A Method and System for Monitoring Sub-Sun-Level Drought Based on Gravity Satellites

Abstract

This application discloses a sub-Sun-level drought monitoring method and system based on gravity satellites, relating to the field of surveying and mapping science and technology. It achieves efficient monitoring through a triple technical architecture: First, it employs a line-of-sight gravity difference algorithm, subtracting the integral reference orbit from the gravity satellite payload ranging data. Based on numerical difference, it obtains the distance acceleration residual, then calculates the line-of-sight gravity difference using a transfer function. The sub-Sun-level mass block of the nadir trajectory is divided into arcs of 2700 epochs (13500 seconds). Combined with Fibonacci point discrete integrals and Tikhonov regularized inversion, it obtains the sub-Sun-level mass change along the orbit. Second, it constructs standardized drought monitoring factors based on the mass change sequence, classifying mild, moderate, and severe drought levels using standard deviation as a threshold. Third, it extracts temporal and spatial features, fuses these features using a fully connected layer, and outputs the drought anomaly probability. The judgment threshold for the anomaly probability is dynamically adjusted, and real-time three-dimensional visualization is achieved through a 5G-A network, improving the accuracy of drought identification.

Description

Technical Field

[0001] This invention relates to the fields of surveying and mapping science and technology, specifically to satellite gravity observation technology, drought disaster monitoring and artificial intelligence data processing methods, and in particular to a sub-sun-level drought monitoring method and system based on gravity satellites. Background Technology

[0002] Satellite gravity observation technology is an important tool for monitoring changes in the Earth's gravitational field, thereby revealing the mass transfer processes within the Earth system. Changes in the Earth's gravitational field are closely related to changes in terrestrial water storage, mass changes in glaciers and ice sheets, and coseismic effects of earthquakes. High-precision gravity field observations can provide a better understanding of the mechanisms of global climate change, water resource distribution, and natural disasters.

[0003] Existing gravity satellite missions have significant limitations in temporal resolution, typically providing monthly (i.e., monthly) global gravity field data. This low temporal resolution limits their application in monitoring short-term mass changes. Monthly data cannot capture rapid processes of Earth's mass change, such as extreme drought events or sharp depletion of groundwater resources. These phenomena usually occur on timescales of days or even hours, and monthly data cannot provide sufficient detail on such timescales.

[0004] Secondly, drought poses a serious threat to agricultural production, the ecological environment, and domestic water supply. High temporal resolution satellite gravity observation technology is of great significance for monitoring drought disasters. Furthermore, by monitoring changes in terrestrial water storage in real time, the severity and extent of drought can be more accurately assessed, providing a scientific basis for disaster early warning and emergency response.

[0005] Therefore, improving the temporal resolution of satellite gravity field solutions and developing technologies that can monitor mass changes on short timescales in real time are not only of great significance to scientific research, but also provide reliable data support for the monitoring and early warning of drought disasters, facilitating the formulation of scientific water resource management policies and mitigating the losses caused by drought disasters. Summary of the Invention

[0006] In view of this, in order to address the technical bottlenecks of existing gravity satellite drought monitoring technologies, such as low temporal resolution, insufficient drought identification accuracy, and low efficiency in processing massive amounts of data, the present invention aims to provide a sub-daily drought monitoring method and system based on gravity satellites with high temporal resolution. This method uses high-temporal-resolution satellite gravity data to invert changes in land surface quality, combines this with a deep learning model to identify drought anomalies, and utilizes a 5G-A (Fifth Generation-Advanced) network to achieve real-time data transmission and visualization. This extends the drought monitoring cycle from monthly to sub-daily levels, significantly enhancing the response capability to sudden droughts. Furthermore, by leveraging deep learning models to integrate spatiotemporal characteristics, the accuracy of drought identification is improved, and the entire process from data inversion to report generation is automated, reducing manual intervention.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] In a first aspect, the present invention provides a sub-Solar drought monitoring method based on a gravity satellite, comprising the following steps:

[0009] Acquire gravity satellite payload data and perform data preprocessing such as gross error removal, resampling, and accelerometer parameter correction;

[0010] Based on the preprocessed gravity satellite payload data and combined with the background force model, the dynamic reference orbit is obtained by using the sliding polynomial integration method. The inter-satellite distance variation converted from the dynamic reference orbit is subtracted from the measured inter-satellite distance variation to obtain the distance variation residual. Then, the distance acceleration residual is calculated by numerical difference, and the distance acceleration residual is converted into the gravity difference in the line-of-sight direction using Fourier transform and transfer function.

[0011] The daily line-of-sight gravity difference is divided into multiple sub-sun level arcs. Along each sub-sun level arc, a disk-shaped mass block is generated at fixed intervals. Each sub-sun level arc corresponds to multiple mass blocks. Fibonacci points are distributed inside the mass blocks as discrete points for integration. A design matrix is ​​constructed through a linear mapping relationship. The Tikhonov regularized inversion method is used to solve the sub-sun level mass change along the track based on the line-of-sight gravity difference vector and the design matrix.

