Early sensing, predicting and early warning method and system for icing of micro-terrain road

Abstract

The invention relates to the technical field of road safety analysis, in particular to a microtopography road icing early sensing, predicting and early warning method and system. The method comprises the following steps: carrying out microtopography road monitoring and correction on target microtopography road area data through a multi-source heterogeneous monitoring device, and generating corrected microtopography road monitoring data; analyzing road surface terrain icing meteorological influence characteristic data based on the corrected micro-terrain road surface monitoring data; performing dynamic downscaling microtopography pavement climate trend simulation processing on the corrected microtopography pavement monitoring data and the road weather forecast data to generate microtopography pavement climate trend simulation data; establishing a micro-terrain pavement icing thickness prediction model; and performing microtopography road icing early sensing prediction processing on the microtopography road climate trend simulation data by using the microtopography road icing thickness prediction model to generate microtopography road icing early sensing prediction data. According to the invention, accurate prediction of early sensing of icing of the micro-terrain road is realized.

Description

Technical Field

[0001] This invention relates to the field of road safety analysis technology, and in particular to a method and system for early perception, prediction and warning of road icing in micro-topography. Background Technology

[0002] Winter icing on micro-topographic roads poses a significant threat to road traffic safety. Due to variations in terrain undulations, vegetation cover, and roadbed materials, the local meteorological environment of micro-topographic roads differs significantly from the broader regional meteorological environment, resulting in icing formation patterns and development trends that differ drastically from those of ordinary roads. In recent years, with the development of intelligent transportation and road meteorological monitoring technologies, road surface condition monitoring equipment, meteorological forecasting models, and icing prediction algorithms have been widely applied in the field of road icing early warning. These technologies have evolved from simple visual monitoring and qualitative judgment to multi-source data fusion and semi-quantitative analysis, providing technical support for the early detection and warning of road icing. However, existing micro-topography road icing early warning technologies do not incorporate targeted node design based on the unique characteristics of micro-topography. This leads to issues such as missed detections in extreme areas and insufficient data representativeness. Abnormal data generated by monitoring nodes due to extreme weather, equipment vibration, and signal interference also affect the accuracy of subsequent icing predictions. Furthermore, the direct use of regional meteorological forecast data for road icing analysis in meteorological simulation fails to accurately simulate the climate development trends of micro-topography road surfaces and makes it difficult to achieve early perception and accurate prediction of icing thickness. Consequently, the early warning and targeted nature of these technologies are insufficient, failing to provide effective data support for road maintenance and traffic management. Summary of the Invention

[0003] Based on this, the present invention provides a method and system for early perception, prediction and warning of road icing in micro-topography, in order to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a method for early sensing, prediction, and warning of road icing in micro-topography includes the following steps: Step S1: Obtain target micro-topography road area data; perform micro-topography road monitoring processing on the target micro-topography road area data using multi-source heterogeneous monitoring equipment to generate micro-topography road monitoring data; Step S2: Correct and adjust the micro-topography road monitoring data for anomalies at the micro-topography road monitoring nodes to generate corrected micro-topography road surface monitoring data; Step S3: Analyze the meteorological impact characteristics of road surface icing based on the corrected micro-topography road surface monitoring data, and generate road surface icing meteorological impact characteristic data; Step S4: Obtain road weather forecast data; Based on the road surface topography icing meteorological impact characteristic data, perform dynamic downscaling micro-topography road surface climate trend simulation processing on the corrected micro-topography road surface monitoring data and road weather forecast data to generate micro-topography road surface climate trend simulation data; Step S5: Establish a micro-topographic road surface icing thickness prediction model based on road surface topographic icing meteorological impact characteristic data and corrected micro-topographic road surface monitoring data; Step S6: Use the micro-topography road surface icing thickness prediction model to perform early sensing and prediction processing on the micro-topography road surface climate trend simulation data, generate early sensing and prediction data of micro-topography road icing, and execute early warning operation of micro-topography road icing based on the early sensing and prediction data of micro-topography road icing.

[0005] This specification provides a micro-topography road icing early detection, prediction, and warning system for performing the micro-topography road icing early detection, prediction, and warning method as described above. The micro-topography road icing early detection, prediction, and warning system includes: The micro-topography road monitoring module is used to acquire data of the target micro-topography road area; it performs micro-topography road monitoring processing on the target micro-topography road area data through multi-source heterogeneous monitoring equipment to generate micro-topography road monitoring data. The micro-topography road surface monitoring anomaly correction module is used to correct and adjust the micro-topography road monitoring node anomalies in the micro-topography road monitoring data and generate corrected micro-topography road surface monitoring data. The topographic icing meteorological impact characteristic analysis module is used to analyze the meteorological impact characteristics of road surface icing based on the corrected micro-topographic road surface monitoring data, and generate road surface topographic icing meteorological impact characteristic data. The climate trend simulation module is used to acquire road weather forecast data; based on the corrected micro-topography pavement monitoring data and road weather forecast data, it performs dynamic downscaling micro-topography pavement climate trend simulation processing to generate micro-topography pavement climate trend simulation data. The icing thickness prediction model building module is used to establish a micro-topographic road surface icing thickness prediction model based on road surface topography icing meteorological impact characteristic data and corrected micro-topographic road surface monitoring data. The micro-topography road icing early warning module is used to perform early sensing and prediction processing of micro-topography road climate trend simulation data using a micro-topography road icing thickness prediction model, generate early sensing and prediction data of micro-topography road icing, and execute micro-topography road icing early warning operations based on the early sensing and prediction data of micro-topography road icing.

[0006] The beneficial effects of this application are as follows: By conducting GIS data collection, grid division, and extreme value node analysis on the target micro-topographic road, this invention solves the inherent pattern of uniform deployment of traditional monitoring equipment. It relies on micro-topographic GIS grid attribute data and extreme value node data to complete the analysis of road micro-topographic characteristics, and accordingly conducts targeted design of multi-source heterogeneous monitoring equipment configuration nodes. This achieves encrypted deployment of equipment in micro-topographic extreme value nodes and areas prone to icing, effectively avoiding the problem of missed monitoring in extreme value areas, and ensuring a high degree of adaptation between the monitoring equipment deployment and micro-topographic features. Through monitoring using the adapted multi-source heterogeneous monitoring equipment, the generated micro-topographic road monitoring data covers multiple dimensions of information, including on-site weather, road surface condition, and icing coverage. The data collection is more targeted and comprehensive, accurately reflecting the actual monitoring characteristics of different areas of the micro-topographic road. By leveraging the attribute data of micro-topographic roads GIS grids, the monitoring data is classified and analyzed for attributes. Clustering and intra-cluster difference analysis of the monitoring data are conducted in conjunction with micro-topographic features, ensuring that anomaly analysis always aligns with the regional characteristics of micro-topographic roads. This avoids the problem of poor adaptability between traditional general anomaly correction methods and micro-topographic scenarios. By determining an adaptive evaluation threshold for micro-topography adaptation, a matching evaluation mapping relationship between micro-topography and monitoring status is constructed using the isolated forest algorithm, forming a customized anomaly analysis model for micro-topography. This makes the judgment criteria for anomaly identification of monitoring nodes more closely aligned with the actual monitoring situation of micro-topographic roads, significantly improving the accuracy of anomaly node identification. Based on the identified anomaly data, the original monitoring data is then specifically corrected, effectively eliminating or correcting anomaly monitoring data caused by extreme weather, equipment vibration, and other factors. The generated corrected micro-topographic road surface monitoring data eliminates errors caused by anomaly data, significantly improving data quality and accuracy. This effectively avoids interference from anomaly data in subsequent stages such as icing meteorological impact characteristic analysis and climate trend simulation, ensuring the accuracy and reliability of subsequent analysis and calculation results. This study precisely extracts two core data categories: road surface topography features and meteorological factor features. Then, through hierarchical analysis, it uncovers the influence patterns of micro-topography on meteorological factors and the correlation between meteorological factors and road icing. This allows for the accurate mining of the coupling effect between micro-topography and meteorological factors, effectively identifying meteorological factors that play a key role in micro-topographic road icing. A secondary analysis of the core icing meteorological factors is conducted using road surface topography meteorological factor influence feature data, integrating micro-topography features to generate road surface topography icing meteorological influence feature data. This data accurately reflects the differentiated influence patterns of each core meteorological factor on road icing under different micro-topographic features, clearly depicting the icing meteorological influence characteristics of micro-topographic roads, and overcoming the limitation of traditional analysis that ignores the coupling relationship between micro-topography and meteorological factors.By analyzing time-series features and meteorological models, the meteorological variation patterns of micro-topographic roads are explored, and a preliminary simulation of macro-climate trends is completed. This is followed by a series of refinements, including micro-topographic accuracy adaptation optimization, terrain downscaling and zonal microclimate simulation, and icing meteorological parameter adaptation optimization. Through iterative optimization of zonal regions using dynamic climate downscaling, the simulation data is cyclically corrected, forming a refined simulation process adapted to micro-topographic characteristics. This overcomes the limitations of existing technologies that directly use large-scale regional meteorological forecast data without targeted downscaling or only perform static downscaling. Through multiple rounds of optimization and zonal iteration, the problem of poor compatibility between large-scale regional meteorological data and micro-topographic road microclimates is effectively solved. This ensures that the simulation process fully reflects the local meteorological characteristics and icing meteorological impact patterns of micro-topographic roads. The generated micro-topographic road surface climate trend simulation data accurately reflects the meteorological development trends of different micro-topographic zones, and the data accuracy and micro-topographic adaptability are significantly improved. This ensures the early and accurate prediction of icing from the meteorological simulation perspective. Based on the meteorological impact data of icing, a micro-topographic meteorological icing regression model reflecting the correlation between meteorological factors and icing is established. Then, through quantitative correlation analysis of road surface moisture and icing, a moisture-icing relationship model that conforms to the physical laws of icing on micro-topographic roads is established. By integrating the two types of models, the prediction relationship of icing thickness is completed. This solves the technical limitations of traditional icing prediction, which only makes qualitative judgments or directly applies flat road models to micro-topographic roads, resulting in poor model adaptability. It uses a statistical model to capture the linear correlation between meteorological factors and micro-topographic icing, and a physical mechanism model to explore the essential laws of icing on micro-topographic roads. This allows the prediction model to fully conform to the icing characteristics of micro-topographic roads and eliminates the dependence on general models. By combining a micro-topographic road surface icing thickness prediction model with micro-topographic road surface climate trend simulation data, early perception and prediction of micro-topographic road icing can be carried out, and early warning operations can be performed. This allows the icing prediction process to fully align with the meteorological development trends and icing patterns of each micro-topographic zone, enabling early prediction of icing thickness, development rate, and distribution characteristics in different micro-topographic areas. This solves the problems of insufficient advance warning, poor targeting, lack of quantitative data support, and mostly simplified general warnings for micro-topographic road icing. It achieves zoned and precise early warning of micro-topographic road icing, clearly grasps the icing risk level and development trend of different micro-topographic areas, significantly improves the early warning, targeting, and effectiveness of micro-topographic road icing, effectively makes up for the shortcomings of existing early warning operations, and enables preventive measures to be taken in advance, effectively reducing the traffic safety risks caused by micro-topographic road icing.

[0007] Therefore, the micro-topography road icing early perception, prediction, and early warning method of this invention incorporates targeted node design for monitoring equipment based on the characteristics and differences of micro-topography, avoiding the problem of missed measurements in extreme areas and significantly improving the representativeness and accuracy of monitoring data. Simultaneously, it constructs an anomaly correction method for monitoring nodes based on micro-topography characteristics, significantly improving the accuracy of anomaly data correction and providing high-quality data source support for subsequent icing prediction-related analysis. Secondly, it performs dynamic downscaling processing of meteorological data for micro-topography roads, effectively improving the adaptability of regional meteorological forecast data to the microclimate of micro-topography roads, and accurately simulating the climate development trend of micro-topography road surfaces. Furthermore, it establishes a quantitative prediction model for icing thickness based on the meteorological impact characteristics of micro-topography icing, breaking through the limitations of traditional qualitative judgment, realizing early perception and accurate prediction of icing thickness on micro-topography roads, significantly improving the advance and targeted nature of micro-topography road icing early warning, effectively avoiding traffic safety hazards caused by micro-topography road icing, and effectively ensuring traffic safety and smooth operation of micro-topography roads in winter. Attached Figure Description

[0008] Figure 1 This is a schematic diagram of the steps of an early sensing, prediction and warning method for road icing in micro-topography according to the present invention; Figure 2 for Figure 1 A detailed flowchart illustrating the implementation steps of step S3. The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0009] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0010] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. Functional entities may be implemented in software, or in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods. The term "and / or" as used herein includes any and all combinations of one or more of the associated items listed.

