CN110929776B - Remote sensing wind field data quality evaluation method and device based on sea surface wind field stability statistical zoning - Google Patents

Abstract

The invention discloses a wind field remote sensing data quality evaluation method and device based on sea surface wind field stability statistics division, comprising the following steps: establishing a wind speed interval division model and a wind direction stable grade classification model; carrying out space-time matching on the obtained actually measured wind field data and the wind field remote sensing data; calculating a wind direction stability coefficient of actually measured wind field data; carrying out wind speed interval division on the actually measured wind field data by using a wind speed interval division model; utilizing a wind direction stability grade classification model to perform wind direction stability grade classification on actually measured wind field data; and for the measured wind field data and the corresponding wind field remote sensing data contained in each wind direction stability grade, performing quality evaluation on the wind field remote sensing data according to the correlation between the measured wind field data and the wind field remote sensing data. The method and the device accurately evaluate the quality of the wind field remote sensing data.

Description

A remote sensing wind field data quality assessment method and device based on sea surface wind field stability statistical zoning

Technical Field

The present invention relates to the field of marine information technology, and specifically to a remote sensing wind field data quality assessment method and device based on sea surface wind field stability statistical zoning.

Background Art

There are significant differences in time and space scales between the sea surface wind field information obtained by microwave remote sensing sensors and the real data observed at actual buoys and other points, which is very unfavorable for the quantitative application of remote sensing inversion information. The main sources of errors and uncertainties in remote sensing products include data source errors, model errors, and errors caused by variations in the time window and spatial scale (temporal and spatial variations) during the verification data matching process. The process of evaluating the quality of data products output by the system through independent methods is called authenticity testing. Its main task is to use field monitoring data and extrapolation models to determine the errors of geophysical data products. The monitoring content includes atmospheric conditions, optical properties of the water-air interface and sub-layer water surface, sea surface meteorological dynamics, and the geophysical quantities to be detected.

Authenticity verification can effectively evaluate the product quality of remote sensing data, thereby improving the reliability of satellite remote sensing data. The international community has always attached great importance to the authenticity verification of ocean remote sensing satellites. As early as 1984, the Committee on Earth Observation Satellites established a calibration and authenticity verification working group to coordinate the authenticity verification of remote sensing satellites in various countries. NASA's Ocean Biology Processing Group (OBPG) also uses global data to carry out continuous authenticity verification during the life cycle of the satellite, and has achieved many beneficial results in the accuracy assessment of ocean remote sensing products, the long-term stability assessment of satellite measurements, and the accuracy verification of satellite in-orbit calibration.

Satellite products and field measured data have different spatiotemporal sampling characteristics. It is necessary to determine a reasonable spatiotemporal matching window for field-measured remote sensing data based on the spatial resolution of satellite products and the spatiotemporal variation and uniformity of water bodies. The internationally accepted principle for determining the spatiotemporal window is: spatial window 3×3 or 5×5 pixels, time window ±3h. However, due to the limitations of remote sensing effective pixel data and sea conditions, this spatiotemporal matching principle is often difficult to strictly implement. Even if the above spatiotemporal matching principle is met, due to the seasonal and regional variation characteristics of the ocean, especially the high dynamic variation characteristics of the wind field, the accuracy verification results are different in different seasons and regions. The credibility and representativeness of the accuracy verification results need to be further evaluated.

Therefore, how to scientifically and effectively select the time and space matching rules for measured and remote sensing data, analyze the representative characteristics of the accuracy evaluation data set, and improve the credibility of the accuracy evaluation results is one of the key scientific issues that need to be solved.

Summary of the invention

The purpose of the present invention is to provide a method and device for evaluating the quality of wind field remote sensing data based on the statistical zoning of sea surface wind field stability, which method and device can accurately evaluate the quality of wind field remote sensing data.

