[0097] It should be noted that the system identifies historical structural information of the same type based on key influencing factors and performs cluster analysis on sample data through deep learning algorithms. Key influencing factors refer to characteristic parameters that have a significant impact on the disaster risk of road, bridge and tunnel structures, such as structural material properties, structural design parameters, traffic loads, environmental corrosion, etc. Deep learning algorithm is a method of data learning and pattern recognition through multi-layer neural networks. It can automatically extract features from a large amount of sample data and perform effective classification and clustering. By applying deep learning algorithms, the system can identify historical structural information with similar key influencing factors, providing a reference for the prediction and analysis of disaster risks.

[0098] Specifically, the system first needs to collect a large amount of historical structural data as sample data, which should contain information on key influencing factors of the structure. Then, select a suitable deep learning algorithm, such as a convolutional neural network (CNN) or a recurrent neural network (RNN), to preprocess and extract features from the sample data. In the preprocessing stage, the data can be normalized to eliminate the influence of different dimensions and magnitudes. In the feature extraction stage, the algorithm automatically identifies key features related to disaster risks. Next, a clustering algorithm is applied to perform cluster analysis on the extracted features to determine the categories of the sample data and the centers of each category. For example, the K-means clustering algorithm can be used to divide the sample data into different categories by calculating the distance between the sample and the cluster center.

[0099] Preferably, the system can use an improved deep learning clustering algorithm to improve the accuracy and efficiency of clustering. For example, based on the K-means clustering algorithm, the concept of fuzzy clustering is introduced to construct a fuzzy matrix to represent the membership relationship between samples and cluster centers. By optimizing the membership in the fuzzy matrix, the weighted sum of the distance between the sample and its cluster center is minimized, thereby achieving effective clustering of sample data.

[0100] Furthermore, an adaptive learning mechanism can be introduced to dynamically adjust the cluster center and membership according to the clustering results to adapt to different types of structural data and disaster risk characteristics. For example, during the clustering process, if it is found that some sample data has a low membership, it means that its similarity with the current cluster center is not high. The system can automatically adjust the position of the cluster center to make it closer to these sample data, thereby improving the accuracy of clustering.

[0101] In some embodiments, the objective function of the deep learning clustering algorithm is:

[0102]

[0103] in, is the fuzzy matrix, Representation sample Belongs to the cluster center The membership degree of is the set of cluster centers, is the fuzzy index, is the sample size, is the number of clusters. By optimizing the objective function, the weighted sum of the distances between the sample and the center of the cluster to which it belongs is minimized, thereby achieving effective clustering of the sample data.

[0104] It should be noted that the system achieves effective clustering of sample data by optimizing the objective function of the deep learning clustering algorithm. The objective function refers to a mathematical expression used to measure the quality of clustering results, and its optimization process aims to minimize the weighted sum of the distances between the sample and the cluster center to which it belongs. The fuzzy matrix is ​​a matrix that represents the membership relationship between samples and cluster centers, where the membership degree represents the degree to which the sample belongs to a certain cluster center. The fuzzy index is a parameter used to adjust the membership calculation method, usually between 1 and infinity. The larger the value, the more fuzzy the membership calculation tends to be. By optimizing the objective function, the system can find the best cluster center and membership degree, so as to accurately divide the sample data into different categories.

[0105] Specifically, the process of optimizing the objective function includes the following steps: first, initialize the fuzzy matrix and cluster center, the initial cluster center can be randomly selected, and the distance between the sample and the cluster center is calculated based on the initial cluster center, and then the membership in the fuzzy matrix is ​​determined. Then, according to the definition of the objective function, the objective function value under the current clustering result is calculated. Next, through an iterative optimization algorithm, such as a gradient descent method or a genetic algorithm, the membership in the cluster center and the fuzzy matrix is ​​updated so that the objective function value gradually decreases. In each iteration, it is necessary to recalculate the distance between the sample and the updated cluster center, and update the membership in the fuzzy matrix. Repeat the iterative process until the objective function value converges to a stable value or reaches a preset number of iterations.

