Multi-source data fusion rainfall prediction method and device based on Beidou B2B signal and computer equipment

Abstract

The invention relates to a multi-source data fusion rainfall prediction method and device based on Beidou B2B signals and computer equipment. The method comprises the following steps: in response to a rainfall prediction request for a target observation area, performing time synchronization operation on each observation device of the multi-source observation system; acquiring a satellite signal received by the Beidou signal receiving equipment, resolving zenith total delay generated when the satellite signal passes through the atmosphere, and converting the zenith total delay into atmospheric precipitable water; inputting the atmospheric precipitable water amount, the radar rain rate of the weather radar, the satellite observation of the meteorological satellite and the ground observation of the ground sensor network as observation vectors into a preset humidity analysis model to obtain a three-dimensional humidity analysis field; and obtaining multi-source historical observation data, and inputting the three-dimensional humidity analysis field and the multi-source historical observation data into a pre-trained rainfall rate field prediction model to obtain a rainfall rate field prediction result in a preset time period. By adopting the method, the rainfall prediction precision can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of weather prediction, in particular to a multi-source data fusion rainfall prediction method and device based on Beidou B2B signals, computer equipment, storage medium and computer program product. BACKGROUND

[0002] With the frequent occurrence of extreme precipitation events, high-precision short-impending rainfall prediction is crucial for disaster prevention and reduction. Traditional rainfall prediction mainly relies on numerical weather prediction models and single observation data (such as weather radar), but the NWP (Numerical Weather Prediction) model has the problem of cumulative error of initial field, and radar observation is easily affected by terrain shielding and attenuation, resulting in high false negative rate of sudden heavy rain.

[0003] Therefore, the rainfall prediction in the prior art has the problem of low accuracy. SUMMARY

[0004] Therefore, it is necessary to provide a multi-source data fusion rainfall prediction method and device based on Beidou B2B signals, computer equipment, computer readable storage medium and computer program product, which can improve the rainfall prediction accuracy of the target observation area.

[0005] In a first aspect, the present application provides a multi-source data fusion rainfall prediction method based on Beidou B2B signals. The method comprises:

[0006] In response to a rainfall prediction request for a target observation area, performing time synchronization operation on each observation device of a multi-source observation system; the target observation area is an observation area after grid processing; the multi-source observation system comprises a weather radar, a meteorological satellite, a ground sensor network, and a Beidou signal receiving device arranged in the target observation area;

[0007] According to the satellite signal received by the Beidou signal receiving device, determine the pseudo-range observation information and the carrier phase observation information, according to the pseudo-range observation information and the carrier phase observation information, solve the zenith total delay generated by the satellite signal passing through the atmosphere, and convert the zenith total delay into atmospheric precipitable water content;

[0008] The atmospheric precipitable water content, the radar rain rate of the weather radar, the satellite observation of the meteorological satellite, and the ground observation of the ground sensor network are input into a preset humidity analysis model as observation vectors to obtain a three-dimensional humidity analysis field;

[0009] Obtain multi-source historical observation data, input the three-dimensional humidity analysis field and the multi-source historical observation data into a pre-trained rainfall rate field prediction model to obtain a rainfall rate field prediction result in a preset time period.

[0010] In one of the embodiments, the method further comprises:

[0011] obtaining a reflected echo signal received by the weather radar;

[0012] determining a radar reflectivity according to the reflected echo signal, and converting the radar reflectivity into a precipitation intensity according to a preset mapping relationship; the preset mapping relationship comprises a power law relationship between the radar reflectivity and the rain rate;

[0013] determining a radar rain rate of the weather radar according to the precipitation intensity.

[0014] In one of the embodiments, under a heavy precipitation condition, the method further comprises:

[0015] obtaining a radar detection distance of the weather radar, and determining an attenuation degree description information of a radar signal of the weather radar according to the radar detection distance and a preset path attenuation coefficient;

[0016] optimizing the radar rain rate of the weather radar according to the attenuation degree description information, to obtain an optimized radar rain rate.

[0017] In one of the embodiments, the converting the zenith total delay into the atmospheric precipitable water comprises:

[0018] adopting a preset delay decomposition model to decompose the zenith total delay into a zenith statics delay and a zenith wet delay;

[0019] converting the zenith wet delay into the atmospheric precipitable water.

[0020] In one of the embodiments, the method further comprises:

[0021] obtaining a multiplicative correction coefficient corresponding to a preset prediction time lag, determining a product between the rainfall rate field prediction result and the multiplicative correction coefficient, to obtain an adjusted prediction result; the multiplicative correction coefficient is fitted by a linear regression slope of a historical same-period prediction and an actual situation;

[0022] obtaining an additive correction term corresponding to the preset prediction time lag, determining a sum between the adjusted prediction result and the additive correction term, to obtain a corrected rainfall rate field prediction result;

[0023] generating a rainfall forecast of the target observation area in the preset time period according to the rainfall rate field prediction result.

