CN117350171B - Mesoscale vortex three-dimensional subsurface structure inversion method and system based on double-flow model - Google Patents
Mesoscale vortex three-dimensional subsurface structure inversion method and system based on double-flow model Download PDFInfo
- Publication number
- CN117350171B CN117350171B CN202311638586.XA CN202311638586A CN117350171B CN 117350171 B CN117350171 B CN 117350171B CN 202311638586 A CN202311638586 A CN 202311638586A CN 117350171 B CN117350171 B CN 117350171B
- Authority
- CN
- China
- Prior art keywords
- sea surface
- dimension
- vortex
- mesoscale vortex
- flow model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 33
- 230000003993 interaction Effects 0.000 claims abstract description 36
- 230000004927 fusion Effects 0.000 claims abstract description 12
- 230000007246 mechanism Effects 0.000 claims abstract description 9
- 238000012549 training Methods 0.000 claims description 14
- 238000011176 pooling Methods 0.000 claims description 8
- 238000000605 extraction Methods 0.000 claims description 6
- 230000006870 function Effects 0.000 claims description 6
- 238000003860 storage Methods 0.000 claims description 5
- 238000011084 recovery Methods 0.000 claims description 4
- 238000011156 evaluation Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 3
- 238000013135 deep learning Methods 0.000 abstract description 2
- 230000008569 process Effects 0.000 description 6
- 238000013528 artificial neural network Methods 0.000 description 4
- 230000009977 dual effect Effects 0.000 description 4
- 239000010410 layer Substances 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 230000003213 activating effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000013079 data visualisation Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007123 defense Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000011065 in-situ storage Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 235000015097 nutrients Nutrition 0.000 description 1
- 238000012950 reanalysis Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 239000002344 surface layer Substances 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
- G06N3/0455—Auto-encoder networks; Encoder-decoder networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/08—Thermal analysis or thermal optimisation
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Geometry (AREA)
- Medical Informatics (AREA)
- Computer Hardware Design (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
The invention provides a mesoscale vortex three-dimensional subsurface structure inversion method and system based on a double-flow model, which relate to the field of deep learning and ocean inversion intersection, and acquire mesoscale vortex sea surface information to be inverted through satellites; inputting the mesoscale vortex sea surface information into a trained double-flow model, and inverting temperature results of the mesoscale vortex at different depths to obtain a mesoscale vortex subsurface temperature profile; the double-flow model introduces a Triplet attention attention mechanism, and adopts a three-branch structure to fuse channel attention and space attention for cross-dimensional interaction; according to the invention, inversion of the mesoscale vortex subsurface structure is realized by adopting the double-flow model, data association among sea surface parameters is explored, relationship models of different parameters and subsurface temperatures are respectively established, the characteristic relationship of multi-source information is fused, the characteristic fusion is realized, the multi-source data is effectively fused, and the inversion effect is improved.
Description
Technical Field
The invention belongs to the field of deep learning and ocean inversion intersection, and particularly relates to a mesoscale vortex three-dimensional subsurface structure inversion method and system based on a double-flow model.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
Mesoscale vortex is a common ocean phenomenon, widely exists in various sea areas, has a closed vortex structure, and has an influence depth of thousands of meters on a vertical structure; meanwhile, the mesoscale vortex contains huge energy to drive the transmission of ocean heat, salinity and nutrient substances, and bring non-negligible influence to the ocean chemistry and biological environment; the kinetic energy of the mesoscale vortex occupies 90% of the kinetic energy of the middle sea and the large sea even, the energy of the mesoscale vortex can cause the disturbance of the underwater three-dimensional sound field, the disturbance of underwater communication and detection and the influence on the safety of submarines, fishing boats and the like, therefore, the sailing safety of the underwater submarines can be effectively ensured by researching the mesoscale vortex three-dimensional structure, and powerful guarantee is provided for marine environment monitoring, national defense and military and the like.
