Road icing disaster risk early warning method and device for road logistics transportation

Abstract

The invention relates to the technical field of risk early warning, and particularly discloses a road icing disaster risk early warning method and device for road logistics transportation. A data processing algorithm based on deep learning is used for carrying out space-time interaction response analysis on traffic blocking live data generated by road icing along a highway and rainfall, air temperature and relative humidity grid point forecast data published by a meteorological department so as to excavate road icing characteristics caused by meteorological elements of the highway; and on the basis, prediction of accumulated icing duration and judgment of a road icing traffic blocking risk level are carried out, so that road icing disaster risk early warning for road logistics transportation is realized. In this way, the accuracy and timeliness of road icing early warning can be remarkably improved, and a more reliable safety guarantee is provided for road logistics transportation.

Description

Technical Field

[0001] The present application relates to the field of risk warning technology, and more specifically, to a method and device for early warning of road icing disaster risks for highway logistics transportation. Background Art

[0002] With the development of economic globalization and the rise of e-commerce, road logistics and transportation are becoming increasingly important in the modern economy. However, meteorological disasters pose a major challenge to the safety and efficiency of road logistics and transportation, especially the frequent road icing in winter, which not only increases the risk of traffic accidents, but also may cause traffic interruptions, seriously affecting the timely delivery of goods.

[0003] Road icing refers to the phenomenon that the water on the road condenses into ice due to the temperature dropping below 0℃. This phenomenon usually occurs after precipitation, when the temperature drops sharply at night, or in the early morning. Road icing greatly reduces the friction between the tires and the road surface, making it difficult to control the vehicle and significantly increasing the braking distance, thus causing traffic accidents. Especially in mountainous areas, bridges and other sections of roads prone to water accumulation, icing is more common and dangerous.

[0004] At present, the research on road icing warning mainly focuses on warning methods based on energy radiation balance and data mining methods. The warning method based on energy radiation balance mainly analyzes the quantitative relationship between road surface temperature and meteorological elements such as air temperature, wind speed, relative humidity, and precipitation, and constructs a statistical model to predict whether the road surface will freeze. Since it mainly relies on road surface temperature to indirectly predict the icing situation, its prediction accuracy and timeliness are subject to certain limitations in practical applications. At the same time, the data mining method mainly establishes a warning model by performing correlation analysis and pattern mining on historical meteorological data and road icing events. However, most of the existing warning models based on data mining only consider a single meteorological element or a simple combination of elements, and lack in-depth analysis and utilization of the complex interactive relationship between multiple meteorological elements, resulting in limited warning effects and insufficient specificity and practicality of warning information.

[0005] Therefore, an optimized road icing disaster risk warning method and device for highway logistics transportation is expected. Summary of the invention

[0006] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a road icing disaster risk warning method and device for highway logistics transportation, which uses a deep learning-based data processing algorithm to conduct spatiotemporal interactive response analysis on the precipitation grid real-time data along the highway and the precipitation, temperature and relative humidity grid forecast data released by the meteorological department, so as to dig out the road icing characteristics caused by meteorological factors, and then on this basis, predict the cumulative icing time and determine the road icing blockage risk level, so as to realize the road icing disaster risk warning for highway logistics transportation. In this way, the accuracy and timeliness of road icing warnings can be significantly improved, providing more reliable safety guarantees for highway logistics transportation.

[0007] According to one aspect of the present application, a road icing disaster risk warning method for highway logistics transportation is provided, which includes: Obtain precipitation grid actual data, precipitation grid forecast data, temperature grid forecast data and relative humidity grid forecast data; Based on the precipitation grid point real-time data, extracting the temporal correlation characteristics of precipitation along the highway to obtain a temporal correlation characteristic coding vector of the precipitation before the forecast; Based on the precipitation grid forecast data, the temperature grid forecast data and the relative humidity grid forecast data, extracting the time series correlation features of meteorological factors along the highway to obtain a joint coding vector of the time series correlation features of the meteorological factor forecast situation; Performing a core feature-driven meteorological element actual-forecast nonlinear interactive response analysis on the pre-forecast precipitation condition temporal correlation feature coding vector and the meteorological factor forecast condition temporal correlation feature joint coding vector to obtain a road icing condition temporal-spatial response interactive coding vector; Based on the spatiotemporal response interactive coding vector of the road icing condition, a road icing blockage risk level is determined.

[0008] According to another aspect of the present application, a road icing disaster risk warning device for highway logistics transportation is provided, which includes: The actual and forecast data acquisition module is used to obtain the actual data of precipitation grid points, the forecast data of precipitation grid points, the forecast data of temperature grid points and the forecast data of relative humidity grid points; A real-time data feature extraction module, used to extract the temporal correlation features of precipitation conditions along the highway based on the real-time data of the precipitation grid points to obtain a temporal correlation feature encoding vector of the precipitation conditions before the forecast; A forecast data feature extraction module, used to extract the time series correlation features of meteorological factors along the highway based on the precipitation grid forecast data, the temperature grid forecast data and the relative humidity grid forecast data to obtain a joint coding vector of the time series correlation features of the meteorological factor forecast situation; An interactive response analysis module is used to perform a core feature-driven meteorological element actual-forecast nonlinear interactive response analysis on the pre-forecast precipitation situation time series correlation feature coding vector and the meteorological factor forecast situation time series correlation feature joint coding vector to obtain a road icing condition spatiotemporal response interactive coding vector; The risk level determination module is used to determine the road icing blockage risk level based on the road icing condition time-space response interactive coding vector.

[0009] Compared with the prior art, the road icing disaster risk warning method and device provided by this application for highway logistics transportation uses a data processing algorithm based on deep learning to perform spatiotemporal interactive response analysis on the actual precipitation grid data along the highway and the precipitation, temperature and relative humidity grid forecast data released by the meteorological department, so as to dig out the road icing characteristics caused by meteorological factors, and then on this basis, predict the cumulative icing time and determine the road icing blocking risk level, so as to realize the road icing disaster risk warning for highway logistics transportation. In this way, the accuracy and timeliness of road icing warning can be significantly improved, providing more reliable safety guarantee for highway logistics transportation. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0011] Figure 1 This is a flow chart of a road icing disaster risk warning method for highway logistics transportation according to an embodiment of the present application.