[0011] To achieve the above objectives, please refer to Figures 1 to 2This invention provides a method and system for early sensing, prediction, and warning of road icing in micro-topography. In the embodiments of this invention, please refer to... Figure 1 The diagram shown is a flowchart illustrating the steps of an early detection, prediction, and warning method for micro-topography road icing according to the present invention. The method includes the following steps: Step S1: Obtain target micro-topography road area data; perform micro-topography road monitoring processing on the target micro-topography road area data using multi-source heterogeneous monitoring equipment to generate micro-topography road monitoring data; In this embodiment of the invention, when acquiring data on the target micro-topographic road area, an airborne lidar and a ground-based mobile measurement system are used in a collaborative manner to comprehensively capture basic information about the entire road surface and the micro-topographic influence areas on both sides of the road. The acquired data covers topographic elevation, surface cover type, road edge, pavement condition, road ancillary facilities, and surrounding surface elements, forming a complete basic dataset for the target area. This ensures that the data can fully reflect the micro-topographic and road engineering characteristics of the target road. After the basic area data is acquired, a multi-source heterogeneous monitoring device is activated to conduct micro-topographic road monitoring and processing. The monitoring device travels at a constant speed along the entire target micro-topographic road according to a preset route. The device is equipped with core modules such as high-definition vision, laser ranging, inertial navigation, meteorological monitoring, and road surface condition monitoring. This enables the simultaneous acquisition of road surface condition, roadbed condition, surrounding micro-topography, near-surface meteorological conditions, and initial icing signs. During the monitoring process, the spatial range and temporal frequency of the acquisition are strictly controlled to ensure that the acquisition range covers the entire road surface and surrounding micro-topographic influence areas of the target road. The acquisition frequency is matched with the device's operating speed to ensure the continuity and integrity of the data acquisition. After the monitoring data collection is completed, the raw data is initially organized and classified according to data type. The collection location, collection time and monitoring node corresponding to the data are marked. Obviously invalid data is removed, and micro-topography road monitoring data is generated, which includes micro-topography feature data, road surface condition data, near-surface meteorological data and initial icing related data.

[0012] Step S2: Correct and adjust the micro-topography road monitoring data for anomalies at the micro-topography road monitoring nodes to generate corrected micro-topography road surface monitoring data; In this embodiment of the invention, the micro-topographic road monitoring data is classified into three main categories based on its purpose and characteristics: meteorological monitoring attributes, road surface condition monitoring attributes, and icing cover monitoring attributes. Each category has clearly defined corresponding monitoring indicators and data characteristics, providing a classification basis for anomaly identification. During the anomaly identification process, combining the micro-topographic features of the micro-topographic road with the temporal variation patterns of the monitoring data, intra-cluster difference analysis and neighborhood comparison analysis are used to identify abnormal nodes in the monitoring data and clarify the anomaly types. These mainly include three categories: anomalies caused by equipment failure, anomalies caused by signal interference, and anomalies caused by sudden changes in the local environment. Among them, equipment failure anomalies are characterized by data remaining unchanged or stabilizing after a sudden change; signal interference anomalies are characterized by data fluctuating drastically and irregularly; and environmental change anomalies are characterized by sudden changes in data that conform to the characteristics of the local micro-topographic environment. Differentiated correction and adjustment methods are adopted for different types of abnormal nodes. For anomalies caused by equipment failure, data interpolation from neighboring normal nodes is used for correction, and different weights are assigned to neighboring nodes based on micro-topographic features to ensure that the correction values ​​conform to the actual micro-topographic conditions of the area where the abnormal node is located. For anomalies caused by signal interference, time-series smoothing is used to eliminate data fluctuation deviations and restore the true trend of data changes. For anomalies caused by sudden changes in the local environment, verification is performed in conjunction with micro-topographic grid attributes. After confirmation, the original data is retained and marked with environmental change indicators to avoid miscorrection. After correction, all correction data undergoes double verification. On the one hand, the numerical range of the data is verified to ensure that it conforms to the actual monitoring patterns of micro-topographic roads. On the other hand, the spatial continuity and temporal consistency of the data are verified to ensure that the data fluctuations of adjacent monitoring nodes conform to the micro-topographic change characteristics, and the changes in time-series data conform to the natural changes in meteorological and road surface conditions. After passing the verification, the correction data is integrated and archived according to attribute type and micro-topographic grid to generate corrected micro-topographic road surface monitoring data.

[0013] Step S3: Analyze the meteorological impact characteristics of road surface icing based on the corrected micro-topography road surface monitoring data, and generate road surface icing meteorological impact characteristic data; In this embodiment of the invention, two types of core feature indicators are extracted from the corrected micro-topographic road surface monitoring data. One type is road surface topographic feature indicators, covering features related to micro-topography such as topographic undulation, road surface slope, vegetation cover, roadbed material, and surface runoff. The other type is road surface meteorological factor feature indicators, covering meteorological elements directly related to road icing such as air temperature, road surface temperature, relative humidity, wind speed, and precipitation. Both types of indicators are quantified to form a feature dataset that can be used for analysis. Subsequently, the road surface topographic feature indicators are grouped. Based on the differences in different topographic features, the target micro-topographic road is divided into different topographic groups. Each group contains several micro-topographic grids to ensure that each group of samples is representative and can reflect the core features of that type of terrain. After grouping, meteorological factor feature analysis is carried out for each type of terrain group. The distribution patterns, trends, and numerical characteristics of various meteorological factors within different terrain groups are compared. The influence degree and direction of topographic features on meteorological factors are quantitatively analyzed to clarify the differentiated performance of meteorological factors under different micro-topographic conditions. Based on this, further correlation analysis of meteorological factors and road icing will be conducted, focusing on meteorological factors that play a key role in icing, exploring the correlation patterns between meteorological factors and icing indicators such as icing thickness and icing coverage area under different micro-topographic conditions, clarifying the indirect impact of topographic features on road icing by influencing meteorological factors, and integrating the meteorological factor characteristics of various topographic groups, the coupling relationship between topography and meteorological factors, and the correlation patterns between meteorological factors and icing to form structured meteorological impact characteristic data of road surface topography icing.

[0014] Step S4: Obtain road weather forecast data; Based on the road surface topography icing meteorological impact characteristic data, perform dynamic downscaling micro-topography road surface climate trend simulation processing on the corrected micro-topography road surface monitoring data and road weather forecast data to generate micro-topography road surface climate trend simulation data; In this embodiment of the invention, authoritative road weather forecast data for the target micro-topographic road area is obtained. The forecast data covers core meteorological indicators for a future period, and the forecast range fully covers the entire road segment of the target micro-topographic road and the surrounding micro-topographic influence area. The forecast data is standardized according to a unified spatial and temporal benchmark to ensure consistency with the benchmark of the calibrated micro-topographic road surface monitoring data, facilitating subsequent data fusion and simulation processing. Subsequently, temporal feature analysis is performed on the calibrated micro-topographic road surface monitoring data to explore the temporal variation patterns, peak and trough distributions, and low-temperature period characteristics of various meteorological indicators, clarifying the temporal differences in meteorological data in the micro-topographic region and providing a temporal basis for meteorological model adaptation. Based on the temporal feature data and road weather forecast data, micro-topographic road surface meteorological model analysis is conducted to establish the adaptation relationship between the meteorological characteristics of the micro-topographic region and the large-scale meteorological forecast data, explore the meteorological change patterns of different micro-topographic groups, quantify the deviation patterns between micro-topographic meteorology and large-scale forecasts, and ensure that subsequent simulation data can both conform to large-scale meteorological trends and accurately reflect the local microclimate differences of the micro-topography. Based on this, a macro-climate trend simulation of micro-topography road surfaces was conducted. Large-scale meteorological forecast data was extended to each grid of the target micro-topography road using interpolation methods. Preliminary corrections were then performed using the established meteorological model to generate preliminary macro-climate trend simulation data. Subsequently, through multiple rounds of accuracy adaptation optimization and dynamic downscaling, the large-scale simulation data was gradually refined to micro-topographic groups. Combined with road surface icing meteorological impact characteristic data, the core icing meteorological parameters in the simulation data were specifically optimized to eliminate simulation biases caused by spatial scale differences. Finally, through iterative optimization of different zones, accurate fusion of large-scale and small-scale simulation data was achieved, ensuring the accuracy and adaptability of the simulation data. After multiple rounds of optimization and verification, micro-topography road surface climate trend simulation data was generated.

[0015] Step S5: Establish a micro-topographic road surface icing thickness prediction model based on road surface topographic icing meteorological impact characteristic data and corrected micro-topographic road surface monitoring data; In this embodiment of the invention, a micro-topographic meteorological icing regression sub-model is constructed based on road surface topographic icing meteorological impact characteristic data. This sub-model uses micro-topographic features and core icing meteorological factors as influencing factors and icing thickness as the prediction target. By quantifying the linear correlation between various factors and icing thickness, it constructs regression relationships adapted to different micro-topographic groups, highlighting the impact of the coupling between micro-topographic features and meteorological factors on icing thickness. This sub-model can achieve preliminary prediction of icing thickness based on meteorological and topographic factors. Simultaneously, based on corrected micro-topographic road surface monitoring data, a micro-topographic road surface moisture-icing relationship sub-model is constructed. This sub-model focuses on road surface moisture as the core material basis of icing, quantifies the intrinsic correlation between road surface moisture content and icing thickness, constructs a piecewise mapping relationship, and highlights the impact of moisture factors on icing thickness, compensating for the shortcoming of the regression sub-model not fully considering the influence of moisture. This sub-model can achieve accurate quantitative prediction of icing thickness based on road surface moisture content. After constructing the two sub-models, a weighted fusion method was used to build a comprehensive prediction model for road icing thickness in micro-topography areas, combining their prediction accuracy. Differential weights were assigned to the two sub-models based on the icing patterns of different micro-topographic regions. In areas with high icing susceptibility, the moisture-icing relationship sub-model had a higher weight, emphasizing the impact of moisture on icing; in areas with low icing susceptibility, the meteorological icing regression sub-model had a higher weight, emphasizing the coupling effect of meteorology and topography. Simultaneously, a micro-topography correction coefficient was introduced, and the comprehensive prediction results were fine-tuned based on the core features of different micro-topographic groups to ensure the model accurately adapts to various micro-topographic conditions, ultimately generating a comprehensive prediction model for road icing thickness in micro-topography areas.

[0016] Step S6: Use the micro-topography road surface icing thickness prediction model to perform early sensing and prediction processing on the micro-topography road surface climate trend simulation data, generate early sensing and prediction data of micro-topography road icing, and execute early warning operation of micro-topography road icing based on the early sensing and prediction data of micro-topography road icing.

[0017] In this embodiment of the invention, micro-topographic road surface climate trend simulation data is input into a micro-topographic road surface icing thickness prediction model. Based on the input meteorological trend parameters and the micro-topographic features of the corresponding grid, the prediction model automatically calls up prediction parameters and weight allocations adapted to the micro-topographic group, calculates the predicted icing thickness value through two sub-models, and then outputs the final icing thickness prediction result through a comprehensive prediction formula. Simultaneously, it outputs the prediction accuracy and the corresponding icing risk level. The risk level is divided according to the predicted icing thickness, clearly defining the degree of icing hazard corresponding to different risk levels. After the prediction of all grids is completed, the prediction results are integrated and archived according to micro-topographic grouping and time dimension to generate early perception and prediction data for micro-topographic road icing. This data clearly marks each grid of the target micro-topographic road, the icing start time, icing duration, peak icing thickness, and icing risk level within a future period. It can accurately reflect the spatial distribution and temporal variation of future micro-topographic road icing, achieving early perception and accurate prediction of road icing. Once the forecast data is generated, micro-topography road icing early warning operations are immediately carried out based on the data. Different early warning plans are formulated according to different icing risk levels. Emergency warnings are issued for high-risk areas, general warnings are issued for medium-risk areas, and alert warnings are issued for low-risk areas. The early warning information clearly marks the warning area, warning period, icing risk level, and prevention suggestions.

[0018] Furthermore, the micro-topography road monitoring data mentioned in step S1 includes on-site meteorological data of micro-topography roads, road surface condition data of micro-topography roads, and road surface icing coverage data of micro-topography roads.

[0019] Furthermore, step S1 includes the following steps: Step S11: Obtain target micro-topography road area data; In this embodiment of the invention, the operation of acquiring target micro-topographic road area data is performed. The acquired data is the full-dimensional basic data of the target micro-topographic road between location A and location B. The data sources are the high-precision topographic mapping results at a scale of 1:5000 from the geographic surveying and mapping department, the historical winter meteorological observation data of the meteorological department's ground observation stations for the past 10 years, the engineering completion data of the road construction department, and the winter icing inspection records of the road maintenance department. The acquired data types include three main categories: first, topographic geographic data, specifically covering spatial characteristic data such as topographic elevation, surface cover type, road direction, road surface slope, and surface catchment area of ​​the entire target road and within 50m on both sides of the road; second, road engineering data, specifically covering engineering characteristic data such as roadbed filling materials, pavement type, number of lanes, roadbed depth, and road drainage system layout; and third, meteorological icing related data, specifically covering data such as the daily minimum temperature distribution, duration of low temperature, precipitation type ratio, snowfall distribution, road icing start time, and icing area distribution of the target area in the past 10 winters. All acquired data are standardized in terms of spatial reference and format. The geographic coordinates of topographic and road engineering data are uniformly converted into the geodetic coordinate system. Meteorological icing-related data are correlated and matched according to the time and spatial dimensions. The corresponding spatial location and time interval are marked for each data item to ensure that the spatial location of various types of data is accurately matched and the time dimension is traceable, thereby eliminating the reference deviation and format difference of data from different sources.