In order to achieve the above-mentioned invention object, the following technical solutions are provided:

A method for evaluating the quality of wind field remote sensing data based on the statistical zoning of sea surface wind field stability, the method comprising the following steps:

Establish a wind speed interval division model based on wind speed;

Determine the wind direction stability coefficient based on historical wind field data, and establish a wind direction stability grade classification model based on the wind direction stability coefficient;

Obtaining measured wind field data and wind field remote sensing data, and performing spatiotemporal matching on the obtained measured wind field data and wind field remote sensing data to obtain the corresponding relationship between the measured wind field data and the wind field remote sensing data;

Calculate the wind direction stability coefficient of measured wind field data;

The wind speed interval division model is used to divide the measured wind field data into wind speed intervals to determine the wind speed interval to which the remote sensing wind field data belongs;

According to the measured wind field data and its wind direction stability coefficient in each wind speed range, the wind direction stability classification model is used to classify the measured wind field data into wind direction stability grades to determine the wind direction stability grade to which the remote sensing wind field data belongs.

For the measured wind field data and the corresponding wind field remote sensing data contained in each wind direction stability level, the quality of the wind field remote sensing data is evaluated according to the error indicators and correlation between the measured wind field data and the wind field remote sensing data.

A wind farm remote sensing data quality assessment device based on sea surface wind farm stability statistical zoning includes a computer memory, a computer processor, and a computer program stored in the computer memory and executable on the computer processor. The computer memory also stores a wind speed interval division model and a wind direction stability grade classification model. When the computer processor executes the computer program, the following steps are implemented:

Obtaining measured wind field data and wind field remote sensing data, and performing spatiotemporal matching on the obtained measured wind field data and wind field remote sensing data to obtain the corresponding relationship between the measured wind field data and the wind field remote sensing data;

Calculate the wind direction stability coefficient of measured wind field data;

Calling the wind speed interval division model to divide the measured wind field data into wind speed intervals to determine the wind speed interval to which the remote sensing wind field data belongs;

According to the measured wind field data and its wind direction stability coefficient in each wind speed range, the wind direction stability classification model is called to classify the measured wind field data into wind direction stability grades to determine the wind direction stability grade to which the remote sensing wind field data belongs;

For the measured wind field data and the corresponding wind field remote sensing data contained in each wind direction stability level, the quality of the wind field remote sensing data is evaluated according to the correlation between the measured wind field data and the wind field remote sensing data.

Compared with the prior art, the present invention has the following beneficial effects:

The present invention starts from the actual problem of the inversion mechanism of wind field remote sensing data and the influence of spatiotemporal variation on verification error, and proposes a comprehensive evaluation method and device for wind field remote sensing data by integrating wind speed and wind direction stability grade zoning. The method and device can accurately evaluate the quality of wind field remote sensing data, and users can scientifically screen and apply remote sensing wind field data according to the evaluation results. The implementation process is feasible, the calculation process is easy to integrate with the program, and the calculation result is accurate and reliable, which has high industry promotion value and application prospects.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

1 is a flow chart of a method for evaluating the quality of wind farm remote sensing data based on statistical zoning of sea surface wind farm stability provided by an embodiment;

FIG2 is a schematic diagram of calculating the temporal wind field stability provided by an embodiment;

FIG3 is a schematic diagram of calculating the stability of a spatial wind field provided by an embodiment;

FIG4 is a schematic diagram of a spatial zoning result of wind direction stability provided by an embodiment;

FIG. 5 is a scatter plot of wind farm data grade assessment provided by an embodiment.

FIG. 6 is a comparison diagram of evaluation results for different wind speed intervals provided in an embodiment.

DETAILED DESCRIPTION

In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific implementation methods described herein are only used to explain the present invention and do not limit the scope of protection of the present invention.

Figure 1 is a flow chart of a method for evaluating the quality of wind farm remote sensing data based on the statistical zoning of wind farm stability on the sea surface provided by an embodiment. As shown in Figure 1, the method for evaluating the quality of wind farm remote sensing data based on the statistical zoning of wind farm stability on the sea surface includes the following steps:

S101, establishing a wind speed interval division model according to the wind speed.

