A method and system for real-time detection and repair of road settlement

Abstract

The present application relates to the technical field of road detection and maintenance, and particularly relates to a road subsidence real-time detection and repair method and system. The method comprises the following steps: obtaining historical road surface data of a road, screening out subsidence abnormal data, establishing a road surface subsidence identification model, obtaining real-time road surface data on the target road based on the sensors embedded along the target road, determining abnormal nodes on the target road, extracting road surface data of the abnormal nodes, inputting the road surface data into the road surface subsidence identification model to obtain the degree of road surface subsidence, establishing repair measures corresponding to different degrees of road surface subsidence, and triggering corresponding repair measures to adaptively repair the subsidence position of the abnormal nodes in response to the output degree of road surface subsidence. The present application can output the degree of subsidence according to the subsidence identification model, so as to trigger differentiated repair strategies, greatly reduce the workload of traditional manual detection, reduce road excavation, timely respond to road defect detection, and quickly restore road traffic.

Description

Technical Field

[0001] This invention relates to the field of road detection technology, and in particular to a method and system for real-time detection and repair of road subsidence. Background Technology

[0002] Current road subsidence detection technology primarily employs a collaborative approach combining manual inspections and specialized equipment. In the inspection phase, manual inspections collect basic data through visual observation and simple tool measurements, while offline equipment such as ground-penetrating radar and laser scanners can identify structural defects in key road sections with millimeter-level precision. Once structural defects are identified, appropriate repair measures must be taken. For pavement repair, engineering techniques such as grouting reinforcement and asphalt repaving have established standardized construction processes. Grouting repair can stabilize the local foundation by filling underground cavities, while repaving can quickly restore the smoothness of the damaged pavement.

[0003] However, existing technologies still have significant limitations: at the detection level, manual inspections are prone to missed detections due to differences in personnel experience, while equipment such as ground-penetrating radar cannot achieve 24 / 7 coverage due to the need for pre-set detection routes; at the repair level, grouting operations rely on heavy machinery, often resulting in long traffic interruptions due to the need for construction barriers, while traditional repaving processes require closed work surfaces and layered compaction, making it difficult to respond promptly to road defect detection and quickly restore road traffic. Summary of the Invention

[0004] The technical problem to be solved by the embodiments of the present invention is to provide a method and system for real-time detection and repair of road subsidence, so as to solve the problem that the existing technology cannot identify road defects in real time and make rapid road surface repairs.

[0005] This invention discloses a method for real-time detection and repair of road subsidence, comprising: Obtain historical road surface data and filter out subsidence anomaly data from the obtained historical road surface data; Temporal features are extracted from the subsidence anomaly data, and a road subsidence identification model is established based on the extracted temporal features. Based on the sensors pre-embedded along the target road, real-time road surface data along the target road is acquired, and abnormal nodes on the target road are determined based on the acquired real-time road surface data. The abnormal node road data is extracted from the real-time road data and input into the road subsidence identification model to output the degree of road subsidence. Establish repair measures corresponding to different degrees of road subsidence. In response to the degree of road subsidence output by the road subsidence identification model, trigger the corresponding repair measures to adaptively repair the subsidence location of abnormal nodes.

[0006] Optionally, the step of establishing a road subsidence identification model based on the extracted time-domain features includes: Based on the preset road subsidence classification criteria, the temporal features of slight subsidence and severe subsidence are extracted from the subsidence anomaly data respectively. All extracted temporal features are divided into training and testing sets. A network architecture is selected to build a basic classification model. All temporal features in the training set are batch-input into the basic classification model for training to obtain the road subsidence recognition model. All time-domain features from the test set are batch-input into the trained road subsidence recognition model for testing, and the model parameters of the road subsidence recognition model are updated through an iterative optimization algorithm.

[0007] Optionally, determining abnormal nodes on the target road based on the acquired real-time road surface data includes: By continuously embedding multiple Type I sensors along the target road, real-time road surface data of each section of the target road is acquired. The real-time road surface data of the road segment area is compared with the preset standard threshold to determine whether the road segment area is an abnormal road segment. Based on the historical road surface data, the locations prone to subsidence in the road section area are determined. If the road section area is determined to be an abnormal road section, the real-time road surface data of the corresponding subsidence location is obtained by the second type of sensor pre-embedded at each subsidence location in the abnormal road section. The real-time road surface data of all prone-to-sinking locations are compared with preset standard thresholds, and abnormal nodes where road surface subsidence occurs are searched out.