[0020] Step S12: Collect micro-topographic road GIS data from the target micro-topographic road area data to obtain micro-topographic road GIS data; In this embodiment of the invention, based on the regularized target micro-topographic road area data, the micro-topographic road GIS data collection operation is carried out. The collection is implemented by a joint collection method of airborne lidar and ground mobile measurement system. The airborne lidar is set with parameters of 200m flight altitude and 30% flight strip overlap to carry out aerial collection. The point cloud collection density is set to 15 point cloud data per square meter to achieve high-precision point cloud collection of the terrain elevation and surface cover of the entire target road. The ground mobile measurement system carries out ground collection along the entire target road at a speed of 5km / h. The collection equipment is equipped with a high-definition visual sensor, a laser rangefinder sensor, and an inertial navigation module to achieve accurate collection of road edge lines, road paving, road ancillary facilities, and surface elements within 50m around the road. The point cloud data collected by airborne lidar undergoes multiple rounds of denoising processing. Invalid point clouds caused by atmospheric scattering and obstacle occlusion are removed by the neighborhood point cloud elevation comparison method. Then, the iterative nearest point algorithm is used to complete the point cloud registration and stitching to generate a continuous three-dimensional point cloud model of the entire target road. Various data collected by the ground mobile measurement system are vectorized. The area features collected by the high-definition visual sensor and the line features collected by the laser rangefinder sensor are converted into vector data that can be recognized by the GIS system. At the same time, the collected geographic information features are processed in layers according to the terrain layer, road layer, and surface cover layer. The spatial resolution of each feature layer is set to 1m. The resulting micro-topographic road GIS data includes two main types: three-dimensional point cloud model and two-dimensional vector data. It accurately restores the spatial characteristics of the terrain, road and surrounding surface of the target micro-topographic road, and realizes comprehensive and high-precision acquisition of the micro-topographic spatial information of the target road.

[0021] Step S13: Perform GIS grid division processing on the micro-topography road GIS data to generate micro-topography road GIS grid data, and perform grid attribute labeling processing on the micro-topography road GIS grid data to generate micro-topography road GIS grid attribute data. In this embodiment of the invention, GIS grid division processing is performed on the micro-topographic road GIS data. According to the square grid division rule, a grid size of 50m×50m is set along the direction of the target micro-topographic road, and the road is divided at equal intervals in the longitudinal and transverse directions. The grid coverage range is the entire road surface of the target road and the micro-topographic influence area within 50m on both sides of the road. Each divided grid is assigned a unique grid code, and the precise spatial location of the grid can be directly traced through the code. At the same time, the spatial coordinates, grid area, and surrounding neighboring grid information of each grid are associated and labeled. The generated micro-topographic road GIS grid data realizes the grid-based spatial segmentation of the target micro-topographic road, transforming the originally continuous micro-topographic area into multiple independent and accurately identifiable spatial units. After completing the grid division, the micro-topographic road GIS grid data is processed for grid attribute labeling. Based on the topographic and road engineering features of the micro-topographic road GIS data, six core micro-topographic feature attributes are extracted from each grid and labeled: topographic relief, vegetation coverage, subgrade material, pavement slope, surface catchment area, and road orientation. Among them, topographic relief is quantitatively labeled according to the difference between the maximum and minimum elevations within the grid; vegetation coverage is quantitatively labeled as a percentage of the ratio of the vegetation projection area within the grid to the total area of ​​the grid; subgrade material is labeled as a fixed category of asphalt concrete, cement concrete, and soil-rock mixture; pavement slope is quantitatively labeled as a percentage value of actual measurement; surface catchment area is quantitatively labeled according to the actual range of natural catchment within the grid; and road orientation is quantitatively labeled according to the azimuth angle based on due north. All attribute labels are presented as quantitative values ​​or fixed categories without ambiguity. The final generated micro-topographic road GIS grid attribute data achieves a precise correspondence between each grid spatial unit and the core micro-topographic feature attributes.

[0022] Step S14: Perform extreme value node analysis on the micro-topography road GIS grid data to generate extreme value node data of the road GIS grid; In this embodiment of the invention, extreme value node analysis of road GIS grids is carried out based on micro-topographic road GIS grid data. The elevation coordinates of the four vertices of each grid in the micro-topographic road GIS grid data are extracted. The elevation change rate of each grid is calculated in both east-west and north-south spatial directions. The maximum value of the elevation change rate in the two directions is then taken as the core elevation gradient value of the grid. The core elevation gradient values ​​of all grids are calculated and labeled to generate grid elevation gradient data. The calculation results of the core elevation gradient value directly reflect the degree of drastic change in terrain elevation of each grid. Subsequently, an elevation gradient threshold of 0.3 is set. Grids with core elevation gradient values ​​exceeding this threshold are identified as terrain elevation gradient abrupt change grids. These grids are high-risk areas for micro-topographic road icing. The geometric center of these grids is directly extracted as the initial extreme value node. Then, spatial redundancy screening is performed on all initial extreme value nodes. The straight-line distance between adjacent initial extreme value nodes is calculated, and redundant extreme value nodes with a spacing of less than 20m are removed to avoid waste of monitoring resources caused by excessively dense extreme value nodes. The selected valid extreme value nodes are sorted in an orderly manner according to the route of the target road. Each valid extreme value node is assigned a unique node code, and core information is marked for each extreme value node, including spatial coordinates, grid code, core elevation gradient value, and distribution of surrounding neighboring grids. The generated road GIS grid extreme value node data enables the accurate location of key nodes of abrupt changes in terrain elevation of the target micro-topography road. These extreme value nodes are prone to forming local microclimates due to terrain features and are key monitoring areas for road icing, providing accurate node location basis for the targeted deployment of subsequent monitoring equipment.

[0023] Step S15: Analyze the micro-topography characteristics of roads using micro-topography road GIS grid attribute data and extreme value node data of road GIS grid, and generate road micro-topography characteristic data; In this embodiment of the invention, based on micro-topographic road GIS grid attribute data and road GIS grid extreme value node data, a comprehensive analysis of road micro-topographic characteristics is conducted from the perspective of the impact of micro-topographic features on road icing. The analysis dimensions are set as five core dimensions: the impact of topographic undulation on cold air accumulation, the impact of vegetation cover on near-surface temperature and humidity, the impact of road surface slope on surface water retention, the impact of surface catchment area on road surface water accumulation, and the impact of abrupt changes in elevation gradient on local microclimate. A quantitative evaluation standard is set for each analysis dimension. For each grid in the micro-topographic road GIS grid attribute data, a quantitative evaluation is conducted according to the five analysis dimensions. Combining the micro-topographic feature attribute values ​​of each dimension, a quantitative score for icing susceptibility is assigned to each grid. The score adopts a standard of 1-10 points, and the score is positively correlated with icing susceptibility. Grids with greater topographic undulation, higher vegetation cover, gentler road surface slope, and larger catchment area have higher icing susceptibility scores. For each extreme node in the road GIS grid extreme node data, an icing susceptibility layer analysis is conducted within a 50m radius around the node, combining its core elevation gradient value and the icing susceptibility score of its respective grid. This is divided into core and secondary influence layers based on distance from the extreme node, with each layer labeled with its icing susceptibility level (high, medium, low). After analyzing all grids and extreme nodes, the results are integrated, labeling the core icing-influencing micro-topographic factors for each grid and extreme node. This identifies the dominant factors influencing icing, such as topographic relief and vegetation cover. The resulting road micro-topographic characteristic data includes the icing susceptibility score and core influencing factors for each grid, as well as the icing susceptibility level and influence layer range for each extreme node. This allows for precise quantification and characteristic analysis of the icing susceptibility in different areas of the target micro-topographic road, ensuring that the design of subsequent monitoring equipment configuration nodes perfectly matches the icing susceptibility characteristics of the micro-topographic road and guaranteeing the targeted and rational deployment of monitoring equipment.

[0024] Step S16: Design the configuration nodes of the multi-source heterogeneous road monitoring equipment based on the road micro-topography characteristic data, and obtain the configuration node data of the multi-source heterogeneous road monitoring equipment; In this embodiment of the invention, the configuration node design of multi-source heterogeneous road monitoring equipment is carried out based on road micro-topographic characteristic data. The deployment principle strictly follows the principle of "densified deployment in areas with high icing susceptibility, uniform deployment in ordinary areas, and separate deployment of extreme value nodes". The deployment level is divided according to the icing susceptibility score in the road micro-topographic characteristic data. For high icing susceptibility grids with an icing susceptibility score of 8-10, the monitoring node deployment spacing is set to 20m; for medium icing susceptibility grids with an icing susceptibility score of 4-7, the monitoring node deployment spacing is set to 50m; and for low icing susceptibility grids with an icing susceptibility score of 1-3, the monitoring node deployment spacing is set to 100m. All extreme value nodes of the road GIS grid are deployed with separate monitoring nodes, and the monitoring nodes of extreme value nodes and the monitoring nodes of the surrounding grids form complementary coverage, with no monitoring blind spots. Based on the monitoring needs of different icing-prone areas, the configuration types of multi-source heterogeneous monitoring equipment are determined. The equipment types include six categories: road surface temperature monitor, roadbed temperature monitor, air temperature and humidity monitor, wind speed and direction monitor, road surface condition monitor, and icing thickness monitor. Among them, the configuration nodes of high-prone grids and extreme value nodes are equipped with a full set of all six types of equipment, the configuration nodes of medium-prone grids are equipped with three core types of equipment: road surface temperature monitor, air temperature and humidity monitor, and road surface condition monitor, and the configuration nodes of low-prone grids are equipped with only two basic types of equipment: road surface temperature monitor and road surface condition monitor. After the design is completed, each configuration node is assigned a unique node code, and five core pieces of information are marked: spatial coordinates, grid code, icing susceptibility level, type of equipment deployed, and number of equipment. At the same time, a distribution map of the monitoring equipment configuration nodes is drawn to clarify the spatial location and equipment configuration of each configuration node. The resulting multi-source heterogeneous road monitoring equipment configuration node data achieves a high degree of adaptation between the monitoring equipment type, deployment density and the icing susceptibility characteristics of micro-topography roads. This ensures that the deployment of monitoring equipment can accurately cover high-incidence icing areas and key extreme value nodes, and guarantees the accuracy and comprehensiveness of subsequent monitoring data collection.

[0025] Step S17: Configure the multi-source heterogeneous road monitoring equipment with node data. After configuration, the multi-source heterogeneous monitoring equipment performs micro-topographic road monitoring processing on the target micro-topographic road area data to generate micro-topographic road monitoring data.

[0026] In this embodiment of the invention, according to the requirements of the configuration node data of the multi-source heterogeneous road monitoring equipment, the on-site deployment and installation of the multi-source heterogeneous monitoring equipment are completed at each configuration node. The equipment installation follows a fixed installation standard. The sensor probe of the road surface temperature monitor is completely attached to the road surface, and the probe is encapsulated with a waterproof and wear-resistant protective sleeve. The sensor of the roadbed temperature monitor is buried in the roadbed at a vertical burial depth of 10cm, and the gap between the sensor and the roadbed is filled with sealing material. The sensors of the air temperature and humidity monitor and the wind speed and direction monitor are installed at a height of 2m above the road, and the sensors avoid the obstruction of road ancillary facilities. The monitoring probes of the road surface condition monitor and the ice thickness monitor face the road surface monitoring area, and the vertical distance between the probe and the road surface is set to 0.5m. All monitoring equipment is installed using low-temperature resistant and vibration-resistant fixed brackets to ensure stable operation of the equipment in the complex road environment in winter. Each monitoring device acquires monitoring data in real time via wired transmission. The collected monitoring data is categorized and integrated according to device type and node code. The final micro-topography road monitoring data includes three main categories: first, on-site meteorological data of micro-topography roads, specifically covering four quantitative indicators: air temperature, air humidity, wind speed, and wind direction; second, road surface condition data of micro-topography roads, specifically covering five fixed condition indicators: dry, wet, waterlogged, snow-covered, and icy; and third, road surface icing coverage data of micro-topography roads, specifically covering two quantitative indicators: icing thickness and icing coverage area. All data are presented as quantitative values ​​or fixed categories, without ambiguity, accurately reflecting the real-time meteorological, road surface, and icing conditions of each monitoring node of the target micro-topography road. This provides comprehensive, accurate, and standardized basic monitoring data for subsequent anomaly correction of micro-topography road monitoring data.

