A highway pavement ice condensation monitoring system

Abstract

The present invention discloses a highway pavement icing monitoring system, which relates to the field of highway traffic safety monitoring technology. The system comprises a data acquisition module that collects environmental data and road icing status information on the highway pavement; and a self-calibration module connected to the data acquisition module that self-calibrates the data acquisition module based on the real-time collected environmental data and road icing status information, correcting sensor errors caused by external environmental factors. By adding a self-calibration function to the monitoring equipment and introducing an intelligent algorithm, the highway pavement icing monitoring system can automatically correct measurement errors caused by external environmental changes in real time. Furthermore, it combines machine learning models to more accurately predict road conditions. This not only improves the system's operational stability in inclement weather but also enables earlier detection of road icing risks, providing more accurate early warning information to traffic management departments and reducing the probability of traffic accidents.

Description

Technical Field

[0001] The present invention relates to the technical field of highway traffic safety monitoring, and in particular to a highway pavement ice condensation monitoring system. Background Art

[0002] Based on existing technological developments, many highway pavement ice monitoring systems have adopted a variety of sensors and technical means to detect pavement icing conditions, including temperature and humidity sensors, infrared sensors, weather stations, and other data acquisition devices. These systems typically provide icing risk warnings by real-time monitoring of pavement temperature, humidity, and meteorological data. However, existing technical solutions still have certain limitations. Although many sensors can detect basic information on pavement temperature and humidity, in severe weather such as blizzards, strong winds, or low temperatures, sensors will be affected by ice and snow accumulation, moisture, or ice, resulting in sensor failure or inaccurate data. To address this issue, some technical solutions have adopted waterproof and anti-freeze designs, or heating devices to ensure the normal operation of sensors under low-temperature conditions, but these solutions generally do not address the long-term reliability issues that may arise in extreme weather conditions. Summary of the Invention

[0003] (1) Technical problems solved

[0004] In view of the shortcomings of the existing technology, the present invention provides a highway pavement ice condensation monitoring system, which solves the long-term reliability problem of sensors that may occur in extreme weather conditions.

[0005] (2) Technical solution

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: A highway pavement ice condensation monitoring system, comprising:

[0007] Data acquisition module, which collects environmental data and road icing status information on highway roads;

[0008] A self-calibration module, connected to the data acquisition module, performs self-calibration on the data acquisition module based on real-time collected environmental data and road icing status information, and corrects sensor errors caused by external environmental factors, such as ice and snow accumulation and moisture;

[0009] a data processing module, connected to the self-calibration module, processing the collected data and analyzing and predicting the road icing condition based on the calibrated data; and further classifying the road icing degree into grades of light icing, moderate icing, and heavy icing by comparing with standards, and determining the icing risk level of the current road section based on the actual data collected and the preset standard data;

[0010] The prediction module is connected to the data processing module. Based on the real-time corrected data, it uses machine learning algorithms to predict the road icing conditions and generate an icing risk assessment report.

[0011] The user interface module provides real-time data on icing risk, weather warnings, and road condition monitoring to traffic management personnel via display screens or mobile devices. This module presents road icing forecasts and risk assessment information to users, facilitating real-time decision-making and adjustment of traffic control measures.

[0012] Among them, the prediction module uses multiple regression analysis and time series prediction algorithms to quantitatively predict the possibility of road icing by analyzing historical data and real-time collected data, and generates an icing warning report.

[0013] Preferably, the data acquisition module includes a high-precision temperature and humidity sensor, an infrared sensor, a meteorological sensor and a heating device. The heating device is used to prevent the sensor from being affected by ice, snow or moisture under severe weather conditions, ensuring that the sensor can operate stably for a long time; the data acquisition module is installed on both sides of the highway pavement or along the streetlight beams and bridges to ensure that the sensor can cover key areas on the road surface and minimize external environmental interference, and obtain road environment data in real time, including: the temperature and humidity sensor regularly collects the current air temperature and humidity data, the infrared sensor collects the road surface temperature in real time, and the meteorological sensor obtains the current meteorological data, including wind speed, air pressure, and precipitation. These data help predict meteorological conditions that may cause icing, including cold waves, snowfall or humidity; the data acquisition module also continuously monitors the road surface temperature and the road icing status through the infrared sensor, and obtains road icing status information.

[0014] Preferably, the self-calibration module includes:

[0015] The sensor error detection unit is responsible for real-time monitoring of the deviation between the data output by the data acquisition module and the preset standard value, and for detecting in real time any sensor output deviation caused by external environmental factors (including ice and snow accumulation, moisture, and temperature changes). The process includes: obtaining real-time data from each sensor from the data acquisition module, including temperature, humidity, and infrared reflectivity information; comparing the real-time collected sensor data with preset standard data. The preset standard value is derived from the factory calibration value of the equipment, historical data, or external environmental standard data, including normal humidity and temperature ranges. If the sensor output data deviates from the preset standard, the system will mark the data as abnormal. When the sensor output data exceeds the allowable error range, the error detection unit will issue an alarm in real time and record the abnormal data to form feedback information for subsequent correction and modification.

[0016] a correction algorithm unit, which automatically adjusts sensor data readings based on feedback from the sensor error detection unit, eliminates errors caused by environmental changes, and ensures the accuracy of sensor data. The correction algorithm adjusts sensor output data based on historical data, current environmental data, and known calibration standards. The process includes: receiving feedback from the sensor error detection unit, analyzing the source of error, which may include ice and snow accumulation and moisture effects, and adjusting the sensor error according to temperature correction and moisture compensation rules based on a physical model using a preset correction algorithm, thereby dynamically correcting the data to make it more consistent with actual road conditions. The preset correction algorithm is based on linear regression, Kalman filtering, or a machine learning model.

[0017] The feedback mechanism unit dynamically adjusts the working parameters of the data acquisition module according to the corrected data to ensure that the entire system can maintain a high-precision working state. This mechanism can automatically adjust the working mode of the acquisition module, including the acquisition frequency and sensor calibration coefficient, to adapt to changes in the external environment. The process includes: after correcting the data, the feedback mechanism unit will adjust the parameters of the data acquisition module according to the corrected information, including increasing the sensor acquisition frequency in icy and snowy weather to better capture subtle changes in road icing; in a humid environment, adjust the sensitivity of the moisture sensor. When the external environment changes (including sudden temperature drops and heavy snowfall), the feedback mechanism will automatically adjust the calibration coefficient or acquisition method to adjust the parameters of the data acquisition module in real time to ensure that the system can continue to work stably. The feedback mechanism automatically adjusts the calibration coefficient or acquisition method to enable the data acquisition module to cope with the constant changes in the external environment.

