Multi-source Data Fusion and Tornado Intelligent Recognition Method

Abstract

The present invention discloses a multi-source data fusion and tornado intelligent recognition method, which relates to the technical field of electrical digital data processing. The multi-source data fusion and tornado intelligent recognition method includes the following steps: obtaining a quality control evaluation index; obtaining a grid accuracy index; obtaining an intelligent recognition evaluation score; obtaining a tornado warning duration. The present invention performs quality control on the obtained dual-polarization phased array radar data to obtain a quality control evaluation index, then inputs the obtained multi-source real-time data into a pre-constructed severe convective weather recognition model to obtain the areas prone to severe convective weather. Finally, it determines whether to obtain radar grid data based on the obtained grid accuracy index, and at the same time determines whether the intelligent recognition is completed according to the obtained intelligent recognition evaluation score, achieving the effect of improving the correlation between the tornado intelligent recognition result and the multi-source data fusion, and solving the problem of low correlation between the tornado intelligent recognition result and the multi-source data fusion in the prior art.

Description

Technical Field

[0001] The present invention relates to the technical field of electrical digital data processing, and particularly to a multi-source data fusion and tornado intelligent recognition method. Background Art

[0002] With the rapid development of information technology and the advent of the big data era, all walks of life are facing the challenges and opportunities of explosive data growth. In the fields of meteorological monitoring, disaster warning, environmental protection, etc., the acquisition, integration, and analysis of multi-source data have become the key to improving prediction accuracy and response speed. Especially in the prediction and identification of extreme weather events such as tornadoes, the traditional single data source method has been difficult to meet the requirements of high precision and real-time. Therefore, the multi-source data fusion technology has emerged. It provides new ideas and methods for tornado intelligent recognition by integrating and comprehensively analyzing data from different channels and different formats.

[0003] In the existing technology, data objects are preferentially selected and preprocessed, then weight assignment and combination are performed on the preprocessed data and input into the numerical weather prediction model and the statistical prediction model to obtain the prediction results. Finally, feature extraction is performed on the prediction results to obtain the feature information before the formation of a tornado, and dynamic monitoring and identification of the tornado are realized based on the feature information.

[0004] For example, an adaptive auxiliary decision-making intelligent method and system based on multi-source dynamic data disclosed in the invention patent announcement with the publication number of CN116842127B includes: respectively performing data processing and data storage on the acquired dynamic data and static data; respectively encoding the static data and the standardized dynamic data based on the text encoder and the visual encoder of the fine-tuned vision-language model to obtain text and image features; obtaining the category with the largest similarity through object recognition, and the category of the target can be obtained by referring to this index.

[0005] For example, a method and system for intelligent target fusion recognition based on radar multi-modal data disclosed in the patent application with the publication number of CN118885974A includes: acquiring the radar multi-modal data of the target to be recognized; inputting the acquired radar multi-modal data into a trained multi-modal radar data target fusion recognition model, and sequentially performing single-modal feature extraction to respectively obtain multiple single-modal features; according to the multiple single-modal features, performing modal similarity feature and eigenfeature learning based on the multi-modal feature learning subspace constraint, compressing the learned multiple modal similarity features to obtain multi-modal similarity features; performing attention fusion on the learned multiple eigenfeatures and the multi-modal similarity features, and outputting the recognition probability to obtain the classification recognition result of the target in the radar multi-modal data of the target to be recognized.

[0006] However, in the process of implementing the inventive technical solution in the embodiments of the present application, it is found that the above technologies have at least the following technical problems:

[0007] In the prior art, statistical prediction models usually rely on historical data and statistical laws, and may not be able to fully capture the complexity and non-linear characteristics of tornadoes. Secondly, the feature extraction algorithm may not be able to accurately capture all the key feature information before the formation of tornadoes, resulting in a decrease in the accuracy of the recognition results. There is a problem that the relevance between the intelligent recognition results of tornadoes and the multi-source data fusion is not high. Summary of the Invention

[0008] The embodiments of the present application provide a multi-source data fusion and intelligent tornado recognition method, which solves the problem that the relevance between the intelligent recognition results of tornadoes and the multi-source data fusion in the prior art is not high, and realizes the improvement of the relevance between the intelligent recognition results of tornadoes and the multi-source data fusion.

[0009] The embodiments of the present application provide a multi-source data fusion and intelligent tornado recognition method, including the following steps: Step 1, obtain the dual-polarization phased array radar data in a preset band and perform quality control to obtain a quality control evaluation index, and at the same time judge whether to obtain multi-source real-time data based on the obtained quality control evaluation index. The quality control evaluation index is used to evaluate the error correction effect of the dual-polarization phased array radar data within a preset time period; Step 2, input the obtained multi-source real-time data into the constructed severe convective weather recognition model to obtain the areas prone to severe convective weather, and at the same time perform nested gridification on the dual-polarization phased array radar data in the areas prone to severe convective weather to obtain a gridification accuracy index. The gridification accuracy index is used to evaluate the accuracy and reliability of the nested gridification of the dual-polarization phased array radar data; Step 3, judge whether to obtain radar grid data based on the obtained gridification accuracy index. If radar grid data is obtained, perform intelligent recognition on the obtained radar grid data to obtain an intelligent recognition evaluation score, and at the same time judge whether the intelligent recognition is completed according to the obtained intelligent recognition evaluation score. The intelligent recognition evaluation score is used to evaluate the matching degree between the radar grid data and the intelligent recognition conditions; Step 4, if the intelligent recognition is completed, design a membership function according to the obtained radar grid data and combine the fuzzy logic method to obtain the tornado warning duration.

[0010] Further, the quality control evaluation index is obtained by processing the signal-to-noise ratio, ground clutter echo removal efficiency, isolated echo removal efficiency, and electromagnetic interference suppression efficiency in the quality control process of the dual-polarization phased array radar data within a preset time period; the ground clutter echo removal efficiency represents the ratio of the difference between the initial ground clutter echo intensity before quality control and the actual ground clutter echo intensity after quality control within a preset time period to the initial ground clutter echo intensity; the isolated echo removal efficiency represents the ratio of the difference between the initial number of isolated echoes before quality control and the actual number of isolated echoes after quality control within a preset time period to the initial number of isolated echoes; the electromagnetic interference suppression efficiency represents the ratio of the difference between the initial electromagnetic interference intensity before quality control and the actual electromagnetic interference intensity after quality control within a preset time period to the initial electromagnetic interference intensity.