Since different wind speeds cause different sea surface roughness, the interference degree to satellite reception signals is also different. The quality assessment of remote sensing wind field data in different wind speed intervals should be treated differently. Therefore, it is necessary to divide the wind speed into intervals. The wind speed interval division model provided in this embodiment is mainly used to determine the wind speed interval to which the wind speed belongs according to the size of the wind speed. Specifically, the wind speed interval division model includes the mapping relationship between the wind speed size and the wind speed interval, as follows:

Wind speed Wind speed range 0～3m/s The first wind speed interval is denoted as Lsp1 3～15m/s The second wind speed interval is denoted as Lsp2 15m/s The third wind speed interval is denoted as Lsp3

S102, determining a wind direction stability coefficient based on historical wind field data, and establishing a wind direction stability grade classification model according to the wind direction stability coefficient.

Wind field data includes wind speed data and wind direction data. When constructing a wind direction stability grade classification model, historical wind field data is used as a data source, that is, the wind direction stability coefficient of the historical wind field data is calculated. In the embodiment, the Wind Sat wind field historical data set is selected to calculate the multi-year spatiotemporal wind direction stability coefficient.

Specifically, the method for determining the wind direction stability coefficient includes:

Calculate the first wind direction stability coefficient of wind field data in time series;

Calculate the second wind direction stability coefficient of wind field data on the spatial grid;

The first wind direction stability coefficient and the second wind direction stability coefficient are merged to determine the wind direction stability coefficient.

As shown in Figure 2, the calculation method of the first wind direction stability coefficient of wind field data in time series is:

As shown in Figure 3, the calculation method of the second wind direction stability coefficient of wind field data on the spatial grid is:

The calculation method for integrating the first wind direction stability coefficient and the second wind direction stability coefficient is:

Among them, S(t) is the first wind direction stability coefficient, u _i , _vi are the longitudinal wind and zonal wind respectively, i is the index of the wind field data sample, t is the total number of wind field data samples, S(j) is the second wind direction stability coefficient, u _j , v _j are the zonal wind and longitudinal wind in the m*m grid window respectively, j is the index of the grid window, and S is the wind direction stability coefficient.

After obtaining the wind direction stability coefficient, a wind direction stability level classification model can be constructed based on the wind direction stability coefficient. When determining the wind direction stability level in different time periods, the grid size of the historical data must be consistent with the grid size of the validation data set. If the grid sizes are inconsistent, a resampling method is required to make the size of the historical grid data validation data consistent. The wind direction stability level classification model contains the mapping relationship between the wind direction stability coefficient and the wind direction stability level, as follows:

Figure 4 is a schematic diagram of the spatial zoning results of wind direction stability.

S103, obtaining measured wind field data and wind field remote sensing data, and performing time-space matching on the obtained measured wind field data and wind field remote sensing data to obtain a corresponding relationship between the measured wind field data and the wind field remote sensing data.

S104, calculating the wind direction stability coefficient of the measured wind field data.

The calculation method of the wind direction stability coefficient of the measured wind field data in S104 is the same as the calculation method of the wind direction stability coefficient in S103, and will not be repeated here.

S105, dividing the measured wind field data into wind speed intervals using a wind speed interval division model to determine the wind speed interval to which the remote sensing wind field data belongs.

S106, for the measured wind field data and the wind direction stability coefficient in each wind speed interval, the wind direction stability classification model is used to classify the measured wind field data into wind direction stability grades to determine the wind direction stability grade to which the remote sensing wind field data belongs.

S107, for the measured wind field data and the corresponding wind field remote sensing data included in each wind direction stability level, quality assessment is performed on the wind field remote sensing data according to the correlation between the measured wind field data and the wind field remote sensing data.

