454 results about "Satellite data" patented technology

The invention discloses a double-layer ionosphere modeling method fusing low-orbit satellite data, which comprises the following steps of: preprocessing ground-based GNSS (Global Navigation Satellite System) data to obtain an STEC expression of a ground-based GNSS; preprocessing the precise orbit determination data of the low-orbit satellite to obtain a VTEC expression of the precise orbit determination data; obtaining an electron density profile product of the radio occultation of the low-orbit satellite, and obtaining VTEC of the radio occultation through profile integration; using two groups of spherical harmonic coefficients to express global VTEC distribution, and constructing an observation equation; and determining weights of different types of data by using Helmert variance component estimation, and finally solving to obtain two groups of spherical harmonic coefficients to generate an ionized layer grid product. Projection errors caused by vertical layering in a traditional single-layer model are weakened, meanwhile, precise orbit determination data and radio occultation data of a low-orbit satellite are combined, the problem that ionosphere TEC cannot be correctly layered in a double-layer model is solved, data of areas where GNSS tracking stations are sparsely distributed are supplemented, and the reliability of the GNSS tracking stations is improved. Therefore, a global ionosphere model with higher precision is obtained.

The invention belongs to the technical field of remote sensing, and relates to a multispectral remote sensing atmospheric correction method based on fusion of a physical model and a neural network, which comprises the following steps: S1, classifying aerosol parameter data obtained based on foundation observation, remote sensing inversion or meteorological model and satellite data fusion inversion by adopting an unsupervised clustering algorithm, extracting a typical aerosol mode with physical representativeness; s2, in combination with the typical aerosol mode, simulating the apparent reflectivity of an observation channel under a plurality of atmospheric states and observation geometric conditions by using a radiation transfer model, generating a lookup table covering a wide parameter space, and constructing a training data set; and S3, constructing a nonlinear regression model, taking the training data set as a training sample, learning a mapping relation among an observation angle, an aerosol condition and surface reflectance, performing atmospheric interference correction on an actual multispectral remote sensing image under a pollution condition, and outputting a surface reflectance result.

A method for transferring seed during a planting operation includes receiving, by a theft detection system, a first equipment identity of first equipment, receiving the seed into a seed bin of second equipment from the first equipment, determining, using a sensing device disposed on the second equipment and communicatively coupled to the theft detection system, a first weight value indicative of weight of the seed received into the seed bin from the first equipment, and at least one of transmitting, by the theft detection system using a satellite data communication protocol, the first weight value and the first equipment identity to a tracking system, and determining, by the theft detection system, whether any of the seed was stolen based on the first weight value and the first equipment identity and selectively generating and transmitting a notification to the tracking system indicating that the seed was stolen.

The invention relates to the technical field of agricultural quantitative remote sensing and intelligent agriculture, and discloses a multi-source remote sensing data radiation correction and chlorophyll content cross-platform inversion method, which comprises the following steps: synchronously obtaining ground actual measurement, unmanned aerial vehicle and satellite remote sensing data, calculating vegetation indexes, and screening sensitive characteristic variables by using Pearson correlation analysis; establishing a radiation correction model by adopting a ratio averaging method, performing radiation normalization processing on satellite data by taking unmanned aerial vehicle data as a reference, and eliminating sensor response differences; an SPAD inversion model is constructed and optimized based on Z-score standardization, the SPAD inversion model is migrated and applied to corrected satellite data, a regional scale chlorophyll distribution diagram is generated through pixel-by-pixel calculation, and then a differentiated farmland management strategy is formulated. According to the method, the difficulty of multi-source data collaboration is solved, the regional scale crop chlorophyll monitoring precision is improved, and scientific decision support is provided for large-area crop fertilization.