[0011] Further, the specific steps for obtaining the grid accuracy index include: E1, real-time monitoring of the nested grid situation of the dual-polarization phased array radar data in the Cartesian coordinate system, and judging whether the number of consistent peak positions within a preset time period is greater than the preset number of consistent peak positions. If so, execute E2; otherwise, re-perform nested gridification; E2, obtain the fraction of the number of consistent peak positions, and at the same time judge whether the time change coefficient is equal to the reference time change coefficient. If so, obtain the gridification success rate and the gridification timestamp, and combine the obtained quality control evaluation index and the fraction of the number of consistent peak positions to obtain the grid accuracy index; otherwise, obtain the gridification timestamp and combine the obtained quality control evaluation index and the fraction of the number of consistent peak positions to obtain the grid accuracy index; the fraction of the number of consistent peak positions represents the ratio of the difference between the number of consistent peak positions and the preset number of consistent peak positions to the preset number of consistent peak positions; the time change coefficient represents the change amplitude of the dual-polarization phased array radar data over time; the gridification success rate represents the ratio of the amount of data successfully nested and gridified within a preset time period to the total amount of dual-polarization phased array radar data to be nested and gridified.

[0012] Further, the specific limit expression of the grid accuracy index is:

[0013] ;

[0014] In the formula, t is the number of the preset time period, , T is the total number of preset time periods, e is the natural constant, represents the grid accuracy index in the process of nested gridification of the dual-polarization phased array radar data in the t-th preset time period, Represents the quality control evaluation index during the quality control process of dual-polarization phased array radar data within the t-th preset time period. Represents the preset quality control evaluation index. Represents the peak position consistency quantity fraction during the nested gridding process of dual-polarization phased array radar data within the t-th preset time period. Represents the peak position consistency quantity during the nested gridding process of dual-polarization phased array radar data within the t-th preset time period. Represents the preset peak position consistency quantity. Represents the gridding success rate during the nested gridding process of dual-polarization phased array radar data within the t-th preset time period. Represents the gridding timestamp during the nested gridding process of dual-polarization phased array radar data within the t-th preset time period. Represents the time variation coefficient during the nested gridding process of dual-polarization phased array radar data within the t-th preset time period. Represents the reference time variation coefficient.

[0015] Furthermore, the intelligent recognition evaluation score is obtained through the following method: The real-time response speed of radar grid data is monitored in real time to obtain the average response speed score during the intelligent recognition process within the preset time period. The average response speed score is the ratio of the difference between the average response speed and the reference average response speed to the average response speed reference deviation; at the end of the preset time period, the quantity of missed recognition data is obtained and it is judged whether the quantity of missed recognition data is less than the maximum allowable quantity of missed recognition data. If so, the missed recognition rate score is obtained and the intelligent recognition of the next time period is executed. Otherwise, the intelligent recognition is performed again. The missed recognition rate score is the ratio of the quantity of missed recognition data to the total quantity of radar grid data to be intelligently recognized; the signal-to-noise ratio during the intelligent recognition process of radar grid data within the preset time period is obtained, and at the same time, the intelligent recognition evaluation score is obtained by combining the obtained average response speed score, missed recognition rate score, and gridding accuracy index.

[0016] Furthermore, the specific process of judging whether the intelligent recognition is completed according to the obtained intelligent recognition evaluation score is as follows: It is judged whether the obtained intelligent recognition evaluation score is greater than the preset intelligent recognition evaluation score: If the obtained intelligent recognition evaluation score is greater than the preset intelligent recognition evaluation score, the intelligent recognition is completed and the radar grid data for which the intelligent recognition is completed is marked as tornado-identified; if the obtained intelligent recognition evaluation score is not greater than the preset intelligent recognition evaluation score, the corresponding radar grid data is marked as not tornado-identified.

[0017] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0018] 1. By obtaining dual-polarization phased array radar data in a preset band and performing quality control to obtain a quality control evaluation index, then inputting the obtained multi-source real-time data into a constructed severe convective weather identification model to obtain areas prone to severe convective weather, and finally judging whether to obtain radar grid data based on the obtained grid accuracy index, and at the same time judging whether the intelligent identification is completed according to the obtained intelligent identification evaluation score, the more accurate acquisition of the intelligent identification evaluation score is realized, and then the improvement of the correlation between the tornado intelligent identification result and multi-source data is realized, effectively solving the problem of low correlation between the tornado intelligent identification result and multi-source data in the prior art.

[0019] 2. By obtaining the number score of consistent peak positions, and at the same time judging whether the time change coefficient is equal to the reference time change coefficient. If so, obtain the grid success rate and grid timestamp, and combine the obtained quality control evaluation index and the number score of consistent peak positions to obtain the grid accuracy index. Otherwise, obtain the grid timestamp and combine the obtained quality control evaluation index and the number score of consistent peak positions to obtain the grid accuracy index, thus realizing the more accurate acquisition of the grid accuracy index, and then realizing the improvement of the accuracy and reliability of nested grid.

[0020] 3. When the amount of missed identification data at the end of the preset time period is less than the maximum allowable amount of missed identification data, obtain the signal-to-noise ratio during the intelligent identification process of the radar grid data within the preset time period, and at the same time combine the obtained average response speed score, missed identification rate score and grid accuracy index to obtain the intelligent identification evaluation score, thus realizing the improvement of the accuracy and reliability of the acquisition of the intelligent identification evaluation score, and then realizing the improvement of the accuracy and real-time performance of tornado intelligent identification. Description of the Drawings

[0021] Figure 1 It is a flowchart of the multi-source data fusion and tornado intelligent identification method provided by the embodiment of the present application;

[0022] Figure 2 It is a flowchart of the multi-source real-time data fusion provided by the embodiment of the present application;

[0023] Figure 3 It is a block diagram of the severe convective weather identification model provided by the embodiment of the present application;

[0024] Figure 4 It is a schematic diagram of the nested grid of dual-polarization phased array radar data provided by the embodiment of the present application;

[0025] Figure 5 It is a comparison diagram of the two-kilometer height effect before and after the nested grid of SA radar data provided by the embodiment of the present application;

[0026] Figure 6This is the roadmap for tornado intelligent recognition provided by the embodiments of this application. Detailed implementation manners

[0027] By providing a multi-source data fusion and tornado intelligent recognition method, the embodiments of this application solve the problem that the relevance between the tornado intelligent recognition result and multi-source data fusion in the prior art is not high. By acquiring dual-polarization phased array radar data in a preset band and performing quality control to obtain a quality control evaluation index, and at the same time judging whether to acquire multi-source real-time data based on the obtained quality control evaluation index, then inputting the obtained multi-source real-time data into the constructed severe convective weather recognition model to obtain the areas prone to severe convective weather. At the same time, nested gridification is performed on the dual-polarization phased array radar data in the areas prone to severe convective weather to obtain a gridification accuracy index. Finally, it is judged whether to acquire radar grid data based on the obtained gridification accuracy index. If radar grid data is acquired, intelligent recognition is performed on the acquired radar grid data to obtain an intelligent recognition evaluation score. At the same time, it is judged whether the intelligent recognition is completed according to the obtained intelligent recognition evaluation score. If the intelligent recognition is completed, a membership function is designed based on the acquired radar grid data and combined with the fuzzy logic method to obtain the tornado warning duration, achieving an improvement in the relevance between the tornado intelligent recognition result and multi-source data fusion.

