A method and device for identifying and predicting light rainfall weather processes

Abstract

The present application discloses a method and device for identifying and predicting micro-rainfall weather processes, which belongs to the field of environmental monitoring technology. The method collects cloud data from Ka-band cloud radar, precipitation data from X-band rain radar, and data from automatic weather stations. A nonlinear model is established, and the data output by the nonlinear model are grouped according to the time axis and radar type. A fitting curve for the micro-rainfall weather process is established to calculate the coefficient of determination and generate eigenvalues, and the prediction result is determined based on the eigenvalues. The output data of the nonlinear model is corrected based on the data from the automatic weather station and the eigenvalues, and the corrected coefficient of determination and eigenvalues ​​are input into the historical data model library as prediction conditions. The judgment results under the same conditions are screened out and compared with the prediction results. If the comparison result is valid, an alarm information push is generated. The method and device for identifying and predicting micro-rainfall weather processes provided by the present application realize the full-process observation and prediction of water vapor-cloud-precipitation, significantly reducing the false alarm rate.

Description

Technical Field

[0001] The present application belongs to the field of environmental monitoring technology, and in particular relates to a method and device for identifying and predicting light rainfall weather processes. Background Art

[0002] At present, the identification of rainfall mainly relies on manual observation, automatic weather stations, X-band rainfall radar monitoring and other means.

[0003] Manual observation mainly relies on experienced meteorological observers to record and update weather data of the current weather process through naked eyes and physical perception. This method is relatively primitive and is limited by the number of personnel and the inability to observe in real time 24 hours a day, so real-time and accuracy cannot be guaranteed.

[0004] Automatic weather stations can simply record the current rainfall process by collecting precipitation, but their accuracy is poor and they cannot make predictions.

[0005] The X-band rain radar operates in the 9.1GHz-9.5GHz frequency band. Although it can detect precipitation particles, it cannot accurately identify small-sized cloud particles due to its long wavelength. In addition, the signal is severely attenuated in heavy rain, resulting in data distortion.

[0006] Currently, the existing technology has the following technical defects:

[0007] 1. Manual observation and automatic station observation depend on the quality and experience of personnel, as well as the working mode and principle of the equipment, and cannot meet the existing observation needs and accuracy.

[0008] 2. X-band rainfall radar observation transmits radar echo data to the client terminal in real time. However, it is often impossible to obtain valid data due to the lack of observation or excessive rain attenuation. Although Ka-band radar can detect cloud particles, the signal attenuation is severe during heavy rainfall.

[0009] 3. Traditional methods can only reflect real-time status and lack the ability to model the micro-rainfall generation / dissipation process based on historical data. Summary of the Invention

[0010] In order to solve the technical defects of the above-mentioned existing technologies, the present application provides a method for identifying and predicting micro-rainfall weather processes. By integrating Ka-band cloud radar, X-band rain radar and automatic weather station data, it realizes the full-process observation and prediction of water vapor-cloud-precipitation, significantly reducing the false alarm rate.

[0011] The technical solution is as follows:

[0012] On the one hand, a method for identifying and predicting a light rainfall weather process is provided, comprising:

[0013] Collect cloud data from Ka-band cloud radar, precipitation data from X-band rain radar, and humidity and hourly precipitation data from automatic weather stations;

[0014] Construct nonlinear models of Ka-band cloud radar and X-band rain radar respectively;

[0015] The data output by the nonlinear model is grouped by time axis and radar type to generate a data set;

[0016] Based on the data set, a fitting curve of the micro-rainfall weather process is established, the determination coefficient is calculated and the characteristic value is generated, and the prediction result is determined according to the characteristic value;

[0017] Correcting the nonlinear model output data according to the automatic weather station data and the eigenvalues, and then calculating the corrected coefficient of determination and eigenvalues;

[0018] The corrected determination coefficient and eigenvalue are input into the historical data model library as the conditions for identifying and predicting the current micro-rainfall weather process. The judgment results in the historical data model library under the same conditions are screened out and compared with the prediction results. If the comparison result is valid, an alarm information push is generated.

[0019] Preferably, the nonlinear model is:

[0020]

[0021] in, is the explained variable, is the explanatory variable, are model parameters, is the disturbance term, the disturbance term is the humidity and hourly precipitation of the automatic weather station, ( ,..., ) is a nonlinear function, It is a distance library based on time axis.

