Meteorological data grid analysis method and system based on big data and machine learning

Abstract

The invention discloses a meteorological data grid analysis method and system based on big data and machine learning, and belongs to the technical field of meteorological data analysis, and the method comprises the steps: constructing multi-source fusion meteorological data based on multi-source meteorological data; gridding the multi-source fusion meteorological data through a three-dimensional grid model considering horizontal geographic features and vertical atmospheric layering features to obtain corresponding grid features; performing feature extraction on the multi-source fusion meteorological data to obtain corresponding meteorological features and time sequence features; and according to the grid characteristics, the meteorological characteristics and the time sequence characteristics, performing analysis through a meteorological data grid analysis model to obtain an analysis result, and outputting the analysis result to a terminal. According to the scheme, data gridding is carried out through multi-source fusion meteorological data in combination with the geographic features and the atmosphere layering features, and grid analysis is carried out in combination with the meteorological features and the time sequence features, so that the reliability of meteorological data analysis can be remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of meteorological data analysis, and in particular relates to a meteorological data grid analysis method and system based on big data and machine learning. Background Art

[0002] With the intensification of global climate change and the frequent occurrence of meteorological disasters, the accuracy and real-time requirements of meteorological data are becoming higher and higher. Meteorological data not only comes from traditional ground observation stations, high-altitude detection, radar, etc., but also covers a variety of data sources such as satellite remote sensing and numerical weather forecast models. These data sources have their own characteristics and contain rich meteorological information, but there are also problems such as inconsistent data formats and uneven data quality. Therefore, how to effectively integrate multi-source meteorological data and improve the accuracy and integrity of data has become an urgent problem to be solved in the meteorological field.

[0003] Most existing meteorological data analyses use a single data source or simple data fusion technology, ignoring the weight differences and data complementarity between different data sources, resulting in limited accuracy and integrity of meteorological data. In addition, most existing meteorological data gridding methods only consider horizontal geographic features, ignoring the impact of vertical atmospheric stratification characteristics on meteorological data, making the spatial distribution expression of meteorological data not detailed and accurate enough. At the same time, in terms of feature extraction, existing technologies often only focus on the basic attributes and spatiotemporal distribution laws of meteorological data, ignoring the correlation and time series characteristics between meteorological elements, thereby limiting the accuracy and timeliness of meteorological forecasts and disaster warnings. These shortcomings limit the application effect of meteorological data analysis in meteorological forecasts and disaster warnings, and the reliability of grid analysis of complex meteorological data is low. Summary of the invention

[0004] The purpose of the present invention is to solve the problem of low reliability of meteorological data grid analysis methods in the prior art, and to provide a meteorological data grid analysis method and system based on big data and machine learning. The method and system fuse multi-source meteorological data, grid the data using a three-dimensional grid model that considers geographical features and atmospheric stratification features, and then performs grid analysis in combination with meteorological features and time series features, thereby significantly improving the reliability of grid analysis of complex meteorological data.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions: According to a first aspect of the present invention, a meteorological data grid analysis method based on big data and machine learning is provided, comprising: Construct multi-source fusion meteorological data based on multi-source meteorological data; The multi-source fusion meteorological data is gridded by a three-dimensional grid model that considers horizontal geographical features and vertical atmospheric stratification features to obtain corresponding grid features; Extracting features from the multi-source fused meteorological data to obtain corresponding meteorological features and time series features; According to the grid characteristics, the meteorological characteristics and the time series characteristics, an analysis is performed through a meteorological data grid analysis model to obtain an analysis result and output it to a terminal.

[0006] Furthermore, the multi-source fusion meteorological data is constructed based on the multi-source meteorological data, including: Collect meteorological data from different data sources; Preprocessing the meteorological data from different data sources, wherein the preprocessing at least includes data cleaning, data standardization and format conversion; Considering the weights of different data sources, the pre-processed meteorological data from different data sources are fused by Kalman filtering to obtain multi-source fused meteorological data.

[0007] Furthermore, the three-dimensional grid model considering horizontal geographical features and vertical atmospheric stratification features is used to grid the multi-source fusion meteorological data to obtain corresponding grid features, including: Determine the horizontal geographic features and vertical atmospheric stratification features corresponding to the multi-source fusion meteorological data, and construct a three-dimensional grid model based on the horizontal geographic features and vertical atmospheric stratification features; The multi-source fusion meteorological data is interpolated into the three-dimensional grid model to obtain an interpolation result, and grid features of the multi-source fusion meteorological data are obtained according to the interpolation result.