[0106] Preferably, the system can adopt some improvement measures to improve the effect and efficiency of objective function optimization. For example, when initializing the fuzzy matrix, a density-based initialization method can be used to select the initial cluster center according to the density distribution of the sample data, so that the initial cluster center is closer to the actual distribution of the sample data, thereby accelerating the convergence speed of the optimization process.

[0107] Furthermore, a local search strategy can be introduced to make local adjustments to the current clustering results during the iterative optimization process, such as exchanging the clustering affiliation of some samples, in order to find a better clustering result. For example, after each iteration, some samples are randomly selected and transferred from the current cluster center to other cluster centers. If the objective function value after the transfer is reduced, the adjustment is retained, otherwise it is restored to its original state. Through these improvement measures, the system can more accurately realize the clustering analysis of sample data and provide more reliable data support for disaster risk assessment of road, bridge and tunnel structures.

[0108] In some embodiments, obtaining a historical reference level of disaster risk for a historical structure of the same type through data analysis includes:

[0109] Calculate the mean, standard deviation, maximum, minimum, and coefficient of variation of sample data;

[0110] The Z-score standardization method is used to standardize the sample data to eliminate the influence of different dimensions and magnitudes;

[0111] Based on the standardized data, the disaster risk score of each sample is calculated. The calculation formula is:

[0112]

[0113] in, is the sample data, is the sample mean, is the sample standard deviation, and are the minimum and maximum values ​​of the samples, respectively. and is the weight coefficient;

[0114] According to the disaster risk score, the sample data are divided into different risk level intervals to obtain the historical reference level of disaster risk for the same type of historical structures.

[0115] It should be noted that the system obtains the historical reference level of disaster risk for historical structures of the same type through data analysis. This process includes statistical analysis and standardization of sample data to eliminate the impact of different dimensions and magnitudes, and calculates the disaster risk score for each sample. The disaster risk score is a quantitative indicator used to measure the disaster risk of sample data, which is usually calculated based on the characteristics and weights of the sample data. By dividing the sample data into different risk level intervals, the system can obtain the historical reference level of disaster risk for historical structures, providing a reference for disaster risk prediction of current structures.

[0116] Specifically, firstly, the sample data is statistically analyzed to calculate the mean, standard deviation, maximum value, minimum value, coefficient of variation and other statistics. These statistics can reflect the distribution characteristics and degree of dispersion of the sample data. Then, the Z-score standardization method is used to standardize the sample data, that is, each sample data is subtracted from the mean and divided by the standard deviation to obtain the standardized data. The standardized data has a unified dimension and magnitude, which is convenient for the subsequent calculation of disaster risk scores. Next, the disaster risk score of each sample is calculated based on the standardized data and the preset weight coefficient. The weight coefficient can be set according to the importance of the sample data and the degree of impact on the disaster risk. For example, a higher weight can be given to key factors such as structural material properties and structural design parameters.

[0117] Preferably, the system can further refine the calculation method of disaster risk score and the division criteria of risk level intervals. For example, when calculating disaster risk score, in addition to considering the mean and standard deviation of sample data, other statistics such as skewness and kurtosis can be introduced to more comprehensively reflect the distribution characteristics of sample data.

[0118] Furthermore, the risk level division criteria can be dynamically adjusted based on actual engineering experience and historical disaster data to make it more in line with the actual conditions of different types of road, bridge and tunnel structures. For example, for some tunnel structures that are susceptible to geological disasters, the risk level division criteria can be set more strictly to improve the accuracy and timeliness of early warning.

[0119] In some embodiments, determining the disaster risk prediction level of the road, bridge and tunnel structure according to the disaster risk reference target and the disaster risk historical reference level includes:

[0120] Calculate the difference between the disaster risk reference target and the disaster risk historical reference level , the calculation formula of difference is:

[0121]

[0122] in, For disaster risk reference targets, is the historical reference level of disaster risk; With the preset threshold For comparison, if , the forecast level is considered to be consistent with the historical reference level, otherwise the forecast level is adjusted according to the difference. The calculation formula is:

[0123]

[0124] in, is the initial prediction level, is the adjustment factor.

