Short-time forecasting and correcting method for visibility of expressway

Abstract

The invention relates to a short-time forecasting and correcting method for visibility of an expressway. The method comprises the following steps: acquiring 10-minute meteorological observation historical data of traffic meteorological observation stations along a measured road section of the expressway; calculating a Pearson's correlation coefficient between the visibility and each meteorological factor, and carrying out significance test; acquiring hour-by-hour real-time meteorological observation data along the measured road section of the expressway; training the LSTM prediction model to obtain visibility prediction results of stake marks per 10 minutes in the future 2 hours; the method comprises the following steps: performing probability correction on a visibility prediction result by using visibility observation data obtained from an automatic weather station, performing trend correction on the visibility prediction result by using visibility data inverted by a video image and obtained along a highway, and correcting a subsequent new visibility prediction result by using a correction result. And obtaining visibility forecasting results of stake marks and 10-minute intervals along the measured road section of the expressway within 0-2 hours of short-time adjacent time.

Description

Technical Field

[0001] The invention relates to a method for forecasting highway visibility, in particular to a method for short-term forecasting and correction of highway visibility. Background Art

[0002] Heavy fog is the main factor that causes highway accidents and traffic obstructions, and the impact of fog on highways is mainly reflected by visibility. Therefore, accurate forecasts of heavy fog and visibility, especially forecasts of highway visibility within the next 2 hours, are of great significance for improving highway driving safety index and highway traffic efficiency. In order to cope with the adverse effects of small-scale, local low visibility events, it is also necessary to carry out visibility forecasts on a pile-by-pile basis. However, mature technical applications in this area are still lacking and urgently need to be researched and developed. Summary of the invention

[0003] The purpose of the present invention is to provide a method for short-term forecasting and correction of highway visibility to improve the traffic safety of highways and meet the development needs of safe operation of highways.

[0004] The object of the present invention is achieved in that:

[0005] A method for short-term forecasting and correction of highway visibility comprises the following steps:

[0006] S1. Obtain the historical meteorological observation data of every 10 minutes for at least one year from the traffic meteorological observation stations along the measured section of the expressway, including visibility, temperature, relative humidity, average wind speed in the past 2 minutes, temperature change in the past 24 hours, maximum temperature in the past 10 minutes, and minimum temperature in the past 10 minutes; among them, other factors except visibility constitute meteorological factors related to visibility prediction; the starting point of each past time is the time point of every 10 minutes.

[0007] S2. Use the correlation analysis method to calculate the Pearson correlation coefficient between visibility and various meteorological factors, and perform a significance test; select meteorological factors with α=0.01 in the significance test and use them as meteorological variables for visibility prediction.

[0008] S3. Obtain hourly real-time meteorological observation data along the measured section of the highway, introduce the visibility changes in 10 minutes, 1 hour, 3 hours and 6 hours, and use them as time series trend characteristic variables; introduce the minimum value and standard deviation of visibility in the last 6 hours, and use them as time series aggregation characteristic variables.

[0009] S4. After standardization, the three types of variables, namely meteorological variables, time series trend characteristic variables and time series aggregation characteristic variables, are used together as variables for visibility prediction to train the LSTM prediction model, and the visibility prediction data for each 10 minutes in the next 2 hours at the location of the traffic meteorological observation points along the tested section of the highway are obtained; the spline interpolation method is used to perform spatial difference on the visibility prediction data to obtain the visibility prediction results for each 10 minutes and each pile number along the tested section of the highway in the next 2 hours.

[0010] S5. Obtain visibility observation data at the time of arrival corresponding to the visibility prediction result from the automatic weather station within 5 km around the measured section of the expressway, and obtain visibility data inverted from video images for each stake number at the time of arrival corresponding to the visibility prediction result from the measured section of the expressway. After filtering out invalid data in the above two groups of data, first use the retained visibility observation data to perform probability correction on the visibility prediction result to obtain the correction probability; then use the retained visibility data inverted from video images to perform trend correction on the visibility prediction result to obtain the visibility change trend; after using the correction probability and visibility change trend to correct the subsequent newly obtained visibility prediction results respectively, the visibility forecast results for each stake number and each 10-minute interval along the measured section of the expressway within the short-term proximity time limit of 0-2 hours are obtained.

[0011] Furthermore, the Pearson correlation coefficient r between visibility and each meteorological factor in step S2 is calculated as follows:

[0012]

[0013] Among them, X and Y are two variables. and is the average value of the two groups of variables, and n is the number of meteorological factors.

