A GNSS-based atmospheric humidity inversion method and system

Abstract

The present invention relates to the technical field of atmospheric humidity retrieval, and relates to a method and system for retrieving atmospheric humidity based on GNSS. The method includes obtaining first information, second information, and a trained first neural network. The first information includes GNSS signals collected at low-altitude stations, and the second information includes surface humidity data around the low-altitude stations; processing the first information by using a polarization filtering algorithm and a beamforming algorithm to obtain processed first information; training a second neural network according to the processed first information and the second information to obtain a trained second neural network; inputting the first information and correction parameters for correcting wet delay into the trained first neural network to obtain an output result; performing water vapor retrieval based on the output result to construct a water vapor field. The present invention effectively improves the accuracy of atmospheric humidity retrieval in low-altitude regions.

Description

Technical Field

[0001] The present invention relates to the technical field of atmospheric humidity retrieval, and in particular, to a method and system for retrieving atmospheric humidity based on GNSS. Background Art

[0002] In the study of the Earth's atmospheric environment, altitude change has a very significant impact on the accuracy of water vapor retrieval. From the perspective of atmospheric physics, the change in altitude is directly related to various physical properties of the atmosphere, and these properties are closely related to the distribution and state of water vapor, thus deeply affecting the accuracy of water vapor retrieval. In high-altitude areas, due to the relatively thin atmosphere, the water vapor content is relatively low, and the atmospheric environment is relatively stable, less affected by factors such as complex terrain and human activities. This enables the GNSS retrieval to maintain relatively stable accuracy in high-altitude areas. However, in low-altitude areas, due to the influence of complex terrain and dense buildings, the existing high-altitude retrieval methods cannot accurately retrieve the atmospheric humidity in low-altitude areas. Summary of the Invention

[0003] The purpose of the present invention is to provide a method and system for retrieving atmospheric humidity based on GNSS to improve the above problems.

[0004] To achieve the above purpose, the embodiments of the present application provide the following technical solutions:

[0005] On the one hand, the embodiments of the present application provide a method for retrieving atmospheric humidity based on GNSS, and the method includes:

[0006] Obtain first information, second information, and a trained first neural network, where the first information includes GNSS signals collected at low-altitude stations, the low-altitude stations include stations corresponding to elevations less than 500 m, the first information is collected through an antenna array set at the low-altitude stations, the second information includes surface humidity data around the low-altitude stations, and the trained first neural network is used to predict the atmospheric weighted average temperature in the low-altitude area;

[0007] Process the first information by using a polarization filtering algorithm and a beamforming algorithm to obtain processed first information;

[0008] Train a second neural network according to the processed first information and the second information to obtain a trained second neural network, and the trained second neural network is used to correct the wet delay;

[0009] Input the first information and the correction parameters for correcting the wet delay into the trained first neural network to obtain an output result, and the output result includes a predicted value of the atmospheric weighted average temperature;

[0010] Perform water vapor inversion based on the output result to construct a water vapor field, which is used to characterize the atmospheric humidity distribution corresponding to the low-altitude area.

[0011] In a second aspect, an embodiment of the present application provides a GNSS-based atmospheric humidity inversion system, which includes:

[0012] An acquisition module, configured to acquire first information, second information, and a trained first neural network. The first information includes GNSS signals collected by low-altitude stations, and the low-altitude stations include stations corresponding to elevations less than 500 m. The first information is collected through an antenna array set at the low-altitude stations. The second information includes surface humidity data around the low-altitude stations. The trained first neural network is used to predict the atmospheric weighted average temperature in the low-altitude area;

[0013] A first processing module, configured to process the first information by using a polarization filtering algorithm and a beamforming algorithm to obtain processed first information;

[0014] A second processing module, configured to train a second neural network according to the processed first information and the second information to obtain a trained second neural network, and the trained second neural network is used to correct the wet delay;

[0015] A third processing module, configured to input the first information and correction parameters for correcting the wet delay into the trained first neural network to obtain an output result, and the output result includes a predicted value of the atmospheric weighted average temperature;

[0016] A fourth processing module, configured to perform water vapor inversion based on the output result to construct a water vapor field, which is used to characterize the atmospheric humidity distribution corresponding to the low-altitude area.