[0022] Preferably, the disturbance term The value of is based on the humidity and hourly precipitation from the automatic weather station.

[0023] Preferably, a weather process fitting curve is established based on the data set, a determination coefficient is calculated and a characteristic value is generated, and a prediction result is determined according to the characteristic value, including:

[0024] Establish a fitting curve and calculate the coefficient of determination ,

[0025] SSE

[0026] SST=

[0027] SSE / SST

[0028] in, is the coefficient of determination, is the explained variable, is the data fitting value, is the data average, SSE is the residual sum of squares, SST is the total sum of squares of deviations;

[0029] According to the coefficient of determination Generate eigenvalues W , and according to the eigenvalue W Determine the prediction result:

[0030] when <0.2, eigenvalue W =1, the prediction result is sunny stage;

[0031] When 0.2≤ <0.4, eigenvalue W =2, the prediction result is the light cloud stage;

[0032] When 0.4≤ <0.6, eigenvalue W =3, the prediction result is cloudy stage;

[0033] When 0.6≤ <0.8, eigenvalue W =4, the prediction result is light rain stage;

[0034] when ≥0.8, eigenvalue W =5, the predicted result is medium to heavy rain.

[0035] Preferably, according to the automatic weather station data and characteristic values W , correcting the nonlinear model output data, and then calculating the corrected determination coefficient and eigenvalue, including:

[0036] When the eigenvalue W =1, the Ka-band echo data is embedded;

[0037] When the eigenvalue W =2, the Ka-band echo data is embedded;

[0038] When the eigenvalue W =3, the Ka band provides the main echo data, and the X band provides the supplementary data;

[0039] When the eigenvalue W =4, Ka-band and X-band provide echo data and correct each other;

[0040] When the eigenvalue W =5, the X-band provides the main echo data and the Ka-band provides the data correction.

[0041] Preferably, when the eigenvalue W =1, the Ka-band clutter is cleared or eliminated through the clutter filtering algorithm and then the Ka-band echo data is implanted.

[0042] Preferably, the clutter filtering algorithm includes an isolated noise point echo elimination algorithm, a floating object clutter filtering algorithm and a clear sky clutter elimination algorithm.

[0043] Preferably, the method further includes: when the accuracy of the prediction result under the same conditions is greater than 80%, the comparison result is valid and an alarm information is generated and pushed.

[0044] Preferably, when the comparison result is valid, the new sample data is stored in the historical data model library to update and improve the historical data model library.

[0045] On the other hand, a radar light rainfall weather process identification and prediction device is provided, which adopts the above-mentioned radar light rainfall weather process identification and prediction method, including:

[0046] The acquisition module is used to collect cloud data from the Ka-band cloud radar, precipitation data from the X-band rain radar, and humidity and hourly precipitation data from the automatic weather station;

[0047] Nonlinear model building module, used to build nonlinear models of Ka-band cloud radar and X-band rain radar respectively;

[0048] The correction module is used to group the data output by the nonlinear model by time axis and radar type to generate a data set; establish a micro-rainfall weather process fitting curve based on the data set, calculate the determination coefficient and generate the eigenvalue, and determine the prediction result based on the eigenvalue; correct the nonlinear model output data based on the automatic weather station data and the eigenvalue, and then calculate the corrected determination coefficient and eigenvalue;

[0049] The comparison module is used to input the corrected determination coefficient and eigenvalue as the conditions for identifying and predicting the current micro-rainfall weather process into the historical data model library, screen out the judgment results in the historical data model library under the same conditions and compare them with the prediction results. If the comparison result is valid, an alarm information push is generated.

