Flood monitoring method based on beidou geo satellite reflected signal

Abstract

The application discloses a kind of based on Beidou GEO satellite reflection signal's flood monitoring method, comprising the following steps: S1, read the original observation value of Beidou GEO satellite on monitoring day;S2, original signal-to-noise ratio is carried out data processing to obtain reflection signal signal-to-noise ratio;S3, reflection signal signal-to-noise ratio amplitude value is obtained by reflection signal signal-to-noise ratio;S4, the method of satellite and frequency is used to establish the reference monitoring model of Beidou GEO satellite based on first monitoring quantity and second monitoring quantity;S5, the first monitoring quantity and second monitoring quantity of Beidou GEO three-frequency signal are monitored to flood by reference monitoring model, and the result is marked according to monitoring threshold value.This method not only eliminates the influence of direct signal, while reflection signal signal-to-noise ratio amplitude value is less affected by environment, and Beidou GEO satellite belongs to stationary satellite, corresponding reflection signal signal-to-noise ratio time series feature is more obvious, so it can effectively improve the accuracy and real-time performance of flood monitoring.

Description

Technical Field

[0001] This invention relates to the field of navigation satellite remote sensing inversion technology, specifically to a flood monitoring method based on the signal-to-noise ratio and amplitude of reflected signals from BeiDou GEO (Geostationary Orbit) satellites. Background Technology

[0002] Floods are characterized by their rapid onset, wide impact, and immense scale. Short-term floods, especially in densely populated urban and rural areas, severely impact people's lives and property, making them one of the major natural disasters threatening human survival and development. Statistics show that in the past two decades, floods have affected over 1 billion people globally, causing economic losses of up to $3 trillion. In my country, due to the distinct characteristics of hydrological and climatic distribution, the rainy season in the south leads to frequent floods, while extreme weather in the north also easily triggers them. For example, in 2021, three periods of extreme heavy rainfall in northern China caused urban flooding, affecting over 3 million people and causing economic losses exceeding 200 billion yuan. Therefore, establishing scientific, effective, and high-precision flood monitoring methods to improve the real-time nature of flood monitoring and protect people's lives and property is an urgent need for flood monitoring and management, and one of the key issues to be addressed in flood early warning and prevention.

[0003] Currently, flood monitoring methods can be mainly divided into three categories. The first category is based on traditional rain gauges, which offer high measurement accuracy and low cost per station. However, this method has a small measurement range, resulting in low temporal and spatial resolution, and is difficult to deploy over a large area. The second category is based on remote sensing imaging satellites, which offer high measurement accuracy and a wide monitoring range. However, remote sensing imaging satellites are greatly affected by weather conditions; severe weather can cause the monitoring method to fail or reduce accuracy, and the monitoring data needs to be transmitted back to the ground for processing, resulting in low real-time performance. The third category is based on navigation satellite systems. This method uses geodetic receivers to receive navigation satellite signals for inversion measurement, offering high measurement accuracy and a wide measurement range. However, existing navigation satellite-based flood monitoring directly models the data using the raw signal-to-noise ratio (SNR), which includes direct sunlight, leading to reduced modeling accuracy. Furthermore, the phase value monitoring method used is greatly affected by the environment and is more sensitive to soil type at the measurement site, resulting in low monitoring accuracy. In addition, existing monitoring methods are often based on IGSO (Inclined Geosynchronous Satellite Orbit) or MEO (Medium Earth Orbit) satellites. However, these two types of satellites are always in motion, resulting in a large signal-to-noise ratio error and reduced accuracy of flood monitoring. Summary of the Invention

[0004] This invention proposes a flood monitoring method based on BeiDou GEO satellite reflected signals. This method not only eliminates the influence of direct signals, but also minimizes the impact of environmental factors on the signal-to-noise ratio amplitude of reflected signals. Furthermore, since BeiDou GEO satellites are geostationary satellites, the time series characteristics of the reflected signal-to-noise ratio are more pronounced. Therefore, this method can effectively improve the accuracy and real-time performance of flood monitoring, as well as ensure the success rate and stability of flood monitoring, providing strong support for flood disaster early warning.

[0005] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows:

[0006] A flood monitoring method based on BeiDou GEO satellite reflected signals includes the following steps:

[0007] S1. Read the original observation values ​​of the BeiDou GEO satellite on the monitoring day, and calculate the elevation angle information and time information at the corresponding epoch. The original observation values ​​include the original signal-to-noise ratio, the original pseudorange, and the original carrier wave.

[0008] S2. The original signal-to-noise ratio is processed to obtain the reflected signal-to-noise ratio, which is then used as the first monitoring quantity.

[0009] S3. Obtain the amplitude value of the reflected signal-to-noise ratio through the reflected signal-to-noise ratio, and use it as the second monitoring quantity;

[0010] S4. A reference monitoring model for BeiDou GEO satellites based on the first and second monitoring quantities is established using a satellite-based and frequency-based method. The reference monitoring model stores the first and second monitoring quantities for the corresponding epochs of each year in the form of a database text on a local computer. It also stores the elevation angle and time information for the corresponding epochs, as well as the corresponding monitoring thresholds.

[0011] S5, together with the first and second monitoring quantities of the Beidou GEO tri-frequency signals, monitors floods through a reference monitoring model and marks the results according to the monitoring threshold.

[0012] Preferably, the method for processing the original signal-to-noise ratio in step S2 is as follows:

[0013] S2-1. Use a Gaussian fitting method to eliminate the influence of the signal-to-noise ratio of the direct signal;

[0014] S2-2. High-frequency random noise is filtered out using an empirical mode decomposition denoising algorithm to obtain a signal that retains only the signal-to-noise ratio of the reflected signal.

[0015] Preferably, the specific method of step S2-1 is as follows:

[0016] S2-1-1. Use the Gaussian fitting algorithm to fit the original signal-to-noise ratio observations to obtain the Gaussian fitting function.

[0017] The process of solving the Gaussian fitting function is as follows: Assume the original signal-to-noise ratio observation dataset is ( x i , S i )( i =1,2,3,...., N ), N Indicates signal length. x Represents the time series of a signal. S This represents the raw signal-to-noise ratio at that time. This dataset can be described using a Gaussian function, with the following formula:

[0018] (1)

[0019] In the formula, S i and x i It is known. exp This represents exponentiation, a common mathematical calculation. a , b and c The unknown parameters represent the peak height, peak position, and half-width of the Gaussian function curve, respectively.