[0017] In a third aspect, an embodiment of the present application provides a GNSS-based atmospheric humidity inversion device, which includes a memory and a processor. The memory is used to store a computer program; the processor is configured to implement the steps of the above-mentioned GNSS-based atmospheric humidity inversion method when executing the computer program.

[0018] In a fourth aspect, an embodiment of the present application provides a readable storage medium, on which a computer program is stored, and the computer program implements the steps of the above-mentioned GNSS-based atmospheric humidity inversion method when being executed by a processor.

[0019] The beneficial effects of the present invention are:

[0020] The present invention processes the first information by using polarization filtering and beamforming algorithms to suppress multipath noise caused by strong reflectors such as buildings and water areas, and then trains the second neural network in combination with the second information. The correction parameter of the wet delay is determined according to the output result of the second neural network, effectively solving the problem of low signal-to-noise ratio caused by low-altitude multipath effects. At the same time, the reflected signals that were traditionally regarded as interference are converted into inversion inputs, improving data utilization. Then, the correction parameter of the wet delay and the first information are input into the trained first neural network to obtain an output result, and a water vapor field is constructed according to the output result to realize the inversion of atmospheric humidity, improving the inversion accuracy in low-altitude areas.

[0021] Other features and advantages of the present invention will be described in the subsequent specification, and, in part, will be obvious from the specification, or can be understood by implementing the embodiments of the present invention. The objectives and other advantages of the present invention can be achieved and obtained by the structures specifically pointed out in the written specification and the drawings. Brief Description of the Drawings

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention, and thus should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0023] Figure 1 It is a schematic flow chart of the GNSS-based atmospheric humidity inversion method described in the embodiments of the present invention.

[0024] Figure 2 It is a schematic structural diagram of the GNSS-based atmospheric humidity inversion system described in the embodiments of the present invention.

[0025] Figure 3 It is a schematic structural diagram of the GNSS-based atmospheric humidity inversion device described in the embodiments of the present invention.

[0026] Labels in the figure: 800, GNSS-based atmospheric humidity inversion device; 801, processor; 802, memory; 803, multimedia component; 804, I / O interface; 805, communication component; 901, acquisition module; 902, first processing module; 903, second processing module; 904, third processing module; 905, fourth processing module. Detailed Embodiments

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. The components of the embodiments of the present invention usually described and illustrated in the accompanying drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0028] It should be noted that similar reference numerals and letters denote similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used for descriptive distinction and cannot be construed as indicating or implying relative importance. Embodiment 1:

[0029] This embodiment provides a method for retrieving atmospheric humidity based on GNSS. It can be understood that in this embodiment, a scenario can be set up, for example, a scenario for detecting the atmospheric humidity corresponding to an area with an altitude less than 500m.

[0030] See Figure 1 , the figure shows that this method includes steps S1, S2, S3, S4, and S5, which specifically include:

[0031] Step S1: Obtain first information, second information, and a trained first neural network. The first information includes GNSS signals collected by low-altitude stations, and the low-altitude stations include stations corresponding to an elevation less than 500m. The first information is collected through an antenna array set up at the low-altitude stations. The second information includes surface humidity data around the low-altitude stations. The trained first neural network is used to predict the weighted average atmospheric temperature in the low-altitude area.

[0032] In this step, a multi-band antenna array is deployed at low-altitude stations (altitude less than 500m) to capture L1 / L2 / L5 band GNSS signals in real time, which are the basic data for subsequent processing and analysis.

[0033] In step S1, it further includes steps S11, S12, S13, S14, and S15, which specifically include:

[0034] Step S11: Obtain the historical data of the high-altitude area, where the historical data of the high-altitude area includes meteorological observation data, GNSS observation data, and time series data;

[0035] Step S12: Construct a feature matrix corresponding to the high-altitude area based on the historical data of the high-altitude area to obtain a second feature matrix;

[0036] In this step, the second feature matrix includes wet delay, atmospheric weighted average temperature, and altitude.