With the development of satellite remote sensing technology, the high-resolution satellite remote sensing technology is widely applied to meteorological hydrology, and vortex feature extraction of the sea surface is mainly based on high-resolution satellite remote sensing observation data at present, but the mode can only acquire vortex surface data and can not directly detect information in the sea. Compared with remote sensing observation, in-situ observation data (such as Argo and CTD data) based on ocean can detect the profile information of the ocean subsurface, but has the problems of uneven space-time distribution, discontinuity, low spatial resolution and the like, and cannot meet the requirements of the dynamic process in the mesoscale eddy; therefore, the prior art cannot reflect the internal structural characteristics of the subsurface layer by using high-resolution satellite data, and prevents further research on a medium-scale vortex three-dimensional structure.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a mesoscale vortex three-dimensional subsurface structure inversion method and a mesoscale vortex three-dimensional subsurface structure inversion system based on a double-flow model.
To achieve the above object, one or more embodiments of the present invention provide the following technical solutions:
the invention provides a mesoscale vortex three-dimensional subsurface structure inversion method based on a double-flow model.
The inversion method of the mesoscale vortex three-dimensional subsurface structure based on the double-flow model comprises the following steps:
acquiring mesoscale vortex sea surface information to be inverted, including sea surface temperature and sea surface height, through a satellite;
inputting the mesoscale vortex sea surface information into a trained double-flow model, and inverting temperature results of the mesoscale vortex at different depths to obtain a mesoscale vortex subsurface temperature profile;
the dual-flow model introduces a Triplet attention attention mechanism, adopts a three-branch structure to fuse channel attention and space attention for cross-dimensional interaction, the first branch establishes characteristic interaction between H dimension and W dimension, the second branch establishes characteristic interaction between C dimension and W dimension, and the third branch establishes characteristic interaction between H dimension and C dimension.
Furthermore, the double-flow model adopts an encoding-decoding structure, the encoding stage is used for extracting the characteristics of sea surface temperature and sea surface height, and the decoding stage is used for characteristic fusion.
Further, the encoding stage takes the sea surface temperature and the sea surface height as inputs to explore the relationship between the sea surface temperature, the sea surface height and the subsurface temperature respectively; the method comprises the steps of carrying out feature extraction on sea surface height and sea surface temperature, simultaneously considering that a mesoscale vortex has complex nonlinear features, inevitably connecting sea surface height information and sea surface temperature information, constructing a data feature fusion network for fusing SSH and SST relationship features, and reducing the risk of overfitting by using jump connection;
and the decoding stage fuses the characteristic output of each layer of the encoding stage and realizes the recovery of resolution by utilizing deconvolution.
Further, the first branch is introduced with an identity residual branch structure, and dimension scaling is realized by fusing average pooling and maximum pooling characteristics through Z-Pool;
the second branch rotates the input feature map tensor along the W axis to realize feature interaction of the C dimension and the W dimension;
and the third branch rotates the input feature map tensor along the H axis to realize feature interaction of the H dimension and the C dimension.
Further, the double-flow model is trained by taking the constructed inversion sample library of the mesoscale vortex subsurface structure as a training data set;
and constructing the mesoscale vortex subsurface structure inversion sample library by taking the sea surface information corresponding to the mesoscale vortex time and the vortex center coordinate position as input parameters and the subsurface temperature data corresponding to the mesoscale vortex time and the vortex center coordinate position as labels.
Further, the loss function of the dual-stream model adopts MAE loss, which represents the sum of absolute differences between a true value and a predicted value, and the formula is as follows:
wherein,for predictive value +.>For a true value, n is the number of temperature fields in a set.
Further, the evaluation criteria of the dual-stream model are compared by using R2, MAE and displayed_variance_score, wherein the R2 is used for measuring the fitting degree of the predicted data, and the formula is expressed as:
wherein f is an estimated value, y is a true value,is the average value of the observed data;
the expanded_variance_score is an interpretation variance, and is used for measuring the interpretation degree of the model on the fluctuation of the data set, and calculating the interpretable variance of the true value and the predicted value, wherein the formula is as follows:
where var is the variance, y is the true value,is a predicted value.