[0014] Furthermore, the significance test in step S2 is performed as follows:

[0015] First calculate according to the following formula:

[0016]

[0017] Where t is the deviation statistic between the sample mean and the population mean, is the sample mean, μ is the population mean, σ Z is the sample standard deviation, m is the sample size;

[0018] Then, by looking up the t value table, determine whether it passes the significance test of α=0.01.

[0019] Furthermore, the standardization processing of the three types of variables in step S4 is as follows:

[0020]

[0021] Among them, V is the standardized variable value, g is the number of variables, C g is the variable sequence of the g-th variable, is the mean value of the g-th variable.

[0022] Furthermore, in step S5, the screening conditions for filtering out invalid data from the acquired visibility observation data and the visibility data inverted from the video image are one of the following: ① the change in visibility between two consecutive minutes is not greater than 500m; ② the minimum duration of visibility is not less than 3min; ③ the relative humidity is greater than 80%; ④ the wind speed is not greater than 4.5m / s.

[0023] Furthermore, in step S5, the method for probability correction of the visibility prediction result is to use the Gmama probability function for correction.

[0024] Furthermore, in step S5, the method for correcting the trend of the visibility prediction result is to use the visibility data inverted from the video image to correct the trend of the visibility prediction result. The specific method is to establish a statistical model:

[0025] y=Ax+B

[0026] Among them, y is the visibility value after trend correction, x is the visibility prediction result, and A and B are coefficients.

[0027] The present invention realizes the refined forecast of visibility along the expressway on a pile-by-pile and 10-minute basis, and can make accurate forecasts of heavy fog and visibility on the expressway, especially making relatively accurate forecasts of visibility on the expressway within the next 2 hours, which is of great significance for improving the driving safety index of the expressway and the traffic efficiency of the expressway, and its prediction results and revised prediction results are better than the prediction and forecast results made by the existing visibility prediction methods. DETAILED DESCRIPTION

[0028] The present invention will be further described in detail below in conjunction with the embodiments.

[0029] The present invention uses a 20-kilometer section of the Beijing-Shijiazhuang section of the Beijing-Hong Kong-Macao Expressway as the measured section, and predicts visibility for each 10 minutes and each pile number in the next 2 hours. The specific prediction method is as follows:

[0030] The first step is to obtain historical meteorological observation data for every 10 minutes from October 2023 to September 2024 for a total of one year from traffic meteorological observation stations along the measured section of the highway, including visibility, temperature, relative humidity, average wind speed in the past 2 minutes, temperature change in the past 24 hours, maximum temperature in the past 10 minutes, and minimum temperature in the past 10 minutes. The starting points of the past 2 minutes, past 10 minutes, and past 24 hours here are all 10-minute information collection time points. Among them, other factors except visibility constitute meteorological factors related to visibility prediction.

[0031] In the second step, first, the Pearson correlation coefficient between visibility and the above-mentioned meteorological factors is calculated using the correlation analysis method. The calculation method of the Pearson correlation coefficient r between visibility and each meteorological factor is as follows:

[0032]

[0033] Among them, X and Y are two variables. and is the average value of the two groups of variables, and n is the number of meteorological factors.

[0034] The calculation results of Pearson correlation coefficient are shown in Table 1.

[0035] Table 1. Pearson correlation coefficient between visibility and various meteorological factors

[0036]

[0037] Note: The data marked with ** have passed the significance test of α=0.01.

[0038] Secondly, a significance test is conducted. The significance test includes the following two sub-steps:

[0039] ①Calculate according to the following formula:

[0040]

[0041] Where t is the deviation statistic between the sample mean and the population mean, is the sample mean, μ is the population mean, σ Z is the sample standard deviation, and m is the sample size.

[0042] ② By looking up the t value table, determine whether it passes the significance test of α=0.01.

[0043] The meteorological factors with α=0.01 in the significance test are selected and used as meteorological variables for visibility prediction. As shown in Table 1, all the six meteorological factors listed have passed the significance test with α=0.01 and can be used as meteorological variables for visibility prediction.

[0044] The third step is to obtain hourly real-time meteorological observation data in October 2024 from the traffic meteorological observation stations along the tested section of the highway as a test. The visibility changes of 10 minutes, 1 hour, 3 hours and 6 hours are introduced as time series trend feature variables. The minimum value and standard deviation of visibility in the last 6 hours are introduced as time series aggregation feature variables.