[0012] Figure 2 This is a data flow diagram of a road icing disaster risk warning method for highway logistics transportation according to an embodiment of the present application.

[0013] Figure 3 This is a flowchart of sub-step S3 of the road icing disaster risk warning method for highway logistics transportation according to an embodiment of the present application.

[0014] Figure 4 This is a flowchart of sub-step S4 of the road icing disaster risk warning method for highway logistics transportation according to an embodiment of the present application.

[0015] Figure 5 This is a flowchart of sub-step S41 of the road icing disaster risk warning method for highway logistics transportation according to an embodiment of the present application.

[0016] Figure 6 This is a flowchart of sub-step S44 of the road icing disaster risk warning method for highway logistics transportation according to an embodiment of the present application.

[0017] Figure 7 This is a flowchart of sub-step S5 of the road icing disaster risk warning method for highway logistics transportation according to an embodiment of the present application.

[0018] Figure 8 It is a block diagram of a road icing disaster risk warning device for highway logistics transportation according to an embodiment of the present application. DETAILED DESCRIPTION

[0019] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0020] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0021] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. Meanwhile, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0022] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0023] It is worth noting that in this application, all actions to obtain data are carried out in compliance with the relevant data protection laws and policies of the country where the data is located, and with the authorization given by the owner of the corresponding device.

[0024] In response to the technical problems described in the above background technology, this application proposes a road icing disaster risk warning method for highway logistics transportation, which uses a data processing algorithm based on deep learning to conduct a spatiotemporal interactive response analysis on the precipitation grid real-time data along the highway and the precipitation, temperature and relative humidity grid forecast data released by the meteorological department, so as to dig out the road icing characteristics caused by meteorological factors, and then on this basis, predict the cumulative icing time and determine the road icing blocking risk level, so as to realize the road icing disaster risk warning for highway logistics transportation. In this way, the accuracy and timeliness of road icing warning can be significantly improved, providing more reliable safety protection for highway logistics transportation.

[0025] Figure 1 This is a flow chart of a road icing disaster risk warning method for highway logistics transportation according to an embodiment of the present application. Figure 2 The following is a data flow diagram of the road icing disaster risk warning method for highway logistics transportation according to an embodiment of the present application. Figure 1 and Figure 2 As shown, the road icing disaster risk warning method for highway logistics transportation includes the following steps: S1, obtaining actual precipitation grid data, precipitation grid forecast data, temperature grid forecast data and relative humidity grid forecast data; S2, based on the actual precipitation grid data, extracting the temporal correlation characteristics of precipitation along the highway to obtain a pre-forecast precipitation temporal correlation characteristic coding vector; S3, based on the precipitation grid forecast data, the temperature grid forecast data and the relative humidity grid forecast data, extracting the temporal correlation characteristics of meteorological factors along the highway to obtain a meteorological factor forecast temporal correlation characteristic joint coding vector; S4, performing a core feature-driven meteorological element actual-forecast nonlinear interactive response analysis on the pre-forecast precipitation temporal correlation characteristic coding vector and the meteorological factor forecast temporal correlation characteristic joint coding vector to obtain a road icing condition spatiotemporal response interactive coding vector; S5, determining the road icing traffic blockage risk level based on the road icing condition spatiotemporal response interactive coding vector.

[0026] In an optional embodiment of the present application, a road icing disaster risk warning method for highway logistics transportation includes the following steps: 1. Acquiring transportation blockage data, actual data on road icing along the highway; actual data on precipitation, temperature, and relative humidity at traffic meteorological stations along the highway; and grid forecast data on precipitation, temperature, and relative humidity; 2. Extracting warning-related feature coding vectors affecting the duration, scope, and intensity of traffic blockage based on the transportation blockage data; 3. Extracting temporal correlation features of precipitation, temperature, and relative humidity along the highway based on actual data on road icing along the highway to obtain a joint coding vector of coupled temporal correlation features of multiple meteorological elements that cause icing-induced obstruction; 4. Introducing meteorological element grid forecast data into the traffic blockage impact-related feature coding vector and the coupled temporal correlation feature joint coding vector of multiple meteorological elements that cause icing-induced obstruction, and conducting a nonlinear interactive response analysis of the actual situation and forecast of road icing meteorological elements driven by core features to obtain a spatiotemporal response interactive coding vector of road icing conditions.

[0027] In the above-mentioned road icing disaster risk warning method for highway logistics transportation, the step S1 obtains precipitation grid actual data, precipitation grid forecast data, temperature grid forecast data and relative humidity grid forecast data. It should be understood that the road icing disaster risk warning is highly dependent on accurate meteorological data, among which meteorological elements such as precipitation, temperature and relative humidity are the core variables that determine whether the road is frozen and the severity of the ice. Specifically, the precipitation grid actual data can present the actual state of precipitation along the current highway, the precipitation grid forecast data can outline the future precipitation trend along the highway, and the temperature and relative humidity grid forecast data play a key role in judging whether precipitation can be converted into ice, the duration of ice, etc. Therefore, this application captures the complex relationship and dynamic changes between different meteorological elements by collecting the actual data of traffic blockage caused by road icing, as well as the actual and forecast data of multi-dimensional meteorological factors along the highway, so as to provide a more accurate prediction basis for road icing warning.

[0028] Specifically, meteorological data comes from a variety of sources, usually including public resources released by the National Meteorological Administration, data interfaces of professional meteorological service providers, and real-time observation data obtained through satellite remote sensing technology. The collection of meteorological data not only relies on fixed observation stations on the ground, but also relies on mobile observation equipment such as vehicle-mounted sensors and drones to more comprehensively cover the different geographical environments in the study area, thereby improving the accuracy and representativeness of the data. In order to establish a stable and reliable data transmission network and ensure that each observation point can upload observation results to the data center in real time, this requires the use of high-speed and stable Internet connections or dedicated communication lines, and the necessary hardware facilities such as servers and storage devices to ensure that data processing capabilities meet the needs. In addition, it is necessary to develop special data collection software or systems to automatically extract the required information from multiple data sources and organize and archive them in a unified format for subsequent use.

