Sea-land-air-space multi-source meteorological detection data fusion method and device and medium

Abstract

The invention discloses a sea-land-air-space multi-source meteorological detection data fusion method and device and a medium. The method comprises the following steps: performing quality control on satellite observation data, performing format conversion, and feeding the satellite observation data to a data receiving end; an airborne platform coordinate system of airborne meteorological radar reflectivity factors, speed and spectral width is converted into a geodetic coordinate system, quality control and format conversion are carried out on data collected by the airborne platform converted into the geodetic coordinate system and sounding data, and then the data and the sounding data are fed to a data receiving end; the data detected by the detection equipment is subjected to quality control and format conversion and then is fed to a data receiving end; and the meteorological sensor data is subjected to quality control and format conversion and then is fed to a data receiving end. According to the invention, the use efficiency of multi-band detection data is improved; meanwhile, scanning data of'points, lines, surfaces and bodies' are converted into three-dimensional lattice point observation data with higher use value.

Description

Technical Field

[0001] The present invention relates to the field of meteorological detection in the meteorological field, and more specifically, to a method, device and medium for fusing sea, land, air and space multi-source meteorological detection data. Background Art

[0002] With the rapid development of meteorological observation systems, more and more observation data are obtained using ground-based automatic weather stations, radars, satellites, etc., and the quality of data simulated by various numerical models is also constantly improving. At the same time, various industries have higher and higher requirements for gridded, spatiotemporally continuous meteorological data products. It is an effective means to use data fusion and data assimilation technology to integrate observation data from multiple sources and multi-mode simulation data to obtain high-precision, high-quality, spatiotemporally continuous multi-source data fusion meteorological grid products.

[0003] The Earth System Research Laboratory of NOAA in the United States has developed the Local Analysis and Prediction System (LAPS), which can obtain three-dimensional cloud fusion grid data by integrating multi-source observation data such as numerical prediction products, ground, sounding, radar, geostationary meteorological satellites, GPS / MET, wind profile radar, and aircraft, providing better initial fields for numerical prediction models and improving the short-term prediction level of numerical prediction models. Other institutions have also developed similar three-dimensional cloud fusion systems, such as the Advanced Regional Prediction System (ARPS) and Rapid Update Cycle (RUC) in the United States, and Integrated Nowcasting through Comprehensive Analysis (INCA) in Austria (Table 1). In recent years, the United States has developed the GSI-Cloud cloud analysis module by embedding ARPS and RUC in the current business assimilation system (GSI), realizing the cloud analysis function, which can provide the initial field containing higher-precision cloud information for the numerical prediction model, which means that three-dimensional cloud fusion will become an important module of the assimilation system.

[0004] Table 1 Overview of foreign fusion numerical models

[0005]

[0006]

[0007] In 2014, the China Meteorological Administration launched the National Meteorological Science and Technology Innovation Project "Quality Control of Meteorological Data and Multi-source Data Fusion and Reanalysis" (hereinafter referred to as the "Innovation Project"), one of the key research tasks is to develop high-quality land, ocean and three-dimensional cloud and rain multi-source data fusion products and related technologies. Relying on the Innovation Project, the National Meteorological Information Center of the China Meteorological Administration has introduced international advanced fusion technology, digested and absorbed it, and independently innovated it, and built the operational Asian China Meteorological Administration Land Data Assimilation System (CMALand Data Assimilaton System, CLDAS) and China Regional Fusion Precipitation Analysis System (CMA Multi-source Precipitation Analysis System, CMPAS), as well as the Global Sea Surface Temperature Fusion Analysis System (CMA Ocean Data Analysis System-SST, CODAS-SST) and China Regional Three-dimensional Cloud Fusion Analysis System (3D Cloud Analysis System, 3DCloudAS). In 2017, the weather forecast service of the China Meteorological Administration was upgraded from the original site forecast to the intelligent grid forecast. A series of multi-source data fusion products (including temperature, precipitation, humidity, wind, total cloud cover, visibility, etc.) have provided intelligent grid forecast business applications by optimizing product timeliness and adjusting grids. In addition, a series of products produced by the China Meteorological Administration's Land Data Assimilation System (CLDAS), including ground meteorological elements and soil temperature and humidity, runoff, evapotranspiration, etc., also provide China Meteorological Administration's smart agriculture business applications.

[0008] Disadvantages of the above fusion systems:

[0009] 1. Several foreign fusion systems are only suitable for data fusion of their own equipment. These systems cannot recognize the data format of my country's meteorological detection equipment. At the same time, their internal data quality control modules are not suitable for data quality control of my country's meteorological detection equipment.

[0010] 2. At present, the domestic fusion system that has been domestically adapted to foreign countries can only be used to integrate data from automatic weather stations, conventional weather radars, and satellite data. The detection data of some new meteorological detection equipment, such as laser wind radar, laser fog radar, millimeter-wave cloud radar, millimeter-wave wind radar, and airborne meteorological radar, cannot be identified and integrated. Summary of the invention

[0011] The present invention is provided to solve the above problems existing in the prior art. Therefore, a method, device and medium for fusing sea, land, air and space multi-source meteorological detection data are needed to solve the problem that the data of existing new meteorological detection equipment such as laser wind radar, laser fog radar, millimeter wave cloud radar, millimeter wave wind radar, airborne meteorological radar, etc. cannot be recognized and processed by the existing fusion system.

[0012] According to a first aspect of the present invention, a method for fusing sea, land, air and space multi-source meteorological detection data is provided, the method comprising:

[0013] After quality control of satellite observation data, the data is converted into a new format and then fed to the data receiving end;

[0014] The airborne platform coordinate system of the airborne meteorological radar reflectivity factor, speed, and spectrum width is converted into the geodetic coordinate system, and the data collected by the airborne platform and the sounding data converted into the geodetic coordinate system are quality controlled and format converted before being fed to the data receiving end;

[0015] The data detected by the detection equipment is quality-controlled and format-converted and then fed to the data receiving end, wherein the detection equipment includes at least one of a ground automatic meteorological observation station, an X-band weather radar, an S-band weather radar, a C-band weather radar, a P-band wind profiler radar, an L-band wind profiler radar, a laser wind measurement radar, a laser fog measurement radar, a millimeter wave cloud radar, and a millimeter wave wind measurement radar;

[0016] The meteorological sensor data is quality controlled and format converted before being fed to a data receiving end. The meteorological sensor data includes data collected by meteorological sensors installed on automatic weather stations, buoys and / or boats.

[0017] Furthermore, the satellite observation data is quality controlled by the following methods:

[0018] Terrain verification: Eliminate satellite observation data with underlying surfaces of land, sea ice, snow, or mixed types;

[0019] Extreme value test: Eliminate satellite observation data with brightness temperature less than 70K or greater than 320K;

[0020] Threshold test: satellite observation data with a brightness temperature standard deviation greater than 6K are eliminated, and satellite observation data with a brightness temperature difference of observed brightness temperature minus simulated brightness temperature greater than 3.5K are eliminated.

[0021] Furthermore, the airborne platform coordinate system of the airborne meteorological radar reflectivity factor, speed, and spectrum width is converted to the geodetic coordinate system by the following method:

[0022] Acquire attitude information of the airborne platform collected by an attitude sensor in real time, wherein the attitude information includes pitch angle, yaw angle and roll angle;

[0023] In the event of jitter or attitude change during flight, the attitude change is determined based on the attitude information acquired in real time, and the coordinate system of the airborne platform is adjusted to keep the antenna stationary relative to the earth;

[0024] Acquiring in real time the operating parameters of the aircraft, the operating parameters including position, velocity and acceleration, the operating parameters being measured relative to a geodetic coordinate system;

[0025] According to the aircraft's operating parameters and the airborne platform's attitude information, the coordinates of the airborne platform are converted into the geodetic coordinate system.