[0008] Furthermore, the determining of horizontal geographic features and vertical atmospheric stratification features corresponding to the multi-source fusion meteorological data, and constructing a three-dimensional grid model based on the horizontal geographic features and vertical atmospheric stratification features, includes: Extracting geographic features of the multi-source fusion meteorological data in the horizontal direction, and setting the horizontal resolution according to the geographic features; Extracting the atmospheric stratification characteristics of the multi-source fusion meteorological data in the vertical direction and determining the vertical resolution of different atmospheric layers; Inputting the horizontal resolution and the vertical resolution into a geographic information system, and performing grid division in the horizontal direction and the vertical direction based on the geographic information system to obtain a plurality of grid units; Geographic coordinates and vertical heights are assigned to each of the grid cells to obtain a three-dimensional grid model.

[0009] Furthermore, interpolating the multi-source fused meteorological data into the three-dimensional grid model to obtain an interpolation result includes: Determine the meteorological element value of a known point and its corresponding longitude, latitude and altitude coordinates based on the multi-source fusion meteorological data; Determine the coordinates of unknown points that need to be interpolated on the three-dimensional network model; Calculate the spatial distance between each unknown point and the known point, and determine the weight according to the inverse distance weighted interpolation principle; The meteorological element values ​​of the unknown points are obtained by performing weighted average calculation on the meteorological element values ​​of the known points and the weights.

[0010] Furthermore, the calculation formula of the meteorological element value is as follows: ; Where: Indicates the meteorological element value of the unknown point, Represents the meteorological element value of a known point, represents the distance between the unknown point and the known point, p represents the weight coefficient, and n represents the number of known points.

[0011] Furthermore, the feature extraction of the multi-source fusion meteorological data to obtain corresponding meteorological features and time series features includes: Determine the basic meteorological elements of the multi-source fusion meteorological data; Extracting the spatiotemporal distribution characteristics of the basic meteorological elements, and extracting composite meteorological characteristics by considering the correlation between the basic meteorological elements; The time series characteristics of the multi-source fusion meteorological data are analyzed by time series analysis method to obtain corresponding time series features.

[0012] Furthermore, the time series characteristics of the multi-source fusion meteorological data are analyzed by a time series analysis method to obtain corresponding time series features, including: Extracting the timestamp corresponding to the multi-source fusion meteorological data; Annotating the multi-source fusion meteorological data based on the timestamp; Constructing a time series model based on the multi-source fusion meteorological data with timestamp annotation; The time series features corresponding to the multi-source fusion meteorological data are generated based on the time series model.

[0013] Further, the analysis results are obtained by analyzing the grid characteristics, the meteorological characteristics and the time series characteristics through a meteorological data grid analysis model, including: Acquire historical meteorological data and label data corresponding to the historical meteorological data; Training a vector machine model by combining the historical meteorological data with the label data to obtain a meteorological data grid analysis model; Inputting the grid characteristics, the meteorological characteristics and the time series characteristics into the meteorological data grid analysis model for analysis to obtain analysis results; The analysis results are output to the terminal in the form of graphs.

[0014] According to another aspect of the present invention, a meteorological data grid analysis system based on big data and machine learning is provided, comprising: A data acquisition module is used to construct multi-source fusion meteorological data based on multi-source meteorological data; A gridding module, used for gridding the multi-source fusion meteorological data by taking into account horizontal geographical features and vertical atmospheric stratification features to obtain corresponding grid features; A feature extraction module is used to extract features from the multi-source fusion meteorological data to obtain corresponding meteorological features and time series features; The grid analysis module is used to analyze the grid characteristics, the meteorological characteristics and the time series characteristics through the meteorological data grid analysis model to obtain the analysis results and output them to the terminal.

[0015] The beneficial effects of the present invention are as follows: by fusing multi-source meteorological data and using the Kalman filter method to perform weight distribution and data fusion on different data sources, the accuracy and integrity of meteorological data are significantly improved. Not only the diversity and complementarity of data are taken into account, but also the optimization and integration of data are realized through scientific algorithms, laying a solid foundation for subsequent meteorological analysis. By introducing a three-dimensional grid model to grid the meteorological data, the horizontal geographical features and the vertical atmospheric stratification features are comprehensively considered, so that the spatial distribution of meteorological data is more refined and accurate. Through grid processing, meteorological data can be more carefully divided and expressed in three-dimensional space, providing more abundant information for meteorological analysis. Through feature extraction technology, meteorological features and time series features are extracted from multi-source fusion meteorological data, which not only reflects the basic attributes and spatiotemporal distribution laws of meteorological data, but also reveals the correlation and time series characteristics between meteorological elements, providing strong support for meteorological forecasting and disaster warning. Using the trained meteorological data grid analysis model, combined with grid features, meteorological features and time series features for comprehensive analysis, accurate meteorological analysis results can be output, improving the efficiency and accuracy of meteorological analysis.