[0045] The fourth step is to standardize the three types of variables: meteorological variables, time series trend characteristic variables, and time series aggregation characteristic variables. The standardization of the three types of variables is as follows:

[0046]

[0047] Among them, V is the standardized variable value, g is the number of variables, C g is the variable sequence of the g-th variable, is the mean value of the g-th variable.

[0048] After standardization, these three types of variables are used together as visibility prediction variables to train the LSTM prediction model. The LSTM prediction model is a commonly used deep learning neural network for time series prediction, which contains three LSTM layers, several hidden layers and one fully connected layer.

[0049] The present invention sets the time step of the LSTM prediction model to 10 minutes, sets the number of hidden layers in the model to 1, sets the number of hidden layer neurons to 64, determines the discard rate to 0.01, sets the training batch size to 128, and sets the number of iterations to 500. The visibility prediction data of 10 minutes each in the next 2 hours at the locations of two sets of traffic meteorological observation points along the tested section of the highway are calculated by using the visibility prediction variables and the LSTM prediction model. The prediction data is compared with the actual observation data as follows:

[0050] According to visibility, it is divided into five levels: less than or equal to 50 meters, greater than 50 meters and less than or equal to 200 meters, greater than 200 meters and less than or equal to 500 meters, greater than 500 meters and less than or equal to 1000 meters, and greater than 1000 meters. The first four levels are considered foggy, and the fifth level is considered fog-free. The forecast level is accurate if it is the same as the observation level. The forecast of foggy but the actual observation of foggy is considered as false alarm, and the forecast of fogless but the actual observation of foggy is considered as missed alarm. The inspection indicators are shown in the following table:

[0051]

[0052] The spline interpolation method is used to perform spatial interpolation on the visibility prediction data in the third step to obtain the visibility prediction results for each 10 minutes and each pile number along the measured section of the highway in the next 2 hours.

[0053] The fifth step is to obtain the visibility observation data at the time of arrival corresponding to the visibility prediction result from two automatic weather stations within 5 km around the measured section of the expressway, and obtain the visibility data inverted from the video image for each stake number at the time of arrival corresponding to the visibility prediction result from the measured section of the expressway. The screening conditions for filtering out invalid data for the obtained visibility observation data and the visibility data inverted from the video image are one of the following: ① The change in visibility between two consecutive minutes is not greater than 500m; ② The minimum visibility duration is not less than 3min; ③ The relative humidity is greater than 80%; ④ The wind speed is not greater than 4.5m / s.

[0054] After filtering out the invalid data in the above two sets of data, the visibility forecast results obtained in the fourth step are first corrected with the retained visibility observation data to obtain the corrected probability. The probability correction method is to use the Gmama probability function for correction. The probability function of the Gmama distribution is:

[0055]

[0056] Among them, Γ(α) is the Gmama function, where α is the shape parameter of Gmama, which indicates the shape of the distribution curve. The larger the value, the flatter the curve and the smaller the probability of small values. β is the scale parameter of the Gmama function. The larger the value, the larger the range of probability function and the more dispersed the distribution.

[0057] The estimation of α and β uses the maximum likelihood method:

[0058]

[0059] Among them, u is the mean of the sequence, σ 2 is the variance of the series.

[0060] Then, the visibility prediction results obtained in the fourth step are trend-corrected using the visibility data inverted from the video images at each stake number along the highway, and the visibility change trend is obtained. The trend correction method is to establish a statistical model:

[0061] y=Ax+B

[0062] Among them, y is the visibility value after trend correction, x is the visibility prediction result, and A and B are coefficients.

[0063] Finally, the revised probability and visibility change trend obtained by the above calculation are used to correct the subsequent newly obtained visibility prediction results respectively. The obtained result is the visibility forecast result for each pile number and each 10-minute interval along the measured section of the highway within the short-term approach time of 0-2 hours.

[0064] After probability correction, the visibility forecast result of the present invention reduces the missed reporting rate by 23% to 13.4% and the false reporting rate by 45% to 12.2%. After trend correction, the fog forecast accuracy below 1000 meters is 63.2%, and the root mean square error is reduced by 12 meters.

[0065] The visibility forecast results published in the article "Research on Low Visibility Classification in Hebei Province Based on CART Decision Tree" in Volume 38, Issue 3 of the Journal of Meteorology and Environment in June 2022 are: missed reporting rate is 51.4% in summer and 15.7% in winter; accuracy rate is 42.1% in summer and 49.8% in winter; false alarm rate is 24.1% in summer and 45.1% in winter.

