A work linkage method and system based on slope retaining wall intelligent monitoring

Abstract

The application provides a kind of based on slope retaining wall intelligent monitoring operation linkage method and system, belong to slope construction technical field, this method includes: obtaining the monitoring data of slope retaining wall, and the monitoring data is preprocessed and feature extraction, obtain monitoring feature vector;Monitoring feature vector is input into pre-trained risk identification model, obtain current risk type and level, and it is input into pre-trained risk trend prediction model with monitoring feature vector, obtain risk evolution trend;Current risk type and level are fused with risk evolution trend, obtain risk warning information;Risk warning information is matched with preset linkage rule, obtain operation linkage strategy, and it is parsed into standardized strategy instruction, strategy instruction is distributed to corresponding target department operation execution system through preset interface protocol and start preset business process, execute corresponding strategy instruction.The application improves the precision of operation linkage.

Description

Technical Field

[0001] This application relates to the field of slope construction technology, and in particular to an operation linkage method and system based on intelligent monitoring of slope retaining walls. Background Technology

[0002] With the widespread application of slope retaining walls in transportation networks and urban infrastructure, their structural stability is crucial to road safety, public travel security, and urban operational order. Existing monitoring methods rely on manual inspections, which suffer from data lag and untimely risk identification. Furthermore, the relatively independent operating systems of various management departments make it difficult to achieve cross-departmental collaborative responses to potential risks such as slope retaining wall deformation and leakage, potentially leading to chain reactions such as traffic congestion and safety accidents.

[0003] Against this backdrop, there is an urgent need for an operational linkage method and system based on intelligent monitoring of slope retaining walls. Summary of the Invention

[0004] To address the aforementioned technical issues, this application provides an operational linkage method and system based on intelligent monitoring of slope retaining walls.

[0005] A first aspect of this application provides a collaborative operation method based on intelligent monitoring of slope retaining walls, comprising: Acquire monitoring data of the slope retaining wall, and preprocess and extract features from the monitoring data to obtain monitoring feature vectors; The monitoring feature vector is input into a pre-trained risk identification model to obtain the current risk type and level; The monitoring feature vector and the current risk type and level are input into a pre-trained risk trend prediction model to obtain the risk evolution trend; By integrating the current risk type and level with the risk evolution trend, risk warning information is obtained; The risk warning information is matched with preset linkage rules to obtain the operation linkage strategy; The operation coordination strategy is parsed into standardized strategy instructions, and these instructions are distributed to the corresponding target department's operation execution system via a preset interface protocol. The target department's operation execution system initiates a preset business process based on the received strategy instructions and executes the corresponding strategy instructions.

[0006] A second aspect of this application provides an operational linkage system based on intelligent monitoring of slope retaining walls, comprising: The data acquisition and processing module is used to acquire monitoring data of the slope retaining wall, and to preprocess and extract features from the monitoring data to obtain monitoring feature vectors. The real-time risk identification module is used to input the monitoring feature vector into the pre-trained risk identification model to obtain the current risk type and level; The risk trend prediction module is used to input the monitoring feature vector and the current risk type and level into a pre-trained risk trend prediction model to obtain the risk evolution trend; The early warning information generation module is used to integrate the current risk type and level with the risk evolution trend to obtain risk early warning information; The linkage strategy matching module is used to match the risk warning information with preset linkage rules to obtain the operation linkage strategy; The instruction parsing and distribution module is used to parse the operation linkage strategy into standardized strategy instructions, and distribute the strategy instructions to the corresponding target department's operation execution system through a preset interface protocol. The target department execution module is used by the target department operation execution system to initiate a preset business process based on the received strategy instructions and execute the corresponding strategy instructions.

[0007] A third aspect of this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the above-described operation linkage method based on intelligent monitoring of slope retaining walls.

[0008] In a fourth aspect of this application, a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described operation linkage method based on intelligent monitoring of slope retaining walls.

[0009] The beneficial effects of the operational linkage method and system based on intelligent monitoring of slope retaining walls provided in this application are as follows: This application generates risk warning information and matches linkage rules by preprocessing, feature extraction, and risk pre-identification and trend prediction of slope retaining wall monitoring data. This forms a standardized operational linkage strategy and distributes it to the corresponding target department's execution system, realizing the full-process intelligent and automated operation linkage of slope retaining wall risk monitoring and multi-departmental operation linkage. It not only detects potential safety hazards in advance through risk pre-identification and trend prediction, avoiding accidents such as traffic interruptions and casualties caused by risk escalation, but also ensures road network traffic safety and public travel safety. Furthermore, by using standardized instructions and interface protocols between various target departments, it breaks down information barriers between target departments, further strengthening the operational linkage of the execution systems of each target department. This application improves the accuracy of operational linkage. Attached Figure Description

[0010] Figure 1A flowchart illustrating an operational linkage method based on intelligent monitoring of slope retaining walls provided in an embodiment of this application; Figure 2 This is a structural block diagram of an operation linkage system based on intelligent monitoring of slope retaining walls provided in an embodiment of this application; Figure 3 This is a schematic block diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0011] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0012] To make the purpose, technical solution, and advantages of this application clearer, the following will be described in conjunction with the appendix. Figure 1-3 The following is an explanation using specific examples.

[0013] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of an operational linkage method based on intelligent monitoring of slope retaining walls provided in this application. The method includes: S101: Obtain monitoring data of the slope retaining wall, and preprocess and extract features from the monitoring data to obtain the monitoring feature vector.

[0014] In this embodiment, the monitoring data includes three categories: the first category is structural deformation data, which includes data such as horizontal / vertical displacement, tilt angle, and wall strain of the retaining wall collected by GNSS positioning equipment, inclinometers, and strain gauges; the second category is environmental impact data, which includes data such as rainfall and groundwater seepage pressure collected by rain gauges and piezometers; and the third category is auxiliary monitoring data, which includes data such as surface cracks of the retaining wall and loosening status of the surrounding soil and rock collected by cameras and 3D laser scanners.

[0015] Due to environmental interference, transmission loss, and other factors, the collected monitoring data contains outliers, missing values, and duplicate data, thus requiring preprocessing. Specifically, firstly, outliers are removed using the 3σ principle; secondly, linear interpolation, mean imputation, or prediction imputation based on time-series models are used to supplement missing values ​​during the data collection process; and finally, data with different dimensions and numerical ranges are converted to a unified scale.

[0016] This embodiment extracts features from the preprocessed monitoring data to identify information factors representing the safety status of the slope retaining wall. Feature extraction is based on engineering experience and data analysis logic to process the data. For example, for structural deformation data, features representing the deformation strength and development trend of the retaining wall are extracted. For instance, continuous displacement data collected by GNSS positioning equipment is calculated as displacement rate and displacement acceleration, and these two features are used to distinguish between slow, uniform deformation and sudden, rapid deformation of the retaining wall. For tilt angle data collected by an inclinometer, the amplitude of tilt angle change is extracted. For wall strain data, the maximum strain value is extracted, representing the wall's stress limit and strain distribution uniformity, used to determine whether there is localized stress concentration in the wall.

[0017] For environmental impact data, extract the impact characteristics of the environmental impact data on the safety of the retaining wall. For example, for rainfall data, calculate the cumulative rainfall and rainfall intensity; for groundwater seepage pressure data, extract the maximum seepage pressure value within a set time period, i.e., the peak seepage pressure, the seepage pressure rise rate, i.e., the increase in seepage pressure per unit time, and the seepage pressure gradient, i.e., the seepage pressure difference at monitoring points at different depths of the retaining wall.

[0018] For visual monitoring data, feature extraction is performed based on image recognition and data analysis technologies. For example, for crack images captured by cameras, the crack outline is identified through edge detection algorithms, and the crack length, crack width, and crack propagation rate are extracted. For surrounding soil and rock data captured by 3D laser scanners, a 3D model is generated through point cloud processing technology, and the loosened area and displacement of the loosened soil and rock are extracted, while the surface roughness change rate is calculated.

[0019] The extracted features are integrated in the order of structural deformation features, environmental impact features, and auxiliary monitoring features to obtain a monitoring feature vector. For example, the feature vector of a certain monitoring section can be represented as: [daily displacement rate, displacement acceleration, dip angle change range, maximum strain value, 24-hour cumulative rainfall, seepage pressure rise rate, crack width, loosened area of ​​soil and rock, ...].