[0050] The technical solution includes at least the following technical effects:

[0051] 1. Overcome the functional limitations of a single radar, such as poor X-band penetration, poor detection capabilities for clouds, fog, and other microparticles, and significant attenuation of heavy rainfall in the Ka-band, to achieve closed-loop monitoring of the entire water vapor-cloud-precipitation process;

[0052] 2. Relying on the real-time matching mechanism of the historical model library, predictions can be made at least 30 minutes in advance, allowing the complete precipitation generation or dissipation process to be seen, reducing the occurrence of misjudgments and false alarms, greatly improving the monitoring of weather processes, and reducing losses caused by misjudgments;

[0053] 3. Through dynamic correction of disturbance terms and dual-band mutual correction strategy, data distortion caused by rain attenuation can be effectively suppressed;

[0054] 4. Innovatively use dual-band radar to establish a nonlinear data model, combine it with automatic weather station data correction, and compare it with the big data model library of historical data. The results are verified and corrected multiple times. The correct results will be sent back to the historical database to improve the sample cases, greatly improving the accuracy of the model and the system's ability to observe and predict disaster weather in real time, reduce unnecessary losses, etc.

[0055] It should be understood that the foregoing general description and the following detailed description are merely illustrative and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0057] Figure 1 A flow chart of a method for identifying and predicting light rainfall weather processes provided by a preferred embodiment;

[0058] Figure 2 A Z fitting curve diagram of echo intensity provided by a preferred embodiment;

[0059] Figure 3 A radial velocity V fitting curve diagram provided for a preferred embodiment;

[0060] Figure 4 Schematic diagram of data complementation when the Ka-band and X-band echo data are provided and corrected with each other when the eigenvalue W=4;

[0061] Figure 5 The diagram shows that when the eigenvalue W=5, the X-band rain radar provides the main echo data and the Ka-band cloud radar provides the data correction.

[0062] Figure 6 A schematic diagram of a device for identifying and predicting light rainfall weather processes provided by a preferred embodiment. DETAILED DESCRIPTION

[0063] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0064] Example 1

[0065] As attached Figure 1 As shown, the present application provides a method for identifying and predicting micro-rainfall weather processes, which is based on the method for identifying and predicting micro-rainfall weather processes based on Ka and X-band radars. A nonlinear data model is established by establishing a dual-band radar, combined with automatic weather station data correction, and compared with a big data model library of historical data. The results are verified and corrected multiple times, and the correct results are transmitted back to the historical database to improve the sample cases, greatly improving the accuracy of the model and the system's ability to observe disaster weather in real time and predict it in advance, reducing unnecessary losses, etc.

[0066] Radar micro-rainfall weather process identification and prediction refers to the critical process before cloud formation and rainfall, and makes rainfall predictions in advance.

[0067] A method for identifying and predicting light rainfall weather processes, the specific steps are as follows:

[0068] Step S1: Collect cloud data from a Ka-band cloud radar, precipitation data from an X-band rain radar, and humidity and hourly precipitation data from an automatic weather station. The Ka-band cloud radar obtains echo intensity Z, radial velocity V, and integrated liquid water content VIL; the X-band rain radar obtains filtered reflectivity R, radial velocity V, and integrated liquid water content VIL; and the automatic weather station obtains relative humidity and hourly precipitation.

[0069] Step S2: Establish nonlinear models for Ka-band cloud radar and X-band rain radar respectively, as follows:

[0070] In the nonlinear model, the echo intensity Z, radial velocity V, and integrated liquid water content VIL data models of Ka-band cloud radar and X-band rain radar are established respectively, and the humidity and hourly precipitation of the automatic weather station are used as disturbance terms. After the real observed data enters the nonlinear model, the processed data is obtained through the influence of the nonlinear model parameters and disturbance terms. .

[0071] The nonlinear data model is:

[0072]

[0073] in, is the explained variable, is the explanatory variable, are model parameters, is the disturbance term, ( ,..., ) is a nonlinear function, It is a distance library based on time axis.

[0074] Step S3: Grouping the data output by the nonlinear model by time axis and radar type to generate a data set. In this embodiment, the time axis refers to grouping the data of the Ka-band cloud radar and the X-band rain radar according to the same time, ensuring that the data observed by these two radars are at the same time, that is, ensuring time synchronization, to avoid large differences in the data observed by the radars at different times.

[0075] Step S4: Establish a fitting curve for the micro-rainfall weather process based on the data set and calculate the coefficient of determination And generate the eigenvalue W, and determine the prediction result based on the eigenvalue. The specific steps are as follows:

[0076] Establish Z, V, VIL fitting curve and calculate the determination coefficient ,

[0077] SSE

[0078] SST=

[0079] SSE / SST

[0080] in, is the coefficient of determination, is the data fitting value, is the data average, SSE is the residual sum of squares, SST is the total sum of squares of deviations, is the explained variable;

[0081] By the coefficient of determination Generate eigenvalues W , according to the eigenvalue W Determine the predicted results.