[0008] Optionally, the establishment of repair measures corresponding to different degrees of road surface subsidence includes: If the road subsidence is minor, microcapsules containing a repair agent are pre-embedded, and according to the identified subsidence-prone locations, the microcapsules are pre-embedded at each subsidence-prone location. Based on the pre-embedded microcapsule, a trigger element for destroying the capsule is preset near the microcapsule, and data interaction between the server and the trigger element is established through a communication module; If the road subsidence is severe, an unmanned repair vehicle equipped with rapid grouting equipment is pre-installed, and a vehicle dispatching network architecture is established to automatically dispatch the unmanned repair vehicle to the abnormal node.

[0009] Optionally, the real-time detection and repair method for road subsidence further includes a method for triggering repair measures when the road subsidence identification model outputs a slight subsidence, including... In response to the road subsidence identification model outputting a slight subsidence, a trigger signal is sent and heating elements deployed near the microcapsule are controlled to heat and destroy the microcapsule; or, In response to the road subsidence identification model outputting a slight subsidence, a trigger signal is sent and piezoelectric elements deployed near the microcapsule are controlled to crush and destroy the microcapsule; or, In response to the output of the road subsidence identification model indicating slight subsidence, a trigger signal is sent and the sensitive material deployed near the microcapsule is controlled to release and dissolve the microcapsule.

[0010] Optionally, the establishment of a vehicle dispatching network architecture to automatically dispatch unmanned repair vehicles to abnormal nodes includes: Based on the identified locations prone to subsidence, coordinate information for each location is obtained using GPS positioning. Each subsidence-prone location along the target road is numbered sequentially, and the number is bound to the coordinate information of the corresponding subsidence-prone location and stored. In response to the severe subsidence identification result output by the road subsidence identification model, the number of the abnormal node is obtained, and the corresponding coordinate information is extracted based on the number of the abnormal node. A vehicle dispatching network architecture is established, the extracted coordinate information is sent to the vehicle dispatching network architecture, and the vehicle dispatching network architecture dispatches unmanned repair vehicles to the abnormal nodes according to the extracted coordinate information.

[0011] Optionally, the step of dispatching an unmanned repair vehicle to the abnormal node by the vehicle dispatching network architecture based on the extracted coordinate information includes: The coordinates of each unmanned repair vehicle are obtained based on GPS positioning; The coordinates of the abnormal nodes are matched one by one with the coordinates of all unmanned repair vehicles, and all unmanned repair vehicles are sorted in order of distance from nearest to farthest. The dispatcher obtains traffic flow information around the nearest unmanned repair vehicle and calculates the travel time of the nearest unmanned repair vehicle to the abnormal node based on the obtained traffic flow information. The calculated travel time is compared with the preset time. If the travel time is greater than the preset time, the dispatcher obtains the surrounding traffic flow information of the second near-unmanned repair vehicle and calculates the travel time of the second near-unmanned repair vehicle to the abnormal node until the calculated travel time is less than the preset time. If the calculated travel time is less than the preset time, the corresponding unmanned repair vehicle will be dispatched to the abnormal node.

[0012] Optionally, the real-time detection and repair method for road subsidence further includes a method for determining the repair priority of multiple abnormal nodes after identifying abnormal nodes on the target road, including: Calculate the first deviation rate between the real-time road surface data and the preset standard threshold for each abnormal road segment; Each abnormal road segment is divided into different repair priorities in descending order of the first deviation rate; Based on the number of abnormal nodes in each abnormal road segment and the degree of subsidence of each abnormal node, a comprehensive abnormality score is calculated for each abnormal node. The repair priority of each abnormal road segment is then dynamically adjusted based on the comprehensive abnormality score. The formula for calculating the comprehensive abnormality score is as follows:

[0013]

[0014] In the formula, Represents the density coefficient of abnormal nodes. This indicates the number of abnormal nodes in the abnormal road segment. This indicates the number of locations prone to subsidence within the abnormal road section. Indicates the overall anomaly score. Indicates the degree of subsidence of abnormal nodes. This represents the first deviation rate between the real-time road surface data of the abnormal road section and the preset standard threshold. and This indicates a fixed weight.

[0015] Optionally, the real-time detection and repair method for road subsidence further includes a method for issuing an early warning after the road subsidence identification model outputs the degree of road subsidence, including: In response to the output of the road subsidence identification model on the degree of road subsidence, the coordinate information, subsidence degree and comprehensive anomaly score of the abnormal node are extracted to generate a structured report; Based on the generated structured report, alerts containing remediation suggestions and different remediation priorities are pushed via client or mobile devices.