The invention relates to the technical field of thunder and lightning early warning, and discloses a thunderstorm early warning system and method applied to a new energy station, and the system comprises a data collection module, a data processing module and an early warning decision module which are arranged in a target new energy station. The multi-source thunderstorm early warning information is sent to a data processing module; wherein the multi-source thunderstorm early warning information comprises lightning data, radar data, satellite data, micrometeorological data and atmospheric electric field data, and the data processing module performs feature extraction on the multi-source thunderstorm early warning information to obtain a first thunderstorm early warning feature vector and a second thunderstorm early warning feature vector; and the early warning decision module performs thunderstorm early warning based on the first thunderstorm early warning feature vector and the second thunderstorm early warning feature vector to obtain thunderstorm early warning information of the new energy station. According to the invention, the accuracy of the thunderstorm early warning information of the new energy station is improved.

The invention relates to the technical field of mine environmental protection, and discloses a gravel aggregate mine carbon emission real-time monitoring and metering method, which comprises the following steps: S1, multi-source sensing, including mining blasting monitoring, ore transportation monitoring and aggregate processing monitoring; s2, edge calculation, including data preprocessing, dynamic compensation calculation and local storage and communication; in the dynamic compensation calculation, the CO2 concentration is output and corrected in real time through a dynamic compensation algorithm, and the dynamic compensation algorithm comprises particulate matter dust drop, temperature and humidity deviation and vibration compensation; and S3, cloud platform processing is carried out, block chain evidence storage, carbon sink deduction calculation and early warning control are carried out, and the carbon sink deduction calculation is vegetation carbon sink absorption amount calculation deduction and waste rock recycling carbon emission reduction calculation deduction according to satellite data. The edge calculation layer adopts a dynamic compensation algorithm, temperature and humidity, vibration and particulate matter compensation items are fused in real time, and the hysteresis bottleneck of accounting is solved; a satellite-ground fusion deduction mechanism is adopted to construct a judicial-level trusted data chain.

The invention discloses a low-altitude aircraft accurate positioning method based on Beidou satellite data, and the method comprises the following steps: collecting the multi-source data of Beidou positioning, IMU, barometer and altimeter in the flight process, and generating an original positioning data set; after multi-source data are aligned, a fusion data sequence is constructed; carrying out error modeling by adopting zero offset correction, drift compensation and scale normalization, and outputting a calibrated data sequence; an improved FGO algorithm is input, factors such as Beidou satellite pseudo-range, IMU integration and process prior are processed in a combined mode, track combined optimization is executed, and a preliminary positioning result is output; performing error correction by combining a ground differential base station and satellite-based enhanced data, and outputting a high-precision positioning result sequence; and abnormal point elimination and trajectory smoothing are realized through consistency discrimination and confidence evaluation, and a final positioning result is obtained. The method improves the positioning precision and robustness of the low-altitude aircraft, and is suitable for high-precision positioning application in a complex environment.

The invention discloses a resource scheduling optimization method for a giant constellation, relates to the technical field of satellite resource allocation of the giant constellation, solves the problem that an existing satellite resource scheduling method is difficult to give consideration to global optimization of time and energy consumption, and aims to improve the execution efficiency of a multi-satellite cooperative task and reduce energy consumption. A real satellite scheduling task is simulated and arranged to obtain satellite data and task data, an evolutionary process of a reinforcement learning guide genetic algorithm is adopted to optimize a task sequence, satellite resource allocation is performed on the optimized task sequence, time and energy consumption of the task sequence are completed, and finally normalization is performed to obtain a comprehensive index. According to the method, a task execution sequence with the minimum time and energy consumption comprehensive index is obtained, so that an optimal scheduling scheme is obtained, and the requirement of satellite scheduling under a giant constellation is met.