[0029] For the acquisition of real-time data of precipitation grid points, the focus is on ensuring the temporal continuity and spatial uniformity of the data. This means that we should not only pay attention to the changes in precipitation along the main traffic arteries, but also take into account the small-scale terrain features in the surrounding areas that may affect the local climate. Combined with the rainfall records with hourly or even minute-level resolutions provided by the ground automatic station network, supplemented by radar echo maps to capture short-term heavy precipitation events, a high-precision precipitation field model is constructed. The discrete observation point data is expanded to the entire study area through the interpolation algorithm to form a complete real-time database of precipitation grid points. This approach not only improves the spatial resolution of the data, but also makes the analysis more nuanced.

[0030] To obtain precipitation grid forecast data, it is necessary to rely on the output products of numerical weather forecast models. There are many mature numerical models available at home and abroad, such as the global forecast system of the European Center for Medium-Range Weather Forecasts (ECMWF) and the GFS model of the National Center for Environmental Prediction (NCEP). These models are based on atmospheric dynamics equations. Through large-scale parallel computing, they simulate the evolution of the global or specific regional atmospheric state in the next few days or even dozens of days. Precipitation forecast field information related to the target highway can be extracted from them, and the spatial resolution and time step can be adjusted as needed to make it more suitable for local application needs. Such forecasts not only take into account the overall trend of the atmosphere, but also pay attention to local details, which improves the accuracy of the forecast.

[0031] The temperature grid forecast data and relative humidity grid forecast data are also derived from the above numerical model. As one of the key factors that determine whether the road surface will freeze, the forecast accuracy of temperature directly affects the quality of the final warning effect. Therefore, when obtaining such data, special attention should be paid to the performance evaluation of the model itself and the optimization of parameter settings. At the same time, considering the interaction between relative humidity and temperature, the two often need to be considered simultaneously and participate in the subsequent risk level classification process. Through statistical analysis of the differences between historical observation data and model forecast results, the forecast products are continuously revised and improved to make them closer to the actual situation. This method emphasizes the importance of data analysis and also demonstrates the application value of scientific methods in solving practical problems. In order to further improve the quality of data, it is also necessary to regularly calibrate and maintain the observation instruments used to ensure the reliability of all measurement results. By integrating multi-source data and using advanced technical means for refined processing, a solid foundation can be laid for subsequent analysis.

[0032] In the above-mentioned road icing disaster risk warning method for highway logistics transportation, the step S2 extracts the temporal correlation characteristics of the precipitation along the highway based on the actual data of the precipitation grid points to obtain the temporal correlation feature encoding vector of the precipitation before the forecast. It should be understood that since the formation of road icing is often associated with the previous precipitation conditions, and the road icing is not formed instantly, the precipitation changes and precipitation period distribution in the previous precipitation process have a significant impact on the starting time of icing, ice thickness accumulation, etc. Therefore, the present application further captures the temporal correlation change pattern of the previous precipitation along the highway by performing a temporal analysis on the actual data of the precipitation grid points. In a specific example of the present application, a temporal encoder based on an LSTM model is used to perform time series encoding on the actual data of the precipitation grid points to obtain the temporal correlation feature encoding vector of the precipitation before the forecast. Specifically, first, the actual data of the precipitation grid points are sorted into a time series in chronological order, and each time step contains the precipitation information at different locations along the highway. Then, the LSTM model is used to perform time series encoding on the sorted time series data. The loop structure and gating mechanism of the LSTM model are used to capture the long-term dependencies in the time series, thereby extracting the dynamic correlation change characteristics of precipitation conditions along the highway over time, and obtaining the time series correlation feature encoding vector of precipitation conditions before forecast, thereby providing strong data support for subsequent road icing warnings.

[0033] In the above-mentioned road icing disaster risk warning method for highway logistics transportation, the step S3 extracts the temporal correlation features of meteorological factors along the highway based on the precipitation grid forecast data, the temperature grid forecast data and the relative humidity grid forecast data to obtain a joint coding vector of the temporal correlation features of the meteorological factor forecast situation. Figure 3 FIG. 4 is a flowchart of sub-step S3 of the road icing disaster risk warning method for highway logistics transportation according to an embodiment of the present application. Figure 3 As shown, the step S3 includes the steps of: S31, using the time series encoder based on the LSTM model to perform time series encoding on the precipitation grid forecast data, the temperature grid forecast data and the relative humidity grid forecast data respectively to obtain a precipitation forecast situation time series correlation feature encoding vector, a temperature forecast situation time series correlation feature encoding vector and a relative humidity forecast situation time series correlation feature encoding vector; S32, fusing the precipitation forecast situation time series correlation feature encoding vector, the temperature forecast situation time series correlation feature encoding vector and the relative humidity forecast situation time series correlation feature encoding vector to obtain the meteorological factor forecast situation time series correlation feature joint encoding vector.

[0034] Specifically, in step S31, the precipitation grid forecast data, the temperature grid forecast data and the relative humidity grid forecast data are respectively time-series encoded using the time-series encoder based on the LSTM model to obtain a precipitation forecast situation time-series associated feature encoding vector, a temperature forecast situation time-series associated feature encoding vector and a relative humidity forecast situation time-series associated feature encoding vector. Specifically, since the changes in precipitation, temperature and relative humidity in the future are key meteorological factors affecting the road icing state, the present application also uses the LSTM model to perform time-series modeling analysis on the precipitation grid forecast data, the temperature grid forecast data and the relative humidity grid forecast data, respectively, so as to utilize the long-term memory capacity of the LSTM model to capture the time-series change trend of precipitation, temperature and relative humidity along the highway in the future, thereby obtaining a precipitation forecast situation time-series associated feature encoding vector, a temperature forecast situation time-series associated feature encoding vector and a relative humidity forecast situation time-series associated feature encoding vector.