[0012] Calculate the average mass change of the mass blocks obtained by inversion within the target watershed. Based on the average mass change of the mass blocks within the target watershed, construct a drought mass change sequence for the watershed. Calculate the standard deviation based on the drought mass change sequence for the watershed. Use the standard deviation as the drought level determination threshold to determine the drought level and construct drought monitoring factors.

[0013] The drought monitoring factors are aligned with the time series of climate and hydrological data, and the improved Transformer architecture is used to extract temporal features. Spatial features are extracted by combining the ResNet convolutional network. The features are fused through fully connected layers, and the drought anomaly probability is output using the Softmax function. The anomaly probability judgment threshold is dynamically adjusted based on the accuracy feedback of historical drought events.

[0014] A comprehensive drought index is generated based on the drought anomaly probability and the sub-sun level orbital mass change data obtained by inversion. A quantitative analysis report is generated based on the comprehensive drought index. Three-dimensional dynamic visualization is achieved by combining geographic information system. The report and data obtained from sub-sun level drought monitoring are transmitted in real time through 5G-A network.

[0015] As a further aspect of the present invention, the data preprocessing includes: removing gross errors, resampling, and correcting accelerometer parameters in the simplified dynamic orbit, accelerometer, and onboard camera data from the gravity satellite payload data.

[0016] As a further aspect of the present invention, a dynamic reference trajectory is obtained using a sliding polynomial integration method, comprising the following steps:

[0017] Input simplified dynamic orbit data, spaceborne camera data, accelerometer data, prior gravity field model, solid tide and solid polar tide model, ocean tide and marine polar tide model, atmospheric tide, atmospheric and oceanic non-tidal model, multibody motion and relativistic effect parameters;

[0018] The acceleration is integrated using the sliding polynomial integration method to obtain the dynamic reference trajectory.

[0019] As a further aspect of the present invention, the distance acceleration residual is converted into the gravity difference along the line of sight using Fourier transform and transfer function, and the calculation formula is as follows:

[0020] ;

[0021] ;

[0022] In the formula, The line-of-sight gravity difference is used to represent the difference in gravity between two gravity satellites in the line-of-sight direction. This is the inverse Fourier transform; This is the transfer function, used to represent the transfer function used to correct frequency response deviation; Fourier transform; Interstellar distance acceleration; For interstellar acceleration rate residuals; This represents the inter-satellite distance acceleration residual; Indicates the signal vibration frequency. Indicates time.

[0023] As a further aspect of the present invention, the daily gravity difference along the line of sight is divided into multiple sub-sunlight-level arc segments, each sub-sunlight-level arc segment having a length of 2700 epochs (one epoch every 5 seconds). Based on each sub-sunlight-level arc segment, at each interval... Generate a diameter of A disk-shaped mass block, the internal distribution distance of the mass block is... The Fibonacci points are used as discrete points for the integration, where the mass block and The mapping relationship between the line-of-sight direction and the gravitational difference at any given moment is as follows:

[0024] ;

[0025] ;

[0026] In the formula, The gravitational constant is... For the first A mass block in Design matrix elements at all times. For the first The mass of each mass block; For the first The total number of Fibonacci points within a mass block; and This indicates that the two gravity satellites are in Under the solid system coordinate; and This indicates that the two gravity satellites are in Under the solid system coordinate; and This indicates that the two gravity satellites are in Under the solid system coordinate; , Indicates the first The first mass block within the mass block The coordinates of a Fibonacci point in the Earth-fixed system; and The first The first mass block within the mass block The distances from each Fibonacci point to Gravity Satellite 1 and Gravity Satellite 2; This represents the unit vector representing the line-of-sight direction between the two gravity satellites; the coordinates of gravity satellite 1 in the Earth-fixed frame are... , , Gravity Satellite 2's coordinates in the Earth-fixed system are: , , .

[0027] As a further aspect of the present invention, the gravity difference along the line of sight and the change in surface mass have a linear relationship as follows:

[0028] In the formula, The gravity difference vector is the line-of-sight direction. For designing the matrix; This is the mass change vector.

[0029] As a further aspect of this invention, the Tikhonov regularized inversion method is employed. Based on the gravity difference vector along the line of sight and the design matrix, the mass change along the orbit of the sub-Sun-class submarine is solved by the Tikhonov regularized method to obtain a stable solution. The solution is then obtained using the least squares method, and the mass change along the orbit of the sub-Sun-class submarine is inverted and expressed as follows:

[0030] ;

[0031] In the formula, As a regularization factor, As a unit array, To design the transpose of the matrix.

[0032] As a further aspect of the present invention, when constructing drought monitoring factors, the standard deviation based on the standardized quality change sequence is calculated. Sequence values Drought at different scales is defined as a situation that persists for more than a week:

[0033] The drought was initially assessed as mild.

[0034] Sequence Values The drought was initially assessed as moderate.