[0026] Furthermore, the quality control of the data collected by the airborne platform and the sounding data converted into the geodetic coordinate system includes:

[0027] The data collected by the airborne platform and the sounding data converted to the geodetic coordinate system are subjected to ground clutter suppression, velocity deambiguation, data consistency assessment, noise and interference removal, clutter and false echo filtering, atmospheric attenuation correction, and error correction and calibration steps;

[0028] The steps of ground clutter suppression include:

[0029] Use space-time two-dimensional adaptive processing and / or digital signal processing techniques to filter out or suppress fixed target clutter and slow target clutter;

[0030] The steps of deblurring velocity include:

[0031] Design staggered dual waveforms A and B, the maximum unambiguous speeds of waveforms A and B are different and mutually prime, and when the fuzzy speeds of the same target speed on wave A and wave B are the same, the target speed is determined to be the true target speed;

[0032] Or in the target tracking process, the actual displacement of the target is determined based on the multi-frame correlation of the target, and matched with the displacement under multiple fuzzy speeds, and the fuzzy speed with the highest matching degree is taken as the real target speed;

[0033] The steps of data consistency assessment include:

[0034] Compare the radar echo strength at different altitudes and time points to ensure the consistency of data in time and space;

[0035] The steps of noise and interference removal include:

[0036] Use filters and / or smoothing algorithms to remove the effects of electromagnetic interference and random noise;

[0037] The steps of clutter and false echo filtering include:

[0038] Remove clutter and false echoes caused by ground objects, birds, and insects contained in radar data;

[0039] The steps of atmospheric attenuation correction include:

[0040] Use atmospheric attenuation models or measured data to correct radar data for atmospheric attenuation;

[0041] The steps of error correction and calibration include:

[0042] Correct the systematic errors of the radar equipment; and calibrate the radar data to ensure that the radar data at different times and locations have a unified scale and standard; the systematic errors of the radar equipment include antenna pointing error and distance error.

[0043] Furthermore, the meteorological sensor data is quality controlled by the following methods:

[0044] Data collection verification, including the use of verification mechanisms during data collection, including sensor calibration, hardware equipment status monitoring, and data collection software stability testing;

[0045] After data collection, data preprocessing is performed, including denoising, smoothing, standardization and normalization operations;

[0046] Data integrity checks, including determining whether data records are complete, timestamps are accurate, and whether data is missing or abnormal;

[0047] Data outlier identification, including the use of statistical methods, comparative analysis methods or expert experience methods to identify outliers that do not conform to conventional meteorological data;

[0048] Data consistency analysis, including comparing data from different sensors or different observation sites to check whether the data are consistent;

[0049] Correct and amend data when outliers or inconsistent data are identified, where correction methods include interpolation, smoothing or substitution using historical data, data from neighboring sites or algorithm models; correction methods include hardware replacement, software upgrade or re-observation for sensor failure or operational error;

[0050] Data storage management, including storing data in preset data storage formats, storage media, backup strategies and security measures;

[0051] Data quality assessment includes determining the overall quality level of the data based on the completeness, accuracy, consistency and availability of the data.

[0052] Furthermore, the data detected by the X-band weather radar, S-band weather radar, and C-band weather radar are quality controlled by the following methods:

[0053] Ground clutter suppression processing includes one or more of digital signal, spatial domain filtering, MTI filter, adaptive frequency domain filter and clutter reduction decision processing steps. When multiple processing steps are used, the results obtained by the multiple processing steps are compared to select the best processing result.

[0054] The digital signal processing steps include: processing the radar echo signal based on median filtering, mean filtering, and adaptive filtering to filter out or suppress fixed target clutter and slow target clutter and retain the moving target echo;

[0055] The spatial filtering processing step includes: performing spatial filtering using echo signals received by multiple antennas to suppress ground clutter from a specific direction, wherein the ground clutter in the specific direction includes echoes from the sky and the ground;

[0056] The MTI filter processing steps include: using the MTI filter to perform a subtraction operation on pulses in adjacent repetition periods to eliminate fixed target echoes and slow-moving clutter, thereby retaining moving target echoes;

[0057] The adaptive frequency domain filter processing steps include: using the adaptive frequency domain filter to adaptively determine the filtering position by analyzing the shape of the spectrum, and restoring the filtered weather signal by a Gaussian fitting method;

[0058] The clutter reduction decision processing steps include: combining the adaptive filter and CMD algorithm to improve the data quality of the weather radar;

[0059] When the target speed observed by the radar exceeds the maximum unambiguous speed of the radar, de-ambiguation processing is performed, wherein the de-ambiguation processing includes one or more combinations of Doppler frequency shift compensation, phase encoding, and repetition frequency diversity;

[0060] The Doppler frequency shift compensation includes: measuring the Doppler frequency shift of the target and calculating the speed of the target according to the radar system parameters, thereby eliminating the error caused by speed ambiguity;

[0061] The phase coding includes: adding a set phase coding sequence to the transmission signal so that the radar can distinguish different speed echoes and solve the speed ambiguity problem;

[0062] The repetition frequency diversity includes: using a plurality of different repetition frequencies so that the radar can simultaneously receive a plurality of echo signals of different speeds, thereby eliminating speed ambiguity;

[0063] In the case of range folding when the radar uses a pulse repetition frequency higher than the set value, random phase encoding, pulse repetition frequency adjustment, scanning at an elevation angle higher than the current elevation angle, or changing the detection location;

[0064] The adjusting of the pulse repetition frequency includes changing the maximum detection range of the radar by changing the pulse repetition frequency, thereby eliminating range folding in the area of ​​interest.

[0065] Furthermore, the data detected by the P-band wind profiler radar, L-band wind profiler radar and millimeter wave radar are quality controlled by ground clutter suppression, median filtering or smoothing filtering.

[0066] The ground clutter suppression includes: designing an IIR notch filter according to the clutter map, and then performing linear interpolation between two edge points of the notch;

[0067] The median filter includes: determining a neighborhood with a certain center point to be processed as the center; sorting the values ​​of each data in the neighborhood, taking the middle value as the new value of the center point data, and using the neighborhood as a window; after the window moves according to time and height, using the median filter to smooth the data;

[0068] The smoothing filter includes: assuming that the observed object is Y, the independent variable corresponding to Y is X, Δx is used as the spacing of the independent variable X, the independent variables with equal spacing distribution are xi (i=1, 2, 3, ..., n), and according to the value of the observed dependent variable Y, the observed data sequence is obtained as yi (i=1, 2, 3, ..., n), and the noise components in the observed data sequence are filtered out by sliding smoothing, wherein the use of sliding smoothing to filter out the noise components in the observed data sequence includes: taking the data of the i-th point and several points nearby, determining a fitted straight line equation according to the least squares principle, and calculating the dependent variable of the i-th point based on the straight line equation as the data value after sliding smoothing.

[0069] Furthermore, the data detected by the laser wind radar is quality controlled by the following methods:

[0070] Processing of noise and missing values:

[0071] Select an M*N sliding window, divide the data in the window into P intervals with equal size, △d is the interval, and the wind speed value at the center of the window is V ij , the P+1th interval is represented by V ij ±△d / 2;

[0072] Count the distribution frequency of wind speed values ​​in different intervals in the M*N window (X1, X2, X3....X P ) and X′, let X MAX It is X1, X2, X3....X P and X′, take VP as X MAX The middle value of the corresponding interval;

[0073] Let the new value of the window center be V ij , noise removal and missing measurement filling are achieved through the following formula:

[0074]

[0075] Where K1 is the threshold for determining whether to remove noise, and K2 is the threshold for determining whether to fill in missing measurements. ij When it is not 0, it means that this is a valid echo point. When X′<=threshold K1, the point is removed; when X′>threshold K1, the corresponding point is retained; when the original value V in the window ij When it is equal to 0, it is an invalid point. MAX >=threshold K2, fill in the gap and assign the center point VP of the interval corresponding to XMAX to V ij ; if X MAX <threshold K2, no filling is performed;

[0076] The data detected by the laser fog radar is quality controlled by the following methods:

[0077] Dead time correction, calculated as:

[0078]

[0079] Where N is the true count rate, M is the measured count rate, and t is the time resolution;

[0080] Geometric overlap factor correction, calculated as:

[0081]