[0035] Specifically, the step S32 is to fuse the precipitation forecast situation time series correlation feature coding vector, the temperature forecast situation time series correlation feature coding vector and the relative humidity forecast situation time series correlation feature coding vector to obtain the meteorological factor forecast situation time series correlation feature joint coding vector. It should be understood that the present application takes into account that the duration of road icing is the result of the synergistic effect of multiple meteorological factors, and a single meteorological factor cannot comprehensively predict the duration of road icing. For example, precipitation provides the material basis for icing, the temperature determines whether precipitation can condense into ice, and the relative humidity affects the speed and stability of ice formation. Therefore, the present application further fuses the precipitation forecast situation time series correlation feature coding vector, the temperature forecast situation time series correlation feature coding vector and the relative humidity forecast situation time series correlation feature coding vector to consider the relationship between the various meteorological factors as a whole, and form a meteorological factor forecast situation time series correlation feature joint coding vector, so as to more accurately describe the comprehensive impact of meteorological conditions on the road icing state.

[0036] In the above-mentioned road icing disaster risk warning method for highway logistics transportation, the step S4 is to perform a nonlinear interactive response analysis of the actual-forecast meteorological elements based on the core feature driven coding vector of the time series correlation feature coding vector of the precipitation before the forecast and the time series correlation feature joint coding vector of the meteorological factor forecast to obtain the time-space response interactive coding vector of the road icing condition. It should be understood that the time series correlation feature coding vector of the precipitation before the forecast describes the time series correlation change of the previous precipitation along the highway, while the time series correlation feature joint coding vector of the meteorological factor forecast comprehensively reflects the time series change trend of meteorological factors such as precipitation, temperature and relative humidity in the future. In this application, by mining the time series core features of the time series correlation feature coding vector of the precipitation before the forecast and the time series correlation feature joint coding vector of the meteorological factor forecast, the nonlinear interactive response analysis of the previous precipitation and the future meteorological factor change trend is carried out, which helps to capture the correlation response relationship of different meteorological elements at different time and space scales, and more deeply understand the formation mechanism and development process of road icing, so as to provide a more effective prediction basis for road icing warning.

[0037] Figure 4 FIG. 4 is a flowchart of sub-step S4 of the road icing disaster risk warning method for highway logistics transportation according to an embodiment of the present application. Figure 4As shown, the step S4 includes the steps of: S41, respectively extracting the core feature information of the time series correlation feature coding vector of the precipitation before forecast and the time series correlation feature joint coding vector of the meteorological factor forecast to obtain the time series core information anchor coding vector of the precipitation before forecast and the time series core information anchor coding vector of the meteorological factor forecast; S42, performing feature granularity response interactive coding on the time series core information anchor coding vector of the precipitation before forecast and the time series core information anchor coding vector of the meteorological factor forecast to obtain the meteorological element actual situation- Forecast time series feature granularity response interactive coding vector; S43, feature value granularity response interactive coding of the pre-forecast precipitation time series core information anchor coding vector and the meteorological factor forecast time series core information anchor coding vector to obtain the meteorological element actual situation-forecast time series feature value granularity response interactive coding vector; S44, fusion of the meteorological element actual situation-forecast time series feature granularity response interactive coding vector and the meteorological element actual situation-forecast time series feature value granularity response interactive coding vector to obtain the road icing condition spatiotemporal response interactive coding vector.

[0038] Figure 5 FIG. 4 is a flowchart of sub-step S41 of the road icing disaster risk warning method for highway logistics transportation according to an embodiment of the present application. Figure 5 As shown, the step S41 includes the steps of: S411, constructing a semantic autocorrelation association matrix of the temporal correlation feature coding vector of the precipitation situation before forecast and the temporal correlation feature joint coding vector of the meteorological factor forecast situation to obtain the temporal autocorrelation association matrix of the precipitation situation before forecast and the temporal autocorrelation association matrix of the meteorological factor forecast situation; S412, performing core information anchor coding based on autocorrelation decoupling on the temporal autocorrelation association matrix of the precipitation situation before forecast and the temporal autocorrelation association matrix of the meteorological factor forecast situation respectively to obtain the temporal core information anchor coding vector of the precipitation situation before forecast and the temporal core information anchor coding vector of the meteorological factor forecast situation.

[0039] More specifically, the step S411 is expressed as follows:

[0040]

[0041] in, Represents the temporal correlation feature coding vector of the precipitation situation before the forecast, Represents the joint coding vector of the temporal correlation features of the meteorological factor forecast, represents the transpose of a vector, represents the time series autocorrelation matrix of the precipitation conditions before the forecast, represents the time series autocorrelation matrix of the meteorological factor forecast, is a linear mapping function.

[0042] That is, the present application reveals the internal correlation structure of the respective coding vectors of the pre-forecast precipitation situation and the meteorological factor forecast situation by constructing the temporal autocorrelation association matrix of the temporal correlation feature coding vector of the pre-forecast precipitation situation and the temporal correlation feature joint coding vector of the meteorological factor forecast situation. This correlation structure reflects the inherent logic and mutual relationship of meteorological elements in the time series, which helps to deeply understand the essential characteristics of these meteorological information. In a specific implementation, the mutual correlation between the component feature parts in the temporal correlation feature coding vector of the pre-forecast precipitation situation and the temporal correlation feature joint coding vector of the meteorological factor forecast situation can be made explicit through the outer product operation, and the temporal autocorrelation association matrix of the pre-forecast precipitation situation and the temporal autocorrelation association matrix of the meteorological factor forecast situation are generated, thereby retaining the global temporal structure of meteorological elements while capturing their internal semantic association patterns, providing clearer structural information for the subsequent extraction of core feature information.

[0043] More specifically, the step S412 is expressed as follows:

[0044]

[0045] in, Indicates the core information anchoring network, represents the decoupling function, , , and They represent the first, second, and third and row vectors, is the number of rows of the time series autocorrelation matrix of the precipitation conditions before the forecast, , , and They represent the first, second, and third and row vectors, and They represent the time series weight matrix of precipitation before forecast and the time series weight matrix of meteorological factor forecast respectively. and They represent the time series bias vector of precipitation before forecast and the time series bias vector of meteorological factor forecast respectively. represents matrix multiplication, express The corresponding pre-forecast precipitation time series importance scoring factor, express The corresponding meteorological factor forecast time series importance scoring factor, Represents the importance score conversion vector of the local time series feature of the precipitation situation before the forecast, Represents the local time series feature importance score conversion vector of the meteorological factor forecast situation, express The corresponding normalized pre-forecast precipitation time series importance score factor, express The corresponding normalized meteorological factor forecast time series importance score factor, represents the sigmoid function, Represents the anchor coding vector of the core information of the precipitation situation before the forecast, Represents the anchor coding vector of the time series core information of the meteorological factor forecast.