[0037] Step S13: Use the second feature matrix corresponding to the high-altitude area to pre-train the first neural network to obtain the pre-trained first neural network;

[0038] In this step, the first neural network uses an LSTM network (128 hidden units) to process the historical data of the high-altitude site, and the input features are the features included in the second feature matrix.

[0039] Step S14: Construct a feature matrix based on the historical data of the low-altitude area and the correction parameters for correcting the wet delay to obtain a third feature matrix;

[0040] Step S15: Train the pre-trained first neural network according to the third feature matrix to obtain the trained first neural network.

[0041] In this step, when training the pre-trained first neural network, freeze the weights of the bottom layer of the LSTM and only train the top fully connected layer. Train the pre-trained first neural network with a small sample of low-altitude historical data to obtain the trained first neural network.

[0042] In this embodiment, since the historical data samples of the low-altitude area are few, first use the historical data of the high-altitude area for pre-training, and then use a small sample of low-altitude historical data to fine-tune the pre-trained first neural network, so as to obtain the trained first neural network, which not only solves the problem of model generalization caused by insufficient low-altitude samples but also improves the accuracy of predicting the atmospheric weighted average temperature in the low-altitude area.

[0043] Step S2: Process the first information using the polarization filtering algorithm and the beamforming algorithm to obtain the processed first information;

[0044] In the step S2, there are also steps S21, S22, S23, S24, and S25, which specifically include:

[0045] Step S21: Preprocess the first information to obtain the first information after removing noise;

[0046] Step S22: Determine the polarization mode corresponding to the GNSS signal according to the first information;

[0047] Step S23: Design a polarization filter according to the polarization mode corresponding to the GNSS signal;

[0048] It can be understood that the signal transmitted by the GNSS satellite has a specific polarization mode. Due to the reflection of the multipath reflected signal by the reflector, its polarization characteristics will change, and there are differences from the polarization characteristics of the direct signal. Therefore, by determining the polarization mode of the collected GNSS signal, the polarization filter can be designed, so as to design a suitable polarization filter, thereby suppressing the multipath reflected signal with different polarization characteristics from the direct signal to the greatest extent.

[0049] Step S24: Process the first information after removing noise according to the polarization filter to obtain the third information;

[0050] In this step, the GNSS direct signal is usually right-handed circular polarization. Design a polarization filter to mainly pass the right-handed circular polarization signal and have a high attenuation for the left-handed circular polarization signal, so as to effectively suppress the possible left-handed circular polarization component in the multipath reflected signal.

[0051] Step S25: Process the third information by using a beamforming algorithm to obtain the processed first information.

[0052] In step S25, it further includes step S251, step S252, step S253, step S254, step S255 and step S256, which specifically include:

[0053] Step S251: Process the third information by using the multiple signal classification algorithm to obtain the arrival angle of the signal;

[0054] In this step, based on the orthogonality of the signal subspace and the noise subspace. First, sample the polarization-filtered signal received by the antenna array to obtain a signal matrix. Then perform eigenvalue decomposition on the signal matrix to decompose it into a signal subspace and a noise subspace. Since the signal subspace is orthogonal to the noise subspace, by constructing a spatial spectrum function and searching for the peak position of the spatial spectrum, the angle corresponding to this peak is the arrival angle of the signal.

[0055] Step S252: Establish an array flow model according to the geometric structure of the antenna array;

[0056] In this step, the array flow model describes the response characteristics of the signal when it arrives at each element of the antenna array. In a specific implementation, the specific representation of the array manifold vector is:

[0057]

[0058] In the above formula, represents the array manifold vector, represents the arrival angle of the signal, represents the amplitude, represents the phase difference.

[0059] Step S253: Determine the time delay of the signal arriving at different array elements according to the arrival angle of the signal and the array flow model;

[0060] In a specific embodiment, when the antenna array is a uniform linear array, taking the first array element as the reference array element, after the arrival angle of the known signal, the wave path difference of each array element relative to the reference array element can be calculated, and then the time delay of each array element relative to the reference array element can be calculated according to the wave path difference of each array element relative to the reference array element.

[0061] Step S254: Determine the phase difference between each array element according to the time delay of the signal arriving at different array elements, and obtain the phase difference information;

[0062] In this step, since the phase of the signal is proportional to the time delay, the phase difference between each array element can be calculated through the time delay.