The invention provides a satellite service data analysis method and system based on a dynamic protocol, belongs to the technical field of satellite real-time data processing, and is used for analyzing satellite data downloaded by a data transmission link so as to complete closed-loop verification on a satellite data transmission core data flow path. The method depends on the support capability of an on-satellite data transmission system for a multi-channel downloading mode, and the independent, lossless and time sequence fidelity binary disk falling capability of ground detection equipment for each channel data. The method comprises the following steps: acquiring original downlink data highly consistent with an on-orbit state; secondly, adopting a scheduler model to carry out structured segmentation and affiliation allocation on the AOS frame after the original data stream is restored; and finally, dynamically loading a corresponding protocol description file for each data unit, and performing field-by-field decoding according to a mapping rule of binary data and readable engineering quantity to generate structured table data. The method can be widely applied to whole-satellite integration test scenes of various remote sensing satellites such as optics, SAR and infrared.

The embodiment of the invention provides a navigation state synchronization method, device and equipment and a computer storage medium, and belongs to the technical field of positioning. Receiving first GNSS satellite data, including a first timestamp; searching first IMU data corresponding to the first timestamp from an IMU data queue; carrying out fusion calculation, and determining a real-time navigation state under the first timestamp; moving a queue head position of the IMU data queue to second IMU data, and initializing an increment integral state; performing incremental integral calculation based on the second IMU data and a target IMU data set in the IMU data queue to obtain a real-time incremental pre-integral; and performing navigation state calculation according to the real-time navigation state and the real-time increment pre-integration under the first timestamp, and determining the current navigation state. According to the method disclosed by the embodiment of the invention, the purposes of accurately fusing navigation combinations and improving the navigation accuracy are achieved.

The invention discloses a fan site selection method and system based on different-mode nested grids, and belongs to the technical field of fan site selection, and the method comprises the steps: firstly constructing a large-region sea ice background field by using global data to preliminarily screen candidate regions, then constructing a small-region high-resolution background field through a grid nesting technology, and carrying out the dynamic simulation, according to the method, a sea ice evolution field is obtained, an initial field is corrected by fusing field and satellite data, errors are quantized, comprehensive scoring is carried out to screen target site selection, finally, simulation parameters are dynamically optimized according to actually measured data, and a final site selection is determined through iterative correction. According to the method, multi-scale evaluation from macroscopic preliminary screening to microcosmic fine screening is realized, and through data fusion and dynamic optimization, the precision and reliability of fan site selection in a complex sea ice environment are remarkably improved.

The invention discloses an accumulated snow coverage analysis method and system and a medium, and the method comprises the steps: obtaining a plurality of pieces of to-be-processed satellite data of a to-be-analyzed region; performing feature extraction on the to-be-processed satellite data to obtain multi-dimensional feature data of each preset pixel; analyzing the multi-dimensional feature data, and determining a first accumulated snow identification result; respectively determining a plurality of target preset pixels geographically covering each preset geographic grid; carrying out statistics on distribution of the first accumulated snow identification results of a plurality of target preset pixels in each preset geographic grid, and determining a state statistical index; constructing a time data sequence of a preset geographic grid; analyzing the time data sequence of each preset geographic grid, and determining a second accumulated snow identification result of each preset geographic grid; and generating an accumulated snow coverage data set of the to-be-analyzed area based on the second accumulated snow identification result. The precision of the accumulated snow coverage data is improved.

The invention discloses a computing power scheduling method for multi-satellite collaborative application. The computing power scheduling method comprises the steps that 1, a ground station generates an undirected connection graph according to a satellite constellation topological structure and a resource state; step 2, the ground station receives a user instruction and carries out task computing power demand estimation; step 3, the ground station adopts a sliding window mode on the undirected connection graph to determine an execution area for executing the task; 4, calculating a central node of the execution area, and taking the central node as a source satellite for executing an information acquisition task; 5, executing an information acquisition task by the source satellite; step 6, completing task data distribution from the source satellite to other satellite nodes in the area by adopting a double-factor sorting scheduling strategy; 7, the satellite node returns a result to the source satellite after completing task processing; and step 8, the source satellite collects the result and sends the result to the ground station. The invention provides a strategy of multi-satellite cooperative processing and resource dynamic scheduling under the condition that the single-satellite capability is limited, and the processing time delay of on-satellite data is reduced.