[0024] In one of the embodiments, the ground sensor network comprises a ground rain gauge network, and the method further comprises:

[0025] acquire rainfall live feedback data sent by the ground rainfall station network;

[0026] determine at least two pre-judgment scores of the rainfall rate field prediction result according to differences between the rainfall live feedback data and the rainfall prediction; the pre-judgment scores are positively correlated with the accuracy of the rainfall rate field prediction result;

[0027] in a case where a number of target scores in the at least two pre-judgment scores is greater than a preset number threshold, trigger incremental training of the pre-trained rainfall rate field prediction model; the target scores are pre-judgment scores less than a preset score threshold in the at least two pre-judgment scores.

[0028] In a second aspect, the present application further provides a multi-source data fusion rainfall prediction device based on a Beidou B2B signal. The device comprises:

[0029] an initialization module configured to perform time synchronization operation on each observation device of a multi-source observation system in response to a rainfall prediction request for a target observation area; the target observation area is an observation area after grid processing; the multi-source observation system comprises a weather radar, a meteorological satellite, a ground sensing network, and a Beidou signal receiving device arranged in the target observation area;

[0030] a calculation module configured to determine pseudo-range observation information and carrier phase observation information according to satellite signals received by the Beidou signal receiving device, calculate zenith total delay generated by the satellite signals passing through the atmosphere according to the pseudo-range observation information and the carrier phase observation information, and convert the zenith total delay into atmospheric precipitable water content;

[0031] a multi-modal fusion module configured to input the atmospheric precipitable water content, radar rain rate of the weather radar, satellite observation of the meteorological satellite, and ground observation of the ground sensing network into a preset humidity analysis model as observation vectors to obtain a three-dimensional humidity analysis field;

[0032] a prediction module configured to acquire multi-source historical observation data, input the three-dimensional humidity analysis field and the multi-source historical observation data into a pre-trained rainfall rate field prediction model to obtain a rainfall rate field prediction result in a preset time period.

[0033] In a third aspect, the present application further provides a computer device. The computer device comprises a memory and a processor, the memory stores a computer program, and the computer program is executed by the processor to implement the steps of the above method.

[0034] In a fourth aspect, the present application further provides a computer readable storage medium. The computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the above method.

[0035] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the steps of the above-described method.

[0036] The aforementioned method, apparatus, computer equipment, storage medium, and computer program product for multi-source data fusion rainfall prediction based on BeiDou B2B signals are applied to a multi-source observation system. This system includes weather radar, meteorological satellites, a ground-based sensor network, and BeiDou signal receiving equipment located in the target observation area. In response to rainfall prediction requests for the target observation area, the various observation devices in the multi-source observation system are synchronized in time. The target observation area is a gridded observation area. Based on the satellite signals received by the BeiDou signal receiving equipment, pseudorange observation information and carrier phase observation information are determined. Based on these pseudorange and carrier phase information, the total zenith delay caused by the satellite signal passing through the atmosphere is calculated and converted into atmospheric precipitable water. The atmospheric precipitable water, radar rainfall rate from the weather radar, satellite observations from the meteorological satellite, and ground observations from the ground-based sensor network are then used as observation vectors input into a preset humidity analysis model to obtain a three-dimensional humidity analysis field. Finally, by acquiring multi-source historical observation data, the three-dimensional humidity analysis... Historical observation data from multiple sources are input into a pre-trained rainfall rate field prediction model to obtain rainfall rate field prediction results within a preset time period. Thus, B2B satellite signals collected by BeiDou signal receiving equipment located in the target observation area are effectively used to invert the atmospheric precipitable water content of the target observation area. Furthermore, by deeply fusing observation data from weather radar, meteorological satellites, and ground sensor networks, a three-dimensional dynamic-thermodynamic consistent humidity analysis field is formed, overcoming data fragmentation barriers and significantly improving the accuracy of the initial field. The pre-trained rainfall rate field prediction model provides a high-precision characterization of the formation and dissipation processes of rainstorm systems, improving the accuracy of precipitation forecasts for the target observation area. Through high-precision water vapor inversion using BeiDou B2B signals, multi-source observation collaborative assimilation, and spatiotemporal deep learning models, the bottlenecks of data fragmentation, lack of physical constraints, and insufficient model generalization ability in traditional rainfall forecasting techniques are systematically solved. This enables high spatiotemporal resolution early warning for the target observation area, significantly improving the accuracy and timeliness of rainstorm forecasts and providing reliable technical support for urban flood control and emergency decision-making. Attached Figure Description

[0037] Figure 1 This is an application environment diagram of a multi-source data fusion rainfall prediction method based on BeiDou B2B signals in one embodiment.

[0038] Figure 2 This is a flowchart illustrating a multi-source data fusion rainfall prediction method based on BeiDou B2B signals in one embodiment.

[0039] Figure 3 This is a schematic diagram of a multi-source data fusion rainfall prediction system based on BeiDou B2B signals in one embodiment.

[0040] Figure 4 This is a flowchart illustrating a multi-source data fusion rainfall prediction method based on BeiDou B2B signals in another embodiment.

[0041] Figure 5 This is a structural block diagram of a multi-source data fusion rainfall prediction device based on BeiDou B2B signals in one embodiment.