The second aspect of the invention provides a mesoscale vortex three-dimensional subsurface structure inversion system based on a double-flow model.
The mesoscale vortex three-dimensional subsurface structure inversion system based on the double-flow model comprises a data acquisition module and a temperature inversion module:
a data acquisition module configured to: acquiring mesoscale vortex sea surface information to be inverted, including sea surface temperature and sea surface height, through a satellite;
a temperature inversion module configured to: inputting the mesoscale vortex sea surface information into a trained double-flow model, and inverting temperature results of the mesoscale vortex at different depths to obtain a mesoscale vortex subsurface temperature profile;
the dual-flow model introduces a Triplet attention attention mechanism, adopts a three-branch structure to fuse channel attention and space attention for cross-dimensional interaction, the first branch establishes characteristic interaction between H dimension and W dimension, the second branch establishes characteristic interaction between C dimension and W dimension, and the third branch establishes characteristic interaction between H dimension and C dimension.
A third aspect of the invention provides a computer readable storage medium having stored thereon a program which when executed by a processor performs the steps in a method for inversion of a mesoscale vortex three-dimensional subsurface structure based on a dual flow model according to the first aspect of the invention.
A fourth aspect of the invention provides an electronic device comprising a memory, a processor and a program stored on the memory and executable on the processor, the processor implementing the steps in a method for inversion of a mesoscale eddy three-dimensional subsurface structure based on a dual flow model according to the first aspect of the invention when the program is executed.
The one or more of the above technical solutions have the following beneficial effects:
according to the invention, through the double-flow neural network structure, the data association among sea surface parameters is discovered, the relation model of different parameters and subsurface temperature is respectively established, the characteristic relation of multi-source information is fused, the characteristic fusion is realized, the multi-source data is effectively fused, and the inversion effect is improved.
The method introduces a Triplet attention attention mechanism, adopts a three-branch structure to fuse the channel attention and the space attention, is more beneficial to realizing cross-feature dimension interaction, improves the mesoscale vortex space feature interaction capability and improves the inversion precision.
The mesoscale vortex three-dimensional subsurface structure inversion scheme developed in the invention is used for exploring the field of ocean subsurface, is more beneficial to improving the working efficiency of ocean field experts and researchers, and has practicability.
Additional aspects of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention.
FIG. 1 is a flow chart of a method of a first embodiment;
FIG. 2 is a flow chart showing the acquisition of a temperature profile of a first embodiment;
FIG. 3 is a block diagram of a dual stream model of the first embodiment;
fig. 4 is a diagram of the structure of Triplet Attention of the first embodiment.
Detailed Description
It should be noted that the following detailed description is illustrative and is intended to provide further explanation of the present application. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments in accordance with the present application. As used herein, the singular is also intended to include the plural unless the context clearly indicates otherwise, and furthermore, it is to be understood that the terms "comprises" and/or "comprising" when used in this specification are taken to specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof.
Example 1
In one embodiment of the present disclosure, a method for inversion of a mesoscale vortex three-dimensional subsurface structure based on a dual-flow model is provided, comprising the steps of:
step S1: acquiring mesoscale vortex sea surface information to be inverted, including sea surface temperature and sea surface height, through a satellite;
step S2: inputting the mesoscale vortex sea surface information into a trained double-flow model, and inverting temperature results of the mesoscale vortex at different depths to obtain a mesoscale vortex subsurface temperature profile;
the dual-flow model introduces a Triplet attention attention mechanism, adopts a three-branch structure to fuse channel attention and space attention for cross-dimensional interaction, the first branch establishes characteristic interaction between H dimension and W dimension, the second branch establishes characteristic interaction between C dimension and W dimension, and the third branch establishes characteristic interaction between H dimension and C dimension.