[0016] This invention also discloses a system employing the above-described real-time detection and repair method for road subsidence, the system comprising: The data filtering module is used to acquire historical road surface data and filter out subsidence anomaly data from the acquired historical road surface data. The model building module is used to extract time-domain features from the subsidence anomaly data and build a road subsidence identification model based on the extracted time-domain features. The abnormal node determination module is used to acquire real-time road surface data along the target road based on sensors pre-embedded along the target road, and to determine abnormal nodes on the target road based on the acquired real-time road surface data. The subsidence degree identification module is used to extract road surface data of abnormal nodes from the real-time road surface data, input them into the road subsidence identification model, and output the degree of road subsidence. The repair trigger module is used to establish repair measures corresponding to different degrees of road subsidence. In response to the degree of road subsidence output by the road subsidence identification model, the corresponding repair measures are triggered to adaptively repair the subsidence location of abnormal nodes.

[0017] Compared with the prior art, the beneficial effects of the real-time detection and repair method and system for road subsidence provided in this invention are as follows: By analyzing historical road surface data and combining it with AI algorithms to extract temporal features, a high-precision subsidence identification model can be established. Furthermore, by collecting real-time road surface data along the target road, abnormal nodes on the target road can be identified. Based on the subsidence degree output by the subsidence identification model, differentiated repair strategies can be triggered. Adaptive repair of the subsidence location of abnormal nodes significantly reduces the workload of traditional manual inspection and enables timely response to road defect detection and rapid restoration of road traffic. Attached Figure Description

[0018] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. In the accompanying drawings: Figure 1 This is a schematic block diagram illustrating the overall steps of the real-time detection and repair method for road subsidence provided in this embodiment of the invention. Figure 2 This is a schematic flowchart illustrating the execution process of the real-time detection and repair method for road subsidence provided in an embodiment of the present invention. Figure 3 A schematic block diagram illustrating the execution flow for searching for abnormal nodes provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the architecture for scheduling unmanned repair vehicles provided in an embodiment of the present invention. Detailed Implementation

[0019] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. The preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0020] This invention discloses a method for real-time detection and repair of road subsidence, such as... Figure 1 As shown, it includes: S1. Obtain historical road surface data and filter out subsidence anomaly data from the obtained historical road surface data; S2. Extract time-domain features from subsidence anomaly data and establish a road subsidence identification model based on the extracted time-domain features; S3. Based on the sensors pre-embedded along the target road, acquire real-time road surface data along the target road, and determine abnormal nodes on the target road based on the acquired real-time road surface data. S4. Extract the road surface data of abnormal nodes from the real-time road surface data and input them into the road surface settlement identification model to output the degree of road surface settlement. S5. Establish repair measures corresponding to different degrees of road subsidence. In response to the degree of road subsidence output by the road subsidence identification model, trigger the corresponding repair measures to adaptively repair the subsidence location of abnormal nodes.

[0021] Through the implementation of the above-described real-time detection and repair method for road subsidence, historical road surface data is analyzed, and temporal features are extracted using AI algorithms to help establish a high-precision subsidence identification model. By collecting real-time road surface data along the target road, abnormal nodes on the target road are identified. Based on the subsidence degree output by the subsidence identification model, differentiated repair strategies are triggered. Adaptive repair of the subsidence location of abnormal nodes significantly reduces the workload of traditional manual detection and enables timely response to road defect detection and rapid restoration of road traffic. Historical road surface data can be obtained from a database of long-term sensor monitoring and recording, or provided by a dedicated municipal department. The road surface data used in this embodiment mainly includes values ​​such as subsidence depth, stress, and strain, and the extracted temporal features include subsidence rate. By identifying and judging the subsidence depth, stress, and strain values ​​through the subsidence identification model, the degree of subsidence at abnormal road surfaces is determined, and corresponding repair measures are matched. Priority repair decisions are made for roads with different subsidence degrees.

[0022] Furthermore, combined Figure 2 As shown, a road surface subsidence identification model is established based on the extracted time-domain features, including: Based on the preset road surface settlement classification criteria, the temporal features of minor settlement and severe settlement were extracted from the settlement anomaly data. All extracted temporal features are divided into training and test sets. A network architecture is selected to build a basic classification model. All temporal features in the training set are batch-input into the basic classification model for training to obtain a road subsidence recognition model. All time-domain features from the test set are batch-input into the trained road subsidence recognition model for testing, and the model parameters of the road subsidence recognition model are updated through an iterative optimization algorithm.