[0063] Step S255: Process the phase difference information by using the minimum variance distortionless response algorithm to obtain the weight coefficient corresponding to each array element;

[0064] In step S255, it further includes steps S2551, S2552, S2553, S2554 and S2555, which specifically include:

[0065] Step S2551: Sample the third information to obtain the discrete samples of the signal;

[0066] Step S2552: Calculate the discrete samples of the signal by using the root mean square algorithm to obtain the amplitude value of the signal;

[0067] Step S2553: Construct an array manifold vector according to the amplitude value of the signal and the phase difference information, and the array manifold vector is used to describe the amplitude and phase relationship of the signals received by each array element when the signal arrives;

[0068] Step S2554: Calculate the covariance matrix corresponding to the third information;

[0069] Step S2555: Calculate according to the array manifold vector and the covariance matrix to obtain the weight coefficient corresponding to each array element.

[0070] In this step, the specific calculation process of the weight coefficient is as follows:

[0071]

[0072] In the above formula, W represents the weight coefficient vector, is the inverse matrix of the covariance matrix, is the array manifold vector, represents the arrival angle of the signal, is the conjugate transpose of the array manifold vector. By solving the above formula, the weight coefficient corresponding to each array element can be obtained. The present invention can minimize the power of the array output on the premise of ensuring that the desired signal passes through without distortion, so as to suppress interference and noise.

[0073] Step S256: Weight the third information received by the antenna array according to the weight coefficient to obtain the processed first information.

[0074] In this embodiment, the calculated weight coefficient is applied to the signal after polarization filtering received by the antenna array. The signals received by each array element are respectively multiplied by the corresponding weight coefficients. By adjusting the amplitude and correcting the phase of the signals of each array element, the signals from the desired direction are in-phase superimposed at the output end of the antenna array, so as to enhance the signal strength in this direction; while the signals from other directions (such as the multi-path reflection signal direction) are cancelled or weakened due to inconsistent phases. In this way, after the weighting process, the residual multi-path reflection signals are further suppressed, and the quality and signal-to-noise ratio of the GNSS signal are improved.

[0075] Step S3: Train the second neural network according to the processed first information and the second information to obtain the trained second neural network, and the trained second neural network is used to correct the wet delay;

[0076] In the step S3, there are also step S31, step S32, step S33, step S34 and step S35, which specifically include:

[0077] Step S31: Extract the signal-to-noise ratio in the processed first information to obtain the signal-to-noise ratio information;

[0078] Step S32: Perform wavelet decomposition on the signal-to-noise ratio information to obtain the decomposed signal information, and the decomposed signal information includes the high-frequency fluctuation part of the reflection signal;

[0079] In this step, wavelet decomposition (using db4 wavelet) is performed on the signal-to-noise ratio information to separate the low-frequency trend term (dominated by the direct signal) and the high-frequency fluctuation term (dominated by the reflection signal), and at the same time, only the reflection signal component is retained, with a frequency of 0.1 - 0.5 Hz.

[0080] Step S33: Perform time-frequency analysis on the decomposed signal information to determine the characteristic information corresponding to the decomposed signal information, where the characteristic information includes amplitude, phase delay, and time delay difference;

[0081] In this step, perform short-time Fourier transform on the decomposed signal information, extract the dominant frequency component of the reflected signal, and then calculate the corresponding amplitude, phase delay, and time delay difference based on the dominant frequency component of the reflected signal.

[0082] Step S34: Establish a first feature matrix according to the characteristic information and the second information, and train a second neural network according to the first feature matrix to obtain a trained second neural network;

[0083] In this step, align the characteristic information and the second information according to the time stamp and jointly form a first feature matrix to train the second neural network, establish a mapping function between the reflection path delay and the surface humidity, so as to obtain a reflection path-humidity mapping model, that is, a trained second neural network.

[0084] Step S35: Correct the wet delay according to the trained second neural network.