[0035] Sequence Values It was determined to be a severe drought.

[0036] As a further aspect of this invention, drought monitoring factors are aligned with time series of climate and hydrological data, and the improved Transformer architecture is used to extract temporal features, combined with a ResNet convolutional network to extract spatial features. The specific process of neural network drought monitoring includes the following steps:

[0037] Temporal feature extraction: A Transformer architecture is adopted, and location encoding and multi-head attention mechanisms are introduced to process time series data of drought monitoring factors;

[0038] Spatial feature extraction: Based on the spatial distribution of mass variation, ResNet convolutional network is used to extract regional drought features;

[0039] Threshold adaptive mechanism: The threshold for judging the probability of anomalies is dynamically adjusted based on the accuracy of historical drought events.

[0040] Secondly, the present invention also provides a sub-Solar drought monitoring system based on a gravity satellite, comprising:

[0041] The data preprocessing module is used to acquire gravity satellite payload data and perform data preprocessing such as gross error removal, resampling, and accelerometer parameter correction.

[0042] The gravity difference calculation module is used to obtain the gravity difference in the line-of-sight direction. Based on preprocessed gravity satellite payload data and combined with the background force model, a dynamic reference orbit is obtained using the sliding polynomial integration method. Then, the range variation residual is obtained by subtracting the reference orbit from the measured range variation. The range acceleration residual is calculated using numerical difference, and then converted into the gravity difference in the line-of-sight direction using Fourier transform and transfer function.

[0043] The mass change inversion module is used to divide the daily line-of-sight gravity difference into multiple sub-sun level arcs. Along each sub-sun level arc, a disk-shaped mass block is generated at fixed intervals. Each sub-sun level arc corresponds to multiple mass blocks. Fibonacci points are distributed inside the mass blocks as discrete points for integration. A design matrix is ​​constructed through a linear mapping relationship, and the Tikhonov regularized inversion method is used to solve the sub-sun level track mass change based on the line-of-sight gravity difference vector and the design matrix.

[0044] The drought monitoring factor construction module is used to calculate the average mass change of the mass blocks obtained by inversion within the target watershed, construct the watershed drought mass change sequence based on the average mass change of the mass blocks within the target watershed, calculate the standard deviation based on the watershed drought mass change sequence, use the standard deviation as the drought level determination threshold to determine the drought level, and construct drought monitoring factors.

[0045] The drought disaster monitoring module is used to align drought monitoring factors with climate and hydrological data time series, extract temporal features from the input improved Transformer architecture, and extract spatial features by combining ResNet convolutional network; it fuses features through fully connected layers, outputs drought anomaly probability using the Softmax function, and dynamically adjusts the anomaly probability judgment threshold based on the accuracy feedback of historical drought events.

[0046] The results output module is used to generate a comprehensive drought index based on the drought anomaly probability and the sub-day-level orbital mass change data obtained by inversion, generate a quantitative analysis report based on the comprehensive drought index, realize three-dimensional dynamic visualization by combining with geographic information system, and transmit the reports and data obtained from sub-day-level drought monitoring in real time through 5G-A network.

[0047] Compared with existing technologies, the sub-Solar drought monitoring method and system based on gravity satellites provided in this invention have the following advantages:

[0048] This invention shortens the processing cycle of gravity satellite data from the traditional monthly calculation to the sub-day level by using line-of-sight gravity difference inversion and sub-day arc segmentation, thereby increasing the monitoring frequency and enabling dynamic tracking of drought evolution. It employs a Transformer architecture neural network to extract temporal features and combines it with a ResNet convolutional network to extract spatial features, integrating drought monitoring factors, climate, and hydrological data to reduce missed detections and false alarms. The physical model based on gravity difference inversion provides interpretable drought monitoring factors, reducing reliance on empirical parameters. Furthermore, it improves numerical stability through Tikhonov regularization-stabilized mass inversion calculations and Fibonacci point discretization integration, avoiding the accumulation of local errors. Sub-day level monitoring provides a critical window for early drought warning, and a quantitative drought index supports refined water resource management, achieving breakthroughs in drought monitoring in terms of temporal resolution, accuracy, efficiency, and adaptability.

[0049] These or other aspects of the invention will become more apparent from the following description of embodiments. It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0050] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. In the drawings:

[0051] Figure 1 This is a flowchart of a sub-sun-level drought monitoring method based on a gravity satellite according to the present invention.

[0052] Figure 2 This is a flowchart illustrating the construction of drought monitoring factors and the generation of drought monitoring information in a sub-sun-level drought monitoring method based on gravity satellites according to the present invention.

[0053] Figure 3 This is a structural block diagram of a sub-Solar drought monitoring system based on a gravity satellite according to the present invention. Detailed Implementation

[0054] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings, but it should be understood that the scope of protection of the present invention is not limited to the specific embodiments.