[0046] That is, by performing feature decoupling, local feature importance evaluation and feature weighted aggregation on the ambient temperature time series feature semantic autocorrelation association matrix and the ambient humidity time series feature semantic autocorrelation association matrix, the most critical and representative core feature information is extracted and redundant information is removed, thereby obtaining more refined and targeted coding vectors, namely the anchor coding vector of the core information of the time series of precipitation before the forecast and the anchor coding vector of the core information of the time series of the meteorological factor forecast, which helps to focus on the core features of the meteorological elements in the subsequent data analysis, reduces the interference of redundant information, and improves the efficiency and accuracy of the subsequent analysis.

[0047] Specifically, the step S42 is expressed by the formula:

[0048] in, represents the hyperbolic tangent function, represents the meteorological element actual-forecast time series interaction feature weight matrix, represents the bias vector of the interaction feature between the actual and forecast time series of meteorological elements, Represents the interactive coding vector of the granular response of the actual-forecast time series characteristics of meteorological elements.

[0049] That is, by further performing feature correlation interaction analysis on the anchor coding vector of the core information of the time series of precipitation before the forecast and the anchor coding vector of the time series of the forecast of meteorological factors, the correlation response relationship of meteorological factors in time and space can be captured, so as to have a deeper understanding of the impact mechanism of meteorological factors on road icing. Here, a neural network is first used to analyze the interaction between the anchor coding vector of the time series of the core information of the time series of precipitation before the forecast and the anchor coding vector of the time series of the core information of the forecast of meteorological factors from the feature level to capture the temporal correlation and mutual influence between the characteristics of different meteorological elements, reveal the synergistic influence of potential meteorological factor combinations on road icing, and then generate the interactive coding vector of the granular response of the actual-forecast time series characteristics of meteorological elements.

[0050] Specifically, the step S43 is expressed by the formula:

[0051] in, Represents the interactive coding vector of the granular response of the time series characteristic values ​​of the actual and forecast meteorological elements.

[0052] Here, in order to have a deeper understanding of the relationship between the actual and forecast meteorological elements and their impact on road icing, this application further analyzes the interaction between the anchor coding vector of the core information of the time series of precipitation before the forecast and the anchor coding vector of the time series of the forecast of meteorological factors from the eigenvalue level. Unlike the feature granularity analysis, the eigenvalue granularity analysis returns to the lower-level numerical operation, and characterizes the temporal interdependence between the two by performing element-by-element interactive response coding on the anchor coding vector of the core information of the time series of precipitation before the forecast and the anchor coding vector of the time series of the forecast of meteorological factors. The generated meteorological element actual-forecast time series eigenvalue granularity response interactive coding vector can more finely reflect the dynamic response relationship between the actual situation of meteorological elements before the forecast and the forecast situation and its potential impact on road icing.

[0053] Figure 6 FIG. 4 is a flowchart of sub-step S44 of the road icing disaster risk warning method for highway logistics transportation according to an embodiment of the present application. Figure 6As shown, the step S44 includes the steps of: S441, performing feature distribution gradient constraint correction based on bidirectional interaction on the meteorological element actual-forecast time series feature granularity response interactive coding vector and the meteorological element actual-forecast time series feature value granularity response interactive coding vector to obtain an optimized meteorological element actual-forecast time series feature granularity response interactive coding vector and an optimized meteorological element actual-forecast time series feature value granularity response interactive coding vector; S442, performing cascade fusion on the optimized meteorological element actual-forecast time series feature granularity response interactive coding vector and the optimized meteorological element actual-forecast time series feature value granularity response interactive coding vector to obtain the road icing condition spatiotemporal response interactive coding vector.

[0054] In particular, here, the present application considers that the difference between the modeling representation of feature granularity interaction and feature value granularity interaction may cause unstable perturbations in the feature manifold interface of the fused road icing condition spatiotemporal response interaction encoding vector with multi-granularity information expression. Therefore, in a preferred example of the present application, the step S441 is expressed by the formula:

[0055] in, yes Middle The eigenvalues ​​at the positions, yes Middle The eigenvalues ​​at the positions, represents the cosine function, yes The corresponding optimized eigenvalue is yes The corresponding optimized eigenvalue.

[0056] That is, based on the correction of the characteristic distribution gradient constraint of the two-way interaction, the interactive coding vector of the characteristic granularity response of the meteorological element actual-forecast time series and the interactive coding vector of the characteristic value granularity response of the meteorological element actual-forecast time series are optimized. The interactive coding vector of the characteristic value granularity response of the meteorological element actual-forecast time series is expressed as an extensibility constraint, and the interface shape perturbation deviation of the overall distribution of the characteristic granularity interaction is modeled based on the growth exponent expression under the extensibility constraint for the overall characteristic interaction distribution growth under the eigenvalue diffusion process. That is, the interaction interface gradient under the eigenvalue granularity is used as the perturbation contribution factor, and the growth exponent stabilization contribution of the gradient diffusion under the mutual dependence of the eigenvalue is determined to achieve the growth mode gradient correction dominated by the perturbation stabilization, and improve the manifold interface stability of the spatial-temporal response interactive coding vector of the road icing condition.

[0057] More specifically, the step S442 is expressed as follows:

[0058] in, Indicates cascade, is the optimized meteorological element actual-forecast time series feature granularity response interactive coding vector, is the optimized interactive coding vector of the granularity response of the meteorological element actual-forecast time series characteristic value, Represents the interactive encoding vector of spatiotemporal responses to road icy conditions.

[0059] That is, in order to integrate the interactive information of meteorological elements at the macro and micro levels and obtain a coding vector that can comprehensively and accurately reflect the spatiotemporal response of road icing conditions, the present application cascades and fuses the optimized meteorological element actual-forecast time series feature granularity response interactive coding vector and the meteorological element actual-forecast time series feature value granularity response interactive coding vector, so as to fully utilize the advantages of different granularity coding vectors, more comprehensively describe the impact of meteorological elements on road icing, and enable the generated road icing condition spatiotemporal response interactive coding vector to more accurately reflect the spatiotemporal changes of road icing conditions.