[0023] Through the implementation of the above-described real-time detection and repair method for road subsidence, the road subsidence classification standard is based on different subsidence thresholds set according to road grade and traffic flow. For example, the subsidence threshold for highways, urban expressways, and urban arterial roads is 5mm, meaning that subsidence of less than 5mm on highways, urban expressways, and urban arterial roads is considered minor subsidence, and subsidence exceeding 5mm is considered severe subsidence; the subsidence threshold for ordinary roads is 10mm, meaning that subsidence of less than 10mm on ordinary roads is considered minor subsidence, and subsidence exceeding 10mm is considered severe subsidence. Preferably, a 7:3 ratio can be used to divide the training set and the test set to ensure a balanced distribution of the two types of subsidence samples. Classification machine learning algorithms such as XGBoost (Extreme Gradient Boosting), Bayes (Bayes Classifier), CNN (Convolutional Neural Network), and SVM (Support Vector Machine) can be selected as the basic classification model to train the training set, and the hyperparameters and model structure of the algorithm can be adjusted to optimize model performance. The preferred convolutional neural network (CNN) algorithm model has an input layer designed based on the shape of the processed action data, such as filtering. If the data is two-dimensional, the input layer should accept two-dimensional input. Convolutional layers extract local features from the input data by adding one or more layers, with appropriate kernel size and number selected. Pooling layers (such as max pooling) are added after each convolutional layer to reduce the spatial dimensionality of the features and decrease computation. Multiple convolutional and pooling layers are typically stacked to form multiple convolutional blocks. After the convolutional layers, one or more fully connected layers are added to learn high-level features and perform classification. The number of neurons in the fully connected layers depends on the needs of the output layer; for example, for multi-class classification problems, the number of neurons in the last layer equals the number of classes. After the last fully connected layer, the output is converted into a probability distribution using the Softmax function (normalized exponential function). This allows for high-precision identification of road subsidence levels using a classification algorithm.

[0024] Furthermore, combined Figure 3 As shown, abnormal nodes on the target road are identified based on the acquired real-time road surface data, including: By continuously embedding multiple Type I sensors along the target road, real-time road surface data of each section of the target road is acquired. The real-time road surface data of the road segment area is compared with the preset standard threshold to determine whether the road segment area is an abnormal road segment. Based on the historical road surface data, the locations prone to subsidence in the road section area are determined. If the road section area is determined to be an abnormal road section, the real-time road surface data of the corresponding subsidence location is obtained by the second type of sensor pre-embedded at each subsidence location in the abnormal road section. The real-time road surface data of all prone-to-sinking locations are compared with preset standard thresholds, and abnormal nodes where road surface subsidence occurs are searched out.

[0025] Through the implementation of the above-described method for real-time detection and repair of road subsidence, the first type of sensor is preferably a fiber optic sensor. Fiber optic sensors possess long-distance continuous monitoring capabilities (covering several kilometers), high sensitivity (detecting micro-strain), and resistance to electromagnetic interference. Deployed continuously at 200-meter intervals, they are suitable for preliminary screening of large-scale road surface deformation. They offer wide coverage and low cost, effectively monitoring global data across the entire road, capturing the overall uneven subsidence trend of the road segment, and quickly identifying suspected abnormal road sections. This avoids the waste of resources associated with high-precision monitoring of all road sections and reduces data processing volume. Furthermore, a preset threshold is used to initially exclude normal areas. Specifically, the stress and strain values ​​in the real-time road surface data are compared with preset standard thresholds. If the stress or strain value of a road section exceeds the preset threshold, the road section is determined to be an abnormal area; otherwise, it is considered a normal area, effectively improving the system's response speed. The second type of sensor is preferably a piezoelectric sensor. Piezoelectric sensors are sensitive to dynamic pressure changes (such as instantaneous roadbed collapse) and have millimeter-level resolution, used for high-precision local monitoring. Deployed at 10-meter intervals, they can cover the microscopic deformation characteristics of easily subsidence locations, such as crack propagation and pore water pressure changes, for node monitoring of each easily subsidence location. When an abnormal road section triggers an early warning, the second type of sensors at each easily subsidence location within the abnormal road section are activated for millimeter-level monitoring. That is, only after an abnormal road section is triggered are the second type of sensors at each easily subsidence location dynamically activated, capturing the core location of the subsidence through high-frequency sampling. The stress and strain values ​​in the real-time pavement data at the easily subsidence location are then compared with preset standard thresholds to pinpoint the specific subsidence anomaly node. This improves the subsidence identification response time, reduces the false alarm rate, and lowers energy consumption and losses during long-term high-load operation. Therefore, in the collaborative mode of dual-sensor monitoring, the monitoring frequency of the entire target road is greatly improved, the time required to capture minor subsidence is effectively shortened, and the sensor deployment density on the target road is reduced, saving costs. Both the first and second type of sensors transmit data to the cloud platform using Low Power Wide Area Network (LPWAN) technology.