[0055] Unless otherwise expressly stated, throughout the specification and claims, the term "comprising" or its variations such as "including" or "comprises" shall be understood to include the stated elements or components without excluding other elements or other components.

[0056] like Figure 1 and Figure 2 As shown, one embodiment of the present invention provides a sub-Sun-level drought monitoring method based on gravity satellites. The method aims to utilize gravity satellite payload data to convert inter-satellite distance variability into line-of-sight gravity difference, then divide this gravity difference into arc segments to construct orbital mass blocks, inverting orbital mass changes. This improves the temporal resolution to the sub-Sun level and constructs sub-Sun-level drought monitoring factors. Furthermore, the present invention proposes a drought disaster monitoring neural network based on sub-Sun-level mass changes for automated, high-precision analysis of drought disasters, and transmits and broadcasts large amounts of data via a 5G-A wireless network, serving as a monitoring and early warning system for disasters. This sub-Sun-level drought monitoring method includes the following steps:

[0057] Step S10: Acquire gravity satellite payload data and perform data preprocessing such as gross error removal, resampling, and accelerometer parameter correction.

[0058] In this step, the data preprocessing includes: removing gross errors, resampling, and correcting accelerometer parameters in the simplified dynamic orbit, accelerometer, and onboard camera data from the gravity satellite payload data, and eliminating interference from solid tides, ocean tides, and atmospheric tides.

[0059] Step S20: Based on the preprocessed gravity satellite payload data and combined with the background force model, the dynamic reference orbit is obtained by using the sliding polynomial integration method. The inter-satellite distance variation converted from the dynamic reference orbit is subtracted from the measured inter-satellite distance variation to obtain the distance variation residual. Then, the distance acceleration residual is calculated by numerical difference, and the distance acceleration residual is converted into the line-of-sight gravity difference using Fourier transform and transfer function.

[0060] In this step, simplified dynamic orbit data, onboard camera data, accelerometer data, prior gravity field model, solid tide and solid polar tide model, ocean tide and marine polar tide model, atmospheric tide, atmospheric and ocean non-tidal model, multibody motion and relativistic effect parameters are input; the acceleration is integrated using the sliding polynomial integral method to obtain the dynamic reference orbit.

[0061] In this embodiment, the distance acceleration residual is converted into the gravity difference along the line of sight using Fourier transform and transfer function. The calculation formula is as follows:

[0062] ;

[0063] ;

[0064] In the formula, The line-of-sight gravity difference is used to represent the difference in gravity between two gravity satellites in the line-of-sight direction. This is the inverse Fourier transform; This is the transfer function, used to represent the transfer function used to correct frequency response deviation; Fourier transform; Interstellar distance acceleration; For interstellar acceleration rate residuals; This represents the inter-satellite distance acceleration residual; Indicates the signal vibration frequency. Indicates time.

[0065] Step S30: Divide the daily line-of-sight gravity difference into multiple sub-sun level arcs. Along each sub-sun level arc, generate a disk-shaped mass block at fixed intervals. Each sub-sun level arc corresponds to multiple mass blocks. Fibonacci points are distributed inside the mass blocks as discrete points for integration. Construct a design matrix through a linear mapping relationship, and use the Tikhonov regularized inversion method to solve for the sub-sun level mass change along the track based on the line-of-sight gravity difference vector and the design matrix.

[0066] In this embodiment, the daily gravity difference along the line of sight is divided into multiple sub-sun level arc segments, each with a length of 2700 epochs, and each epoch is 5 seconds long. Based on each sub-sun level arc segment, the distance between each arc segment is... Generate a diameter of A disk-shaped mass block, the internal distribution distance of the mass block is... The Fibonacci points are used as discrete points for the integration, where the mass block and The mapping relationship between the line-of-sight direction and the gravitational difference at any given moment is as follows:

[0067] ;

[0068] ;

[0069] In the formula, The gravitational constant is... For the first A mass block in Design matrix elements at all times. For the first The mass of each mass block; For the first The total number of Fibonacci points within a mass block; and This indicates that the two gravity satellites are in Under the solid system coordinate; and This indicates that the two gravity satellites are in Under the solid system coordinate; and This indicates that the two gravity satellites are in Under the solid system coordinate; , Indicates the first The first mass block within the mass block The coordinates of a Fibonacci point in the Earth-fixed system; and The first The first mass block within the mass block The distances from each Fibonacci point to Gravity Satellite 1 and Gravity Satellite 2; This represents the unit vector representing the line-of-sight direction between the two gravity satellites; the coordinates of gravity satellite 1 in the Earth-fixed frame are... , , Gravity Satellite 2's coordinates in the Earth-fixed system are: , , .

[0070] Among them, the linear relationship between the gravity difference along the line of sight and the change in surface mass is as follows:

[0071] In the formula, The gravity difference vector is the line-of-sight direction. For designing the matrix; This is the mass change vector.