[0060] In the above-mentioned road icing disaster risk warning method for highway logistics transportation, the step S5 determines the road icing blocking risk level based on the road icing condition spatiotemporal response interactive coding vector. Figure 7 FIG. 5 is a flowchart of sub-step S5 of the road icing disaster risk warning method for highway logistics transportation according to an embodiment of the present application. Figure 7 As shown, the step S5 includes the steps of: S51, inputting the spatiotemporal response interactive coding vector of the road icing condition into the trained road icing determination module to obtain a predicted value of the cumulative icing duration; S52, determining the road icing blocking risk level based on the predicted value of the cumulative icing duration.

[0061] Specifically, in step S51, the road icing condition spatiotemporal response interactive coding vector is input into the trained road icing determination module to obtain the predicted value of the cumulative icing duration. Specifically, the road icing determination module is constructed based on a neural network, and according to the conditional characteristics of the actual data of traffic blockage caused by road icing, after training and optimization with a large amount of sample data, the associated mapping relationship between the meteorological condition characteristics of road icing and the cumulative icing duration is effectively learned. After receiving the road icing condition spatiotemporal response interactive coding vector as input, the road icing determination module deeply understands the spatiotemporal changes of road icing conditions by comprehensively considering the change trends of meteorological factors such as the previous precipitation of the highway, future precipitation, temperature and relative humidity contained in the vector, and uses the learned mapping relationship to predict the cumulative icing duration along the highway, thereby providing key decision support information for whether highway logistics transportation is blocked.

[0062] Specifically, the step S52 determines the road icing blockage risk level based on the predicted value of the cumulative icing duration. It should be understood that different degrees of road icing will have different impacts on traffic. For example, slight, short-term icing may cause vehicles to travel slowly, while severe icing may cause the road to be completely blocked for a long time. Therefore, the present application further divides the risk level according to the predicted cumulative icing duration, so that the transportation department can quickly determine the severity of the road conditions and take targeted measures in a timely manner, such as closing roads, speed limits, and strengthening patrols, so as to minimize the impact of road icing on traffic safety and protect the lives and property of the public. In a specific example of the present application, the step S52 includes: in response to the predicted value of the cumulative icing duration being in the range of 1-6 hours, the road icing blockage risk level is "Level I, blue, the risk of road blockage caused by road icing is medium, which will affect the traffic efficiency"; in response to the predicted value of the cumulative icing duration being in the range of 7-12 hours, the road icing blockage risk level is "Level II, yellow, the risk of road blockage caused by road icing is high, which greatly affects the traffic efficiency"; in response to the predicted value of the cumulative icing duration being in the range of 13-18 hours, the road icing blockage risk level is "Level III, orange, the risk of road blockage caused by road icing is high, which greatly affects the traffic efficiency"; in response to the predicted value of the cumulative icing duration being in the range of 19-24 hours, the road icing blockage risk level is "Level IV, red, the risk of road blockage caused by road icing is extremely high, which greatly affects the traffic efficiency".

[0063] Specifically, the predicted value of the cumulative icing duration is an important indicator for assessing the road icing situation and its impact. Converting the cumulative icing duration into a specific traffic blockage risk level is conducive to the traffic management department taking corresponding measures to ensure traffic safety. When the predicted value of the cumulative icing duration falls within the range of 1-6 hours, it means that the risk of road blockage caused by road icing is at a medium level, which is defined as Level I, blue warning. In this case, although the traffic efficiency will be affected, the overall risk is controllable. At this time, relevant departments need to be vigilant, prepare necessary emergency supplies, and strengthen monitoring efforts to ensure that they can respond quickly in the event of an emergency.

[0064] When the predicted value of the cumulative icing duration is in the range of 7-12 hours, it indicates that the risk of road blockage caused by road icing is high, and this situation is marked as Level II, yellow warning. Compared with Level I warning, the road traffic efficiency is more affected under Level II warning, and stricter control measures are required. To this end, the traffic management department should consider implementing measures such as speed limits or temporary closures of some sections of the road, while increasing the frequency of road inspections, clearing snow and ice in a timely manner, and reducing the probability of traffic accidents. In addition, it is necessary to release warning information to the public through various channels to remind drivers to pay attention to driving safety and avoid unnecessary travel.

[0065] If the predicted value of the cumulative icing duration is within the range of 13-18 hours, it means that the risk of road blockage caused by road icing has reached a high risk level, corresponding to Level III, orange warning. In this case, the road traffic efficiency is greatly affected, which may cause large-scale traffic congestion and even accidents. Therefore, more decisive measures must be taken, such as fully closing dangerous sections of the road and forcing vehicles to detour. At the same time, it is necessary to activate emergency plans and organize forces to carry out de-icing operations to ensure that traffic order is restored as soon as possible. For residents and businesses in the affected areas, it is particularly important to make preparations for material reserves and personnel evacuation in advance.

[0066] When the predicted cumulative icing duration exceeds 19 hours and reaches 24 hours, it indicates that the risk of road blockage caused by road icing is extremely high, that is, Level IV, red warning. This is the most serious risk level, indicating that road traffic is almost completely paralyzed, which not only seriously affects daily travel, but may also lead to a series of chain reactions such as supply chain disruptions. In the face of such a severe situation, in addition to the immediate implementation of all the aforementioned measures, it is also necessary to coordinate resources from all parties.

[0067] Furthermore, when determining the risk level of road icing, it is necessary to comprehensively consider the local geographical environment characteristics and historical cases. For example, due to the complex terrain and drastic temperature changes on mountainous roads, it is more likely to form localized severe icing; on the main roads within the city, due to the large traffic volume, even slight icing may cause major traffic accidents. Therefore, differentiated warning standards and response plans can be formulated for different types of roads to achieve local conditions and improve the effectiveness and adaptability of the entire warning system.