[0028] The technical solutions in the embodiments of this application are to solve the problem that the relevance between the tornado intelligent recognition result and multi-source data fusion is not high. The general idea is as follows:

[0029] By acquiring dual-polarization phased array radar data in a preset band and performing quality control to obtain a quality control evaluation index, then inputting the obtained multi-source real-time data into the constructed severe convective weather recognition model to obtain the areas prone to severe convective weather, and finally judging whether to acquire radar grid data based on the obtained gridification accuracy index, and at the same time judging whether the intelligent recognition is completed according to the obtained intelligent recognition evaluation score, the effect of improving the relevance between the tornado intelligent recognition result and multi-source data fusion is achieved.

[0030] To better understand the above technical solutions, the above technical solutions will be described in detail below in conjunction with the accompanying drawings of the specification and specific implementation manners.

[0031] Such as Figure 1As shown in the figure, it is a flowchart of the multi-source data fusion and tornado intelligent recognition method provided by the embodiment of the present application. The method includes the following steps: Step 1, obtain the dual-polarization phased array radar data of the preset band and perform quality control (perform differential phase correction on the dual-polarization phased array radar data through the differential phase quality control algorithm built in the radar system, that is, remove the backward differential phase error existing in the measured differential phase value of the dual-polarization phased array radar data) to obtain a quality control evaluation index. At the same time, based on the obtained quality control evaluation index, judge whether to obtain multi-source real-time data. The preset band includes the X band and the C band. Quality control is used to correct the errors in the dual-polarization phased array radar data, and the quality control evaluation index is used to evaluate the error correction effect of the dual-polarization phased array radar data within the preset time period; Step 2, input the obtained multi-source real-time data into the constructed severe convective recognition model to obtain the areas prone to severe convective weather. At the same time, perform nested gridification on the dual-polarization phased array radar data in the areas prone to severe convective weather to obtain a gridification accuracy index. The gridification accuracy index is used to evaluate the accuracy and reliability of the nested gridification of the dual-polarization phased array radar data; Step 3, based on the obtained gridification accuracy index, judge whether to obtain radar grid data. If radar grid data is obtained, perform intelligent recognition on the obtained radar grid data to obtain an intelligent recognition evaluation score. At the same time, according to the obtained intelligent recognition evaluation score, judge whether the intelligent recognition is completed. The intelligent recognition evaluation score is used to evaluate the matching degree between the radar grid data and the intelligent recognition conditions; Step 4, if the intelligent recognition is completed, design a membership function according to the obtained radar grid data and combine it with the fuzzy logic method to obtain the tornado warning duration. The membership function includes the TVS membership function and the TDS membership function. The TVS membership function is used for tornado classification warning to obtain the tornado warning release time and for tornado touchdown warning to obtain the tornado formation time. The tornado warning duration is the duration between the tornado warning release time and the tornado formation time.

[0032] In this embodiment, the dual-polarization phased array radar data includes reflectivity factors of horizontal polarization and vertical polarization (used to reflect the intensity and distribution of radar echoes, which helps to identify the echo characteristics of tornadoes), radial velocity (indicating the movement direction and speed of the radar echo relative to the radar station), differential reflectivity factor (used to reflect the differences in the size and shape of precipitation particles, which helps to distinguish different types of precipitation, such as raindrops and hailstones, so as to more accurately identify tornadoes), and differential propagation phase shift (indicating the phase change of electromagnetic waves during propagation due to the scattering and absorption of precipitation particles); the multi-source real-time data is used to verify the accuracy of the real-time observation data; the real-time observation data includes radar echo data, meteorological element observation data (such as temperature, humidity, air pressure, wind direction, and wind speed), and geographic information data (such as terrain, terrain undulation amplitude, and vegetation coverage); the severe convection identification model is used to identify the occurrence probability of severe convection weather (such as tornadoes) in the coverage area of the dual-polarization phased array radar; the area prone to severe convection weather refers to the area where the occurrence probability of severe convection weather is greater than the preset occurrence probability; the nested gridification is used to convert the dual-polarization phased array radar data into a regular grid form, and the radar grid data (representing the dual-polarization phased array radar data after the nested gridification) is used to reflect the distribution information of meteorological elements in the geographical space.

[0033] It should be added that, as Figure 2 shown, it is the flowchart of the multi-source real-time data fusion provided by the embodiment of the present application. The multi-source real-time data includes, but is not limited to, sounding data, surface automatic station data, satellite data, and numerical forecast data. The multi-source real-time data fusion is the fusion among sounding data, surface automatic station data, satellite data, and numerical forecast data. Among them, the sounding data has the characteristic of high vertical resolution and usually includes temperature, humidity, air pressure, wind direction, and wind speed at different levels from the ground to high altitude, which is mainly measured in real time by high-altitude observation equipment (such as meteorological sounding balloons equipped with meteorological sensors); the surface automatic station data has the characteristics of high density and high frequency and is mainly provided by surface automatic meteorological observation stations distributed around the world (equipped with temperature sensors, humidity sensors, barometers, and pressure sensors for real-time monitoring of surface meteorological elements); the satellite data has the characteristics of wide coverage and high observation frequency and is provided by meteorological satellites (usually equipped with remote sensing instruments of visible light, infrared, and microwave) for real-time monitoring of cloud amount, cloud type, precipitation, radiation, and reflection characteristics of the ocean and the surface globally; the numerical forecast data has the characteristics of high accuracy and long time and is used to predict the changes of meteorological elements in the future for a period of time, which is mainly provided by the meteorological numerical forecasting system.