[0026] Furthermore, establish corresponding repair measures for different degrees of road surface subsidence, including: If the road surface subsidence is minor, microcapsules containing repair agents are pre-installed, and microcapsules are pre-embedded at each subsidence-prone location based on the identified subsidence-prone locations. Based on the pre-embedded microcapsules, a trigger element for destroying the capsule is preset near the microcapsules, and data interaction between the server and the trigger element is established through a communication module; If the road subsidence is severe, an unmanned repair vehicle equipped with rapid grouting equipment is pre-installed, and a vehicle dispatching network architecture is established to automatically dispatch the unmanned repair vehicle to the abnormal node.

[0027] Furthermore, the real-time detection and repair method for road subsidence also includes a method for triggering repair measures when the road subsidence identification model outputs a slight subsidence, including... In response to the road subsidence identification model outputting a slight subsidence, a trigger signal is sent and heating elements deployed near the microcapsules are controlled to heat and destroy the microcapsules; or, In response to the road subsidence identification model outputting a slight subsidence, a trigger signal is sent and piezoelectric elements deployed near the microcapsule are controlled to crush and destroy the microcapsule; or, In response to the road subsidence recognition model outputting a slight subsidence, a trigger signal is sent and sensitive materials deployed near the microcapsules are controlled to release and dissolve the microcapsules.

[0028] Through the implementation of the above-described method for real-time detection and repair of road subsidence, the microcapsule shell is made of polyurethane material, and the interior is filled with epoxy resin repair agent. When the repair agent flows out from the ruptured microcapsule, it can fill cracks in the road surface where there is slight subsidence. At this time, the repair agent reacts chemically with the road surface material, and after curing, it can quickly restore the road surface smoothness. That is, the epoxy resin repair agent undergoes a cross-linking reaction with the components in the road surface material, forming a strong polymer network. The repair agent typically cures initially within 10-30 minutes and fully cures within 24 hours. The repaired road surface strength reaches more than 95% of the original road surface, ensuring traffic safety.

[0029] Preferably, the microcapsule diameter is 0.1mm-0.5mm, and the filling amount of the repair agent is 100g-200g per square meter. When the trigger element is a heating element, a resistance wire is preferred. When the trigger signal arrives, the heating element is energized, softening and rupturing the microcapsule shell. When the trigger element is a piezoelectric element, it generates mechanical pressure upon the trigger signal, damaging the microcapsule shell. When the trigger element is a chemically sensitive material, a pH-sensitive material is preferred. Upon the trigger signal, it releases a chemical substance, such as an acid or alkali, to dissolve the microcapsule shell. The trigger signal can be transmitted to the trigger element via a wireless network. For example, an electrical signal can be transmitted via a wire to the trigger element near the microcapsule; or a wireless signal can be used to activate the trigger element near the microcapsule via radio frequency or infrared signals.

[0030] Furthermore, combined Figure 4 As shown, a vehicle dispatching network architecture is established to automatically dispatch unmanned repair vehicles to abnormal nodes, including: Based on the identified locations prone to subsidence, coordinate information for each location is obtained using GPS positioning. Each subsidence-prone location along the target road is numbered sequentially, and the number is bound to the coordinate information of the corresponding subsidence-prone location and stored. In response to the severe subsidence identification result output by the road subsidence identification model, the numbers of abnormal nodes are obtained, and the corresponding coordinate information is extracted based on the numbers of abnormal nodes. A vehicle dispatching network architecture is established. The extracted coordinate information is sent to the vehicle dispatching network architecture, which then dispatches the unmanned repair vehicle to the abnormal node based on the extracted coordinate information.

[0031] Furthermore, the vehicle dispatching network architecture dispatches unmanned repair vehicles to the abnormal nodes based on the extracted coordinate information, including: The coordinates of each unmanned repair vehicle are obtained based on GPS positioning; The coordinates of the abnormal nodes are matched one by one with the coordinates of all unmanned repair vehicles, and all unmanned repair vehicles are sorted in order of distance from nearest to farthest. The dispatcher obtains traffic flow information around the nearest unmanned repair vehicle and calculates the travel time of the nearest unmanned repair vehicle to the abnormal node based on the obtained traffic flow information. The calculated travel time is compared with the preset time. If the travel time is greater than the preset time, the dispatcher obtains the surrounding traffic flow information of the second near-unmanned repair vehicle and calculates the travel time of the second near-unmanned repair vehicle to the abnormal node until the calculated travel time is less than the preset time. If the calculated travel time is less than the preset time, the corresponding unmanned repair vehicle will be dispatched to the abnormal node.