[0020] S102: Input the monitoring feature vector into the pre-trained risk identification model to obtain the current risk type and level.

[0021] In this embodiment, the risk identification model is based on a hybrid model that integrates gradient boosting trees and an attention mechanism. The training process of the risk identification model is based on historical monitoring data. Specifically, firstly, monitoring data of slope retaining walls in different regions and under different geological conditions over the past 5 years, along with corresponding actual safety event records, are collected to construct a training dataset. The training dataset is then divided into a training set and a validation set in a 7:3 ratio. Secondly, feature enhancement processing, including feature crossing and noise addition, is performed on the training set data. Finally, the hyperparameters of the gradient boosting tree and attention mechanism are optimized using a grid search method. For example, candidate value ranges and combinations are set for the hyperparameters of the gradient boosting tree and attention mechanism. Then, the risk level identification accuracy and risk type recall rate on the validation set are used as the core objective functions. By traversing all hyperparameter combinations and iteratively training, the identification performance of the risk identification model under different parameter combinations is continuously evaluated until the accuracy and performance of the risk identification model on the validation set tend to stabilize, ultimately obtaining the optimal hyperparameter combination. The hyperparameters of the gradient boosting tree include the learning rate, tree depth, subsample ratio, and regularization coefficient, which determine the ability of the gradient boosting tree to fit the monitored feature vector; the hyperparameters of the attention mechanism include the attention weight coefficient, the hidden dimension of the attention layer, and the dropout rate.

[0022] In the reasoning stage of the risk identification model in this embodiment, the input of the risk identification model is the monitoring feature vector. By calculating the feature weights layer by layer, the three types of features—structural deformation, environmental impact, and auxiliary monitoring—are comprehensively and weighted for evaluation. For example, when structural deformation features and environmental impact features simultaneously meet their corresponding risk thresholds, the risk identification model will increase the weights of these features through an attention mechanism, thereby improving the sensitivity of identifying high-risk levels. For cases where a single feature is close to its corresponding risk threshold but does not meet the risk judgment criteria, a judgment is made based on the established safety redundancy standards in the engineering field.

[0023] The risk identification model outputs the current risk type and risk level. The current risk type includes deformation risk, seepage risk, and soil / rock loosening risk; the risk level includes Level I (safe), Level II (low risk), Level III (medium risk), and Level IV (high risk), with a corresponding feature threshold range for each level. For example, when the displacement acceleration in the monitored feature vector is less than 0.1 mm / d... 2 When the rate of increase in seepage pressure is less than 0.5 kPa / h, the rate of crack propagation is less than 0.1 mm / day, and there are no other abnormal characteristics, a Level I safety rating is issued; when the displacement acceleration is between 0.1 and 0.3 mm / day... 2If the seepage pressure peak exceeds 80% of the preset limit, or the crack width exceeds 0.3mm, a Level II low-risk rating is output, and so on. The risk thresholds corresponding to each feature are shown in Table 1, "Risk Level Feature Thresholds and Determination Table for Slope Retaining Walls."

[0024] Table 1. Thresholds for Risk Level Characteristics of Slope Retaining Walls and Their Determination Risk level Feature threshold Level I <![CDATA[1. Displacement acceleration is less than 0.1 mm / d 2 2. The rate of increase in seepage pressure is less than 0.5 kPa / h 3. The crack propagation rate is less than 0.1 mm / day]]> Level II <![CDATA[1. Displacement acceleration ∈ [0.1, 0.3) mm / d 2 2. The peak seepage pressure is greater than or equal to 80% of the preset limit value. 3. The crack width is greater than 0.3 mm (either condition can be used for determination)]]> Level III <![CDATA[1. Displacement acceleration ∈ [0.3, lower limit of Class IV threshold) mm / d 2 2. Peak seepage pressure is greater than or equal to 90% of the preset limit 3. Crack propagation rate ∈ [0.1, lower limit of Class IV threshold) mm / day]]> Level IV <![CDATA[1. The displacement acceleration is greater than or equal to 0.5 mm / d 2 (referring to the industry critical value) 2. The peak seepage pressure is greater than or equal to 100% of the preset limit value. 3. The crack propagation rate is greater than or equal to 0.3 mm / day (any one of the conditions is satisfied for determination)]]> S103: Input the monitoring feature vector and the current risk type and level into the pre-trained risk trend prediction model to obtain the risk evolution trend.

[0025] In this embodiment, the monitoring feature vector and the current risk type and level output by the risk identification model are input into the pre-trained risk trend prediction model. By mining the dynamic evolution law of time-series monitoring features through the risk identification model and embedding the risk transmission logic in the engineering field, the risk evolution trend of the slope retaining wall is obtained.

[0026] The risk trend prediction model employs a hybrid model combining Long Short-Term Memory (LSTM) networks and Gated Recurrent Units (GRUs). The first layer of this model is an LSTM feature extraction layer, responsible for capturing the long-term temporal dependencies of monitoring feature vectors and uncovering the dynamic changes of three types of features over time: structural deformation, environmental impact, and auxiliary monitoring. The second layer is a GRU feature enhancement and fusion layer, which uses a lightweight gating structure to reduce the dimensionality and enhance the deep features output by the LSTM network. Simultaneously, based on the risk transmission rules in slope retaining wall engineering, it couples data patterns with engineering logic. The input to the risk trend prediction model consists of monitoring feature vectors and current risk type and level data, forming the input matrix. The output of the risk trend prediction model is the risk evolution trend of the slope retaining wall, which specifically includes the risk level change trend and the risk type transmission path.

[0027] The risk trend prediction model is trained using time-series monitoring data of slope retaining walls under different geological conditions and risk scenarios, along with corresponding risk evolution records. The dataset is labeled with monitoring feature time-series sequences, current risk type and level, and future risk evolution results. Then, based on chronological order, it is divided into training, validation, and test datasets in a 6:2:2 ratio. Subsequently, the labeled datasets undergo sliding window processing to generate samples with a fixed time step. During the training process, Bayesian optimization is used to iteratively optimize the model's parameters. The parameters for the long short-term memory network feature extraction layer include the number of hidden layer neurons (set to 64 / 128 / 256 candidate values) and the time step length (selected as 12 / ...). The parameters for the gated recurrent unit feature enhancement and fusion layer are set as follows: 24 / 48 monitoring cycles, forget gate bias coefficient (range 0.1-0.5), weight coefficients of update gate and reset gate (range 0.01-0.1), dropout rate (set to 0.2-0.5 to prevent overfitting), learning rate of risk trend prediction model (using a step decay strategy, initial value set to 0.001), batch size (set to 32 / 64), etc. Training parameters are also set. The training uses the mean absolute error (MAE) of risk trend prediction and the accuracy of risk level prediction on the validation set as the dual objective function. Iterative training is performed until the performance of risk trend prediction model converges. Finally, the generalization ability of the model is verified through the test set so that the prediction results of risk trend prediction model are consistent with engineering practice.

[0028] S104: Integrate the current risk type and level with the risk evolution trend to obtain risk warning information.

[0029] In this embodiment, a base weight W1 is assigned based on the current risk type and level, and a base weight W2 is assigned to the risk evolution trend, satisfying W1 + W2 = 1. The baseline value of W1 is positively correlated with the current risk level. For example, the baseline values ​​of W1 corresponding to Level I (safe), Level II (low risk), Level III (medium risk), and Level IV (high risk) are 0.7, 0.6, 0.5, and 0.8, respectively. W2 is 1 − W1, used to characterize the corrective effect of the trend on the early warning result. Based on W1 and W2, the current risk type and level and the risk evolution trend are weighted and fused to obtain risk early warning information. The risk early warning information includes the early warning level, risk type, risk impact scope, and time window for risk escalation / degradation. The warning levels are divided into four levels based on a dynamic weighted fusion of the current risk type and level, and the future risk evolution trend: General Warning, Yellow Warning, Orange Warning, and Red Warning. A General Warning corresponds to a current Level I safety status with a stable risk trend; a Yellow Warning corresponds to Level II low risk or a trend towards escalating to medium risk; an Orange Warning corresponds to Level III medium risk with a short-term risk escalation to high risk; and a Red Warning corresponds to Level IV high risk or an existing risk deterioration trend exceeding a preset risk deterioration trend threshold. The preset risk deterioration trend threshold is set based on the requirements of the slope retaining wall engineering safety specifications, historical risk event evolution data, and the rate of change of key characteristics output by the risk trend prediction model. Specifically, it is a critical value determined based on the increase in driving characteristics such as displacement acceleration, seepage pressure rise rate, and crack propagation rate per unit time, and the degree of hazard of the risk type.