[0082] Where P(R,λ) is the original echo signal; P′(R,λ) is the signal corrected by the geometric overlap factor; R is the height from the radar; λ is the laser emission wavelength, and O(R) is the geometric overlap factor;

[0083] Background noise subtraction, the calculation formula is:

[0084] P″(R,λ)=P(R,λ)-P bg (R,λ)

[0085] Among them, P(R,λ) is the original echo signal; P″(R,λ) is the signal after background noise subtraction, P bg (R,λ) is the signal of background noise;

[0086] Smoothing the echo signal;

[0087] Cloud and aerosol classification: Calculate the slope of the corrected signal curve and determine the cloud layer or aerosol layer based on the positive and negative changes in the slope value;

[0088] Data quality verification: The quality of the original echo data is verified according to the RCS curve score, the signal-to-noise ratio of the original echo signal, and the consistency with the atmospheric molecular signal. The corresponding quality control identification code and quality control type code are given. The consistency with the atmospheric molecular signal means that above the boundary layer, in the area where the aerosol content in the atmosphere is lower than the set threshold, when the echo signal is set to be only related to the density of the molecules, the molecular backscattering signal calculated according to the lidar equation and the curve after the real echo is subtracted from the background and corrected for the square of the distance are consistent with the atmospheric molecular signal within the set error after taking the log, and then it is determined to be consistent.

[0089] According to a second aspect of the present invention, a device for fusing sea, land, air and space multi-source meteorological detection data is provided, the device comprising:

[0090] A first quality control module is configured to perform quality control on the satellite observation data, convert the format of the data, and then feed the data to the data receiving end;

[0091] The second quality control module is configured to convert the airborne platform coordinate system of the airborne meteorological radar reflectivity factor, speed, and spectrum width into a geodetic coordinate system, perform quality control on the data collected by the airborne platform converted into the geodetic coordinate system and the sounding data, perform format conversion, and then feed them to the data receiving end;

[0092] A third quality control module is configured to perform quality control and format conversion on the data detected by the detection equipment and then feed it to the data receiving end, wherein the detection equipment includes at least one of a ground automatic meteorological observation station, an X-band weather radar, an S-band weather radar, a C-band weather radar, a P-band wind profiler radar, an L-band wind profiler radar, a laser wind measurement radar, a laser fog measurement radar, a millimeter wave cloud radar, and a millimeter wave wind measurement radar;

[0093] The fourth quality control module is configured to perform quality control and format conversion on the meteorological sensor data and then feed it to the data receiving end. The meteorological sensor data includes data collected by meteorological sensors installed on automatic weather stations, buoys and / or boats.

[0094] According to a third aspect of the present invention, there is provided a readable storage medium storing one or more programs, wherein the one or more programs can be executed by one or more processors to implement the method as described above.

[0095] The present invention has at least the following beneficial effects:

[0096] 1. Add the reading interface, data conversion program and quality control program for the laser wind radar radial velocity Vr and signal-to-noise ratio SNR data, so that the laser wind radar radial velocity Vr data can be imported into the LAPS numerical model.

[0097] 2. Add the reading interface, data conversion program and quality control program for the visibility VIS data of the laser fog radar, so that the two types of data of the laser fog radar can be imported into the small and medium-scale forecast numerical model.

[0098] 3. Add the reading interface, data conversion program, and quality control program of millimeter-wave cloud radar reflectivity factor data, so that the millimeter-wave cloud radar reflectivity factor data can be imported into the small and medium-scale forecast numerical model.

[0099] 4. Add the millimeter-wave wind radar velocity data reading interface, data conversion program, and quality control program so that the millimeter-wave wind radar velocity data can be imported into the small and medium-scale forecast numerical model.

[0100] 5. Add the reading interface, data conversion program and quality control program of airborne meteorological radar reflectivity factor and radial velocity data, so that the airborne meteorological radar reflectivity factor and radial velocity data can be imported into the small and medium-scale forecast numerical model.

[0101] 6. Overall, the present invention realizes the numerical fusion of observation data from various meteorological detection equipment on land, sea, air and space rather than a simple "optimal selection" of multi-source observation data; improves the utilization efficiency of detection data of multiple bands (X, S, C, L, P, millimeter wave, laser); at the same time, converts "point, line, surface, and volume" scanning data into more valuable three-dimensional grid observation data. BRIEF DESCRIPTION OF THE DRAWINGS

[0102] Figure 1 A schematic diagram showing the structure and working principle of a sea, land, air and space multi-source meteorological detection data fusion system according to an embodiment of the present invention is shown.

[0103] Figure 2 The present invention shows an overall flow chart of a method for fusing multi-source meteorological detection data from land, sea, air and space according to an embodiment of the present invention.

[0104] Figure 3 A quality control flow chart of satellite observation data according to an embodiment of the present invention is shown;

[0105] Figure 4 A flow chart showing the conversion of airborne platform coordinates into geodetic coordinates according to an embodiment of the present invention is shown;

[0106] Figure 5 A flow chart of airborne weather radar data quality control according to an embodiment of the present invention is shown;

[0107] Figure 6 It shows a data quality control flow chart of an automatic weather station / buoy / ship weather sensor according to an embodiment of the present invention;

[0108] Figure 7 A flow chart of X / S / C band weather radar data quality control according to an embodiment of the present invention is shown;

[0109] Figure 8 A flow chart of L / P band wind profile radar / millimeter wave wind radar data quality control according to an embodiment of the present invention is shown;

[0110] Fig. 9 A flowchart of laser fog radar data quality control according to an embodiment of the present invention is shown;

[0111] Fig.10 A structural diagram of a device for fusion of sea, land, air and space multi-source meteorological detection data according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0112] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. The embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings and specific embodiments, but are not intended to limit the present invention. For the various steps described herein, if there is no necessity for a causal relationship between each other, the order in which they are described as examples herein should not be regarded as a limitation, and those skilled in the art should know that they can be adjusted in order, as long as the logic between them is not destroyed, resulting in the inability to implement the entire process.

[0113] The embodiment of the present invention provides a multi-source meteorological detection data fusion system for land, sea, air and space. The system structure is as follows: Figure 1 As shown, the system can collect, control and integrate data from marine meteorological and hydrological instruments, marine buoys, marine meteorological radars, land X-band weather radars, S-band weather radars, cloud radars, three-dimensional laser wind radars, automatic weather stations, airborne meteorological radars, sounding data, and Fengyun satellites and sunflower-8 satellites in the sky, to obtain meteorological element information such as three-dimensional wind analysis fields, three-dimensional cloud analysis fields, three-dimensional water vapor analysis fields, three-dimensional temperature analysis fields and two-dimensional ground analysis fields, truly realizing the integration of data from meteorological equipment with different bands and detection principles.

[0114] The embodiment of the present invention also provides a method for fusing sea, land, air and space multi-source meteorological detection data. The method can be based on Figure 1 The fusion system shown is implemented. Figure 2As shown, it is a flow chart of the method for fusion of multi-source meteorological detection data from sea, land, air and space. The method includes the following steps S100 to S400, which are described in detail as follows:

[0115] Step S100, quality control is performed on the satellite observation data, the format is converted, and the data is fed to the data receiving end.

[0116] Step S200, convert the airborne platform coordinate system of the airborne meteorological radar reflectivity factor, speed, and spectral width into a geodetic coordinate system, perform quality control on the data collected by the airborne platform converted into the geodetic coordinate system and the sounding data, perform format conversion, and then feed them to the data receiving end.

[0117] Step S300, performing quality control and format conversion on the data detected by the detection equipment and then feeding it to the data receiving end, wherein the detection equipment includes at least one of a ground automatic meteorological observation station, an X-band weather radar, an S-band weather radar, a C-band weather radar, a P-band wind profiler radar, an L-band wind profiler radar, a laser wind measurement radar, a laser fog measurement radar, a millimeter wave cloud radar, and a millimeter wave wind measurement radar;

[0118] Step S400, quality control and format conversion are performed on the meteorological sensor data and then the data is fed to the data receiving end. The meteorological sensor data includes data collected by meteorological sensors installed on automatic weather stations, buoys and / or boats.