[0016] The above invention content is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Other features, objects and advantages of the present invention will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings. The drawings are only for the purpose of illustrating preferred embodiments and are not to be considered as limiting the present invention. Also, the same reference symbols are used throughout the drawings to represent the same parts.

[0018] Figure 1 A flowchart of a meteorological data grid analysis method based on big data and machine learning disclosed in an embodiment of the present invention; Figure 2 A schematic diagram of the structure of a meteorological data grid analysis system based on big data and machine learning disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to make the objectives, technical solutions 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 method described herein is only an optimal embodiment of the present invention, which is only used to explain the present invention and does not limit the scope of protection of the present invention. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0020] Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flow charts. Although the flow charts describe the operations (or steps) as sequential processes, many of the operations (or steps) therein may be performed in parallel, concurrently, or simultaneously. In addition, the order of the operations may be rearranged. A process may be terminated when its operations are completed, but may also have additional steps not included in the accompanying drawings; a process may correspond to a method, function, procedure, subroutine, subprogram, etc.

[0021] Example: According to a first aspect of this embodiment, a meteorological data grid analysis method based on big data and machine learning is provided, such as Figure 1 As shown, including: S101, constructing multi-source fusion meteorological data based on multi-source meteorological data; Furthermore, multi-source fusion meteorological data is constructed based on multi-source meteorological data, including: Collect meteorological data from different data sources; Preprocess meteorological data from different data sources, which at least includes data cleaning, data standardization and format conversion; Considering the weights of different data sources, the preprocessed meteorological data from different data sources are fused through the Kalman filtering method to obtain multi-source fused meteorological data.

[0022] Specifically, multiple meteorological data sources are selected, such as data from the Meteorological Bureau, data from commercial meteorological service companies, and data collected in real time through IoT devices (such as weather stations, sensors, etc.). Meteorological data is obtained from the above data sources through API interfaces, database queries, file transfers, etc. Then the obtained meteorological data is preprocessed. First, duplicate records are checked and deleted to ensure the uniqueness of the data set. For missing data, a method of filling (such as using the average value, median, etc. of the data before and after) is selected for processing according to the characteristics and importance of the data. Outliers are identified and removed through statistical methods. Meteorological data of the same type (such as temperature, humidity, etc.) in different data sources are converted to the same dimension and unit. The data is normalized so that the data value is between 0-1. The data formats provided by different data sources are converted to a unified format, such as CSV, JSON, etc. Combining multiple factors such as the credibility, accuracy, and real-time of different data sources, a weight is assigned to each data source. The weight is determined by expert scoring combined with historical data verification. Then the Kalman filter is initialized, and parameters such as the initial state vector, state covariance matrix, and observation noise covariance matrix are set. According to the state vector and state transfer matrix of the previous time step, the state vector and state covariance matrix of the current time step are predicted. The state vector and state covariance matrix are updated using the observation value and observation noise covariance matrix of the current time step. Repeat the above prediction and update steps until the data of all time steps are processed to obtain multi-source fusion meteorological data. Multi-source fusion meteorological data combines information from multiple data sources.

[0023] In the above implementation, data cleaning can remove errors and outliers and improve the accuracy of data; data standardization can unify data from different data sources to the same scale, which is convenient for subsequent data processing and analysis; format conversion ensures that data from different data sources can be compatible and integrated with each other. The Kalman filter method can be used to optimally estimate the pre-processed meteorological data according to the weights and uncertainties of different data sources. It can make full use of the advantages of each data source while reducing the influence of noise and errors, so as to obtain more accurate and reliable multi-source fusion meteorological data. In the process of data fusion, the weights of different data sources are determined according to factors such as their accuracy, reliability and timeliness. By setting the weights reasonably, it can be ensured that the fused data can reflect the true situation of each data source while avoiding conflicts and contradictions between data. Multi-source fusion meteorological data has higher timeliness and accuracy, can more accurately reflect the current meteorological conditions and future trends, and help improve the accuracy and timeliness of weather forecasts.

[0024] S103, gridding the multi-source fusion meteorological data by a three-dimensional grid model that considers horizontal geographical features and vertical atmospheric stratification features to obtain corresponding grid features; Furthermore, the multi-source fusion meteorological data is gridded by a three-dimensional grid model that considers horizontal geographical features and vertical atmospheric stratification features to obtain corresponding grid features, including: Determine the horizontal geographic features and vertical atmospheric stratification features corresponding to the multi-source fusion meteorological data, and construct a three-dimensional grid model based on the horizontal geographic features and vertical atmospheric stratification features; The multi-source fusion meteorological data is interpolated into a three-dimensional grid model to obtain an interpolation result, and the grid characteristics of the multi-source fusion meteorological data are obtained according to the interpolation result.