[0018] Preferably, the correction algorithm unit analyzes the source of the error. According to the preset correction algorithm, the process of dynamically correcting the data includes: receiving feedback information from the sensor error detection unit, the feedback content includes the deviation between the sensor output data and the preset standard data, specifically including the deviation amplitude and error type of the data. The error type includes temperature deviation and moisture deviation. The feedback information is the basis for subsequent error analysis. The correction algorithm unit performs a preliminary analysis on the received feedback data to identify the source of the error. For temperature error analysis, the output value of the temperature sensor is checked to determine whether the temperature sensor error is caused by the external environment. The external environment includes ice and snow accumulation and sudden drop in temperature. For moisture error analysis, the output value of the moisture sensor is checked to analyze whether the error is caused by changes in environmental humidity or precipitation. For example, in humid weather, moisture may affect the accuracy of the moisture sensor. The error source is analyzed by a physical model to obtain an analysis result. The correction algorithm unit applies the physical model to deeply analyze the error source. Based on the physical model, a temperature error analysis is performed. When the temperature sensor data is low, it may be due to ice and snow accumulation. For errors caused by accumulation or frost, the correction algorithm applies the physical principles of the temperature change model and the heat conduction model, combined with environmental factors, to infer the actual temperature value and compare it with the sensor output to confirm the source of the error. Environmental factors include air temperature, humidity, and wind speed. Based on the moisture compensation rule, moisture error analysis is performed. The error of the moisture sensor is usually caused by excessive humidity or meteorological changes. The correction algorithm uses the moisture compensation model to compare the current ambient humidity with the sensor output value to determine whether the moisture sensor is affected by moisture changes and make appropriate corrections. The correction algorithm confirms the source of the error based on the analysis results. Temperature error confirmation: The correction algorithm uses the temperature model to confirm whether ice and snow accumulation or frost is the main source of temperature sensor error. If it is determined that the error is caused by ice and snow accumulation, the correction algorithm will generate a correction strategy and adjust the sensor data. The temperature model includes the temperature change model and the heat conduction model. Moisture error confirmation: The correction algorithm uses the moisture compensation model to confirm whether moisture changes have caused errors in the moisture sensor. If it is determined to be a moisture error, the correction algorithm will make corrections based on the actual environmental data and compensation rules.

[0019] Once the source of the error is confirmed, the correction algorithm unit will formulate a specific correction strategy and dynamically adjust the sensor data to perform temperature correction. If the error comes from ice and snow accumulation or frost, the correction algorithm will infer the true temperature value based on the environmental data and dynamically adjust the output of the temperature sensor through the physical model to perform moisture correction. For the error of the moisture sensor, the correction algorithm will use the moisture compensation rule to adjust the output of the moisture sensor to ensure that it is consistent with the actual humidity value. As the system runs for a long time, the correction algorithm will self-optimize based on the continuously accumulated historical data. The correction algorithm unit will eventually output the adjusted sensor data and pass this data to the data acquisition module to replace the original erroneous data. These corrected data provide accurate data support for subsequent data processing, icing risk analysis and prediction.

[0020] Preferably, the process of the correction algorithm unit calculating the real temperature value according to the environmental data and dynamically adjusting the output of the temperature sensor through the physical model includes:

[0021] The original output data of the temperature sensor is obtained from the temperature and humidity sensor and the infrared sensor. This data may be biased due to ice and snow accumulation and frost. At the same time, the surrounding environmental parameter data, including air temperature, humidity, and wind speed, are collected as a reference for temperature correction. The collected environmental data and temperature sensor data are filtered, smoothed, and denoised to remove noise and ensure data stability and reliability. The temperature data collected by each sensor and the environmental parameter data are integrated to form a complete environmental data set for subsequent processing.

[0022] Apply a physical model to calculate the true temperature. Select a physical model suitable for temperature correction, such as a mathematical model based on heat conduction or temperature attenuation, and determine key parameters in the model, including the ice and snow accumulation influence coefficient and the ambient temperature compensation coefficient. Using the integrated environmental data, i.e., the environmental data set, as input, the physical model is used to calculate the true temperature value of the temperature sensor before it is potentially affected by ice, snow, or frost. The model automatically calculates the correction factor by comparing the ambient temperature with the sensor reading, and then obtains the true temperature value based on the correction factor.

[0023] Based on the difference between the inferred true temperature value and the raw data from the temperature sensor, the required dynamic adjustment amount is calculated. Using preset adjustment rules, linear or nonlinear mapping, the output of the temperature sensor is dynamically adjusted to update the raw data to the corrected true temperature value. The adjustment process is continuously executed in the real-time data stream to ensure that the sensor output can continue to reflect the corrected temperature information when the environment changes, where environmental changes include sudden temperature drops and changes in ice and snow accumulation. The temperature data (true temperature value) corrected by the physical model is fed back to the data acquisition module to replace the original erroneous data for subsequent data processing and icing risk prediction. The data before and after correction, the correction factors, and the model parameter information are stored in the database for long-term optimization and further correction of the model.

[0024] Preferably, the correction algorithm unit calculates the true temperature value based on the environmental data and dynamically adjusts the output of the temperature sensor through the physical model. The process includes: the correction algorithm unit first receives feedback data from the moisture sensor from the sensor error detection unit. The feedback data includes deviation information between the moisture sensor output and a preset standard humidity value. The deviation information involves the specific value of the data being higher or lower and its fluctuation range. The feedback data is analyzed to determine whether the abnormal output of the moisture sensor is caused by external environmental humidity factors (such as high humidity environment, rainfall, dew or condensation). This process is determined by comparing the sensor data with the equipment factory calibration value, historical data or external environmental standard humidity data. Environmental parameter data related to humidity measurement are collected synchronously, including air temperature and air pressure. These data are filtered, smoothed and integrated to form a complete environmental data set, which provides necessary reference for subsequent moisture compensation. According to the pre-set moisture compensation rules, the integrated environmental data is used to calculate the compensation factor of the moisture sensor data. The compensation rule takes into account the influence of ambient temperature and air pressure factors on the moisture sensor reading, and determines the compensation coefficient through linear or nonlinear mapping relationship, so as to estimate the deviation of the moisture sensor caused by moisture. The calculated compensation factor is used to dynamically correct the original humidity data. The adjusted data can more accurately reflect the actual ambient humidity.