[0042] Figure 6 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0044] This application provides a multi-source data fusion rainfall prediction method based on BeiDou B2B signals, which can be applied to, for example... Figure 1 In the application environment shown, computer device 102 communicates with multi-source observation system 104 via a network. The multi-source observation system includes weather radar, meteorological satellites, a ground-based sensor network, and BeiDou signal receiving equipment located in the target observation area. The ground-based sensor network may include a ground-based meteorological sensor network and a ground-based rainfall sensor network. The BeiDou signal receiving equipment can refer to a BeiDou B2B receiver. In practical applications, in response to a rainfall prediction request for the target observation area, computer device 102 performs time synchronization operations on each observation device of the multi-source observation system 104. Based on the satellite signals received by the BeiDou signal receiving equipment, computer device 102 determines pseudorange observation information and carrier phase observation information. Based on the pseudorange observation information and carrier phase observation information, it calculates the total zenith delay caused by the satellite signal passing through the atmosphere and converts the total zenith delay into atmospheric precipitable water. Computer device 102 inputs atmospheric precipitable water, radar rainfall rate from weather radar, satellite observations from meteorological satellites, and ground observations from the ground-based sensor network as observation vectors into a preset humidity analysis model to obtain a three-dimensional humidity analysis field. Computer device 102 acquires multi-source historical observation data, inputs the three-dimensional humidity analysis field and the multi-source historical observation data into a pre-trained rainfall rate field prediction model, and obtains the rainfall rate field prediction results within a preset time period. Computer device 102 can be, but is not limited to, various personal computers, laptops, and servers. The server can be a standalone server or a server cluster composed of multiple servers.

[0045] In one embodiment, such as Figure 2 As shown, a multi-source data fusion rainfall prediction method based on BeiDou B2B signals is provided, and this method is applied to... Figure 1 Taking a computer device as an example, the explanation includes the following steps:

[0046] Step S202: In response to the rainfall prediction request for the target observation area, time synchronization operation is performed on each observation device of the multi-source observation system.

[0047] The target observation area is the gridded observation area. The observation area can refer to the region where rainfall needs to be predicted. In practical applications, the observation area can be divided into multiple grids using Kriging interpolation to smooth the spatial data, thus achieving gridding of the target observation area. Kriging interpolation... It can be represented as:

[0048]

[0049] In the formula, For Kriging weights, Given the location, The value is known.

[0050] In specific implementation, such as Figure 3 As shown, the computer equipment can respond to a rainfall forecast request for the target observation area by activating and initializing each observation device in the multi-source observation system; among other things, it can perform time synchronization operations on each observation device. Specifically, it can use the BeiDou satellite time as a standard for time reference correction based on the clock synchronization method of each observation device to achieve clock synchronization of each observation device.

[0051] Step S204: Based on the satellite signals received by the BeiDou signal receiving equipment, determine the pseudorange observation information and carrier phase observation information. Based on the pseudorange observation information and carrier phase observation information, calculate the total zenith delay caused by the satellite signal passing through the atmosphere, and convert the total zenith delay into atmospheric precipitable water.

[0052] In specific implementation, such as Figure 3 As shown, computer equipment acquires satellite signals received by the BeiDou signal receiving device. Based on these signals, pseudorange observation information and carrier phase observation information can be determined. Specifically, the distance to the satellite is estimated by measuring the signal propagation time. Precise positioning is achieved by measuring the received carrier phase difference. In practical applications, pseudorange observation information includes pseudorange observation values, which can be expressed as:

[0053]

[0054] in, The pseudorange observation value of the BeiDou signal receiving device r at the frequency f of satellite s. Let c be the actual geometric distance from the satellite to the receiver, and c be the speed of light. For receiver clock bias For satellite clock bias, For tropospheric delay, For ionospheric delay, and These are the receiver and satellite hardware delays, respectively. This is pseudorange observation noise.

[0055] In addition, carrier phase observation needs to consider the integer ambiguity problem. The carrier phase observation information can be represented as:

[0056]

[0057] In the formula, For wavelength, The carrier phase observation value of the BeiDou signal receiving device r for the frequency f of satellite s. For integer ambiguity, This is phase observation noise.

[0058] Then, the computer equipment can calculate the total zenith delay (ZTD) caused by the satellite signal passing through the atmosphere based on pseudorange and carrier phase observation information. Specifically, the computer equipment can use precise point positioning (PPP) to perform positioning and estimate the ZTD using precise ephemeris and precise clock bias data, obtain the ZTD caused by the satellite signal passing through the atmosphere, and convert the ZTD into atmospheric precipitable water.

[0059] Optionally, satellites with an elevation angle below 10 degrees can be removed, and cycle slips can be detected, marked, or repaired using the MW (Melbourne-Wübbena) combination (a linear combination of dual-frequency pseudorange and carrier phase).

[0060] Optionally, the conversion of total zenith delay into atmospheric precipitable water includes: using a preset delay decomposition model to decompose total zenith delay into zenith static delay and zenith wet delay; and converting zenith wet delay into atmospheric precipitable water.

[0061] The preset delay decomposition model can refer to the tropospheric delay correction model, such as the Saastamoinen model.