The embodiment of the invention provides a bridge monitoring satellite data quality control method based on signal-to-noise ratio grid comparison, and belongs to the technical field of satellites. The method comprises the following steps: carrying out pseudo-range single-point positioning calculation according to an observation value file and a broadcast ephemeris file to obtain observation data of each satellite; performing grid division on the satellite sky map according to a preset elevation angle interval and an azimuth angle interval to obtain a standard grid; on the basis of a standard grid, independently stacking observation data of each observation station to obtain a signal-to-noise ratio grid model corresponding to each observation station; performing preferential fusion according to the signal-to-noise ratio grid models of all the observation stations to obtain a high-quality grid model; performing data comparison on the signal-to-noise ratio grid model of each observation station and the high-quality grid model to obtain an available data area of each observation station; and performing quality control on the current observation data according to the available data area to obtain high-quality data. According to the invention, the observation area which is not subjected to interspersed shielding interference can be identified, and data quality control is realized.

The invention discloses a carbon emission metering method and metering system based on multi-source remote sensing and multi-modal learning, and relates to the technical field of carbon emission prediction, and the method comprises the following steps: processing multi-source remote sensing data through a data processing module; performing feature extraction on the multi-source remote sensing data through a multi-modal depth modeling module; carrying out multi-modal fusion on the extracted features through a multi-modal depth modeling module; carbon emission estimation is carried out through a multi-modal deep modeling module; outputting a carbon emission estimation result in real time through a carbon emission report module; by combining the carbon satellite data and the power emission data, the carbon emission data can be calculated and output in real time, the accuracy of estimating the carbon emission of the power plant is improved, environmental monitoring, policy making and carbon market management are supported, and powerful data support is provided for coping with climate changes and carbon emission reduction policies.

The invention relates to the technical field of meteorological observation, in particular to a solar radiation prediction method and device based on multi-source satellite data fusion, an electronic device and a storage medium. And an end-to-end deep learning prediction model is constructed, so that a dark start prediction blind area caused by visible light deficiency before and after sunrise is effectively overcome, continuous prediction in all days, especially in a dawn period, is realized, and the prediction coverage capability is expanded. And secondly, regional adaptability parameter calibration is carried out on the physical inversion model, so that the localization precision of a historical reference radiation field is improved, and a more reliable training basis is provided for the prediction model. Finally, the output solar radiation prediction product with high temporal-spatial resolution and physical consistency can directly support key businesses such as photovoltaic power station output prediction and power grid climbing management, and the practicability and reliability of the prediction result are enhanced.

The invention relates to the technical field of satellite-borne edge computing, and provides a satellite-borne distributed efficient data distribution system which comprises a computing processing module which is provided with a DDS and is used for processing satellite-borne computing tasks; the containerization module is used for packaging the application program and the DDS into a container; the dynamic unicast routing module is used for carrying out data routing among heterogeneous equipment and planning a data transmission path so as to deal with the situation that links are uncertainly connected and disconnected in a complex space environment; the ddrouter virtual multicast routing module is used for realizing forwarding of multicast routing by software and overcoming the dependence of a DDS service discovery stage on a hardware multicast routing function under a multi-hop network; the QoS configuration module is used for flexibly configuring QoS and multiple transmission modes to adapt to data stream transmission requirements under different task scenes; and the distributed consistency guarantee module is used for keeping the satellite data consistency through leader election and a log replication mechanism. The method aims to solve the problems of low data distribution efficiency, poor consistency, difficulty in heterogeneous adaptation and the like in a satellite-borne edge computing scene.