[0125] It should be noted that the system determines the disaster risk prediction level of road, bridge and tunnel structures by calculating the difference between the disaster risk reference target and the disaster risk historical reference level. The difference refers to the degree of difference between two risk levels, usually expressed as a numerical value. The disaster risk reference target is a risk level predicted based on current monitoring data and disaster risk influencing factors, while the disaster risk historical reference level is a risk level obtained by analyzing historical structural data of the same type. By comparing the difference between the two levels, the system can determine whether the risk status of the current structure is consistent with the historical data, and thus adjust the prediction level to improve accuracy.

[0126] Specifically, the calculation formula of the difference is:

[0127]

[0128] During the calculation process, it is necessary to convert the disaster risk reference target and the disaster risk historical reference level into comparable numerical forms. For example, the risk level can be divided into different numerical intervals, such as 1 for low risk, 2 for medium risk, and 3 for high risk. Then, the specific value of the difference is calculated according to the above formula. Next, the calculated difference is compared with the preset threshold value. The threshold value can be determined based on engineering experience and historical data. For example, it can be set to 0.1 or 0.2. If the difference is less than or equal to the threshold, the predicted level is considered to be consistent with the historical reference level; otherwise, the predicted level needs to be adjusted according to the difference.

[0129] Preferably, the system can further refine the adjustment method of the prediction level. For example, when adjusting the prediction level, an adjustment coefficient can be introduced to dynamically adjust the value of the adjustment coefficient according to the size of the difference. The adjustment coefficient can be set as a function proportional to the difference, for example, , where k is a constant. Then, the adjusted forecast level is calculated based on the adjustment coefficient and the initial forecast level, and the formula is: initial forecast level + adjustment coefficient. In addition, the forecast level can also be adjusted comprehensively in combination with other influencing factors, such as environmental changes, traffic load changes, etc., to improve the accuracy and reliability of the forecast.

[0130] In some embodiments, the disaster risk of the road, bridge and tunnel structure is predicted based on the influencing factors of the disaster risk to obtain a disaster risk reference target including:

[0131] Based on the influencing factors of the disaster risk, a disaster risk assessment model based on deep learning optimization is used to predict the disaster risk of the road, bridge and tunnel structure to obtain a disaster risk reference target, wherein the disaster risk assessment model maps the input vector to a high-dimensional feature space through nonlinear mapping, and constructs an optimal classification hyperplane in the high-dimensional feature space. The classification function of the disaster risk assessment model is similar to a neural network, and the output is a linear combination of intermediate nodes, wherein each intermediate node corresponds to a support vector.

[0132] It should be noted that the system uses a disaster risk assessment model based on deep learning optimization to predict the disaster risk of road, bridge and tunnel structures. Deep learning optimization refers to the improvement and optimization of traditional disaster risk assessment models through deep learning algorithms to improve the prediction accuracy and robustness of the model. The disaster risk assessment model is a mathematical model used to analyze and predict the risk level of a structure under different disaster conditions, usually including structures such as input layer, hidden layer and output layer. Through deep learning optimization, the model can better capture the interaction between complex environmental factors and structural characteristics, thereby more accurately predicting disaster risks.

[0133] Specifically, the disaster risk assessment model based on deep learning optimization can be implemented through the following steps: First, collect a large amount of historical monitoring data and disaster event data as training samples, including stress and strain data, environmental temperature, humidity and wind speed data, image and video data, and disaster event records. Then, select a suitable deep learning algorithm, such as convolutional neural network (CNN) or long short-term memory network (LSTM), to extract features and recognize patterns from the training samples.

[0134] More specifically, during the model training process, it is necessary to set appropriate network structure parameters, such as the number of layers, the number of neurons in each layer, the type of activation function, etc., as well as optimize algorithm parameters, such as learning rate, number of iterations, etc. Through training, the model can learn the complex relationship between different factors and construct the optimal classification hyperplane for disaster risk prediction of new input data.

[0135] Preferably, the system can further improve the performance of the disaster risk assessment model. For example, during the model training process, data enhancement technology can be used to increase the diversity of samples and improve the generalization ability of the model by rotating, scaling, flipping and other operations on the training samples.

[0136] Furthermore, an attention mechanism can be introduced to enable the model to pay more attention to the characteristic factors that have a greater impact on disaster risks, thereby improving the accuracy of predictions. For example, for bridge structures, attention weights can be set to make the model pay more attention to key factors such as the stress-strain data of the main beam and the deformation of the support. An integrated learning method can also be used to combine multiple deep learning models, and the prediction results of multiple models can be integrated through voting or averaging to improve the overall prediction performance and robustness.