[0020] a , b and c These three parameters can be solved by the following process: Taking the natural logarithm of both sides of equation (1), equation (1) can be transformed into:

[0021] (2)

[0022] make: , , , Then equation (2) can be transformed into a quadratic polynomial fitting function, expressed as follows:

[0023] (3)

[0024] Considering all data and measurement errors, the results are presented in matrix form as follows:

[0025] (4)

[0026] The above formula can be simplified to:

[0027] (5)

[0028] In the formula, Z Given a quantity, from Calculations show that X Given a quantity, from x i Therefore, without considering overall measurement error, E Under the influence of various factors, the fitting constants can be obtained according to the least squares principle. b 0, b 1, b 2 matrices B The generalized least squares solution is expressed as:

[0029] (6)

[0030] In the formula, X and Z All are known quantities. T The matrix transpose operation is represented by equation (6), and the fitting constant can be obtained by equation (6). b 0, b 1, b 2. Furthermore, the unknown parameters can be obtained using the following formula. a , b , c :

[0031] (7)

[0032] In equation (7) b 0, b 1, b 2 has been obtained through equation (6), e Since is a natural constant, the unknown parameter can be obtained through equation (7). a , b , c Therefore, we can obtain the Gaussian function fitted by the original signal-to-noise ratio data, i.e., equation (1).

[0033] S2-1-2. Using the Gaussian fitting function obtained in the previous step, calculate the fitting value for each epoch, expressed as: S G The calculation process can be expressed as:

[0034] (8)

[0035] In equation (8) a , b , c It has already been obtained through equation (7), and x i Since it represents a time series, the fitted value at the corresponding time point can be calculated. S G ;

[0036] S2-1-3, Utilizing the original signal-to-noise ratio S d+r+n Subtract the fitted value calculated in the previous step S G The result is a signal containing only the signal-to-noise ratio of the reflected signal and random noise, expressed by the formula:

[0037] S r+n = S d+r+n - S G (9)

[0038] The above process can eliminate the influence of the direct signal's signal-to-noise ratio from the original signal-to-noise ratio observations, resulting in a signal that contains only the reflected signal's signal-to-noise ratio and random noise. S r+n .

[0039] Preferably, the specific method of step S2-2 is as follows:

[0040] S2-2-1, Signal Decomposition

[0041] For a given signal S r+n (x The process involves finding the locations of all local maxima and minima on the signal, and then using an interpolation function to generate the given signal from all the maxima and minima. S r+n ( x The upper and lower envelopes of the expression are given by the expression, where the upper envelope is generated by maximum interpolation and is represented as follows: a 1. The lower envelope is generated by minimum interpolation, denoted as: b 1,

[0042] The average of the two obtained envelopes is expressed as:

[0043] (10)

[0044] In the formula a 1 and b 1. That is, the upper and lower envelopes are generated by the interpolation function, and then the given signal is used. S r+n ( x Subtracting the value obtained from equation (10) The first set of residuals was obtained. , is represented as:

[0045] (11)

[0046] The result calculated in judgment (11) Does it meet the two assumptions of the intrinsic mode function? If not, then... The above decomposition process is repeated as a new input signal until the decomposed signal is obtained. This continues until the two assumptions regarding the intrinsic mode functions are met.

[0047] Assuming decomposition to the th k After that It satisfies the two assumptions of the intrinsic mode function (IMF), the first k The decomposition is represented as follows:

[0048] (12)

[0049] Then at this time That is S r+n ( x The first intrinsic mode function (IMF) derived from the decomposition, denoted as IMF1, has the shortest period and the highest frequency among all subsequent IMFs derived from the decomposition.

[0050] Subtracting the first intrinsic mode function (IMF1) from the original signal yields a new data sequence with IMF1 removed. r 1. For this new sequencer 1. Continue repeating the process of obtaining IMF1 to obtain the second intrinsic mode function IMF2, and then use... r Subtracting the second intrinsic mode function (IMF2) from 1 yields a new data sequence. r 2; Repeat this calculation until the last one. r m The process continues until it can no longer be decomposed; the above process can be represented as:

[0051] (13)

[0052] at this time, r m That is the original signal S r+n ( x The residual signal remaining after empirical mode decomposition is expressed as: Res ,

[0053] S2-2-2, Signal Reconstruction

[0054] After the above empirical mode decomposition of the signal, each intrinsic mode function (IMF) represents an independent frequency component of the original signal, arranged in descending order of frequency. Therefore, all the separated IMFs and the residuals can be used to... Res The signal is reconstructed to obtain the denoised signal. The reconstruction formula is as follows:

[0055] (14)

[0056] After the above process, high-frequency random noise can be effectively filtered out, resulting in a signal that retains only the signal-to-noise ratio of the reflected signal. , represented as This signal is used as the primary monitoring quantity.

[0057] Preferably, the method for obtaining the amplitude value of the reflected signal-to-noise ratio in step S3 is as follows:

[0058] S3-1. The signal-to-noise ratio (SNR) of a linearly reflected signal in dB-Hz is converted to the SNR of a periodically reflected signal in volts / volts using a linear-to-periodic conversion model. The conversion formula for the linear-to-periodic conversion model is expressed as follows:

[0059] (15)

[0060] In the formula, This is a linear signal that contains only the signal-to-noise ratio of the reflected signal, measured in dB-Hz. Sr_v The signal-to-noise ratio of the periodic reflected signal, expressed in volts / volts, is obtained after the linear-periodic transformation model.

[0061] S3-2. Determining the effective reflectivity of a satellite using the Lomb-Scargle spectral analysis method. ;

[0062] S3-3. The Gaussian particle filter algorithm is used to calculate the signal-to-noise ratio amplitude of the reflected signal. The specific formula is as follows:

[0063]

[0064] In the formula, A and The amplitude and phase of the signal-to-noise ratio represent the unknowns in the formula. S r_v It is the signal-to-noise ratio of the periodically reflected signal. The satellite has high effective reflectivity. λ It is the wavelength of the signal, which is a known quantity. EL That is the corresponding elevation angle.