[0025] Preferably, the prediction module adopts multiple regression analysis and time series prediction algorithm, and predicts the possibility of road icing by analyzing historical data and real-time collected data, and generates an icing warning report; the process includes: extracting historical environmental data, sensor data and previous road icing records from the database to form the basis for model training, obtaining real-time temperature, humidity, infrared sensor data, and meteorological parameters after self-calibration through the data acquisition module, including wind speed, air pressure, and precipitation, performing missing value processing, filtering, smoothing and normalization on historical and real-time data to ensure data consistency and comparability, using a multiple regression model (linear or nonlinear), with various environmental parameters and sensor data as independent variables, and road icing status or icing degree as dependent variables, to establish a prediction model, using historical data to train the model, determining the regression coefficients of each variable, capturing the relationship between each factor and icing risk, using cross-validation and residual analysis to test the accuracy of the model, adjusting the model structure and parameters when necessary, and obtaining multiple regression analysis results, For the dynamic characteristics of real-time data, ARIMA and LSTM time series prediction algorithms are used to analyze the trends, cyclical changes and short-term fluctuations of the data. Historical data is combined to extract long-term trends, seasonal changes and the impact of emergencies, providing a time series reference for subsequent predictions. Based on current data and historical trends, the probability or risk value of road icing within a certain period of time in the future is predicted to obtain the time series prediction results. The multivariate regression analysis results are weighted and fused with the time series prediction results to form a comprehensive icing risk index. Based on the comprehensive risk index, the warning threshold is set and the road icing risk is divided into mild, moderate and severe to facilitate intuitive judgment. The prediction results, risk classification and trend analysis information are organized to generate a detailed icing warning report, which includes current risk assessment, future risk trends and recommended traffic management measures. The predicted data is intuitively displayed through charts and curves to facilitate traffic management personnel to quickly understand the risk situation. The warning report is transmitted to the traffic management department in real time through the user interface module, display screen or mobile terminal to ensure timely response.

[0026] Preferably, the process of weighted fusion of the multivariate regression analysis results and the time series prediction results to form a comprehensive icing risk index is as follows: using historical data and real-time environmental data, a multivariate regression model is used to establish the relationship between the environmental parameters of temperature, humidity, wind speed, and air pressure and the road icing risk, and generate a multivariate regression prediction result R reg , which reflects the icing risk level under current environmental conditions. Time series prediction algorithms, ARIMA and LSTM, are used to model the time dynamic changes of real-time data to obtain the time series prediction result R time , reflecting the trend change of icing risk in the future. To ensure the consistency of the numerical range of the output of different models, R reg and R timeNormalization is performed separately to map the data to the range of 0 to 1 for subsequent weighted fusion calculations. According to the prediction accuracy of each model in historical data and real-time environmental feedback, the relative performance of the multiple regression model and the time series prediction model is evaluated to determine the weighting coefficients α and β, where α + β = 1. The prediction results of the two models are combined according to the weighting coefficients to obtain a unified index that can reflect the current and future road icing risks: the comprehensive icing risk index;

[0027] According to the comprehensive icing risk index R total Based on the numerical range, the risk level division criteria are preset. 0 ≤ R total <T1: Low icing risk, T1 ≤ R total <T2: Moderate icing risk, R total ≥ T2: High icing risk. The thresholds T1 and T2 are adjusted according to historical data statistics and actual application requirements to ensure accurate and effective risk classification. According to the comprehensive risk index R total and its corresponding risk level, a detailed icing warning report is generated. The report content includes the current risk assessment, future trend prediction, impact analysis of each environmental factor, and recommended traffic management measures. Through the user interface module, the warning report is timely conveyed to the traffic management department in an intuitive form of charts and curves;

[0028] Among them, T1: The demarcation threshold between low icing risk and moderate icing risk, T2: The demarcation threshold between moderate icing risk and high icing risk.

[0029] Preferably, the data processing module includes the following steps:

[0030] S1: Compare the temperature, humidity, and infrared sensor data collected in real time with the preset standard temperature, humidity, and infrared data. By comparing, determine whether there is an icing risk on the road surface and preliminarily classify it into low icing, moderate icing, or high icing levels;

[0031] S2: Evaluate the moderately icing sections according to the meteorological data collected in real time, and combine the historical icing data to correct the preliminary classification results. Among them, the meteorological data includes wind speed and air pressure;

[0032] S3: Finally confirm the icing levels of all sections, generate an icing risk assessment report, and feedback it to the traffic management personnel through the user interface module.

[0033] Preferably, when the data processing module corrects the moderately icing sections, the real-time traffic data of the vehicle detour rate and vehicle deceleration gradient are used to perform a secondary correction on the icing level, and the corrected result more accurately reflects the actual situation of road icing;

[0034] When the data processing module corrects moderately icy sections, it uses real-time traffic data such as vehicle detour rates and vehicle deceleration gradients to perform secondary corrections on the icing level. The process is as follows:

[0035] The data processing module first compares temperature, humidity, and infrared sensor data with pre-set standards, preliminarily classifying the icing risk of each road section into light, moderate, and severe. For sections initially identified as "moderately icy," further refined assessments are considered necessary to more accurately reflect actual road conditions.

[0036] Using cameras or vehicle detection equipment installed on the road, the proportion of vehicles choosing to detour when passing through the road section is monitored in real time; this data reflects drivers' intuitive response to icy road conditions. Using high-speed cameras or sensor systems, the deceleration of vehicles passing through the road section (such as the average acceleration of vehicle deceleration) is recorded in real time to reflect the impact of road abnormalities on driving behavior.

[0037] The system presets standard vehicle detour rates and standard vehicle deceleration gradient values, which can be determined based on historical data, measured statistics or relevant traffic safety regulations; the data processing module compares the vehicle detour rates and deceleration gradients collected in real time with the corresponding standard values: if the actual vehicle detour rate is significantly higher than the standard value, it means that drivers on this road section generally take detour measures due to icing, indicating that the risk may be higher than the initial assessment; if the vehicle deceleration gradient increases significantly, it indicates that the vehicle needs to decelerate sharply on this road section, further proving that the degree of road icing may be high, and a comparison result is obtained; based on the comparison result, the system The system uses a preset correction formula to make a secondary correction to the initial moderate icing level. When the calculation results show that the comprehensive risk index increases, the system adjusts the moderate-risk road section to a severe icing risk. Conversely, if the risk index decreases, it may be adjusted to a mild icing risk. After the secondary correction, the data processing module generates a final comprehensive icing risk index, which fully reflects the relationship between road icing conditions and real-time traffic reactions. Based on the comprehensive icing risk index, the system further classifies road section risks and updates icing warning reports to ensure that traffic management personnel obtain accurate risk assessment information.