[0137] The above-mentioned embodiments of the present invention have the following beneficial effects: the system of the present invention can realize comprehensive monitoring and intelligent analysis of road, bridge and tunnel structures. By collecting and integrating various types of historical monitoring data, such as stress and strain, ambient temperature and humidity, wind speed, and image and video data, the system can form a detailed structural health database. This provides a solid data foundation for the accurate prediction of disaster risks. In addition, the system uses deep learning and edge-cloud collaborative technology to carry out detailed regional division of road, bridge and tunnel structures and identify key disaster risk influencing factors in different regions. This refined analysis method can improve the accuracy of disaster risk prediction, enable the early warning system to detect potential safety hazards earlier, and take timely countermeasures.

[0138] The system can also determine the historical reference level of disaster risk by analyzing the same type of historical structural information, and dynamically adjust the prediction level in combination with the current disaster risk reference target. This method can not only improve the accuracy of prediction, but also provide scientific decision-making support for the maintenance and management of road, bridge and tunnel structures. By optimizing the disaster risk assessment model and adjusting the early warning strategy, the system can effectively reduce the economic losses and social impacts caused by disasters, improve the safety and reliability of transportation infrastructure, and provide strong technical support for transportation management departments.

[0139] like Figure 2 As shown, in some embodiments, a road, bridge and tunnel disaster intelligent analysis device 200 based on deep learning and edge-cloud collaboration includes:

[0140] The data acquisition module 201 is used to collect historical monitoring data of road, bridge and tunnel structures to form a structural health database, wherein the historical monitoring data includes stress and strain data, environmental temperature, humidity and wind speed data, and image and video data;

[0141] An influencing factor identification module 202 is used to divide the road, bridge and tunnel structure into regions according to the monitoring characteristics of the road, bridge and tunnel structure, and identify the influencing factors of disaster risks for different regions of the road, bridge and tunnel structure;

[0142] The risk prediction module 203 predicts the disaster risk of the road, bridge and tunnel structure based on the influencing factors of the disaster risk to obtain a disaster risk reference target;

[0143] A historical risk acquisition module 204 is used to set key influencing factors of the same type of road, bridge and tunnel structures based on the influencing factors of the disaster risk, identify the same type of historical structure information based on the key influencing factors, and then obtain the disaster risk historical reference level of the same type of historical structure through data analysis;

[0144] The prediction target acquisition module 205 is used to determine the disaster risk prediction level of the road, bridge and tunnel structure according to the disaster risk reference target and the disaster risk historical reference level.

[0145] It can be understood that the modules recorded in the intelligent analysis device for road, bridge and tunnel disasters based on deep learning and edge-cloud collaboration 200 are similar to those in the reference Figure 1 The steps described above correspond to the various steps in the intelligent analysis system for road, bridge and tunnel disasters based on deep learning and edge-cloud collaboration. Therefore, the operations, features and beneficial effects described above for the intelligent analysis system for road, bridge and tunnel disasters based on deep learning and edge-cloud collaboration are also applicable to the intelligent analysis device 200 for road, bridge and tunnel disasters based on deep learning and edge-cloud collaboration and the modules contained therein, and will not be repeated here.

[0146] Reference below Figure 3 , which shows a schematic diagram of an electronic device structure 300 suitable for implementing some embodiments of the present invention. The electronic devices in some embodiments of the present invention may include, but are not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 3 The terminal device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0147] like Figure 3As shown, the electronic device 300 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage device 308 to a random access memory (RAM) 303. In the RAM 303, various programs and data required for the operation of the electronic device 300 are also stored. The processing device 301, the ROM 302, and the RAM 303 are connected to each other via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0148] Typically, the following devices may be connected to the I / O interface 305: an input device 306 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 307 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 308 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 309. The communication device 309 may allow the electronic device 300 to communicate with other devices wirelessly or by wire to exchange data. Figure 3 The electronic device 300 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead. Figure 3 Each block shown in the figure may represent one device, or may represent multiple devices as required.