[0065] Preferably, in step S3-2, the effective reflectance of the satellite is high. Solution method:

[0066] S3-2-1. Determining the signal-to-noise ratio of periodic reflection signals using the Lomb-Scargle spectral analysis method. S r_v Frequency at maximum power The specific process is as follows:

[0067] The power spectrum of the signal is determined using the Lomb-Sacragle spectral analysis method, and the formula is as follows:

[0068] (17)

[0069] In the formula, P x ( f ) is a frequency of f The power of the periodic signal; S r_v ( x () represents the signal-to-noise ratio of the reflected signal. N For sequence length, τ As a time-shift invariant, it can be calculated using the following formula:

[0070]

[0071] The following can be obtained using equations (17-18). P x( f When the peak value is reached, the corresponding ;

[0072] S3-2-2, Utilizing and satellite effective reflection high The relationship between the effective reflectance of the satellite and the signal-to-noise ratio frequency of the reflected signal is used to determine the effective reflectance of the satellite. The relationship between the effective reflectance of the satellite and the effective reflectance is expressed as follows:

[0073]

[0074] In the formula, λ The wavelength of the signal is a known quantity; the effective reflectivity of the satellite can be calculated using the above process. .

[0075] Preferably, in step S4, the monitoring thresholds are that the fluctuation of the first monitoring quantity does not exceed 2dB-Hz and the fluctuation of the second monitoring quantity does not exceed 0.5.

[0076] Preferably, step S5 includes the following steps:

[0077] S5-1. Flood monitoring is conducted using the signal-to-noise ratio of three-frequency reflected signals from the BeiDou GEO satellite, and the monitoring results are marked accordingly. The marking levels include primary marking and secondary marking.

[0078] S5-2. Flood monitoring is carried out using the amplitude value of the signal-to-noise ratio of the three-frequency reflected signals based on the BeiDou GEO satellite, and the monitoring results are marked. The marking levels include primary marking and secondary marking.

[0079] S5-3. The flood monitoring results of the first and second monitoring quantities are processed together. If the first and second monitoring quantities have a first-level mark, it is considered that there is a flood risk and a first-level flood warning is required. If the first and second monitoring quantities do not have a first-level mark but have a second-level mark, a second-level flood warning is required. If neither has a mark, no warning is issued.

[0080] Preferably, the method for monitoring flooding in step S5-1 is as follows:

[0081] First, the raw signal-to-noise ratio at the current epoch is obtained. The first monitoring quantity at the current epoch is calculated through step S2, and the elevation angle information and time information at that epoch are calculated.

[0082] Secondly, the calculated elevation angle information and time information of the three frequencies of BeiDou GEO satellite, namely the B1 band, B2 band and B3 band, are input into the reference monitoring model to search for the first monitoring quantity in the corresponding reference value. The first monitoring quantity at the current epoch is compared with the first monitoring quantity in the reference value. If the fluctuation of the first monitoring quantity at the current epoch is greater than 2dB-Hz, flooding exists, and it is marked.

[0083] Finally, statistical analysis is performed based on the epoch marking results. If the marked epochs account for more than 80% of the total number of epochs, it is considered that there is a strong flood correlation and a first-level marking is made. If the marked epochs account for 60%-80% of the total number of epochs, it is considered that there is a relatively strong flood correlation and a second-level marking is made. If the marked epochs account for less than 60%, no marking is made. Finally, the monitoring results are saved and output.

[0084] Preferably, the method for monitoring flooding in step S5-2 is as follows:

[0085] First, the amplitude value of the reflected signal-to-noise ratio at the current epoch is calculated by step S3 based on the reflected signal-to-noise ratio calculated in step S5-1, which is the second detection quantity at the current epoch. Then, the elevation angle and time information at that epoch are calculated.

[0086] Secondly, the calculated elevation angle information and time information of the three frequencies of BeiDou GEO satellite, namely the B1 band, B2 band and B3 band, are input into the reference monitoring model to search for the second monitoring quantity in the corresponding reference value. The calculated second monitoring quantity at the current epoch is compared with the second monitoring quantity in the reference value. If the amplitude fluctuation of the reflected signal-to-noise ratio at the current epoch is greater than 0.5, then flooding exists, and it is marked.

[0087] Finally, statistical analysis is performed based on the epoch marking results. If the marked epochs account for more than 80% of the total number of epochs, it is considered that there is a strong flood correlation and a first-level marking is made. If the marked epochs account for 60%-80% of the total number of epochs, it is considered that there is a relatively strong flood correlation and a second-level marking is made. If the marked epochs account for less than 60%, no marking is made. Finally, the monitoring results are saved and output.

[0088] This invention has the following characteristics and beneficial effects:

[0089] By adopting the above technical solution, flood monitoring can be carried out using existing BeiDou navigation GEO satellites. Compared with existing methods based on traditional rain gauges, this solves the problem of small monitoring range. Compared with current remote sensing imaging satellite measurement methods, it solves the problems of being greatly affected by the environment and low real-time performance. Compared with existing navigation system satellite measurement methods, it solves the problems of low monitoring accuracy caused by orbital motion of IGSO and MEO satellites, as well as the problems of low monitoring accuracy and unstable performance caused by ignoring the influence of direct signal-to-noise ratio and the amplitude information of reflected signal-to-noise ratio. Furthermore, joint monitoring using BeiDou GEO three-frequency signals can ensure monitoring success rate and stability, thus meeting the requirements of high-precision, high-real-time, and high-stability flood monitoring. Considering the problem of low monitoring accuracy caused by direct signal-to-noise ratio, a Gaussian fitting method is used to remove the direct signal-to-noise ratio from the original signal-to-noise ratio observations. A high-frequency random noise is filtered out using an empirical mode decomposition denoising algorithm to extract the high-precision reflected signal-to-noise ratio as the first detection quantity. A linear-to-periodic signal-to-noise ratio (SNR) conversion model is used to transform the SNR of linear reflected signals (in dB-Hz) into the SNR of periodic reflected signals (in volts / volts). The effective reflectance height of the satellite is calculated using Lomb-Scargle spectral analysis to eliminate the influence of non-uniform elevation angle characteristics. The amplitude of the reflected signal SNR is calculated using a Gaussian particle filter algorithm, which serves as the second detection parameter for flood monitoring, improving monitoring accuracy. Joint monitoring using the B1, B2, and B3 bands of the BeiDou GEO satellites further ensures the success rate and accuracy of flood monitoring. Attached Figure Description

[0090] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0091] Figure 1 Flowchart of flood monitoring method based on BeiDou GEO satellite reflected signals.

[0092] Figure 2 Flowchart of a flood monitoring method based on the signal-to-noise ratio of BeiDou GEO satellite reflected signals.