[0027] Furthermore, step S14 includes the following steps: Step S141: Perform grid elevation gradient analysis based on micro-topography road GIS grid data to generate grid elevation gradient data; In this embodiment of the invention, grid elevation gradient analysis is performed based on micro-topographic road GIS grid data. This analysis uses a 50m × 50m GIS grid as the basic analysis unit. Precise elevation coordinates of the four vertices and geometric center of each grid are extracted from the micro-topographic road GIS grid data. Then, single-dimensional elevation gradients are calculated along both the longitudinal and transverse spatial dimensions of the road. The longitudinal elevation gradient is the ratio of the elevation difference between the two vertices in the north-south direction of the grid to the longitudinal side length of the grid; the transverse elevation gradient is the ratio of the elevation difference between the two vertices in the east-west direction of the grid to the transverse side length of the grid. The results of both dimensions are presented as percentage values, directly reflecting the degree of drastic change in terrain elevation in the corresponding direction. After completing the single-dimensional elevation gradient calculation, the root mean square algorithm is used to calculate the comprehensive elevation gradient value of each grid. The comprehensive elevation gradient value is the square root of the sum of the squares of the longitudinal and transverse elevation gradient values. This value is the core indicator for determining abrupt changes in grid terrain undulation. Simultaneously, the dominant direction of the elevation gradient for each grid is marked, i.e., the direction with the larger longitudinal or transverse elevation gradient value, clarifying the main spatial direction of the grid terrain abrupt change. After calculating the vertical, horizontal, and comprehensive elevation gradient values ​​and labeling the dominant direction for all GIS grids in sequence, a unique grid code is matched for each grid. The grid code, the elevation coordinates of the four vertices and the center, the vertical elevation gradient value, the horizontal elevation gradient value, the comprehensive elevation gradient value, and the dominant direction of the elevation gradient are associated and integrated to generate grid elevation gradient data, which is structured terrain feature data.

[0028] Step S142: Extract the grid center elevation gradient extreme value nodes from the grid elevation gradient data to obtain the grid center elevation gradient extreme value node data; In this embodiment of the invention, grid center elevation gradient extreme node extraction is performed on grid elevation gradient data to obtain grid center elevation gradient extreme node data. The extraction operation first sets the comprehensive elevation gradient value of 0.3 as the extreme value judgment threshold. GIS grids with a comprehensive elevation gradient value exceeding 0.3 in the grid elevation gradient data are judged as terrain elevation gradient abrupt change grids. Such grids are highly prone to micro-topographic road icing due to severe terrain undulations. The geometric center of such grids is directly extracted as the initial elevation gradient extreme node to ensure that the node position accurately corresponds to the core area of ​​terrain abrupt change. For all extracted initial extreme value nodes, each node is first labeled with its grid code, comprehensive elevation gradient value, longitudinal and lateral elevation gradient values, and spatial geographic coordinates. The coordinate data maintains the same spatial benchmark as the elevation coordinates in the grid elevation gradient analysis. Then, spatial redundancy screening of the initial extreme value nodes is carried out. The straight-line spatial distance between two adjacent initial extreme value nodes is calculated. If the straight-line distance between adjacent nodes is less than 20m, only the node with the larger comprehensive elevation gradient value is retained, and redundant nodes with smaller comprehensive elevation gradient values ​​are removed. This avoids excessive density of extreme value nodes, which would waste resources for subsequent monitoring equipment deployment, while ensuring that the retained nodes are the core locations with the most significant terrain abrupt changes. After completing the redundancy screening, the remaining valid initial extreme value nodes are initially classified. According to the comprehensive elevation gradient value, the nodes are divided into two intervals: 0.3-0.5 and above 0.5, corresponding to different degrees of terrain abrupt changes, respectively. This generates grid center elevation gradient extreme value node data.

[0029] Step S143: Perform road GIS grid extreme node integration processing on the grid center elevation gradient extreme node data to generate road GIS grid extreme node data.

[0030] In this embodiment of the invention, extreme value node data of the elevation gradient at the center of the grid are integrated and processed using road GIS grid extreme value nodes. Spatial coverage verification of the extreme value nodes is performed to check whether extreme value nodes have been extracted for all terrain abrupt change grids with a comprehensive elevation gradient value exceeding 0.3, ensuring that no extreme value nodes for terrain abrupt change grids are missed and guaranteeing comprehensive coverage of terrain abrupt change areas by extreme value nodes. After verification, core attribute information is added to each extreme value node, including the dominant direction of the node's elevation gradient, the neighboring grid code within a 50m radius, and the terrain abrupt change type. The terrain abrupt change type is divided into three categories based on the comprehensive elevation gradient value and terrain features: steep slope type, valley boundary type, and plateau edge type. These are fixed category labels based on actual terrain features. Then, a basic score for icing susceptibility is assigned to each node based on the comprehensive elevation gradient value: nodes in the 0.3-0.5 range are scored 8 points, and nodes in the range above 0.5 are scored 9 points. The score directly reflects the degree of icing risk in the node area. Finally, all extreme value nodes are divided into first-level extreme value nodes and second-level extreme value nodes according to their comprehensive elevation gradient values, generating road GIS grid extreme value node data.

[0031] Furthermore, step S2 includes the following steps: Step S21: Based on the micro-topography road GIS grid attribute data, classify the micro-topography road monitoring attribute types of the micro-topography road monitoring data, and generate micro-topography road monitoring attribute type classification data; In this embodiment of the invention, the micro-topographic road monitoring data is classified into micro-topographic road monitoring attribute types based on the micro-topographic road GIS grid attribute data. The classification is based on three core attributes in the micro-topographic road GIS grid attribute data: roadbed material, road surface slope, and vegetation coverage. Combined with the monitoring index characteristics of the micro-topographic road monitoring data, all monitoring data are divided into three major categories of core attribute types. The first category is meteorological monitoring attributes, covering four quantitative monitoring indicators: air temperature, air humidity, wind speed, and wind direction. This type of data mainly reflects the local meteorological conditions within the grid area and directly affects the formation and development of road icing; the second... The first category is the road surface condition monitoring attribute category, which covers five fixed state indicators: dry, wet, waterlogged, snow-covered, and icy. This type of data directly reflects the real-time condition of the road surface and is the core basis for judging whether the road is icy. The second category is the icing coverage monitoring attribute category, which covers two quantitative monitoring indicators: icing thickness and icing coverage area. This type of data directly quantifies the severity of road icing. After the classification is completed, the monitoring indicators, data formats, and numerical ranges of each attribute type are standardized. The generated micro-topography road monitoring attribute type classification data clarifies the attribute attribution, core indicators, and associated micro-topography grid information of each type of monitoring data.

[0032] Step S22: Perform attribute type feature analysis on the micro-topography road monitoring attribute type classification data to generate micro-topography road monitoring attribute type feature data; In this embodiment of the invention, for meteorological monitoring attribute data, the focus is on analyzing the temporal variation characteristics and spatial distribution characteristics of various indicators. Temporally, data are collected at a frequency of one minute, and the hourly and daily numerical variation ranges, averages, and extreme values ​​are statistically analyzed to clarify the fluctuation patterns of meteorological data during low-temperature periods (temperature ≤ 0℃). Spatially, the differences in meteorological data within different micro-topographic grids are analyzed by combining the topographic relief and vegetation coverage in the grid attributes. For road surface condition monitoring attribute data, the focus is on analyzing the conversion frequency and duration of various conditions, and the conversion patterns of road surface conditions under different micro-topographic conditions are analyzed by combining the roadbed material and road surface slope in the grid attributes. For icing coverage monitoring attribute data, the focus is on analyzing the growth rate of icing thickness and the changing trend of coverage area, and the differences in icing characteristics within different micro-topographic grids are analyzed by combining the topographic relief and surface catchment area in the grid attributes. After the analysis, the spatiotemporal variation patterns, spatial distribution differences, numerical ranges, and micro-topography adaptation characteristics of various attributes are integrated. Each attribute type is assigned a feature code, and the corresponding micro-topography influencing factors, numerical fluctuation thresholds, and state transition standards are labeled. The generated micro-topography road monitoring attribute type feature data accurately depicts the core characteristics of various monitoring data and their correlation with micro-topography.

[0033] Step S23: Perform clustering and segmentation processing on the feature data of each attribute type of micro-topography road monitoring to generate micro-topography road monitoring feature cluster data; In this embodiment of the invention, clustering is performed on the feature data of each attribute type based on the micro-topography road monitoring attribute type. The clustering process first integrates the feature data of the three major attribute types and extracts the core feature indicators corresponding to each monitoring node, including the fluctuation range of meteorological data, the frequency of road surface condition transition, the rate of increase in ice thickness, and the associated micro-topography grid attribute indicators (topographic relief, vegetation coverage, and roadbed material). These indicators are used as the core basis for clustering. A clustering method based on feature similarity is adopted to calculate the feature similarity between different monitoring nodes. The similarity calculation adopts the Euclidean distance algorithm, and the smaller the distance, the higher the similarity. The clustering was fixed at 6 clusters, strictly adhering to the coupling relationship between micro-topography type and monitoring characteristics. Cluster 1 corresponds to valley grids, high vegetation coverage, small meteorological data fluctuations, and roads prone to dampness and icing; Cluster 2 corresponds to steep slope grids, low vegetation coverage, large wind speed fluctuations, and roads prone to water runoff; Cluster 3 corresponds to flat slope grids, asphalt concrete subgrade, stable meteorological data, and smooth road surface condition transitions; Cluster 4 corresponds to plateau edge grids, moderate topographic relief, moderate meteorological data fluctuations, and moderate icing thickness; Cluster 5 corresponds to soil-rock mixed subgrade grids, moderate vegetation coverage, roads prone to water accumulation, and large icing coverage area; Cluster 6 corresponds to elevation gradient extreme value node grids, abrupt topographic changes, large meteorological data fluctuations, and high icing risk. After clustering, each cluster was validated by calculating the characteristic mean of all monitoring nodes within the cluster. Abnormal candidate nodes whose characteristic values ​​deviated from the cluster mean by more than 10% were removed to ensure the consistency of features within the cluster. At the same time, the characteristic difference between clusters was calculated to ensure that the difference between clusters was greater than 30% to avoid feature overlap between clusters. After verification, clustered data of micro-topographic road monitoring features are obtained.

[0034] Step S24: Perform intra-cluster differential feature analysis on the micro-topography road monitoring feature cluster data to generate intra-cluster differential feature data of the micro-topography road neighborhood; In this embodiment of the invention, the intra-cluster difference characteristics of micro-topographic road monitoring feature cluster data are analyzed to clarify the neighborhood relationships of each monitoring node within each cluster. Each monitoring node is considered as the core, and monitoring nodes within a 50m radius are designated as neighboring nodes, ensuring that all neighboring nodes belong to the same cluster (as cluster division ensures consistency in micro-topography). Real-time monitoring data and feature data of all monitoring nodes within each cluster are then extracted. The deviation between the data of each monitoring node within the cluster and the mean of the corresponding attribute within the cluster is calculated according to attribute type. The deviation is calculated using the formula (node ​​data - cluster mean) / cluster mean × 100%, quantitatively reflecting the degree of difference between the node data and the overall characteristics within the cluster. The larger the absolute value of the deviation, the more likely the node data is to be abnormal. Simultaneously, the data differences between each node and its neighboring nodes are analyzed, and the difference between the node data and the mean of the neighboring node data is calculated, with the positive and negative directions and absolute values ​​of the differences marked. During the analysis, the deviation and neighborhood difference of each node are recorded according to the attribute type. Combined with the micro-topography grid attributes, different difference judgment benchmarks are set for different clusters, and the difference level of each node is marked (normal, slight difference, significant difference). Among them, the absolute value of deviation ≤ the benchmark is normal, the benchmark value < the absolute value of deviation ≤ 1.5 times the benchmark is slight difference, and the absolute value of deviation > 1.5 times the benchmark is significant difference, generating the difference feature data within the micro-topography road neighborhood cluster.

[0035] Step S25: Analyze the abnormal state of micro-topographic road monitoring nodes based on the differential feature data within the micro-topographic road neighborhood cluster, and generate abnormal state data of micro-topographic road monitoring nodes; In this embodiment of the invention, data is categorized based on the attribute types of micro-topographic road monitoring. Dynamic anomaly thresholds are set for each cluster and each attribute type, with the thresholds determined by a combination of intra-cluster difference judgment criteria and micro-topographic features. Subsequently, an improved isolated forest algorithm is used for anomaly analysis. The algorithm assigns weights to micro-topographic features to the decision tree segmentation nodes. Through weight allocation, anomaly identification is made more closely aligned with the micro-topographic scenario. For example, under the same deviation, the anomaly judgment priority for soil-rock mixed subgrade nodes is higher than that for asphalt concrete subgrade nodes. During the analysis, the deviation and neighborhood difference values ​​from the intra-cluster difference feature data of the micro-topographic road are input into the algorithm. Anomaly judgment is performed on each significantly different node to clarify the anomaly type, which is divided into three fixed anomaly types: equipment failure (data mutation, continuous), signal interference (data fluctuation, irregular), and environmental mutation (caused by sudden changes in the local micro-topographic environment). Environmental mutation types require further verification based on micro-topographic grid attributes, such as whether it is a terrain mutation node or whether there is water accumulation. After the judgment is completed, each abnormal node is labeled with an abnormal code, abnormal type, abnormal start time, abnormal duration, peak deviation, and peak neighborhood difference. At the same time, the degree of abnormality (mild, moderate, and severe) is determined. Mild abnormality is 1.5-2 times the absolute value of deviation, moderate is 2-3 times the threshold, and severe is >3 times the threshold. The abnormal status data of micro-topography road monitoring nodes is generated, which includes detailed information of all abnormal nodes.