As an example, from the viewpoint of model training, a specific implementation of a method for inverting a mesoscale vortex three-dimensional subsurface structure based on a dual-flow model is described, and the specific process is shown in fig. 1:
step one, constructing a mesoscale vortex subsurface structure inversion sample library
In this step, a data sample is acquired, and sea surface data and subsurface data are fused. The sea surface data is derived from a Meta3.2dt data set in an AVISO official network, the spatial resolution is 0.25 degrees multiplied by 0.25 degrees, the time resolution is daily average, a certain area (0-30 degrees N,105-130 degrees E) is selected, and key information such as the longitude and latitude coordinates of the vortex center, the vortex type and time of the area are obtained; the sea surface information is derived from the Cobini Marine Environmental Monitoring Service (CMEMS), and the data products are selected as follows: global Ocean Physics Reanalysis the data ID is GLOBAL_MULTIYEAR_PHY_001_030, which has spatial and temporal resolutions of 1/12 DEG x 1/12 DEG and daily average, respectively, and the subsurface temperature profile data is derived from this data.
And (3) carrying out data processing on the acquired data, wherein the process is as shown in fig. 2, vortex center position coordinates and time information are acquired from the mesoscale vortex data of AVISO, a cyclone and anti-cyclone sample library is established, the information is utilized to correspond to the ocean surface layer information of Goobaini to realize data alignment, grids with the size of 4 multiplied by 4 are selected, sea surface height and sea surface temperature data corresponding to the grids and temperature profile data of the subsurface layer are obtained, and the mesoscale vortex subsurface structure inversion sample library is constructed.
Step two, realizing uniform resolution and default processing by data preprocessing
And (3) carrying out data preprocessing on the constructed sample library, unifying the resolution ratio to be 0.25 degrees multiplied by 0.25 degrees by using a linear interpolation method, and carrying out 0-supplementing processing on the missing values.
Step three, inputting the sample library into a double-flow model
And inputting the processed sample library into a double-flow neural network structure for model training and verification, outputting a model after model parameter updating and optimizing, inputting test set sea surface parameter information into the output model for reasoning, and obtaining a mesoscale vortex sub-table three-dimensional temperature inversion result.
Specifically, the dual-flow model adopts an encoding-decoding structure, the model structure is shown in fig. 3, the encoding stage is used for a characteristic extraction module of sea surface height SSH and sea surface temperature SST, and the decoding stage realizes characteristic fusion.
In the encoding stage, double-flow data SST and SSH are input first, and the relation among sea surface temperature, sea surface height and subsurface temperature is explored respectively. The method is characterized in that the sea surface height and the sea surface temperature are extracted, the fact that the mesoscale vortex has complex nonlinear characteristics is considered, the sea surface height information and the sea surface temperature information are necessarily related, a data characteristic fusion network is constructed for fusing the relation characteristics of SSH and SST, and meanwhile jump connection is utilized to reduce the fitting risk. In the process, SSH and SST are input into two branches of a neural network as double streams, modeling is carried out with a subsurface temperature field respectively, feature extraction is carried out among the same branch of the neural network, feature interaction is carried out among different branches, and data feature fusion is achieved.
The coding stage realizes the feature map to be smaller through a deepened network, extracts data information with different resolutions, and realizes the resolution recovery by utilizing deconvolution while the decoding stage fuses the feature output of each layer of the coding stage; and Triplet attention attention mechanism is introduced into the model, and a three-branch structure is adopted to fuse the channel attention and the space attention, so that cross-dimension interaction is realized.
The present embodiment employs Triplet attention for achieving cross-dimensional interactions, as shown in fig. 4, by capturing interactions of the spatial dimension and the input tensor channel dimension to achieve fusion of channel attention and spatial attention.