[0093] Figure 3 Flowchart of a flood monitoring method based on the signal-to-noise ratio amplitude of reflected signals from BeiDou GEO satellites. Detailed Implementation

[0094] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0095] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0096] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0097] This invention provides a flood monitoring method based on BeiDou GEO satellite reflected signals, such as... Figure 1 As shown, it includes the following steps:

[0098] S1. Read the original observation values ​​of the BeiDou GEO satellite on the monitoring day, and calculate the elevation angle information and time information at the corresponding epoch. The original observation values ​​include the original signal-to-noise ratio, the original pseudorange, and the original carrier wave.

[0099] Specifically, the raw observations were acquired by continuously collecting raw data from BeiDou GEO satellites on non-flood days using a conventional geodetic receiver. This included signal-to-noise ratio, pseudorange, carrier phase observations, and satellite ephemeris data. Simultaneously, the elevation angle, azimuth angle, and epoch time were calculated for each epoch to provide auxiliary parameters for establishing a flood monitoring model.

[0100] It should be noted that the signal-to-noise ratio (SNR) can be directly extracted from the observation file, and the elevation angle and time information at the corresponding epoch can be calculated using pseudorange, carrier phase, and satellite ephemeris. This process is common knowledge in the positioning field, and the calculation process will not be detailed here. For ease of explanation later, the raw SNR extracted here will be used in subsequent steps. S d+r+n This indicates that the calculated elevation angle is expressed using... EL express.

[0101] S2. The original signal-to-noise ratio is processed to obtain the reflected signal-to-noise ratio, which is then used as the first monitoring quantity.

[0102] Understandably, the extracted raw signal-to-noise ratio S d+r+n This includes the signal-to-noise ratio of the direct signal (expressed as...). S d ), reflected signal-to-noise ratio (expressed as S r ), and random noise (represented as S n Therefore, the original signal-to-noise ratio needs to be processed. The specific processing method is as follows:

[0103] S2-1. Use a Gaussian fitting method to eliminate the influence of the signal-to-noise ratio of the direct signal.

[0104] S2-1-1. Use the Gaussian fitting algorithm to fit the observed values ​​of the original signal-to-noise ratio to obtain the Gaussian fitting function.

[0105] The solution process for the Gaussian fitting function is as follows: Assume the original signal-to-noise ratio observation dataset is ( x i , S i )( i =1,2,3,...., N ), N Indicates signal length. x Represents the time series of a signal. S This represents the raw signal-to-noise ratio at that time. This dataset can be described using a Gaussian function, with the following formula:

[0106] (1)

[0107] In the formula, S i and x i It is known. exp This represents exponentiation, a common mathematical calculation. a , b and cThe unknown parameters represent the peak height, peak position, and half-width of the Gaussian function curve, respectively.

[0108] a , b and c These three parameters can be solved by the following process: Taking the natural logarithm of both sides of equation (1), equation (1) can be transformed into:

[0109] (2)

[0110] make: , , , Then equation (2) can be transformed into a quadratic polynomial fitting function, expressed as follows:

[0111] (3)

[0112] Considering all data and measurement errors, the results are presented in matrix form as follows:

[0113] (4)

[0114] The above formula can be simplified to:

[0115] (5)

[0116] In the formula, Z Given a quantity, from Calculations show that X Given a quantity, from x i Therefore, without considering overall measurement error, E Under the influence of various factors, the fitting constants can be obtained according to the least squares principle. b 0, b 1, b 2 matrices B The generalized least squares solution is expressed as:

[0117] (6)

[0118] In the formula, X and Z All are known quantities. T The matrix transpose operation is represented by equation (6), and the fitting constant can be obtained by equation (6). b 0, b 1, b 2. Furthermore, the unknown parameters can be obtained using the following formula. a , b , c :

[0119] (7)

[0120] In equation (7) b 0, b 1, b 2 has been obtained through equation (6), e Since is a natural constant and is common mathematical knowledge, it will not be explained here. Therefore, the unknown parameter can be obtained through equation (7). a , b , c Therefore, we can obtain the Gaussian function fitted by the original signal-to-noise ratio data, i.e., equation (1).

[0121] S2-1-2. Using the Gaussian fitting function obtained in the previous step, calculate the fitting value for each epoch, expressed as: S G The calculation process can be expressed as:

[0122] (8)

[0123] In equation (8) a , b , c It has already been obtained through equation (7), and x i Since it represents a time series, the fitted value at the corresponding time point can be calculated. S G ;

[0124] S2-1-3, Utilizing the original signal-to-noise ratio S d+r+n Subtract the fitted value calculated in the previous step S G The result is a signal containing only the signal-to-noise ratio of the reflected signal and random noise, expressed by the formula:

[0125] S r+n = S d+r+n - S G (9)

[0126] The above process can eliminate the influence of the direct signal's signal-to-noise ratio from the original signal-to-noise ratio observations, resulting in a signal that contains only the reflected signal's signal-to-noise ratio and random noise. S r+n .

[0127] S2-2. High-frequency random noise is filtered out using an empirical mode decomposition denoising algorithm to obtain a signal that retains only the signal-to-noise ratio of the reflected signal.

[0128] Understandably, Empirical Mode Decomposition (EMD) essentially smooths a noisy signal by cyclically decomposing it into Intrinsic Mode Functions (IMFs) and residuals until the resulting IMFs meet the following two assumptions: 1) The number of zero-crossing points within the curve plotted by the IMFs must differ from the number of all extreme points on the curve by no more than one. 2) The sum of the values ​​of the upper and lower envelopes at the same point in the curve plotted by the IMFs is zero. Once these two conditions are met, the decomposition stops. The resulting IMFs are then used to reconstruct the signal, achieving denoising. This method effectively filters out random noise while preserving the useful signal to the greatest extent possible. The denoising algorithm based on EMFs mainly consists of two parts:

[0129] S2-2-1, Signal Decomposition

[0130] For a given signal S r+n ( x The process involves finding the locations of all local maxima and minima on the signal, and then using an interpolation function to generate the given signal from all the maxima and minima. S r+n ( x The upper and lower envelopes of the expression are given by the expression, where the upper envelope is generated by maximum interpolation and is represented as follows: a 1. The lower envelope is generated by minimum interpolation, denoted as: b 1. It should be noted that the interpolation function can use Lagrange interpolation or cubic spline interpolation. Both of these methods are common knowledge in the field of mathematics and will not be described here.