[0034] Specifically, the specific construction steps of the severe convective identification model are as follows: According to the convective characteristics of meso - scale and micro - scale, select no less than 100 representative parameters from the historical severe convective database as the input of the severe convective identification model, including the atmospheric structure state parameters at different heights of each grid point, gridded radar parameters, as well as the spatio - temporal variables of these parameters and cloud - top temperature. As shown in Table 1, it is a statistical table of some input parameters of the severe convective identification model provided by the embodiments of the present application:

[0035] Table 1 Statistical table of some input parameters of the severe convective identification model

[0036] Abbreviation Name Abbreviation Name K K Index TQC Vent Pipe Parameter JI Jefferson Index Ls Dry Warm Cap Index SI Showalter Index CCL Convective Condensation Level Faust Faust Index CCL_T Temperature at Convective Condensation Level ICC Barber Convective Instability Index CCL_mod Modified Convective Condensation Level LIMax Maximum Lifting Index Tg Convective Temperature BI Convective Stability Index SWEAT Severe Weather Threat Index MDPI Potential Downburst Index Wd_S Static Conditional Stability Teffer Teffer Index SSI Storm Intensity Index ChTT Charba Total Index SWISS00 Swiss Thunderstorm 1 Hd_020 Mixed Phase Layer SWISS12 Swiss Thunderstorm 2 Hd_204 BB Growth Layer Shr Coarse Richardson Number Shear ZH 0 Degree Height SRH Storm Relative Helicity ZH20 -20 Height Z Echo Intensity ZH30 -30 Height ET Echo Top Height ZH40 -40 Height VIL Vertical Integrated Liquid Water Content Integra1Q Integrated Moisture Content of the Whole Layer

[0037] Form a severe convective training data set from the parameters in Table 1 (70% for training and 30% for testing), and use the deep learning algorithm in machine learning algorithms for cross - testing and training to establish a severe convective identification model for the development of severe convective weather such as thunderstorm gales and tornadoes with self - learning functions. As Figure 3 shown, it is the block diagram of the severe convective identification model provided by the embodiments of the present application. Through the establishment and application of this model, the intelligent identification of severe convective weather such as tornadoes has been successfully achieved. More importantly, this model can also be fused and correlated with multi - source data, thereby improving the fusion and correlation of the intelligent identification results of tornadoes with multi - source data.

[0038] Furthermore, the quality control evaluation index is obtained by processing the signal - to - noise ratio (the ratio of signal power to noise power, that is, in the limiting expression of the quality control evaluation index ), the ground clutter removal efficiency (that is, in the limiting expression of the quality control evaluation index ), the isolated echo removal efficiency (that is, in the limiting expression of the quality control evaluation index ), and the electromagnetic interference suppression efficiency (that is, in the limiting expression of the quality control evaluation index ) of the dual - polarization phased - array radar data during the quality control process within a preset time period; the ground clutter removal efficiency represents the ratio of the difference between the initial ground clutter intensity before quality control and the actual ground clutter intensity after quality control within a preset time period of the dual - polarization phased - array radar data to the initial ground clutter intensity; the isolated echo removal efficiency represents the ratio of the difference between the initial number of isolated echoes before quality control and the actual number of isolated echoes after quality control within a preset time period of the dual - polarization phased - array radar data to the initial number of isolated echoes; the electromagnetic interference suppression efficiency represents the ratio of the difference between the initial electromagnetic interference intensity before quality control and the actual electromagnetic interference intensity after quality control within a preset time period of the dual - polarization phased - array radar data to the initial electromagnetic interference intensity.

[0039] Among them, the signal-to-noise ratio is measured in real time by a signal-to-noise ratio tester, the ground echo intensity and the electromagnetic interference intensity are measured in real time by an optical fiber microcrack detector, and the ground echo, isolated echo and electromagnetic wave are obtained by a radar receiver in the radar system.

[0040] In this embodiment, the specific limiting expression of the quality control evaluation index is:

[0041] ;

[0042] ;

[0043] ;

[0044] ;

[0045] In the formula, t is the number of the preset time period, , T is the total number of the preset time periods, e is the natural constant, represents the quality control evaluation index in the quality control process of the dual-polarization phased array radar data in the t-th preset time period, 1 represents the signal-to-noise ratio in the quality control process of the dual-polarization phased array radar data in the preset time period, represents the ground echo removal efficiency of the dual-polarization phased array radar data in the t-th preset time period, represents the initial ground echo intensity of the dual-polarization phased array radar data before quality control in the t-th preset time period, represents the actual ground echo intensity of the dual-polarization phased array radar data after quality control in the t-th preset time period, represents the isolated echo removal efficiency of the dual-polarization phased array radar data in the t-th preset time period, represents the initial number of isolated echoes of the dual-polarization phased array radar data before quality control in the t-th preset time period, represents the actual number of isolated echoes of the dual-polarization phased array radar data after quality control in the t-th preset time period, represents the electromagnetic interference suppression efficiency of the dual-polarization phased array radar data in the t-th preset time period, represents the initial electromagnetic interference intensity of the dual-polarization phased array radar data before quality control in the t-th preset time period, represents the actual electromagnetic interference intensity of the dual-polarization phased array radar data after quality control in the t-th preset time period.

[0046] It should be understood that the quality control evaluation index increases with the increase of the signal-to-noise ratio and decreases with the increase of the ground clutter removal efficiency, isolated clutter removal efficiency, and electromagnetic interference suppression efficiency. It should be noted that (1) the signal-to-noise ratio also indirectly affects the value of the electromagnetic interference suppression efficiency. When the signal-to-noise ratio increases, it indicates that the radar system's ability to identify signals is enhanced, which means that the radar system can more effectively distinguish radar signals from electromagnetic interference signals, thereby improving the electromagnetic interference suppression efficiency.

[0047] (2) The ground clutter removal efficiency also indirectly affects the value of the isolated clutter removal efficiency. When the ground clutter removal efficiency increases, the remaining ground clutter in the dual-polarization phased array radar data is less, which helps to reduce the occurrence of isolated clutter because isolated clutter sometimes results from improper processing of ground clutter, such as incomplete removal of ground clutter or misidentification as target echoes. Therefore, improving the ground clutter removal efficiency can indirectly improve the isolated clutter removal efficiency.

[0048] By considering the above indirect influence mechanism, it helps to more accurately identify tornadoes, reduce false alarms and missed detections, thereby improving the reliability of the intelligent recognition results, and further achieving an improvement in the correlation between the intelligent recognition results of tornadoes and multi-source data fusion, effectively solving the problem of low correlation between the intelligent recognition results of tornadoes and multi-source data fusion in the prior art.