[0119] It should be noted that the data receiving end mentioned above can be a small and medium scale numerical model LAPS / STMAS or ARPS, and steps S100-S400 are parallel steps when implemented, that is, the serialization here in this embodiment is only one of the implementation orders. In other implementations, steps S100-S400 can be any execution order, and each step is used to perform quality control and format conversion on different data, that is, the observation data of various meteorological detection equipment on land, sea, air and space are quality controlled and format converted and then sent to the small and medium scale numerical model LAPS / STMAS or ARPS, and then a three-dimensional wind field, a three-dimensional temperature field, a three-dimensional cloud analysis field, a three-dimensional water vapor analysis field and a two-dimensional ground analysis field will be obtained, that is, the acquisition of basic meteorological elements from observation data of multi-source detection equipment is completed, and the conversion of "point, line and surface" observation data to three-dimensional observation data is completed.

[0120] The following embodiments of the present invention will specifically introduce the process of quality control of different types of data.

[0121] like Figure 3 As shown in the figure, the quality control of satellite observation data is achieved through the following steps:

[0122] Step S301, terrain inspection: remove observation data with underlying surfaces of land, sea ice, snow, and mixed types.

[0123] Step S302, extreme value test: eliminate data with brightness temperature less than 70K or greater than 320K.

[0124] Step S303, threshold test: remove data with a brightness temperature standard deviation greater than 6K, and remove data with a brightness temperature greater than 3.5K minus the simulated brightness temperature.

[0125] Step S304, system bias correction: These errors are often comparable to the typical errors of the atmospheric temperature field in the short-term forecast of the numerical model, so the corresponding bias correction must be performed before using the variational method to assimilate satellite observation data. Bias correction can be divided into two categories: one is the scanning bias correction caused by the satellite itself, and the other is the air mass bias correction caused by the change of the atmospheric transmittance itself.

[0126] like Figure 4 As shown in the figure, the conversion of airborne platform coordinates to geodetic coordinates mainly depends on the attitude sensor and navigation equipment in the airborne equipment. The conversion steps are as follows:

[0127] In step S401, the airborne platform first obtains its own attitude information, such as pitch angle, yaw angle and roll angle, through its built-in attitude sensor, which describes the direction and attitude of the aircraft relative to the earth coordinate system.

[0128] Step S402: When the aircraft vibrates or changes its attitude during flight, the attitude sensor monitors these changes in real time and adjusts the coordinate system of the airborne platform to ensure that the antenna can remain stationary relative to the ground.

[0129] Step S403: The navigation device on board the aircraft monitors the operating parameters of the aircraft in real time, such as position, speed, acceleration, etc. These parameters are measured relative to the earth coordinate system.

[0130] Step S404: The navigation device converts the coordinates of the airborne platform according to the operating parameters of the aircraft and the data of the attitude sensor. The purpose of the conversion is to convert the coordinates of the airborne platform into coordinates in the geodetic coordinate system so as to align and compare with the ground or other reference systems.

[0131] It should be noted that the accuracy and precision of coordinate transformation are affected by many factors, such as sensor accuracy, navigation equipment performance, and data processing algorithms. Therefore, when performing coordinate transformation, it is necessary to select appropriate sensors and navigation equipment, and adopt appropriate algorithms and technologies to ensure the accuracy and reliability of the conversion results.

[0132] like Figure 5 As shown in Figure 1, the steps for quality control of airborne weather radar data are as follows:

[0133] Step S501, ground clutter suppression:

[0134] 1) One method is to use space-time two-dimensional adaptive processing (STAP). Adaptive processing can achieve effective matching with complex external environments and can compensate for the influence of system errors to a certain extent, thereby greatly improving the performance of the system. However, adaptive processing often requires the system to flexibly form multiple beams and calculate adaptive weights online in real time, which involves a large amount of calculation.

[0135] In order to reduce the amount of calculation, digital beam forming (DBF) technology and very large scale integrated circuit (VLSI) can be used. The development of these technologies provides a guarantee for accurately controlling the space-time adaptive weights and accelerating the processing speed, thus providing favorable conditions for the practical application of space-time two-dimensional adaptive processing.

[0136] 2) Clutter suppression Digital signal processing technology can also be used to filter out or suppress fixed target clutter and slow target clutter, thereby retaining the moving target echo. This helps to improve the radar's ability to detect targets in a strong clutter background.

[0137] Step S502, deblurring the speed.

[0138] There are two methods to deambiguate velocity in airborne weather radar:

[0139] 1) Design staggered dual waveforms A and B. The maximum unambiguous speeds of the two waveforms are different and mutually prime. Then the fuzzy speeds of the same target speed on wave A and wave B are not exactly the same, and the same one is the true target speed. This method can be solved using the remainder theorem.

[0140] 2) Speed ​​deblurring is not performed during detection, but rather during target tracking. The actual displacement of the target is determined based on the multi-frame correlation of the target, and matched with the displacement under various blurred speeds. The one with the highest matching degree is the real target speed.

[0141] Step S503: data consistency assessment.

[0142] Data consistency assessment mainly checks whether there are contradictions or inconsistencies in radar data from different sources or at different time points. This requires comparative analysis of the data, such as comparing radar echo intensity at different altitudes and different time points, to ensure the consistency of the data in time and space.

[0143] Step S504: noise and interference removal.

[0144] There are often various noises and interferences in radar data, such as electromagnetic interference, random noise, etc. These noises and interferences will affect the accuracy and reliability of radar data. Therefore, it is necessary to remove or reduce the impact of these noises and interferences through filters, smoothing algorithms, etc.

[0145] Step S505: filtering out clutter and false echoes.

[0146] Radar data may contain clutter and false echoes caused by ground objects, flying birds, insects, etc. These clutter and false echoes will interfere with the radar's detection and identification of meteorological targets. Therefore, it is necessary to filter out these clutter and false echoes through appropriate algorithms and technologies to improve the quality of radar data.

[0147] Step S506: atmospheric attenuation correction.

[0148] During the propagation process, radar signals will be attenuated by the atmosphere, which will cause the radar echo intensity to weaken. Therefore, it is necessary to perform atmospheric attenuation correction on the radar data to restore the true intensity of the radar echo. This usually requires correction through atmospheric attenuation models or measured data.

[0149] Step S507: error correction and calibration.

[0150] Error correction and calibration are the last step in radar data quality control. This includes correcting the systematic errors of radar equipment, such as antenna pointing error, distance error, etc., and calibrating radar data to ensure that radar data at different times and locations have a unified scale and standard. Error correction and calibration can further improve the accuracy and reliability of radar data.

[0151] like Figure 6 As shown, the quality control of automatic weather station / buoy / ship weather sensor data includes the following steps:

[0152] Step S601: data collection and verification.

[0153] Data collection is the primary task of automatic weather stations / buoys / ship weather sensors. Ensuring the accuracy of the collected data is the basis of data quality control. During the data collection process, the verification mechanism includes sensor calibration, hardware equipment status monitoring, data collection software stability testing, etc., to ensure that the observation data is reliable at the source.

[0154] Step S602: data preprocessing.

[0155] After data collection, data preprocessing is a necessary step. Preprocessing includes operations such as denoising, smoothing, standardization, and normalization to eliminate or reduce non-meteorological interference in the data, such as sensor noise, electromagnetic interference, etc.

[0156] Step S603: data integrity check.

[0157] Data integrity check is a systematic and comprehensive check of observation data, including whether the data records are complete, whether the timestamps are accurate, data missing and abnormal, etc. If the data is found to be incomplete, it needs to be supplemented or marked in time.

[0158] Step S604: identifying data outliers.

[0159] Outlier identification is to identify outliers that are obviously inconsistent with conventional meteorological data through statistical methods, comparative analysis methods or expert experience. These outliers may be caused by sensor failure, environmental interference or human operation errors.

[0160] Step S605: data consistency analysis.

[0161] Data consistency analysis is to check whether the data is consistent by comparing data from different sensors or different observation sites. This analysis helps to find inconsistencies and potential errors in the data.

[0162] Step S606: data calibration and correction.