[0038] (3) Beneficial effects

[0039] The present invention provides a highway pavement ice condensation monitoring system. It has the following beneficial effects:

[0040] This highway pavement icing monitoring system adds a self-calibration function to the monitoring equipment, especially when detecting the impact of environmental factors such as humidity and temperature on sensor performance. By introducing intelligent algorithms, the monitoring system can automatically correct measurement errors caused by changes in the external environment in real time, and combine machine learning models to make more accurate predictions of road conditions. This not only improves the system's operating stability in severe weather, but also enables earlier detection of the risk of road icing, thereby providing traffic management departments with more accurate early warning information and reducing the probability of traffic accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a schematic diagram of the framework of a highway pavement ice condensation monitoring system according to the present invention;

[0042] Figure 2 Schematic diagram of the flow of the prediction module of the present invention;

[0043] Figure 3 Schematic diagram of the flow of the data processing module and the secondary correction of the data processing module of the present invention;

[0044] Figure 4 Schematic diagram of the flow of the self-calibration module and the correction algorithm unit of the present invention. DETAILED DESCRIPTION

[0045] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0046] See also Figures 1 to 4 The present invention provides a technical solution: a highway pavement ice condensation monitoring system, comprising:

[0047] Data acquisition module, which collects environmental data and road icing status information on highway roads;

[0048] A self-calibration module, connected to the data acquisition module, performs self-calibration on the data acquisition module based on real-time collected environmental data and road icing status information, and corrects sensor errors caused by external environmental factors, such as ice and snow accumulation and moisture;

[0049] The data processing module is connected to the self-calibration module to process the collected data and analyze and predict the road icing conditions based on the corrected data. It also classifies the road icing severity into light icing, moderate icing, and heavy icing levels based on standard comparisons. The actual data collected is compared with preset standard data to determine the icing risk level of the current road section.

[0050] The prediction module is connected to the data processing module. Based on the real-time corrected data, it uses machine learning algorithms to predict the road icing conditions and generate an icing risk assessment report.

[0051] The user interface module provides real-time data on icing risk, weather warnings, and road condition monitoring to traffic management personnel via display screens or mobile devices. This module presents road icing forecasts and risk assessment information to users, facilitating real-time decision-making and adjustment of traffic control measures.

[0052] The prediction module uses multiple regression analysis and time series prediction algorithms to quantitatively predict the possibility of road icing by analyzing historical data and real-time data, and generates an icing warning report.

[0053] When the data processing module corrects moderately icy sections, it uses real-time traffic data such as vehicle detour rate and vehicle deceleration gradient to make a secondary correction to the icing level. The corrected result more accurately reflects the actual situation of road icing.

[0054] The data acquisition module includes a high-precision temperature and humidity sensor, an infrared sensor, a meteorological sensor, and a heating device. The heating device is used to prevent the sensor from being affected by ice, snow, or moisture in severe weather conditions, ensuring that the sensor can operate stably for a long time. The data acquisition module is installed on both sides of the highway or along the crossbeams of streetlights and bridges to ensure that the sensor can cover key areas on the road surface and minimize external environmental interference. It obtains road environment data in real time, including: the temperature and humidity sensor regularly collects the current air temperature and humidity data, especially in cold weather and environments with large humidity changes. These data are crucial for judging whether there is a potential risk of icing on the road surface, especially when the temperature is close to freezing point, and humidity changes also affect the road surface. Infrared sensors collect road surface temperature in real time to detect icing conditions. When the road surface temperature drops below freezing, the infrared sensor can detect significant temperature differences, thereby providing important icing status information. Infrared sensors are installed on the road surface or slightly higher to ensure that the road surface status can be fully monitored. They can sense road surface temperature changes in real time, especially when icing. Infrared sensors can identify whether the road surface is frozen. Meteorological sensors obtain current meteorological data, including wind speed, air pressure, and precipitation. These data help predict meteorological conditions that may cause icing, including cold waves, snowfall, or humidity. The data acquisition module also continuously monitors the road surface temperature through infrared sensors, monitors the road surface icing status, and obtains road surface icing status information.

[0055] The self-calibration module includes:

[0056] The sensor error detection unit is responsible for real-time monitoring of the deviation between the data output by the data acquisition module and the preset standard value, and for detecting in real time any sensor output deviation caused by external environmental factors, including ice and snow accumulation, moisture, and temperature changes. The process includes: obtaining real-time data from each sensor from the data acquisition module, including temperature, humidity, and infrared reflectivity information; comparing the real-time collected sensor data with preset standard data. The preset standard value is derived from the factory calibration value of the equipment, historical data, or external environmental standard data, including normal humidity and temperature ranges. If the sensor output data deviates from the preset standard, the system will mark the data as abnormal. When the sensor output data exceeds the allowable error range, the error detection unit will issue an alarm in real time and record the abnormal data to form feedback information for subsequent correction and revision.

[0057] The correction algorithm unit automatically adjusts the reading of the sensor data based on the feedback information provided by the sensor error detection unit, eliminates the error caused by environmental changes, and ensures the accuracy of the sensor data. The correction algorithm adjusts the sensor output data according to historical data, current environmental data, and known calibration standards. The process includes: receiving feedback from the sensor error detection unit, analyzing the source of the error, which includes ice and snow accumulation and moisture influence. According to the preset correction algorithm, the sensor error is adjusted according to the temperature correction and moisture compensation rules based on the physical model, and the data is dynamically corrected to make it more consistent with the actual road conditions. The preset correction algorithm is based on linear regression, Kalman filtering or machine learning model. For ice and snow accumulation correction, if the temperature sensor causes low data due to ice and snow accumulation, the correction algorithm calculates the actual temperature value through the temperature change model and adjusts it. For moisture correction, moisture affects the accuracy of the moisture sensor. The correction algorithm combines the actual environmental humidity data and the standard humidity data to correct the sensor error. The corrected data is transmitted to the data acquisition module to replace the erroneous data value.

[0058] The feedback mechanism unit dynamically adjusts the working parameters of the data acquisition module according to the corrected data to ensure that the entire system can maintain a high-precision working state. This mechanism can automatically adjust the working mode of the acquisition module, including the acquisition frequency and sensor calibration coefficient, to adapt to changes in the external environment. The process includes: after correcting the data, the feedback mechanism unit will adjust the parameters of the data acquisition module according to the corrected information, including increasing the sensor acquisition frequency in icy and snowy weather to better capture subtle changes in road icing; in a humid environment, it will adjust the sensitivity of the moisture sensor. When the external environment changes, including sudden temperature drops and heavy snowfall, the feedback mechanism will automatically adjust the calibration coefficient or acquisition method to adjust the parameters of the data acquisition module in real time to ensure that the system can continue to work stably. The feedback mechanism automatically adjusts the calibration coefficient or acquisition method to enable the data acquisition module to cope with the continuous changes in the external environment. As the system runs for a long time, the feedback mechanism unit also continuously optimizes and adjusts the strategy through the accumulation of historical data. Based on long-term calibration data, the feedback mechanism can self-optimize the data acquisition module to further improve monitoring accuracy.

[0059] The entire self-calibration module's operating process includes: the sensor error detection unit promptly detects errors caused by changes in the external environment by comparing real-time data with standard values; the correction algorithm unit automatically adjusts the sensor data based on error feedback using physical models or algorithms to eliminate errors caused by ice and snow accumulation and moisture in the environment; the feedback mechanism unit adjusts the operating parameters of the data acquisition module in real time based on the corrected data to ensure the long-term stability and accuracy of the system. Through this mechanism, the self-calibration module can ensure that the highway pavement ice monitoring system maintains high precision under various environmental conditions and provides timely and accurate pavement icing risk assessments.