[0072] In this embodiment, the Tikhonov regularized inversion method is adopted. Based on the gravity difference vector along the line of sight and the design matrix, the mass change along the orbit of the sub-Sun-class is solved by the Tikhonov regularized method to obtain a stable solution. The least squares method is then used to solve for the mass change along the orbit of the sub-Sun-class, which is expressed as:

[0073] ;

[0074] In the formula, As a regularization factor, As a unit array, To design the transpose of the matrix.

[0075] Step S40: Calculate the average mass change of the mass blocks obtained from the inversion within the target watershed. Based on the average mass change of the mass blocks within the target watershed, construct a drought mass change sequence for the watershed. Calculate the standard deviation based on the drought mass change sequence for the watershed. Use the standard deviation as the drought level determination threshold to determine the drought level and construct drought monitoring factors.

[0076] In this embodiment, when constructing drought monitoring factors, the standard deviation based on the standardized quality change sequence is calculated. Sequence values Drought at different scales is defined as a situation that persists for more than a week:

[0077] The drought was initially assessed as mild.

[0078] Sequence Values The drought was initially assessed as moderate.

[0079] Sequence Values It was determined to be a severe drought.

[0080] Step S50: Align drought monitoring factors with climate and hydrological data time series, input the improved Transformer architecture to extract temporal features, and combine ResNet convolutional network to extract spatial features; fuse features through fully connected layers, use the Softmax function to output drought anomaly probability, and dynamically adjust the anomaly probability judgment threshold based on the accuracy feedback of historical drought events.

[0081] In this embodiment, drought monitoring factors are aligned with time series of climate and hydrological data, and an improved Transformer architecture is used to extract temporal features. Spatial features are then extracted using a ResNet convolutional network. The specific process of neural network drought monitoring includes the following steps:

[0082] Temporal feature extraction: A Transformer architecture is adopted, and location encoding and multi-head attention mechanisms are introduced to process time series data of drought monitoring factors;

[0083] Spatial feature extraction: Based on the spatial distribution of mass variation, ResNet convolutional network is used to extract regional drought features;

[0084] Threshold adaptive mechanism: The threshold for judging the probability of anomalies is dynamically adjusted based on the accuracy of historical drought events.

[0085] In this embodiment, the neural network drought disaster monitoring network first performs data input and preprocessing, standardizing climate and hydrological data and aligning the time series with drought monitoring factors. Based on the polygon information of the mass blocks, it extracts and calculates regional feature data within the mass blocks to ensure spatial consistency. Then, it extracts temporal features using a Transformer architecture incorporating a temporal attention mechanism to capture anomalous changes in drought monitoring factors and climate and hydrological data. Location encoding is introduced to enhance the model's understanding of the time series sequence; a multi-head attention mechanism is used to extract multi-scale temporal features, improving the model's ability to capture the drought evolution process. For spatial feature extraction, ResNet convolution is used based on the spatial distribution of drought monitoring factors to extract spatial features of drought disasters. Combining this with the spatial distribution of climate and hydrological data enhances the model's ability to identify regional drought characteristics. For anomaly pattern recognition, the extracted temporal and spatial features are input into a fully connected layer for feature fusion. The Softmax function outputs the anomaly probability to determine whether a drought anomaly has occurred. Then, the anomaly probability judgment threshold is dynamically adjusted based on historical drought events and real-time monitoring results. A feedback mechanism is introduced to optimize the threshold setting by combining user input and model output, improving monitoring accuracy. Finally, based on anomaly probability and quality change data, a comprehensive drought index is designed to quantify drought severity. The drought index includes temporal and spatial dimensions. Information such as the drought index, anomaly occurrence time, location, and severity is visualized, and combined with a geographic information system to generate a quantitative analysis report and a 3D dynamic visualization map.

[0086] Step S60: Generate a comprehensive drought index based on the drought anomaly probability and the sub-sun level orbital mass change data obtained from inversion; generate a quantitative analysis report based on the comprehensive drought index; achieve three-dimensional dynamic visualization by combining with a geographic information system; and transmit the report and data obtained from sub-sun level drought monitoring in real time through a 5G-A network.

[0087] The sub-day drought monitoring method using gravity satellites in this invention addresses the technical bottlenecks of existing gravity satellite drought monitoring technologies, such as low temporal resolution, insufficient drought identification accuracy, and low efficiency in processing massive amounts of data. Based on gravity satellite payload data and monthly solution data, it employs a line-of-sight gravity difference algorithm combined with adaptive filtering technology to invert high temporal resolution surface quality changes along the orbital direction, shortening the monitoring cycle to the sub-day level and constructing drought monitoring factors. A spatiotemporally coupled deep learning framework is constructed, employing an improved Transformer architecture neural network and introducing an attention mechanism to achieve efficient feature extraction and drought anomaly pattern recognition from time-series data of drought monitoring factors. A 5G-A transmission network is integrated to transmit quality change monitoring information in real time, and a quantitative analysis report and a three-dimensional dynamic visualization map are generated in conjunction with a geographic information system, highlighting the spatiotemporal distribution of water resource changes.