[0049] Further, the specific process of determining whether to obtain multi-source real-time data based on the obtained quality control evaluation index is as follows: Determine whether the obtained quality control evaluation index is greater than the preset quality control evaluation index. If the obtained quality control evaluation index is greater than the preset quality control evaluation index, it indicates that the result of quality control meets the expected requirements and multi-source real-time data is obtained. If the obtained quality control evaluation index is not greater than the preset quality control evaluation index, it indicates that the result of quality control does not meet the expected requirements. At this time, multi-source real-time data is not obtained and the preset band parameters are adjusted, and quality control is performed again according to the adjusted preset band parameters until the obtained quality control evaluation index is greater than the preset quality control evaluation index. The preset band parameters include the center frequency of the band, the radar scanning mode (such as sector scanning, conical scanning), and the pulse repetition frequency.

[0050] In this embodiment, the preset quality control evaluation index is represented by the result of summing and averaging the historical quality control evaluation indexes in the preset database during the historical quality control process. Suppose that during a certain quality control process, the radar system uses the following preset band parameters: center frequency: 10 GHz; scanning mode: sector scanning; pulse repetition frequency: 1 kHz. At this time, the obtained quality control evaluation index is 1.2 (the preset quality control evaluation index is 1.5).

[0051] To improve the quality control effect, the preset band parameters can be adjusted as follows: adjust the center frequency from 10 GHz to 12 GHz; adjust the scanning mode from sector scanning to conical scanning; adjust the pulse repetition frequency from 1 kHz to 1.5 kHz or 2 kHz; it should be noted that the above adjustments are a complete set of preset band parameter adjustments. After each complete set of preset band parameter adjustments, quality control needs to be performed again to obtain a new quality control evaluation index. If the new quality control evaluation index reaches or exceeds 1.5, the adjustment is successful; otherwise, continue to adjust. Through the above steps of adjusting the preset band parameters, the detection effect of the radar system can be gradually optimized, the accuracy of quality control can be improved, and thus a more accurate evaluation of the quality control effect can be achieved.

[0052] Further, the specific steps for nesting and gridding the dual-polarization phased array radar data in areas prone to severe convective weather are as follows: generate radar information based on the quality-controlled dual-polarization phased array radar data, and at the same time use an adaptive algorithm to fuse the generated radar information into a Cartesian coordinate system to form three-dimensional grid point (including longitude, latitude, and altitude) information (with uniform spatial resolution); the radar information is used to visually describe the weather phenomena in the coverage area of the dual-polarization phased array radar; the Cartesian coordinate system is used to visually represent the positioning and distribution of the dual-polarization phased array radar data in three-dimensional space. Among them, the Cartesian coordinate system consists of three mutually perpendicular number axes, representing longitude (X-axis), latitude (Y-axis), and altitude (Z-axis) respectively. In this coordinate system, each point can be represented by a triple (X, Y, Z) to indicate its position.

[0053] As Figure 4 shown, it is a schematic diagram of the nesting and gridding of the dual-polarization phased array radar data provided by the embodiment of the present application. After nesting and gridding, the resolution of the dual-polarization phased array radar data is significantly improved, which helps to more clearly observe the fine internal structure of clouds in severe convective weather, and improves the accuracy and reliability of obtaining the internal microphysical structure of severe convective cloud systems (such as supercells and mesocyclones that produce tornadoes), laying a foundation for intelligent tornado identification.

[0054] Taking the SA radar data obtained by the Guangzhou SA (CINRAD-SA, China New Generation Weather Radar) radar as an example, as Figure 5 shown, it is a comparison chart of the two-kilometer altitude effect before and after nesting and gridding of the SA radar data provided by the embodiment of the present application. Among them, figure (a) is before nesting and gridding, and figure (b) is after nesting and gridding. By Figure 5It can be seen that the convective cloud structure after nested gridification is clearer and more accurate than that before nested gridification. In this example, nested gridification helps to improve the accuracy and reliability of dual-polarization phased array radar data acquisition, thus realizing the adaptive nesting of dual-polarization phased array radar data.

[0055] Furthermore, the specific steps for obtaining the gridification accuracy index include: E1, when the obtained quality control evaluation index is greater than the preset quality control evaluation index, the nested gridification situation of dual-polarization phased array radar data in the Cartesian coordinate system is monitored in real time, and it is judged whether the number of consistent peak positions within the preset time period is greater than the preset number of consistent peak positions. If so, execute E2; otherwise, re-perform nested gridification; E2, obtain the fraction of the number of consistent peak positions (i.e., in the limiting expression of the gridification accuracy index), and at the same time judge whether the time change coefficient is equal to the reference time change coefficient. If so, obtain the gridification success rate (i.e., in the limiting expression of the gridification accuracy index) and the gridification timestamp (i.e., in the limiting expression of the gridification accuracy index), and combine the obtained quality control evaluation index (i.e., in the limiting expression of the gridification accuracy index) and the fraction of the number of consistent peak positions to obtain the gridification accuracy index; otherwise, obtain the gridification timestamp and combine the obtained quality control evaluation index and the fraction of the number of consistent peak positions to obtain the gridification accuracy index.

[0056] Among them, the fraction of the number of consistent peak positions represents the ratio of the difference between the number of consistent peak positions (indicating the number of successful matches between the peaks of dual-polarization phased array radar data and the corresponding reference peak positions in the Cartesian coordinate system) and the preset number of consistent peak positions to the preset number of consistent peak positions; the time change coefficient represents the variation amplitude of dual-polarization phased array radar data over time; the reference time change coefficient represents the variation amplitude of reference dual-polarization phased array radar data over time; the gridification success rate represents the ratio of the amount of data successfully nested and gridified by dual-polarization phased array radar data within the preset time period to the total amount of dual-polarization phased array radar data to be nested and gridified; the gridification timestamp is obtained from the actual nested gridification duration (i.e., the total duration of completing nested gridification) of dual-polarization phased array radar data and the reference nested gridification duration.

[0057] The specific limiting expression of the gridification accuracy index is:

[0058] ;

[0059] ;

[0060] ;

[0061] ;

[0062] where t is the serial number of the preset time period, , T is the total number of preset time periods, e is the natural constant, represents the grid accuracy index during the nested grid process of dual-polarization phased array radar data in the t-th preset time period, represents the quality control evaluation index during the quality control process of dual-polarization phased array radar data in the t-th preset time period, represents the preset quality control evaluation index, represents the peak position consistency quantity fraction during the nested grid process of dual-polarization phased array radar data in the t-th preset time period, represents the number of consistent peak positions during the nested grid process of dual-polarization phased array radar data in the t-th preset time period, represents the preset number of consistent peak positions, represents the grid success rate during the nested grid process of dual-polarization phased array radar data in the t-th preset time period, represents the amount of successfully nested grid data during the nested grid process of dual-polarization phased array radar data in the t-th preset time period, represents the total amount of dual-polarization phased array radar data to be nested and gridded, represents the grid timestamp during the nested grid process of dual-polarization phased array radar data in the t-th preset time period, represents the actual nested grid duration during the nested grid process of dual-polarization phased array radar data in the t-th preset time period, represents the reference nested grid duration, represents the time change coefficient during the nested grid process of dual-polarization phased array radar data in the t-th preset time period, represents the reference time change coefficient.