[0030] Specifically, the basic weights W1 and W2 are adjusted based on the risk evolution trend and risk type. For example, the adjustment of the basic weights W1 and W2 based on the risk evolution trend includes: if the risk evolution trend is risk escalation, such as from Level II to Level III or from Level III to Level IV, then an increase coefficient k1 (k1=1.2−1.5, the coefficient is larger the shorter the escalation time window) is applied to W2, and after adjustment W2′=W2×k1, W1′=1−W2′ is simultaneously reduced, indicating the urgency of risk escalation; if the risk evolution trend is risk maintenance, then the weights remain unchanged at the baseline value; if the risk evolution trend is risk deterioration, then a decrease coefficient k2 (k2=0.6−0.8) is applied to W2, and after adjustment W2′=W2×k2, W1′ is 1-W2′. Furthermore, the correction of the basic weights W1 and W2 based on the risk evolution trend includes: if the current risk type is a high-risk type (structural deformation risk, overall slip risk), an additional risk type correction coefficient k3 (k3=1.1) is configured to further amplify the proportion of high-risk types in the integrated assessment; if the current risk type is a general-risk type (local crack risk, slight seepage risk), k3 correction is not used. For example, assuming that the current risk type of a monitoring section is Level II low risk, and the risk evolution trend is to upgrade to Level III medium risk within 24 hours, after the baseline weight and trend adjustment, the weight of the risk evolution trend W2′=0.5. Since this risk type belongs to the high-risk type, k3 is used for secondary correction, and after correction W2′′=0.5×1.1=0.55), while the weight of the current risk level becomes 1-0.55=0.45.

[0031] S105: Match the risk warning information with the preset linkage rules to obtain the operation linkage strategy.

[0032] In this embodiment, firstly, the risk warning information is structurally analyzed to obtain structured elements, and then matched with linkage rules based on the structured elements to obtain matching results. Specifically, when the matching result is a unique rule, the unique rule is used as the operation linkage strategy. When the matching result is multiple rules, the multiple rules are prioritized based on preset priority calculation rules, and the multiple rules are sorted to obtain a rule set, which is used as the operation linkage strategy. When no rule is matched, the spatiotemporal characteristics, risk type, and impact range of the risk warning information are extracted as multi-dimensional feature vectors. Based on the multi-dimensional feature vectors, a multi-dimensional search is performed in a preset historical strategy case library to obtain K historical similar cases corresponding to each feature vector, where K is a preset positive integer.

[0033] Secondly, data on the handling effects of K similar historical cases are extracted to form an initial strategy set. This initial strategy set is then subjected to multi-dimensional analysis to obtain the results. Based on the effectiveness values ​​of the effectiveness analysis results within the multi-dimensional analysis, a subset of candidate strategies with effectiveness values ​​greater than a preset first threshold is selected from the initial strategy set. Subsequently, based on the strategy cost analysis and strategy execution efficiency analysis results from the multi-dimensional analysis, the timeliness of each strategy in the candidate strategy subset is calculated to obtain its comprehensive effectiveness per unit time. Simultaneously, the matching degree between each strategy in the candidate strategy subset and the risk warning information is evaluated based on the strategy fitness analysis results from the multi-dimensional analysis. Based on the comprehensive effectiveness and matching degree, the reorganization weight of each strategy in the candidate strategy subset is determined.

[0034] Subsequently, the roulette wheel selection algorithm is used to extract several strategies from the candidate strategy subset based on each recombination weight using sampling with replacement, resulting in a candidate strategy sequence. The candidate strategy sequence is then sorted and grouped based on a preset strategy logic constraint rule base to obtain a preliminary strategy recombination sequence.

[0035] Finally, after obtaining the preliminary strategy reorganization sequence, the comprehensive effectiveness of the preliminary strategy reorganization sequence is calculated to obtain the comprehensive effectiveness prediction score; the candidate sequence with the highest comprehensive effectiveness prediction score is selected, and after logical coherence verification and departmental execution feasibility verification, the operation linkage strategy is obtained.

[0036] S106: Parse the operation linkage strategy into standardized strategy instructions, and distribute the strategy instructions to the corresponding target department's operation execution system through a preset interface protocol.

[0037] In this embodiment, standardized strategy instructions refer to the detailed instructions obtained after breaking down the operation linkage strategy, which are uniform in format, clear in semantics, and can be identified and executed by the operation execution systems of each target department; the operation execution systems of each target department refer to the dedicated business systems or equipment control systems used by each target department to receive instructions, execute operations, and provide progress feedback, including the traffic control and dispatch system of the traffic police department, the engineering machinery and equipment dispatch system (bulldozers) of the construction department, and the emergency command and dispatch system of the urban operation management / emergency management department, etc.

[0038] First, the operational coordination strategy is standardized and its instructions are broken down. For example, based on the five-tuple structure of departmental responsibility, execution action, time requirements, quantitative indicators, and safety regulations, the comprehensive handling requirements in the strategy are broken down into detailed standardized instructions. For instance, the traffic control strategy is broken down into specific instructions for the traffic police department: "Close XX road section - complete within 1 hour - set up warning signs - comply with the 'Road Traffic Safety Emergency Response Specifications'." The emergency reinforcement strategy is broken down into quantitative instructions for the construction department: "Grouting reinforcement of the retaining wall at the XX monitoring section - start within 6 hours - grouting pressure greater than or equal to 0.8MPa - follow the slope reinforcement construction safety regulations." The instruction requiring rapid clearing of loose soil and rock around the retaining wall and widening of the emergency work area is broken down into instructions for the equipment dispatch department: "Coordinate bulldozer operation - arrive at the designated monitoring section within 30 minutes - clear loose accumulation at the toe of the retaining wall greater than or equal to 20m." 3 - A specific instruction to maintain a 5-meter safe distance from the retaining wall during operations. Additionally, each instruction is assigned a unique identifier (ID), instruction priority, execution time limit, and other metadata.

[0039] Secondly, based on the technical specifications of the different target department's operation execution systems, an interface protocol adapted to the target department's operation execution system is used for instruction encapsulation and distribution. The interface protocol adopts an industrial-grade standard communication protocol. For the execution systems of construction machinery equipment such as bulldozers, real-time data such as the bulldozer's location, fuel level, and fault status are simultaneously retrieved during instruction encapsulation, ensuring that the target equipment for the strategy instruction is in a workable state. During the instruction encapsulation process, the strategy instructions are converted into structured data in JSON / XML format according to the interface protocol requirements. The structured data includes fields such as instruction identifier, execution department code, action parameters, time constraints, and feedback requirements.

[0040] Finally, after the strategy instructions are distributed, a monitoring and feedback mechanism for the distribution status is established. Specifically, the system listens for receipts from the target department's work execution system in real time. If no receipt is received within a preset time, a second distribution and alarm notification are automatically triggered. During the execution of instructions by the target department, execution progress data is transmitted back in real time via an interface protocol. The execution progress data includes 80% completion of road closure, commencement of grouting reinforcement, and clearing of 15m of loose material by bulldozers. 3 5m remaining 3 If an execution anomaly occurs, an anomaly alarm is immediately generated and pushed to the command and dispatch center. Simultaneously, the instruction distribution strategy is adjusted according to the pre-set plan, such as coordinating backup bulldozer support or optimizing the work route. The timeframe for not responding within the pre-set timeframe is determined based on a comprehensive analysis of the target department's business response characteristics, instruction priority, equipment dispatch distance, and the average response time of historical cases.

[0041] S107: The target department's operation execution system initiates a preset business process based on the received policy instructions and executes the corresponding policy instructions.

[0042] In this embodiment, after receiving the policy instruction, the target department's operation execution system first verifies the completeness, priority, and adaptability of the policy instruction. The verification includes checking whether the instruction identifier is unique, whether the execution time limit is reasonable, and whether the operation parameters comply with equipment safety specifications. Once the verification is successful, the system triggers the business process execution policy instruction built into the target department's operation execution system.

