55 results about "Gravity gradient" patented technology

The invention relates to the technical field of fault identification, in particular to a three-dimensional fault occurrence intelligent identification method and system based on multi-source remote sensing, and the method comprises the steps: obtaining a high-resolution optical image, a penetrating radar profile and gravity gradient data of a target region, and generating a three-dimensional geological response body fusing earth surface and underground information; inputting the three-dimensional geological response body into a vertical adaptive enhancement module, and constructing a unified structured fault feature tensor; and inputting the fault feature tensor into a dynamic curvature constrained curved surface reconstruction module to generate a fault space curved surface, and resolving trend, tendency and dip angle change functions of the curved surface at different depths. According to the three-dimensional fault occurrence intelligent identification method based on multi-source remote sensing, the high-resolution optical image, the penetrating radar profile and the gravity gradient data are fused, a structured three-dimensional geological response body is constructed, the cooperative expression ability of earth surface-underground fault information is remarkably improved, and the fault occurrence identification accuracy is improved. And high-fidelity modeling of complex fault bending and abrupt change areas is realized.

The invention relates to a geological fracture recognition method based on multi-source geophysical data and a computer system, and solves the problem of high-precision recognition of tiny fractures and complex fracture zones. The method comprises the following steps: acquiring gravitational field model data, full-tensor gravity gradient data and aeromagnetic data of a target area, and carrying out gridding processing, pole conversion processing and edge expansion preprocessing on the data; performing boundary identification calculation on the preprocessed data by adopting a traditional physical field multi-derivative algorithm to generate a first group of boundary information A; carrying out boundary recognition calculation on the preprocessed data by adopting a gravity and full-tensor gravity gradient combined derivative algorithm (an enhanced total horizontal gradient derivative algorithm, a combined derivative algorithm and a combined oblique derivative algorithm) to generate a second group of boundary information B, and carrying out boundary recognition calculation on the preprocessed data by adopting an image edge detection algorithm to generate a second group of boundary information B; generating a third group of boundary information C; and carrying out weighted fusion on the three groups of boundary information to obtain a final fracture identification result.

The invention discloses a spacecraft in-orbit operation health assessment method, and relates to the technical field of space flight and aviation. Environmental data such as gravity gradient, radiation intensity and dust particle density of an asteroid surface are collected in real time through a multi-mode sensor array, modeling and prediction are carried out by adopting a dynamic Bayesian network in combination with Kalman filtering, and the health assessment accuracy of the spacecraft in-orbit operation is improved. The method comprises the following steps: dynamically generating environment characteristic parameters including a risk level and an environment change trend, and in health state prediction, simulating stress distribution in a low-gravitation environment through digital twinborn body and finite element analysis, predicting a fatigue state and a failure risk of a contact device, and accurately evaluating the fatigue life of a high-stress area. The quantification capability of the failure risk is improved through the risk probability model; besides, according to a health prediction result, landing parameters including a damping mode and a landing speed are dynamically optimized, and a dynamic planning method is used for minimizing impact force and stress load, so that the detector can be in stable contact in a low-gravity environment.

The invention belongs to the technical field of gravity data noise reduction, discloses a gravity data noise reduction method based on a U-Net network, and provides a U-Net network structure fusing a multi-scale feature fusion module and an attention mechanism aiming at the noise reduction problem of gravity and gravity gradient data. And a physical consistency constraint is introduced in a model training process. A Laplace equation met by the gravitational field potential in geophysics is used as prior information to be fused into a loss function, and a Laplace item is designed. The operation effectively improves the physical interpretability and generalization ability of the noise reduction result of the neural network. Meanwhile, the network structure adopts multi-scale feature fusion and an attention mechanism, and information of different spatial scale features and key spatial positions in complex geological signals is better captured.

The invention discloses a mountainous area surface matrix physical property prediction method, and relates to the technical field of geological exploration, and the method comprises the steps: obtaining mountainous area surface point cloud data, separating vegetation from ground points through a point cloud classification algorithm, and generating a digital elevation model and a digital surface model; based on the digital elevation model and the digital surface model, a quantum gravity gradiometer is adopted to move along a contour line for measurement, a gravity gradient tensor is obtained, terrain interference signals are eliminated, and a vertical density profile is obtained through inversion; identifying a density abnormal region according to the vertical density profile, acquiring an undisturbed soil column sample by adopting a drilling unit, determining a grain composition and moisture characteristic curve, performing DNA extraction and microorganism sequencing on the undisturbed soil column sample, and calculating the relative abundance of the nitro spirillum phylum. According to the method, the dynamic coupling model is constructed through fusion of quantum gravity gradient measurement and microbiological omics data, and the problem of insufficient vertical density-permeability coefficient coupling modeling precision of a traditional method is solved.

The invention discloses a method for monitoring the relative gravity of surface subsidence in a coal mining process, and the method comprises the steps: arranging a quantum gravity base station at the periphery of a subsidence area, deploying a relative gravimeter in a basin according to a grid, arranging a fiber grating gravity sensing chain in an underground crossheading, and carrying a cold atom gravity gradiometer by adopting an unmanned plane for periodic scanning. A gravitational field-mass migration model is established based on real-time data, coal mining volume, rock stratum density change and earth surface gravity anomaly are associated, geological background interference is eliminated through wavelet multi-scale decomposition, and a gravity change value is optimized through hydrological correction in combination with water level and porosity. And when the corrected gravity change exceeds a threshold value, calculating a rock stratum fracture azimuth angle by using a gravity gradient tensor, inverting an equivalent mining thickness based on the gravity change, dynamically predicting a surface subsidence form in combination with a subsidence basin model, constructing a state vector in combination with gravity, gradient and deformation data, and realizing multi-source fusion prediction through a filtering algorithm.

The invention belongs to the technical field of intelligent management, and discloses an unmanned aerial vehicle intelligent flight control system for low-altitude traffic management. The data acquisition module is used for acquiring gyroscope data, gravity gradient data and atomic clock time; the reference construction module is used for analyzing and obtaining initial positioning coordinates through a gravity matching positioning algorithm based on the gravity gradient data and the dynamic gravity anomaly graph; the coupling positioning module is used for analyzing according to the initial positioning coordinate, the gravity gradient data and the environment data to obtain a positioning error vector; the communication disaster recovery module is used for analyzing according to the initial positioning coordinate and the positioning error vector to obtain a gravity positioning result; the path re-planning module is used for analyzing according to a pre-stored gravity anomaly graph and a gravity positioning result to obtain an optimized flight path; according to the method, the problem of positioning failure caused by GPS signal loss in a strong interference environment is effectively solved, and high-precision positioning independent of GPS, real-time error correction and optimal path planning adaptive to a gravity field are realized.

The present application relates to a kind of rotation accelerometer gravity gradiometer attitude error post-processing compensation method, comprising the following steps: step 1, original motion data is acquired using the inertial measurement unit fixed with gravity gradient sensor, and the positioning and speed information provided by radio navigation are combined to complete original data acquisition;Step 2, the attitude error of gravity gradient sensor in the measurement process is optimally estimated;Step 3, based on the coordinate transformation matrix of gravity gradient tensor under different coordinate systems, the error compensation of gravity gradient measurement value is carried out, and finally the gravity gradient measurement data after attitude error compensation is obtained.The present application can be used in dynamic measurement environment, and the measurement information of carrier motion is measured using gyroscope and accelerometer, the attitude error of gravity gradient measurement component is estimated and compensated offline with high precision, and then the dynamic measurement precision of rotation accelerometer gravity gradiometer is improved.

The present invention discloses a method for generating a dynamic coupling model. The method comprises setting benchmark points and arranging observation piers; transmitting benchmark gravity values ​​to the observation piers; collecting spatial pose change data and three-dimensional coordinate data of markers; obtaining relative gravity joint measurement data through relative gravity joint measurement; calculating the three-dimensional displacement of each observation pier based on the spatial pose change data; fusing the benchmark gravity values, relative gravity joint measurement data, and three-dimensional coordinate data to construct a gravity gradient field tensor model; systematically integrating the three-dimensional displacement, gravity gradient field tensor model, and three-dimensional coordinate data to form a spatial correlation framework; and performing a joint adjustment based on a conditional least squares model to constrain the physical coupling mechanism of displacement and gravity gradient to generate a dynamic coupling model of the interaction between the deformation field and the gravity field. The method achieves the unification of the spatiotemporal benchmarks of displacement and gravity data, eliminates rotation errors and local interference, enhances the reliability of the gravity gradient field, and physically constrains the coupling mechanism of deformation and gravity.

The invention belongs to the technical field of uranium ore geological exploration, and particularly relates to a sandstone uranium ore excretion structure authenticity discrimination method based on gravity gradient deep constraint, which comprises the following steps: step 1, deep fracture ridge extraction based on gravity data; 2, constructing a'deep part-earth surface 'structure projection buffer area; 3, removing a micro-landform catchment area based on a DEM (Digital Elevation Model); 4, extracting earth surface remote sensing abnormity; and step five, space coupling and excretion zone authenticity discrimination are carried out. According to the method, by constructing a deep fracture-earth surface projection buffer area model and combining with micro-landform analysis, earth surface non-structural ponding and salt accumulation interference is effectively eliminated, and an effective discharge zone with a deep root is accurately locked.

The invention discloses a three-component optical fiber Sagnac gravity gradiometer and a three-component gravity gradient measurement method. The three-component optical fiber Sagnac gravity gradiometer comprises a joint resolving unit and at least two optical fiber Sagnac gravity gradiometers, the optical fiber Sagnac gravity gradiometer comprises a gravity gradiometer body and a Sagnac interferometer, and an optical fiber ring is installed on a rotor of the gravity gradiometer body and used for obtaining the angular acceleration when the rotor rotates. Rotors of the two optical fiber Sagnac gravity gradiometers are parallel to the central axis of the ground and are orthogonal, and the two optical fiber Sagnac gravity gradiometers are used for respectively generating rotary motion under the action of different components of the gravity gradient; and the resolving unit is used for resolving three gravity gradient components according to the angular accelerations acquired by the two optical fiber Sagnac gravity gradiometers. According to the invention, the three diagonal components of the gravity gradient can be measured at the same time.

The invention relates to the technical field of oil and gas resource exploration data processing, in particular to a three-dimensional gravity gradient inversion method for oil and gas resource exploration, a medium and equipment. The method comprises the following steps: acquiring gravity gradient plane two-dimensional observation data of an exploration target area; setting an inversion constraint function; and performing inversion iterative operation to obtain an underground three-dimensional density structure, and delineating a space boundary of a salt hill, a rock cover or an oil storage structure according to the structure. According to the method, a regularization objective function is constructed based on the L1 norm of the model and the total variation TV1 of the model, an L1-TV1 weight factor is introduced and reasonably selected, and a density model vector after depth weighting is used in the total variation of the model. The method can effectively overcome the shielding effect of the shallow rock cover on the deep salt nucleus in salt hill detection, improve the depth resolution and the boundary depiction definition, reduce the inversion multiplicity of solutions, and provide accurate guidance for oil-gas exploration and drilling risk avoidance.

The application relates to a configuration design method of a three-dimensional orbiting sail with posture stability, and belongs to the field of spacecraft attitude dynamics and structure design.The application is realized by the following method: considering the environmental perturbation of atmospheric resistance, atmospheric resistance moment and gravity gradient moment, an orbit attitude coupling dynamics model of a three-dimensional orbiting sail system is established based on a position vector and Euler angles; the three-dimensional orbiting sail and a spacecraft body are regarded as rigid bodies, and the windward area is considered to be blocked by airflow so as to more accurately describe the orbit movement and attitude movement of the system in the orbiting process; the attitude movement is simplified for the attitude dynamics of the system; the attitude stability of the system in the orbiting process is analyzed by using a phase plane method, a Lyapunov energy function method and other stability analysis methods of nonlinear dynamics systems; the relationship between important structure parameters such as a cone angle and a support rod length of the three-dimensional orbiting sail and the attitude stability is further constructed; and the three-dimensional orbiting sail satisfying the preset attitude stability requirement is obtained based on the relationship, so that the configuration design is realized.

The invention belongs to the technical field of precision measurement, and relates to an indoor movable base testing device for a horizontal tensor gravity gradiometer and application of the indoor movable base testing device. The device comprises a control processing module, and a gravity gradient standard field generation module and a motion simulation module which are respectively connected with the control processing module, the motion simulation module and the gravity gradient standard field generation module are both fixed on a foundation, and the gravity gradient standard field generation module comprises an electric displacement table; a pair of attraction source assemblies arranged in an orthogonal mode is fixed to the electric displacement table. The horizontal tensor gravity gradiometer to be measured is arranged on the motion simulation module and located between the pair of gravitation source assemblies; the control processing module is used for driving the motion simulation module to simulate the environment of the movable base, and is also used for enabling the gravitation source assembly to translate on the electric displacement table and generating a gravity gradient standard field dynamically matched with the motion of the gravitation source assembly. According to the invention, the performance verification of the horizontal tensor gravity gradiometer under the condition of simulating the movable base can be realized, and the test cost and period are reduced.

The present application relates to a kind of gravity gradient satellite segmented optimal attitude tracking control method, the present application is aimed at satellite platform, coiled stretch arm and the gravity gradient system of end payload three compositions, adopt dumbbell model to describe its swing motion in coplanar orbit transfer, establish the attitude dynamics model of gravity gradient satellite;Orbit dynamics model is established based on electric propulsion coplanar orbit transfer.Segmented optimal trajectory of gravity gradient satellite under the action of electric propulsion is obtained by designing optimal controller, while realizing the orbit transfer between two coplanar circular orbits, the stability and robustness of system are enhanced.On this basis, based on finite time tracking error function, a kind of segmented adaptive terminal sliding mode control method is designed, the problem of finite time convergence and uncertainty boundary is solved.The present application is specific object in the gravity gradient satellite of coplanar orbit transfer, the control method proposed is conducive to the improvement of control performance, easy to realize.

The invention discloses a gravity gradient layering-overall inversion method based on a radial power spectrum, and relates to the technical field of gravity gradient inversion. The method comprises the following steps: acquiring gravity gradient data of a target area, and preprocessing the gravity gradient data; performing wavelet multi-scale decomposition on the preprocessed gravity gradient data to obtain wavelet detail signals of different orders; performing burial depth estimation on the wavelet detail signals of different orders to extract depth features of sources corresponding to the signals of each order; and performing layered-overall simulated annealing inversion based on the depth features of the source corresponding to each order of signal, and outputting an inversion result. According to the method, the radial power spectrum method can be applied to the gravity gradient signals, the problem of incomplete signal separation caused by wavelet multi-scale analysis is effectively solved, and the inversion accuracy is improved.

The invention discloses an edge detection filter based on a TANH function and an arcsine function, and the filter determines Bouguer gravity anomaly data through obtaining free air gravity anomaly data, terrain correction values and Bouguer correction values of a plurality of gravity points of a research region. Calculating Gaussian envelope gradient data based on the gravity gradient tensor component data and Gaussian envelope, and obtaining an edge detection filtering result according to the Gaussian envelope gradient data and TANH function filtering; noise interference can be effectively suppressed by adopting a double-constraint mechanism, a generated boundary result is clear and continuous, and the problems of boundary divergence, positioning fuzziness, form distortion and the like existing in a traditional method are effectively solved; anomalous bodies with different depths are synchronously and accurately identified, stable detection performance can be maintained under complex geological conditions, and generation of false boundaries under the condition of coexistence of positive and negative anomalies is reduced; the reliable boundary recognition capability can be kept under the strong noise background, manual intervention in the preprocessing link is reduced, and the practical application convenience and the result reliability are improved.

This invention provides a method, system, and medium for predicting earthquake source locations using TLF-GV signals, belonging to the field of earthquake prediction technology. The method includes: constructing a U-Net++ model, where multiple encoding blocks are used to sequentially generate feature maps of different spatial scales based on the gradient tensors extracted from TLF-GV; multiple decoding blocks are used to decode and output earthquake source probability heatmaps based on the feature maps at different spatial scales; obtaining active fault distribution maps, and constructing a geological prior penalty term based on the distance from each grid cell in the earthquake source probability heatmap to the nearest active fault, and constructing a loss function by combining the cross-entropy loss between the probability heatmap and the actual source mask; constructing a training set, using the concatenated features of the gradient tensors extracted from the gravity gradient grid and the fault binary mask as input to train and obtain the prediction model. This method utilizes the spatial grid features of the TLF-GV gravity gradient, combined with the small-scale feature capture capability of U-Net++, to effectively improve the accuracy of earthquake source location prediction.

This invention discloses a multi-source fusion monitoring method based on gravity gradient tensor and sedimentary evolution, belonging to the field of water conservancy engineering monitoring technology, including the following steps: Data acquisition: acquiring gravity / gravity gradient data, surface geometric deformation data, and spectral data. Spatial constraint: constructing a dynamic mask using geometric deformation data to constrain the spatial boundary of gravity inversion. Physical enhancement: performing gradient processing or fusion on gravity data to obtain high-resolution mass distribution. Density correction: calculating the dynamic density of sediments over time based on a sedimentary evolution model, and correcting the background fluid density based on spectral data. Coupled solution: substituting the above parameters into the coupling equation to solve for the volume or mass of sediments. This invention can achieve long-term, continuous, non-contact, and quantitative accurate monitoring of reservoir capacity changes and sediment deposition volume through the synergistic fusion of satellite gravity observation, airborne gravity gradient tensor observation, optical remote sensing, and sedimentary physical evolution models.

The application provides a CO2 storage monitoring method and system based on space-air-ground-well multi-source information. The application combines space-air-ground-well multi-source information, comprehensively interprets multi-source data obtained by synthetic aperture radar in space, air or semi-air transient electromagnetic in the sky, gravity gradient detection on the ground, and resistivity method and microseismic monitoring in the well. Reasonable use of different monitoring technologies plays a role of mutual assistance and mutual complementation, and then CO2 migration in the injection well and the stratum can be effectively monitored.

The application discloses a shaft elevation transfer method based on gravity measurement, comprising the following steps: arranging an upper known point and a lower to-be-measured point in a shaft, and obtaining a measured gravity value and a normal height of the upper known point; calculating the normal gravity of the upper known point according to the geodetic coordinates of the upper known point, and obtaining the disturbing gravity of the upper known point according to the measured gravity value and the normal gravity; setting auxiliary points above the plumb of each of the upper known point and the lower to-be-measured point, and obtaining the disturbing gravity gradient of each measuring point according to the gravity observation value of each measuring point and the corresponding auxiliary point; constructing a height difference transfer constraint model based on the disturbing gravity and the disturbing gravity gradient of the upper known point and iteratively solving to obtain the geodetic height difference between the upper known point and the lower to-be-measured point; and calculating the normal height difference between the upper known point and the lower to-be-measured point based on the geodetic height difference, and determining the normal height of the lower to-be-measured point according to the normal height of the upper known point and the normal height difference.

The invention discloses a multi-source submarine topography reconstruction method and system based on a frequency domain attention neural network, and relates to the technical field of submarine topography reconstruction. The method comprises the following steps: acquiring GEBCO water depth data, ocean gravity parameter abnormal data, vertical gravity gradient data, north component data of vertical deviation and east component data of vertical deviation, and constructing a five-channel input tensor; performing standardization processing on the five-channel input tensor, constructing a data set, and randomly dividing the data set into a training set and a test set through uniform sampling; processing the training set by using a frequency domain attention model to obtain fusion features, and inputting the fusion features into a predictor to generate a water depth estimated value; wherein the frequency domain attention model combines fast Fourier transform preprocessing with an attention mechanism, and the predictor is composed of a full connection layer and a ReLU activation function. According to the invention, the Fourier transform mechanism is introduced into the neural network architecture, the attention in the frequency domain is realized, and the submarine topography reconstruction precision is improved.

The invention provides a three-dimensional gravity gradient tensor inversion method with physical information constraints. Firstly, the coordinate system is established according to the gravity gradient tensor data and the underground space is divided, and the coordinates of the center point of the cube are extracted and normalized; then, a seven-layer fully connected neural network is constructed to output the network output parameters; an then the gravity gradient tensor data are calculated on the observation surface, the objective function is constructed by observation data and calculation data, including a data term and a regularization item; and finally, the objective function is automatically differentiated to update the network parameters and optimize the network output parameters until the maximum training times are reached. The invention overcomes challenges in existing methods, including over-reliance on initial parameters, over-simplification of the model and assumptions, insufficient ability to deal with nonlinear relationships, and poor generalization performance.

This invention provides a method for determining fault dip using gravity data. The method calculates the total horizontal gravity gradient based on the Bouguer gravity anomaly values ​​at various gravity measurement points above the fault and plots a total horizontal gravity gradient profile. A semi-extreme line segment is then drawn on the profile, and the direction of the vector from the maximum point to the midpoint of the semi-extreme line segment is taken as the fault dip. This method for determining fault dip using gravity data is a simple, fast, and intuitive research approach. It enables the rapid acquisition of fault dip information using an economical and efficient gravity exploration method, providing a reliable basis for studying fault-related geological structures.

The application discloses a gravity and gravity gradient integrated measuring device and method based on atomic interference, and the measuring device comprises a Raman laser emitter and a structural framework. Raman laser emitted by the Raman laser emitter passes through an upper atomic interferometer and a lower atomic interferometer in sequence and then is vertically incident on a Raman laser reflector. The bottom of the Raman laser reflector is fixedly connected with the top of a seismometer through an adapter, and the structural framework is installed on a rotating platform. The measuring method comprises the following steps: firstly, the orientation angles corresponding to the maximum value and the minimum value of the gravity gradient are determined; at a new measuring point, the orientation angle of the measuring device is adjusted to the orientation angles corresponding to the maximum value and the minimum value of the gravity gradient; the original gravity measurement value of the upper atomic interferometer, the original gravity measurement value of the lower atomic interferometer and the original measurement value of the gravity gradient are obtained and corrected. The measuring repeatability problem caused by the Coriolis error in the measuring process is effectively solved.

This invention is applicable to the fields of gravimetry and gravity exploration, providing a spatial domain multi-parameter analytical forward modeling method for gravity using a rotating rectangular prism. The method includes the following steps: establishing an observation coordinate system and a local coordinate system of the source; providing parameters for the rectangular prism model and the observation system; defining the attitude of the rotating prism using Euler angles, determining the rotation matrix, and completing the coordinate transformation; calculating the analytical solutions for the three-directional vectors of the gravity field and the analytical solutions for each component of the gravity gradient tensor in the spatial domain of the source local coordinate system; and deriving the three-directional vector data of the gravity field in the observation coordinate system, as well as the data for each component of the gravity gradient tensor in the observation coordinate system. This invention is of great significance for constructing high-precision, full-space multi-parameter gravity field models. It offers high accuracy and fast computation, and can be flexibly applied to gravity field numerical simulation, qualitative analysis of gravity data, and gravity field modeling, effectively serving the teaching and research fields of gravimetry and gravity exploration.

The invention relates to a gravity gradiometer comprising at least two vehicles (V1, V2), each comprising an inertial measurement unit (UMI1, UMI2) and a control unit (COM1, COM2) for controlling the vehicles along paths having separate segments for performing inertial-measurement-unit reversals and coupled segments for exchanging inertial measurements. At least one of the electronic control units is arranged so to determine, in particular from the exchanged measurements: a transformation matrix between the inertial reference frames, a correction of the linear-acceleration measurements based on an observed accelerometric model, a correction of the angular-velocity measurements based on an observed gyrometric model, and a gradient of the gravitational field. The invention also relates to a navigation system comprising such a gravity gradiometer.

The invention belongs to the technical field of airborne gravity measurement, and particularly relates to an airborne gravity gradient measurement data multi-source error modeling method, which comprises the following steps: constructing a transformation matrix of a measurement coordinate system and a carrier coordinate system; correcting the transformation matrix to obtain an error-containing transformation matrix; converting the gravity gradient tensor under the carrier coordinate system into a gravity gradient tensor under a measurement coordinate system by using the error-containing transformation matrix; converting the vibration error from a carrier coordinate system to a measurement coordinate system by using an error-containing transformation matrix; calculating a self-gradient error through mapping based on the mass distribution of each part of the airplane; describing random noise of a measurement system by adopting a Gaussian white noise model; and adding the gravity gradient tensor under the measurement coordinate system, the vibration error under the measurement coordinate system, the self-gradient error and the random noise of the measurement system to obtain a multi-source error model. Weak noise interference can be effectively detected, the approximation degree of an output result and actual measurement data is high, and the aviation gravity gradient measurement precision is greatly improved.

The invention provides a TLF-GV signal correlation seismic source position prediction method and system and a medium, and belongs to the technical field of seismic forecasting, and the method comprises the steps: constructing a U-Net + + model, and enabling a plurality of coding blocks to be used for generating feature maps of different spatial scales according to gradient tensors extracted by TLF-GV in sequence; the plurality of decoding blocks are used for decoding and outputting a seismic source probability thermodynamic diagram according to the feature maps of different spatial scales; obtaining an active fault distribution diagram, constructing a geological priori penalty term according to the distance from each grid unit in the seismic source probability thermodynamic diagram to the nearest active fault, and constructing a loss function in combination with cross entropy loss between the probability thermodynamic diagram and a real source mask; and constructing a training set, taking gradient tensor extracted by the gravity gradient grid and fault binary mask splicing features as input, and training to obtain a prediction model. According to the method, the spatial grid features of the TLF-GV gravity gradient are utilized, and the small-scale feature capture capability of U-Net + + is combined, so that the accuracy of seismic source position prediction is effectively improved.

The invention discloses an external coincidence precision evaluation method and equipment for a gravity gradient vertical horizontal component, and the method comprises the steps: enabling a shipborne cold atom absolute gravimeter and a full-tensor gravity gradiometer to be installed on the same ship, controlling a measurement ship to sail at a constant speed along a preset measurement line in the north-south direction or the east-west direction, and carrying out the synchronous measurement, obtaining gravity vertical component data output by a shipborne cold atom absolute gravimeter and vertical horizontal component data output by a full-tensor gravity gradiometer; performing filtering processing on the gravity vertical component data and the vertical horizontal component data, and performing point-by-point calculation on the filtered gravity vertical component data and the corresponding position data by adopting a midpoint difference method to obtain a calculation value of a vertical horizontal component; and respectively counting the total number of effective measuring points of the measuring lines in the north-south direction and the east-west direction, and calculating the root-mean-square error of the vertical horizontal component data and the corresponding calculation value, the root-mean-square error being the external coincidence precision of the vertical horizontal component of the full-tensor gravity gradiometer.