[0062] like Figure 3As shown, in the process of converting the total zenith delay (ZTD) into atmospheric precipitable water, the computer equipment can use tropospheric delay correction models such as the Saastamoinen model to decompose the total zenith delay (ZTD) into the zenith static delay (ZHD) and the zenith wet delay (ZWD), and convert the zenith wet delay (ZWD) into atmospheric precipitable water (PWV), while performing quality control.

[0063] Step S206: Input the atmospheric precipitable water, radar rainfall rate of weather radar, satellite observation of meteorological satellite, and ground observation of ground sensor network as observation vectors into the preset humidity analysis model to obtain a three-dimensional humidity analysis field.

[0064] In specific implementation, such as Figure 3 As shown, the computer equipment can also obtain radar rainfall rate through weather radar, satellite observations through the data acquisition interface of meteorological satellites, and ground observations through ground sensor networks. Then, atmospheric precipitable water, radar rainfall rate, satellite observations, and ground observations are constructed as observation vectors and input into a preset humidity analysis model. Using the 3D-Var method, the difference between the background field and the observation field is minimized by the 3D-Var objective function to generate a physically consistent three-dimensional humidity analysis field.

[0065] In practical applications, radar rainfall rate can be the rainfall rate field retrieved by radar; satellite observation can refer to satellite infrared brightness temperature; and ground observation can refer to ground temperature and humidity observations. The observation vector can be represented as:

[0066]

[0067] In the formula, For the observation vector, This refers to atmospheric precipitable water. For the rainfall rate field retrieved by radar, For satellite infrared brightness temperature, For ground humidity observation, For ground humidity observation.

[0068] Furthermore, the 3D-Var objective function can be expressed as:

[0069]

[0070] In the formula, Let B be the background field, B be the background error covariance matrix, R be the observation error covariance matrix, and H be the observation operator.

[0071] Step S208: Obtain multi-source historical observation data, input the three-dimensional humidity analysis field and multi-source historical observation data into the pre-trained rainfall rate field prediction model, and obtain the rainfall rate field prediction results within the preset time period.

[0072] Among them, multi-source historical observation data includes at least one of radar historical observation data, satellite historical observation data, or ground historical observation data.

[0073] Among them, the pre-trained rainfall rate field prediction model can refer to the convolutional Earthformer spatiotemporal deep learning model.

[0074] The preset time period can refer to the next 0 to 6 hours.

[0075] In specific implementation, such as Figure 3 As shown, computer equipment can acquire multi-source historical observation data, input the three-dimensional humidity analysis field and multi-source historical observation data into a pre-trained rainfall rate field prediction model, and obtain the rainfall rate field prediction results within a preset time period, such as the high-resolution rainfall rate field for the next 0 to 6 hours.

[0076] The aforementioned multi-source data fusion rainfall prediction method based on BeiDou B2B signals is applied to a multi-source observation system, which includes weather radar, meteorological satellites, a ground sensor network, and BeiDou signal receiving equipment located in the target observation area. In response to rainfall prediction requests for the target observation area, the various observation devices in the multi-source observation system are synchronized in time. The target observation area is a gridded observation area. Based on the satellite signals received by the BeiDou signal receiving equipment, pseudorange observation information and carrier phase observation information are determined. Based on the pseudorange and carrier phase observation information, the total zenith delay caused by the satellite signal passing through the atmosphere is calculated and converted into atmospheric precipitable water. Then, the radar rainfall rate from the weather radar, satellite observations from the meteorological satellites, and ground observations from the ground sensor network are used as observation vectors input to a preset humidity analysis model to obtain a three-dimensional humidity analysis field. Finally, by acquiring multi-source historical observation data, the three-dimensional humidity analysis field and the multi-source historical observation data are input to a pre-defined model. The trained rainfall rate field prediction model yields rainfall rate field prediction results within a preset time period. Thus, the atmospheric precipitable water content in the target observation area is effectively retrieved from B2B satellite signals collected by BeiDou signal receiving equipment located in the target observation area. Furthermore, by deeply fusing observation data from weather radar, meteorological satellites, and ground sensor networks, a three-dimensional dynamic-thermodynamic consistent humidity analysis field is created, overcoming data fragmentation barriers and significantly improving the accuracy of the initial field. The pre-trained rainfall rate field prediction model also provides a high-precision characterization of the formation and dissipation processes of rainstorm systems, enhancing the accuracy of precipitation forecasts for the target observation area. Through high-precision water vapor inversion using BeiDou B2B signals, multi-source observation collaborative assimilation, and spatiotemporal deep learning models, the bottlenecks of data fragmentation, lack of physical constraints, and insufficient model generalization ability in traditional rainfall forecasting techniques are systematically addressed. This enables high spatiotemporal resolution early warning for the target observation area, significantly improving the accuracy and timeliness of rainstorm forecasts and providing reliable technical support for urban flood control and emergency decision-making.

[0077] In another embodiment, the method further includes: acquiring the reflected echo signal received by the weather radar; determining the radar reflectivity based on the reflected echo signal, and converting the radar reflectivity into precipitation intensity according to a preset mapping relationship; and determining the radar rainfall rate of the weather radar based on the precipitation intensity.

[0078] Among them, the preset mapping relationship includes the power law relationship between radar reflectivity and rainfall rate (ZR power law relationship).