Specifically, given an input tensorThe method is transferred to three branches of a Triplet attention module, and in the first branch, an identity residual error branch structure is introduced, and dimension scaling is achieved by fusing average pooling and maximum pooling features through Z-Pool. Performing the Batch Norm normalization processing after convolution (Conv) with the size of 7 multiplied by 7, generating an attention weight through a Sigmoid activation function, and multiplying the attention weight with an identity residual branch point to obtain a first branch output; wherein Z-Pool is responsible for connecting the maximum pooling kernel average pooling feature of the dimension, reducing the 0 th dimension of the tensor to 2 nd dimensionThe depth of the actual tensor rich features is reduced while the actual tensor rich features are reserved, so that the purpose of light weight is achieved, and the Z-Pool formula can be expressed as follows:
the Sigmoid activation function formula can be expressed as:
inputting feature map tensors in the second branchRotating 90 counter-clockwise along the W-axis, achieving a dimension transformation tensor denoted +.>Dimension scaling is realized after Z-Pool to obtain +.>And performing normalization to obtain ++7 matrix>By activating the function with Sigmoid and then rotating 90 deg. clockwise along the W-axis to obtain the output->. Similarly, in the third branch, the H-axis is rotated to establish feature interaction between the H dimension and the C dimension.
Step four, training and verifying, iterating and optimizing the model
In this embodiment, the model training and model verification output the subsurface temperature structure. And training the double-flow model for inversion of the subsurface temperature by taking the preprocessed mesoscale vortex subsurface structure inversion sample library as a training set to obtain an inversion model after the training is completed. In model training, the experiment set batch_size was 64, epoch was 50, the optimizer was Adam, and the learning rate was 1e-3.
In the training process, a loss function adopts Mean Absolute Error (MAE) loss to represent the sum of absolute differences between a true value and a predicted value, and is used for model optimization in the training process to enable the predicted result to approach to a target result, wherein the model is expressed as the following formulaFor predictive value +.>For a true value, n is the number of temperature fields in a set:
the evaluation standard is used for evaluating inversion effects and verifying experimental results; r2, MAE and displayed_variance_score are adopted for comparison, wherein R2 is used for measuring the fitting degree of predicted data, the value is between [0,1], and the formula can be expressed as follows:
wherein f is an estimated value, y is a true value,is the average of the observed data.
The expanded_variance is an interpretation variance used for measuring the interpretation degree of the model on the fluctuation of the data set, and the interpretation variance of the true value and the predicted value can be calculated, and the value is between 0 and 1]Where var is the variance, y is the true value,is a predicted value.
Step five, obtaining a trained model, inputting test data, and outputting a subsurface temperature structure
And (3) taking the model after iterative optimization training as an inference model, inputting sea surface information (SSH, SST) in test data into the inference model to obtain temperature results under different depths of the mesoscale vortex of the reverse performance, and carrying out data visualization by utilizing a matplotlib library.
Example two
In one embodiment of the disclosure, a mesoscale vortex three-dimensional subsurface structure inversion system based on a dual-flow model is provided, which comprises a data acquisition module and a temperature inversion module:
a data acquisition module configured to: acquiring mesoscale vortex sea surface information to be inverted, including sea surface temperature and sea surface height, through a satellite;
a temperature inversion module configured to: inputting the mesoscale vortex sea surface information into a trained double-flow model, and inverting temperature results of the mesoscale vortex at different depths to obtain a mesoscale vortex subsurface temperature profile;
the dual-flow model introduces a Triplet attention attention mechanism, adopts a three-branch structure to fuse channel attention and space attention for cross-dimensional interaction, the first branch establishes characteristic interaction between H dimension and W dimension, the second branch establishes characteristic interaction between C dimension and W dimension, and the third branch establishes characteristic interaction between H dimension and C dimension.
Example III
An object of the present embodiment is to provide a computer-readable storage medium.
A computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps in a method for inversion of a mesoscale eddy three-dimensional subsurface structure based on a dual flow model as described in embodiment one of the present disclosure.