[0085] In step S35, it further includes steps S351, S352, S353, S354, and S355, which specifically include:

[0086] Step S351: Obtain real-time characteristic information;

[0087] Step S352: Send the real-time characteristic information to the trained second neural network to obtain a fourth piece of information, where the fourth piece of information includes a predicted value of the surface humidity of the reflection point;

[0088] Step S353: Calculate according to the fourth piece of information and the measured surface humidity value of the reflection point at the station to obtain a correction parameter for the wet delay;

[0089] In this step, the specific calculation formula for the correction parameter of the wet delay is:

[0090] ;

[0091] In the above formula, represents the correction parameter of the wet delay, represents the predicted value of the surface humidity of the reflection point, and represent model coefficients, which are used to adjust the value.

[0092] Step S355: Correct the wet delay according to the correction parameter of the wet delay.

[0093] In this step, the specific process of correcting the wet delay is as follows:

[0094] ;

[0095] In the above formula, is the corrected wet delay, is the wet delay before correction, represents the correction parameter of the wet delay, represents the elevation correction factor. For the low altitude area, the corresponding H is 1.02 - 1.05, compensating for the water vapor density gradient near the surface.

[0096] Step S4: Input the first information and the correction parameter for correcting the wet delay into the trained first neural network to obtain an output result, where the output result includes the predicted value of the atmospheric weighted average temperature;

[0097] Step S5: Perform water vapor inversion based on the output result to construct a water vapor field, and the water vapor field is used to characterize the atmospheric humidity distribution corresponding to the low altitude area.

[0098] In step S5, it further includes step S51, step S52, step S53 and step S54, which specifically include:

[0099] Step S51: Obtain the density information of water, the water vapor gas constant and the atmospheric refraction constant;

[0100] Step S52: Calculate according to the density information of water, the water vapor gas constant, the atmospheric refraction constant and the output result to obtain a conversion coefficient;

[0101] In this step, the specific calculation process of the conversion coefficient is:

[0102] ;

[0103] In the above formula, represents the conversion coefficient, represents the density of water, represents the water vapor gas constant, and both represent the atmospheric refraction constant. Among them, is related to the combined refraction effect of dry air and water vapor in the atmosphere, is related to the refraction effect of water vapor, represents the output result.

[0104] Step S53: Calculate according to the conversion coefficient and the corrected wet delay to obtain the precipitable water in the atmosphere;

[0105] In this step, the precipitable water vapor can be obtained by multiplying the conversion coefficient by the corrected wet delay.

[0106] Step S54: Use Kriging interpolation to process the precipitable water vapor corresponding to all stations in the low altitude area to construct a water vapor field.

[0107] In this step, Kriging interpolation is performed on the precipitable water vapor of all GNSS stations in the low altitude area to generate a water vapor distribution map with a resolution of 1 km, visually showing the distribution of water vapor in the area, and providing detailed data and visual information for meteorological analysis and decision-making.

[0108] It should be noted that the water vapor field needs to be further optimized. Embodiment 2:

[0109] As Figure 2 shown, this embodiment provides a GNSS-based atmospheric humidity inversion system, which includes an acquisition module 901, a first processing module 902, a second processing module 903, a third processing module 904, and a fourth processing module 905, specifically including:

[0110] The acquisition module 901 is used to acquire the first information, the second information, and the trained first neural network. The first information includes GNSS signals collected by low altitude stations, and the low altitude stations include stations corresponding to elevations less than 500 m. The first information is collected through an antenna array set at the low altitude stations. The second information includes surface humidity data around the low altitude stations. The trained first neural network is used to predict the atmospheric weighted average temperature in the low altitude area;

[0111] The first processing module 902 is used to process the first information by using a polarization filtering algorithm and a beamforming algorithm to obtain the processed first information;

[0112] The second processing module 903 is used to train the second neural network according to the processed first information and the second information to obtain the trained second neural network. The trained second neural network is used to correct the wet delay;

[0113] The third processing module 904 is used to input the first information and the correction parameters for correcting the wet delay into the trained first neural network to obtain an output result, and the output result includes a predicted value of the atmospheric weighted average temperature;

[0114] The fourth processing module 905 is used to perform water vapor inversion according to the output result to construct a water vapor field, and the water vapor field is used to characterize the atmospheric humidity distribution corresponding to the low altitude area.