[0032] By implementing the above-described real-time detection and repair method for road subsidence, the coordinates of subsidence-prone locations are obtained based on GPS positioning, ensuring that repair vehicles can accurately navigate to the abnormal nodes and avoiding positioning errors caused by traditional manual marking. Furthermore, by sorting according to the dual constraints of "distance priority + time tolerance," unmanned repair vehicles near abnormal nodes can be selected, and dynamic adjustments based on traffic flow are made to prioritize the unmanned repair vehicles with the fastest estimated arrival time. Thus, by dynamically planning the scheduling of unmanned repair vehicles, the entire process from identifying road subsidence anomalies to the deployment of unmanned repair vehicles is significantly shortened, resulting in a significant improvement in efficiency compared to manual scheduling. In multi-vehicle collaborative mode, scheduling tasks for multiple abnormal nodes can be processed in parallel, improving resource utilization. The preset time threshold can be set according to road grade differences. For example, the threshold is 20 minutes for main roads and 30 minutes for secondary roads, avoiding scheduling failures caused by a single standard.

[0033] Furthermore, the real-time detection and repair method for road subsidence also includes a method for determining the repair priority of multiple abnormal nodes after identifying abnormal nodes on the target road, including: Calculate the first deviation rate between the real-time road surface data and the preset standard threshold for each abnormal road segment; Each abnormal road segment is divided into different repair priorities in descending order of the first deviation rate; Based on the number of abnormal nodes in each abnormal road segment and the degree of subsidence of each abnormal node, a comprehensive anomaly score is calculated for each abnormal node. The repair priority of each abnormal road segment is then dynamically adjusted based on the comprehensive anomaly score. The formula for calculating the comprehensive anomaly score is as follows:

[0034]

[0035] In the formula, Represents the density coefficient of abnormal nodes. This indicates the number of abnormal nodes in the abnormal road segment. This indicates the number of locations prone to subsidence within the abnormal road section. Indicates the overall anomaly score. Indicates the degree of subsidence of abnormal nodes. This represents the first deviation rate between the real-time road surface data of the abnormal road section and the preset standard threshold. and This indicates a fixed weight.

[0036] Through the implementation of the above-described real-time detection and repair method for road subsidence, the deviation rate between the real-time stress and deformation data of each road segment and a preset threshold is first calculated. For example, a strain value exceeding the threshold by 10% is considered a Level 1 anomaly, and exceeding 5% is considered a Level 2 anomaly. Then, within the abnormal road segment, the deformation data of the abnormal nodes are weighted and calculated to generate a comprehensive anomaly score. Based on this comprehensive anomaly score, the repair priority of each abnormal road segment is dynamically adjusted. For example, when the comprehensive anomaly score is greater than the threshold, a Level 1 response is directly triggered. Furthermore, by setting different thresholds, abnormal situations are classified into different levels, which helps to manage anomalies hierarchically, prioritize the handling of more severe anomalies, ensure that resources are allocated to the road segments most in need of repair, improve the efficiency of road maintenance, reduce interference from human factors, and improve the accuracy and reliability of decision-making.

[0037] Furthermore, the real-time detection and repair method for road subsidence also includes a method for issuing early warnings after the road subsidence recognition model outputs the degree of road subsidence, including: In response to the output of the road subsidence identification model, the coordinate information, subsidence degree, and comprehensive anomaly score of abnormal nodes are extracted to generate a structured report. Based on the generated structured report, alerts containing repair suggestions and different repair priorities are pushed to clients or mobile devices.

[0038] Through the implementation of the above-described real-time detection and repair method for road subsidence, the generated structured reports present the road subsidence situation in an intuitive and easy-to-understand manner, facilitating rapid access to key information for relevant personnel. By providing the coordinates of abnormal nodes, the method accurately locates subsidence areas requiring attention, improving work efficiency. Timely push notifications of early warning information via client or mobile devices ensure that relevant personnel receive alerts anytime, anywhere, enhancing the convenience and flexibility of information transmission, preventing further expansion of road subsidence, and aiding in rapid response and problem-solving. The pushed warning information includes repair suggestions and different repair priorities, providing decision support for road management departments, helping to formulate reasonable repair plans, ensuring that resources are allocated first to the most urgent and critical repair tasks, and improving work efficiency.