[0066] By comparison, it can be seen that the present invention has obvious advantages in both false negative rate and accuracy rate.

Claims

1. A method for short-term forecasting and correction of highway visibility, characterized in that: The following steps are involved: S1. Obtain the 10-minute meteorological observation history data of the traffic meteorological observation stations along the measured section of the expressway for at least one year, including visibility, temperature, relative humidity, average wind speed in the past 2 minutes, temperature change in the past 24 hours, maximum temperature in the past 10 minutes, and minimum temperature in the past 10 minutes; among them, other factors except visibility constitute meteorological factors related to visibility prediction; S2. Calculate the Pearson correlation coefficient between visibility and various meteorological factors using correlation analysis method, and conduct significance test; select meteorological factors with α = 0.01 in the significance test, and use them as meteorological variables for visibility prediction; S3. Obtain hourly real-time meteorological observation data along the measured section of the expressway, introduce the visibility changes in 10 minutes, 1 hour, 3 hours and 6 hours, and use them as time series trend characteristic variables; introduce the minimum value and standard deviation of visibility in the last 6 hours, and use them as time series aggregation characteristic variables; S4. After standardizing the meteorological variables, time series trend characteristic variables, and time series aggregation characteristic variables, the three types of variables are used together as variables for visibility prediction to train the LSTM prediction model, and obtain the visibility prediction data for every 10 minutes in the next 2 hours at the location of the traffic meteorological observation point along the tested section of the expressway; use the spline interpolation method to perform spatial difference on the visibility prediction data, and obtain the visibility prediction results for every 10 minutes and pile number in the next 2 hours along the tested section of the expressway; S5. Obtain visibility observation data at the time of arrival corresponding to the visibility prediction result from the automatic weather station within 5 km around the measured section of the expressway, and obtain visibility data inverted from video images for each stake number at the time of arrival corresponding to the visibility prediction result from the measured section of the expressway. After filtering out invalid data in the above two groups of data, first use the retained visibility observation data to perform probability correction on the visibility prediction result to obtain the correction probability; then use the retained visibility data inverted from video images to perform trend correction on the visibility prediction result to obtain the visibility change trend; after using the correction probability and visibility change trend to correct the subsequent newly obtained visibility prediction results respectively, the visibility forecast results for each stake number and each 10-minute interval along the measured section of the expressway within the short-term proximity time limit of 0-2 hours are obtained.

2. The method for short-term forecasting and correction of highway visibility according to claim 1 is characterized in that: The Pearson correlation coefficient r between visibility and each meteorological factor in step S2 is calculated as follows: Among them, X and Y are two variables. and is the average value of the two groups of variables, and n is the number of meteorological factors.

3. The method for short-term forecasting and correction of highway visibility according to claim 1 is characterized in that: The significance test in step S2 is performed as follows: First calculate according to the following formula: Where t is the deviation statistic between the sample mean and the population mean, is the sample mean, μ is the population mean, σ Z is the sample standard deviation, m is the sample size; Then, by looking up the t value table, determine whether it passes the significance test of α=0.

01.

4. The method for short-term forecasting and correction of highway visibility according to claim 1 is characterized in that: The standardization of the three types of variables in step S4 is as follows: Among them, V is the standardized variable value, g is the number of variables, C g is the variable sequence of the g-th variable, is the mean value of the g-th variable.

5. The method for short-term forecasting and correction of highway visibility according to claim 1 is characterized in that: In step S5, the screening conditions for filtering out invalid data from the acquired visibility observation data and the visibility data inverted from the video image are one of the following: ① the change in visibility between two consecutive minutes is not greater than 500m; ② the minimum visibility duration is not less than 3min; ③ the relative humidity is greater than 80%; ④ the wind speed is not greater than 4.5m / s.

6. The method for short-term forecasting and correction of highway visibility according to claim 1 is characterized in that: The method for making probability correction to the visibility prediction result in step S5 is to use the Gmama probability function for correction.

7. The method for short-term forecasting and correction of highway visibility according to claim 1 is characterized in that: The method for correcting the trend of the visibility prediction result in step S5 is to use the visibility data inverted from the video image to correct the trend of the visibility prediction result. The specific method is to establish a statistical model: y=Ax+B Among them, y is the visibility value after trend correction, x is the visibility prediction result, and A and B are coefficients.

Images