[0063] In this embodiment, the preset number of consistent peak positions is represented by the result of summing and averaging the historical number of consistent peak positions of dual-polarization phased array radar data in the historical time period in the preset database, the reference time change coefficient is represented by the result of summing and averaging the historical time change coefficients of dual-polarization phased array radar data in the historical time period in the preset database, and the reference nested grid duration is represented by the result of summing and averaging the historical nested grid durations of dual-polarization phased array radar data in the preset database, which improves the accuracy and reliability of obtaining the grid accuracy index.

[0064] It should be understood that when When the gridification accuracy index increases with the increase of the quality control evaluation index, the fraction of the number of consistent peak positions, and the gridification success rate, and decreases with the increase of the gridification timestamp. Among them, the fraction of the number of consistent peak positions increases with the increase of the number of consistent peak positions, the gridification success rate increases with the increase of the amount of data successfully nested in the gridification, and the gridification timestamp decreases with the increase of the actual nested gridification duration.

[0065] It should be noted that (1) the quality control evaluation index also indirectly affects the value of the fraction of the number of consistent peak positions. Suppose there are two sets of data (i.e., dual-polarization phased array radar data) for tornado intelligent recognition. The quality control evaluation index of the first set of data is 1.6, and the quality control evaluation index of the second set of data is 1.4. At this time, in the process of identifying the peak position in the Cartesian coordinate system, the accuracy of identifying the peak position of the tornado by the first set of data is greater than that of the first set of data. This means that with the improvement of the quality control evaluation index, the fraction of the number of consistent peak positions will also increase accordingly.

[0066] (2) The fraction of the number of consistent peak positions also indirectly affects the values of the gridification success rate and the gridification timestamp. When the fraction of the number of consistent peak positions increases, it means that there are more aligned peak positions, that is, the accuracy of the peak position is improved, providing a more accurate reference point for nested gridification, and the corresponding gridification success rate also increases accordingly.

[0067] At this time, due to the improvement of the accuracy of peak position recognition, the nested gridification will be more efficient, thus reducing the time required for nested gridification (i.e., the gridification timestamp decreases). Through the above indirect influence mechanism, not only the accuracy of tornado intelligent recognition is improved, but also the correlation of multi-source data fusion is enhanced, and then the improvement of the correlation between the tornado intelligent recognition result and multi-source data fusion is realized, effectively solving the problem of low correlation between the tornado intelligent recognition result and multi-source data fusion in the prior art.

[0068] Further, the specific process of determining whether to obtain radar grid data based on the obtained gridification accuracy index is as follows: Determine whether the obtained gridification accuracy index is within the range of the gridification accuracy threshold:

[0069] If the obtained gridification accuracy index is within the range of the gridification accuracy threshold (including the cases of being equal to the maximum and minimum values of the historical gridification accuracy index), then complete the nested gridification and obtain the radar grid data; if the obtained gridification accuracy index is not within the range of the gridification accuracy threshold, then do not obtain the radar grid data and continue with the nested gridification until the obtained gridification accuracy index is within the range of the gridification accuracy threshold and then obtain the radar grid data.

[0070] In this embodiment, the grid accuracy threshold range represents the range corresponding to the maximum and minimum values of the historical grid accuracy metrics of the dual-polarized phased array radar data in the preset database during the historical time period; in this example, by setting the grid accuracy threshold range, it is ensured that only the radar grid data meeting the expected accuracy (set by the preset personnel) is acquired and stored, which helps to avoid acquiring and storing low-quality radar grid data under unnecessary circumstances, thereby saving storage space and processing resources.

[0071] Furthermore, the intelligent recognition evaluation score is obtained through the following method: when it is determined whether the acquired grid accuracy metric is within the grid accuracy threshold range, the real-time response speed of the radar grid data is monitored in real time to obtain the average response speed score during the intelligent recognition process within the preset time period (i.e., in the limiting expression of the intelligent recognition evaluation score), and the average response speed score is the ratio of the difference between the average response speed (recorded in real time by the deployed speed sensor and used to reflect the response rate of the radar system to the radar grid data) and the reference average response speed to the average response speed reference deviation; at the end of the preset time period, the amount of missed recognition data is acquired and it is determined whether the amount of missed recognition data is less than the maximum allowable amount of missed recognition data. If so, the missed recognition rate score (i.e., in the limiting expression of the intelligent recognition evaluation score) is obtained and the intelligent recognition for the next time period is executed. Otherwise, the intelligent recognition is performed again. The missed recognition rate score is the ratio of the amount of missed recognition data to the total amount of radar grid data to be intelligently recognized; the signal-to-noise ratio of the radar grid data during the intelligent recognition process within the preset time period is obtained (i.e., in the limiting expression of the intelligent recognition evaluation score), and at the same time, in combination with the acquired average response speed score, missed recognition rate score, and grid accuracy metric (i.e., in the limiting expression of the intelligent recognition evaluation score), the intelligent recognition evaluation score is obtained.

[0072] In this embodiment, the specific limiting expression of the intelligent recognition evaluation score is:

[0073] ;

[0074] ;

[0075] ;

[0076] In the formula, t is the number of the preset time period, , T is the total number of preset time periods, e is the natural constant, represents the intelligent recognition evaluation score during the intelligent recognition process of the radar grid data in the t-th preset time period, Indicates the grid accuracy index during the nested grid process of dual-polarization phased array radar data within the t-th preset time period. Indicates the range of grid accuracy thresholds. Indicates the signal-to-noise ratio during the intelligent recognition process of radar grid data within the preset time period. Indicates the average response speed score during the intelligent recognition process of radar grid data within the t-th preset time period. Indicates the average response speed during the intelligent recognition process of radar grid data within the t-th preset time period. Indicates the reference average response speed. Indicates the reference deviation of the average response speed. Indicates the missed recognition rate score during the intelligent recognition process of radar grid data within the t-th preset time period. Indicates the amount of missed recognition data during the intelligent recognition process of radar grid data within the t-th preset time period. Indicates the total amount of radar grid data to be intelligently recognized. Indicates the maximum allowable amount of missed recognition data.