[0068] In addition, in order to ensure that warning information can be quickly and accurately transmitted to every relevant person, it is essential to establish an efficient information dissemination mechanism. Using modern information technology, such as mobile Internet, social media platforms, and radio stations, to release warning information through multiple channels can ensure that drivers are aware of the latest road conditions in a timely manner and plan their trips reasonably. At the same time, carrying out traffic safety publicity and education activities to enhance the public's safety awareness and self-protection ability is also one of the effective ways to reduce the rate of road icing accidents.

[0069] In summary, the road icing disaster risk warning method for highway logistics transportation based on the embodiment of the present application is explained, which uses a data processing algorithm based on deep learning to perform spatiotemporal interactive response analysis on the precipitation grid real-time data along the highway and the precipitation, temperature and relative humidity grid forecast data released by the meteorological department, so as to dig out the road icing characteristics caused by meteorological factors, and then on this basis, predict the cumulative icing time and determine the road icing blocking risk level, so as to realize the road icing disaster risk warning for highway logistics transportation. In this way, the accuracy and timeliness of road icing warning can be significantly improved, providing more reliable safety protection for highway logistics transportation.

[0070] Furthermore, a road icing disaster risk warning device for highway logistics transportation is also provided.

[0071] Figure 8 FIG. 1 is a block diagram of a road icing disaster risk warning device for highway logistics transportation according to an embodiment of the present application. Figure 8As shown, according to the embodiment of the present application, the road icing disaster risk warning device 100 for highway logistics transportation includes: a real-time and forecast data acquisition module 110, which is used to acquire the real-time data of precipitation grid points, the forecast data of precipitation grid points, the forecast data of temperature grid points and the forecast data of relative humidity grid points; a real-time data feature extraction module 120, which is used to extract the temporal correlation features of precipitation conditions along the highway based on the real-time data of precipitation grid points to obtain the temporal correlation feature encoding vector of precipitation conditions before forecast; a forecast data feature extraction module 130, which is used to extract the temporal correlation feature encoding vector of precipitation conditions before forecast based on the forecast data of precipitation grid points, the forecast data of temperature grid points and the forecast data of relative humidity grid points; Relative humidity grid forecast data, extracting the temporal correlation characteristics of meteorological factors along the highway to obtain a joint coding vector of the temporal correlation characteristics of the meteorological factor forecast situation; an interactive response analysis module 140, used to perform a core feature-driven meteorological element actual-forecast nonlinear interactive response analysis on the temporal correlation characteristic coding vector of the pre-forecast precipitation situation and the temporal correlation characteristic joint coding vector of the meteorological factor forecast situation to obtain a spatiotemporal response interactive coding vector of road icing conditions; a risk level determination module 150, used to determine the road icing blockage risk level based on the spatiotemporal response interactive coding vector of the road icing conditions.

[0072] Here, those skilled in the art can understand that the specific operations of each module in the above-mentioned road icing disaster risk warning device for highway logistics transportation have been described in the above reference. Figures 1 to 7 The road icing disaster risk warning method for road logistics transportation has been introduced in detail, and therefore, its repeated description will be omitted.

[0073] The basic principle of the present invention is described above in conjunction with specific embodiments. However, it should be pointed out that the advantages, strengths, effects, etc. mentioned in the present invention are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. must be possessed by each embodiment of the present invention. In addition, the specific details of the above embodiments are only for the purpose of illustration and facilitation of understanding, rather than limitation, and the above details do not limit the present invention to being implemented by adopting the above specific details.

[0074] In the above embodiments, the description of each embodiment has its own emphasis. For the parts that are not described or recorded in detail in a certain embodiment, please refer to the relevant description of other embodiments. In the several embodiments provided by the present invention, it should be understood that the disclosed system and method can be implemented in other ways. For example, the system embodiment described above is only schematic. For example, the unit division is only a logical function division, and there may be other division methods in actual implementation. The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the scheme of this embodiment.

[0075] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference to a figure in a claim should not be considered as limiting the claim to which it relates.

[0076] In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. The multiple units stated in the system claims can also be implemented by one unit through software or hardware.

[0077] Finally, it should be noted that the above description has been given for the purpose of illustration and description. In addition, the above embodiments are only used to illustrate the technical solution of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A road icing disaster risk warning method for highway logistics transportation, characterized in that: include: Obtain precipitation grid actual data, precipitation grid forecast data, temperature grid forecast data and relative humidity grid forecast data; Based on the precipitation grid point real-time data, extracting the temporal correlation characteristics of precipitation along the highway to obtain a temporal correlation characteristic coding vector of the precipitation before the forecast; Based on the precipitation grid forecast data, the temperature grid forecast data and the relative humidity grid forecast data, extracting the time series correlation features of meteorological factors along the highway to obtain a joint coding vector of the time series correlation features of the meteorological factor forecast situation; Performing a core feature-driven meteorological element actual-forecast nonlinear interactive response analysis on the pre-forecast precipitation condition temporal correlation feature coding vector and the meteorological factor forecast condition temporal correlation feature joint coding vector to obtain a road icing condition temporal-spatial response interactive coding vector; Based on the spatiotemporal response interactive coding vector of the road icing condition, a road icing blockage risk level is determined.

2. The road icing disaster risk warning method for highway logistics transportation according to claim 1 is characterized in that: Based on the actual precipitation grid point data, the temporal correlation features of the precipitation along the highway are extracted to obtain a temporal correlation feature coding vector of the precipitation before the forecast, including: The precipitation grid point actual data is time-series encoded using a time series encoder based on the LSTM model to obtain a time series associated feature encoding vector of the precipitation situation before the forecast.

3. The road icing disaster risk warning method for highway logistics transportation according to claim 2 is characterized in that: Based on the precipitation grid forecast data, the temperature grid forecast data and the relative humidity grid forecast data, the temporal correlation features of meteorological factors along the highway are extracted to obtain a joint coding vector of the temporal correlation features of the meteorological factor forecast situation, including: Using the LSTM model-based time series encoder to perform time series encoding on the precipitation grid forecast data, the temperature grid forecast data, and the relative humidity grid forecast data, respectively, to obtain a precipitation forecast situation time series associated feature encoding vector, a temperature forecast situation time series associated feature encoding vector, and a relative humidity forecast situation time series associated feature encoding vector; The precipitation forecast situation time series correlation feature coding vector, the temperature forecast situation time series correlation feature coding vector and the relative humidity forecast situation time series correlation feature coding vector are integrated to obtain the meteorological factor forecast situation time series correlation feature joint coding vector.