[0036] Step S26: Perform micro-topography road monitoring correction processing on the micro-topography road monitoring data through the abnormal state data of the micro-topography road monitoring nodes to generate corrected micro-topography road surface monitoring data.

[0037] In this embodiment of the invention, micro-topographic road monitoring data is corrected using abnormal state data from micro-topographic road monitoring nodes. The correction is categorized by anomaly type. For equipment failure-type anomalies, a neighborhood node interpolation method is used. The average of the concurrent monitoring data from three normal nodes within the same cluster within a 50m radius of the fault node is calculated as the correction value. Weights are assigned based on micro-topographic features during the interpolation process to ensure the correction value closely matches the micro-topographic conditions of the fault node. For signal interference-type anomalies, a time-series smoothing correction method is used, selecting normal monitoring data from 30 minutes before and after the faulty period of the fault node. According to the data, a smooth time-series curve is generated through fitting, and the fitted values ​​corresponding to abnormal periods are extracted as correction values. During the fitting process, the fluctuation range of the curve is controlled to match the time change characteristics of the monitoring data of the cluster. For example, the fluctuation range of the fitted curve for humidity data of the valley cluster is ≤5%. For abnormal nodes with sudden environmental changes, no interpolation or smoothing correction is performed. Instead, the authenticity of the sudden environmental change is verified by combining the micro-topography grid attributes with the data of neighboring nodes in the same period. If it is indeed a local environmental change (such as local snowfall or rime), the original data is retained and the environmental change label is marked. If it is a false change (such as sensor misjudgment), the mean of neighboring nodes is used for correction. After the correction is completed, all corrected data are subjected to double verification. The first verification is the numerical range verification to ensure that the corrected data is within the normal numerical range of the characteristic data of this attribute type. For example, the corrected air temperature value is between -15℃ and 10℃ (the normal range of micro-topography roads in winter). The second verification is the micro-topography adaptation verification to ensure that the corrected data matches the micro-topography characteristics of the grid to which the node belongs. For example, the humidity correction value of the valley grid is not less than 70%, and the wind speed correction value of the steep slope grid is not less than 1.2m / s. If the verification fails, the data is readjusted until the verification meets the standards, and then the corrected micro-topographic pavement monitoring data is generated.

[0038] Furthermore, step S25 includes the following steps: Step S251: Determine the adaptive evaluation threshold for micro-topography and road monitoring status based on the data classification according to the micro-topography road monitoring attribute type, and generate the micro-topography-monitoring difference status evaluation threshold. In this embodiment of the invention, historical valid samples of various attribute monitoring data within each cluster are extracted (excluding obviously abnormal data initially identified). Then, the basic threshold is dynamically adjusted based on the micro-topographic grid attribute characteristics, and different micro-topographic factors are assigned adjustment coefficients. Among them, the terrain undulation adjustment coefficient is 0.8-1.2 (the greater the terrain undulation, the larger the adjustment coefficient and the wider the threshold range, such as 1.2 for steep slope cluster and 0.8 for valley cluster), the roadbed material adjustment coefficient is 0.9-1.1 (1.1 for soil-rock mixed roadbed and 0.9 for asphalt concrete roadbed), and the vegetation coverage adjustment coefficient is 0.95-1.05 (the higher the vegetation coverage, the smaller the adjustment coefficient and the stricter the threshold). The adjustment formula is: adaptive threshold = basic threshold × comprehensive micro-topographic adjustment coefficient (the comprehensive adjustment coefficient is the weighted average of the adjustment coefficients of each micro-topographic factor, with the weights allocated according to terrain undulation 0.4, roadbed material 0.3, and vegetation coverage 0.3). Simultaneously, in conjunction with the needs of micro-topography road icing early warning, the threshold range of core attributes (icing cover monitoring attributes) was tightened and calibrated. The basic threshold was adjusted to the mean of the attribute within the cluster ± 2.5 × the standard deviation of the attribute within the cluster, ensuring that no anomalies in icing-related data were missed. After the threshold calculation was completed, micro-topography-monitoring difference status assessment threshold data was generated.

[0039] Step S252: Establish the matching evaluation mapping relationship between the terrain and monitoring status of micro-terrain roads using the preset isolated forest algorithm through the micro-terrain-monitoring difference status evaluation threshold, and generate a micro-terrain road terrain-monitoring status matching evaluation model. In this embodiment of the invention, the basic structure of the pre-defined isolated forest algorithm is clearly defined. This algorithm contains 100 decision trees, each constructed by randomly selecting features and randomly splitting nodes. The core flaw is that all monitoring features are assigned equal weights, without considering the impact of micro-topographic features on icing warnings, resulting in low accuracy in identifying anomalies in micro-topographic areas. The improvement process first extracts the core micro-topographic grid attributes (topographic relief, roadbed material, and vegetation coverage) of each micro-topographic cluster. Combining the core requirements of micro-topographic road icing warnings, fixed weights are assigned to micro-topographic features and monitoring difference features, with a total weight of 1. Among them, topographic relief has a weight of 0.3 (directly affecting local meteorological conditions, thus affecting the stability of monitoring data and the probability of road icing), roadbed material has a weight of 0.25 (affecting road surface temperature conduction and road surface condition, indirectly affecting the accuracy of monitoring data), vegetation coverage has a weight of 0.2 (affecting near-surface temperature and humidity distribution, leading to fluctuations in monitoring data), and monitoring data difference features have a weight of 0.25 (directly reflecting the degree of deviation between node data and the overall features within the cluster). Subsequently, a matching evaluation mapping relationship between micro-topography and monitoring status was established, accurately mapping the adaptive evaluation threshold to the anomaly level, defining three mapping intervals: when the monitoring data deviation × corresponding attribute weight ≤ the lower limit of the adaptive threshold, it is mapped to "normal state"; when the lower limit of the adaptive threshold < the monitoring data deviation × corresponding attribute weight ≤ the upper limit of the adaptive threshold, it is mapped to "slightly abnormal state"; when the monitoring data deviation × corresponding attribute weight > the upper limit of the adaptive threshold, it is mapped to "significantly abnormal state". This mapping relationship was also integrated into the decision tree splitting nodes of the isolated forest algorithm, modifying the decision tree splitting rules to prioritize micro-topography features and monitoring features with higher weights. After the model was built, the model was trained and calibrated using pre-prepared sample data (normal monitoring data within the cluster, simulated abnormal data, and historical real abnormal data). During training, the depth of the decision tree was adjusted (fixed at 8 layers) and the number of leaf nodes (each leaf node contains at least 5 samples) to ensure that the model did not overfit or underfit, generating a micro-topography road terrain-monitoring status matching evaluation model.

[0040] Step S253: Use the micro-topography road terrain-monitoring status matching evaluation model to analyze the abnormal status of micro-topography road monitoring nodes in the neighborhood cluster of differential characteristic data, and generate abnormal status data of micro-topography road monitoring nodes.

[0041] In this embodiment of the invention, the micro-topographic feature weights and micro-topographic-monitoring difference status assessment thresholds of the cluster to which the node belongs are matched based on the node encoding. Then, a comprehensive anomaly score is calculated for each node according to the attribute type. The comprehensive anomaly score = (deviation of each attribute × corresponding attribute weight) + (neighborhood difference of each attribute × neighborhood weight 0.2), where the neighborhood weight is fixed at 0.2 to enhance the influence of neighborhood differences on anomaly identification. Subsequently, the comprehensive anomaly score is compared with the adaptive threshold of the corresponding attribute of the cluster to determine the anomaly level of each node: if the comprehensive anomaly score ≤ the lower limit of the adaptive threshold, it is determined to be a normal state; if the lower limit of the adaptive threshold < the comprehensive anomaly score ≤ the upper limit of the adaptive threshold, it is determined to be a slightly abnormal state; if the comprehensive anomaly score > the upper limit of the adaptive threshold, it is determined to be a significantly abnormal state. Based on the anomaly level assessment, and combined with the micro-topographic grid attributes and monitoring data change characteristics, the anomaly type is accurately determined, and the criteria for three fixed anomaly types are clarified: equipment failure anomalies, characterized by continuously unchanged monitoring data (e.g., temperature data remaining at the same value for 10 consecutive minutes) or continuous stability after a sudden change (e.g., ice thickness suddenly increasing from 0 mm to 5 mm and then remaining unchanged), and the corresponding micro-topographic area has no obvious environmental abrupt change characteristics (e.g., not a topographic extreme node, no water accumulation); signal interference anomalies, characterized by drastic and irregular fluctuations in monitoring data (e.g., wind speed data fluctuating from 2 m / s to 10 m / s within 1 minute, and then suddenly dropping to 1 m / s). The deviation and neighborhood difference values ​​alternate between positive and negative, and the fluctuation frequency exceeds 3 times per minute; environmental mutation anomalies are characterized by sudden but regular changes in monitoring data, and the corresponding micro-topography areas have obvious environmental mutation conditions (such as terrain extreme nodes, valley water catchment areas, humidity data rises sharply and then stabilizes in the high value range), while the neighboring node data have no obvious anomalies or only slight fluctuations. For each node judged as abnormal, further check whether there are similar nodes in its cluster with the same anomaly. If multiple nodes have the same anomaly at the same time, it is necessary to re-judge it as a signal interference anomaly (excluding single device failure), and generate micro-topography road monitoring node abnormal status data.

[0042] Furthermore, as an embodiment of the present invention, reference is made to... Figure 2 As shown, Figure 1 A detailed flowchart illustrating the implementation steps of step S3 is provided in this embodiment. Step S3 includes: Step S31: Extract pavement topographic feature data and pavement meteorological factor feature data from the corrected micro-topography pavement monitoring data; In this embodiment of the invention, specific extraction indicators for two types of feature data are defined. Five core indicators are extracted from road surface topographic feature data: topographic relief, road surface slope, vegetation coverage, roadbed material, and surface catchment area. These are all key topographic features that directly affect the formation and development of micro-topographic road icing. Six core indicators are extracted from road surface meteorological factor feature data: air temperature, road surface temperature, relative humidity, wind speed, wind direction, and precipitation. These are all meteorological elements directly related to road icing, with road surface temperature being the core indicator for icing determination. During the extraction process, each set of feature data is labeled with the corresponding micro-topographic GIS grid code, monitoring node code, and extraction time to ensure the correspondence between topographic feature data, meteorological factor feature data, grid attributes, and monitoring nodes, avoiding data gaps. Ultimately, road surface topographic feature data and road surface meteorological factor feature data are obtained, accurately covering the core feature indicators related to micro-topographic road icing and quantitatively presenting the topographic and meteorological differences between different grids.

[0043] Step S32: Analyze the influence characteristics of meteorological factors on road surface based on road surface topographic feature data and road surface meteorological factor feature data, and generate road surface topographic meteorological factor influence feature data. In this embodiment of the invention, the road surface topographic feature data is grouped according to its core indicators, resulting in five groups: topographic relief (<5m, 5-10m, >10m), road surface slope (<3°, 3°-8°, >8°), vegetation coverage (<30%, 30%-70%, >70%), roadbed material (asphalt concrete, cement concrete, soil-rock mixture), and surface catchment area (<100㎡, 100-200㎡, >200㎡). Each group contains several micro-topographic GIS grids to ensure sufficient and representative sample size. After grouping, for each group, various statistical indicators of road surface meteorological factor feature data are calculated. Through inter-group comparison, the differences in the same meteorological factor under different topographic conditions are analyzed. Simultaneously, the correlation between topographic feature indicators and meteorological factor indicators is calculated to quantify the influence of topography on meteorological factors. In the specific analysis, the influence of various terrain types on meteorological factors was clarified: the greater the terrain undulation, the lower the air temperature and road surface temperature; for every 10m increase in undulation, the average air temperature decreases by 0.6℃, the average road surface temperature decreases by 0.8℃, and the temperature fluctuation range increases by 15%; the greater the road surface slope, the greater the average wind speed; for every 5° increase in slope, the average wind speed increases by 0.7m / s, and water is less likely to remain on the road surface, resulting in a decrease in the average relative humidity of 8%-12%; the higher the vegetation cover, the greater the average relative humidity of 10m / s. A 20% increase leads to a 7% increase in average relative humidity and a 10% decrease in temperature fluctuation, as vegetation can cause cold air to accumulate. The temperature conduction rate of a soil-rock mixed subgrade is slower than that of an asphalt concrete subgrade; under the same air temperature, the surface temperature of a soil-rock mixed subgrade is 0.5-0.9℃ lower. A larger surface catchment area results in a higher average relative humidity; for every 100㎡ increase in catchment area, the average relative humidity increases by 5%, and the road surface is prone to water accumulation, indirectly affecting road surface temperature changes. Ultimately, this generates data on the impact characteristics of road surface topographic meteorological factors.