[0077] Among them, the reference average response speed is represented by the result of summing and averaging the historical average response speeds of radar grid data in the preset database within the historical time period, the reference deviation of the average response speed is represented by the result of summing and averaging the historical average response speed deviations of radar grid data in the preset database within the historical time period, and the maximum allowable amount of missed recognition data is the maximum value of the missed recognition data amounts of radar grid data in the preset database (i.e., the historical database after completing intelligent recognition) within the historical time period.

[0078] Specifically, assuming that the range of grid accuracy thresholds is between 1 and 2, (assuming that the total amount of radar grid data to be intelligently recognized within the preset time period is 100), the maximum allowable amount of missed recognition data is 10, and the change statistical table of the intelligent recognition evaluation score is shown in Table 2:

[0079] Table 2 Change Statistical Table of Intelligent Recognition Evaluation Score

[0080]

[0081] It should be understood that from the first group of data and the second group of data in Table 2, it can be seen that the intelligent recognition evaluation score decreases as the missed recognition rate score increases, and from the third group of data, the fourth group of data, and the fifth group of data in Table 2, it can be seen that it increases as the grid accuracy index, the signal-to-noise ratio, and the average response speed score increase. Among them, the average response speed score increases as the average response speed deviation (i.e., ) increases, and the missed recognition rate score increases as the amount of missed recognition data increases.

[0082] It should be noted that: (1) The grid accuracy index also indirectly affects the value of the average response speed score. When the grid accuracy index decreases, it indicates that the deviation between the dual-polarization phased array radar data after nested gridification and the radar grid data increases. This deviation may cause misjudgment or delay in the radar system when identifying tornadoes, and further lead to a delay in the response rate of the radar system to the radar grid data (i.e., the average response speed deviation increases, and the corresponding average response speed score also increases accordingly).

[0083] (2) The average response speed score also indirectly affects the value of the missed identification rate score. When the average response speed score increases, it indicates that the radar system needs more time to analyze and identify after receiving the radar grid data, which may cause some tornado events to be missed during the identification process, further exacerbating the risk of missed identification (i.e., increasing the number of missed identifications of radar grid data).

[0084] From the above analysis, it can be seen that there is an interaction relationship among the grid accuracy index, the average response speed score, and the missed identification rate score. By considering the above indirect influence mechanism, the relevance of these indexes can be improved, which helps to optimize the performance of the tornado intelligent identification system, and further improves the relevance between the tornado intelligent identification result and the multi-source data fusion, effectively solving the problem of low relevance between the tornado intelligent identification result and the multi-source data fusion in the prior art.

[0085] Further, the specific process of determining whether the intelligent identification is completed according to the obtained intelligent identification evaluation score is as follows: Determine whether the obtained intelligent identification evaluation score is greater than the preset intelligent identification evaluation score. If the obtained intelligent identification evaluation score is greater than the preset intelligent identification evaluation score, the intelligent identification is completed and the radar grid data for which the intelligent identification is completed is marked as having identified a tornado. If the obtained intelligent identification evaluation score is not greater than the preset intelligent identification evaluation score, the corresponding radar grid data is marked as not having identified a tornado.

[0086] In this embodiment, the preset intelligent identification evaluation score is represented by the result of summing and averaging the historical intelligent identification evaluation scores of the radar system in the historical time period in the preset database. As Figure 6 shown, it is the roadmap of tornado intelligent identification provided by the embodiment of the present application. The intelligent identification result and the data processing status are visually displayed through the display screen of the radar system, which helps the decision maker quickly understand the current situation of tornado intelligent identification, realizes automated and intelligent data management, and further improves the relevance between the tornado intelligent identification result and the multi-source data fusion.

[0087] In summary, in the embodiments of the present application, by obtaining dual-polarization phased array radar data in a preset waveband and performing quality control to obtain a quality control evaluation index, then inputting the obtained multi-source real-time data into a constructed severe convective weather identification model to obtain areas prone to severe convective weather, and finally determining whether to obtain radar grid data based on the obtained grid accuracy index, and at the same time determining whether intelligent identification is completed according to the obtained intelligent identification evaluation score, the more accurate acquisition of the intelligent identification evaluation score is realized, and further the improvement of the correlation between the tornado intelligent identification result and multi-source data is realized, effectively solving the problem of low correlation between the tornado intelligent identification result and multi-source data in the prior art.

[0088] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0089] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, and the combination of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0090] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means implements the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0091] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide means for implementing the functions specified in Figure 1steps of one process or multiple processes and / or boxes Figure 1 steps of the functions specified in one box or multiple boxes.

[0092] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications to these embodiments once they learn the basic creative concepts. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications falling within the scope of the present invention.

[0093] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these changes and modifications.

Claims

1. Multi-source data fusion and tornado intelligent identification method, characterized in that: The following steps are involved: Step 1: Acquire dual-polarization phased array radar data of a preset band and perform quality control to obtain a quality control evaluation index, and determine whether to acquire multi-source real-time data based on the obtained quality control evaluation index, wherein the quality control evaluation index is used to evaluate the error correction effect of the dual-polarization phased array radar data within a preset time period; Step 2: Input the acquired multi-source real-time data into the established severe convection identification model to obtain the severe convective weather prone area, and at the same time, nested gridding the dual-polarization phased array radar data in the severe convective weather prone area to obtain a gridding accuracy index, which is used to evaluate the accuracy and reliability of nested gridding of dual-polarization phased array radar data; Step 3: determine whether to obtain radar grid point data based on the obtained grid accuracy index; if radar grid point data is obtained, then perform intelligent recognition on the obtained radar grid point data to obtain an intelligent recognition evaluation score; and determine whether intelligent recognition is completed based on the obtained intelligent recognition evaluation score, wherein the intelligent recognition evaluation score is used to evaluate the degree of match between the radar grid point data and the intelligent recognition condition; Step 4: If intelligent identification is completed, the membership function is designed according to the acquired radar grid data and the tornado warning duration is obtained by combining the fuzzy logic method; The specific steps of obtaining the grid accuracy index include: E1, real-time monitoring of the nested gridding of the dual-polarization phased array radar data in the Cartesian coordinate system, and judging whether the number of consistent peak positions within a preset time period is greater than the preset number of consistent peak positions. If so, execute E2, otherwise re-nest gridding; E2, obtain the peak position consistent number score, and determine whether the time variation coefficient is equal to the reference time variation coefficient. If so, obtain the gridding success rate and gridding timestamp and combine the obtained quality control evaluation index and the peak position consistent number score to obtain the gridding accuracy index. Otherwise, obtain the gridding timestamp and combine the obtained quality control evaluation index and the peak position consistent number score to obtain the gridding accuracy index. The peak position consistent number score represents the ratio of the difference between the peak position consistent number and the preset peak position consistent number to the preset peak position consistent number; The time variation coefficient represents the variation amplitude of the dual-polarization phased array radar data over time; The gridding success rate indicates the ratio of the amount of dual-polarization phased array radar data successfully gridded within a preset time period to the total amount of dual-polarization phased array radar data to be gridded.