[0131] Furthermore, the average of the two obtained envelopes is calculated and expressed as:

[0132] (10)

[0133] In the formula a 1 and b 1. That is, the upper and lower envelopes are generated by the interpolation function, and then the given signal is used. S r+n ( x Subtracting the value obtained from equation (10) The first set of residuals was obtained. , is represented as:

[0134] (11)

[0135] The result calculated in judgment (11) Does it meet the two assumptions of the intrinsic mode function? If not, then... The above decomposition process is repeated as a new input signal until the decomposed signal is obtained. This continues until the two assumptions regarding the intrinsic mode functions are met.

[0136] Assuming decomposition to the th k After that It satisfies the two assumptions of the intrinsic mode function (IMF), the first k The decomposition is represented as follows:

[0137] (12)

[0138] Then at this time That is S r+n ( x The first intrinsic mode function (IMF) derived from the decomposition, denoted as IMF1, has the shortest period and the highest frequency among all subsequent IMFs derived from the decomposition.

[0139] Subtracting the first intrinsic mode function (IMF1) from the original signal yields a new data sequence with IMF1 removed. r 1. For this new sequence r 1. Continue repeating the process of obtaining IMF1 to obtain the second intrinsic mode function IMF2, and then use... r Subtracting the second intrinsic mode function (IMF2) from 1 yields a new data sequence. r 2; Repeat this calculation until the last one. r m The process continues until it can no longer be decomposed; the above process can be represented as:

[0140] (13)

[0141] at this time, r m That is the original signal S r+n ( x The residual signal remaining after empirical mode decomposition is expressed as: Res .

[0142] S2-2-2, Signal Reconstruction

[0143] After the above empirical mode decomposition of the signal, each intrinsic mode function (IMF) represents an independent frequency component of the original signal, arranged in descending order of frequency. Therefore, all the separated IMFs and the residuals can be used to... Res The signal is reconstructed to obtain the denoised signal. The reconstruction formula is as follows:

[0144] (14)

[0145] After the above process, high-frequency random noise can be effectively filtered out, resulting in a signal that retains only the signal-to-noise ratio of the reflected signal. , represented as This signal is used as the primary monitoring quantity.

[0146] S3. Obtain the amplitude value of the reflected signal signal-to-noise ratio through the reflected signal signal-to-noise ratio, and use it as the second monitoring quantity.

[0147] The obtained high-precision signal-to-noise ratio (SNR) of the reflected signal is further processed. A linear-to-periodic SNR conversion model is used to convert the linear reflected signal SNR in dB-Hz to a periodic reflected signal SNR in volts / volts. The effective satellite reflectance height is calculated using Lomb-Scargle spectral analysis to eliminate the influence of non-uniform elevation angle characteristics. The amplitude of the reflected signal SNR is calculated using a Gaussian particle filter algorithm, which serves as the second monitoring parameter for flood monitoring.

[0148] Specifically, the method for obtaining the amplitude value of the reflected signal-to-noise ratio:

[0149] S3-1. The signal-to-noise ratio (SNR) of a linearly reflected signal in dB-Hz is converted to the SNR of a periodically reflected signal in volts / volts using a linear-to-periodic conversion model. The conversion formula for the linear-to-periodic conversion model is expressed as follows:

[0150] (15)

[0151] In the formula, This is a linear signal that contains only the signal-to-noise ratio of the reflected signal, measured in dB-Hz. The signal-to-noise ratio of the periodic reflected signal, expressed in volts / volts, is obtained after the linear-periodic transformation model.

[0152] S3-2. Determining the effective reflectivity of a satellite using the Lomb-Scargle spectral analysis method. .

[0153] The process mainly consists of two parts. The first part is to determine the signal-to-noise ratio of the periodic reflection signal using the Lomb-Scargle spectral analysis method. The frequency at which the power is at its maximum (expressed as: The second part is to utilize... And satellite effective reflectivity (expressed as: The effective reflectance of the satellite is determined by the relationship between the two parameters. The specific process is as follows:

[0154] S3-2-1. Determining the signal-to-noise ratio of periodic reflection signals using the Lomb-Scargle spectral analysis method. S r_v Frequency at maximum power The specific process is as follows:

[0155] The power spectrum of the signal is determined using the Lomb-Sacragle spectral analysis method, and the formula is as follows:

[0156] (16)

[0157] In the formula, P x ( f ) is a frequency of f The power of the periodic signal; S r_v ( x () represents the signal-to-noise ratio of the reflected signal. N For sequence length, τ As a time-shift invariant, it can be calculated using the following formula:

[0158]

[0159] The following can be obtained using equation (16-17). P x ( f When the peak value is reached, the corresponding ;

[0160] S3-2-2, Utilizing and satellite effective reflection high The relationship between the effective reflectance of the satellite and the signal-to-noise ratio frequency of the reflected signal is used to determine the effective reflectance of the satellite. The relationship between the effective reflectance of the satellite and the effective reflectance is expressed as follows:

[0161]

[0162] In the formula, It has been calculated using formula (16-17). λ This refers to the wavelength of the signal, which is a known quantity. For example, the wavelength of the BeiDou B1 band is 19.2 cm, which is common knowledge in the field. The effective reflectivity of the satellite can be calculated through the above process.

[0163] S3-3. The Gaussian particle filter algorithm is used to calculate the signal-to-noise ratio amplitude of the reflected signal. The specific formula is as follows:

[0164]

[0165] In the formula, A and The amplitude and phase of the signal-to-noise ratio represent the unknowns in the formula. S r_v It is the signal-to-noise ratio of the periodically reflected signal. The satellite has high effective reflectivity. λ It is the wavelength of the signal, which is a known quantity. EL That is the corresponding elevation angle.

[0166] Understandably, the above equation can be solved using the least squares algorithm, Kalman filter, or Gaussian particle filter algorithm.

[0167] Considering the non-uniform distribution of the elevation angle, this embodiment employs the Gaussian particle filter algorithm to improve the accuracy of parameter calculation. The Gaussian particle filter algorithm combines the advantages of particle filtering and Gaussian filtering. It approximates the system state distribution using a single Gaussian distribution within the basic framework of particle filtering, thereby obtaining the optimal parameter solution. This algorithm is a mature one, and many open-source codes can be implemented in the Matlab platform. Therefore, only a simplified algorithm flow is described here, as follows:

[0168] Assume a length of N of m Gaussian random variable X The probability density can be written as:

[0169]

[0170] In the formula, ∑ is the signal mean, and ∑ is the variance matrix of the signal. T This indicates the transpose operation. exp These are exponential operations, and both are considered basic mathematical concepts. Let's assume... k- The conditional probability distribution can be obtained at time 1. The approximation is as follows:

[0171]

[0172] When new observations are input, the conditional probability distribution can be obtained through the prediction and update steps of particle filtering. The approximation is as follows:

[0173]

[0174] In most cases, the mean in the above formula and variance Since the analytical expression cannot be obtained, it is necessary to use the particles and their weights in particle filtering to find the answer. and The estimated value. Therefore, assume the signal length is... NThe execution process of the Gaussian particle filter algorithm can be summarized as follows:

[0175] 1) From The sample set is obtained by sampling from the middle. ;

[0176] 2) From respectively ( i =1,2,..., N Sampling is performed from ) to obtain the sample set. ;

[0177] 3) Calculate the sample mean and variance , means as follows:

[0178]

[0179] In the formula, T This is the transpose operation, a common operation in linear algebra.