[0043] Different types of execution systems have different logics for business process initiation and strategy instruction execution. For example, for the traffic control and dispatch system of the traffic police department, after receiving instructions such as road closure and traffic diversion, it retrieves electronic map data of dangerous road sections, generates a control route planning scheme, and sends strategy instructions to the front-line execution terminal, and links with road monitoring equipment to provide real-time feedback on the control site status. For the construction machinery and equipment dispatch system (bulldozer control terminal) of the construction department, after receiving the bulldozer dispatch operation instruction, it matches the bulldozer equipment in standby state and sends control instructions including quantitative parameters such as operation location, operation route, amount of clearing work, and safe operation distance. After receiving the instructions, the bulldozer control terminal is either manually executed by the operator or automatically completed based on the operation route in the strategy instruction to clear the loose rock and soil at the toe of the slope. In addition, during the operation of the bulldozer control terminal, the progress of the project completion and the operating status of the bulldozer equipment are transmitted back in real time; for the emergency command and dispatch system of the urban operation management / emergency management department, after receiving the strategy instructions, it initiates a multi-department collaborative command process, summarizes the progress feedback data of each execution system in real time, and generates a visualized disposal progress ledger to achieve overall control of the entire operation process.

[0044] During the execution of strategy instructions, each target department's operation execution system is equipped with an abnormal handling mechanism: if abnormal situations such as strategy instruction execution timeout, equipment failure, or sudden changes in the operating environment occur, the current process is suspended and an abnormal alarm message is generated and pushed to the emergency command and dispatch system through a preset interface protocol; at the same time, the backup plan activation logic is triggered. For example, if a bulldozer suddenly fails, a backup bulldozer in the nearby area is dispatched to take over the operation.

[0045] As can be seen from the above, this application generates risk warning information and matches linkage rules by preprocessing, feature extraction, risk pre-identification, and trend prediction of slope retaining wall monitoring data. This forms a standardized operation linkage strategy and distributes it to the corresponding target department's execution system, realizing the intelligent and automated full-process linkage of slope retaining wall risk monitoring and multi-department operation. It not only detects potential safety hazards in advance through risk pre-identification and trend prediction, avoiding accidents such as traffic interruption and casualties caused by risk expansion, but also ensures the safety of road network traffic and public travel. Furthermore, by using standardized instructions and interface protocols between various target departments, it breaks down information barriers between target departments, further strengthens the operation linkage of the execution systems of various target departments, and improves the speed of risk response and the scientific nature of decision-making.

[0046] In one embodiment of this application, risk warning information is matched with preset linkage rules to obtain an operational linkage strategy, including: Structural analysis of risk warning information yields structured elements; Matching results are obtained by matching structured elements with linkage rules; When the matching result is a unique rule, then that unique rule is used as the job linkage strategy; When the matching result is multiple rules, the multiple rules are prioritized based on the preset priority calculation rules, and the multiple rules are sorted to obtain a rule set, which is then used as the job linkage strategy. If no rule is matched, a job linkage strategy is generated based on a multi-dimensional feature matching strategy.

[0047] In this embodiment, the risk warning information is first analyzed in a standardized structure, and structured elements such as warning level, risk type, scope of impact, and risk evolution time window are proposed. At the same time, each element is quantitatively labeled and semantically normalized, such as labeling the specific road segment or monitoring section coordinates corresponding to the scope of risk impact, and the start and end duration of the risk evolution time window.

[0048] Subsequently, the structured elements are matched with rule entries in a pre-defined linkage rule base. This linkage rule base is built upon the "Technical Specification for Building Slope Engineering," successful historical emergency response cases, and industry expert experience. The linkage rule base contains a one-to-one mapping relationship between multiple structured element combinations and standardized response actions. During the matching process between structured elements and rule entries in the linkage rule base, targeted matching is performed on structured elements such as warning levels and risk types; fuzzy matching is used on structured elements such as impact range boundaries and evolution time windows to obtain the matching results.

[0049] Secondly, based on the different matching results, a differentiated operation linkage strategy generation logic is adopted. For example, when the matching result is a unique rule, the responsible department, execution action, time threshold, safety regulations, and other contents corresponding to the unique rule are integrated to obtain the operation linkage strategy. When the matching result is multiple rules, the multiple rules are quantitatively sorted based on the preset priority calculation rules. For example, the priority calculation uses the warning level weight (40%), risk hazard degree weight (30%), and impact range weight (30%) as the core dimensions. The priority score of each rule is obtained by weighted summation, and then sorted from high to low based on the priority scores to form an ordered rule set. This rule set is used as the operation linkage strategy. When no rule is matched, the multi-dimensional feature matching strategy is triggered to generate the operation linkage strategy.

[0050] As can be seen from the above, this embodiment extracts structured elements from risk warning information, providing a unified and unambiguous judgment benchmark for rule matching, thereby improving the pertinence and accuracy of the generated operational linkage strategy. Based on the differentiated processing logic of the matching results, it can quickly output a standardized strategy when a unique rule is matched, and form an ordered rule set by priority sorting when multiple rules are matched. At the same time, it triggers a multi-dimensional feature matching strategy when no rule is matched, which enhances the adaptability to unconventional and sudden risk scenarios and avoids the blank area in the generation of operational linkage strategy.

[0051] In one embodiment of this application, a job linkage strategy is generated based on a multi-dimensional feature matching strategy, including: Extract multi-dimensional feature vectors of risk warning information, including its spatiotemporal characteristics, risk type, and scope of impact. Based on multi-dimensional feature vectors, a multi-dimensional search is performed in a pre-set historical strategy case library to obtain K historical similar cases corresponding to each feature vector, where K is a pre-set positive integer; Extract the treatment effect data from K similar historical cases to form an initial strategy set; A multi-dimensional analysis was conducted on the initial strategy set to obtain multi-dimensional analysis results, including strategy effectiveness analysis, strategy cost analysis, strategy execution efficiency analysis, and strategy fitness analysis. Based on the results of multi-dimensional analysis, the initial strategy set was reorganized to obtain the operation linkage strategy.

[0052] In this embodiment, firstly, the feature dimensions of risk warning information are extracted, and a multi-dimensional feature vector that can characterize the current risk scenario is constructed. The multi-dimensional feature vector includes three types of elements: spatiotemporal characteristics (location of the monitoring section where the risk occurs, geological conditions, and risk evolution time window), risk type (specific types and sub-subtypes such as deformation risk, seepage risk, and soil loosening risk), and impact range (length of the threatened road section, distribution of surrounding structures, and population density).

[0053] Secondly, based on the constructed multi-dimensional feature vectors, a multi-dimensional similarity search is performed in a pre-defined historical strategy case library. The historical strategy case library includes slope retaining wall treatment cases under different geological conditions and risk levels, and each case is labeled with complete risk characteristics, treatment strategies, and execution effect correlation information.

[0054] In the process of multi-dimensional similarity retrieval, the cosine similarity algorithm is used to calculate the similarity score between the current multi-dimensional feature vector and the feature vector of historical cases in the database, and to select the top K historical similar cases with the highest similarity. Here, K is a preset positive integer with a value range of 5-10, and its specific value is dynamically adjusted according to the size of the case database.

[0055] Subsequently, details of the handling strategies and corresponding handling effect data were extracted from K similar historical cases and integrated to obtain an initial strategy set. The handling strategy details include the responsible department, execution actions, equipment scheduling plan, time thresholds, etc., while the handling effect data includes indicators such as risk control success rate, resource consumption cost, handling time, and secondary risk occurrence rate.

[0056] Next, a multi-dimensional quantitative analysis was conducted on the initial strategy set to obtain multi-dimensional analysis results. Specifically, the analysis included four categories: the first category was strategy effectiveness analysis, which used the risk control success rate of historical cases as an indicator to evaluate the effectiveness of the strategy in handling similar risks; the second category was strategy cost analysis, which statistically analyzed the consumption of resources such as manpower, equipment, and materials in the cases to calculate the economic cost of the strategy; the third category was strategy execution efficiency analysis, which measured the entire process time from strategy initiation to risk stabilization control and calculated the timeliness of the strategy; and the fourth category was strategy adaptability analysis, which compared the spatiotemporal characteristics, risk intensity, and resource allocation differences between historical cases and the current risk scenario to evaluate the scenario matching degree of the strategy.