[0025] Specifically, identify the main geographical features of the target area, such as mountains, rivers, cities, etc. Use geographic information system (GIS) data to determine the latitude and longitude range, terrain undulations, landform types, etc. of the target area. According to the resolution and accuracy requirements of meteorological data, divide the target area into several horizontal grid units, each with the same latitude and longitude range. Analyze the atmospheric structure of the target area and determine the atmospheric stratification in the vertical direction, including the troposphere, mesosphere, and stratosphere. According to meteorological principles and actual needs, divide the vertical direction into several altitude layers, each with a specific pressure range or altitude interval. Determine the change law and characteristics of meteorological elements (such as temperature, humidity, wind speed, etc.) in each altitude layer. Design a three-dimensional grid model by combining horizontal geographical features and vertical atmospheric stratification characteristics. The grid model consists of horizontal grid units and vertical altitude layers to form grid nodes in three-dimensional space. Determine the coordinates of the grid nodes, including longitude, latitude, and altitude. Interpolate multi-source fusion meteorological data into the three-dimensional grid model. Consider the spatiotemporal distribution characteristics and change laws of meteorological data during the interpolation process to ensure the accuracy and reliability of the interpolation results. According to the interpolation results, the meteorological element values ​​of each grid node are extracted to form a grid feature data set.

[0026] In the above implementation, the three-dimensional grid model constructed by the horizontal geographic features and vertical atmospheric stratification features corresponding to the multi-source fusion meteorological data can more finely describe the spatial distribution and changes of the meteorological data, thereby improving the spatial resolution of the data. By interpolating the multi-source fusion meteorological data into the three-dimensional grid model, a more accurate and detailed data distribution can be obtained. The interpolation process takes into account the spatial correlation and continuity of the data, so that the results are more in line with the actual meteorological conditions. By mapping the meteorological data to the grid, the spatial distribution and change trend of the meteorological data can be effectively determined.

[0027] Furthermore, the horizontal geographic features and vertical atmospheric stratification features corresponding to the multi-source fusion meteorological data are determined, and a three-dimensional grid model is constructed based on the horizontal geographic features and the vertical atmospheric stratification features, including: Extract the geographical features of multi-source fusion meteorological data in the horizontal direction and set the horizontal resolution according to the geographical features; Extract the atmospheric stratification characteristics in the vertical direction of multi-source fusion meteorological data and determine the vertical resolution of different atmospheric layers; The horizontal resolution and the vertical resolution are input into the geographic information system, and grid division is performed in the horizontal direction and the vertical direction based on the geographic information system to obtain a number of grid units; Assign geographic coordinates and vertical height to each grid cell to obtain a three-dimensional grid model.

[0028] Specifically, the geographic information system is used to extract the geographical features of multi-source fusion meteorological data in the horizontal direction, including terrain undulation, landform type, urban distribution, etc. According to the geographical features, the boundaries and internal features of the target area are determined. According to the accuracy requirements of the meteorological data and the geographical features of the target area, the horizontal resolution is set. The horizontal resolution is expressed as the side length or area of ​​the grid unit. The change characteristics of the multi-source fusion meteorological data in the vertical direction are analyzed, including temperature gradient, humidity change, wind speed distribution, etc. According to the principles of meteorology, the boundaries and characteristics of different atmospheric layers are determined. According to the characteristics of atmospheric stratification and the vertical resolution requirements of meteorological data, the vertical resolution of each atmospheric layer is determined. The vertical resolution can be expressed as the thickness or height range of the altitude layer. The horizontal resolution and vertical resolution are input into the geographic information system, and the grid is divided in the horizontal direction and vertical direction according to the set horizontal resolution and vertical resolution in the geographic information system. In the horizontal direction, several horizontal grid units are divided according to the boundaries and internal features of the target area. In the vertical direction, several vertical altitude layers are divided according to the boundaries and vertical resolution of the atmospheric layer. Each grid unit is assigned a unique geographic coordinate (longitude, latitude) and vertical height. Ensure that the geographic coordinates and vertical height of each grid unit match its position in three-dimensional space. In the geographic information system, the divided grid units and the assigned geographic coordinates and vertical height information are integrated together to generate a three-dimensional grid model.