[0163] After identifying outliers or inconsistent data, data correction and revision are required. Correction methods include interpolation, smoothing or substitution using historical data, data from neighboring sites or algorithm models. Correction involves hardware replacement, software upgrade or re-observation for sensor failures or operational errors.

[0164] Step S607: data storage management.

[0165] Data storage management involves data storage format, storage media, backup strategies, and security measures. An efficient and stable data storage solution should be selected to ensure long-term preservation and easy retrieval of data. At the same time, data should be backed up regularly to prevent data loss or damage.

[0166] Step S608: data quality assessment.

[0167] Data quality assessment is a summary and feedback of the entire data quality control process. By evaluating indicators such as data integrity, accuracy, consistency, and availability, we can understand the overall quality level of the data. At the same time, based on the evaluation results, we continuously optimize and improve data quality control methods to improve data reliability and accuracy.

[0168] In short, the quality control of automatic weather station / buoy / ship meteorological sensor data is a systematic and comprehensive work involving multiple aspects and links. By implementing effective quality control measures, the quality of observation data can be improved, providing more accurate and reliable data support for meteorological forecasting, climate research and other fields.

[0169] X / S / C band weather radar data quality control is an important part of ensuring the accuracy and reliability of radar data. Figure 7 As shown in the figure, X / S / C band weather radar data quality control includes the following steps:

[0170] Step S701: ground clutter suppression.

[0171] Digital signal processing technology: Through digital signal processing algorithms such as median filtering, mean filtering, adaptive filtering, etc., radar echo signals are processed to filter out or suppress fixed target clutter and slow target clutter, and retain moving target echoes. This method can effectively reduce the interference of ground clutter in the signal processing stage.

[0172] Spatial filtering technology: This technology uses the echo signals received by multiple antennas to perform spatial filtering to suppress ground clutter from a specific direction. This method can effectively distinguish between echoes from the sky and the ground, further improving the radar's target detection accuracy.

[0173] MTI filter: MTI filter, or moving target indication filter, is a radar clutter suppression method. It eliminates fixed target echoes and slow-moving clutter by subtracting pulses in adjacent repetition cycles, thereby retaining moving target echoes. MTI filters can be divided into single delay line cancellers, dual delay line cancellers, and multi-delay line cancellers, with performance improving in turn, but the structural complexity also increases accordingly.

[0174] Adaptive frequency domain filter: Adaptive frequency domain filter analyzes the shape of the spectrum, adaptively determines where to filter, and restores the filtered weather signal through methods such as Gaussian fitting. This method can retain useful weather information to the greatest extent while filtering out clutter.

[0175] Clutter reduction decision algorithm (CMD): CMD algorithm is an algorithm for identifying ground clutter, which can effectively identify and suppress ground clutter. By combining adaptive filters with CMD algorithm, the data quality of X-band weather radar can be further improved.

[0176] Step S702, deblurring the speed.

[0177] When the target speed observed by the radar exceeds the maximum unambiguous speed of the radar, speed ambiguity occurs, causing the radar to be unable to accurately measure the target's speed.

[0178] The basic principle of velocity deambiguation technology is to eliminate the error caused by velocity ambiguity by processing radar echo signals. The specific implementation method can adopt technical means such as Doppler frequency shift compensation, phase coding, and repetition frequency diversity.

[0179] Doppler shift compensation is a method to eliminate velocity ambiguity. It measures the Doppler shift of the target and calculates the target's velocity based on the radar system parameters, thereby eliminating the error caused by velocity ambiguity. Phase coding technology adds a specific phase coding sequence to the transmitted signal, allowing the radar to distinguish different velocity echoes, thereby solving the velocity ambiguity problem. Repetition frequency diversity technology uses multiple different repetition frequencies to enable the radar to simultaneously receive multiple echo signals of different speeds, thereby eliminating velocity ambiguity.

[0180] Step S703, folding the back distance.

[0181] Weather radar range defolding is the process by which the radar signal processor eliminates range folding. When the radar uses a higher pulse repetition frequency (PRF), its maximum detection range becomes shorter. In this case, the radar may receive a strong echo from a target outside its maximum detection range and mistakenly believe that the target is within its maximum detection range, resulting in range folding.

[0182] In order to eliminate this distance folding phenomenon, the radar signal processor will use a series of technical means, such as random phase coding technology, adjusting the pulse repetition frequency (PRF), selecting a higher elevation scanning or changing the detection location, etc. The purpose of these methods is to change the maximum detection range (Rmax) of the radar, thereby eliminating the distance folding phenomenon in the area of ​​concern.

[0183] Among them, adjusting the pulse repetition frequency (PRF) can change the maximum detection range (Rmax) of the radar by changing the PRF, thereby eliminating range folding in the area of ​​interest. In addition, selecting a higher elevation scanning angle can also help overcome the range folding problem, because a higher elevation angle will shorten the detection range of the radar, thereby reducing the possibility of range folding.

[0184] It should be noted that the implementation of the back-off range folding technology needs to take into account multiple factors such as radar system parameters, target characteristics, and environmental conditions. Therefore, in practical applications, it is necessary to select the appropriate back-off range folding method according to the specific situation, and perform corresponding optimization and adjustment to achieve the best effect.

[0185] The data quality control methods for L-band wind profile radar, P-band wind profile radar, and millimeter-wave wind radar (vertical profile scanning mode) are the same. Figure 8 As shown in the figure, the L / P band wind profile radar / millimeter wave wind radar data quality control includes the following steps:

[0186] Step S801: ground clutter suppression.

[0187] According to the clutter map, an IIR notch filter is designed, and then linear interpolation is performed between the two edge points of the notch. This interpolation can make up for the filtering loss of the weather echo signal spectrum caused by removing the spectral component. However, in some cases, when the spectrum of the weather echo signal is very narrow and very close to the ground clutter spectrum, choosing a wide filter will still attenuate some weather echo signals. And sometimes, because the atmospheric signal is completely covered by the ground interference clutter, this method will lose the atmospheric signal.

[0188] Step S802: median filtering.

[0189] Median filtering is a nonlinear smoothing filtering method. It is a neighborhood operation, similar to convolution, but not a weighted summation calculation. Its main principle is: first determine a neighborhood with a certain data center to be processed; then sort the values ​​of each data in the neighborhood, and take the middle value as the new value of the center point data. The neighborhood here is usually called a window; when the window moves according to time and height, the median filtering algorithm can be used to smooth the data well. This paper first performs median filtering on the detection data at the same height over time, and then filters the detection data at the same time with height.

[0190] This algorithm tests the spatiotemporal continuity of the data, and it is effective based on the assumption that the outlier ratio is small, so its disadvantage is that if there are a large number of continuous outliers in the data, this algorithm will consider these outliers to be valid.

[0191] Step S803: smoothing filtering.

[0192] Sliding average filtering is a linear sliding smoothing method. Let the observed object be Y, the corresponding independent variable be X, Δx be the spacing of the independent variable X, the equally spaced independent variables be x1, x2, ..., xi, ..., xn, at these values, observe the values ​​of the dependent variable Y, and the obtained observed data sequence is y1, y2, ..., yi, ..., yn. The noise components in yi (i = 1, 2, 3, ..., n) can be filtered out by applying sliding smoothing. The principle is: take the data of the i-th point and several nearby points, determine a fitted straight line equation according to the least squares principle, and then calculate the dependent variable of the i-th point as the data value after sliding smoothing. Generally speaking, there are three-point linear sliding smoothing, five-point linear sliding smoothing, seven-point linear sliding smoothing and m-point linear sliding smoothing. Generally speaking, the larger the m value, the stronger the suppression of high-frequency noise.