[0082]

[0083] like Figure 2 and Figure 3 The fitting curve diagram shown in the figure, wherein the horizontal axis X is the distance library, the vertical axis Y is the height value of the space where the sample is located, and the fitting curves are the echo intensity Z fitting curve, the radial velocity V fitting curve, and the integrated liquid water content VIL fitting curve.

[0084] Step S5: Based on the automatic weather station data and characteristic values W , correcting the nonlinear model output data, and then calculating the corrected determination coefficient and eigenvalue;

[0085] In sunny weather, Ka-band radar can detect clear sky clutter. In light cloud, Ka-band radar can also detect clearly. In cloudy or light rain, Ka-band radar can detect multi-layer cloud structure efficiently and accurately, while X-band can only detect weak echo signals in light rain. In moderate to heavy rain, due to the shorter wavelength of Ka-band, the rain attenuation is larger and the detection capability is weakened, while X-band radar is more accurate. Therefore, the source data is corrected through the discrimination of eigenvalues ​​and the assistance of automatic stations. In this process, the data of automatic weather stations are only provided for reference. The generated eigenvalues ​​are confirmed. W Is there a large deviation to confirm the eigenvalue W authentic reliability.

[0086] The correction method is:

[0087] When the eigenvalue W=1, the echo data is completely embedded in the echo map of the Ka-band cloud radar. You can first select the clutter filtering algorithm to clear the clutter display or eliminate the Ka-band clutter and then embed the Ka-band echo data.

[0088] Clutter filtering algorithms include isolated / noise point echo elimination algorithm, floating object clutter filtering algorithm and clear sky clutter elimination algorithm.

[0089] This embodiment uses an isolated / noise point echo cancellation algorithm. Insects fly at low altitudes and are generally within detection blind spots, so they can be ignored. Birds and airplanes generally have larger RCSs, but their radial dimensions generally do not exceed 30 meters, corresponding to one or two range bins. This manifests as one or two isolated strong points on the echo waveform, which can be removed using a low-pass filter. This embodiment uses the "K-domain frequency method," which uses the continuous variation of the statistical characteristics of adjacent regions to filter out some isolated noise points or radial interference, which has a positive effect on subsequent calculations.

[0090] When the eigenvalue W = 2, the echo data is implanted by the Ka-band cloud radar;

[0091] When the eigenvalue W=3, the Ka-band provides the main echo data, and the X-band provides the supplementary data, which mainly supplements the data lost due to the wavelength characteristics of the Ka-band, such as the bright band data of the zero-degree layer and the ice-water mixture data above the zero-degree layer.

[0092] When the eigenvalue W = 4, the Ka- and X-band radar data provide echo data and correct each other;

[0093] like Figure 4 As shown in the figure, when the eigenvalue W=4, it is the light rain stage. At this time, the Ka-band radar will be affected by the rain particles due to its wavelength characteristics and will be attenuated to a certain extent. However, the X-band rain radar has a weak echo intensity due to the light rainfall, and is greatly affected by the clutter interference. Therefore, at this stage, the two radar data need to complement and correct each other. The specific method is that the X-band radar signal echo signal, in the precipitation zone ( Figure 4 The orange-red portion in the middle is used to supplement the attenuation of the Ka radar echo. The blue-green area uses the Ka echo to interpolate the X-band echo signal to supplement the weak X-band detection area. The yellow area is the particle area between clouds and rain, which is corrected by particle echo intensity and radial velocity.

[0094] like Figure 5 As shown in the figure, when the eigenvalue W=5, the X-band rain radar provides the main echo data, and the Ka-band cloud radar provides data correction, with the X-band as the main echo and the Ka-band as the correction. During the medium and heavy rain stage, the X-band rain radar detected local rainfall clusters in real time ( Figure 5 However, due to the limitations of the operating system and wavelength, clouds that did not form precipitation or clouds that dissipated after precipitation could not be fully monitored. At this time, the Ka-band radar cloud image was added. After data supplementation and correction, the complete cloud layer and rainfall distribution can be seen. Based on the analysis of the corrected data, the complete echo map can be seen. Figure 5 The location and distribution of clouds and rain clusters, as well as the process and direction of particle formation into clouds and rain, can be seen in the system, realizing a complete closed loop of real-time monitoring, real-time analysis, and early prediction.