[0115] In a specific embodiment of the present disclosure, the first processing module further includes a first processing unit, a second processing unit, a third processing unit, a fourth processing unit, and a fifth processing unit, which specifically include:

[0116] The first processing unit is configured to preprocess the first information to obtain the first information after removing noise;

[0117] The second processing unit is configured to determine the polarization mode corresponding to the GNSS signal according to the first information;

[0118] The third processing unit is configured to design a polarization filter according to the polarization mode corresponding to the GNSS signal;

[0119] The fourth processing unit is configured to process the first information after removing noise according to the polarization filter to obtain the third information;

[0120] The fifth processing unit is configured to process the third information by using a beamforming algorithm to obtain the processed first information.

[0121] In a specific embodiment of the present disclosure, the fifth processing unit further includes a sixth processing unit, a seventh processing unit, an eighth processing unit, a ninth processing unit, a tenth processing unit, and an eleventh processing unit, which specifically include:

[0122] The sixth processing unit is configured to process the third information by using a multiple signal classification algorithm to obtain the arrival angle of the signal;

[0123] The seventh processing unit is configured to establish an array flow model according to the geometric structure of the antenna array;

[0124] The eighth processing unit is configured to determine the time delay of the signal arriving at different array elements according to the arrival angle of the signal and the array flow model;

[0125] The ninth processing unit is configured to determine the phase difference between each array element according to the time delay of the signal arriving at different array elements to obtain phase difference information;

[0126] The tenth processing unit is configured to process the phase difference information by using a minimum variance distortionless response algorithm to obtain the weight coefficient corresponding to each array element;

[0127] The eleventh processing unit is configured to weight the third information received by the antenna array according to the weight coefficient to obtain the processed first information.

[0128] It should be noted that for the system in the above embodiments, the specific manner in which each module performs operations has been described in detail in the embodiments related to the method, and will not be elaborated here. Example 3:

[0129] Corresponding to the above method embodiments, in this embodiment, a GNSS-based atmospheric humidity inversion device is also provided. A GNSS-based atmospheric humidity inversion device described below can be correspondingly referred to the GNSS-based atmospheric humidity inversion method described above.

[0130] Figure 3 It is a block diagram of a GNSS-based atmospheric humidity inversion device 800 shown according to an exemplary embodiment. As Figure 3 shown, the GNSS-based atmospheric humidity inversion device 800 may include: a processor 801, a memory 802. The GNSS-based atmospheric humidity inversion device 800 may further include one or more of a multimedia component 803, an I / O interface 804, and a communication component 805.

[0131] Among them, the processor 801 is used to control the overall operation of the GNSS-based atmospheric humidity inversion device 800 to complete all or part of the steps in the above GNSS-based atmospheric humidity inversion method. The memory 802 is used to store various types of data to support the operation of the GNSS-based atmospheric humidity inversion device 800. These data may include, for example, instructions for any application or method operating on the GNSS-based atmospheric humidity inversion device 800, as well as application-related data, such as contact data, sent and received messages, pictures, audio, video, and so on. The memory 802 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, a magnetic disk, or an optical disc. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signals may be further stored in the memory 802 or sent through the communication component 805. The audio component further includes at least one speaker for outputting audio signals. The I / O interface 804 provides an interface between the processor 801 and other interface modules, and the other interface modules may be a keyboard, a mouse, buttons, etc. These buttons may be virtual buttons or physical buttons. The communication component 805 is used for wired or wireless communication between the GNSS-based atmospheric humidity inversion device 800 and other devices. Wireless communication, such as Wi-Fi, Bluetooth, near field communication (NFC), 2G, 3G, or 4G, or a combination of one or more of them. Accordingly, the communication component 805 may include: a Wi-Fi module, a Bluetooth module, and an NFC module.

[0132] In an exemplary embodiment, the GNSS-based atmospheric humidity inversion device 800 can be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components, and is used to execute the above-mentioned GNSS-based atmospheric humidity inversion method.