[0149] Furthermore, the storage medium of the embodiment of the present application stores program instructions that can implement all the above methods, wherein the program instructions can be stored in the above storage medium in the form of a software product, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or terminal devices such as a computer, a server, a mobile phone, and a tablet.

[0150] The above descriptions are only some preferred embodiments of the present invention and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present invention is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above inventive concept. For example, the above features are replaced with (but not limited to) technical features with similar functions disclosed in the embodiments of the present invention.

Claims

1. An intelligent analysis system for road, bridge and tunnel disasters based on deep learning and edge-cloud collaboration, characterized in that: include: Collect historical monitoring data of road, bridge and tunnel structures to form a structural health database, wherein the historical monitoring data includes stress and strain data, environmental temperature, humidity and wind speed data, and image and video data; According to the monitoring characteristics of the road-bridge-tunnel structure, the road-bridge-tunnel structure is divided into regions, and the influencing factors of disaster risks are identified for different regions of the road-bridge-tunnel structure; Based on the influencing factors of the disaster risk, the disaster risk of the road, bridge and tunnel structure is predicted to obtain a disaster risk reference target; Setting key influencing factors for structures of the same type based on the influencing factors of the disaster risk, identifying historical structural information of the same type based on the key influencing factors, and then obtaining historical reference levels of disaster risks for historical structures of the same type through data analysis; and determining the disaster risk prediction level of the road, bridge and tunnel structure based on the disaster risk reference target and the disaster risk historical reference level; Based on the influencing factors of disaster risk, the disaster risk of road, bridge and tunnel structures is predicted, and the disaster risk reference targets include: The disaster risks of the key parts, secondary parts and environmental impact areas of the road, bridge and tunnel structures are predicted separately to obtain the disaster risk levels of different areas; and synthesizing the disaster risk levels of the different areas to obtain the total disaster risk level of the road, bridge and tunnel structure as the disaster risk reference target; Identifying the same type of historical structural information based on the key influencing factors includes: Based on the key influencing factors, all sample data are searched, and a deep learning algorithm is applied to find the number of clusters with the smallest error square sum J value and the corresponding initial cluster center; After the initial clustering center is determined, the categories of sample data and the centers of each category are finally determined according to the improved deep learning clustering algorithm, wherein the key influencing factors of the key parts include structural material properties, structural design parameters and traffic loads; The key influencing factors of the secondary parts include environmental corrosion and historical disaster records; The key factors affecting the environmental impact area include geographical location, topography and environment; The objective function of the deep learning clustering algorithm is: ； in, is the fuzzy matrix, Represents a sample, Belongs to the cluster center The membership degree of is the set of cluster centers, is the fuzzy index, is the sample size, is the number of clusters; by optimizing the objective function, the weighted sum of the distances between the sample and the center of its cluster is minimized, thereby achieving effective clustering of the sample data.

2. The system according to claim 1, characterized in that After determining the disaster risk prediction level of road, bridge and tunnel structures, it further includes: Combined with the regional division, the disaster risk design level of different regions is compared with the disaster risk prediction level to determine their deviation and impact ratio, and the areas with greater monitoring flexibility are adjusted first to control the deviation within the preset error range; And use the system to predict the disaster risk design levels of different types of road, bridge and tunnel structures.

3. The system according to claim 1, characterized in that Identify the disaster risk factors in different areas of road, bridge and tunnel structures including: Determine the monitoring characteristics of the structure based on the image and video data of the road, bridge and tunnel structure; Decomposing the monitoring part of the road-bridge-tunnel structure into key parts, secondary parts and environmental impact areas according to the monitoring characteristics of the road-bridge-tunnel structure; And identify the influencing factors of the disaster risk for the key parts, secondary parts and environmental impact areas of the road, bridge and tunnel structures, which include geographical location, topography, environment, wind speed, temperature, geology, structural material properties, structural design parameters, traffic loads, environmental corrosion and historical disaster records.