The invention discloses a severe convective cloud cluster identification method and system based on satellite remote sensing data, and relates to the field of meteorological remote sensing and disastrous weather monitoring, and the method comprises the steps: carrying out the time-space registration of multi-spectral satellite data to form a multi-channel time sequence, introducing an environment parameter matched with the multi-channel time sequence, and carrying out the recognition of a severe convective cloud cluster; calculating an adaptive dynamic threshold to stably extract candidate cloud cluster objects; fusing the multi-spectral visual features and the environmental physical features to construct a convection complexity index, and generating node features; and further constructing a graph structure by taking the candidate cloud clusters as nodes and taking spatial proximity as edges, inputting a physical constraint graph convolution model, performing gating constraint by environmental physical characteristics in neighborhood aggregation, and outputting the severe convection probability, severity score and geographical range of each cloud cluster. Therefore, by adopting a fusion mechanism of environmental physical priori and graph structure reasoning, the adaptation capability and the output consistency of the model under different climate backgrounds are improved, and the reliability and the interpretability of an identification result are enhanced on the whole.

The embodiment of the invention provides a satellite navigation ranging link integrity monitoring method and device, a computer readable storage medium, an electronic device and a computer program product, and the method comprises the steps: screening out at least one first satellite meeting a preset monitoring condition from a preset satellite set according to satellite data broadcasted by a satellite-based augmentation system; calculating a first deviation of a satellite signal of the first satellite at the first site according to the satellite data; calculating a second deviation of the satellite signal of the first satellite at the first station according to the actual observation data of the first station; and determining a consistency index of the first satellite according to the first deviation and the second deviation, and determining whether the integrity of the navigation ranging link of the first satellite reaches the standard based on the consistency index. Therefore, signal anomalies are captured on different scales, limitation of a single monitoring method is avoided, monitoring precision is improved, and accurate monitoring of integrity of a satellite navigation ranging link is realized.

The invention provides a prior information constraint-based Antarctic ice cover material balance inversion method and device, and the method comprises the steps: obtaining satellite height measurement data, and extracting an elevation parameter of an Antarctic ice cover region; preprocessing the gravity satellite data; performing inversion material balance of physical priori constraint, including dividing the Antarctic ice cover into Mascon units which are geographically aligned with the ice flow basin, constructing a regularization constraint matrix of space differentiation based on elevation variation variance of satellite altimetry data, dynamically optimizing regularization parameters in combination with generalized cross validation criterion, and solving mass change; and outputting the Antarctic ice cover material balance result. According to the method, regularization is carried out by using the uncertainty difference of the quantized area of the height measurement data, so that the inversion precision and the spatial resolution are remarkably improved, and GIA model errors and data noise interference are effectively inhibited.

The invention relates to the technical field of data fusion, and particularly provides a multi-scale eco-hydrological parameter joint inversion method and system based on unmanned aerial vehicle and satellite-ground data fusion, and the method comprises the steps: obtaining multispectral, hyperspectral and thermal infrared data of a research region through a high-resolution sensor carried by an unmanned aerial vehicle; the method comprises the following steps: acquiring unmanned aerial vehicle data, acquiring satellite remote sensing data, performing radiation correction on the unmanned aerial vehicle data and the satellite data, and performing feature extraction and space-time fusion on the unmanned aerial vehicle data and the satellite data by using a deep learning algorithm to generate high-resolution fusion data; through a multi-resolution analysis or spatial interpolation technology, coordinating differences of unmanned aerial vehicle data and satellite data in space and time scales, and constructing a multi-scale data set; based on a physical model and a data driving model, combined with fusion data, an inversion result is visually displayed, and the method has the effects of meeting the hydrological monitoring requirement of high time resolution, and particularly providing real-time monitoring data in the hydrological process of rapid change.

The invention relates to the field of meteorological prediction, in particular to a multispectral-based all-day sea fog detection method, and solves the problems of failure of a morning and evening period detection algorithm, insufficient sea fog and low cloud distinguishing precision, split algorithm flow and lack of systematicness in the current sea fog detection field. The method specifically comprises the following steps: acquiring satellite data, preprocessing the satellite data, and calculating a solar zenith angle SZA of each pixel; distributing each pixel to a corresponding processing mode according to the value of the SZA; on the basis of the distributed processing mode, adaptive spectrum preliminary screening of the corresponding mode is executed, and a preliminary screening binary mask is generated; based on the processing mode, performing regional texture consistency analysis on the primarily screened binary mask to effectively relieve low cloud interference; and performing post-processing optimization on the reserved sea fog area, and outputting a final sea fog detection result. High-precision and high-stability detection of all-day sea fog can be achieved, good performance is still kept in the morning and night transition period, and sea fog and low clouds can be effectively distinguished.