[0133] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided. When the program instructions are executed by a processor, the steps of the above-mentioned GNSS-based atmospheric humidity inversion method are implemented. For example, the computer-readable storage medium can be the above-mentioned memory 802 including program instructions, and the above program instructions can be executed by the processor 801 of the GNSS-based atmospheric humidity inversion device 800 to complete the above-mentioned GNSS-based atmospheric humidity inversion method. Embodiment 4:

[0134] Corresponding to the above method embodiment, a readable storage medium is also provided in this embodiment. A readable storage medium described below can be correspondingly referred to with a GNSS-based atmospheric humidity inversion method described above.

[0135] A readable storage medium has a computer program stored thereon. When the computer program is executed by a processor, the steps of the GNSS-based atmospheric humidity inversion method in the above method embodiment are implemented.

[0136] The readable storage medium can specifically be various readable storage media such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc that can store program codes.

[0137] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

[0138] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A GNSS-based atmospheric humidity inversion method, characterized in that: include: Acquire first information, second information, and a trained first neural network, wherein the first information includes GNSS signals collected by low-altitude sites, the low-altitude sites include sites corresponding to elevations less than 500 m, the first information is collected by antenna arrays provided at the low-altitude sites, the second information includes surface humidity data around the low-altitude sites, and the trained first neural network is used to predict the weighted average temperature of the atmosphere in the low-altitude area; Processing the first information using a polarization filtering algorithm and a beamforming algorithm to obtain processed first information; Training a second neural network according to the processed first information and the second information to obtain a trained second neural network, wherein the trained second neural network is used to correct wet delay; Inputting the first information and a correction parameter for correcting wet delay into the trained first neural network to obtain an output result, wherein the output result includes a predicted value of the atmospheric weighted average temperature; Water vapor inversion is performed according to the output result to construct a water vapor field, and the water vapor field is used to characterize the atmospheric humidity distribution corresponding to the low altitude area.

2. The GNSS-based atmospheric humidity inversion method according to claim 1, characterized in that: The first information is processed by using a polarization filtering algorithm and a beamforming algorithm to obtain the processed first information, including: Preprocessing the first information to obtain first information after noise is removed; Determine a polarization mode corresponding to the GNSS signal according to the first information; Designing a polarization filter according to the polarization mode corresponding to the GNSS signal; Processing the first information after the noise is removed according to the polarization filter to obtain third information; The third information is processed using a beamforming algorithm to obtain the processed first information.

3. The GNSS-based atmospheric humidity inversion method according to claim 2, characterized in that: Processing the third information by using a beamforming algorithm to obtain the processed first information includes: Processing the third information using a multiple signal classification algorithm to obtain an arrival angle of the signal; An array flow model is established according to the geometric structure of the antenna array; Determine the time delay of the signal arriving at different array elements according to the arrival angle of the signal and the array flow model; Determine the phase difference between the array elements according to the time delay of the signal reaching different array elements, and obtain phase difference information; The phase difference information is processed using the minimum variance distortion-free response algorithm to obtain the weight coefficient corresponding to each array element; The third information received by the antenna array is weighted according to the weight coefficient to obtain the processed first information.

4. The GNSS-based atmospheric humidity inversion method according to claim 3, characterized in that: The phase difference information is processed using the minimum variance distortion-free response algorithm to obtain the weight coefficient corresponding to each array element, including: Sampling the third information to obtain discrete samples of the signal; Calculating discrete samples of the signal using a root mean square algorithm to obtain an amplitude value of the signal; constructing an array manifold vector according to the amplitude value of the signal and the phase difference information, wherein the array manifold vector is used to describe the relationship between the amplitude and phase of the signal received by each array element when the signal arrives; Calculate the covariance matrix corresponding to the third information; Calculation is performed based on the array manifold vector and the covariance matrix to obtain a weight coefficient corresponding to each array element.

5. The GNSS-based atmospheric humidity inversion method according to claim 1, characterized in that: The second neural network is trained according to the processed first information and the second information to obtain a trained second neural network, including: Extracting a signal-to-noise ratio from the processed first information to obtain signal-to-noise ratio information; Performing wavelet decomposition on the signal-to-noise ratio information to obtain decomposed signal information, wherein the decomposed signal information includes a high-frequency fluctuation part of the reflected signal; Performing time-frequency analysis on the decomposed signal information to determine characteristic information corresponding to the decomposed signal information, wherein the characteristic information includes amplitude, phase delay, and time delay difference; Establishing a first feature matrix according to the feature information and the second information, and training a second neural network according to the first feature matrix to obtain a trained second neural network; The wet delay is corrected according to the trained second neural network.