4. The road icing disaster risk warning method for highway logistics transportation according to claim 3 is characterized in that: The time series correlation feature coding vector of the precipitation situation before the forecast and the time series correlation feature joint coding vector of the meteorological factor forecast situation are subjected to the nonlinear interactive response analysis of the actual situation of the meteorological element-forecast situation driven by the core feature to obtain the time-space response interactive coding vector of the road icing condition, including: Respectively extract the core feature information of the time series correlation feature coding vector of the precipitation before forecast and the time series correlation feature joint coding vector of the meteorological factor forecast to obtain the time series core information anchor coding vector of the precipitation before forecast and the time series core information anchor coding vector of the meteorological factor forecast; Performing feature granularity response interactive coding on the core information anchoring coding vector of the precipitation situation before forecast and the core information anchoring coding vector of the meteorological factor forecast situation to obtain the feature granularity response interactive coding vector of the meteorological element actual-forecast time series; Performing characteristic value granularity response interactive coding on the core information anchoring coding vector of the time series of the precipitation situation before the forecast and the core information anchoring coding vector of the time series of the meteorological factor forecast situation to obtain the characteristic value granularity response interactive coding vector of the meteorological element actual situation-forecast time series; The meteorological element actual-forecast time series characteristic granularity response interactive coding vector and the meteorological element actual-forecast time series characteristic value granularity response interactive coding vector are fused to obtain the road icing condition spatiotemporal response interactive coding vector.

5. The road icing disaster risk warning method for highway logistics transportation according to claim 4 is characterized in that: The core feature information of the time series correlation feature coding vector of the precipitation before forecast and the time series correlation feature joint coding vector of the meteorological factor forecast are respectively extracted to obtain the time series core information anchor coding vector of the precipitation before forecast and the time series core information anchor coding vector of the meteorological factor forecast, including: Constructing a semantic autocorrelation association matrix of the time series correlation feature coding vector of the precipitation situation before the forecast and the time series correlation feature joint coding vector of the meteorological factor forecast situation to obtain a time series autocorrelation association matrix of the precipitation situation before the forecast and a time series autocorrelation association matrix of the meteorological factor forecast situation; The pre-forecast precipitation situation time series autocorrelation association matrix and the meteorological factor forecast situation time series autocorrelation association matrix are respectively subjected to core information anchor coding based on autocorrelation decoupling to obtain the pre-forecast precipitation situation time series core information anchor coding vector and the meteorological factor forecast situation time series core information anchor coding vector.

6. The road icing disaster risk warning method for highway logistics transportation according to claim 5 is characterized in that: The method of fusing the meteorological element actual-forecast time series characteristic granularity response interactive coding vector and the meteorological element actual-forecast time series characteristic value granularity response interactive coding vector to obtain the road icing condition spatiotemporal response interactive coding vector includes: Performing a characteristic distribution gradient constraint correction based on bidirectional interaction on the meteorological element actual-forecast time series characteristic granularity response interactive coding vector and the meteorological element actual-forecast time series characteristic value granularity response interactive coding vector to obtain an optimized meteorological element actual-forecast time series characteristic granularity response interactive coding vector and an optimized meteorological element actual-forecast time series characteristic value granularity response interactive coding vector; The optimized meteorological element actual-forecast time series feature granularity response interactive coding vector and the optimized meteorological element actual-forecast time series feature value granularity response interactive coding vector are cascaded and fused to obtain the road icing condition spatiotemporal response interactive coding vector.

7. The road icing disaster risk warning method for highway logistics transportation according to claim 6 is characterized in that: Determining the road icing blockage risk level based on the road icing condition spatiotemporal response interactive coding vector includes: Inputting the road icing condition spatiotemporal response interactive coding vector into a trained road icing determination module to obtain a predicted value of the cumulative icing duration; The road icing blockage risk level is determined based on the predicted value of the accumulated icing duration.

8. The road icing disaster risk warning method for highway logistics transportation according to claim 7 is characterized in that: Determining the road icing blockage risk level based on the predicted value of the accumulated icing duration includes: In response to the predicted value of the cumulative icing duration being within the range of 1-6 hours, the road icing blockage risk level is "Level I, blue, the risk of road blockage caused by road icing is medium, which will affect traffic efficiency"; In response to the predicted value of the cumulative icing duration being within the range of 7-12 hours, the road icing blockage risk level is "Level II, yellow, the risk of road blockage caused by road icing is high, which greatly affects traffic efficiency"; In response to the predicted value of the cumulative icing duration being in the range of 13-18 hours, the road icing blockage risk level is "Level III, orange, the risk of road blockage caused by road icing is high, which greatly affects traffic efficiency"; In response to the predicted value of the cumulative icing duration being in the range of 19-24 hours, the road icing blockage risk level is "Level IV, red, the risk of road blockage caused by road icing is extremely high, which greatly affects traffic efficiency".

9. A road icing disaster risk warning device for highway logistics transportation, characterized in that: include: The actual and forecast data acquisition module is used to obtain the actual data of precipitation grid points, the forecast data of precipitation grid points, the forecast data of temperature grid points and the forecast data of relative humidity grid points; A real-time data feature extraction module, used to extract the temporal correlation features of precipitation conditions along the highway based on the real-time data of the precipitation grid points to obtain a temporal correlation feature encoding vector of the precipitation conditions before the forecast; A forecast data feature extraction module, used to extract the time series correlation features of meteorological factors along the highway based on the precipitation grid forecast data, the temperature grid forecast data and the relative humidity grid forecast data to obtain a joint coding vector of the time series correlation features of the meteorological factor forecast situation; An interactive response analysis module is used to perform a core feature-driven meteorological element actual-forecast nonlinear interactive response analysis on the pre-forecast precipitation situation time series correlation feature coding vector and the meteorological factor forecast situation time series correlation feature joint coding vector to obtain a road icing condition spatiotemporal response interactive coding vector; The risk level determination module is used to determine the road icing blockage risk level based on the road icing condition time-space response interactive coding vector.

Images