[0057] Finally, based on the results of multi-dimensional analysis, the initial strategy set was reorganized to obtain the job linkage strategy.

[0058] From the above, it can be concluded that this embodiment constructs a multi-dimensional feature vector based on the spatiotemporal characteristics, risk type, and impact scope of risk warning information. By retrieving similarity data from a historical strategy case library and drawing on successful historical handling experience, it avoids the blindness in strategy formulation. Furthermore, by conducting multi-dimensional quantitative analysis of the effectiveness, cost, efficiency, and adaptability of the initial strategy set, it achieves the identification and selection of the merits of historical strategies. Simultaneously, based on the analysis results, it reorganizes strategies, integrating the handling advantages of different historical cases, solving the problem of strategy generation when there are no pre-set rules for matching, filling the gap in handling strategies for unconventional and sudden risk scenarios, and further improving the comprehensiveness and intelligence level of slope retaining wall risk management.

[0059] In one embodiment of this application, based on the results of multi-dimensional analysis, the initial strategy set is reorganized to obtain a job linkage strategy, including: Based on the validity value of the validity analysis results in the multi-dimensional analysis results, a subset of candidate strategies with validity values ​​greater than a preset first threshold is selected from the initial strategy set; Based on the strategy cost analysis results and strategy execution efficiency analysis results in the multi-dimensional analysis results, the timeliness calculation is performed on each strategy in the candidate strategy subset to obtain its comprehensive efficiency within a unit timeliness. The degree of matching between each strategy in the candidate strategy subset and the risk warning information is evaluated based on the strategy fitness analysis results in the multi-dimensional analysis results. Based on comprehensive effectiveness and matching degree, the reorganization weight of each strategy in the candidate strategy subset is determined; Based on the recombination weights, the roulette wheel selection algorithm is used to perform multiple samplings with replacement from the candidate strategy subset to form a candidate strategy sequence. Based on a pre-defined policy logic constraint rule base, the candidate policy sequences are sorted and grouped to obtain a preliminary policy recombination sequence; The overall effectiveness of the initial strategy recombination sequence is calculated to obtain the overall effectiveness prediction score; The candidate sequence with the highest comprehensive performance prediction score is selected, and after logical coherence verification and departmental execution feasibility verification, the operation linkage strategy is obtained.

[0060] In this embodiment, firstly, using strategy effectiveness as the screening criterion, strategies with effectiveness values ​​greater than a preset first threshold are selected from the initial strategy set based on the effectiveness analysis results in the multi-dimensional analysis results, forming a subset of candidate strategies. The setting of this first threshold is based on the average success rate of historical cases and engineering safety standards.

[0061] Secondly, the evaluation of the candidate strategy subset includes two aspects: On the one hand, based on the strategy cost analysis results (data on resource consumption such as manpower, equipment, and materials) and the strategy execution efficiency analysis results (the time from strategy initiation to risk stabilization), the comprehensive efficiency of each candidate strategy within a unit of time is obtained through the formula: comprehensive efficiency = effectiveness value / (handling time × resource cost coefficient). This comprehensive efficiency within a unit of time achieves a dual measurement of the strategy's economy and timeliness. On the other hand, based on the strategy adaptability analysis results, the degree of matching between each strategy in the candidate strategy subset and the current risk warning information is quantitatively evaluated from the dimensions of the risk scenario's spatiotemporal characteristics, risk intensity, and resource allocation conditions.

[0062] Subsequently, based on the comprehensive effectiveness per unit time and the degree of scenario matching, a weighted assignment method is used to determine the reorganization weight of each strategy in the candidate strategy subset. The weight ratio of comprehensive effectiveness and matching degree can be adjusted according to the actual needs of the project. Based on this reorganization weight, a roulette wheel selection algorithm is used to perform multiple samplings with replacement from the candidate strategy subset. The sampling probability is positively correlated with the strategy reorganization weight, thereby selecting the disposal actions and forming a candidate strategy sequence.

[0063] Next, a pre-defined policy logic constraint rule base is invoked to optimize the candidate policy sequence. This rule base is constructed based on the engineering execution logic and departmental collaboration specifications for slope and retaining wall risk management, such as constraints like monitoring and early warning before on-site handling, personnel evacuation before equipment operation, and emergency hazard mitigation before reinforcement and repair. Based on the policy logic constraint rule base, the handling actions in the candidate policy sequence are ordered chronologically and grouped by responsible departments to obtain the execution order and collaborative relationships of each action, generating a preliminary policy reorganization sequence.

[0064] Finally, the preliminary strategy reorganization sequence undergoes comprehensive performance estimation and dual verification. For example, firstly, based on indicators such as the effectiveness, cost, and efficiency of each action in the sequence, a comprehensive performance estimation score is calculated for the preliminary strategy reorganization sequence. The candidate sequence with the highest comprehensive performance estimation score is selected, and then its logical coherence and departmental execution feasibility are verified. The logical coherence verification verifies whether the sequence of actions conforms to engineering logic and risk evolution patterns, avoiding process conflicts. The departmental execution feasibility verification assesses the strategy's implementation capability by considering the actual situation of each responsible department, such as staffing, equipment standby status, and response radius. After dual verification and without objection, the final operational coordination strategy is obtained.

[0065] From the above, it can be concluded that this embodiment uses a first threshold to screen out inefficient strategies, enabling candidate strategies to possess basic risk management capabilities. Simultaneously, it calculates the comprehensive efficiency per unit time based on cost and efficiency, and determines the reorganization weight by associating it with the degree of scenario matching, thus achieving precise quantification of strategy value. Furthermore, through the roulette wheel algorithm and the strategy logic constraint rule library, and based on logical coherence verification and departmental execution feasibility verification, an operational linkage strategy is obtained. This strategy integrates the advantages of handling historical cases and is highly adapted to the actual needs of the current risk scenario, effectively improving the implementability of the operational linkage strategy and the accuracy of risk management.

[0066] In addition, feedback data of the operation linkage strategy generated under the first threshold is obtained, and the feedback data is quantitatively evaluated to obtain the actual effectiveness value of the strategy; the actual effectiveness value of the strategy is compared with the comprehensive effectiveness prediction score before the corresponding strategy is generated to calculate the effectiveness deviation value; based on the effectiveness deviation value, the rationality of the current first threshold setting is judged; if the effectiveness deviation value continues to be greater than the preset deviation upper limit, the first threshold is adaptively lowered; if the effectiveness deviation value continues to be less than the preset deviation lower limit, the first threshold is adaptively raised.

[0067] The magnitude of the first threshold adjustment is based on the absolute value of the performance deviation. The actual performance value of the strategy is a comprehensive performance score derived from quantitative analysis and weighted calculation of on-site feedback data collected after the operational linkage strategy is issued and executed. This score objectively measures the actual effect of the strategy in reality. The comprehensive performance prediction is divided into ideal and optimal strategy performance values. The performance deviation is the difference or ratio between the actual performance value and the comprehensive performance prediction score. It quantifies the gap between the predicted and actual results, and its calculation formula is actual value - predicted score or (actual value - predicted score) / predicted score. The preset upper and lower limits of the deviation are set based on the statistical distribution (mean ± standard deviation) of historical performance deviation values ​​and the tolerance for prediction errors (acceptable maximum optimistic / pessimistic deviation).

[0068] In one embodiment of this application, the task linkage strategy is parsed into standardized strategy instructions, and the strategy instructions are distributed to the corresponding target department's task execution system through a preset interface protocol, including: Based on the operation linkage strategy, the target department's operation execution system is determined from the preset department system responsibility-action mapping table, and its corresponding instruction coding standard and data interface protocol are determined. Based on the instruction coding standard, the natural language descriptions or logical parameters in the operation linkage strategy are converted into structured instruction data packets that conform to the target department's operation execution system. The structured instruction data packet is sent to the corresponding target execution department system via a data interface protocol.