[0193] The quality control of laser wind radar data is achieved through the following scheme a or scheme b:

[0194] a. Processing of missing values

[0195] There are often noise and missing measurements in radar base data. Common noises include the following categories: isolated points, singular points, and radial strip clutter. These are mostly pulse noise and Gaussian noise caused by airplanes, bird flocks, and radio interference. In addition, they also include clutter that is not removed cleanly after being processed by the radar's own signal processor. Missing measurements include missing points, missing small blocks, and missing radials. These missing measurements and noise will affect the radar's subsequent algorithms to a certain extent, so it is necessary to give priority to them. Commonly used noise removal and missing measurement filling methods include median filtering and sliding average. These algorithms lack an effective threshold to measure the scale of removal or filling in data processing. There may be a lot of noise around the missing values. At this time, filling will cause the noise range to expand; and the data determined to be valid when removing noise may itself be noise, which will result in the noise not being effectively removed. This embodiment uses a statistically based algorithm to process the velocity field data. The specific steps are as follows:

[0196] (1) Select an M*N sliding window and divide the data in the window into P intervals of equal size. △d is the interval interval. The wind speed value at the center of the window is V ij , the P+1th interval is represented by V ij ±△d / 2.

[0197] (2) Count the distribution frequency of wind speed values ​​in different intervals within the M*N window (X1, X2, X3...X P ) and X′, let X MAX It is X1, X2, X3....X P and X′, take VP as X MAX The middle value of the corresponding interval.

[0198] (3) Let the new value of the window center be V ij , then the noise removal and missing measurement filling can be realized as follows, where K1 is the threshold for whether to remove noise, K2 is the threshold for whether to fill the missing measurement, and "0" is used instead of no echo. That is, when the original value V of the window center point is ij When it is not 0, it means that this is a valid echo point. When X′<=threshold K1, it means that there are few echo points with similar values ​​around the center point of the window, and the point is a singular point and needs to be removed; when X′>threshold K1, it means that there are many echo points with similar values ​​around it, and the point is judged to be a normal point in the window and needs to be retained. ij When it is equal to 0, it is an invalid point. At this time, it should be judged whether it is a missing point and whether it needs to be filled. MAX When the point is greater than or equal to the threshold value K2, it indicates that the frequency of the point in the window is relatively large, and there are many valid points around it, which can be filled. MAX The center point VP of the corresponding interval is assigned to Vij ;X MAX When <threshold K2, it indicates that there is no echo or very few echo points in the window, and there are not many valid points around it, so the point is considered to be an invalid point and does not need to be filled.

[0199]

[0200] b. Median filtering.

[0201] The median filter is a classic nonlinear signal processing technology. Its basic idea is to process data based on the sorting theory. It can effectively suppress noise interference while protecting the edge of the signal. The signal processing process is: define a sliding window with an odd length, and replace the value of the center point of the window with the median of the values ​​of the other points in the neighborhood of the window. Suppose there is a set of atmospheric lidar data y(n), the specific process of the median filter is: define an initial sliding window with a length of (2w+1), then the output value y^(n) of the nth sampling point in the filtered lidar data y^ can be calculated by the following formula:

[0202] y(n)=med[y(nw),...,y(n),...,y(n+w)]

[0203] In the formula, med() is the median function. It can be seen that this filtering method is relatively simple and easy to implement. Its filtering effect is closely related to the window size. If the window is too small, it may cause more residual noise in the signal and the filtering effect is not obvious enough. However, if the window is too large, it may cause excessive smoothing of the signal, loss of important detail information, and cause de-noising. Therefore, based on the characteristic that the noise of the laser radar echo signal gradually increases with the increase of distance, the choice of window size is generally gradually increased.

[0204] like Fig. 9 As shown in the figure, the quality control of laser fog radar data includes the following steps:

[0205] Step S901, dead time correction.

[0206] The laser radar system consists of a laser emission system, an optical receiving system, a photoelectric conversion and data acquisition system, a signal processing system, etc. In the avalanche diode in the photoelectric conversion link, the upper limit of the photon counting mode count rate depends on the pulse pair resolution, that is, the minimum time interval at which each pulse can be separated. The reciprocal of the pulse pair resolution is the maximum count rate. However, since most events in the light counting area usually occur randomly, the counting pulses may overlap. Taking into account the possibility of pulse overlap, the actual maximum count rate is about one-tenth of the above calculated value. If we denote the true count rate as N, the measured count rate as M, and the time resolution as t, then the true count rate N can be expressed as:

[0207]

[0208] For example, the time resolution of a laser radar is 100ns, and the cumulative sampling times in one minute is 58400. Therefore, the measured count rate M1 = Data1 / (58400*100), the real count rate N1 = M1 / (1-M1·t), t is the dead time, generally provided by the manufacturer. After obtaining the real count rate, the real photon number Data1 = N1*100*58400 can be obtained. The original data of the radar is corrected according to this formula.

[0209] Step S902: geometric overlap factor correction.

[0210] Generally speaking, in the blind zone, the beam emitted by the laser is not in the receiving field of the telescope; in the transition zone, the atmospheric echo signal cannot be completely received by the telescope, resulting in the received echo signal not being consistent with the actual echo signal. The ratio of the number of photons received by the telescope to the actual number of photons is called the geometric overlap factor. This situation needs to be corrected during data processing, which is the geometric overlap factor correction.

[0211] The geometric overlap factor is a parameter that approaches 1 as the distance increases. It is generally provided by the manufacturer. If the geometric overlap factor is O(R), the corrected signal is:

[0212]

[0213] Wherein, P(R, λ) is the original echo signal; P′(R, λ) is the signal corrected by the geometric overlap factor; R is the height from the radar, in m; λ is the laser emission wavelength, in nm.

[0214] Step S903: background noise subtraction.

[0215] Background noise is caused by stray light in the atmosphere. The data measured by lidar contains not only signals such as aerosols and clouds, but also sky background noise signals.

[0216] According to the vertical distribution characteristics of the atmosphere, the high-altitude atmosphere is generally clean and contains almost no aerosols or clouds. Therefore, the high-altitude atmospheric echo signal is generally used to subtract background noise. The average value of the 1-2km high-altitude signal is selected as the background noise, and the real atmospheric echo signal is subtracted from the original echo. For example, if the range of the laser radar is 60km, the average value of the signal within any 2km between 50-60km can be selected as the background signal for subtraction. The calculation method is shown in the following formula:

[0217] P""(R,λ)=P(R,λ)-P bg (R,λ)

[0218] Among them, P(R,λ) is the original echo signal; P″(R,λ) is the signal after background noise subtraction, P bg (R,λ) is the signal of background noise.

[0219] Step S904: data smoothing processing.

[0220] Since the atmospheric state changes at any time and is affected by the random errors of the detection equipment, the original signal output by the lidar acquisition card has jitter. In this case, the echo signal can be smoothed to reduce signal jitter.

[0221] Data smoothing is to average the echo signals at several adjacent distance points, and the distance corresponding to the average echo value is also the average of these distances. The number of points for calculating the average value depends on the actual signal quality, such as 3-point smoothing, 5-point smoothing, etc. Data smoothing also changes the distance resolution of the echo signal. For example, if the distance resolution of a group of aerosol lidar echo signals is △r, after N-point smoothing, its distance resolution is N*△r.

[0222] Step S905, cloud and aerosol classification.

[0223] Clouds and aerosols have different optical properties. Choosing different aerosol lidar comparisons during data processing will have a significant impact on the inversion results. Therefore, cloud and aerosol data need to be distinguished.

[0224] Since the backscattering coefficient of clouds is much larger than that of aerosols, when the laser encounters clouds during transmission in the atmosphere, the range correction signal (RCS) will increase rapidly. The slope of the RCS curve is calculated, and the cloud layer or aerosol layer can be determined based on the positive and negative changes in the slope value.

[0225] If, starting from a certain height point, the slope of the RCS curve is continuously negative, then changes from negative to positive, and then continuously positive again, then this point is judged as the boundary between layers. After the entire curve is divided into layers according to the method, the interval range of each layer can be obtained, and then each layer is scored based on several parameters such as the maximum slope, minimum slope, and aspect ratio of each layer. If the score exceeds a certain threshold, it can be judged as a cloud layer, otherwise it is an aerosol layer or uncertain.

[0226] The data with clouds identified are marked and removed in the subsequent inversion of aerosol extinction coefficient and backscattering coefficient.

[0227] Step S906: data quality verification.

[0228] Taking into account the RCS curve score, the signal-to-noise ratio of the original echo signal, and the consistency of the fitting with the atmospheric molecule signal, the quality of the original echo data is checked and the corresponding quality control identification code and quality control type code are given.