[0060] The correction algorithm unit analyzes the source of the error. According to the preset correction algorithm, the process of dynamically correcting the data includes: receiving feedback information from the sensor error detection unit. The feedback content includes the deviation between the sensor output data and the preset standard data, specifically including the deviation amplitude and error type of the data. The error type includes temperature deviation and moisture deviation. The feedback information is the basis for subsequent error analysis. The correction algorithm unit performs a preliminary analysis on the received feedback data to identify the source of the error. For temperature error analysis, the output value of the temperature sensor is checked to determine whether the temperature sensor error is caused by the external environment. The external environment includes ice and snow accumulation and a sudden drop in temperature. For example, if the output value of the temperature sensor is significantly lower than the actual temperature, it may be an error caused by ice and snow accumulation. For moisture error analysis, the output value of the moisture sensor is checked to analyze whether the error is caused by changes in environmental humidity or precipitation. For example, in humid weather, moisture may affect the accuracy of the moisture sensor. The error source is analyzed through a physical model to obtain an analysis result. The correction algorithm unit applies the physical model to deeply analyze the error source. Based on the physical model, temperature error analysis is performed. When the temperature sensor When sensor data is low, it may be due to errors caused by ice and snow accumulation or frost. The correction algorithm applies the physical principles of the temperature change model and the heat conduction model, combined with environmental factors, to calculate the actual temperature value and compare it with the sensor output to confirm the source of the error. Environmental factors include air temperature, humidity, and wind speed. Based on the moisture compensation rules, moisture error analysis is performed. Moisture sensor errors are usually caused by excessive humidity or meteorological changes. The correction algorithm uses the moisture compensation model to compare the current ambient humidity with the sensor output value to determine whether the moisture sensor is affected by moisture changes and make appropriate corrections. The correction algorithm confirms the error source based on the analysis results. Temperature error confirmation: The correction algorithm uses the temperature model to confirm whether ice and snow accumulation or frost is the main source of temperature sensor error. If the error is determined to be caused by ice and snow accumulation, the correction algorithm will generate a correction strategy to adjust the sensor data. The temperature model includes the temperature change model and the heat conduction model. Moisture error confirmation: The correction algorithm uses the moisture compensation model to confirm whether moisture changes have caused error in the moisture sensor. If it is determined to be a moisture error, the correction algorithm will make corrections based on the actual environmental data and compensation rules.

[0061] Once the source of the error is confirmed, the correction algorithm unit will formulate a specific correction strategy and dynamically adjust the sensor data to perform temperature correction. If the error comes from ice and snow accumulation or frost, the correction algorithm will infer the true temperature value based on the environmental data and dynamically adjust the output of the temperature sensor through the physical model to perform moisture correction. For the error of the moisture sensor, the correction algorithm will use the moisture compensation rule to adjust the output of the moisture sensor to ensure that it is consistent with the actual humidity value. As the system runs for a long time, the correction algorithm will self-optimize based on the continuously accumulated historical data. Through long-term historical data, the error analysis and correction strategy will be continuously optimized to enable the system to adapt to changes under different environmental conditions and provide a more accurate correction algorithm. The optimized correction strategy will be fed back to the data acquisition module to further improve the system's operating stability and data accuracy. The correction algorithm unit will finally output the adjusted sensor data and pass this data to the data acquisition module to replace the original erroneous data. These corrected data provide accurate data support for subsequent data processing, icing risk analysis and prediction.

[0062] The correction algorithm unit calculates the actual temperature value based on the environmental data and dynamically adjusts the output of the temperature sensor through the physical model. The process includes:

[0063] The original output data of the temperature sensor is obtained from the temperature and humidity sensor and the infrared sensor. This data may be biased due to ice and snow accumulation and frost. At the same time, the surrounding environmental parameter data, including air temperature, humidity, and wind speed, are collected as a reference for temperature correction. The collected environmental data and temperature sensor data are filtered, smoothed, and denoised to remove noise and ensure data stability and reliability. The temperature data collected by each sensor and the environmental parameter data are integrated to form a complete environmental data set for subsequent processing.

[0064] Apply a physical model to calculate the true temperature. Select a mathematical model based on heat conduction or temperature attenuation and determine key model parameters, including the ice and snow accumulation coefficient and the ambient temperature compensation coefficient. Using the integrated environmental data, i.e., the environmental dataset, as input, the physical model calculates the true temperature value of the temperature sensor before it is potentially affected by ice, snow, or frost. The model automatically calculates a correction factor by comparing the ambient temperature with the sensor reading, and uses this correction factor to determine the true temperature value.

[0065] Based on the difference between the inferred true temperature value and the original data from the temperature sensor, the required dynamic adjustment amount is calculated. Using preset adjustment rules, linear or nonlinear mapping, the output of the temperature sensor is dynamically adjusted to update the original data to the corrected true temperature value. The adjustment process is continuously executed in the real-time data stream to ensure that when the environment changes, the sensor output can continue to reflect the corrected temperature information. Among them, environmental changes include sudden temperature drops and changes in ice and snow accumulation. The true temperature value corrected by the physical model is fed back to the data acquisition module to replace the original erroneous data for subsequent data processing and icing risk prediction. The data before and after correction, the correction factor and model parameter information are stored in the database to facilitate long-term optimization and further correction of the model.

[0066] The correction algorithm unit calculates the true temperature value based on the environmental data and dynamically adjusts the output of the temperature sensor through the physical model. The process includes: the correction algorithm unit first receives the feedback data of the moisture sensor from the sensor error detection unit. The feedback data includes the deviation information between the moisture sensor output and the preset standard humidity value. The deviation information involves the specific value of the data that is too high or too low and its fluctuation range. The feedback data is analyzed to determine whether the abnormal output of the moisture sensor is caused by a high humidity environment, rainfall, dew or condensation phenomenon in the external environment humidity factors. This process is determined by comparing the sensor data with the factory calibration value of the equipment, historical data or external environment standard humidity data, and synchronously collecting environmental parameter data related to humidity measurement. The environmental parameter data include air temperature and air pressure, and these data are filtered, smoothed and integrated to form a complete environmental data set. , providing necessary reference for subsequent moisture compensation. According to the pre-set moisture compensation rules, the compensation factor of the moisture sensor data is calculated using the integrated environmental data. The compensation rule takes into account the influence of ambient temperature and air pressure factors on the moisture sensor readings, and determines the compensation coefficient through linear or nonlinear mapping relationships, thereby estimating the deviation of the moisture sensor caused by moisture. The calculated compensation factor is used to dynamically correct the original humidity data. The adjusted data can more accurately reflect the actual ambient humidity. The correction process is executed in real time to ensure that when the humidity conditions change, the moisture sensor output can be updated in real time to keep consistent with the actual humidity. The humidity data corrected by the moisture compensation rule is fed back to the data acquisition module, and the data before and after correction, compensation factors and environmental parameter information are stored to facilitate subsequent continuous optimization and adjustment of the moisture compensation rules and correction algorithms.