In the embodiment, the quality assessment of the wind farm remote sensing data based on the correlation between the measured wind farm data and the wind farm remote sensing data includes:

Calculate the average absolute relative error, root mean square error, average relative deviation and correlation coefficient between the measured wind field data and the wind field remote sensing data:

和

分别为风场遥感数据和实测风场数据的算术平均值；Among them, RE is the mean absolute relative error, RMSE is the root mean square error, BLAS is the average relative deviation, γ is the correlation coefficient, the value range of γ is between -1 and +1, M is the number of data, _ri is the wind field remote sensing data, _si is the measured wind field data,

and

are the arithmetic mean values of wind field remote sensing data and measured wind field data respectively;

The mean absolute relative error, root mean square error, mean relative deviation and correlation coefficient are used as four evaluation indicators, and the quality of wind farm remote sensing data is determined by combining the four evaluation indicators.

The embodiment also provides a wind farm remote sensing data quality assessment device based on the statistical zoning of sea surface wind farm stability, comprising a computer memory, a computer processor, and a computer program stored in the computer memory and executable on the computer processor, wherein the computer memory also stores a wind speed interval division model and a wind direction stability grade classification model, and the computer processor implements the following steps when executing the computer program:

Obtaining measured wind field data and wind field remote sensing data, and performing spatiotemporal matching on the obtained measured wind field data and wind field remote sensing data to obtain the corresponding relationship between the measured wind field data and the wind field remote sensing data;

Calculate the wind direction stability coefficient of measured wind field data;

Calling the wind speed interval division model to divide the measured wind field data into wind speed intervals to determine the wind speed interval to which the remote sensing wind field data belongs;

According to the measured wind field data and its wind direction stability coefficient in each wind speed range, the wind direction stability classification model is called to classify the measured wind field data into wind direction stability grades to determine the wind direction stability grade to which the remote sensing wind field data belongs;

For the measured wind field data and the corresponding wind field remote sensing data contained in each wind direction stability level, the quality of the wind field remote sensing data is evaluated according to the correlation between the measured wind field data and the wind field remote sensing data.

In the wind field remote sensing data quality assessment device, the wind speed interval division model is established according to the wind speed size, and the wind speed interval division model includes a mapping relationship between the wind speed size and the wind speed interval, as follows:

Wind speed Wind speed range 0～3m/s The first wind speed interval is denoted as Lsp1 3～15m/s The second wind speed interval is denoted as Lsp2 More than 15m/s The third wind speed interval is denoted as Lsp3

In the wind field remote sensing data quality assessment device, the construction method of the wind direction stability grade classification model is: determine the wind direction stability coefficient based on historical wind field data, and establish the wind direction stability grade classification model according to the wind direction stability coefficient.

In the wind field remote sensing data quality assessment device, the construction method of the wind speed interval division model and the wind direction stability grade classification model is the same as the construction method of the wind speed interval division model and the wind direction stability grade classification model in the above-mentioned wind field remote sensing data quality assessment method based on the statistical zoning of sea surface wind field stability, and will not be repeated here.

The above method and device can accurately evaluate the quality of wind field remote sensing data, and users can scientifically screen and apply the remote sensing wind field data based on the evaluation results.

Experimental example

In the experimental example, the measured wind speed and direction data are the historical data of the National Data Buoy Center (NDBC) of the United States (the website information is as follows: https://www.ndbc.noaa.gov). The remote sensing data are the microwave remote sensing data of the Wind Sat sea surface wind field provided by the Earth Microwave Data Center (the website information is as follows: http://www.remss.com/missions/windsat/). The microwave remote sensing data of the Wind Sat sea surface wind field to be evaluated are the daily wind speed and direction data in 2014, and the historical remote sensing data are the microwave remote sensing data of the Wind Sat sea surface wind field from 2004 to 2013. The verification data set obtained by the spatiotemporal matching of the measured wind field data and the wind field remote sensing data is 1328 pairs.

The wind speed interval division model is used to divide 1328 pairs of wind field data. The division results are shown in Table 1:

Table 1

Wind speed range 0～3m/s 3～15m/s >15m/s Number of data pairs 122 776 430

Then, the wind direction stability level is divided according to the measured wind field data and wind direction stability coefficient in each wind speed range. Figure 5 is a scatter plot of the verification data level evaluation in the experimental example.