[0095] Step S6: The corrected coefficient of determination and eigenvalue are input into the historical data model library as the conditions for identifying and predicting the current light rainfall weather process. The judgment results in the historical data model library under the same conditions are screened and compared with the prediction results. If the comparison result is valid, an alarm message is generated and pushed. The historical data model library is a large data model composed of historical data.

[0096] The specific method is to input the current conditions for identifying and predicting light rainfall weather (the characteristic values ​​and determination coefficients in step S4) into the historical data model library. Through query and matching, the accuracy of all the same judgment results that meet the conditions in the historical data model library (the same conditions, i.e., the same characteristic values ​​and determination coefficients) are counted. If the accuracy is greater than 80%, the judgment result is valid, an alarm information push is generated, and the new sample data is stored in the historical data model library. That is, the new sample data is transmitted back to the historical data model library for case filling.

[0097] The conditions for identifying and predicting the current light rainfall weather process and the determination coefficient =0.1, eigenvalue W=1, input into the historical data model library, through query and matching, match the conditions ( <0.2, W=1), there are 200 data, of which 186 corresponding judgment results (weather stage) are sunny, and 14 corresponding judgment results are partly cloudy. The accuracy of the judgment result is 93%, that is, the accuracy of the judgment result is greater than 80%. The judgment result is valid and the alarm information is pushed.

[0098] Alarm information push refers to the push of rainfall forecast information. Specifically, after the system generates the judgment result, it reports the current result of the weather and makes an advance prediction of the future rainfall process. The results will be fed back to customers in three forms: pop-up window on the software terminal interface, SMS push, and email. The push information includes the current meteorological conditions observation, rainfall trend forecast, disaster warning, etc.

[0099] Through collaborative observation by radars of different systems, the shortcomings of a single radar's single function are made up, and the strengths of different radar systems are utilized to the maximum extent, realizing the observation and prediction of the entire process of water vapor-cloud-precipitation. In terms of time, it can provide an early warning of at least half an hour in advance, and the complete precipitation generation or dissipation process can be seen, reducing the occurrence of misjudgments and false alarms, which can greatly improve the monitoring of weather processes and reduce the losses caused by misjudgments.

[0100] Example 2

[0101] like Figure 6 As shown, a radar light rainfall weather process identification and prediction device is provided, which adopts the radar light rainfall weather process identification and prediction method in embodiment 1, including:

[0102] The acquisition module 101 is used to collect cloud data from the Ka-band cloud radar, precipitation data from the X-band rain radar, and humidity and hourly precipitation data from the automatic weather station;

[0103] A nonlinear model construction module 102 is used to construct nonlinear models of Ka-band cloud detection radar and X-band rain detection radar respectively;

[0104] Correction module 103 is configured to group the data output by the nonlinear model by time axis and radar type to generate a data set; establish a fitting curve for the micro-rainfall weather process based on the data set, calculate a coefficient of determination and generate an eigenvalue, determine a prediction result based on the eigenvalue, correct the nonlinear model output data based on the automatic weather station data and the eigenvalue, and then calculate the corrected coefficient of determination and eigenvalue;

[0105] The comparison module 104 is used to input the corrected determination coefficient and eigenvalue as the conditions for identifying and predicting the current micro-rainfall weather process into the historical data model library, screen out the judgment results in the historical data model library under the same conditions and compare them with the prediction results. If the comparison result is valid, an alarm information push is generated.

[0106] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, apparatuses, or computer program products. Therefore, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0107] The present application is described with reference to the flowcharts and / or block diagrams of the methods, apparatus (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0108] These computer program instructions may 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 produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0109] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0110] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present application.

[0111] Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is intended to include these modifications and variations.