[0067] The prediction module uses multiple regression analysis and time series prediction algorithms to predict the possibility of road icing by analyzing historical data and real-time collected data, and generate an icing warning report; the process includes: extracting historical environmental data, sensor data and previous road icing records from the database to form the basis for model training, obtaining real-time temperature, humidity, infrared sensor data, and meteorological parameters after self-calibration through the data acquisition module, including wind speed, air pressure, and precipitation, and performing missing value processing, filtering, smoothing and normalization on historical and real-time data to ensure data consistency and comparability, using a multiple nonlinear regression model with various environmental parameters and sensor data as independent variables and road icing status or icing degree as dependent variables to establish a prediction model, using historical data to train the model, determining the regression coefficients of each variable, and capturing the relationship between each factor and icing risk, using cross-validation and residual analysis to test the accuracy of the model, adjusting the model structure and parameters when necessary, and obtaining multiple regression analysis results, and using ARIMA and LSTM time series prediction algorithms to analyze data trends, periodic changes and short-term trends based on the dynamic characteristics of real-time data. The system combines historical data to extract long-term trends, seasonal changes, and the impact of sudden events, providing a time-series reference for subsequent forecasts. Based on current data and historical trends, it predicts the probability or risk value of road icing within a certain period of time in the future, obtaining a time-series forecast result. The multivariate regression analysis results are weighted and fused with the time-series forecast results to form a comprehensive icing risk index. Based on the comprehensive risk index, warning thresholds are set, and the road icing risk is classified into mild, moderate, and severe levels to facilitate intuitive judgment. The forecast results, risk classification, and trend analysis information are organized to generate a detailed icing warning report, which includes current risk assessment, future risk trends, and recommended traffic management measures. The forecast data is intuitively displayed through charts and graphs, allowing traffic management personnel to quickly understand the risk situation. The warning report is transmitted to the traffic management department in real time through the user interface module, display screen, or mobile terminal to ensure timely response. The forecast results, actual road conditions, and environmental parameters are recorded in real time to form a new historical data set. The multivariate regression model and time-series forecast model are regularly retrained and parameter adjusted using the latest data to continuously optimize forecast accuracy and response speed.

[0068] The process of weighted fusion of the multiple regression analysis results and the time series prediction results to form a comprehensive icing risk index is as follows: using historical data and real-time environmental data, the relationship between the environmental parameters of temperature, humidity, wind speed, and air pressure and the road icing risk is established through a multiple regression model, and the multiple regression prediction result R is generated. reg , which reflects the icing risk level under current environmental conditions. The time series prediction algorithm ARIMA is used to model the time dynamic changes of real-time data and obtain the time series prediction result R time, reflecting the trend change of icing risk in a future period. To ensure the consistency of the numerical ranges of different model outputs, R reg and R time are normalized respectively, mapping the data to the interval of 0 to 1 for subsequent weighted fusion calculation. According to the prediction accuracy of each model in historical data and real-time environmental feedback, the relative performance of the multiple regression model and the time series prediction model is evaluated to determine the weighted coefficients α and β, where α + β = 1, which is a fixed weight. An adaptive weight strategy can also be adopted: under certain climate conditions, if the multiple regression prediction is more stable, α can be set larger; otherwise, the proportion of β is increased. The comprehensive icing risk index R total is calculated using the following fusion formula:

[0069] ;

[0070] This formula synthesizes the prediction results of the two models according to the weighted coefficients, obtaining a unified index that can reflect the current and future road icing risks: the comprehensive icing risk index;

[0071] According to the numerical range of the comprehensive icing risk index R total , the risk level division criteria are preset. 0 ≤ R total < T1: low icing risk, T1 ≤ R total < T2: medium icing risk, R total ≥ T2: high icing risk. The thresholds T1 and T2 are adjusted according to historical data statistics and actual application requirements to ensure the accuracy and effectiveness of risk classification. According to the comprehensive risk index R total and its corresponding risk level, a detailed icing warning report is generated. The report content includes current risk assessment, future trend prediction, impact analysis of various environmental factors, and recommended traffic management measures. Through the user interface module, the warning report is timely conveyed to the traffic management department in the intuitive forms of charts and curves. Among them, T1: the demarcation threshold between low icing risk and medium icing risk, T2: the demarcation threshold between medium icing risk and high icing risk. During the operation of the system, the prediction results and the actual road conditions are continuously recorded to form new historical data. These data are used to retrain the multiple regression and time series prediction models regularly, and at the same time, the weighted coefficients are dynamically adjusted to ensure the continuous improvement of the accuracy and adaptability of the comprehensive prediction and achieve the self-optimization of the model;

[0072] Through the above process, the results of multiple regression analysis and time series prediction are effectively weighted and fused to form a comprehensive icing risk index, providing accurate and timely risk assessment data for highway road icing monitoring and warning;

[0073] The data processing module includes a data preprocessing unit, which is used to filter, smooth and denoise the collected raw data to ensure data quality for subsequent analysis and prediction.

[0074] The data processing module includes the following steps:

[0075] S1: Compares the real-time collected temperature, humidity, and infrared sensor data with the preset standard temperature, humidity, and infrared data to determine whether there is a risk of road icing and preliminarily classify the road as light icing, moderate icing, or heavy icing.

[0076] S2: Evaluate the moderately icy road sections based on real-time meteorological data, and revise the preliminary classification results based on historical icing data. Meteorological data includes wind speed and air pressure.

[0077] S3: Final confirmation of the icing levels of all road sections, generation of an icing risk assessment report and feedback to traffic management personnel through the user interface module.

[0078] It should be further explained that, in the specific implementation process, the data processing module uses the real-time traffic data of vehicle detour rate and vehicle deceleration gradient to make a secondary correction to the icing level when correcting the moderately icy road section. The process is as follows: the data processing module first compares the data of temperature, humidity and infrared sensors with the preset standards, and preliminarily divides the icing risk of each road section into light, moderate and severe; for the road sections preliminarily determined to be "moderately icy", it is considered necessary to further refine the assessment of these areas in order to more accurately reflect the actual road conditions; through the cameras or vehicle detection equipment installed on the road, the proportion of vehicles that choose to detour when passing through the road section is monitored in real time; the data reflects the driver's intuitive response to the road icing conditions; the use of high-speed cameras or sensor systems records the real-time movement of vehicles passing through the road section. The deceleration situation at that time, including the average acceleration of vehicle deceleration, is used to reflect the impact of road abnormalities on driving behavior; the system presets standard vehicle detour rate and standard vehicle deceleration gradient values, which can be determined based on historical data, actual measurement statistics or relevant traffic safety regulations; the data processing module compares the vehicle detour rate and deceleration gradient collected in real time with the corresponding standard values: if the actual vehicle detour rate is significantly higher than the standard value, it means that drivers in this section of road generally take detour measures due to icing, indicating that the risk may be higher than the preliminary assessment; if the vehicle deceleration gradient increases significantly, it indicates that the vehicle needs to decelerate sharply in this section of road, further proving that the degree of road icing may be high, and a comparison result is obtained; based on the comparison result, the system uses a preset correction formula to make a secondary correction to the preliminary moderate icing level; the correction formula is as follows:

[0079] ;

[0080] Among them, Rinitial is the preliminary classification of icing risk level; W real With D real are the real-time vehicle detour rate and vehicle deceleration gradient respectively; W standard With D standard is the preset standard value; k1 and k2 are correction coefficients, which are set according to actual conditions;

[0081] When the calculation results show that the comprehensive risk index increases, the system adjusts the moderate risk section to a severe icing risk; conversely, if the risk index decreases, it may be adjusted to a mild icing risk; after the second correction, the data processing module generates the final comprehensive icing risk index, which fully reflects the relationship between road icing conditions and real-time traffic response; based on the comprehensive icing risk index, the system further classifies the road section risk and updates the icing warning report to ensure that traffic management personnel obtain accurate risk assessment information; the corrected icing risk level is fed back to the user interface module through the data interaction module and displayed in real time on the monitoring terminal or mobile device; the system also records all secondary correction data and related traffic data to provide a basis for subsequent long-term analysis and model optimization, thereby continuously improving the prediction accuracy and risk assessment accuracy; through the above process, the data processing module uses real-time traffic data such as vehicle detour rate and vehicle deceleration gradient to make secondary corrections to the road sections initially determined to be moderately iced, ensuring that the icing risk level more accurately reflects the actual road conditions and provides more accurate warning information to the traffic management department;

[0082] Through steps S1 to S3, the system not only makes a preliminary division based on real-time data and standard data, but can also make corrections and corrections based on further environmental factors, ultimately deriving an accurate icing level; by comparing actual data with preset standard data, the system can divide road icing conditions into three levels: light, moderate, and heavy icing; this comparison method enables the monitoring system to dynamically assess icing risks and ensure that the icing conditions of each road section can be accurately reflected; through this detailed technical description, combined with standard comparison and correction steps, the accuracy and reliability of highway road icing monitoring can be effectively improved.

[0083] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a..." does not exclude the presence of additional identical elements in the process, method, article, or device comprising the element.

[0084] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A highway pavement ice condensation monitoring system, characterized in that: include: Data acquisition module, which collects environmental data and road icing status information on highway roads; The self-calibration module is connected to the data acquisition module and performs self-calibration on the data acquisition module based on the real-time collected environmental data and road icing status information to correct sensor errors caused by external environmental factors. The self-calibration module includes: The sensor error detection unit is responsible for real-time monitoring of the deviation between the data output by the data acquisition module and the preset standard value. It obtains real-time data from each sensor from the data acquisition module and compares the real-time data with the preset standard data. If the sensor output data deviates from the preset standard, the data is marked as abnormal, an alarm is issued in real time, and the abnormal data is recorded to form feedback information; A correction algorithm unit automatically adjusts sensor data readings based on feedback from the sensor error detection unit to eliminate errors caused by environmental changes. The process includes: receiving feedback from the sensor error detection unit, analyzing the source of the error, and adjusting the sensor error according to temperature correction and moisture compensation rules based on the physical model using a preset correction algorithm to correct the data. The feedback mechanism unit dynamically adjusts the working parameters of the data acquisition module according to the corrected data. After the data is corrected, the parameters of the data acquisition module are adjusted according to the corrected information. When the external environment changes, the parameters of the data acquisition module are adjusted in real time by automatically adjusting the calibration coefficient or the acquisition method. a data processing module, connected to the self-calibration module, processing the collected data and analyzing and predicting the road icing condition based on the calibrated data; and further classifying the road icing degree into grades of light icing, moderate icing, and heavy icing by comparing with standards, and determining the icing risk level of the current road section based on the actual data collected and the preset standard data; The prediction module is connected to the data processing module. Based on the real-time corrected data, it uses machine learning algorithms to predict the road icing conditions and generate an icing risk assessment report. A user interface module provides real-time data on icing risk, weather warnings, and road condition monitoring to traffic management personnel via display screens or mobile devices, presenting road icing forecasts and risk assessment information to users; Among them, the prediction module uses multiple regression analysis and time series prediction algorithms to predict the possibility of road icing by analyzing historical data and real-time collected data, and generates an icing warning report.

2. The highway pavement ice condensation monitoring system according to claim 1, characterized in that: The data acquisition module includes a temperature and humidity sensor, an infrared sensor, a meteorological sensor and a heating device; The data acquisition module is installed on both sides of the highway pavement or along the crossbeams of streetlights and bridges to obtain real-time road environment data, including: the temperature and humidity sensor collects the current air temperature and humidity data, the infrared sensor collects the road surface temperature, and the meteorological sensor obtains the current meteorological data, including wind speed, air pressure, and precipitation. The data acquisition module also continuously monitors the road surface temperature through the infrared sensor, monitors the road surface icing status, and obtains road surface icing status information.

3. The highway pavement ice condensation monitoring system according to claim 2, characterized in that: The correction algorithm unit analyzes the source of the error. According to a preset correction algorithm, the process of dynamically correcting data includes: receiving feedback information, performing preliminary analysis, identifying the source of the error, checking the output value of the temperature sensor to determine whether the temperature sensor error is caused by the external environment, checking the output value of the humidity sensor to determine whether the error is caused by changes in ambient humidity or precipitation, analyzing the source of the error through a physical model, and obtaining an analysis result. When the temperature sensor data is low, the correction algorithm applies a temperature change model and a heat conduction model, combined with environmental factors, to infer the actual temperature value, and compares it with the sensor output to confirm the source of the error. The correction algorithm uses a humidity compensation model to compare the current ambient humidity with the sensor output value to determine whether the humidity sensor is affected by humidity changes and perform corrections. The correction algorithm confirms the source of the error based on the analysis result, and uses the temperature model to determine whether ice and snow accumulation or frost is the main source of the temperature sensor error. If it is determined that the error is caused by ice and snow accumulation, the correction algorithm generates a correction strategy and adjusts the sensor data. Using the humidity compensation model, the correction algorithm confirms whether humidity changes have caused errors in the humidity sensor. If it is determined to be a humidity error, the correction algorithm performs corrections based on actual environmental data and compensation rules. If the error comes from ice and snow accumulation or frost, the true temperature value is calculated based on the environmental data, and the output of the temperature sensor is dynamically adjusted through the physical model. For the error of the moisture sensor, the moisture compensation rule is used to adjust the output of the moisture sensor to ensure that it is consistent with the actual humidity value. The adjusted sensor data is output and passed to the data acquisition module to replace the original erroneous data.