[0079] In practice, computer equipment can deduce the rainfall rate field from the reflected echo signals collected by weather radar. Specifically, the computer equipment can acquire the reflected echo signals received by the weather radar and perform signal processing on the reflected echo signals to obtain the radar reflectivity.

[0080] Then, asFigure 3 As shown, the computer equipment can convert radar reflectivity into precipitation intensity according to a preset mapping relationship; based on the precipitation intensity, the radar rainfall rate of the weather radar is determined. The preset mapping relationship can refer to a conversion model between radar reflectivity and precipitation intensity (rainfall rate, R), which can be expressed as:

[0081]

[0082] Where Z is radar reflectivity, R is rainfall rate, and a and b are empirical constants, where a can take the value 200 and b can take the value 1.6.

[0083] The technical solution of this embodiment obtains the reflected echo signal received by the weather radar; determines the radar reflectivity based on the reflected echo signal, and converts the radar reflectivity into precipitation intensity according to a preset mapping relationship; and determines the radar rainfall rate of the weather radar based on the precipitation intensity. This can effectively determine the radar rainfall rate of the target observation area based on the reflected echo signal received by the weather radar.

[0084] In another embodiment, under heavy precipitation conditions, the method further includes: acquiring the radar detection range of the weather radar, and determining the attenuation degree description information of the radar signal of the weather radar based on the radar detection range and a preset path attenuation coefficient; optimizing the radar rainfall rate of the weather radar based on the attenuation degree description information to obtain the optimized radar rainfall rate.

[0085] In practical implementation, under conditions of heavy precipitation, when determining radar reflectivity based on reflected echo signals, the computer equipment needs to consider the attenuation effect on the radar signal and use a path integral attenuation model (PIA) for correction. Specifically, the computer equipment can obtain the radar detection range of the weather radar and use the path integral attenuation model (PIA) to determine the attenuation level of the radar signal based on the radar detection range and a preset path attenuation coefficient.

[0086] Then, the computer equipment can optimize the radar rainfall rate of the weather radar based on the attenuation description information to obtain the optimized radar rainfall rate. In practical applications, the path integral attenuation model can be expressed as:

[0087]

[0088] In the formula, denoted as path attenuation coefficient, and r as radar detection range.

[0089] The technical solution of this embodiment obtains the radar detection range of the weather radar and determines the attenuation degree description information of the radar signal of the weather radar according to the radar detection range and the preset path attenuation coefficient; according to the attenuation degree description information, the reflected echo signal is adjusted, so as to effectively consider the attenuation effect of the radar signal under heavy precipitation conditions and more accurately deduce the accurate radar rainfall field based on the reflected echo signal received by the weather radar.

[0090] In another embodiment, the method further includes: obtaining a multiplicative correction coefficient corresponding to a preset forecast lead time, determining the product between the rainfall rate field prediction result and the multiplicative correction coefficient, and obtaining an adjusted prediction result; the multiplicative correction coefficient is obtained by fitting the slope of a linear regression between historical forecasts and actual conditions; obtaining an additive correction term corresponding to a preset forecast lead time, determining the sum between the adjusted prediction result and the additive correction term, and obtaining a corrected rainfall rate field prediction result; and generating a rainfall forecast for the target observation area within a preset time period based on the rainfall rate field prediction result.

[0091] In practical implementation, computer equipment can also perform systematic error correction, spatial consistency optimization, and probabilistic uncertainty estimation on the predicted rainfall rate field to generate precipitation forecast products; specifically, such as Figure 3 As shown, the computer equipment can obtain the multiplication correction coefficient corresponding to the preset forecast lead time, determine the product between the rainfall rate field prediction result and the multiplication correction coefficient, and obtain the adjusted prediction result; then, the computer equipment obtains the addition correction term corresponding to the preset forecast lead time, determines the sum between the adjusted prediction result and the addition correction term, and obtains the corrected rainfall rate field prediction result.

[0092] The corrected rainfall rate field prediction result can be expressed as:

[0093]

[0094] In the formula, To provide a forecast timeframe, For the corrected Rainfall rate field at time (corrected rainfall rate field prediction result). for The time-sensitivity multiplication correction coefficient is obtained by fitting the slope of a linear regression between historical forecasts and actual conditions. For the future Rainfall rate at any time for Additive correction items for timeliness.

[0095] Then, the computer equipment generates a rainfall forecast for the target observation area within a preset time period based on the rainfall rate field prediction results.

[0096] The technical solution of this embodiment obtains the multiplication correction coefficient corresponding to the preset forecast lead time, determines the product between the rainfall rate field prediction result and the multiplication correction coefficient, and obtains the adjusted prediction result. It also obtains the addition correction term corresponding to the preset forecast lead time, determines the sum between the adjusted prediction result and the addition correction term, and obtains the corrected rainfall rate field prediction result. Based on the rainfall rate field prediction result, a rainfall forecast for the target observation area within a preset time period is generated. This can effectively perform systematic error correction, spatial consistency optimization, and probabilistic uncertainty estimation on the predicted rainfall rate field.