[0039] This invention also discloses a system employing the above-mentioned real-time detection and repair method for road subsidence. The system includes: The data filtering module is used to acquire historical road surface data and filter out subsidence anomaly data from the acquired historical road surface data. The model building module is used to extract time-domain features from subsidence anomaly data and build a road subsidence identification model based on the extracted time-domain features. The abnormal node determination module is used to acquire real-time road surface data along the target road based on sensors pre-embedded along the target road, and to determine abnormal nodes on the target road based on the acquired real-time road surface data. The subsidence degree identification module is used to extract road surface data of abnormal nodes from real-time road surface data and input them into the road subsidence identification model, and output the degree of road subsidence. The repair trigger module is used to establish repair measures corresponding to different degrees of road subsidence. In response to the degree of road subsidence output by the road subsidence identification model, the corresponding repair measures are triggered to adaptively repair the subsidence location of abnormal nodes.

[0040] The present invention also discloses a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the above-described method for real-time detection and repair of road subsidence.

[0041] The present invention also discloses a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the above-described real-time detection and repair method for road subsidence.

[0042] This invention is described based on flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to specific embodiments. It should be understood that each block of the flowcharts and / or block diagrams, and combinations of blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing device, generate instructions for implementing the flowcharts and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0043] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0044] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0045] It should be understood that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them. Those skilled in the art can modify the technical solutions described in the above embodiments, or make equivalent substitutions for some of the technical features; and all such modifications and substitutions should fall within the protection scope of the present invention.

Claims

1. A method for real-time detection and repair of road subsidence, characterized in that, The real-time detection and repair method for road subsidence includes: Obtain historical road surface data and filter out subsidence anomaly data from the obtained historical road surface data; Temporal features are extracted from the subsidence anomaly data, and a road subsidence identification model is established based on the extracted temporal features. Based on the sensors pre-embedded along the target road, real-time road surface data along the target road is acquired, and abnormal nodes on the target road are determined based on the acquired real-time road surface data. The abnormal node road data is extracted from the real-time road data and input into the road subsidence identification model to output the degree of road subsidence. Establish repair measures corresponding to different degrees of road subsidence. In response to the degree of road subsidence output by the road subsidence identification model, trigger the corresponding repair measures to adaptively repair the subsidence location of abnormal nodes. The step of establishing a road subsidence identification model based on the extracted time-domain features includes: Based on the preset road subsidence classification criteria, the temporal features of slight subsidence and severe subsidence are extracted from the subsidence anomaly data respectively. All extracted temporal features are divided into training and testing sets. A network architecture is selected to build a basic classification model. All temporal features in the training set are batch-input into the basic classification model for training to obtain the road subsidence recognition model. All time-domain features from the test set are batch-input into the trained road subsidence recognition model for testing, and the model parameters of the road subsidence recognition model are updated through an iterative optimization algorithm. The step of determining abnormal nodes on the target road based on the acquired real-time road surface data includes: By continuously embedding multiple Type I sensors along the target road, real-time road surface data of each section of the target road is acquired. The real-time road surface data of the road segment area is compared with the preset standard threshold to determine whether the road segment area is an abnormal road segment. Based on the historical road surface data, the locations prone to subsidence in the road section area are determined. If the road section area is determined to be an abnormal road section, the real-time road surface data of the corresponding subsidence location is obtained by the second type of sensor pre-embedded at each subsidence location in the abnormal road section. The real-time road surface data of all prone-to-sinking locations in the abnormal road section are compared with preset standard thresholds, and abnormal nodes where road surface subsidence occurs are searched out.

2. The method for real-time detection and repair of road subsidence according to claim 1, characterized in that, The established repair measures corresponding to different degrees of road surface subsidence include: If the road subsidence is minor, microcapsules containing a repair agent are pre-embedded, and according to the identified subsidence-prone locations, the microcapsules are pre-embedded at each subsidence-prone location. Based on the pre-embedded microcapsule, a trigger element for destroying the capsule is preset near the microcapsule, and data interaction between the server and the trigger element is established through a communication module; If the road subsidence is severe, an unmanned repair vehicle equipped with rapid grouting equipment is pre-installed, and a vehicle dispatching network architecture is established to automatically dispatch the unmanned repair vehicle to the abnormal node.

3. The method for real-time detection and repair of road subsidence according to claim 2, characterized in that, The real-time detection and repair method for road subsidence also includes a method for triggering repair measures when the road subsidence identification model outputs a slight subsidence, including... In response to the road subsidence identification model outputting a slight subsidence, a trigger signal is sent and heating elements deployed near the microcapsule are controlled to heat and destroy the microcapsule; or, In response to the road subsidence identification model outputting a slight subsidence, a trigger signal is sent and piezoelectric elements deployed near the microcapsule are controlled to crush and destroy the microcapsule; or, In response to the output of the road subsidence identification model indicating slight subsidence, a trigger signal is sent and the sensitive material deployed near the microcapsule is controlled to release and dissolve the microcapsule.