[0069] In this embodiment, firstly, the handling actions, risk spatial location information, and estimated impact range are extracted from the operational coordination strategy. Then, the handling actions are matched with the departmental system responsibility-action mapping table to obtain a set of candidate departmental operational execution systems associated with the handling actions. Based on the risk spatial location information and estimated impact range, a geographic information system service is invoked to overlay spatial jurisdiction onto the set of candidate departmental operational execution systems to obtain spatially matched departments. Simultaneously, a directed graph model is constructed for each handling action in the operational coordination strategy, and the actions are sorted using a topological sorting algorithm to obtain the logical dependencies between them. By summarizing the spatially matched departments and logical dependencies, the target department's operational execution system is obtained, and the instruction encoding specifications and data interface protocols corresponding to each target execution department in the target department's operational execution system are acquired. The target department's operational execution system includes multiple target execution departments.

[0070] Secondly, based on the instruction coding standard, the operation linkage strategy is standardized, parsed, and structurally transformed. This includes the comprehensive handling requirements described in natural language within the operation linkage strategy (e.g., dispatching a bulldozer to clear loose material at the toe of the slope at monitoring section XX within 30 minutes, with a depth greater than or equal to 20m). 3 The XX retaining wall grouting reinforcement operation will be initiated within 6 hours (grouting pressure greater than or equal to 0.8MPa). Based on the five-tuple structure of execution subject - core action - time threshold - quantitative indicator - safety specifications, the operation will be broken down into subdivided execution items. For the logical parameters in the operation linkage strategy (instruction priority, execution sequence constraints, equipment safety operation thresholds), numerical annotation and labeling will be performed according to coding specifications (red warnings will be labeled as priority 1, and yellow warnings as priority 2). Finally, a structured instruction data package that meets the semantic parsing requirements of the target department's operation execution system will be formed. This structured instruction data package is semantically clear, parameterized, and formatted, and can be recognized by the target department's operation execution system. It includes core fields such as a unique instruction identifier (ID), execution department code, action parameter details, time constraints, safety operation specifications, and feedback data requirements.

[0071] Finally, a communication link is established through the data interface protocol corresponding to each target execution department in the target department's work execution system. Before distributing the structured instruction data packet, the structured instruction data packet undergoes dual security processing. Specifically, SSL encryption technology is used to ensure privacy and security during data transmission, and an MD5 checksum mechanism is used to verify the integrity of the structured instruction data packet. For the control system of special equipment such as bulldozers, real-time status data of the equipment (current location, fuel level, fault alarm information) is retrieved synchronously before distribution, so that the instructions related to the control system of the special equipment in the structured instruction data packet are issued to the bulldozer equipment in the standby state. The structured instruction data packet is then sent to the corresponding target execution department system through its dedicated communication link and adapted interface protocol. At the same time, a distribution status monitoring mechanism is activated to monitor the reception and acknowledgment of the target execution department system in real time.

[0072] As can be seen from the above, this embodiment determines the target execution system and its corresponding coding specifications and interface protocols by using a preset departmental system responsibility-action mapping table. This avoids compatibility issues in cross-system instruction transmission and improves the accuracy of structured instruction data packet distribution. Furthermore, converting the natural language descriptions and logical parameters in the strategy into structured instruction data packets reduces the parsing cost and risk of erroneous execution of the target department's operation execution system. At the same time, sending structured instruction data packets through an adapted interface protocol ensures the stability and security of structured instruction data packet transmission, further improving the multi-departmental collaborative efficiency and strategy execution synergy in slope retaining wall risk management.

[0073] In one embodiment of this application, based on a job linkage strategy, the target department job execution system is determined from a preset department system responsibility-action mapping table, and its corresponding instruction coding specification and data interface protocol are determined, including: Extract the response actions, spatial location information of risks, and estimated impact range of the coordinated operation strategy; Match the action to the departmental system responsibility-action mapping table to obtain a set of candidate departmental operation execution systems associated with the action; Based on risk spatial location information and estimated impact range, geographic information system services are invoked to overlay spatial jurisdiction on the candidate department operation execution system set to obtain spatially matched departments. A directed graph model is constructed for each action in the operation linkage strategy, and the actions are sorted using a topological sorting algorithm to obtain the logical dependencies between the actions. By summarizing the spatial matching departments and logical dependencies, the target department's operation execution system is obtained; Based on the target department operation execution system, obtain the instruction coding specifications and data interface protocols corresponding to each target execution department in the target department operation execution system, where the target department operation execution system includes multiple target execution departments.

[0074] In this embodiment, firstly, the handling actions, risk spatial location information, and estimated impact range are extracted from the operational coordination strategy. The handling actions include specific execution actions such as traffic control, bulldozing slope toe clearing, retaining wall grouting reinforcement, and personnel evacuation; the risk spatial location information includes the road section where the slope retaining wall is located, the precise coordinates of the monitoring section, and the geographical boundary of the risk-occurring area; the estimated impact range includes the scope of the threatened surrounding roads, structures, and areas of personnel activity.

[0075] Secondly, the actions are matched one by one with a pre-defined departmental system responsibility-action mapping table. This departmental system responsibility-action mapping table predefines the association between various actions and the corresponding responsible departments and dedicated operation execution systems. For example, traffic control actions correspond to the traffic dispatch system of the traffic police department, bulldozer equipment operations correspond to the special equipment control system of the construction department, and emergency reinforcement actions correspond to the equipment dispatch system of the geotechnical engineering operation department. Through the mapping of actions and responsibilities, a preliminary set of candidate departmental operation execution systems associated with each action is obtained.

[0076] Next, based on the spatial location information of the risk and the estimated impact range, a spatial jurisdiction overlay analysis was conducted using Geographic Information System (GIS) services. By spatially overlaying and verifying the coordinates of the risk area and the boundary of the impact range with the administrative jurisdiction and operational responsibility area of ​​each candidate department, departments and systems that do not have jurisdiction over the area or the scope of their operations were eliminated, and spatially matching departments that simultaneously meet the requirements of functional matching and spatial jurisdiction matching were selected.

[0077] After obtaining the spatially matched departments, this embodiment performs directed graph modeling and logical sorting of all actions in the operational linkage strategy. Specifically, each action is abstracted as an independent node in a directed graph. Based on a preset action constraint rule base, the logical relationships between actions are identified and directed edges are established to construct a directed graph model. The nodes are linearly sorted using a topological sorting algorithm to obtain the logical dependencies between actions, providing a basis for the timing of multi-department collaborative execution.

[0078] Secondly, the logical dependencies between spatial matching departments and handling actions are summarized and integrated. For example, based on the collaborative execution requirements of each handling action in the logical dependencies, all responsible departments and corresponding operation execution systems required to complete the entire process are identified, and duplicate or redundant department systems are eliminated. Finally, the target department operation execution systems that include all handling actions, meet spatial jurisdiction requirements, and adapt to logical collaboration requirements are determined. The target department operation execution systems include multiple target execution departments and their dedicated systems, such as the traffic control and dispatch system of the traffic police department, the special equipment control system of the construction department, and the emergency command and dispatch system of emergency management.

[0079] Finally, based on the target department's operation execution system, the instruction coding specifications and data interface protocols corresponding to each target execution department in the target department's operation execution system are obtained. The instruction coding specifications characterize the field definitions, data types, format requirements, and semantic rules of the structured instruction data packets. For example, bulldozer operation instructions in structured instruction data packets include specific fields such as operation location coordinates, cleanup volume, and safety distance. The data interface protocols are adapted according to the differences in department system types. For example, general business systems adapt to the RESTful API protocol, special equipment control systems such as bulldozers adapt to the ModbusTCP industrial-grade protocol, and IoT terminals adapt to the MQTT lightweight protocol.

[0080] From the above, it can be concluded that this embodiment, by extracting the disposal actions, risk spatial location, and impact scope, and based on the departmental system responsibility-action mapping table and GIS spatial jurisdiction overlay analysis, not only improves the adaptability of the target department's functions and the accuracy of territorial jurisdiction, but also avoids departmental overlap or jurisdictional misalignment. Directed graph modeling and topological sorting of the disposal actions yields the logical dependencies of multi-departmental collaboration, providing a basis for the timing planning of subsequent structured instruction data packet distribution. Furthermore, by retrieving the target department's operation execution system's proprietary instruction coding specifications and data interface protocols, the accuracy, collaboration, and feasibility of structured instruction data packet distribution are improved.