[0097] In another embodiment, the method further includes: acquiring real-time rainfall feedback data sent by a network of ground rain gauges; determining at least two prediction scores for the rainfall rate field prediction result based on the difference between the real-time rainfall feedback data and the rainfall forecast; the prediction scores are positively correlated with the accuracy of the rainfall rate field prediction result; if the number of target scores in the at least two prediction scores is greater than a preset threshold, triggering incremental training of the pre-trained rainfall rate field prediction model; the target score is the prediction score that is less than a preset score threshold among the at least two prediction scores.

[0098] The ground-based sensor network includes a network of ground-based rain gauges.

[0099] In specific implementation, such as Figure 3 As shown, the computer equipment can acquire real-time rainfall feedback data sent by the ground rain gauge network and determine at least two prediction scores for the rainfall rate field prediction result based on the differences between the real-time rainfall feedback data and the rainfall forecast. The prediction scores are positively correlated with the accuracy of the rainfall rate field prediction result. Specifically, when the prediction score is the Critical Success Index (CSI), the computer equipment can determine the number of hits, false alarms, and missed alarms in the rainfall rate field prediction result based on the differences between the real-time rainfall feedback data and the rainfall forecast, and determine the Critical Success Index (CSI) (i.e., the prediction score) based on the number of hits, false alarms, and missed alarms. In practical applications, the Critical Success Index can be expressed as:

[0100]

[0101] Then, the computer device can determine the predicted score with a preset score threshold as an abnormal score (e.g., a critical success index less than 0.4) among at least two predicted scores; if the number of consecutive occurrences of the critical success index less than 0.4 exceeds a preset number threshold (e.g., 3 times), incremental training of the pre-trained rainfall rate field prediction model (Earthformer model) is triggered.

[0102] The technical solution of this embodiment can determine the risk level of rainstorm disasters based on precipitation forecast products, generate early warning signals and visualization products, and optimize the rainfall rate field prediction model in real time through an online feedback mechanism, forming a positive cycle that continuously improves the accuracy of precipitation prediction.

[0103] In another embodiment, such as Figure 4 As shown, a multi-source data fusion rainfall prediction method based on BeiDou B2B signals is provided, and this method is applied to... Figure 1 Taking a computer device as an example, the explanation includes the following steps:

[0104] Step S402: Power-on initialization of the multi-source sensor system and grid division of the spatial monitoring area.

[0105] Step S404: Collect multi-source positioning data in parallel and evaluate the quality and reliability of BeiDou satellite signals in real time.

[0106] Step S406: Construct a core model for deep integrated navigation solution and data fusion based on the extended Kalman filter framework.

[0107] Step S408: During periods when the BeiDou positioning signal quality is good, perform online dynamic calibration and compensation of the odometer measurement parameters.

[0108] Step S410: Based on the fused high-precision real-time pose information, predict the motion trajectory and analyze the behavior of the monitored target.

[0109] Step S412: Based on the motion trajectory prediction results, adaptively trigger the pre-activation and intelligent scheduling instructions of the video surveillance equipment.

[0110] Step S414: Conduct a systematic quantitative evaluation of the overall performance of the video linkage control and optimize the closed-loop feedback parameters.

[0111] Step S416: Maintain basic positioning and linkage functions in extreme environments where satellite navigation signals are continuously lost or severely interfered with.

[0112] It should be noted that the specific limitations of the above steps can be found in the above description of the specific limitations of a multi-source data fusion rainfall prediction method based on BeiDou B2B signals.

[0113] The aforementioned method, apparatus, computer equipment, storage medium, and computer program product for multi-source data fusion rainfall prediction based on BeiDou B2B signals are applied to a multi-source observation system. This system includes weather radar, meteorological satellites, a ground-based sensor network, and BeiDou signal receiving equipment located in the target observation area. In response to rainfall prediction requests for the target observation area, the various observation devices in the multi-source observation system are synchronized in time. The target observation area is a gridded observation area. Based on the satellite signals received by the BeiDou signal receiving equipment, pseudorange observation information and carrier phase observation information are determined. Based on these pseudorange and carrier phase information, the total zenith delay caused by the satellite signal passing through the atmosphere is calculated and converted into atmospheric precipitable water. The atmospheric precipitable water, radar rainfall rate from the weather radar, satellite observations from the meteorological satellite, and ground observations from the ground-based sensor network are then used as observation vectors input into a preset humidity analysis model to obtain a three-dimensional humidity analysis field. Finally, by acquiring multi-source historical observation data, the three-dimensional humidity analysis... Historical observation data from multiple sources are input into a pre-trained rainfall rate field prediction model to obtain rainfall rate field prediction results within a preset time period. Thus, B2B satellite signals collected by BeiDou signal receiving equipment located in the target observation area are effectively used to invert the atmospheric precipitable water content of the target observation area. Furthermore, by deeply fusing observation data from weather radar, meteorological satellites, and ground sensor networks, a three-dimensional dynamic-thermodynamic consistent humidity analysis field is formed, overcoming data fragmentation barriers and significantly improving the accuracy of the initial field. The pre-trained rainfall rate field prediction model provides a high-precision characterization of the formation and dissipation processes of rainstorm systems, improving the accuracy of precipitation forecasts for the target observation area. Through high-precision water vapor inversion using BeiDou B2B signals, multi-source observation collaborative assimilation, and spatiotemporal deep learning models, the bottlenecks of data fragmentation, lack of physical constraints, and insufficient model generalization ability in traditional rainfall forecasting techniques are systematically solved. This enables high spatiotemporal resolution early warning for the target observation area, significantly improving the accuracy and timeliness of rainstorm forecasts and providing reliable technical support for urban flood control and emergency decision-making.