2. The multi-source data fusion and tornado intelligent identification method according to claim 1, characterized in that: The dual-polarization phased array radar data includes reflectivity factors, radial velocities, differential reflectivity factors, and differential propagation phase shifts of horizontal polarization and vertical polarization; The severe convective weather prone area refers to an area where the probability of occurrence of severe convective weather is greater than a preset probability of occurrence; The nested gridding is used to convert the dual-polarization phased array radar data into a regular grid form.

3. The multi-source data fusion and tornado intelligent identification method according to claim 1, characterized in that: The quality control evaluation index is obtained by obtaining the signal-to-noise ratio, ground object echo removal efficiency, isolated echo removal efficiency and electromagnetic interference suppression efficiency of the dual-polarization phased array radar data during the quality control process within a preset time period; The ground object echo removal efficiency represents the ratio of the difference between the initial ground object echo intensity before quality control of the dual-polarization phased array radar data within a preset time period and the actual ground object echo intensity after quality control within the preset time period to the initial ground object echo intensity; The isolated echo removal efficiency represents the ratio of the difference between the initial isolated echo number before quality control of the dual-polarization phased array radar data within a preset time period and the actual isolated echo number after quality control within the preset time period to the initial isolated echo number; The electromagnetic interference suppression efficiency indicates a ratio of a difference between an initial electromagnetic interference intensity of dual-polarization phased array radar data before quality control within a preset time period and an actual electromagnetic interference intensity after quality control within the preset time period to the initial electromagnetic interference intensity.

4. The multi-source data fusion and tornado intelligent identification method according to claim 1, characterized in that: The specific process of judging whether to acquire multi-source real-time data based on the acquired quality control evaluation index is as follows: Determine whether the obtained quality control evaluation index is greater than the preset quality control evaluation index: If the obtained quality control evaluation index is greater than the preset quality control evaluation index, multi-source real-time data is obtained; If the obtained quality control evaluation index is not greater than the preset quality control evaluation index, the multi-source real-time data is not obtained and the preset band parameters are adjusted; The preset band parameters include the center frequency of the band, the scanning mode of the radar, and the pulse repetition frequency.

5. The multi-source data fusion and tornado intelligent identification method as claimed in claim 1, characterized in that: The specific steps of nesting and gridding the dual-polarization phased array radar data in the area prone to severe convective weather are as follows: Generate radar information based on the dual-polarization phased array radar data after quality control, and form three-dimensional grid point information based on it; The radar information is used to visually describe weather phenomena in the coverage area of ​​the dual-polarization phased array radar.

6. The multi-source data fusion and tornado intelligent identification method according to claim 1, characterized in that: The specific limiting expression of the grid accuracy index is: ； Where t is the number of the preset time period, , T is the total number of preset time periods, e is a natural constant, It represents the gridding accuracy index of the dual-polarization phased array radar data in the nested gridding process within the t-th preset time period. represents the quality control evaluation index of the dual-polarization phased array radar data in the quality control process in the t-th preset time period, represents the preset quality control evaluation index, It represents the number of consistent peak positions of dual-polarization phased array radar data in the nested gridding process in the t-th preset time period, It represents the number of peak position coincidences of the dual-polarization phased array radar data in the nested gridding process in the t-th preset time period, Indicates the preset peak position consistency number, represents the success rate of gridding during the nested gridding process of dual-polarization phased array radar data in the tth preset time period, Represents the gridded timestamp of the dual-polarization phased array radar data during the nested gridding process in the tth preset time period, It represents the time variation coefficient of the dual-polarization phased array radar data in the process of nested gridding in the t-th preset time period, Indicates the reference time variation coefficient.

7. The multi-source data fusion and tornado intelligent identification method as claimed in claim 1, characterized in that: The specific process of determining whether to obtain radar grid data based on the obtained grid accuracy index is as follows: Determine whether the obtained grid accuracy index is within the grid accuracy threshold range: If the obtained gridding accuracy index is within the gridding accuracy threshold range, the nested gridding is completed and the radar grid data is obtained; If the obtained gridding accuracy index is not within the gridding accuracy threshold range, the radar grid data will not be obtained and the nested gridding will continue.

8. The multi-source data fusion and tornado intelligent identification method as claimed in claim 1, characterized in that: The intelligent recognition evaluation score is obtained by the following method: Real-time monitoring of the real-time response speed of radar grid point data to obtain an average response speed score in the intelligent recognition process within a preset time period, wherein the average response speed score is a ratio of a difference between the average response speed and a reference average response speed to a reference deviation of the average response speed; At the end of the preset time period, the amount of missed identification data is obtained and it is determined whether the amount of missed identification data is less than the maximum allowable amount of missed identification data. If so, the missed identification rate score is obtained and intelligent identification of the next time period is performed, otherwise, intelligent identification is performed again. The missed identification rate score is the ratio of the amount of missed identification data to the total amount of radar grid point data to be intelligently identified. The signal-to-noise ratio of radar grid data in the intelligent recognition process within a preset time period is obtained, and the intelligent recognition evaluation score is obtained by combining the obtained average response speed score, missed recognition rate score and grid accuracy index.

9. The multi-source data fusion and tornado intelligent identification method as claimed in claim 1, characterized in that: The specific process of judging whether the intelligent recognition is completed according to the obtained intelligent recognition evaluation score is as follows: Determine whether the obtained intelligent recognition evaluation score is greater than the preset intelligent recognition evaluation score: If the obtained intelligent recognition evaluation score is greater than the preset intelligent recognition evaluation score, the intelligent recognition is completed and the radar grid point data that has completed the intelligent recognition is marked as a recognized tornado; If the obtained intelligent recognition evaluation score is not greater than the preset intelligent recognition evaluation score, the corresponding radar grid point data is marked as an unidentified tornado.

Images