[0088] This invention improves the temporal resolution of drought monitoring to the sub-Sun timescale based on gravity satellite technology; it accurately identifies drought-related anomalies through a neural network model; and the 5G-A transmission module significantly reduces data latency. It is particularly suitable for dynamically capturing the evolution of drought processes, providing scientific observational data support at the sub-Sun timescale for drought early warning, water resource management, and climate research.

[0089] See Figure 3 As shown, another embodiment of the present invention also provides a sub-Sun-level drought monitoring system based on a gravity satellite, the sub-Sun-level drought monitoring system comprising:

[0090] The data preprocessing module is used to acquire gravity satellite payload data from external data storage and perform data preprocessing such as gross error removal, resampling, and accelerometer parameter correction.

[0091] The gravity difference calculation module is used to obtain the gravity difference in the line-of-sight direction. Based on preprocessed gravity satellite payload data and combined with the background force model, a dynamic reference orbit is obtained using the sliding polynomial integration method. Then, the range variation residual is obtained by subtracting the reference orbit from the measured range variation. The range acceleration residual is calculated using numerical difference, and then converted into the gravity difference in the line-of-sight direction using Fourier transform and transfer function.

[0092] The mass change inversion module is used to divide the daily line-of-sight gravity difference into multiple sub-day-level arc segments. Along each sub-day-level arc segment, a disk-shaped mass block is generated at fixed intervals. Each sub-day-level arc segment corresponds to multiple mass blocks. Fibonacci points are distributed inside the mass blocks as discrete points for integration. A design matrix is ​​constructed through a linear mapping relationship, and the Tikhonov regularized inversion method is used to solve for the sub-day-level track-side mass change based on the line-of-sight gravity difference vector and the design matrix. Intermediate and historical data can be stored in the internal memory of the sub-day-level drought monitoring system.

[0093] The drought monitoring factor construction module is used to calculate the average mass change of the mass blocks obtained by inversion within the target watershed. Based on the average mass change of the mass blocks within the target watershed, a drought mass change sequence of the watershed is constructed. The standard deviation of the drought mass change sequence of the watershed is calculated. The standard deviation is used as the drought level determination threshold to determine the drought level. A standardized drought monitoring factor that can determine the mild, moderate and severe drought levels is constructed.

[0094] The drought disaster monitoring module is used to align drought monitoring factors with climate and hydrological data time series, extract temporal features from the input improved Transformer architecture, and extract spatial features by combining ResNet convolutional network; it fuses features through fully connected layers, outputs drought anomaly probability using the Softmax function, and dynamically adjusts the anomaly probability judgment threshold based on the accuracy feedback of historical drought events.

[0095] The results output module is used to generate a comprehensive drought index based on the drought anomaly probability and the sub-day-level orbital mass change data obtained by inversion. Based on the comprehensive drought index, a quantitative analysis report is generated. Combined with the geographic information system engine, the geographic information system realizes three-dimensional dynamic visualization of drought level, spatial distribution, etc., forming an intuitive interface for visual display of decision results. At the same time, by integrating a 5G-A network drought monitoring and response unit, the monitoring data is transmitted in real time through the 5G-A network, and the monitoring report and early warning information are transmitted to water resource management platforms at all levels with a latency of less than milliseconds.

[0096] To ensure the efficient and stable operation of the core processing flow, the sub-day drought monitoring system in this embodiment also includes:

[0097] External data storage: used to persistently store raw, unprocessed satellite payload data streams as a system data source;

[0098] Sub-day drought monitoring internal memory: serving as a high-speed cache and temporary database, used to store preprocessed intermediate data, inverted quality change sequences, historical monitoring factors, and model parameters, providing fast data access support for real-time computing and model training;

[0099] System Development and Maintenance Interface: This interface provides an integrated development environment for IDE program maintenance. It supports updating, debugging, and optimizing regularization parameters, neural network structures, and threshold adjustment strategies within the system, ensuring continuous improvement and adaptability of the system.

[0100] Drought Disaster Assessment Terminal: This terminal receives and comprehensively analyzes information such as drought index, spatial extent, duration, and evolution trend provided by the results output module. Based on a pre-set assessment model, it comprehensively evaluates the severity, impact range, and potential risks of drought events, generating a disaster assessment report to provide direct support for disaster prevention and mitigation decision-making.

[0101] The sub-Sun-level drought monitoring system in this embodiment can combine GNSS (Global Navigation Satellite System) data to jointly retrieve sub-Sun-level mass change data, thereby improving the spatial resolution, accuracy, and reliability of sub-Sun-level mass change data. During the data training and decision-making process of the deep learning model, in addition to climate, hydrological, and mass change data, more types of data sources can be integrated, such as satellite remote sensing data and social media data. Through multimodal fusion technology of image, text, and time-series data, complex environmental changes can be captured from different perspectives, enhancing generalization capabilities.