4. The system according to claim 1, characterized in that The historical reference levels of disaster risk for the same type of historical structures obtained through data analysis include: Calculate the mean, standard deviation, maximum, minimum, and coefficient of variation of sample data; The Z-score standardization method is used to standardize the sample data to eliminate the influence of different dimensions and magnitudes; Based on the standardized data, the disaster risk score of each sample is calculated. The calculation formula is: ； in, is the sample data, is the sample mean, is the sample standard deviation, and are the minimum and maximum values ​​of the samples, respectively. and is the weight coefficient; According to the disaster risk score, the sample data are divided into different risk level intervals to obtain the historical reference level of disaster risk for the same type of historical structures.

5. The system according to claim 4, characterized in that According to the disaster risk reference target and the disaster risk historical reference level, the disaster risk prediction level of the road, bridge and tunnel structure is determined including: Calculate the difference between the disaster risk reference target and the disaster risk historical reference level , the calculation formula of difference is: ； in, For disaster risk reference targets, is the historical reference level of disaster risk; With the preset threshold For comparison, if , the forecast level is considered to be consistent with the historical reference level, otherwise the forecast level is adjusted according to the difference. The calculation formula is: ； in, is the initial prediction level, is the adjustment factor.

6. The system according to claim 1, characterized in that Based on the influencing factors of the disaster risk, the disaster risk of the road, bridge and tunnel structure is predicted, and the disaster risk reference targets include: Based on the influencing factors of the disaster risk, a disaster risk assessment model based on deep learning optimization is used to predict the disaster risk of the road, bridge and tunnel structure to obtain a disaster risk reference target, wherein the disaster risk assessment model maps the input vector to a high-dimensional feature space through nonlinear mapping, and constructs an optimal classification hyperplane in the high-dimensional feature space. The classification function of the disaster risk assessment model is similar to a neural network, and the output is a linear combination of intermediate nodes, wherein each intermediate node corresponds to a support vector.

7. An intelligent analysis device for road, bridge and tunnel disasters based on deep learning and edge-cloud collaboration, characterized in that: include: A data acquisition module is used to collect historical monitoring data of road, bridge and tunnel structures to form a structural health database, wherein the historical monitoring data includes stress and strain data, environmental temperature, humidity and wind speed data, and image and video data; An influencing factor identification module is used to divide the road, bridge and tunnel structure into regions according to the monitoring characteristics of the road, bridge and tunnel structure, and identify the influencing factors of disaster risks for different regions of the road, bridge and tunnel structure; A risk prediction module predicts the disaster risk of the road, bridge and tunnel structure based on the influencing factors of the disaster risk to obtain a disaster risk reference target; Based on the influencing factors of disaster risk, the disaster risk of road, bridge and tunnel structures is predicted, and the disaster risk reference targets include: The disaster risks of the key parts, secondary parts and environmental impact areas of the road, bridge and tunnel structures are predicted separately to obtain the disaster risk levels of different areas; and synthesizing the disaster risk levels of the different areas to obtain the total disaster risk level of the road, bridge and tunnel structure as the disaster risk reference target; A historical risk acquisition module is used to set key influencing factors of the same type of road, bridge and tunnel structures based on the influencing factors of the disaster risk, identify the same type of historical structure information based on the key influencing factors, and then obtain the historical reference level of disaster risk of the same type of historical structure through data analysis; Identifying the same type of historical structural information based on the key influencing factors includes: Based on the key influencing factors, all sample data are searched, and a deep learning algorithm is applied to find the number of clusters with the smallest error square sum J value and the corresponding initial cluster center; After the initial clustering center is determined, the categories of sample data and the centers of each category are finally determined according to the improved deep learning clustering algorithm, wherein the key influencing factors of the key parts include structural material properties, structural design parameters and traffic loads; The key influencing factors of the secondary parts include environmental corrosion and historical disaster records; The key factors affecting the environmental impact area include geographical location, topography and environment; The objective function of the deep learning clustering algorithm is: ； in, is the fuzzy matrix, Represents a sample, Belongs to the cluster center The membership degree of is the set of cluster centers, is the fuzzy index, is the sample size, is the number of clusters; by optimizing the objective function, the weighted sum of the distances between the sample and the center of the cluster to which it belongs is minimized, thereby achieving effective clustering of the sample data; And a prediction target acquisition module is used to determine the disaster risk prediction level of the road, bridge and tunnel structure according to the disaster risk reference target and the disaster risk historical reference level.

Images