[0029] In the above implementation, by extracting the geographical features of multi-source fusion meteorological data in the horizontal direction, such as terrain, landform, water system, etc., a reasonable horizontal resolution can be set according to these features, which helps to more accurately describe the distribution and changes of meteorological data in the horizontal direction. In the vertical direction, by extracting atmospheric stratification features, such as the vertical distribution of meteorological elements such as temperature, humidity, and wind speed, the vertical resolution of different atmospheric layers can be determined. The identification of such vertical stratification features helps to understand the atmospheric structure more deeply. The horizontal resolution and the vertical resolution are input into the geographic information system, and the grid division is performed in the horizontal direction and the vertical direction respectively based on the geographic information system, so that a number of grid units can be obtained, which takes into account both the horizontal geographical features and the vertical atmospheric stratification features, making the grid division more reasonable and practical. The geographic coordinates and vertical height are assigned to each grid unit to obtain a three-dimensional grid model, so that the meteorological data is accurately located and expressed in the three-dimensional space. The meteorological data based on the three-dimensional grid model has higher spatial resolution and accuracy, can more accurately reflect the distribution and changes of meteorological elements in the three-dimensional space, helps to improve the refinement level of weather forecasts, and provides more accurate and reliable meteorological information.

[0030] Furthermore, the multi-source fusion meteorological data is interpolated into the three-dimensional grid model to obtain the interpolation results, including: Determine the meteorological element values ​​of known points and their corresponding longitude, latitude and altitude coordinates based on multi-source fusion meteorological data; Determine the coordinates of unknown points that need to be interpolated on the three-dimensional network model; Calculate the spatial distance between each unknown point and the known point, and determine the weight according to the inverse distance weighted interpolation principle; The meteorological element values ​​of the unknown points are obtained by performing weighted average calculation on the meteorological element values ​​and weights of the known points.

[0031] Specifically, determine the meteorological element values ​​(such as temperature, humidity, wind speed, etc.) of the known points and their corresponding longitude, latitude, and altitude coordinates. On the three-dimensional grid model, determine the coordinates of the unknown points that need to be interpolated. The weights are calculated according to the inverse distance weighted interpolation principle, and the weights are inversely proportional to the spatial distance. That is, the closer the spatial distance of the known point is, the greater its contribution to the unknown point, and the greater its weight.

[0032] Among them, the weight calculation formula can be expressed as: ; Where: Indicates i The weight of the known points, Indicates unknown point i The distance from a known point, k is the power parameter; According to the meteorological element values ​​and weights of known points, weighted average calculation is performed to obtain the meteorological element values ​​of unknown points.

[0033] Among them, the weighted average calculation formula can be expressed as: ; Where: Z Indicates the meteorological element value of the unknown point, Indicates i Meteorological element values ​​of known points; The above interpolation calculation is performed on all unknown points in the three-dimensional grid model to obtain the meteorological element value of each unknown point. The interpolation results are integrated together to form a meteorological element distribution map on the three-dimensional grid model.

[0034] In the above implementation, the meteorological element values ​​of the known points and their coordinate information can be used to deduce the meteorological element values ​​of the unknown points through the inverse distance weighted interpolation principle, which can fill the data gaps and improve the spatial integrity and continuity of the meteorological data. Interpolation on the three-dimensional grid model can ensure that the interpolation results are more accurate and reliable in space. Because the three-dimensional grid model has taken into account the horizontal geographical features and the vertical atmospheric stratification features, the interpolation results can better reflect the true distribution of meteorological elements in three-dimensional space. The inverse distance weighted interpolation principle determines the weight according to the spatial distance between the unknown point and the known point. The closer the known point is, the greater the influence on the unknown point is. The spatial relationship between the known points can be fully considered, making the interpolation result more accurate. Through weighted average calculation, the meteorological element values ​​of the unknown points can be obtained, and the information of multiple known points can be comprehensively considered, reducing the error caused by a single known point and improving the reliability of the interpolation result.

[0035] Furthermore, the calculation formula of the meteorological element value is as follows: ; Where: Indicates the meteorological element value of the unknown point, Represents the meteorological element value of a known point, p represents the weight coefficient, n Represents the number of known points.

[0036] S105, extracting features from the multi-source fused meteorological data to obtain corresponding meteorological features and time series features; Furthermore, feature extraction is performed on multi-source fusion meteorological data to obtain corresponding meteorological features and time series features, including: Determine the basic meteorological elements of multi-source fusion meteorological data; Extract the spatiotemporal distribution characteristics of basic meteorological elements, and extract composite meteorological characteristics by considering the correlation between basic meteorological elements; The time series analysis method is used to analyze the time series characteristics of multi-source fusion meteorological data to obtain the corresponding time series characteristics.