Example IV
An object of the present embodiment is to provide an electronic apparatus.
The electronic device comprises a memory, a processor and a program stored on the memory and capable of running on the processor, wherein the processor realizes the steps in the inversion method of the mesoscale vortex three-dimensional subsurface structure based on the double-flow model according to the embodiment I of the disclosure when executing the program.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (7)
1. The inversion method of the mesoscale vortex three-dimensional subsurface structure based on the double-flow model is characterized by comprising the following steps of:
acquiring mesoscale vortex sea surface information to be inverted, including sea surface temperature and sea surface height, through a satellite;
inputting the mesoscale vortex sea surface information into a trained double-flow model, and inverting temperature results of the mesoscale vortex at different depths to obtain a mesoscale vortex subsurface temperature profile;
the double-flow model introduces a Triplet attention attention mechanism, adopts a three-branch structure to fuse channel attention and space attention for cross-dimensional interaction, the first branch establishes characteristic interaction between H dimension and W dimension, the second branch establishes characteristic interaction between C dimension and W dimension, and the third branch establishes characteristic interaction between H dimension and C dimension;
the double-flow model adopts a coding-decoding structure, wherein a coding stage is used for extracting characteristics of sea surface temperature and sea surface height, and a decoding stage is used for characteristic fusion;
the encoding stage takes sea surface temperature and sea surface height as input to explore the relationship among the sea surface temperature, the sea surface height and the subsurface temperature respectively; the method comprises the steps of carrying out feature extraction on sea surface height and sea surface temperature, simultaneously considering that a mesoscale vortex has complex nonlinear features, inevitably connecting sea surface height information and sea surface temperature information, constructing a data feature fusion network for fusing SSH and SST relationship features, and reducing the risk of overfitting by using jump connection;
the decoding stage fuses the characteristic output of each layer of the encoding stage and realizes the recovery of resolution by utilizing deconvolution;
the double-flow model is trained by taking the constructed mesoscale vortex subsurface structure inversion sample library as a training data set;
and constructing the mesoscale vortex subsurface structure inversion sample library by taking the sea surface information corresponding to the mesoscale vortex time and the vortex center coordinate position as input parameters and the subsurface temperature data corresponding to the mesoscale vortex time and the vortex center coordinate position as labels.
2. The inversion method of a mesoscale vortex three-dimensional subsurface structure based on a double-flow model according to claim 1, wherein the first branch is introduced with an identity residual branch structure, and dimension scaling is realized by fusing average pooling and maximum pooling features through Z-Pool;
the second branch rotates the input feature map tensor along the W axis to realize feature interaction of the C dimension and the W dimension;
and the third branch rotates the input feature map tensor along the H axis to realize feature interaction of the H dimension and the C dimension.
3. The method for inversion of a mesoscale vortex three-dimensional subsurface structure based on a dual-flow model according to claim 1, wherein the loss function of the dual-flow model adopts MAE loss, which represents the sum of absolute differences between a true value and a predicted value, and the formula is:
，
wherein,for predictive value +.>For a true value, n is the number of temperature fields in a set.
4. Such as weightThe inversion method of a mesoscale vortex three-dimensional subsurface structure based on a double-flow model as claimed in claim 1, wherein the evaluation criteria of the double-flow model are as followsComparison of MAE and displained_variance_score, saidFor measuring the fitting degree of the predicted data, the formula is as follows:
，
wherein f is an estimated value, y is a true value,is the average value of the observed data;
the expanded_variance_score is an interpretation variance, and is used for measuring the interpretation degree of the model on the fluctuation of the data set, and calculating the interpretable variance of the true value and the predicted value, wherein the formula is as follows:
，
where var is the variance, y is the true value,is a predicted value.