[0229] Among them, the signal-to-noise ratio is the ratio of signal to noise, and the calculation formula is:

[0230]

[0231] As mentioned above, P bg (R,λ) is the background noise signal, and P(R,λ) is the original echo signal.

[0232] Among them, the consistency of fitting with the atmospheric molecular signal means that above the boundary layer, where the aerosol content in the atmosphere is almost zero, the echo signal can be assumed to be only related to the density of the molecules. Therefore, the molecular backscatter signal calculated according to the lidar equation and the curve after the real echo minus the background and distance square correction should be basically consistent after taking the log, that is, they are consistent.

[0233] The embodiment of the present invention also provides a device for fusing sea, land, air and space multi-source meteorological detection data. Fig.10 As shown, the device 1000 includes:

[0234] The first quality control module 1001 is configured to perform quality control on the satellite observation data, convert the format of the data, and then feed the data to the data receiving end;

[0235] The second quality control module 1002 is configured to convert the airborne platform coordinate system of the airborne meteorological radar reflectivity factor, speed, and spectrum width into a geodetic coordinate system, perform quality control on the data collected by the airborne platform and the sounding data converted into the geodetic coordinate system, perform format conversion, and then feed the data to the data receiving end;

[0236] The third quality control module 1003 is configured to perform quality control and format conversion on the data detected by the detection equipment and then feed it to the data receiving end, wherein the detection equipment at least includes one of a ground automatic meteorological observation station, an X-band weather radar, an S-band weather radar, a C-band weather radar, a P-band wind profiler radar, an L-band wind profiler radar, a laser wind measurement radar, a laser fog measurement radar, a millimeter wave cloud radar, and a millimeter wave wind measurement radar;

[0237] The fourth quality control module 1004 is configured to perform quality control and format conversion on the meteorological sensor data and then feed it to the data receiving end. The meteorological sensor data includes data collected by meteorological sensors installed on automatic weather stations, buoys and / or boats.

[0238] It should be noted that the various device structures described in this embodiment belong to the same technical concept as the previously described method, and achieve the same technical effect through the same principle, which will not be repeated here.

[0239] An embodiment of the present invention further provides a readable storage medium, which stores one or more programs. The one or more programs can be executed by one or more processors to implement the method described in any of the above embodiments.

[0240] In addition, although exemplary embodiments have been described herein, the scope includes any and all embodiments based on the present invention with equivalent elements, modifications, omissions, combinations (e.g., various embodiments intersecting schemes), adaptations or changes. The elements in the claims will be interpreted broadly based on the language adopted in the claims, and are not limited to the examples described in this specification or during the implementation of this application, and the examples will be interpreted as non-exclusive. Therefore, this specification and examples are intended to be considered as examples only, and the true scope and spirit are indicated by the following claims and the full scope of their equivalents.

[0241] The above description is intended to be illustrative rather than restrictive. For example, the above examples (or one or more of them) can be used in combination with each other. For example, those of ordinary skill in the art can use other embodiments when reading the above description. In addition, in the above-mentioned specific embodiments, various features can be grouped together to simplify the present invention. This should not be interpreted as a feature of an invention that is not claimed for protection being necessary for any claim. On the contrary, the subject matter of the present invention may be less than all the features of the embodiments of a specific invention. Thus, the following claims are incorporated into the specific embodiments as examples or embodiments, wherein each claim is independently used as a separate embodiment, and it is considered that these embodiments can be combined with each other in various combinations or arrangements. The scope of the present invention should be determined with reference to the full scope of the equivalent forms of the attached claims and these claims.

Claims

1. A method for fusing sea, land, air and space multi-source meteorological detection data, characterized in that: The method comprises: After quality control of satellite observation data, the data is converted into a new format and then fed to the data receiving end; The airborne platform coordinate system of the airborne meteorological radar reflectivity factor, speed, and spectrum width is converted into the geodetic coordinate system, and the data collected by the airborne platform and the sounding data converted into the geodetic coordinate system are quality controlled and format converted before being fed to the data receiving end; The data detected by the detection equipment is quality-controlled and format-converted and then fed to the data receiving end, wherein the detection equipment includes at least one of a ground automatic meteorological observation station, an X-band weather radar, an S-band weather radar, a C-band weather radar, a P-band wind profiler radar, an L-band wind profiler radar, a laser wind measurement radar, a laser fog measurement radar, a millimeter wave cloud radar, and a millimeter wave wind measurement radar; The meteorological sensor data is quality controlled and format converted before being fed to a data receiving end. The meteorological sensor data includes data collected by meteorological sensors installed on automatic weather stations, buoys and / or boats.

2. The method according to claim 1, characterized in that The satellite observation data is quality controlled by the following methods: Terrain verification: Eliminate satellite observation data with underlying surfaces of land, sea ice, snow, or mixed types; Extreme value test: Eliminate satellite observation data with brightness temperature less than 70K or greater than 320K; Threshold test: satellite observation data with a brightness temperature standard deviation greater than 6K are eliminated, and satellite observation data with a brightness temperature difference of observed brightness temperature minus simulated brightness temperature greater than 3.5K are eliminated.

3. The method according to claim 1, characterized in that The airborne platform coordinate system of the airborne meteorological radar reflectivity factor, speed, and spectrum width is converted to the geodetic coordinate system by the following method: Acquire attitude information of the airborne platform collected by an attitude sensor in real time, wherein the attitude information includes pitch angle, yaw angle and roll angle; In the event of jitter or attitude change during flight, the attitude change is determined based on the attitude information acquired in real time, and the coordinate system of the airborne platform is adjusted to keep the antenna stationary relative to the earth; Acquiring in real time the operating parameters of the aircraft, the operating parameters including position, velocity and acceleration, the operating parameters being measured relative to a geodetic coordinate system; According to the aircraft's operating parameters and the airborne platform's attitude information, the coordinates of the airborne platform are converted into the geodetic coordinate system.

4. The method according to claim 1, characterized in that: Quality control of data collected by the airborne platform and sounding data converted to the geodetic coordinate system includes: The data collected by the airborne platform and the sounding data converted to the geodetic coordinate system are subjected to ground clutter suppression, velocity deambiguation, data consistency assessment, noise and interference removal, clutter and false echo filtering, atmospheric attenuation correction, and error correction and calibration steps; The steps of ground clutter suppression include: Use space-time two-dimensional adaptive processing and / or digital signal processing techniques to filter out or suppress fixed target clutter and slow target clutter; The steps of deblurring velocity include: Design staggered dual waveforms A and B, the maximum unambiguous speeds of waveforms A and B are different and mutually prime, and when the fuzzy speeds of the same target speed on wave A and wave B are the same, the target speed is determined to be the true target speed; Or in the target tracking process, the actual displacement of the target is determined based on the multi-frame correlation of the target, and matched with the displacement under multiple fuzzy speeds, and the fuzzy speed with the highest matching degree is taken as the real target speed; The steps of data consistency assessment include: Compare the radar echo strength at different altitudes and time points to ensure the consistency of data in time and space; The steps of noise and interference removal include: Use filters and / or smoothing algorithms to remove the effects of electromagnetic interference and random noise; The steps of clutter and false echo filtering include: Remove clutter and false echoes caused by ground objects, birds, and insects contained in radar data; The steps of atmospheric attenuation correction include: Use atmospheric attenuation models or measured data to correct radar data for atmospheric attenuation; The steps of error correction and calibration include: Correct the systematic errors of the radar equipment; and calibrate the radar data to ensure that the radar data at different times and locations have a unified scale and standard; the systematic errors of the radar equipment include antenna pointing error and distance error.