[0044] Step S33: Perform correlation analysis on the road meteorological factor characteristic data and icing to generate road meteorological factor-icing correlation data; In this embodiment of the invention, core evaluation indicators related to icing are clearly defined. Three indicators—road surface icing thickness, icing coverage area, and icing duration—are selected as dependent variables for correlation analysis. Six indicators from the road meteorological factor characteristic data—air temperature, road surface temperature, relative humidity, wind speed, wind direction, and precipitation—are selected as independent variables to ensure that all analyzed indicators are directly related to icing. The analysis sample consists of continuous monitoring data from the low-temperature period in winter, with a sample size of no less than 1000 sets, covering icing conditions under different micro-topographic grids and meteorological conditions to ensure the comprehensiveness and representativeness of the sample. The correlation analysis uses the Pearson correlation coefficient method to calculate the correlation coefficient between each type of independent variable (meteorological factor) and each dependent variable (icing index). The correlation coefficient ranges from -1 to 1, with positive numbers indicating positive correlation (the larger the factor value, the more severe the icing) and negative numbers indicating negative correlation (the larger the factor value, the milder the icing). A larger absolute value indicates a stronger correlation. During the analysis, a fixed standard for classifying correlation levels was established: an absolute value of correlation coefficient > 0.7 indicates strong correlation, 0.5-0.7 indicates moderate correlation, < 0.5 indicates weak correlation, and an absolute value of correlation coefficient < 0.3 indicates no correlation. Factors with no correlation were directly eliminated and not included in subsequent analysis, thus generating correlation data between road meteorological factors and icing.

[0045] Step S34: Perform a characteristic analysis of road icing meteorological factors based on the correlation data between road meteorological factors and icing, and generate characteristic data of road icing meteorological factors. In this embodiment of the invention, strongly correlated factors (road surface temperature, air temperature, relative humidity) and moderately correlated factors (precipitation) in the road surface meteorological factor-icing correlation data are retained, while weakly correlated and uncorrelated factors are removed. The analysis focuses on four core icing meteorological factors, with each factor's analysis centered on icing early warning requirements, clearly defining specific characteristic parameters and variation patterns without ambiguity. For road surface temperature (a strongly correlated factor), the analysis focuses on its icing critical threshold, temporal variation characteristics, and impact on icing thickness: the critical icing threshold for road surface temperature is defined as 0℃. When the road surface temperature is ≤0℃ and lasts for ≥2 hours, icing begins; ≤-3℃ and lasts for ≥4 hours, the icing thickness is ≥1mm, reaching the warning threshold. In terms of temporal variation, the period from 23:00 to 7:00 the next day in winter is the lowest point for road surface temperature, during which the probability of icing is highest, with temperature fluctuations ≤0.5℃ / hour. After the road surface temperature rises above 0℃, the icing begins to melt at a melting rate of 0.2mm / hour. For air temperature (a strongly correlated factor), the focus is on analyzing its correlation characteristics with road surface temperature, and the relationship between the duration of low temperatures and icing: the difference between air temperature and road surface temperature is fixed at 0.3-0.8℃. When the air temperature is lower than the road surface temperature, the difference increases as the air temperature decreases. If the duration of air temperature ≤0℃ is ≥6 hours, road surface icing will inevitably occur. For every 2-hour increase in duration, the icing thickness increases by 0.4mm. If the duration of air temperature ≤-5℃ is ≥3 hours, the icing thickness is ≥1.5mm, which is considered high-risk icing. For relative humidity (a strongly correlated factor), the focus is on analyzing its critical threshold, coupling characteristics with temperature, and its impact on icing coverage area. Relative humidity and temperature are negatively correlated; the lower the temperature, the stronger the impact of relative humidity on icing. For precipitation (a moderately correlated factor), the focus is on analyzing its precipitation type, precipitation threshold, and coupling effect with temperature. The generated meteorological factor characteristic data for road icing accurately characterizes the core features of the four core icing meteorological factors and their impact on icing.

[0046] Step S35: Analyze the meteorological impact characteristics of road surface icing on the meteorological factor characteristic data of road surface topography using the road surface topography meteorological factor impact characteristic data, and generate road surface topography icing meteorological impact characteristic data.

[0047] In this embodiment of the invention, road surface topographic meteorological factor influence characteristic data (the influence law of topography on meteorological factors) and road surface icing meteorological factor characteristic data (the influence characteristics of core meteorological factors on icing) are integrated. Using micro-topographic GIS grids as the basic analysis unit, coupled analysis is conducted by grouping according to topographic features to analyze the differentiated influence of meteorological factors on icing under different topographic conditions. The specific analysis process is quantified and clearly defined. Combining the coupling relationship between topography and meteorological factors, specific influence laws are explored: In the topographic relief group, grids with topographic relief > 10m (valleys, plateau edges) have road surface temperatures 1.2-1.8℃ lower than grids with topographic relief < 5m. At the same air temperature (-3℃), the icing thickness in grids with topographic relief > 10m is 0.5-0.7mm thicker than that in grids with topographic relief < 5m, and the icing duration is increased by 3-4 hours. This is because large topographic relief easily leads to cold air accumulation, strengthening the influence of meteorological factors on icing. In the road surface slope group, gentle slope grids with a slope < 3° are prone to water retention on the road surface. Under humidity (85%) and road surface temperature (-2℃), the icing thickness of the gentle slope grid is 0.3-0.5 mm thicker than that of the steep slope grid with a slope greater than 8°, and the icing coverage area is 20% larger. However, the steep slope grid has higher wind speeds, and water is easily washed away, resulting in a relatively thinner icing thickness. In the vegetation coverage group, the relative humidity of the grid with vegetation coverage >70% is 15-20% higher than that of the grid with <30%. At the same air temperature (-4℃), the icing rate of the high vegetation coverage grid is 0.2 mm / h faster than that of the low vegetation coverage grid, and the icing duration is increased by 2 hours. This is because the vegetation blocks the cold air, which is trapped, increasing the relative humidity and intensifying the icing process. For each type of terrain and meteorological factor coupling combination, the core icing characteristic parameters (icing thickness, coverage area, growth rate, and duration) are calculated, and the coupling influence coefficient is defined, which is the quantification degree of the influence of terrain features on meteorological factors. At the same time, the icing risk level (high, medium, low) of each coupling combination is marked, generating road surface terrain icing meteorological influence characteristic data.

[0048] Furthermore, step S4 includes the following steps: Step S41: Obtain road weather forecast data; In this embodiment of the invention, the road weather forecast data is obtained from an authoritative meteorological agency in the area where the target micro-topography road is located. The forecast range fully covers the entire route of the target micro-topography road from point A to point B and the surrounding 5km area, and the forecast duration is set to the next 7 days. The forecast data includes six core meteorological indicators, namely air temperature, relative humidity, wind speed, wind direction, precipitation, and road surface temperature.

[0049] Step S42: Perform time series feature analysis on the corrected micro-topography pavement monitoring data to generate micro-topography pavement monitoring time series feature data; In this embodiment of the invention, the four core meteorological indicators—air temperature, road surface temperature, relative humidity, and wind speed—in the micro-topographic road surface monitoring data are sequentially subjected to time-series characteristic statistics and analysis. At a collection frequency of one minute, the mean, maximum, minimum, and fluctuation amplitude of the monitoring data within each hour are statistically analyzed, and the fluctuation coefficient of the hourly data is calculated to quantify the stability of the hourly meteorological data. At a daily scale, the daily temperature peak and trough values ​​are statistically analyzed, and the duration and average temperature of the daily low-temperature period are calculated to clarify the time-series characteristics of the daily high-incidence period of icing. Simultaneously, time-series difference analysis is conducted for each micro-topographic cluster to compare the time-series characteristic patterns of different clusters. For each type of meteorological indicator and each micro-topographic cluster, time-series characteristic parameters are labeled, including the hourly fluctuation amplitude threshold, daily peak and trough time, low-temperature duration threshold, and temperature recovery rate (0.2-0.3℃ / hour), clarifying the core patterns of time-series changes, and using the time-series characteristic data of micro-topographic road surface monitoring.

[0050] Step S43: Analyze the micro-topography road surface meteorological model based on the time-series characteristic data of micro-topography road surface monitoring and road weather forecast data to generate micro-topography road surface meteorological model data; In this embodiment of the invention, based on the spatial correspondence between the micro-topography GIS grid and the road meteorological forecast grid, each micro-topography cluster is matched to the corresponding large-scale forecast grid (10km×10km) to ensure that the meteorological time-series characteristics of each cluster are accurately correlated with the forecast data of the corresponding forecast grid. Subsequently, adaptability analysis is carried out by micro-topography cluster and meteorological index, calculating the deviation between the monitoring time-series mean of each cluster and the forecast mean of the corresponding forecast grid during the same period, and quantifying the deviation pattern. Based on the quantified deviation, the meteorological change pattern of each micro-topography cluster is explored, and the core parameters of the pattern are identified, including the deviation coefficient, the time lag time, and the fluctuation adaptability coefficient. The deviation coefficient is the ratio of the monitoring mean to the forecast mean (a fixed value), the time lag time is the time that the micro-topography meteorological change lags behind the large-scale forecast (30 minutes for valley cluster, 15 minutes for steep slope cluster, and no lag for flat slope cluster), and the fluctuation adaptability coefficient is the ratio of the monitoring fluctuation amplitude to the forecast fluctuation amplitude (1.2 for valley cluster, 1.5 for steep slope cluster, and 0.9 for flat slope cluster). Meanwhile, by combining the low-temperature period patterns in the time-series characteristic data of micro-topography road surface monitoring, the time adaptation parameters of the meteorological model are adjusted to generate meteorological model data for micro-topography road surfaces.

[0051] Step S44: Perform macro-climate trend simulation processing on micro-topography road surface based on micro-topography road surface meteorological model data to generate macro-climate trend simulation data for micro-topography road surface. In this embodiment of the invention, a micro-topographic road GIS grid (50m×50m) is used as the basic simulation unit. A bilinear interpolation method is employed to interpolate large-scale road meteorological forecast data (10km×10km) to each 50m×50m micro-topographic grid, achieving spatial extension of large-scale data to the micro-topographic grid. During the interpolation process, the time scale of the forecast data is retained (1 hour / interpolation), ensuring that each micro-topographic grid has a corresponding interpolated forecast value every hour. Subsequently, combined with micro-topographic road meteorological model data, the interpolated forecast values ​​of each micro-topographic grid are model-corrected. The interpolated values ​​are adjusted according to the model parameters adapted to the micro-topographic cluster, and the simulation timestamp of the corresponding grid is adjusted according to the time lag (30 minutes for valley clusters). The wind speed interpolation value for steep slope clusters is adjusted by a deviation coefficient of 1.3, and the temperature fluctuation amplitude is adjusted by a fluctuation adaptation coefficient of 1.5, ensuring that the corrected simulation values ​​conform to the meteorological characteristics of the micro-topography. During the simulation, the focus was on correcting core meteorological indicators related to road icing (air temperature, road surface temperature, and relative humidity). The temporal distribution of simulated values ​​was adjusted by incorporating the low-temperature periods observed in the time-series data of micro-topographic road surface monitoring. After the simulation, the generated preliminary simulation data was validated for errors. One hundred evenly distributed micro-topographic monitoring nodes were selected, and the simulated values ​​from the same period were compared with the measured values ​​from the corrected micro-topographic road surface monitoring data. The simulation error (absolute value) for each indicator was calculated. The simulation errors for air temperature, road surface temperature, relative humidity, and wind speed were controlled to be ≤2.0℃, ≤1.8℃, ≤7%, and ≤0.8m / s, respectively. If the error exceeded the threshold, the interpolation method and model correction parameters were readjusted until all indicator errors met the requirements, thus generating macro-climate trend simulation data for micro-topographic road surfaces.

[0052] Step S45: Perform micro-topography accuracy adaptation and optimization processing on the micro-topography road surface macro-climate trend simulation data to generate optimized topography road surface macro-climate trend simulation data; In this embodiment of the invention, five core attributes (terrain relief, road surface slope, vegetation coverage, roadbed material, and surface catchment area) are extracted from the micro-topographic road GIS grid attribute data. These attributes are grouped by value, and a fixed accuracy optimization coefficient is assigned to each group. This coefficient is determined based on the influence of terrain on meteorological factors in the road surface topographic meteorological factor influence characteristic data. Subsequently, optimization is performed on each micro-topographic grid individually. The simulated value of each grid is multiplied by the corresponding attribute's optimization coefficient. Simultaneously, the correlation between terrain and meteorological factors in the road surface topographic meteorological factor influence characteristic data is considered to perform a secondary adjustment on the core icing meteorological indicators. During the optimization process, special attention is paid to the simulation accuracy of grids with high icing susceptibility. The simulation data for these grids undergoes additional verification to ensure that the road surface temperature simulation error is ≤1.5℃ and the relative humidity simulation error is ≤5%. After optimization, error verification is performed again. Using the 100 monitoring nodes from step S44, the optimized simulated values ​​are compared with the measured values, and the error is calculated and corrected to generate optimized topographic road surface macroclimate trend simulation data.

[0053] Step S46: Based on the optimized topographic and road surface macroclimate trend simulation data, perform microclimate trend simulation processing on the downscaled topographic zoning to generate topographic zoning microclimate trend simulation data; In this embodiment of the invention, the micro-topography monitoring clusters are divided into 6 terrain downscaling zones. Each zone contains several micro-topography GIS grids, and the core micro-topography features of each zone are clearly defined (e.g., cluster 1 is a valley zone, cluster 2 is a steep slope zone). The optimized terrain and road surface macroclimate trend simulation data are split by zone, with each zone corresponding to a set of optimized simulation data. Subsequently, a refined downscaling simulation is carried out for each zone. The 50m×50m grid is further subdivided into 10m×10m smaller grids. The Kriging interpolation method is used to interpolate the optimized simulation data of the 50m grid within the zone to the 10m smaller grid. During the interpolation process, a time scale of 1 hour / time is retained to ensure that each 10m smaller grid has a corresponding refined simulation value for each hour. After interpolation, the simulated values ​​of the 10m grid are corrected for regional characteristics based on the core micro-topographic features of each region. For example, for the 10m grid of the valley region (cluster 1), if it is located at the bottom of the valley (topographic relief > 12m), the simulated air temperature is reduced by 0.3℃ and the relative humidity is increased by 2%. For the 10m grid of the steep slope region (cluster 2), if the slope is > 10°, the simulated wind speed is increased by 0.2m / s. For the 10m grid of the flat slope region (cluster 3), the simulated values ​​maintain interpolation accuracy, only adjusting the fluctuation range to ≤ 0.1℃ / hour. During the simulation, the focus is on simulating the local microclimate differences within the region. For example, in the valley region, the temperature and humidity differences between the bottom and edge of the valley; in the steep slope region, the wind speed differences between the top and bottom of the slope. This ensures that the simulation data can capture subtle meteorological changes within the region and generate simulated microclimate trends for the topographic regions.

[0054] Step S47: Adapt and optimize the icing meteorological parameters of the terrain zoning microclimate trend simulation data by using the road surface icing meteorological impact characteristic data to generate terrain zoning microclimate trend optimized simulation data; In this embodiment of the invention, the core icing meteorological parameters are clearly defined, and four parameters directly related to icing—air temperature, road surface temperature, relative humidity, and precipitation—are selected as optimization targets. Parameters that are unrelated to or weakly related to icing (wind direction) are eliminated. Focusing on each micro-topographic zone, and combining the coupling influence law in the road surface topography icing meteorological influence characteristic data, zone-specific optimization rules and optimization coefficients are set. The specific optimization process is carried out by zone, and the optimization rules for each zone strictly adhere to the coupling relationship between its topography and icing meteorological factors: Valley zone (cluster 1), taking into account the characteristics of large topographic relief, easy accumulation of cold air, and high relative humidity, the simulated road surface temperature value is optimized with a coefficient of 0.96 (simulated value × 0.96) to ensure that the simulated road surface temperature value closely matches the actual high icing susceptibility of valleys. When the simulated air temperature is ≤0℃, the simulated relative humidity value is further increased by 3%, and the simulated precipitation value is optimized with a coefficient of 1.05 to strengthen the impact of precipitation on icing; Steep slope zone (cluster 2), taking into account the characteristics of large slope, high wind speed, and easy runoff of water, when the simulated wind speed is >3m / s, the simulated road surface temperature value is optimized with a coefficient of 1.02 (simulated value × 1.02), the simulated relative humidity value is optimized with a coefficient of 0.98 to reduce the impact of humidity on icing, and the simulated precipitation value is optimized with a coefficient of 0.97 to weaken the effect of water accumulation on icing. During the optimization process, the focus is on the simulation parameters near the critical threshold of icing. For example, when the simulated air temperature is close to 0℃ (-0.5℃ to 0.5℃), the simulated road surface temperature is precisely fine-tuned to ensure that the simulated value can accurately reflect whether the road surface has reached the critical state of icing. When the simulated air temperature is ≤0℃ and the relative humidity is ≥80%, the simulated road surface temperature is further reduced by 0.2℃ to match the actual meteorological conditions for icing formation, and to generate optimized simulation data of microclimate trends for terrain zones.

[0055] Step S48: Perform dynamic climate downscaling and iterative optimization on the micro-topography road surface macro-climate trend simulation data by optimizing the micro-climate trend simulation data of the terrain zoning, and generate micro-topography road surface climate trend simulation data.

[0056] In this embodiment of the invention, the iteration rules and parameters are clearly defined, and the number of iterations is set to 3 rounds, with each round of iteration carried out according to micro-topography zones. In the first round of iteration, the microclimate trend optimization simulation data (10m small grid) of the terrain zone is matched with the macroclimate trend simulation data of the micro-topography road surface (50m grid) by spatial coordinates. The mean of the optimized simulation values ​​of all 10m small grids within each 50m grid is calculated, and this mean is used as the correction value of that 50m grid. The original macroclimate simulation value is replaced with the correction value, and the correction deviation (correction value - original simulation value) is calculated. The average correction deviation of each zone is marked. In the second round of iteration, based on the correction deviation of the first round, the 10m small grid optimization simulation values ​​of each zone are adjusted. Adjustments are made, with the adjustment range being 30% of the average correction deviation for the corresponding partition. Then, the correction value for the 50m grid is calculated, replacing the macro-climate simulation value. The correction deviation is calculated again to ensure that the deviation is reduced by more than 50% compared to the first round. In the third iteration, the second round is repeated, with the adjustment range being 20% ​​of the average correction deviation for the second round. The final correction value is calculated and replaced with the macro-climate simulation value. After completing three iterations, the average correction deviation for each partition is ensured to be ≤0.3℃ (temperature), ≤1% (humidity), ≤0.1m / s (wind speed), and ≤0.1mm (precipitation). During the iteration process, the focus is on the iteration accuracy of zones with high icing susceptibility (valleys, mixed soil and rock roadbeds). Additional deviation checks are added to these zones to ensure that the simulation data after iteration can accurately reflect their icing meteorological trends. For example, the simulation error of road surface temperature after iteration in valley zones is ≤0.7℃, and the simulation error of relative humidity after iteration is ≤2%. The final generated micro-topographic road surface climate trend simulation data takes into account both large-scale meteorological trends and local microclimate details of micro-topography. It has high simulation accuracy and strong adaptability, and includes the simulation values ​​of core meteorological indicators for each micro-topographic grid and every hour, which can accurately reflect the meteorological development trends of different micro-topographic regions.

[0057] Furthermore, step S5 includes the following steps: Step S51: Based on the meteorological impact characteristics of road surface topography icing, establish a multi-source logistic regression mapping relationship between micro-topography meteorology and icing, and generate a micro-topography meteorological icing regression model; In this embodiment of the invention, the core variables of the regression model are clearly defined. The independent variables are selected from four core icing meteorological factors (air temperature, road surface temperature, relative humidity, and precipitation) and three key micro-topographic features (topographic relief, roadbed material, and vegetation cover). Roadbed material is quantified using coding, while the other independent variables are all quantitative values. The dependent variable is road surface icing thickness, which is a continuous variable to ensure that the regression model can quantitatively predict icing thickness. The samples are selected from continuous monitoring data during low-temperature periods in winter and data on the meteorological impact of road surface icing, divided into training and validation samples in a 7:3 ratio. In the regression analysis, a multi-source logistic regression algorithm was used to calculate the regression coefficient and significance level of each independent variable. The significance level was set at 0.05, and irrelevant factors with a significance level > 0.05 were eliminated (ultimately, all selected factors passed the significance test). A fixed regression equation was determined: Ice thickness = α × air temperature + β × road surface temperature + γ × relative humidity + δ × precipitation + ε × topographic relief + ζ × roadbed material code + η × vegetation cover + κ, where α, β, γ, δ, ε, ζ, ... η is the regression coefficient, and κ is the constant term. The regression coefficients and constant term differ for each micro-topographic region (to match the micro-topographic characteristics of the region). For example, for the valley region, α=-0.32, β=-0.45, γ=0.06, δ=0.08, ε=0.05, ζ=0.12, η=0.03, κ=2.15; for the flat slope region, α=-0.28, β=-0.40, γ=0.05, δ=0.07, ε=0.02, ζ=0.10, η=0.02, κ=1.85. After the model is built, it is calibrated using validation samples, and the regression coefficients are adjusted to ensure the goodness of fit, thus generating a micro-topographic meteorological icing regression model.

[0058] Step S52: Perform quantitative correlation analysis on moisture and icing of micro-topographic pavement based on the corrected micro-topographic pavement monitoring data to generate moisture-icing correlation data for micro-topographic pavement. In this embodiment of the invention, specific analytical indicators are defined. The pavement moisture index is selected as the surface moisture content, extracted by conversion from pavement condition monitoring data in the corrected micro-topographic pavement monitoring data. The icing index is selected as ice thickness and ice coverage area, ensuring that all indicators are quantifiable core parameters. The analysis samples are selected from continuous corrected monitoring data during low-temperature periods in winter (air temperature ≤ 0℃), with a sample size of no less than 1000 sets, covering 6 micro-topographic zones, different pavement conditions (damp, waterlogged), and different moisture ranges, ensuring that the samples can comprehensively reflect the correlation characteristics between moisture and icing. Invalid samples with a moisture content ≤ 2% (dry state, no icing) are removed. The analysis was conducted by dividing the data into micro-topographical zones and moisture intervals. First, multiple moisture intervals were defined. For samples within each interval, the average ice thickness and average ice coverage area were statistically analyzed. The correlation coefficients between moisture content and ice thickness and coverage area were calculated (using the Pearson correlation coefficient method). The correlation patterns were clarified: when the moisture content was 3%-5%, ice formation only occurred when the air temperature was ≤-3℃ for ≥2 hours, with an average ice thickness of 0.3-0.5mm and an average coverage area of ​​10%-20%, resulting in a correlation coefficient of 0.68; when the moisture content was 5%-10%, ice formation occurred when the air temperature was ≤-1℃, with an average ice thickness of 0.6-1.2mm and an average coverage area of ​​30%-50%, resulting in a correlation coefficient of 0.82, and so on. The higher the moisture content, the larger the correlation coefficient. Simultaneously, the correlation differences between different micro-topographic zones were analyzed: the valley zone has a lower moisture content threshold (freezing occurs at 3%), and at the same moisture content, the ice thickness is 0.2-0.4 mm thicker than that of the flat slope zone; in the soil-rock mixed subgrade zone, at the same moisture content, the ice thickness is 0.3-0.5 mm thicker than that of the asphalt concrete subgrade zone (poor subgrade insulation, moisture easily freezes); in the steep slope zone, at the same moisture content, the ice thickness is 0.1-0.2 mm thinner than that of the valley zone (higher wind speed, faster moisture evaporation). After the analysis, micro-topographic pavement moisture-icing correlation data were generated.

[0059] Step S53: Establish a quantitative mapping relationship between moisture and icing on micro-topographic pavements using moisture-icing correlation data, and generate a moisture-icing relationship model for micro-topographic pavements. In this embodiment of the invention, the mapping relationship type is first clarified. Combining the correlation patterns of micro-topographic pavement moisture-icing data, a piecewise linear mapping method is adopted (different mapping coefficients for different moisture intervals) to construct a quantitative mapping equation between moisture content and icing thickness. The core form of the mapping equation is: Icing thickness = k × pavement moisture content + b, where k is the mapping coefficient and b is the correction coefficient. The values ​​of k and b vary with the micro-topographic zones and moisture intervals to ensure that the mapping relationship conforms to the correlation patterns under different conditions. After the mapping equation is determined, model calibration is carried out. Using S52 validation samples (200 sets), the pavement moisture content in the samples is substituted into the mapping equation to calculate the predicted icing thickness. The predicted icing thickness is compared with the measured icing thickness, and the values ​​of k and b are adjusted to ensure the prediction error for each zone and each moisture interval. Simultaneously, applicable conditions are set for the model, clarifying that the mapping equation only takes effect when the air temperature is ≤0℃ and the pavement moisture content is ≥3%. When the air temperature is >0℃ or the moisture content is <3%, the model outputs an icing thickness of 0mm (no icing), ensuring the applicability of the model. Generate a moisture-icing relationship model for micro-topographic road surfaces.

[0060] Step S54: Based on the micro-topography meteorological icing regression model and the micro-topography road moisture-icing relationship model, establish the prediction relationship for micro-topography road icing thickness and generate a micro-topography road icing thickness prediction model.