[0037] Specifically, the geographic information system is used to analyze the spatial distribution of basic meteorological elements, such as temperature gradients, humidity zones, precipitation areas, etc. The changing trends of meteorological elements are analyzed in time series, such as daily changes, seasonal changes, and interannual changes. The correlation between basic meteorological elements, such as the correlation between temperature and humidity, is calculated to identify possible composite meteorological features, such as the heat and humidity index. Based on the correlation between basic meteorological elements, composite meteorological features are constructed, including heat and humidity index, dew point temperature, wind shear, etc. The time series analysis technology is applied to analyze the time series characteristics of multi-source fusion meteorological data. The time series analysis technology of this embodiment can use autoregressive models, sliding average models, exponential smoothing methods, etc. to extract periodicity, trend, seasonality, volatility, and anomalies. Evaluate the contribution of time series features to meteorological forecasting and interpretation, and select features that have a significant impact on the forecasting model.

[0038] In the above implementation, the basic meteorological elements are determined: by processing the multi-source fusion meteorological data, its basic meteorological elements, such as temperature, humidity, wind speed, precipitation, etc., can be clarified. These basic meteorological elements are the basis of meteorological analysis and can more effectively understand meteorological phenomena and predict weather changes. Extracting the spatiotemporal distribution characteristics of the basic meteorological elements can reveal the changing laws of meteorological elements in space and time. In addition to the basic meteorological elements, the correlation between them is also considered to extract composite meteorological features. For example, the combination of temperature and humidity can reflect the state of the air, and the combination of wind speed and wind direction can reflect the characteristics of the wind field. These composite meteorological features can provide more comprehensive meteorological information and help to deeply understand meteorological phenomena. By analyzing the time series characteristics of multi-source fusion meteorological data through the time series analysis method, the corresponding time series characteristics can be obtained, which can reflect the changing laws of meteorological elements over time, such as seasonal changes, periodic changes, etc., and can improve the accuracy and timeliness of meteorological forecasts. Further, the time series characteristics of multi-source fusion meteorological data are analyzed by the time series analysis method to obtain the corresponding time series characteristics, including: Extract the timestamp corresponding to the multi-source fusion meteorological data; Annotate multi-source fusion meteorological data based on timestamp; Construct a time series model based on multi-source fused meteorological data with timestamp annotations; Generate time series features corresponding to multi-source fusion meteorological data based on the time series model.

[0039] Specifically, the time series model ARIMA model (differential integrated moving average autoregressive model) of this embodiment first integrates meteorological data from different data sources by timestamp to form a data set containing a complete time series. Duplicate, abnormal or missing timestamps and data points are removed, and timestamp labels are added to each data point. The ARIMA model is trained using multi-source fused meteorological data with timestamp annotations to capture the time series characteristics in the data. The trained time series model is used to extract time series features from the multi-source fused meteorological data.

[0040] In the above implementation, meteorological data is annotated based on timestamps, making the data clearer and more orderly in time series, which helps to improve the accuracy and efficiency of data analysis. Using the annotated time series data to build a model can more accurately capture the time series laws and trends in the data, which helps to more deeply understand the dynamic changes of meteorological data. Based on the time series model, time series features corresponding to multi-source fusion meteorological data can be generated. These features cover multiple aspects such as the time series, periodicity, and trend of the data. This comprehensive feature extraction helps to more comprehensively grasp the changing laws of meteorological data. Time series analysis can accurately extract time series features in the data, avoiding the errors and uncertainties that may exist in traditional methods.

[0041] S107, analyzing the grid characteristics, meteorological characteristics and time series characteristics through the meteorological data grid analysis model to obtain analysis results and output them to the terminal.

[0042] Furthermore, according to the grid characteristics, meteorological characteristics and time series characteristics, the meteorological data grid analysis model is used to analyze and obtain the analysis results, including: Obtain historical meteorological data and label data corresponding to the historical meteorological data; The vector machine model is trained by combining historical meteorological data with label data to obtain a meteorological data grid analysis model; Inputting grid characteristics, meteorological characteristics and time series characteristics into the meteorological data grid analysis model for analysis to obtain analysis results; The analysis results are output to the terminal in the form of graphs.