[0114] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0115] Based on the same inventive concept, this application also provides a BeiDou B2B signal-based multi-source data fusion rainfall prediction device for implementing the aforementioned multi-source data fusion rainfall prediction method based on BeiDou B2B signals. The solution provided by this device is similar to the implementation described in the above method. Therefore, the specific limitations of one or more embodiments of the BeiDou B2B signal-based multi-source data fusion rainfall prediction device provided below can be found in the limitations of the BeiDou B2B signal-based multi-source data fusion rainfall prediction method described above, and will not be repeated here.

[0116] In one embodiment, such as Figure 5 As shown, a multi-source data fusion rainfall prediction device based on BeiDou B2B signals is provided, comprising:

[0117] Initialization module 510 is used to perform time synchronization operation on each observation device of the multi-source observation system in response to a rainfall forecast request for the target observation area; the target observation area is an observation area that has been gridded; the multi-source observation system includes weather radar, meteorological satellite, ground sensor network, and Beidou signal receiving equipment set in the target observation area;

[0118] The calculation module 520 is used to determine pseudorange observation information and carrier phase observation information based on the satellite signals received by the Beidou signal receiving device, calculate the total zenith delay caused by the satellite signal passing through the atmosphere based on the pseudorange observation information and the carrier phase observation information, and convert the total zenith delay into atmospheric precipitable water.

[0119] The multimodal fusion module 530 is used to input the atmospheric precipitable water, the radar rainfall rate of the weather radar, the satellite observation of the meteorological satellite, and the ground observation of the ground sensor network as observation vectors into a preset humidity analysis model to obtain a three-dimensional humidity analysis field.

[0120] The prediction module 540 is used to acquire multi-source historical observation data, input the three-dimensional humidity analysis field and the multi-source historical observation data into a pre-trained rainfall rate field prediction model, and obtain the rainfall rate field prediction result within a preset time period.

[0121] In one embodiment, the device is used to acquire reflected echo signals received by the weather radar; determine radar reflectivity based on the reflected echo signals, and convert the radar reflectivity into precipitation intensity according to a preset mapping relationship; the preset mapping relationship includes a power law relationship between radar reflectivity and rainfall rate; and determine the radar rainfall rate of the weather radar based on the precipitation intensity.

[0122] In one embodiment, under heavy precipitation conditions, the device is used to acquire the radar detection range of the weather radar, and determine the attenuation degree description information of the radar signal of the weather radar based on the radar detection range and a preset path attenuation coefficient; based on the attenuation degree description information, the radar rainfall rate of the weather radar is optimized to obtain the optimized radar rainfall rate.

[0123] In one embodiment, the calculation module 520 is used to decompose the total zenith delay into zenith static delay and zenith wet delay using a preset delay decomposition model; and to convert the zenith wet delay into atmospheric precipitable water.

[0124] In one embodiment, the device is used to obtain a multiplicative correction coefficient corresponding to a preset forecast lead time, determine the product between the rainfall rate field prediction result and the multiplicative correction coefficient, and obtain an adjusted prediction result; the multiplicative correction coefficient is obtained by fitting the slope of a linear regression between historical forecasts and actual conditions; obtain an additive correction term corresponding to a preset forecast lead time, determine the sum between the adjusted prediction result and the additive correction term, and obtain the corrected rainfall rate field prediction result; and generate a rainfall forecast for the target observation area during the preset time period based on the rainfall rate field prediction result.

[0125] In one embodiment, the ground sensor network includes a ground rain gauge network, and the device is used to acquire real-time rainfall feedback data sent by the ground rain gauge network; determine at least two prediction scores for the rainfall rate field prediction result based on the difference between the real-time rainfall feedback data and the rainfall forecast; the prediction scores are positively correlated with the accuracy of the rainfall rate field prediction result; if the number of target scores among the at least two prediction scores is greater than a preset threshold, incremental training of the pre-trained rainfall rate field prediction model is triggered; the target score is the prediction score among the at least two prediction scores that is less than a preset score threshold.

[0126] The modules in the aforementioned multi-source data fusion rainfall prediction device based on BeiDou B2B signals can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0127] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 6As shown, the computer device includes a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a multi-source data fusion rainfall prediction method based on BeiDou B2B signals. The display screen can be an LCD screen or an e-ink display screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0128] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0129] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, which, when executed by the processor, causes the processor to perform the steps of the aforementioned multi-source data fusion rainfall prediction method based on BeiDou B2B signals. The steps of this multi-source data fusion rainfall prediction method based on BeiDou B2B signals can be the steps in one of the multi-source data fusion rainfall prediction methods based on BeiDou B2B signals described in the various embodiments above.