[0102] This invention discloses a sub-day-level drought monitoring method and system based on gravity satellites. By employing line-of-sight gravity difference inversion and sub-day-level arc segmentation, the processing cycle of gravity satellite data is shortened from the traditional monthly calculation to the sub-day level, increasing the monitoring frequency and enabling dynamic tracking of drought evolution. Furthermore, a Transformer architecture neural network is used to extract temporal features, combined with a ResNet convolutional network to extract spatial features, integrating drought monitoring factors, climate, and hydrological data to reduce missed detections and false alarms. The physical model based on gravity difference inversion provides interpretable drought monitoring factors, reducing dependence on empirical parameters. Tikhonov regularization stabilizes the quality inversion calculation, and Fibonacci point discretization integration improves numerical stability and avoids the accumulation of local errors. Sub-day-level monitoring provides a critical window for early drought warning, and quantitative drought indices support refined water resource management, achieving breakthroughs in drought monitoring in terms of temporal resolution, accuracy, efficiency, and adaptability.

[0103] The foregoing description of specific exemplary embodiments of the invention is for illustrative and explanatory purposes. These descriptions are not intended to limit the invention to the precise forms disclosed, and it will be apparent that many changes and variations can be made in accordance with the foregoing teachings. The exemplary embodiments were chosen and described in order to explain the specific principles of the invention and its practical application, thereby enabling those skilled in the art to implement and utilize various different exemplary embodiments of the invention, as well as various different choices and variations. The scope of the invention is intended to be defined by the claims and their equivalents.

Claims

1. A sub-Solar drought monitoring method based on gravity satellites, characterized in that, Includes the following steps: Acquire gravity satellite payload data and perform data preprocessing such as gross error removal, resampling, and accelerometer parameter correction; Based on the preprocessed gravity satellite payload data and combined with the background force model, the dynamic reference orbit is obtained by using the sliding polynomial integration method. The inter-satellite distance variation converted from the dynamic reference orbit is subtracted from the measured inter-satellite distance variation to obtain the distance variation residual. Then, the distance acceleration residual is calculated by numerical difference, and the distance acceleration residual is converted into the gravity difference in the line-of-sight direction using Fourier transform and transfer function. The daily line-of-sight gravity difference is divided into multiple sub-sun level arcs. Along each sub-sun level arc, a disk-shaped mass block is generated at fixed intervals. Each sub-sun level arc corresponds to multiple mass blocks. Fibonacci points are distributed inside the mass blocks as discrete points for integration. A design matrix is ​​constructed through a linear mapping relationship. The Tikhonov regularized inversion method is used to solve the sub-sun level mass change along the track based on the line-of-sight gravity difference vector and the design matrix. Calculate the average mass change of the mass blocks obtained by inversion within the target watershed. Based on the average mass change of the mass blocks within the target watershed, construct a drought mass change sequence for the watershed. Calculate the standard deviation based on the drought mass change sequence for the watershed. Use the standard deviation as the drought level determination threshold to determine the drought level and construct drought monitoring factors. The drought monitoring factors are aligned with the time series of climate and hydrological data, and the improved Transformer architecture is used to extract temporal features. Spatial features are extracted by combining the ResNet convolutional network. The features are fused through fully connected layers, and the drought anomaly probability is output using the Softmax function. The anomaly probability judgment threshold is dynamically adjusted based on the accuracy feedback of historical drought events. A comprehensive drought index is generated based on the drought anomaly probability and the sub-sun level orbital mass change data obtained by inversion. A quantitative analysis report is generated based on the comprehensive drought index. Three-dimensional dynamic visualization is achieved by combining geographic information system. The report and data obtained from sub-sun level drought monitoring are transmitted in real time through 5G-A network.

2. The sub-Solar drought monitoring method based on gravity satellite as described in claim 1, characterized in that, The data preprocessing includes: removing gross errors, resampling, and correcting accelerometer parameters in the simplified dynamic orbit, accelerometer, and onboard camera data from the gravity satellite payload data.

3. The sub-Sun-level drought monitoring method based on gravity satellite as described in claim 2, characterized in that, The dynamic reference orbit is obtained using the sliding polynomial integration method, including the following steps: Input simplified dynamic orbit data, spaceborne camera data, accelerometer data, prior gravity field model, solid tide and solid polar tide model, ocean tide and marine polar tide model, atmospheric tide, atmospheric and oceanic non-tidal model, multibody motion and relativistic effect parameters; The acceleration is integrated using the sliding polynomial integration method to obtain the dynamic reference trajectory.

4. The sub-Sun-level drought monitoring method based on gravity satellite as described in claim 3, characterized in that, The distance acceleration residual is converted into the gravity difference along the line of sight using Fourier transform and transfer function. The calculation formula is as follows: ； ； In the formula, The difference in gravity along the line of sight; This is the inverse Fourier transform; For transfer functions; Fourier transform; Interstellar distance acceleration; For interstellar acceleration rate residuals; This represents the inter-satellite distance acceleration residual; Indicates the signal vibration frequency. Indicates time.