[0081] In one embodiment of this application, a directed graph model is performed on each action in the job linkage strategy, and the actions are sorted using a topological sorting algorithm to obtain the logical dependencies between the actions, including: The various actions in the operation linkage strategy are abstracted into nodes of a directed graph, resulting in a set of action nodes; Based on the set of action nodes and the preset action constraint rule base, the prerequisite constraints, resource dependency constraints and security mutual exclusion constraints between each action node are identified to obtain the set of constraint relationships. Based on the set of constraint relationships, corresponding directed edges are established between the corresponding action nodes to obtain a directed graph model; Based on the preset priority rules for handling actions, priority weights are calculated for parallel handling action nodes in the directed graph model that have no direct dependencies. Based on the directed graph model and priority weights, a linear execution sequence is obtained through topological sorting algorithm. By reverse analysis of the predecessor and successor relationships between adjacent action nodes in a linear execution sequence, the logical dependencies between the actions are obtained.

[0082] In this embodiment, firstly, the various disposal actions decomposed from the operation linkage strategy are abstracted, and each independent disposal action (traffic control, bulldozer slope clearing, retaining wall grouting reinforcement, personnel evacuation, etc.) is defined as an independent action node in a directed graph, and all independent action nodes together constitute an action node set.

[0083] Secondly, based on the set of action nodes, a pre-defined action constraint rule base is retrieved. This rule base is based on three types of constraint logic in the slope retaining wall risk management scenario: the first type is the prerequisite constraint (traffic control must be completed before bulldozer operations can commence), the second type is the resource dependency constraint (grouting reinforcement requires priority for equipment scheduling and material transportation), and the third type is the safety mutual exclusion constraint (heavy equipment operations are prohibited in areas where personnel have not been evacuated). By matching the rules with the action constraint rule base, the constraint relationships between various action nodes are identified, forming a set of constraint relationships. Based on the set of constraint relationships, directed edges are established between action nodes with dependencies. The direction of the directed edges strictly corresponds to the execution order of the actions. For example, a traffic control node pointing to a bulldozer slope toe clearing node indicates the pre- and post-action relationships of this action node, resulting in a directed graph model representing the logical relationships of all management actions.

[0084] Next, for parallel action nodes in the directed graph model that have no direct dependencies, the priority weight of each parallel action node is calculated by weighted scoring based on the preset action priority rules (with risk mitigation urgency, personnel safety priority, and equipment operation efficiency as core indicators). This priority weight represents the priority execution order of the parallel actions.

[0085] Subsequently, the directed graph model is combined with priority weights, and a topological sorting algorithm is used to linearly sort all action nodes. During the execution of the topological sorting algorithm, action nodes without prerequisite dependencies are traversed first, and then subsequent related nodes are processed in turn. At the same time, parallel nodes are sorted based on priority weights, and finally a linear execution sequence that meets the logical constraints and priority requirements is obtained.

[0086] Finally, the linear execution sequence is reverse-analyzed to sort out the predecessor and successor relationships between adjacent action nodes. For example, traffic control is the predecessor action of bulldozer slope clearing, and bulldozer slope clearing is the predecessor action of retaining wall grouting reinforcement. In this way, the logical link of the linear sequence is restored, and the complete and accurate logical dependency relationship between each disposal action is finally obtained.

[0087] This embodiment also includes adjusting the identification parameters in the risk identification model. Specifically, a validation set is constructed based on historical risk identification data. The validation set includes monitoring feature vector samples labeled with the actual risk type and level. The risk identification model is evaluated using the validation set, and the precision and recall of the identification parameters are calculated at different identification thresholds. The harmonic mean of precision and recall, F1, is used as the evaluation index to plot the correspondence curve between the identification parameters and the F1 value. Based on the correspondence curve, the threshold interval that makes the F1 value globally optimal is identified and located. If the threshold interval includes multiple consecutive values, the median of the interval is selected as the updated identification parameter. If the threshold interval is unclear, the identification parameters are adjusted based on the business tolerance of risk warning, prioritizing recall or precision. The threshold interval is a continuous range of threshold values ​​on the correspondence curve between the identification threshold and the F1 value that makes the F1 value optimal. The recall priority principle prioritizes false alarms over false negatives, while the precision priority principle prioritizes reliable alarms to avoid resource waste.

[0088] As can be seen from the above, this embodiment abstracts each disposal action into directed graph nodes and identifies three types of constraints—prerequisites, resource dependencies, and security mutual exclusion—based on the action constraint rule base, thus constructing logical connections that fit the actual engineering situation. By calculating the priority weights of parallel action nodes and then generating a linear execution sequence through a topological sorting algorithm, the logical rigor of action execution is ensured, and the priority order of parallel tasks is clarified, avoiding process conflicts and resource waste when multiple departments collaborate. At the same time, by reverse parsing the predecessor and successor relationships of the linear sequence, the logical dependency relationship of disposal actions is formed, improving the collaborative efficiency and safety controllability of slope retaining wall risk disposal operations.

[0089] Corresponding to the operation linkage method based on intelligent monitoring of slope retaining walls in the above embodiment, Figure 2 This is a structural block diagram of an operational linkage system based on intelligent monitoring of slope retaining walls, provided as an embodiment of this application. For ease of explanation, only the parts relevant to the embodiment of this application are shown. References Figure 2 The operation linkage system 20 based on intelligent monitoring of slope retaining walls includes: a data acquisition and processing module 21, a real-time risk identification module 22, a risk trend prediction module 23, an early warning information generation module 24, a linkage strategy matching module 25, an instruction parsing and distribution module 26, and a target department execution module 27.

[0090] Among them, the data acquisition and processing module 21 is used to acquire monitoring data of the slope retaining wall, and to preprocess and extract features from the monitoring data to obtain monitoring feature vectors. The real-time risk identification module 22 is used to input the monitoring feature vector into the pre-trained risk identification model to obtain the current risk level; The risk trend prediction module 23 is used to input the monitoring feature vector and the current risk level into the pre-trained risk trend prediction model to obtain the risk evolution trend; The early warning information generation module 24 is used to obtain risk early warning information by combining the current risk level with the risk evolution trend; The linkage strategy matching module 25 is used to match risk warning information with preset linkage rules to obtain the operation linkage strategy; The instruction parsing and distribution module 26 is used to parse the operation linkage strategy into standardized strategy instructions and distribute the strategy instructions to the corresponding target department's operation execution system through a preset interface protocol. The target department execution module 27 is used by the target department operation execution system to initiate a preset business process based on the received strategy instructions and execute the corresponding strategy instructions.

[0091] See Figure 3 , Figure 3 This is a schematic block diagram of an electronic device provided according to an embodiment of this application. Figure 3 The electronic device 300 in this embodiment may include one or more processors 301, one or more input devices 302, one or more output devices 303, and one or more memories 304. The processors 301, input devices 302, output devices 303, and memories 304 communicate with each other via a communication bus 305. The memories 304 store computer programs, including program instructions. The processors 301 execute the program instructions stored in the memories 304. Specifically, the processors 301 are configured to invoke the program instructions to perform the functions of the modules in the aforementioned device embodiments, for example... Figure 2 The functions of the data acquisition and processing module 21, real-time risk identification module 22, risk trend prediction module 23, early warning information generation module 24, linkage strategy matching module 25, instruction parsing and distribution module 26, and target department execution module 27 are shown.

[0092] It should be understood that, in the embodiments of this application, the processor 301 may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0093] Input device 302 may include a touchpad, a fingerprint sensor (for collecting the user's fingerprint information and fingerprint orientation information), a microphone, etc., and output device 303 may include a display (LCD, etc.), a speaker, etc.

[0094] The memory 304 may include read-only memory and random access memory, and provides instructions and data to the processor 301. A portion of the memory 304 may also include non-volatile random access memory. For example, the memory 304 may also store device type information.

[0095] In specific implementations, the processor 301, input device 302, and output device 303 described in the embodiments of this application can execute the implementation methods described in any embodiment of the operation linkage method based on intelligent monitoring of slope retaining walls provided in the embodiments of this application, or they can execute the implementation methods of the electronic devices described in the embodiments of this application, which will not be repeated here.