[0180] 4) From the importance probability density Samples drawn from .generally It can be adopted or approximate.

[0181] 5) Utilize the obtained observations y k Calculate each x k ( i The weight of ) The calculation formula is expressed as follows:

[0182]

[0183] 6) Normalized weights get The calculation formula is expressed as follows:

[0184]

[0185] 7) Utilize Calculate the mean and variance The calculation formula is expressed as follows:

[0186] (26)

[0187] Thus obtain k time The approximation is as follows:

[0188]

[0189] In the formula, the mean It can be seen as k time The estimated value.

[0190] In this specific case... x These are time series values ​​that include unknown variables. y This refers to the corresponding known quantities. In multidimensional situations... x and y It exists in matrix form. Therefore, using the Gaussian particle filter algorithm described above, the amplitude value of the signal-to-noise ratio of the reflected signal in formula (19) can be accurately calculated. A And phase value. In this example embodiment, only the amplitude value is used. A The amplitude value of the reflected signal-to-noise ratio. A As the second monitoring quantity.

[0191] S4. A reference monitoring model for BeiDou GEO satellites based on the first and second monitoring quantities is established using a satellite-based and frequency-based method. The reference monitoring model stores the first and second monitoring quantities for the corresponding epochs of each year in the form of a database text on a local computer. It also stores the elevation angle and time information for the corresponding epochs, as well as the corresponding monitoring thresholds. The monitoring thresholds are that the fluctuation of the first monitoring quantity does not exceed 2dB-Hz, and the fluctuation of the second monitoring quantity does not exceed 0.5dB-Hz.

[0192] Understandably, a reference monitoring model is established (the model is stored on the local computer in the form of a database). Using the calculated first and second monitoring quantities, along with corresponding auxiliary information, a flood monitoring model is established based on the signal-to-noise ratio (SNR) and amplitude of the reflected signal-to-noise ratio from the BeiDou GEO satellite. Reference monitoring models for the first and second monitoring quantities are established separately for each BeiDou GEO satellite and frequency, and stored according to the time, elevation angle, and other information of the corresponding epoch, laying the foundation for subsequent flood monitoring. Specifically, in this embodiment, the storage method is as follows: the first detection quantity at 12:12:30 (taking a data sampling rate of 30 seconds as an example) on the B1 band of the BeiDou B01 satellite is S. B01_B1 The second detection quantity is SA B01_B1 The corresponding elevation angle is: EL B01_B1 Time information, such as 12:12:30 in this example, is represented by the symbol: T B01_B1 .

[0193] The S5 system, along with the first and second monitoring quantities of the combined BeiDou GEO tri-frequency signals, monitors flooding using a reference monitoring model and labels the results based on monitoring thresholds. Specifically, for example... Figure 2 As shown,

[0194] S5-1. Flood monitoring is carried out using the signal-to-noise ratio of three-frequency reflected signals from the BeiDou GEO satellite, and the monitoring results are marked. The marking levels include primary marking and secondary marking.

[0195] First, the raw signal-to-noise ratio at the current epoch is obtained. Step S2 is then used to calculate the first monitoring value at the current epoch, and the elevation angle and time information at that epoch are calculated to provide a benchmark for subsequent search of the reference model. The elevation angle calculation process is relatively simple and is a standard method in this field, so it will not be described here.

[0196] Secondly, the calculated elevation angle and time information of the three frequencies of BeiDou GEO satellite, namely the B1 band, B2 band and B3 band, are input into the reference monitoring model to search for the first monitoring quantity in the corresponding reference value. The first monitoring quantity at the current epoch is compared with the first monitoring quantity in the reference value. If the fluctuation of the first monitoring quantity at the current epoch is greater than 2dB-Hz, flooding exists, and it is marked.

[0197] Finally, statistical analysis is performed based on the epoch marking results. If the marked epochs account for more than 80% of the total number of epochs, it is considered that there is a strong flood correlation and a first-level marking is made. If the marked epochs account for 60%-80% of the total number of epochs, it is considered that there is a relatively strong flood correlation and a second-level marking is made. If the marked epochs account for less than 60%, no marking is made. Finally, the monitoring results are saved and output.

[0198] S5-2. Flood monitoring is conducted using the amplitude value of the signal-to-noise ratio of the three-frequency reflected signals from the BeiDou GEO satellite, and the monitoring results are marked. The marking levels include primary marking and secondary marking; specifically, such as... Figure 3 As shown,

[0199] First, the amplitude value of the reflected signal-to-noise ratio at the current epoch is calculated by step S3 based on the reflected signal-to-noise ratio calculated in step S5-1, which is the second detection quantity at the current epoch. Then, the elevation angle and time information at that epoch are calculated.

[0200] Secondly, the calculated elevation angle information and time information of the three frequencies of BeiDou GEO satellite, namely the B1 band, B2 band and B3 band, are input into the reference monitoring model to search for the second monitoring quantity in the corresponding reference value. The calculated second monitoring quantity at the current epoch is compared with the second monitoring quantity in the reference value. If the amplitude fluctuation of the reflected signal-to-noise ratio at the current epoch is greater than 0.5, then flooding exists, and it is marked.

[0201] Finally, statistical analysis is performed based on the epoch marking results. If the marked epochs account for more than 80% of the total number of epochs, it is considered that there is a strong flood correlation and a first-level marking is made. If the marked epochs account for 60%-80% of the total number of epochs, it is considered that there is a relatively strong flood correlation and a second-level marking is made. If the marked epochs account for less than 60%, no marking is made. Finally, the monitoring results are saved and output.