Claims

1. A method for identifying and predicting light rainfall weather processes, characterized in that: include: Collect cloud data from Ka-band cloud radar, precipitation data from X-band rain radar, and humidity and hourly precipitation data from automatic weather stations; Construct nonlinear models of Ka-band cloud radar and X-band rain radar respectively; The data output by the nonlinear model is grouped by time axis and radar type to generate a data set; Based on the data set, a fitting curve of the micro-rainfall weather process is established, the coefficient of determination is calculated and the eigenvalue is generated, and the prediction result is determined according to the eigenvalue, where: According to the coefficient of determination Generate eigenvalues W , and according to the eigenvalue W Determine the predicted results; when <0.2, eigenvalue W =1, the prediction result is sunny stage; When 0.2≤ <0.4, eigenvalue W =2, the prediction result is the light cloud stage; When 0.4≤ <0.6, eigenvalue W =3, the prediction result is cloudy stage; When 0.6≤ <0.8, eigenvalue W =4, the prediction result is light rain stage; when ≥0.8, eigenvalue W =5, the prediction result is medium to heavy rain stage; According to the automatic weather station data and eigenvalues, the nonlinear model output data is corrected, and then the corrected determination coefficient and eigenvalue are calculated, where: When the eigenvalue W =1, Ka-band echo data is embedded; When the eigenvalue W =2, Ka-band echo data is embedded; When the eigenvalue W =3, the Ka band provides the main echo data, and the X band provides the supplementary data; When the eigenvalue W =4, Ka-band and X-band provide echo data and correct each other; When the eigenvalue W =5, the X-band provides the main echo data, and the Ka-band provides the data correction; The corrected determination coefficient and eigenvalue are input into the historical data model library as the conditions for identifying and predicting the current micro-rainfall weather process. The judgment results in the historical data model library under the same conditions are screened out and compared with the prediction results. If the comparison result is valid, an alarm information push is generated.

2. The method for identifying and predicting light rainfall weather processes according to claim 1, characterized in that: The nonlinear model is: in, is the explained variable, is the explanatory variable, are model parameters, is the disturbance term, the disturbance term is the humidity and hourly precipitation of the automatic weather station, ( ,..., ) is a nonlinear function, It is a distance library based on time axis.

3. The method for identifying and predicting light rainfall weather processes according to claim 1, characterized in that: Establish a weather process fitting curve based on the data set, including: Establish a fitting curve and calculate the coefficient of determination , SSE SST= SSE / SST in, is the coefficient of determination, is the explained variable, is the data fitting value, is the data average, SSE is the residual sum of squares, SST is the total sum of squares of deviations.

4. The method for identifying and predicting light rainfall weather processes according to claim 1, wherein: When the eigenvalue W =1, the Ka-band echo data is implanted after clearing or eliminating the Ka-band clutter through the clutter filtering algorithm.

5. The method for identifying and predicting light rainfall weather processes according to claim 4, characterized in that: The clutter filtering algorithm includes an isolated noise point echo elimination algorithm, a floating object clutter filtering algorithm and a clear sky clutter elimination algorithm.

6. The method for identifying and predicting light rainfall weather processes according to claim 1, characterized in that: Also includes: When the accuracy of the prediction result under the same conditions is greater than 80%, the comparison result is valid and an alarm information push is generated.

7. The method for identifying and predicting light rainfall weather processes according to claim 6, characterized in that: When the comparison result is valid, the new sample data is stored in the historical data model library to update and improve the historical data model library.

8. A radar light rainfall weather process identification and prediction device, using the radar light rainfall weather process identification and prediction method according to any one of claims 1 to 7, characterized in that: It includes: The acquisition module is used to collect cloud data from the Ka-band cloud radar, precipitation data from the X-band rain radar, and humidity and hourly precipitation data from the automatic weather station; Nonlinear model building module, used to build nonlinear models of Ka-band cloud radar and X-band rain radar respectively; The correction module is used to group the data output by the nonlinear model by time axis and radar type to generate a data set; Establish a fitting curve for the micro-rainfall weather process based on the data set, calculate the coefficient of determination and generate the eigenvalue, determine the prediction result based on the eigenvalue, correct the nonlinear model output data based on the automatic weather station data and the eigenvalue, and then calculate the corrected coefficient of determination and eigenvalue; The comparison module is used to input the corrected determination coefficient and eigenvalue as the conditions for identifying and predicting the current micro-rainfall weather process into the historical data model library, screen out the judgment results in the historical data model library under the same conditions and compare them with the prediction results. If the comparison result is valid, an alarm information push is generated.

Images