[0043] Specifically, historical meteorological data and corresponding label data are obtained from the database. Label data is the classification of meteorological phenomena (such as sunny days, rainy days, snowy days, etc.), or the threshold judgment of meteorological elements (such as high temperature, low temperature, heavy rain, etc.). The historical meteorological data is preprocessed, including data cleaning, missing value filling, outlier processing, etc., to ensure the accuracy and consistency of the data. Grid features (such as geographical location, terrain, etc.), meteorological features (such as temperature, humidity, wind speed, etc.) and time series features (such as periodicity and trend of time series, etc.) are extracted from the preprocessed historical meteorological data. The support vector machine (SVM) algorithm is used to train the model in combination with the extracted features and label data. The performance of the model is optimized by adjusting the C parameter, kernel function parameter and other parameters of the model. The trained model is verified using the validation set data to evaluate the accuracy and reliability of the model. According to the verification results, the model is adjusted and optimized. The grid features, meteorological features and time series features of the data to be analyzed are integrated to form a complete feature vector. The integrated feature vector is input into the trained meteorological data grid analysis model for prediction analysis. The prediction results output by the model are post-processed, including threshold judgment, classification label conversion, etc., to obtain the final analysis results. According to the analysis results, appropriate chart types are used for visualization. Finally, the drawn charts are output to the terminal for users to view and analyze.

[0044] In the above implementation, historical meteorological data and corresponding label data are obtained for training vector machine models, thereby obtaining a meteorological data grid analysis model, ensuring that the model can learn the inherent laws and characteristics of meteorological data. By continuously optimizing model parameters and training processes, the accuracy and efficiency of the model can be further improved. Inputting grid features, meteorological features, and time series features into the meteorological data grid analysis model for analysis can make full use of these feature information, capture the spatial and temporal variation of meteorological data, provide more comprehensive meteorological information, and help to more accurately understand and predict meteorological phenomena. Outputting the analysis results to the terminal in the form of charts can intuitively display the changing trends and characteristics of meteorological data, which helps meteorological workers to understand the analysis results more quickly and make accurate judgments and decisions.

[0045] According to a second aspect of the present invention, a meteorological data grid analysis system based on big data and machine learning is provided. Figure 2 As shown, including: A data acquisition module 202 is used to construct multi-source fusion meteorological data based on multi-source meteorological data; A gridding module 204 is used to grid the multi-source fusion meteorological data by taking into account the horizontal geographical features and the vertical atmospheric stratification features to obtain corresponding grid features; The feature extraction module 206 is used to extract features from the multi-source fused meteorological data to obtain corresponding meteorological features and time series features; The grid analysis module 208 is used to analyze the grid characteristics, meteorological characteristics and time series characteristics through the meteorological data grid analysis model to obtain the analysis results and output them to the terminal.

[0046] The beneficial effects of this embodiment are as follows: by fusing multi-source meteorological data and considering the weights of different data sources, the Kalman filter method is used for data fusion, thereby improving the accuracy and completeness of meteorological data. A three-dimensional grid model is used to grid the meteorological data, and the horizontal geographical features and vertical atmospheric stratification features are comprehensively considered, so that the spatial distribution of meteorological data is more refined and accurate. Through feature extraction technology, meteorological features and time series features are extracted from multi-source fused meteorological data, providing richer information for meteorological analysis. Through the application of time series analysis method, the time evolution law of meteorological data is revealed, and the reliability of meteorological forecasts is improved. Using the trained meteorological data grid analysis model, combined with grid features, meteorological features and time series features for comprehensive analysis, accurate meteorological analysis results can be output.

[0047] The above specific implementation is a preferred implementation of a meteorological data grid analysis method and system based on big data and machine learning of the present invention. It is not intended to limit the specific implementation scope of the present invention. The scope of the present invention includes but is not limited to this specific implementation. All equivalent changes made in accordance with the shape and structure of the present invention are within the protection scope of the present invention.

Claims

1. A meteorological data grid analysis method based on big data and machine learning, characterized in that: include: Construct multi-source fusion meteorological data based on multi-source meteorological data; The multi-source fusion meteorological data is gridded by a three-dimensional grid model that considers horizontal geographical features and vertical atmospheric stratification features to obtain corresponding grid features; Extracting features from the multi-source fused meteorological data to obtain corresponding meteorological features and time series features; According to the grid characteristics, the meteorological characteristics and the time series characteristics, an analysis is performed through a meteorological data grid analysis model to obtain an analysis result and output it to a terminal.

2. The meteorological data grid analysis method based on big data and machine learning according to claim 1 is characterized in that: The method of constructing multi-source fusion meteorological data based on multi-source meteorological data includes: Collect meteorological data from different data sources; Preprocessing the meteorological data from different data sources, wherein the preprocessing at least includes data cleaning, data standardization and format conversion; Considering the weights of different data sources, the pre-processed meteorological data from different data sources are fused by Kalman filtering to obtain multi-source fused meteorological data.