[0096] In another embodiment of this application, a computer-readable storage medium is provided. This computer-readable storage medium stores a computer program, which includes program instructions. When executed by a processor, the program instructions implement all or part of the processes in the methods described above. Alternatively, the computer program can instruct related hardware to complete the process. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include any entity or device capable of carrying computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0097] The computer-readable storage medium can be an internal storage unit of the electronic device in any of the foregoing embodiments, such as a hard disk or memory of the electronic device. The computer-readable storage medium can also be an external storage device of the electronic device, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the electronic device. Furthermore, the computer-readable storage medium can include both internal and external storage units of the electronic device. The computer-readable storage medium is used to store computer programs and other programs and data required by the electronic device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0098] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0099] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the electronic devices and units described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0100] In the several embodiments provided in this application, it should be understood that the disclosed electronic devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces or units, or it may be an electrical, mechanical, or other form of connection.

[0101] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of this application, depending on actual needs.

[0102] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0103] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A collaborative operation method based on intelligent monitoring of slope retaining walls, characterized in that, include: Acquire monitoring data of the slope retaining wall, and preprocess and extract features from the monitoring data to obtain monitoring feature vectors; The monitoring feature vector is input into a pre-trained risk identification model to obtain the current risk type and level; The monitoring feature vector and the current risk type and level are input into a pre-trained risk trend prediction model to obtain the risk evolution trend; By integrating the current risk type and level with the risk evolution trend, risk warning information is obtained; The risk warning information is matched with preset linkage rules to obtain the operation linkage strategy; The operation linkage strategy is parsed into standardized strategy instructions, and the strategy instructions are distributed to the corresponding target department's operation execution system through a preset interface protocol; The target department's operation execution system initiates a preset business process based on the received strategy instructions and executes the corresponding strategy instructions.

2. The operational linkage method based on intelligent monitoring of slope retaining walls according to claim 1, characterized in that, The step of matching the risk warning information with preset linkage rules to obtain an operational linkage strategy includes: Structural analysis is performed on the aforementioned risk warning information to obtain structured elements; The structured elements are matched with the linkage rule engine to obtain the matching result; When the matching result is a unique rule, then the unique rule is used as the job linkage strategy; When the matching result is multiple rules, the multiple rules are prioritized based on the preset priority calculation rules, and the multiple rules are sorted to obtain a rule set, which is then used as the job linkage strategy. If no rule is matched, the job linkage strategy is generated based on the multi-dimensional feature matching strategy.

3. The operational linkage method based on intelligent monitoring of slope retaining walls according to claim 2, characterized in that, The method for generating the job linkage strategy based on multi-dimensional feature matching includes: Extract multi-dimensional feature vectors of the spatiotemporal characteristics, risk type, and scope of impact of the risk warning information; Based on the multidimensional feature vector, a multidimensional search is performed in the preset historical strategy case library to obtain K historical similar cases corresponding to each feature vector, where K is a preset positive integer; Extract the treatment effect data from the K historical similar cases to form an initial strategy set; A multi-dimensional analysis is performed on the initial strategy set to obtain multi-dimensional analysis results, including strategy effectiveness analysis, strategy cost analysis, strategy execution efficiency analysis, and strategy fitness analysis. Based on the results of the multi-dimensional analysis, the initial strategy set is reorganized to obtain the job linkage strategy.

4. The operation linkage method based on intelligent monitoring of slope retaining walls according to claim 3, characterized in that, Based on the multi-dimensional analysis results, the initial strategy set is reorganized to obtain the job linkage strategy, including: Based on the validity value of the validity analysis results in the multi-dimensional analysis results, a subset of candidate strategies with validity values ​​greater than a preset first threshold is selected from the initial strategy set; Based on the strategy cost analysis results and strategy execution efficiency analysis results in the multi-dimensional analysis results, the timeliness calculation is performed on each strategy in the candidate strategy subset to obtain its comprehensive efficiency within a unit timeliness. The degree of matching between each strategy in the candidate strategy subset and the risk warning information is evaluated based on the strategy fitness analysis results in the multi-dimensional analysis results. Based on the overall effectiveness and the matching degree, the reorganization weight of each strategy in the candidate strategy subset is determined; Based on the recombination weights, the roulette wheel selection algorithm is used to perform multiple samplings with replacement from the candidate strategy subset to form a candidate strategy sequence. The candidate strategy sequences are sorted and grouped based on a preset strategy logic constraint rule base to obtain a preliminary strategy recombination sequence; The overall effectiveness of the preliminary strategy recombination sequence is calculated to obtain an overall effectiveness prediction score; The candidate sequence with the highest comprehensive performance prediction score is selected, and after logical coherence verification and departmental execution feasibility verification, the operation linkage strategy is obtained.

5. The operational linkage method based on intelligent monitoring of slope retaining walls according to claim 1, characterized in that, The step of parsing the operation linkage strategy into standardized strategy instructions and distributing the strategy instructions to the corresponding target department's operation execution system through a preset interface protocol includes: Based on the aforementioned operation linkage strategy, the target department's operation execution system is determined from the preset department system responsibility-action mapping table, and its corresponding instruction coding standard and data interface protocol are determined. Based on the instruction encoding standard, the natural language description or logical parameters in the operation linkage strategy are converted into structured instruction data packets that conform to the target department's operation execution system. The structured instruction data packet is sent to the corresponding target execution department system through the data interface protocol.

6. The operational linkage method based on intelligent monitoring of slope retaining walls according to claim 5, characterized in that, Based on the aforementioned task linkage strategy, the target department's task execution system is determined from a preset department system responsibility-action mapping table, and its corresponding instruction coding standard and data interface protocol are determined, including: Extract the handling actions, risk spatial location information, and estimated impact range of the operation linkage strategy; The action is matched with the department system responsibility-action mapping table to obtain a set of candidate department operation execution systems associated with the action. Based on the risk spatial location information and the estimated impact range, geographic information system services are invoked to overlay spatial jurisdiction over the candidate department operation execution system set to obtain spatially matched departments. A directed graph model is constructed for each action in the operation linkage strategy, and the actions are sorted using a topological sorting algorithm to obtain the logical dependencies between the actions. The spatial matching departments and the logical dependencies are summarized to obtain the target department's operation execution system; Based on the target department operation execution system, obtain the instruction encoding specification and data interface protocol corresponding to each target execution department in the target department operation execution system, wherein the target department operation execution system includes multiple target execution departments.

7. The operational linkage method based on intelligent monitoring of slope retaining walls according to claim 6, characterized in that, The process of modeling a directed graph of each action in the operation linkage strategy and sorting them using a topological sorting algorithm to obtain the logical dependencies between the actions includes: The various actions in the operation linkage strategy are abstracted into nodes of a directed graph to obtain a set of action nodes; Based on the set of action nodes and the preset action constraint rule base, the prerequisite constraints, resource dependency constraints and security mutual exclusion constraints between each action node are identified to obtain a set of constraint relationships. Based on the set of constraints, corresponding directed edges are established between the corresponding action nodes to obtain a directed graph model; Based on the preset priority rules for handling actions, priority weights are calculated for parallel handling action nodes in the directed graph model that have no direct dependencies. Based on the directed graph model and the priority weights, a linear execution sequence is obtained by processing with a topological sorting algorithm; By reverse analysis of the predecessor and successor relationships between adjacent action nodes in the linear execution sequence, the logical dependencies between the disposal actions are obtained.

8. A collaborative operation system based on intelligent monitoring of slope retaining walls, characterized in that, include: The data acquisition and processing module is used to acquire monitoring data of the slope retaining wall, and to preprocess and extract features from the monitoring data to obtain monitoring feature vectors. The real-time risk identification module is used to input the monitoring feature vector into the pre-trained risk identification model to obtain the current risk type and level; The risk trend prediction module is used to input the monitoring feature vector and the current risk type and level into a pre-trained risk trend prediction model to obtain the risk evolution trend; The early warning information generation module is used to integrate the current risk type and level with the risk evolution trend to obtain risk early warning information; The linkage strategy matching module is used to match the risk warning information with preset linkage rules to obtain the operation linkage strategy; The instruction parsing and distribution module is used to parse the operation linkage strategy into standardized strategy instructions, and distribute the strategy instructions to the corresponding target department's operation execution system through a preset interface protocol. The target department execution module is used by the target department operation execution system to initiate a preset business process based on the received strategy instructions and execute the corresponding strategy instructions.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 7.

Images