[0202] S5-3. The flood monitoring results of the first and second monitoring quantities are processed together. If the first and second monitoring quantities have a first-level mark, it is considered that there is a flood risk and a first-level flood warning is required. If the first and second monitoring quantities do not have a first-level mark but have a second-level mark, a second-level flood warning is required. If neither has a mark, no warning is issued.

[0203] After processing by the above flood monitoring algorithm, the problems of low monitoring accuracy caused by orbital fluctuations in existing IGSO and MEO satellites can be effectively solved. It can also solve the problems of low monitoring accuracy and unstable monitoring performance caused by ignoring the influence of direct signals and the effect of the signal-to-noise ratio amplitude of reflected signals, thus providing a guarantee for timely and accurate flood monitoring.

[0204] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings, but the present invention is not limited to the described embodiments. For those skilled in the art, various changes, modifications, substitutions, and variations can be made to these embodiments, including components, without departing from the principles and spirit of the present invention, and these variations still fall within the protection scope of the present invention.

Claims

1. A flood monitoring method based on BeiDou GEO satellite reflected signals, characterized in that, Includes the following steps: S1. Read the original observation values ​​of the BeiDou GEO satellite on the monitoring day, and calculate the elevation angle information and time information at the corresponding epoch. The original observation values ​​include the original signal-to-noise ratio, the original pseudorange, and the original carrier wave. S2. The original signal-to-noise ratio is processed to obtain the reflected signal-to-noise ratio, which is then used as the first monitoring quantity. In step S2, the method for processing the raw signal-to-noise ratio is as follows: S2-1. Use a Gaussian fitting method to eliminate the influence of the signal-to-noise ratio of the direct signal; The specific method for step S2-1 is as follows: S2-1-1. Use the Gaussian fitting algorithm to fit the original signal-to-noise ratio observations to obtain the Gaussian fitting function. The process of solving the Gaussian fitting function is as follows: Assume the original signal-to-noise ratio observation dataset is (x... i ,S i )in, i = 1, 2, 3, ..., N, where N represents the signal length, x represents the time series of the signal, and S represents the raw signal-to-noise ratio at that time. This dataset is described using a Gaussian function, with the following formula: （1） In the formula, S i and x i It is known that exp represents exponential operation, and a, b, and c are unknown parameters, representing the peak height, peak position, and half-width of the Gaussian function curve, respectively. The three parameters a, b, and c are obtained through the following process: Taking the natural logarithm of both sides of equation (1), equation (1) is transformed into: （2） make: , , , Then equation (2) is transformed into a quadratic polynomial fitting function, expressed as follows: （3） Considering all data and measurement errors, the results are presented in matrix form as follows: （4） The above formula simplifies to: （5） In the formula, Z is a known quantity, derived from... The calculation shows that X is a known quantity, and from x... i Therefore, without considering the influence of the overall measurement error E, the generalized least squares solution of the matrix B composed of the fitting constants b0, b1, and b2, obtained according to the least squares principle, is expressed as: (6) In the formula, X and Z are known quantities, and T represents the matrix transpose operation. The fitting constants b0, b1, and b2 are obtained through formula (6), and the unknown parameters a, b, and c are further obtained through the following formulas: （7） In equation (7), b0, b1, and b2 have been obtained through equation (6), and e is a natural constant. Therefore, the unknown parameters a, b, and c can be obtained through equation (7), and then the Gaussian function fitted by the original signal-to-noise ratio data can be obtained, which is equation (1). S2-1-2. Using the Gaussian fitting function obtained in the previous step, calculate the fitting value at each epoch, denoted as S. G The calculation process is expressed as follows: （8） In equation (8), a, b, and c have already been obtained from equation (7), while x i Since this represents a time series, we calculate the fitted value S at the corresponding time point. G ; S2-1-3, Utilizing the original signal-to-noise ratio S d+r+n Subtract the fitted value S calculated in the previous step G The result is a signal containing only the signal-to-noise ratio of the reflected signal and random noise, expressed by the formula: S r+n =S d+r+n -S G (9) The above process eliminates the influence of the direct signal-to-noise ratio from the original signal-to-noise ratio observations, resulting in a signal S that contains only the reflected signal-to-noise ratio and random noise. r+n ; S2-2. High-frequency random noise is filtered out using an empirical mode decomposition denoising algorithm to obtain a signal that retains only the signal-to-noise ratio of the reflected signal. The specific method for step S2-2 is as follows: S2-2-1, Signal Decomposition For a given signal S r+n (x) is processed to find the locations of all local maxima and minima on the signal, and then all the maxima and minima are used to generate the given signal S through an interpolation function. r+n The upper and lower envelopes of (x) are generated by interpolation of the maximum value, denoted as a1, and the lower envelope is generated by interpolation of the minimum value, denoted as a1. , The average of the two obtained envelopes is expressed as: （10） In the formula, a1 and That is, the upper and lower envelopes are generated by the interpolation function, and then the given signal S is used. r+n (x) minus the result of equation (10) The first set of residuals was obtained. , is represented as: （11） The result calculated in judgment (11) Does it meet the two assumptions of the intrinsic mode function? If not, then... The above decomposition process is repeated as a new input signal until the decomposed signal is obtained. This continues until the two assumptions regarding the intrinsic mode functions are met. Assuming the decomposition reaches the k-th iteration... The k-th decomposition, satisfying the two assumptions of the Intrinsic Mode Function (IMF), is expressed as: （12） Then at this time That is S r+n The first eigenmode function decomposed from (x), denoted as IMF1, has the shortest period and the highest frequency among all subsequent eigenmode functions decomposed from x. Subtracting the first intrinsic mode function (IMF1) from the original signal yields a new data sequence r1. The process of obtaining IMF1 is repeated on this new sequence r1 to obtain the second IMF2. Then, subtracting the second IMF2 from r1 yields a new data sequence r2. This calculation is repeated until the final r... m The process continues until it can no longer be decomposed; the above process can be represented as: (13) At this time, r m That is the original signal S r+n (x) The residual signal remaining after empirical mode decomposition is represented as Res; S2-2-2, Signal Reconstruction After the above empirical mode decomposition of the signal, each intrinsic mode function (IMF) represents an independent frequency component in the original signal, arranged in descending order of frequency. Therefore, the signal is reconstructed using all the separated IMFs and the residual Res to obtain the denoised signal. The reconstruction formula is as follows: （14） After the above process, high-frequency random noise is effectively filtered out, resulting in a signal that retains only the signal-to-noise ratio of the reflected signal. , represented as And use this signal as the first monitoring quantity; S3. Obtain the amplitude value of the reflected signal-to-noise ratio through the reflected signal-to-noise ratio, and use it as the second monitoring quantity; In step S3, the method for obtaining the amplitude value of the reflected signal-to-noise ratio is as follows: S3-1. The signal-to-noise ratio (SNR) of a linearly reflected signal in dB-Hz is converted to the SNR of a periodically reflected signal in volts / volts using a linear-to-periodic conversion model. The conversion formula for the linear-to-periodic conversion model is expressed as follows: （15） In the formula, This is a linear signal that contains only the signal-to-noise ratio of the reflected signal, measured in dB-Hz. The signal-to-noise ratio of the periodic reflected signal, expressed in volts / volts, is obtained after the linear-periodic transformation model. S3-2. Determining the effective reflectivity of a satellite using the Lomb-Scargle spectral analysis method. ; In step S3-2, the effective reflectivity of the satellite Solution method: S3-2-1. Determine the signal-to-noise ratio S of the periodic reflection signal using the Lomb-Scargle spectral analysis method. r_v Frequency at maximum power The specific process is as follows: The power spectrum of the signal is determined using the Lomb-Sacragle spectral analysis method, and the formula is as follows: In the formula, P x (f) is the power of a periodic signal with frequency f; S r_v (x) represents the signal-to-noise ratio of the reflected signal, N is the sequence length, and τ is the time-shift invariant, which is calculated using the following formula: P can be obtained using equations (17) and (18). x (f) corresponds to the peak value ; S3-2-2, Utilizing and satellite effective reflection high The relationship between the effective reflectance of the satellite and the signal-to-noise ratio frequency of the reflected signal is used to determine the effective reflectance of the satellite. The relationship between the effective reflectance of the satellite and the effective reflectance is expressed as follows: (19) In the formula, λ is the wavelength of the signal, which is a known quantity. The effective reflection height of the satellite is calculated through the above process. ; S3-3. The Gaussian particle filter algorithm is used to calculate the signal-to-noise ratio amplitude of the reflected signal. The specific formula is as follows: (16) In the formula, and The amplitude and phase of the signal-to-noise ratio represent the unknowns in the formula. It is the signal-to-noise ratio of the periodically reflected signal. λ is the effective reflection height of the satellite, λ is the wavelength of the signal, which is a known quantity, and EL is the corresponding elevation angle; S4. A reference monitoring model for BeiDou GEO satellites based on the first and second monitoring quantities is established using a satellite-based and frequency-based method. The reference monitoring model stores the first and second monitoring quantities for the corresponding epochs of each year in the form of a database text on a local computer. It also stores the elevation angle and time information for the corresponding epochs, as well as the corresponding monitoring thresholds. S5, together with the first and second monitoring quantities of the Beidou GEO tri-frequency signals, monitors floods through a reference monitoring model and marks the results according to the monitoring threshold.