3. The meteorological data grid analysis method based on big data and machine learning according to claim 1 is characterized in that: The method of gridding the multi-source fusion meteorological data by taking into account the horizontal geographical features and the vertical atmospheric stratification features of the three-dimensional grid model to obtain corresponding grid features includes: Determine the horizontal geographic features and vertical atmospheric stratification features corresponding to the multi-source fusion meteorological data, and construct a three-dimensional grid model based on the horizontal geographic features and vertical atmospheric stratification features; The multi-source fusion meteorological data is interpolated into the three-dimensional grid model to obtain an interpolation result, and grid features of the multi-source fusion meteorological data are obtained according to the interpolation result.

4. The meteorological data grid analysis method based on big data and machine learning according to claim 3 is characterized in that: The determining of horizontal geographic features and vertical atmospheric stratification features corresponding to the multi-source fusion meteorological data, and constructing a three-dimensional grid model based on the horizontal geographic features and vertical atmospheric stratification features, includes: Extracting geographic features of the multi-source fusion meteorological data in the horizontal direction, and setting the horizontal resolution according to the geographic features; Extracting the atmospheric stratification characteristics of the multi-source fusion meteorological data in the vertical direction and determining the vertical resolution of different atmospheric layers; Inputting the horizontal resolution and the vertical resolution into a geographic information system, and performing grid division in the horizontal direction and the vertical direction based on the geographic information system to obtain a plurality of grid units; Geographic coordinates and vertical heights are assigned to each of the grid cells to obtain a three-dimensional grid model.

5. The meteorological data grid analysis method based on big data and machine learning according to claim 3 is characterized in that: The interpolating the multi-source fused meteorological data into the three-dimensional grid model to obtain an interpolation result includes: Determine the meteorological element value of a known point and its corresponding longitude, latitude and altitude coordinates based on the multi-source fusion meteorological data; Determine the coordinates of unknown points that need to be interpolated on the three-dimensional network model; Calculate the spatial distance between each unknown point and the known point, and determine the weight according to the inverse distance weighted interpolation principle; The meteorological element values ​​of the unknown points are obtained by performing weighted average calculation on the meteorological element values ​​of the known points and the weights.

6. The meteorological data grid analysis method based on big data and machine learning according to claim 5 is characterized in that: The calculation formula of the meteorological element value is as follows: ； Where: Indicates the meteorological element value of the unknown point, Represents the meteorological element value of a known point, represents the distance between the unknown point and the known point, p represents the weight coefficient, and n represents the number of known points.

7. The meteorological data grid analysis method based on big data and machine learning according to claim 1 is characterized in that: The extracting features of the multi-source fusion meteorological data to obtain corresponding meteorological features and time series features includes: Determine the basic meteorological elements of the multi-source fusion meteorological data; Extracting the spatiotemporal distribution characteristics of the basic meteorological elements, and extracting composite meteorological characteristics by considering the correlation between the basic meteorological elements; The time series characteristics of the multi-source fusion meteorological data are analyzed by time series analysis method to obtain corresponding time series features.

8. The meteorological data grid analysis method based on big data and machine learning according to claim 7 is characterized in that: The time series characteristics of the multi-source fusion meteorological data are analyzed by the time series analysis method to obtain corresponding time series features, including: Extracting the timestamp corresponding to the multi-source fusion meteorological data; Annotating the multi-source fusion meteorological data based on the timestamp; Constructing a time series model based on the multi-source fusion meteorological data with timestamp annotation; The time series features corresponding to the multi-source fusion meteorological data are generated based on the time series model.

9. The meteorological data grid analysis method based on big data and machine learning according to claim 1, characterized in that: The analyzing results obtained by analyzing the grid characteristics, the meteorological characteristics and the time series characteristics through a meteorological data grid analysis model include: Acquire historical meteorological data and label data corresponding to the historical meteorological data; Training a vector machine model by combining the historical meteorological data with the label data to obtain a meteorological data grid analysis model; Inputting the grid characteristics, the meteorological characteristics and the time series characteristics into the meteorological data grid analysis model for analysis to obtain analysis results; The analysis results are output to the terminal in the form of graphs.

10. A meteorological data grid analysis system based on big data and machine learning, applicable to the meteorological data grid analysis method based on big data and machine learning as claimed in any one of claims 1 to 9, characterized in that: include: A data acquisition module is used to construct multi-source fusion meteorological data based on multi-source meteorological data; A gridding module, used for gridding the multi-source fusion meteorological data by taking into account horizontal geographical features and vertical atmospheric stratification features to obtain corresponding grid features; A feature extraction module is used to extract features from the multi-source fusion meteorological data to obtain corresponding meteorological features and time series features; The grid analysis module is used to analyze the grid characteristics, the meteorological characteristics and the time series characteristics through the meteorological data grid analysis model to obtain the analysis results and output them to the terminal.

Images