6. The GNSS-based atmospheric humidity inversion method according to claim 1, characterized in that: Acquiring the trained first neural network includes: Acquire historical data of a high altitude area, wherein the historical data of the high altitude area includes meteorological observation data, GNSS observation data, and time series data; constructing a feature matrix corresponding to the high-altitude area according to the historical data of the high-altitude area to obtain a second feature matrix; Pre-training the first neural network using the second feature matrix corresponding to the high altitude area to obtain a pre-trained first neural network; A characteristic matrix is ​​constructed according to historical data of low altitude areas and correction parameters for correcting wet delay to obtain a third characteristic matrix; The pre-trained first neural network is trained according to the third characteristic matrix to obtain a trained first neural network.

7. The GNSS-based atmospheric humidity inversion method according to claim 1, characterized in that: Perform water vapor inversion according to the output results to construct a water vapor field, including: Obtain water density information, water vapor gas constant and atmospheric refraction constant; Calculating according to the water density information, the water vapor gas constant, the atmospheric refraction constant and the output result to obtain a conversion coefficient; Calculating according to the conversion coefficient and the corrected wet delay to obtain atmospheric precipitable water; Kriging interpolation is used to process the atmospheric precipitable water corresponding to all stations in the low-altitude area to construct the water vapor field.

8. A GNSS-based atmospheric humidity inversion system, characterized in that: include: an acquisition module, configured to acquire first information, second information, and a trained first neural network, wherein the first information includes GNSS signals collected by low-altitude sites, the low-altitude sites include sites corresponding to elevations less than 500 m, the first information is collected by antenna arrays provided at the low-altitude sites, the second information includes surface humidity data around the low-altitude sites, and the trained first neural network is used to predict the weighted average temperature of the atmosphere in the low-altitude area; A first processing module, configured to process the first information by using a polarization filtering algorithm and a beamforming algorithm to obtain processed first information; A second processing module is used to train a second neural network according to the processed first information and the second information to obtain a trained second neural network, wherein the trained second neural network is used to correct wet delay; a third processing module, configured to input the first information and a correction parameter for correcting wet delay into the trained first neural network to obtain an output result, wherein the output result includes a predicted value of the atmospheric weighted average temperature; The fourth processing module is used to perform water vapor inversion according to the output result to construct a water vapor field, and the water vapor field is used to characterize the atmospheric humidity distribution corresponding to the low altitude area.

9. The GNSS-based atmospheric humidity inversion system according to claim 8, characterized in that: The first processing module comprises: A first processing unit, configured to preprocess the first information to obtain first information after noise is removed; A second processing unit, configured to determine a polarization mode corresponding to the GNSS signal according to the first information; A third processing unit, configured to design a polarization filter according to a polarization mode corresponding to the GNSS signal; a fourth processing unit, configured to process the first information after the noise is removed according to the polarization filter to obtain third information; The fifth processing unit is used to process the third information by using a beamforming algorithm to obtain the processed first information.

10. The GNSS-based atmospheric humidity inversion system according to claim 9, characterized in that: The fifth processing unit comprises: a sixth processing unit, configured to process the third information using a multiple signal classification algorithm to obtain an arrival angle of the signal; a seventh processing unit, configured to establish an array flow model according to a geometric structure of the antenna array; an eighth processing unit, configured to determine a time delay of a signal arriving at different array elements according to an arrival angle of the signal and the array flow model; A ninth processing unit, configured to determine the phase difference between the array elements according to the time delay of the signal reaching different array elements, and obtain phase difference information; a tenth processing unit, configured to process the phase difference information using a minimum variance distortion-free response algorithm to obtain a weight coefficient corresponding to each array element; The eleventh processing unit is used to weight the third information received by the antenna array according to the weight coefficient to obtain the processed first information.

Images