[0061] In this embodiment of the invention, fixed weights are assigned based on the prediction accuracy of the two models. The weight of the micro-topographic meteorological icing regression model (emphasizing the coupling of meteorology and topography) is 0.45, and the weight of the micro-topographic road moisture-icing relationship model (emphasizing the physical mechanism of moisture) is 0.55. This weight allocation is determined based on the model validation accuracy (the moisture-icing model has slightly higher prediction accuracy, so its weight is higher), and the weights for different micro-topographic zones are slightly adjusted. Subsequently, a comprehensive prediction equation is constructed, with the core form as follows: Final icing thickness prediction value = (meteorological icing regression model prediction value × meteorological model weight) + (moisture-icing relationship model prediction value × moisture model weight) + λ, where λ is the micro-topographic correction coefficient, ranging from 0.02 to 0.08, determined by the core characteristics of the micro-topographic zone (the greater the topographic relief and the higher the vegetation coverage, the larger the λ value). For example, λ = 0.07 for the valley zone, λ = 0.05 for the steep slope zone, λ = 0.03 for the flat slope zone, and λ = 0.08 for the soil-rock mixed roadbed zone. After the prediction model was built, model calibration was carried out. 150 sets of calibration monitoring data (covering all micro-topographic zones and different meteorological and moisture conditions) were selected during the low-temperature period of winter. The data were substituted into two basic models to calculate their respective predicted values, which were then substituted into the comprehensive prediction equation. The final predicted values ​​were compared with the measured icing thickness, and the weights and λ value were adjusted to ensure that the overall prediction error of the model was ≤0.3mm, the goodness of fit R²≥0.95, and the prediction accuracy for each micro-topographic zone ≥98%. Simultaneously, input and output parameters were set for the model. Input parameters included core meteorological factors (air temperature, road surface temperature, relative humidity, precipitation), micro-topographic features (topographic relief, roadbed material, vegetation cover), and road surface moisture content. The output parameter was the predicted road surface icing thickness, along with the prediction accuracy and icing risk level. The final micro-topographic road surface icing thickness prediction model integrates the coupled influence of meteorology and topography with the physical mechanism of moisture, accurately adapting to the icing patterns of different micro-topographic regions and achieving quantitative prediction of road icing thickness in micro-topographic areas.

[0062] This specification provides a micro-topography road icing early detection, prediction, and warning system for performing the micro-topography road icing early detection, prediction, and warning method as described above. The micro-topography road icing early detection, prediction, and warning system includes: The micro-topography road monitoring module is used to acquire data of the target micro-topography road area; it performs micro-topography road monitoring processing on the target micro-topography road area data through multi-source heterogeneous monitoring equipment to generate micro-topography road monitoring data. The micro-topography road surface monitoring anomaly correction module is used to correct and adjust the micro-topography road monitoring node anomalies in the micro-topography road monitoring data and generate corrected micro-topography road surface monitoring data. The topographic icing meteorological impact characteristic analysis module is used to analyze the meteorological impact characteristics of road surface icing based on the corrected micro-topographic road surface monitoring data, and generate road surface topographic icing meteorological impact characteristic data. The climate trend simulation module is used to acquire road weather forecast data; based on the corrected micro-topography pavement monitoring data and road weather forecast data, it performs dynamic downscaling micro-topography pavement climate trend simulation processing to generate micro-topography pavement climate trend simulation data. The icing thickness prediction model building module is used to establish a micro-topographic road surface icing thickness prediction model based on road surface topography icing meteorological impact characteristic data and corrected micro-topographic road surface monitoring data. The micro-topography road icing early warning module is used to perform early sensing and prediction processing of micro-topography road climate trend simulation data using a micro-topography road icing thickness prediction model, generate early sensing and prediction data of micro-topography road icing, and execute micro-topography road icing early warning operations based on the early sensing and prediction data of micro-topography road icing.

[0063] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0064] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for early sensing, prediction, and warning of road icing in micro-topography, characterized in that, Includes the following steps: Step S1: Obtain target micro-topography road area data; Micro-topographic road monitoring data is generated by processing the target micro-topographic road area data using multi-source heterogeneous monitoring equipment. Step S2: Correct and adjust the micro-topography road monitoring node anomalies in the micro-topography road monitoring data to generate corrected micro-topography road surface monitoring data; Step S3: Based on the corrected micro-topographic pavement monitoring data, analyze the meteorological impact characteristics of pavement topography on icing, and generate pavement topography icing meteorological impact characteristic data; Step S4: Obtain road weather forecast data; Based on the road surface topography icing meteorological impact characteristic data, perform dynamic downscaling micro-topography road surface climate trend simulation processing on the corrected micro-topography road surface monitoring data and road weather forecast data to generate micro-topography road surface climate trend simulation data; Step S5: Establish a micro-topographic road surface icing thickness prediction model based on road surface topographic icing meteorological impact characteristic data and corrected micro-topographic road surface monitoring data; Step S6: Use the micro-topography road surface icing thickness prediction model to perform early sensing and prediction processing on the micro-topography road surface climate trend simulation data, generate early sensing and prediction data of micro-topography road icing, and execute early warning operation of micro-topography road icing based on the early sensing and prediction data of micro-topography road icing.

2. The method for early sensing, prediction, and warning of road icing in micro-topography according to claim 1, characterized in that, The micro-topography road monitoring data mentioned in step S1 includes on-site meteorological data of micro-topography roads, road surface condition data of micro-topography roads, and road surface icing coverage data of micro-topography roads.

3. The method for early sensing, prediction, and warning of road icing in micro-topography according to claim 1, characterized in that, Step S1 includes the following steps: Step S11: Obtain target micro-topography road area data; Step S12: Collect micro-topographic road GIS data from the target micro-topographic road area data to obtain micro-topographic road GIS data; Step S13: Perform GIS grid division processing on the micro-topography road GIS data to generate micro-topography road GIS grid data, and perform grid attribute labeling processing on the micro-topography road GIS grid data to generate micro-topography road GIS grid attribute data. Step S14: Perform extreme value node analysis on the micro-topography road GIS grid data to generate extreme value node data of the road GIS grid; Step S15: Analyze the micro-topography characteristics of roads using micro-topography road GIS grid attribute data and extreme value node data of road GIS grid, and generate road micro-topography characteristic data; Step S16: Design the configuration nodes of the multi-source heterogeneous road monitoring equipment based on the road micro-topography characteristic data, and obtain the configuration node data of the multi-source heterogeneous road monitoring equipment; Step S17: Configure the multi-source heterogeneous road monitoring equipment with node data. After configuration, the multi-source heterogeneous monitoring equipment performs micro-topographic road monitoring processing on the target micro-topographic road area data to generate micro-topographic road monitoring data.

4. The method for early sensing, prediction, and warning of road icing in micro-topography according to claim 3, characterized in that, Step S14 includes the following steps: Step S141: Perform grid elevation gradient analysis based on micro-topography road GIS grid data to generate grid elevation gradient data; Step S142: Extract the grid center elevation gradient extreme value nodes from the grid elevation gradient data to obtain the grid center elevation gradient extreme value node data; Step S143: Perform road GIS grid extreme node integration processing on the grid center elevation gradient extreme node data to generate road GIS grid extreme node data.

5. The method for early sensing, prediction, and warning of road icing in micro-topography according to claim 3, characterized in that, Step S2 includes the following steps: Step S21: Based on the micro-topography road GIS grid attribute data, classify the micro-topography road monitoring attribute types of the micro-topography road monitoring data, and generate micro-topography road monitoring attribute type classification data; Step S22: Perform attribute type feature analysis on the micro-topography road monitoring attribute type classification data to generate micro-topography road monitoring attribute type feature data; Step S23: Perform clustering and segmentation processing on the feature data of each attribute type of micro-topography road monitoring to generate micro-topography road monitoring feature cluster data; Step S24: Perform intra-cluster differential feature analysis on the micro-topography road monitoring feature cluster data to generate intra-cluster differential feature data of the micro-topography road neighborhood; Step S25: Analyze the abnormal state of micro-topographic road monitoring nodes based on the differential feature data within the micro-topographic road neighborhood cluster, and generate abnormal state data of micro-topographic road monitoring nodes; Step S26: Perform micro-topography road monitoring correction processing on the micro-topography road monitoring data through the abnormal state data of the micro-topography road monitoring nodes to generate corrected micro-topography road surface monitoring data.

6. The method for early sensing, prediction, and warning of road icing in micro-topography according to claim 5, characterized in that, Step S25 includes the following steps: Step S251: Determine the adaptive evaluation threshold for micro-topography and road monitoring status based on the data classification according to the micro-topography road monitoring attribute type, and generate the micro-topography-monitoring difference status evaluation threshold. Step S252: Establish the matching evaluation mapping relationship between the terrain and monitoring status of micro-terrain roads using the preset isolated forest algorithm through the micro-terrain-monitoring difference status evaluation threshold, and generate a micro-terrain road terrain-monitoring status matching evaluation model. Step S253: Use the micro-topography road terrain-monitoring status matching evaluation model to analyze the abnormal status of micro-topography road monitoring nodes in the neighborhood cluster of micro-topography roads, and generate abnormal status data of micro-topography road monitoring nodes.

7. The method for early sensing, prediction, and warning of road icing in micro-topography according to claim 1, characterized in that, Step S3 includes the following steps: Step S31: Extract pavement topographic feature data and pavement meteorological factor feature data from the corrected micro-topography pavement monitoring data; Step S32: Analyze the influence characteristics of meteorological factors on road surface based on road surface topographic feature data and road surface meteorological factor feature data, and generate road surface topographic meteorological factor influence feature data. Step S33: Perform correlation analysis on the road meteorological factor characteristic data and icing to generate road meteorological factor-icing correlation data; Step S34: Perform a characteristic analysis of road icing meteorological factors based on the correlation data between road meteorological factors and icing, and generate characteristic data of road icing meteorological factors. Step S35: Analyze the meteorological impact characteristics of road surface icing on the meteorological factor characteristic data of road surface topography using the road surface topography meteorological factor impact characteristic data, and generate road surface topography icing meteorological impact characteristic data.

8. The method for early sensing, prediction, and warning of road icing in micro-topography according to claim 1, characterized in that, Step S4 includes the following steps: Step S41: Obtain road weather forecast data; Step S42: Perform time series feature analysis on the corrected micro-topography pavement monitoring data to generate micro-topography pavement monitoring time series feature data; Step S43: Analyze the micro-topography road surface meteorological model based on the time-series characteristic data of micro-topography road surface monitoring and road weather forecast data to generate micro-topography road surface meteorological model data; Step S44: Perform macro-climate trend simulation processing on micro-topography road surface based on micro-topography road surface meteorological model data to generate macro-climate trend simulation data for micro-topography road surface. Step S45: Perform micro-topography accuracy adaptation and optimization processing on the micro-topography road surface macro-climate trend simulation data to generate optimized topography road surface macro-climate trend simulation data; Step S46: Based on the optimized topographic and road surface macroclimate trend simulation data, perform microclimate trend simulation processing on the downscaled topographic zoning to generate topographic zoning microclimate trend simulation data; Step S47: Adapt and optimize the icing meteorological parameters of the terrain zoning microclimate trend simulation data by using the road surface icing meteorological impact characteristic data to generate terrain zoning microclimate trend optimized simulation data; Step S48: Perform dynamic climate downscaling and iterative optimization on the micro-topography road surface macro-climate trend simulation data by optimizing the micro-climate trend simulation data of the terrain zoning, and generate micro-topography road surface climate trend simulation data.

9. The method for early sensing, prediction, and warning of road icing in micro-topography according to claim 1, characterized in that, Step S5 includes the following steps: Step S51: Based on the meteorological impact characteristics of road surface topography icing, establish a multi-source logistic regression mapping relationship between micro-topography meteorology and icing, and generate a micro-topography meteorological icing regression model; Step S52: Perform quantitative correlation analysis on moisture and icing of micro-topographic pavement based on the corrected micro-topographic pavement monitoring data to generate moisture-icing correlation data for micro-topographic pavement. Step S53: Establish a quantitative mapping relationship between moisture and icing on micro-topographic pavements using moisture-icing correlation data, and generate a moisture-icing relationship model for micro-topographic pavements. Step S54: Based on the micro-topography meteorological icing regression model and the micro-topography road moisture-icing relationship model, establish the prediction relationship for micro-topography road icing thickness and generate a micro-topography road icing thickness prediction model.

10. A micro-terrain road icing early sensing, prediction and early warning system, characterized in that, For executing the micro-topography road icing early sensing, prediction, and early warning method as described in claim 1, the micro-topography road icing early sensing, prediction, and early warning system includes: The micro-topography road monitoring module is used to acquire data of the target micro-topography road area; it performs micro-topography road monitoring processing on the target micro-topography road area data through multi-source heterogeneous monitoring equipment to generate micro-topography road monitoring data. The micro-topography road surface monitoring anomaly correction module is used to correct and adjust the micro-topography road monitoring node anomalies in the micro-topography road monitoring data and generate corrected micro-topography road surface monitoring data. The topographic icing meteorological impact characteristic analysis module is used to analyze the meteorological impact characteristics of road surface icing based on the corrected micro-topographic road surface monitoring data, and generate road surface topographic icing meteorological impact characteristic data. The climate trend simulation module is used to acquire road weather forecast data; based on the corrected micro-topography pavement monitoring data and road weather forecast data, it performs dynamic downscaling micro-topography pavement climate trend simulation processing to generate micro-topography pavement climate trend simulation data. The icing thickness prediction model building module is used to establish a micro-topographic road surface icing thickness prediction model based on road surface topography icing meteorological impact characteristic data and corrected micro-topographic road surface monitoring data. The micro-topography road icing early warning module is used to perform early sensing and prediction processing of micro-topography road climate trend simulation data using a micro-topography road icing thickness prediction model, generate early sensing and prediction data of micro-topography road icing, and execute micro-topography road icing early warning operations based on the early sensing and prediction data of micro-topography road icing.

Images