The invention discloses a high-resolution tomographic SAR four-dimensional imaging method for space-time baseline decoupling processing, and the method comprises the following steps: 1, reconstructing original satellite data into uniform baseline observation data, and solving the noise-free estimation of the uniform baseline observation data through the low-rank characteristic of a Hankel matrix; step 2, inverting elevation information based on noise-free estimation, taking the elevation information as priori, constructing a deformation spectrum estimation dictionary through space-time decoupling, and performing deformation information estimation; and step 3, matching the estimated elevation and deformation quantity by using a maximum likelihood method so as to realize four-dimensional imaging. According to the method, the algorithm complexity of four-dimensional imaging is reduced exponentially while high-precision imaging is realized, and differential tomography SAR four-dimensional imaging result verification is carried out based on domestic land exploration No.1 satellite data.

The invention discloses a deep learning CO2 near-real-time inversion method and system fused with multi-scale characteristics, and the method comprises the steps: inputting the observed spectral data into an inversion model, carrying out the in-orbit near-real-time calculation, and outputting a satellite data inversion result; the inversion model comprises a two-stage full connection projection, a Transform encoder, a learnable attention pooling and a regression decoder which are connected in sequence; wherein each layer of the encoder comprises two sequentially connected sub-layers, namely a multi-head self-attention layer and a feedforward network layer, a residual connection layer and a normalization layer are introduced behind each sub-layer, in the multi-head self-attention layer, rotation position coding is applied to Q and K before dot product attention is entered, and relative displacement is implicitly injected into attention weight calculation through phase rotation; according to the inversion method, satellite observation is taken as a characteristic, and ground in-situ observation is taken as a supervision CO2 data driving inversion thought, so that the calculation efficiency is maintained, and meanwhile, the modeling capability of a nonlinear relationship between characteristics is improved.

The invention discloses an earth surface deformation monitoring and early warning method based on unmanned aerial vehicle and satellite data fusion, and relates to the technical field of remote sensing image processing. The method comprises the following steps: S1, acquiring remote sensing data; s2, preprocessing remote sensing data; s3, constructing an earth surface three-dimensional point cloud model; s4, ground deformation monitoring and geological disaster judgment; s5, disaster scale and trend analysis; and S6, issuing early warning and taking measures. Based on time sequence three-dimensional point cloud data fused by an unmanned aerial vehicle and a satellite, through multi-feature boundary optimization, multi-algorithm weighted fusion, multi-source data verification, smooth terrain emphasizing on triangular prism method improvement and complex terrain emphasizing on TIN difference method, sparse data is supplemented in combination with space-time weight interpolation, and scene limitation of a single algorithm is avoided.

The invention provides a FY-4 offshore cloud bottom height correction method considering a buoy ceilometer, and belongs to the technical field of offshore cloud bottom height correction.According to the method, the cloud bottom height data of a buoy ceilometer and a FY-4 satellite are collected, sea surface reflection pollution correction is carried out on the satellite data, and a cloud layer structure is identified by using K-means clustering; extracting multi-scale fractal dimension features of buoy and satellite data and calculating a difference parameter, organizing correction data into a four-dimensional correction tensor, decomposing and extracting a low-rank structure by adopting CANDECOMP, and inputting a decomposition result and the fractal dimension difference parameter into a cloud base height bidirectional optimization model; dynamic balance between precision improvement and space-time stability maintenance is realized through a game mechanism of a precision optimization sub-model and a stability optimization sub-model in the model, and the technical problem that the marine cloud bottom height precision and the space-time stability of satellite inversion are difficult to consider at the same time is solved.