4. The method for real-time detection and repair of road subsidence according to claim 2, characterized in that, The establishment of a vehicle dispatching network architecture to automatically dispatch unmanned repair vehicles to abnormal nodes includes: Based on the identified locations prone to subsidence, coordinate information for each location is obtained using GPS positioning. Each subsidence-prone location along the target road is numbered sequentially, and the number is bound to the coordinate information of the corresponding subsidence-prone location and stored. In response to the severe subsidence identification result output by the road subsidence identification model, the number of the abnormal node is obtained, and the corresponding coordinate information is extracted based on the number of the abnormal node. A vehicle dispatching network architecture is established, the extracted coordinate information is sent to the vehicle dispatching network architecture, and the vehicle dispatching network architecture dispatches unmanned repair vehicles to the abnormal nodes according to the extracted coordinate information.

5. The method for real-time detection and repair of road subsidence according to claim 4, characterized in that, The process of dispatching unmanned repair vehicles to the abnormal nodes based on the extracted coordinate information by the vehicle dispatching network architecture includes: The coordinates of each unmanned repair vehicle are obtained based on GPS positioning; The coordinates of the abnormal nodes are matched one by one with the coordinates of all unmanned repair vehicles, and all unmanned repair vehicles are sorted in order of distance from nearest to farthest. The dispatcher obtains traffic flow information around the nearest unmanned repair vehicle and calculates the travel time of the nearest unmanned repair vehicle to the abnormal node based on the obtained traffic flow information. The calculated travel time is compared with the preset time. If the travel time is greater than the preset time, the dispatcher obtains the surrounding traffic flow information of the second near-unmanned repair vehicle and calculates the travel time of the second near-unmanned repair vehicle to the abnormal node until the calculated travel time is less than the preset time. If the calculated travel time is less than the preset time, the corresponding unmanned repair vehicle will be dispatched to the abnormal node.

6. The method for real-time detection and repair of road subsidence according to any one of claims 4 or 5, characterized in that, The real-time detection and repair method for road subsidence also includes a method for determining the repair priority of multiple abnormal nodes after identifying abnormal nodes on the target road, including: Calculate the first deviation rate between the real-time road surface data and the preset standard threshold for each abnormal road segment; Each abnormal road segment is divided into different repair priorities in descending order of the first deviation rate; Based on the number of abnormal nodes in each abnormal road segment and the degree of subsidence of each abnormal node, a comprehensive abnormality score is calculated for each abnormal node. The repair priority of each abnormal road segment is then dynamically adjusted based on the comprehensive abnormality score. The formula for calculating the comprehensive abnormality score is as follows: In the formula, Represents the density coefficient of abnormal nodes. This indicates the number of abnormal nodes in the abnormal road segment. This indicates the number of locations prone to subsidence within the abnormal road section. Indicates the overall anomaly score. Indicates the degree of subsidence of abnormal nodes. This represents the first deviation rate between the real-time road surface data of the abnormal road section and the preset standard threshold. and This indicates a fixed weight.

7. The method for real-time detection and repair of road subsidence according to claim 6, characterized in that, The real-time detection and repair method for road subsidence also includes a method for issuing early warnings after the road subsidence recognition model outputs the degree of road subsidence, including: In response to the output of the road subsidence identification model on the degree of road subsidence, the coordinate information, subsidence degree and comprehensive anomaly score of the abnormal node are extracted to generate a structured report; Based on the generated structured report, alerts containing remediation suggestions and different remediation priorities are pushed via client or mobile devices.

8. A system, characterized in that, The system employing the real-time detection and repair method for road subsidence according to any one of claims 1-7 comprises: The data filtering module is used to acquire historical road surface data and filter out subsidence anomaly data from the acquired historical road surface data. The model building module is used to extract time-domain features from the subsidence anomaly data and build a road subsidence identification model based on the extracted time-domain features. The abnormal node determination module is used to acquire real-time road surface data along the target road based on sensors pre-embedded along the target road, and to determine abnormal nodes on the target road based on the acquired real-time road surface data. The subsidence degree identification module is used to extract road surface data of abnormal nodes from the real-time road surface data, input them into the road subsidence identification model, and output the degree of road subsidence. The repair trigger module is used to establish repair measures corresponding to different degrees of road subsidence. In response to the degree of road subsidence output by the road subsidence identification model, the corresponding repair measures are triggered to adaptively repair the subsidence location of abnormal nodes.

Images