2. The flood monitoring method based on BeiDou GEO satellite reflected signals according to claim 1, characterized in that, In step S4, the monitoring thresholds are that the fluctuation of the first monitoring quantity does not exceed 2dB-Hz and the fluctuation of the second monitoring quantity does not exceed 0.

5.

3. The flood monitoring method based on BeiDou GEO satellite reflected signals according to claim 1, characterized in that, Step S5 includes the following steps: S5-1. Flood monitoring is conducted using the signal-to-noise ratio of three-frequency reflected signals from the BeiDou GEO satellite, and the monitoring results are marked accordingly. The marking levels include primary marking and secondary marking. S5-2. Flood monitoring is carried out using the amplitude value of the signal-to-noise ratio of the three-frequency reflected signals based on the BeiDou GEO satellite, and the monitoring results are marked. The marking levels include primary marking and secondary marking. S5-3. The flood monitoring results of the first and second monitoring quantities are processed together. If the first and second monitoring quantities have a first-level mark, it is considered that there is a flood risk and a first-level flood warning is required. If the first and second monitoring quantities do not have a first-level mark but have a second-level mark, a second-level flood warning is required. If neither has a mark, no warning is issued.

4. The flood monitoring method based on BeiDou GEO satellite reflected signals according to claim 3, characterized in that, The method for monitoring flooding in step S5-1 is as follows: First, the raw signal-to-noise ratio at the current epoch is obtained. The first monitoring quantity at the current epoch is calculated through step S2, and the elevation angle information and time information at that epoch are calculated. Secondly, the calculated elevation angle and time information of the three frequencies of BeiDou GEO satellite, namely the B1 band, B2 band and B3 band, are input into the reference monitoring model to search for the first monitoring quantity in the corresponding reference value. The first monitoring quantity at the current epoch is compared with the first monitoring quantity in the reference value. If the fluctuation of the first monitoring quantity at the current epoch is greater than 2dB-Hz, flooding exists, and it is marked. Finally, statistical analysis is performed based on the epoch marking results. If the marked epochs account for more than 80% of the total number of epochs, it is considered that there is a strong flood correlation and a first-level marking is made. If the marked epochs account for 60%-80% of the total number of epochs, it is considered that there is a relatively strong flood correlation and a second-level marking is made. If the marked epochs account for less than 60%, no marking is made. Finally, the monitoring results are saved and output.

5. The flood monitoring method based on BeiDou GEO satellite reflected signals according to claim 4, characterized in that, The method for monitoring flooding in step S5-2 is as follows: First, the amplitude value of the reflected signal-to-noise ratio at the current epoch is calculated by step S3 based on the reflected signal-to-noise ratio calculated in step S5-1, which is the second detection quantity at the current epoch. Then, the elevation angle and time information at that epoch are calculated. Secondly, the calculated elevation angle information and time information of the three frequencies of BeiDou GEO satellite, namely the B1 band, B2 band and B3 band, are input into the reference monitoring model to search for the second monitoring quantity in the corresponding reference value. The calculated second monitoring quantity at the current epoch is compared with the second monitoring quantity in the reference value. If the amplitude fluctuation of the reflected signal-to-noise ratio at the current epoch is greater than 0.5, then flooding exists, and it is marked. Finally, statistical analysis is performed based on the epoch marking results. If the marked epochs account for more than 80% of the total number of epochs, it is considered that there is a strong flood correlation and a first-level marking is made. If the marked epochs account for 60%-80% of the total number of epochs, it is considered that there is a relatively strong flood correlation and a second-level marking is made. If the marked epochs account for less than 60%, no marking is made. Finally, the monitoring results are saved and output.

Images