5. The sub-Sun-level drought monitoring method based on gravity satellite as described in claim 4, characterized in that, The daily line-of-sight gravity difference is divided into multiple sub-sun level arcs, each with a length of 2700 epochs, and each epoch lasts 5 seconds. Based on each sub-sun level arc, the interval distance is... Generate a diameter of A disk-shaped mass block, the internal distribution distance of the mass block is... The Fibonacci points are used as discrete points for the integration, where the mass block and The mapping relationship between the line-of-sight direction and the gravitational difference at any given moment is as follows: ； ； In the formula, The gravitational constant is... For the first A mass block in Design matrix elements at all times. For the first The mass of each mass block; For the first The total number of Fibonacci points within a mass block; and This indicates that the two gravity satellites are in Under the solid system coordinate; and This indicates that the two gravity satellites are in Under the solid system coordinate; and This indicates that the two gravity satellites are in Under the solid system coordinate; , Indicates the first The first mass block within the mass block The coordinates of a Fibonacci point in the Earth-fixed system; and The first The first mass block within the mass block The distances from each Fibonacci point to Gravity Satellite 1 and Gravity Satellite 2; This represents the unit vector representing the line-of-sight direction between the two gravity satellites; the coordinates of gravity satellite 1 in the Earth-fixed frame are... , , Gravity Satellite 2's coordinates in the Earth-fixed system are: , , .

6. The sub-Sun-level drought monitoring method based on gravity satellite as described in claim 5, characterized in that, The gravitational difference along the line of sight has a linear relationship with changes in surface mass as follows: In the formula, The gravity difference vector is the line-of-sight direction. For designing the matrix; This is the mass change vector.

7. The sub-Sun-level drought monitoring method based on gravity satellite as described in claim 6, characterized in that, Using the Tikhonov regularized inversion method, based on the gravity difference vector along the line of sight and the design matrix, the mass variation along the orbit of the sub-Sun-class submarine is solved by the Tikhonov regularized method to obtain a stable solution, which is then solved using the least squares method. The mass variation along the orbit of the sub-Sun-class submarine is obtained by inversion, expressed as: ； In the formula, As a regularization factor, As a unit array, To design the transpose of the matrix.

8. The sub-Solar drought monitoring method based on gravity satellite as described in claim 7, characterized in that, When constructing drought monitoring factors, the standard deviation based on the standardized quality change series is calculated. Sequence values Drought at different scales is defined as a situation that persists for more than a week: The drought was initially assessed as mild. Sequence Values The drought was initially assessed as moderate. Sequence Values It was determined to be a severe drought.

9. A sub-Solar drought monitoring system based on a gravity satellite, characterized in that, For performing the method as described in any one of claims 1-8, the sub-day drought monitoring system comprises: The data preprocessing module is used to acquire gravity satellite payload data and perform data preprocessing such as gross error removal, resampling, and accelerometer parameter correction. The gravity difference calculation module is used to obtain the gravity difference in the line-of-sight direction. Based on preprocessed gravity satellite payload data and combined with the background force model, a dynamic reference orbit is obtained using the sliding polynomial integration method. Then, the range variation residual is obtained by subtracting the reference orbit from the measured range variation. The range acceleration residual is calculated using numerical difference, and then converted into the gravity difference in the line-of-sight direction using Fourier transform and transfer function. The mass change inversion module is used to divide the daily line-of-sight gravity difference into multiple sub-sun level arcs. Along each sub-sun level arc, a disk-shaped mass block is generated at fixed intervals. Each sub-sun level arc corresponds to multiple mass blocks. Fibonacci points are distributed inside the mass blocks as discrete points for integration. A design matrix is ​​constructed through a linear mapping relationship, and the Tikhonov regularized inversion method is used to solve the sub-sun level track mass change based on the line-of-sight gravity difference vector and the design matrix. The drought monitoring factor construction module is used to calculate the average mass change of the mass blocks obtained by inversion within the target watershed, construct the watershed drought mass change sequence based on the average mass change of the mass blocks within the target watershed, calculate the standard deviation based on the watershed drought mass change sequence, use the standard deviation as the drought level determination threshold to determine the drought level, and construct drought monitoring factors. The drought disaster monitoring module is used to align drought monitoring factors with climate and hydrological data time series, extract temporal features from the input improved Transformer architecture, and extract spatial features by combining ResNet convolutional network; it fuses features through fully connected layers, outputs drought anomaly probability using the Softmax function, and dynamically adjusts the anomaly probability judgment threshold based on the accuracy feedback of historical drought events. The results output module is used to generate a comprehensive drought index based on the drought anomaly probability and the sub-day-level orbital mass change data obtained by inversion, generate a quantitative analysis report based on the comprehensive drought index, realize three-dimensional dynamic visualization by combining with geographic information system, and transmit the reports and data obtained from sub-day-level drought monitoring in real time through 5G-A network.

Images