Finally, for the measured wind field data and the corresponding wind field remote sensing data contained in each wind direction stability level, the quality of the wind field remote sensing data is evaluated based on the correlation between the measured wind field data and the wind field remote sensing data.

The statistical results of wind direction are: the RE of Level 1 is 18.36%, the largest is Level 2, the result is 20.54%, and the overall RE is 19.59. The RMSE of Level 1 is 27.71, and the largest is Level 2, the result is 33.09%, and the overall RMSE is 31.14. The statistical results of wind speed are: the RE of Level 1 is 19.22%, the largest is Level 2, the result is 23.70%, and the overall RE is 19.88. The minimum RMSE is Level 2, the result is 1.59, the maximum is Level 3, the result is 1.97, and the overall RMSE is 1.74.

Figure 6 shows the results of the matching analysis of measured and remote sensing data in different wind speed ranges. It can be seen from the figure that the data test accuracy is the best when the wind speed is 3-15m/s and the zoning level is Level 2, with small RE and BIAS and the largest correlation coefficient. Therefore, using the variation level zoning results to evaluate the representativeness of the validation data set can more effectively and comprehensively evaluate the accuracy results of remote sensing products, reveal the error sources of remote sensing products, and provide guidance for the optimization of the spatiotemporal strategy of measured sample points (navigation, buoys).

The specific implementation methods described above provide a detailed description of the technical solutions and beneficial effects of the present invention. It should be understood that the above is only the most preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, supplements and equivalent substitutions made within the scope of the principles of the present invention should be included in the protection scope of the present invention.

Claims (8)

1. A method for evaluating the quality of wind field remote sensing data based on the statistical zoning of sea surface wind field stability, characterized in that the method comprises the following steps: Establish a wind speed interval division model based on wind speed; Determine the wind direction stability coefficient based on historical wind field data, and establish a wind direction stability grade classification model based on the wind direction stability coefficient; Obtaining measured wind field data and wind field remote sensing data, and performing spatiotemporal matching on the obtained measured wind field data and wind field remote sensing data to obtain the corresponding relationship between the measured wind field data and the wind field remote sensing data; Calculating the wind direction stability coefficient of the measured wind field data, including: calculating a first wind direction stability coefficient of the wind field data in a time series; calculating a second wind direction stability coefficient of the wind field data in a spatial grid; fusing the first wind direction stability coefficient and the second wind direction stability coefficient to determine the wind direction stability coefficient; The wind speed interval division model is used to divide the measured wind field data into wind speed intervals to determine the wind speed interval to which the remote sensing wind field data belongs; According to the measured wind field data and wind direction stability coefficient in each wind speed range, the wind direction stability classification model is used to classify the measured wind field data into wind direction stability grades to determine the wind direction stability grade to which the remote sensing wind field data belongs. For the measured wind field data and the corresponding wind field remote sensing data contained in each wind direction stability level, the quality of the wind field remote sensing data is evaluated according to the error indicators and correlation between the measured wind field data and the wind field remote sensing data. 2. The method for evaluating the quality of wind field remote sensing data based on the statistical zoning of sea surface wind field stability according to claim 1, characterized in that the wind speed interval division model includes a mapping relationship between wind speed magnitude and wind speed interval, as follows: Wind speed Wind speed range 0～3m/s The first wind speed interval is denoted as Lsp1 3～15m/s The second wind speed interval is denoted as Lsp2 More than 15m/s The third wind speed interval is denoted as Lsp3 . 3. The method for evaluating the quality of wind field remote sensing data based on the statistical zoning of sea surface wind field stability according to claim 1, characterized in that the calculation method of the first wind direction stability coefficient of the wind field data in the time series is:

The calculation method of the second wind direction stability coefficient of the wind field data on the spatial grid is:

The calculation method for fusing the first wind direction stability coefficient and the second wind direction stability coefficient is:

Among them, S(t) is the first wind direction stability coefficient, u _i , _vi are the longitudinal wind and zonal wind respectively, i is the index of the wind field data sample, t is the total number of wind field data samples, S(j) is the second wind direction stability coefficient, u _j , v _j are the zonal wind and longitudinal wind in the m*m grid window respectively, j is the index of the grid window, and S is the wind direction stability coefficient. 4. The method for evaluating the quality of wind field remote sensing data based on the statistical zoning of sea surface wind field stability according to claim 1, characterized in that the wind direction stability grade classification model comprises a mapping relationship between the wind direction stability coefficient and the wind direction stability grade, as follows: Wind direction stability coefficient Wind direction stability level 0～30 Wind direction stability level I 30～70 Wind direction stability level II 70～100 Wind direction stability level III . 5. The method for evaluating the quality of wind farm remote sensing data based on the statistical zoning of sea surface wind farm stability according to claim 1, characterized in that the quality evaluation of wind farm remote sensing data based on the correlation between the measured wind farm data and the wind farm remote sensing data comprises: Calculate the average absolute relative error, root mean square error, average relative deviation and correlation coefficient between the measured wind field data and the wind field remote sensing data:

Among them, RE is the mean absolute relative error, RMSE is the root mean square error, BLAS is the average relative deviation, γ is the correlation coefficient, the value range of γ is between -1 and +1, M is the number of data, _ri is the wind field remote sensing data, _si is the measured wind field data,

and

are the arithmetic mean values of wind field remote sensing data and measured wind field data respectively; The mean absolute relative error, root mean square error, mean relative deviation and correlation coefficient are used as four evaluation indicators, and the quality of wind farm remote sensing data is determined by combining the four evaluation indicators. 6. A wind farm remote sensing data quality assessment device based on sea surface wind farm stability statistical zoning, comprising a computer memory, a computer processor, and a computer program stored in the computer memory and executable on the computer processor, characterized in that the computer memory also stores a wind speed interval division model and a wind direction stability grade classification model, and the computer processor implements the following steps when executing the computer program: Obtaining measured wind field data and wind field remote sensing data, and performing spatiotemporal matching on the obtained measured wind field data and wind field remote sensing data to obtain the corresponding relationship between the measured wind field data and the wind field remote sensing data; Calculating the wind direction stability coefficient of the measured wind field data, including calculating a first wind direction stability coefficient of the wind field data in a time series; calculating a second wind direction stability coefficient of the wind field data in a spatial grid; fusing the first wind direction stability coefficient and the second wind direction stability coefficient to determine the wind direction stability coefficient; Calling the wind speed interval division model to divide the measured wind field data into wind speed intervals to determine the wind speed interval to which the remote sensing wind field data belongs; According to the measured wind field data and its wind direction stability coefficient in each wind speed range, the wind direction stability classification model is called to classify the measured wind field data into wind direction stability grades to determine the wind direction stability grade to which the remote sensing wind field data belongs; For the measured wind field data and the corresponding wind field remote sensing data contained in each wind direction stability level, the quality of the wind field remote sensing data is evaluated according to the correlation between the measured wind field data and the wind field remote sensing data. 7. The wind field remote sensing data quality assessment device based on sea surface wind field stability statistical zoning according to claim 6, characterized in that the wind speed interval division model is established according to the wind speed size, and the wind speed interval division model includes a mapping relationship between the wind speed size and the wind speed interval, as follows: Wind speed Wind speed range 0～3m/s The first wind speed interval is denoted as Lsp1 3～15m/s The second wind speed interval is denoted as Lsp2 More than 15m/s The third wind speed interval is denoted as Lsp3 . 8. The wind field remote sensing data quality assessment device based on sea surface wind field stability statistical zoning according to claim 6, characterized in that the wind direction stability grade classification model is constructed by: The wind direction stability coefficient is determined based on historical wind field data, and a wind direction stability grade classification model is established according to the wind direction stability coefficient.

Images