4. The highway pavement ice condensation monitoring system according to claim 3, characterized in that: The process of the correction algorithm unit calculating the true temperature value based on the environmental data and dynamically adjusting the output of the temperature sensor through the physical model includes: The original output data of the temperature sensor is obtained from the temperature and humidity sensor and the infrared sensor, and the surrounding environmental parameter data is collected at the same time. The collected environmental data and temperature sensor data are filtered, smoothed and denoised to remove noise. The temperature data collected by each sensor and the environmental parameter data are integrated to form an environmental data set; Select a physical model suitable for temperature correction, determine key parameters in the model based on a mathematical model of heat conduction or temperature attenuation, use the environmental data set as input, and calculate the actual temperature value of the temperature sensor before it is affected by ice, snow, or frost based on the physical model. By comparing the ambient temperature with the sensor reading, calculate the correction factor, and then obtain the actual temperature value based on the correction factor. Based on the difference between the actual temperature value and the original data of the temperature sensor, the dynamic adjustment amount is calculated. The output of the temperature sensor is dynamically adjusted using the preset adjustment rules, and the original data is updated to the corrected actual temperature value. The temperature data corrected by the physical model is fed back to the data acquisition module to replace the original data with errors.

5. The highway pavement ice condensation monitoring system according to claim 4, characterized in that: The correction algorithm unit calculates the true temperature value based on the environmental data and dynamically adjusts the output of the temperature sensor through the physical model. The process includes: receiving feedback data from the moisture sensor, analyzing the feedback data, judging whether the abnormal output of the moisture sensor is caused by external environmental moisture factors, and making a judgment by comparing the sensor data with the factory calibration value of the equipment, historical data or external environmental standard humidity data. The environmental parameter data related to the humidity measurement are simultaneously collected, filtered, smoothed and integrated to form an environmental data set. According to a pre-set moisture compensation rule, the integrated environmental data is used to calculate the compensation factor of the moisture sensor data, and the compensation coefficient is determined through a linear or nonlinear mapping relationship. The deviation of the moisture sensor caused by the influence of moisture is estimated, and the original humidity data is corrected using the compensation factor.

6. The highway pavement ice condensation monitoring system according to claim 5, characterized in that: The prediction module uses multiple regression analysis and time series prediction algorithms to analyze historical data and real-time collected data to predict the possibility of road icing and generate an icing warning report; Historical environmental data, sensor data and previous road icing records are extracted from the database to form the basis for model training. Real-time temperature, humidity, infrared sensor data and meteorological parameters after self-calibration are obtained through the data acquisition module. Missing value processing, filtering, smoothing and normalization are performed on historical and real-time data. A multivariate regression model is used to establish a prediction model with various environmental parameters and sensor data as independent variables and road icing status as dependent variable. The model is trained using historical data to determine the regression coefficients of each variable and capture the relationship between each factor and icing risk. Cross-validation and residual analysis are used to test the accuracy of the model to obtain the results of multivariate regression analysis. In view of the dynamic characteristics of real-time data, ARIMA and LSTM time series prediction algorithms analyze data trends, cyclical changes, and short-term fluctuations, and combine historical data to extract long-term trends, seasonal changes, and the impact of emergencies. Based on current data and historical trends, the probability or risk value of road icing is predicted to obtain time series prediction results. The multivariate regression analysis results are weighted and fused with the time series prediction results to form a comprehensive icing risk index. Based on the comprehensive risk index, the warning threshold is set, and the road icing risk is divided into mild, moderate, and severe. The prediction results, risk classification, and trend analysis information are organized to generate a warning report, which is displayed through charts and graphs. The warning report is transmitted to the traffic management department through the user interface module.

7. The highway pavement ice condensation monitoring system according to claim 6, characterized in that: The multivariate regression analysis results are weighted and integrated with the time series prediction results to form a comprehensive icing risk index. The relationship between temperature, humidity, wind speed, air pressure and road icing risk is established through a multivariate regression model using historical data and real-time environmental data to generate a multivariate regression prediction result R. reg , using the time series prediction algorithm to model the time dynamic changes of real-time data and obtain the time series prediction result R time , for R reg and R time Normalization processing was performed separately. Based on the prediction accuracy of each model in historical data and real-time environmental feedback, the performance of the multivariate regression model and the time series prediction model was evaluated. The weighting coefficients α and β were determined, where α + β = 1. The prediction results of the two models were combined according to the weighting coefficients to obtain a comprehensive icing risk index. According to the comprehensive icing risk index R total Within the numerical range, the risk level classification standard is preset. 0 ≤ R total <T1: Mild icing risk, T1 ≤ R total <T2: Moderate icing risk, R total ≥ T2: Severe icing risk. According to the comprehensive risk index R total and its corresponding risk level, an icing warning report is generated and conveyed to the traffic management department in the form of charts and curves through the user interface module.

8. The highway pavement ice condensation monitoring system according to claim 7, characterized in that: The data processing module includes the following steps: S1: Compares the real-time collected temperature, humidity, and infrared sensor data with the preset standard temperature, humidity, and infrared data to determine whether there is a risk of road icing and preliminarily classify the road as light icing, moderate icing, or heavy icing. S2: Evaluate the moderately icy road sections based on real-time meteorological data, and revise the preliminary classification results based on historical icing data. Meteorological data includes wind speed and air pressure. S3: Final confirmation of the icing levels of all road sections, generation of an icing risk assessment report and feedback to traffic management personnel through the user interface module.

9. The highway pavement ice condensation monitoring system according to claim 8, characterized in that: When the data processing module corrects the moderately icy road section, it uses the real-time traffic data of the vehicle detour rate and the vehicle deceleration gradient to make a secondary correction to the icing level, and compares the data of the temperature, humidity and infrared sensors with the preset standards to preliminarily divide the icing risk of each road section into light, moderate and severe. It uses the camera or vehicle detection equipment installed on the road surface to monitor in real time the proportion of vehicles that choose to detour when passing through the road section, and uses the high-speed camera or sensor system to record in real time the deceleration of the vehicle when passing through the road section. It presets the standard vehicle detour rate and the standard vehicle deceleration gradient value, and compares the real-time collected vehicle detour rate and deceleration gradient with the corresponding standard values ​​respectively: if the actual vehicle detour rate is higher than the standard value, it indicates that the risk is higher than the preliminary assessment; if the vehicle deceleration gradient increases significantly, the degree of icing on the road surface is high, and a comparison result is obtained; based on the comparison result, the preliminary moderate icing level is corrected for a secondary time by using the correction formula; when the calculation result shows that the comprehensive risk index has increased, the moderate risk road section is adjusted to a severe icing risk; On the contrary, if the risk index decreases, it is adjusted to a slight icing risk; After the second correction, a comprehensive icing risk index is generated. According to the comprehensive icing risk index, the road section risk is graded and the icing warning report is updated.

Images