The invention relates to the technical field of side slope deformation space-time prediction and multi-source data fusion, in particular to a side slope deformation space-time prediction method and device based on multi-scale data fusion. The method comprises the following steps: acquiring slope synthetic aperture radar interference data and global navigation satellite data, determining a measuring point displacement time sequence and coordinates according to the data, and aligning a displacement direction with a time sequence window; extracting time sequence features of a displacement time sequence in the global navigation satellite data and generating a feature matrix, and fusing the feature matrix with the synthetic aperture radar interference data to obtain a feature fusion matrix; an intra-set adjacency matrix of the former measuring point is constructed according to the measuring point coordinates, an inter-set adjacency matrix of the two types of data measuring points is constructed through a Thiessen polygon graph, and the intra-set adjacency matrix and the inter-set adjacency matrix are fused to form a fused adjacency matrix. And based on the feature fusion matrix and the fusion adjacency matrix, aggregating space neighborhood information to generate a displacement time sequence prediction result of the slope global measurement point. Therefore, the problems of insufficient space-time correlation excavation, low slope global deformation prediction precision and the like in related technologies are solved.

The invention discloses a multi-spectrum and SAR cooperative sensing sea reclamation increment remote sensing detection method, and belongs to the technical field of land change remote sensing monitoring. According to the scheme, multi-spectrum and SAR satellite data are fused, the data advantages of optical images and synthetic aperture radar images are effectively combined, and through the wave band operation and threshold segmentation technology, the real-time performance of the sea reclamation increment is improved. According to the method, more detailed and reliable sea reclamation pattern spot information can be obtained, the spatial distribution and dynamic change of suspected sea reclamation pattern spots can be accurately identified, and the limitation that traditional optical remote sensing is interfered by cloudy and rainy weather is effectively overcome. The method is used for a computer program to execute and complete multi-source data fusion, change detection and sea reclamation suspected pattern spot detection and analysis, and can be widely applied to the fields of coastal zone ecological environment protection, coastal city planning management, ocean resource sustainable utilization, relevant policy making and the like. And a core technical support is provided for full-process monitoring and scientific decision making of the reclamation activity.

The invention discloses a satellite multi-physics field prediction method and system based on structured self-supervised training. The method comprises the steps of providing a satellite data set, providing a progressive prediction network model of a branch-trunk structure, and executing structured self-supervised training. And the progressive prediction network model takes the high-fidelity reference field as a starting point, plans a path according to the parameter variation amplitude, and iteratively outputs a physical field prediction result. Meanwhile, a high-fidelity external reference field can be accessed in the reasoning process according to a new working condition or a high-reliability demand scene, and the local prediction precision is improved. The system comprises a data preparation module, a model training module and a progressive prediction module. According to the method and the system, the prediction speed and the prediction precision of satellite multi-physical field prediction can be remarkably improved, and the core technical problem of error accumulation in a long-range and multi-step prediction task of an existing AI agent model is solved.

The invention discloses a lightning-to-ground lightning activity track correction method and device, electronic equipment and a storage medium, and is used for solving the technical problem that a current lightning-to-ground lightning activity track determination method depends on a single lightning positioning data source and is low in precision. The method comprises the following steps: acquiring meteorological data of a target area and lightning positioning data for lightning ground lightning activities in the target area; wherein the meteorological data comprises satellite data and radar data; determining a lightning initial centroid trajectory according to the lightning positioning data; generating a first centroid correction track corresponding to the satellite data, and generating a second centroid correction track corresponding to the radar data; evaluating the first centroid correction track and the second centroid correction track according to a pre-constructed track evaluation model, and outputting a track correction strategy; and correcting the lightning initial centroid trajectory according to a trajectory correction strategy to obtain a lightning corrected centroid trajectory of the lightning ground-to-ground lightning activity in the target area.