[0130] In one embodiment, a computer-readable storage medium is provided, storing a computer program that, when executed by a processor, causes the processor to perform the steps of the above-described method for multi-source data fusion rainfall prediction based on BeiDou B2B signals. The steps of this method for multi-source data fusion rainfall prediction based on BeiDou B2B signals can be the steps in one of the above-described embodiments of the method.

[0131] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, causes the processor to perform the steps of the above-described method for multi-source data fusion rainfall prediction based on BeiDou B2B signals. The steps of this method for multi-source data fusion rainfall prediction based on BeiDou B2B signals can be the steps in one of the above-described embodiments of the method.

[0132] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0133] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic resistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0134] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0135] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A multi-source data fusion rainfall prediction method based on BeiDou B2B signals, characterized in that, The method includes: In response to a request for rainfall forecasting of a target observation area, time synchronization is performed on each observation device of the multi-source observation system; the target observation area is an observation area that has been gridded; the multi-source observation system includes weather radar, meteorological satellites, ground sensor networks, and BeiDou signal receiving equipment installed in the target observation area; Based on the satellite signals received by the BeiDou signal receiving device, pseudorange observation information and carrier phase observation information are determined. Based on the pseudorange observation information and the carrier phase observation information, the total zenith delay caused by the satellite signal passing through the atmosphere is calculated, and the total zenith delay is converted into atmospheric precipitable water. The atmospheric precipitable water, the radar rainfall rate of the weather radar, the satellite observations of the meteorological satellite, and the ground observations of the ground sensor network are used as observation vectors input into a preset humidity analysis model to obtain a three-dimensional humidity analysis field. Acquire multi-source historical observation data, input the three-dimensional humidity analysis field and the multi-source historical observation data into a pre-trained rainfall rate field prediction model, and obtain the rainfall rate field prediction results within a preset time period.

2. The method according to claim 1, characterized in that, The method further includes: Acquire the reflected echo signal received by the weather radar; Based on the reflected echo signal, the radar reflectivity is determined, and the radar reflectivity is converted into precipitation intensity according to a preset mapping relationship; the preset mapping relationship includes the power law relationship between radar reflectivity and rainfall rate. The radar rainfall rate of the weather radar is determined based on the precipitation intensity.

3. The method according to claim 2, characterized in that, Under conditions of heavy precipitation, the method further includes: The radar detection range of the weather radar is obtained, and the attenuation degree description information of the radar signal of the weather radar is determined based on the radar detection range and the preset path attenuation coefficient. Based on the attenuation description information, the radar rainfall rate of the weather radar is optimized to obtain the optimized radar rainfall rate.

4. The method according to claim 1, characterized in that, The process of converting the total zenith delay into atmospheric precipitable water includes: Using a preset delay decomposition model, the total zenith delay is decomposed into zenith static delay and zenith wet delay. The zenith wet delay is converted into atmospheric precipitable water.

5. The method according to claim 1, characterized in that, The method further includes: Obtain the multiplication correction coefficient corresponding to the preset forecast lead time, determine the product between the rainfall rate field prediction result and the multiplication correction coefficient, and obtain the adjusted prediction result; the multiplication correction coefficient is obtained by fitting the slope of the linear regression between historical forecasts and actual conditions. Obtain the additive correction term corresponding to the preset forecast lead time, determine the sum between the adjusted prediction result and the additive correction term, and obtain the corrected rainfall rate field prediction result; Based on the rainfall rate field prediction results, a rainfall forecast for the target observation area during the preset time period is generated.

6. The method according to claim 5, characterized in that, The ground-based sensor network includes a ground-based rain gauge network, and the method further includes: Obtain the real-time rainfall feedback data sent by the ground rain gauge network; Based on the difference between the actual rainfall feedback data and the rainfall forecast, at least two prediction scores are determined for the rainfall rate field prediction result; the prediction scores are positively correlated with the accuracy of the rainfall rate field prediction result. If the number of target scores in the at least two predicted scores is greater than a preset threshold, incremental training of the pre-trained rainfall rate field prediction model is triggered; the target score is the predicted score that is less than the preset score threshold among the at least two predicted scores.

7. A multi-source data fusion rainfall prediction device based on BeiDou B2B signals, characterized in that, The device includes: An initialization module is used to perform time synchronization operations on each observation device of the multi-source observation system in response to a rainfall forecast request for the target observation area; the target observation area is an observation area that has been gridded; the multi-source observation system includes weather radar, meteorological satellites, ground sensor networks, and BeiDou signal receiving equipment set in the target observation area; The calculation module is used to determine pseudorange observation information and carrier phase observation information based on the satellite signals received by the Beidou signal receiving device, calculate the total zenith delay caused by the satellite signal passing through the atmosphere based on the pseudorange observation information and the carrier phase observation information, and convert the total zenith delay into atmospheric precipitable water. The multimodal fusion module is used to input the atmospheric precipitable water, the radar rainfall rate of the weather radar, the satellite observations of the meteorological satellite, and the ground observations of the ground sensor network as observation vectors into a preset humidity analysis model to obtain a three-dimensional humidity analysis field. The prediction module is used to acquire multi-source historical observation data, input the three-dimensional humidity analysis field and the multi-source historical observation data into the pre-trained rainfall rate field prediction model, and obtain the rainfall rate field prediction results within a preset time period.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images