5. The method according to claim 1, characterized in that The meteorological sensor data is quality controlled by the following methods: Data collection verification, including the use of verification mechanisms during data collection, including sensor calibration, hardware equipment status monitoring, and data collection software stability testing; After data collection, data preprocessing is performed, including denoising, smoothing, standardization and normalization operations; Data integrity checks, including determining whether data records are complete, timestamps are accurate, and whether data is missing or abnormal; Data outlier identification, including the use of statistical methods, comparative analysis methods or expert experience methods to identify outliers that do not conform to conventional meteorological data; Data consistency analysis, including comparing data from different sensors or different observation sites to check whether the data are consistent; Correct and amend data when outliers or inconsistent data are identified, where correction methods include interpolation, smoothing or substitution using historical data, data from neighboring sites or algorithm models; correction methods include hardware replacement, software upgrade or re-observation for sensor failure or operational error; Data storage management, including storing data in preset data storage formats, storage media, backup strategies and security measures; Data quality assessment includes determining the overall quality level of the data based on the completeness, accuracy, consistency and availability of the data.

6. The method according to claim 1, characterized in that The data detected by X-band weather radar, S-band weather radar and C-band weather radar are quality controlled by the following methods: Ground clutter suppression processing includes one or more of digital signal, spatial domain filtering, MTI filter, adaptive frequency domain filter and clutter reduction decision processing steps. When multiple processing steps are used, the results obtained by the multiple processing steps are compared to select the best processing result. The digital signal processing steps include: processing the radar echo signal based on median filtering, mean filtering, and adaptive filtering to filter out or suppress fixed target clutter and slow target clutter and retain the moving target echo; The spatial filtering processing step includes: performing spatial filtering using echo signals received by multiple antennas to suppress ground clutter from a specific direction, wherein the ground clutter in the specific direction includes echoes from the sky and the ground; The MTI filter processing steps include: using the MTI filter to perform a subtraction operation on pulses in adjacent repetition periods to eliminate fixed target echoes and slow-moving clutter, thereby retaining moving target echoes; The adaptive frequency domain filter processing steps include: using the adaptive frequency domain filter to adaptively determine the filtering position by analyzing the shape of the spectrum, and restoring the filtered weather signal by a Gaussian fitting method; The clutter reduction decision processing steps include: combining the adaptive filter and CMD algorithm to improve the data quality of the weather radar; When the target speed observed by the radar exceeds the maximum unambiguous speed of the radar, de-ambiguation processing is performed, wherein the de-ambiguation processing includes one or more combinations of Doppler frequency shift compensation, phase encoding, and repetition frequency diversity; The Doppler frequency shift compensation includes: measuring the Doppler frequency shift of the target and calculating the speed of the target according to the radar system parameters, thereby eliminating the error caused by speed ambiguity; The phase coding includes: adding a set phase coding sequence to the transmission signal so that the radar can distinguish different speed echoes and solve the speed ambiguity problem; The repetition frequency diversity includes: using a plurality of different repetition frequencies so that the radar can simultaneously receive a plurality of echo signals of different speeds, thereby eliminating speed ambiguity; In the case of range folding when the radar uses a pulse repetition frequency higher than the set value, random phase encoding, pulse repetition frequency adjustment, scanning at an elevation angle higher than the current elevation angle, or changing the detection location; The adjusting of the pulse repetition frequency includes changing the maximum detection range of the radar by changing the pulse repetition frequency, thereby eliminating range folding in the area of ​​interest.

7. The method according to claim 1, characterized in that The data detected by P-band wind profiler radar, L-band wind profiler radar and millimeter wave radar are quality controlled by ground clutter suppression, median filtering or smoothing filtering. The ground clutter suppression includes: designing an IIR notch filter according to the clutter map, and then performing linear interpolation between two edge points of the notch; The median filter includes: determining a neighborhood with a certain center point to be processed as the center; sorting the values ​​of each data in the neighborhood, taking the middle value as the new value of the center point data, and using the neighborhood as a window; after the window moves according to time and height, using the median filter to smooth the data; The smoothing filter includes: assuming that the observed object is Y, the independent variable corresponding to Y is X, Δx is used as the spacing of the independent variable X, the independent variables with equal spacing distribution are xi (i=1, 2, 3, ..., n), and according to the value of the observed dependent variable Y, the observed data sequence is obtained as yi (i=1, 2, 3, ..., n), and the noise components in the observed data sequence are filtered out by sliding smoothing, wherein the use of sliding smoothing to filter out the noise components in the observed data sequence includes: taking the data of the i-th point and several points nearby, determining a fitted straight line equation according to the least squares principle, and calculating the dependent variable of the i-th point based on the straight line equation as the data value after sliding smoothing.

8. The method according to claim 1, characterized in that The data detected by the laser wind radar is quality controlled by the following methods: Processing of noise and missing values: Select an M*N sliding window, divide the data in the window into P intervals with equal size, △d is the interval, and the wind speed value at the center of the window is V ij , the P+1th interval is represented by V ij ±△d / 2; Count the distribution frequency of wind speed values ​​in different intervals in the M*N window (X1, X2, X3....X P ) and X′, let X MAX It is X1, X2, X3....X P and X′, take VP as X MAX The middle value of the corresponding interval; Let the new value of the window center be V ij , noise removal and missing measurement filling are achieved through the following formula: Where K1 is the threshold for determining whether to remove noise, and K2 is the threshold for determining whether to fill in missing measurements. ij When it is not 0, it means that this is a valid echo point. When X′<=threshold K1, the point is removed; when X′>threshold K1, the corresponding point is retained; when the original value V in the window ij When it is equal to 0, it is an invalid point. MAX >=threshold K2, fill in the gap and assign the center point VP of the interval corresponding to XMAX to V ij If X MAX <threshold K2, no filling is performed; The data detected by the laser fog radar is quality controlled by the following methods: Dead time correction, calculated as: Where N is the true count rate, M is the measured count rate, and t is the time resolution; Geometric overlap factor correction, calculated as: Where P(R, λ) is the original echo signal; P′(R, λ) is the signal corrected by the geometric overlap factor; R is the height from the radar; λ is the laser emission wavelength, and O(R) is the geometric overlap factor; Background noise subtraction, the calculation formula is: P″(R,λ)=P(R,λ)-P bg (R,λ) Among them, P(R,λ) is the original echo signal; P′′(R,λ) is the signal after background noise subtraction, P bg (R,λ) is the signal of background noise; Smoothing the echo signal; Cloud and aerosol classification: Calculate the slope of the corrected signal curve and determine the cloud layer or aerosol layer based on the positive and negative changes in the slope value; Data quality verification: The quality of the original echo data is verified according to the RCS curve score, the signal-to-noise ratio of the original echo signal, and the consistency with the atmospheric molecular signal. The corresponding quality control identification code and quality control type code are given. The consistency with the atmospheric molecular signal means that above the boundary layer, in the area where the aerosol content in the atmosphere is lower than the set threshold, when the echo signal is set to be only related to the density of the molecules, the molecular backscattering signal calculated according to the lidar equation and the curve after the real echo is subtracted from the background and corrected for the square of the distance are consistent with the atmospheric molecular signal within the set error after taking the log, and then it is determined to be consistent.

9. A device for fusion of multi-source meteorological detection data from land, sea, air and space, characterized in that: The device comprises: A first quality control module is configured to perform quality control on the satellite observation data, convert the format of the data, and then feed the data to the data receiving end; The second quality control module is configured to convert the airborne platform coordinate system of the airborne meteorological radar reflectivity factor, speed, and spectrum width into a geodetic coordinate system, perform quality control on the data collected by the airborne platform converted into the geodetic coordinate system and the sounding data, perform format conversion, and then feed them to the data receiving end; A third quality control module is configured to perform quality control and format conversion on the data detected by the detection equipment and then feed it to the data receiving end, wherein the detection equipment includes at least one of a ground automatic meteorological observation station, an X-band weather radar, an S-band weather radar, a C-band weather radar, a P-band wind profiler radar, an L-band wind profiler radar, a laser wind measurement radar, a laser fog measurement radar, a millimeter wave cloud radar, and a millimeter wave wind measurement radar; The fourth quality control module is configured to perform quality control and format conversion on the meteorological sensor data and then feed it to the data receiving end. The meteorological sensor data includes data collected by meteorological sensors installed on automatic weather stations, buoys and / or boats. 10 . A non-transitory computer-readable storage medium storing instructions, which, when executed by a processor, perform the method according to claim 1 .

Images