5. The mesoscale vortex three-dimensional subsurface structure inversion system based on the double-flow model is characterized by comprising a data acquisition module and a temperature inversion module:
a data acquisition module configured to: acquiring mesoscale vortex sea surface information to be inverted, including sea surface temperature and sea surface height, through a satellite;
a temperature inversion module configured to: inputting the mesoscale vortex sea surface information into a trained double-flow model, and inverting temperature results of the mesoscale vortex at different depths to obtain a mesoscale vortex subsurface temperature profile;
the double-flow model introduces a Triplet attention attention mechanism, adopts a three-branch structure to fuse channel attention and space attention for cross-dimensional interaction, the first branch establishes characteristic interaction between H dimension and W dimension, the second branch establishes characteristic interaction between C dimension and W dimension, and the third branch establishes characteristic interaction between H dimension and C dimension;
the double-flow model adopts a coding-decoding structure, wherein a coding stage is used for extracting characteristics of sea surface temperature and sea surface height, and a decoding stage is used for characteristic fusion;
the encoding stage takes sea surface temperature and sea surface height as input to explore the relationship among the sea surface temperature, the sea surface height and the subsurface temperature respectively; the method comprises the steps of carrying out feature extraction on sea surface height and sea surface temperature, simultaneously considering that a mesoscale vortex has complex nonlinear features, inevitably connecting sea surface height information and sea surface temperature information, constructing a data feature fusion network for fusing SSH and SST relationship features, and reducing the risk of overfitting by using jump connection;
the decoding stage fuses the characteristic output of each layer of the encoding stage and realizes the recovery of resolution by utilizing deconvolution;
the double-flow model is trained by taking the constructed mesoscale vortex subsurface structure inversion sample library as a training data set;
and constructing the mesoscale vortex subsurface structure inversion sample library by taking the sea surface information corresponding to the mesoscale vortex time and the vortex center coordinate position as input parameters and the subsurface temperature data corresponding to the mesoscale vortex time and the vortex center coordinate position as labels.
6. An electronic device, comprising:
a memory for non-transitory storage of computer readable instructions; and
a processor for executing the computer-readable instructions,
wherein the computer readable instructions, when executed by the processor, perform the method of any of the preceding claims 1-4.
7. A storage medium, characterized by non-transitory storing computer readable instructions, wherein the computer readable instructions, when executed by a computer, perform the instructions of the method of any of claims 1-4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311638586.XA CN117350171B (en) | 2023-12-04 | 2023-12-04 | Mesoscale vortex three-dimensional subsurface structure inversion method and system based on double-flow model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311638586.XA CN117350171B (en) | 2023-12-04 | 2023-12-04 | Mesoscale vortex three-dimensional subsurface structure inversion method and system based on double-flow model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117350171A CN117350171A (en) | 2024-01-05 |
| CN117350171B true CN117350171B (en) | 2024-03-12 |
Family
ID=89363438
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311638586.XA Active CN117350171B (en) | 2023-12-04 | 2023-12-04 | Mesoscale vortex three-dimensional subsurface structure inversion method and system based on double-flow model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117350171B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117633712B (en) * | 2024-01-24 | 2024-04-19 | 国家卫星海洋应用中心 | Sea level height data fusion method, device and equipment based on multi-source data |
| CN117935023B (en) * | 2024-03-22 | 2024-05-28 | 青岛哈尔滨工程大学创新发展中心 | Method and system for identifying mesoscale vortex and analyzing features based on attention mechanism |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112102325A (en) * | 2020-09-17 | 2020-12-18 | 中国科学院海洋研究所 | Ocean abnormal mesoscale vortex identification method based on deep learning and multi-source remote sensing data |
| CN116486075A (en) * | 2023-03-31 | 2023-07-25 | 重庆大学 | HRNet-based multi-scale strong fusion semantic segmentation method for extracting ground features of remote sensing image |
| CN116612396A (en) * | 2023-05-26 | 2023-08-18 | 中国海洋大学 | Ocean surface temperature sequence prediction method based on space-time double-flow non-stationary sensing |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9310517B2 (en) * | 2011-11-07 | 2016-04-12 | The United States Of America As Represented By The Secretary Of The Army | Method and system for determining accuracy of a weather prediction model |
| US20190258925A1 (en) * | 2018-02-20 | 2019-08-22 | Adobe Inc. | Performing attribute-aware based tasks via an attention-controlled neural network |
-
2023
- 2023-12-04 CN CN202311638586.XA patent/CN117350171B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112102325A (en) * | 2020-09-17 | 2020-12-18 | 中国科学院海洋研究所 | Ocean abnormal mesoscale vortex identification method based on deep learning and multi-source remote sensing data |
| CN116486075A (en) * | 2023-03-31 | 2023-07-25 | 重庆大学 | HRNet-based multi-scale strong fusion semantic segmentation method for extracting ground features of remote sensing image |
| CN116612396A (en) * | 2023-05-26 | 2023-08-18 | 中国海洋大学 | Ocean surface temperature sequence prediction method based on space-time double-flow non-stationary sensing |
Non-Patent Citations (2)
| Title |
|---|
| The observed signature of mesoscale eddies in sea surface temperature and the associated heat transport;Ute Hausmann 等;《 Deep Sea Research Part I: Oceanographic Research Papers》;20121231;全文 * |
| 面向海洋的多模态智能计算：挑战、进展和展望;聂婕 等;《中国图象图形学报》;20220930;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117350171A (en) | 2024-01-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117350171B (en) | Mesoscale vortex three-dimensional subsurface structure inversion method and system based on double-flow model | |
| Wang et al. | Deep learning for daily precipitation and temperature downscaling | |
| CN116524361A (en) | Remote sensing image change detection network and detection method based on double twin branches | |
| CN115147731A (en) | SAR image target detection method based on full-space coding attention module | |
| CN114662788B (en) | Seawater quality three-dimensional time-space sequence multi-parameter accurate prediction method and system | |
| CN113627093B (en) | Underwater mechanism trans-scale flow field characteristic prediction method based on improved Unet network | |
| CN116306318B (en) | Three-dimensional ocean thermal salt field forecasting method, system and equipment based on deep learning | |
| CN113449640A (en) | Remote sensing image building semantic segmentation edge optimization method based on multitask CNN + GCN | |
| CN115908772A (en) | Target detection method and system based on Transformer and fusion attention mechanism | |
| Zhang et al. | FODA: Building change detection in high-resolution remote sensing images based on feature–output space dual-alignment | |
| Ran et al. | RGAN: Rethinking generative adversarial networks for cloud removal | |
| CN112949944A (en) | Underground water level intelligent prediction method and system based on space-time characteristics | |
| Fu et al. | CAL-Net: Conditional Attention Lightweight Network for In-Orbit Landslide Detection | |
| Yang et al. | Establishment of leakage detection model for oil and gas pipeline based on VMD-MD-1DCNN | |
| Xin et al. | Digitalization system of ancient architecture decoration art based on neural network and image features | |
| CN115393735A (en) | Remote sensing image building extraction method based on improved U-Net | |
| Zhang et al. | PEViz: an in situ progressive visual analytics system for ocean ensemble data | |
| Feng et al. | An underwater image enhancement strategy based on pyramid attention mechanism | |
| He et al. | A TMA-Seq2seq Network for Multi-Factor Time Series Sea Surface Temperature Prediction. | |
| US20230306557A1 (en) | Semantic learning-based down-sampling method of point cloud data of aircraft | |
| CN117250657B (en) | Seismic data reconstruction denoising integrated method | |
| Yao et al. | Multivariate deep learning for reconstruction of spatial missing climate data | |
| CN117975174B (en) | Three-dimensional digital core reconstruction method based on improvement VQGAN | |
| Li et al. | Intelligent Detection of Marine Organisms with Deep Learning based on Attention Mechanism | |
| El Montassir et al. | HyPhAI v1. 0: Hybrid Physics-AI architecture for cloud cover nowcasting |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |