275 results about "Sea ice" patented technology

A method for retrieving tropospheric wet delay and atmospheric water vapor content over polar sea ice with TDS-1 satellite grazing angle spaceborne GNSS-R is provided, including: Si, obtaining TDS-1 GNSS-R raw intermediate frequency signal data, VMF3 grid data, GPT3 grid data and ERA5 data; S2, correcting an error of tropospheric wet delay estimation of grazing angle spaceborne GNSS-R; S3, constructing a grazing angle spaceborne GNSS-R tropospheric wet delay estimation model; S4, calculating grazing angle spaceborne GNSS-R ZWD; S5, calculating a Tm value of a target point based on GPT3 model, substituting the Tm value into a conversion factor II, and combining calculated GNSS-R ZWD to obtain a GNSS-R IWV estimated value; and S6, verifying inversion performance of GNSS-R ZWD and integrated water vapor (IWV) by using reference data.

The invention discloses an euphausia superba fishery change prediction method based on an ocean circulation-sea ice-species distribution coupling model, and relates to the technical field of ocean science. Euphausia superba fishery historical observation data, satellite remote sensing data and reanalysis data are collected, euphausia superba resources and environmental data are integrated, and the euphausia superba fishery change prediction method based on the ocean circulation-sea ice-species distribution coupling model is obtained. Constructing a polar region resource environment database covering total elements of physics-biology-environment; and based on an FVCOM model, constructing an ocean circulation-sea ice coupling model of a surrounding sea area of the peninsula of the south pole, and simulating and verifying physical environment characteristics of a target sea area. By constructing the ocean circulation-sea ice-species distribution coupling model, integrating multi-source data and analyzing a cooperative driving mechanism of a multi-scale physical process to euphausia superba resource change, compared with a traditional method, the change of an euphausia superba fishery can be simulated and predicted more accurately, and the method has the advantages of being high in practicability and the like. And the problem of inaccurate prediction caused by limited space-time coverage rate of observation data is effectively solved.

The invention provides a north pole atmosphere coupling forecasting method capable of automatically fusing sea ice concentration and thickness, which belongs to the technical field of meteorology, and comprises the following steps: firstly, acquiring GFS background field data and correcting by applying an ice-sea heat exchange correction equation; and then sea ice concentration and thickness information is extracted, and physical consistency verification is carried out based on a thermodynamic equilibrium equation. Performing multi-element fusion through an Arctic ice sea assimilation model, calculating the sea ice surface temperature by applying a polar region sea ice heat flux equation, and establishing a sea ice concentration thickness boundary layer interaction equation set to describe the modulation effect of sea ice on a boundary layer; the method comprises the core steps of performing deep fusion on various sea ice features by applying a feature fusion model, adaptively adjusting the weight based on an ice-gas interface balance index, generating an optimized mode initial field and boundary conditions, and finally setting a parameter scheme to execute an integral program to obtain a forecasting result. The technical problem of low forecasting accuracy caused by insufficient physical consistency of sea ice information is solved.

The invention provides a sea ice area extraction method and device based on remote sensing, which are applied to the field of seawater detection, and the method comprises the steps: obtaining a remote sensing image of a target area; pixels of the remote sensing image are partitioned based on the normalized vegetation index, the pixels with the normalized vegetation index smaller than a preset vegetation index threshold value are divided into first type regions, and the first type regions comprise at least one of a water body region and a sea ice region; determining a normalized deicing index of the first type region; performing normalized deicing index threshold segmentation on the first type region through a maximum between-class variance method to obtain a sea ice classification threshold; and filtering and post-processing the remote sensing image based on the sea ice classification threshold, and taking the first type region of which the normalized deicing index is greater than the sea ice classification threshold as a sea ice region. The method is suitable for sea ice monitoring in a large area, and sea ice information can be efficiently and accurately extracted.

The invention discloses a near-shore sea ice remote sensing extraction method based on FY-3 / MERSI satellite data, and the method comprises the following steps: S1, carrying out the preprocessing of an obtained MERSI image, and obtaining an MERSI image only containing sea ice and seawater; s2, establishing a coastal sea ice remote sensing recognition algorithm according to the difference of the spectral characteristics of sea ice and seawater in each channel of the MERSI image, and evaluating the precision and accuracy of the coastal sea ice remote sensing recognition algorithm; s3, according to a coastal sea ice remote sensing recognition algorithm, extracting coastal sea ice information from the MERSI image of each day; and carrying out 10-day averaging and monthly averaging on the daily sea ice data extracted in each month to generate a sea ice ten-day product and a monthly product. According to the invention, automatic identification of high-spatial-resolution sea ice information of a high-latitude offshore area in winter can be realized.

The invention discloses a highly dense broken ice image segmentation method based on an iteration MGAC and an SAM model, and the method comprises the steps: carrying out the sea ice instance segmentation of a preprocessing image based on the SAM model, and obtaining a sea ice mask image corresponding to the preprocessing image; according to the binary image and the sea ice mask image, obtaining an initial sea ice residual region image which is not identified by the SAM model; obtaining initial sea ice residual region images under different gray threshold values to obtain an initial seed mask graph; taking the initial sea ice residual region image and the initial seed mask image as inputs of a preset MGAC contour model, and obtaining an MGAC sea ice recognition result based on a multi-round iteration partitioning mechanism; and performing union operation on the MGAC sea ice identification result and the sea ice mask image to obtain a crushed ice segmentation mask result. The method solves the problem that the existing method is insufficient in structure extraction precision and boundary integrity of the dense broken ice area.

The invention discloses a polar region sea area ice area route dynamic planning method based on multi-source data, and the method comprises the following steps: obtaining sea ice, ocean current and wind field historical data of a target channel sea area, and carrying out the preprocessing; inputting a preprocessing result into a constructed sea ice forecasting model to obtain a sea ice concentration and sea ice thickness distribution forecasting result of a planning day; combining the forecast result with ocean current and wind field forecast data, and utilizing a pre-established random forest navigational speed reasoning model to obtain target channel sea area planning daily navigational speed space distribution data; and carrying out route planning with minimum time cost through an improved A * algorithm to obtain a dynamic planning result of the target channel sea area route. The invention relates to the technical field of ship navigation in a sea ice area, and relates to the technical field of ship navigation in the sea ice area, a polar sea area navigation speed background field is deduced according to a fitting relationship through random forest modeling and fitting of the relationship between the navigation speed and sea ice, ocean current and a wind field, and dynamic route planning with minimum time cost is carried out.

The invention provides a polar region multi-source remote sensing sea wave significant wave height optimization interpolation fusion method and device, and belongs to the field of remote sensing, and the method comprises the steps: carrying out the preprocessing of collected sea ice marginal region and open sea surface data; north-pole geometric adaptation correction is carried out, system deviation of multiple satellite sensors is eliminated, and consistency and accuracy of data are ensured. Real-time wave height forecast is used as a background field, a weighting function is constructed in combination with a satellite observation error covariance matrix, dynamic background field fusion is carried out, and the precision and reliability of a data sparse region are improved; according to different characteristics of a sea ice marginal region and an open sea surface, different weights are set, a transition weight function is introduced, partition fusion is carried out, boundary discontinuity is avoided, and therefore the final fusion wave height is obtained; the robustness of the fusion method is verified through K-fold cross validation, and meanwhile precision verification is conducted in combination with multi-dimensional data. According to the method, the problem of insufficient accuracy verification spatial representativeness caused by rare north pole in-situ observation quantity and concentrated distribution is solved.

The invention discloses a ship route planning method and system in a polar region sea ice environment and a program product. The ship route planning method comprises the steps of preprocessing satellite remote sensing radiation intensity data, performing image reconstruction of inter-ice water channel detection through an improved Vision Transform model, planning and visualizing a ship route and the like. The deep features of the image blocks are extracted through a Transformer encoder which comprises a multi-head self-attention mechanism and is processed by a multi-layer perceptron; and the Transform encoder firstly performs layer normalization on the input, and then performs multi-head attention mechanism and multiple linear transformation and regularization processing. And realizing path planning in an inter-ice water channel environment by fusing a dynamic planning cost function through a fast marching method. Compared with a traditional algorithm, the method has the advantages that the calculation precision is higher in the same scene, the calculation efficiency is greatly improved, and the planning result obtained after multiple costs are comprehensively considered is given instead of pursuing the shortest distance.

The invention discloses a polar region sea ice semantic segmentation method based on a vision-language model, a sea ice segmentation region is specified through a language, and the method comprises the following steps: obtaining a visible light polar region sea ice data set; encoding the visible light polar region sea ice data set by using a vision-language model to obtain a text embedding vector and an image embedding vector; fusing the text embedding vector and the image embedding vector to obtain image semantic information and text semantic information; and based on the text semantic information, decoding the image semantic information to obtain a polar region sea ice segmentation mask. According to the method, the target area can be selected and segmented more flexibly, and different types of ice can be segmented more accurately and efficiently. Besides, through calculation of a mutual attention mechanism of the text and the image, the text and the image can be mutually fused, and the text description can be accurately utilized to select and segment the corresponding image area.

The invention provides a method for calculating fatigue by adopting near-field dynamics under a complex ice condition based on propeller icebreaking, which is used for processing excitation data of an ice-induced load on a propeller of an icebreaker, and comprises the following steps: firstly, constructing a sea ice environment, an elastic propeller constructed by a plurality of units is simulated by adopting a finite element method, and the continuous interaction between the elastic propeller and ice blocks in the rotating process is considered. The contact algorithm adopts an element surface finite element (FEM) method to process force transmission from a force bearing point to a force bearing surface so as to influence the force bearing state of each node of the propeller, the force is transmitted to element stress through the node effect, and the ice load borne by the propeller structure is obtained and verified.

The invention provides an SAR image polar region sea ice detection method and system based on deep learning and a medium, the land and the water area are segmented by using a water area detection algorithm based on a DeeplabV3 network, the segmentation precision of the DeeplabV3 network is improved by means of cavity convolution, a spatial pyramid pooling module and a decoder, and the calculation and memory overhead is reduced by using a lightweight MobileNet; then, a sea ice detection algorithm based on a short-term dense cascading STDC network is used for precisely segmenting sea ice and water in a water area, the STDC network reduces the calculation complexity through specific module design and keeps multi-scale information, and meanwhile, multiple optimization means are adopted in decoder design to improve the detection performance; in addition, data is prepared and augmented, and the generalization ability of the model is enhanced. The sea ice detection precision and efficiency are improved, and the method is suitable for resource-limited environments such as on-satellite environments and the like.

The invention discloses a low-frequency sound propagation loss forecasting method suitable for a polar region sea area, and the method comprises the steps: building a two-dimensional ice layer model of a forecast sea area according to remote sensing data based on the setting that sea ice of the polar region sea area is a uniform elastic medium; calculating hydrological conditions of the ocean sound channel of the polar sea area; establishing a forecast sea area two-dimensional seabed model according to the seabed topography sediment database; an under-ice sound field model is constructed based on a three-dimensional cylindrical coordinate system, sound pressure from unit intensity sound sources with different sizes and depths to a receiving point in a forecast sea area is calculated through a frequency domain finite element method in combination with a PML technology, and sound propagation loss of a needed position is obtained through calculation. According to the method, the application limitation of traditional ray and normal wave theories under complex boundary conditions is broken through through a finite element method, a synergistic effect mechanism of actual ice layer parameter space change, submarine topography fluctuation and sound velocity profile gradient characteristics is systematically analyzed, and sound field modeling is more accurately carried out on different ice layers and submarine topographies in the Chukong sea region all year round; the influence of sound wave propagation is simulated.

The invention discloses a polar region ship navigation planning method based on sea ice prediction and multi-stage path optimization, and the method comprises the steps: obtaining a navigation risk map based on a to-be-navigated ship according to the updated single-day sea ice data in a time group; the method comprises the steps of obtaining a candidate path set based on a K shortest path algorithm, determining final U candidate paths according to path similarity of a result set of the candidate paths, further obtaining comprehensive scores of the paths in the result set, and finally determining a final driving path of a to-be-sailed ship according to the comprehensive score of each final candidate path. And completing the path planning of the polar region ship navigation. According to the method, the time, the risk and the path smoothness under the distance suboptimal constraint are taken as optimization targets, and similarity calculation is performed on the paths in the result set of the obtained candidate paths, so that diversified risk avoiding scheme paths can be provided. According to the method, the prediction time range is sliced, so that the high-frequency environment updating requirement in the prior art is met.

This paper constructs a remote sensing sea ice dataset and, based on the characteristics of sea ice images, improves the YOLOv5 model by adding the SE attention mechanism, improving the SPP pooling pyramid structure, and replacing the SiLU activation function with the FReLU. The model's confidence is verified using evaluation metrics such as accuracy, recall, F1-score, and mean average precision, ultimately resulting in an optimal sea ice target detection model. The test set is input into the target detection model to obtain recognition results, extracting information such as the location and size of the sea ice. Based on this extracted information, a grid map is created. The starting and ending points are determined based on the ship's own position, and a path is planned using the Theta* algorithm. The generated path is evaluated using metrics such as route distance, average offset distance, number of ship turns, and number of sea ice avoidances. This paper implements sea ice recognition at a remote sensing scale, constructs simulated navigation scenarios, and plans the optimal navigation path. This paper can improve the safety of ships sailing in ice areas and provide technical support for collision avoidance and navigation decision-making assistance in ice areas.

An SAR sea ice remote sensing image segmentation method based on an MAE and a ViT model comprises the following steps that an unmarked AI4Arctic data set is used, a mask auto-encoder MAE is used for pre-training a visual Transform model, a picture is segmented into patches, then random masking is carried out, sine and cosine position embedding is adopted for carrying out position coding on visible patches, and a pre-trained ViT model is obtained; three segmentation labels of sea ice concentration SIC, sea ice development stage SOD and floating ice size FLOE are extracted from AI4Arctic original NetCDF ice map data, a single-channel grey-scale map is converted into a three-channel PNG image to serve as a fine tuning data set, fine tuning training of an encoder-decoder architecture is carried out in an end-to-end mode, any weight is not frozen in training, and the fine tuning data set serves as a fine tuning data set. And a multi-task ViT prediction model taking SIC, SOD and FLOE as segmentation targets is obtained. The ViT of the invention can better capture SAR image features, improve the accuracy of sea ice segmentation, significantly reduce the demand for training computing power and dependence on data annotation, and improve the speed of model training.

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 embodiment of the invention relates to an offshore photovoltaic sea ice identification method and system, a program product and equipment, and relates to the technical field of offshore photovoltaic, and the method comprises the steps: obtaining an unmanned plane multispectral image, a satellite remote sensing image, photovoltaic panel power generation data and underwater sonar data of an offshore photovoltaic site; global image features and multi-scale features are extracted from the unmanned aerial vehicle multispectral image and fused, first image features are determined, second image features are extracted from the satellite remote sensing image, time sequence features are extracted from photovoltaic panel power generation data, and ice anchor features are extracted from underwater sonar data; fusing the first image feature, the second image feature, the time sequence feature and the ice anchor feature through a plurality of image convolutional layers in an image neural network to obtain a fused feature; and the sea ice probability of each pixel point in the unmanned aerial vehicle multispectral image and the satellite remote sensing image is determined according to the fusion features and the external environment parameters, and a sea ice recognition result is determined. The sea ice can be accurately identified.

The invention belongs to the technical field of prediction of the thickness of sea ice in the arctic region, and particularly relates to an intelligent prediction method for the thickness of sea ice in a polar region based on multi-layer stacking space-time Transform. The method comprises the following steps: step 1, based on polar region sea ice thickness reanalysis data information of a target sea area, extracting a feature matrix in ocean SIT reanalysis data and constructing a space-time sequence forecast supervision data set; 2, preprocessing data of a background field and a target field; 3, an HSS-Transform model is constructed, and the HSS-Transform model is constructed; 4, inputting the sea ice thickness background field space-time sequence into an HSS-Transform model to obtain a sea ice thickness prediction space-time sequence in the next 15 days; step 5, carrying out reverse normalization operation on the obtained prediction space-time sequence, and restoring the land point; according to the method, the spatial-temporal characteristics of the sea ice thickness data can be extracted and processed at the same time on multiple time scales and spatial scales, the extension period intelligent prediction of the polar region sea ice thickness spatial-temporal sequence is realized, and the prediction level of the sea ice thickness is remarkably improved.

The invention provides an ice layer thickness identification and stress sensing cooperative navigation monitoring method and system, and relates to the technical field of ship navigation safety monitoring. The method comprises the following steps: acquiring and preprocessing sea ice images around a ship in real time; a self-built data set is made through labeling, a deep learning model is selected for segmentation, segmentation results of the snow accumulation layer and the ice layer are obtained, mask pixel values of the snow accumulation layer and the ice layer are obtained, and then the actual thickness is calculated; acquiring real-time stress data; using a timestamp synchronization mechanism to match the sea ice image of the time corresponding frame and the stress data; carrying out the normalization processing of the sea ice image and the stress data, carrying out the weight fusion of the normalized data, and calculating a comprehensive risk coefficient; a risk early warning level is output according to the interval where the coefficient is located; and when the coefficient exceeds a set threshold value, sound-light alarm and interface warning information are sent out. According to the invention, through collaborative analysis and risk assessment of multi-dimensional data, the sailing safety of the ship in the polar region and the sea ice region is improved.

The invention relates to the technical field of meteorological model optimization and polar region ocean data processing, in particular to a method and system for optimizing a COARE model based on sea ice density, and the method comprises the steps: obtaining the historical actual measurement data of the polar region ocean, calculating the real roughness and the model roughness outputted by the COARE model, and determining the roughness difference value; dividing a wind speed interval, and fitting and establishing a correction function of a difference value between the sea ice concentration and the roughness; and introducing the sea ice concentration into the COARE model, judging the interval according to the wind speed, calling a corresponding function to correct the roughness of the model, and outputting the ocean-atmosphere flux conforming to the polar region sea ice condition to complete the model optimization. According to the method, the sea ice concentration is introduced, and the wind speed partition correction relation is established, so that the roughness of the polar region under the sea ice coverage condition is dynamically corrected by the COARE model, the limitation of the original model in the aspects of wind speed applicability and polar region adaptability is broken through, and the accuracy and practicability of polar region ocean-atmosphere flux calculation are remarkably improved.

The embodiment of the invention provides a self-adaptive energy recovery system, device and method, and the system comprises an environment sensing module which is used for collecting external environment parameters and device operation state data; the data analysis module is connected with the environment sensing module and used for processing the collected data and obtaining sea ice height information, threat level and position information of the energy recovery device through analysis; the control execution module is connected with the data analysis module and used for generating a control instruction according to the sea ice height information, the threat level and the position information of the energy recovery device; the energy recovery device is connected with the control execution module and is used for executing the control instruction and carrying out position adjustment and / or energy recovery operation; and the energy storage module is connected with the energy recovery device and is used for storing the recovered electric energy. External information can be sensed in real time, self-adaptive adjustment can be made, damage to a wind power airport is reduced, and secondary utilization of sea ice mechanical energy can be achieved.

The invention discloses an ice breaking channel sea ice resistance calculation method for first ice breaking of layer ice, which comprises the following steps: acquiring and acquiring a broken ice distribution video image in an ice breaking channel to obtain a broken ice area feature map; obtaining a fitting probability density function used for obtaining a crushed ice size probability density distribution histogram, crushed ice shape feature distribution and crushed ice density of crushed ice distribution in the icebreaking channel according to the crushed ice region feature map, and further constructing a crushed ice channel sea ice resistance calculation optimization model related to crushed ice channel sea ice parameters; based on the sea ice resistance calculation optimization model of the crushed ice channel, realizing icebreaking piloting guidance of the ship in the crushed ice channel; the problem that an existing method cannot accurately obtain sea ice parameters of an ice breaking channel for first ice breaking of layer ice, and then cannot effectively pilot and guide ice breaking of a ship in the ice breaking channel is solved.

The invention relates to the technical field of intelligent transportation, in particular to a state space modeling method for ship ice region navigation decision model training, which comprises the following steps of: constructing a sea ice dynamic model, and estimating an unobservable sea ice state by using unscented Kalman filtering to obtain a hidden variable vector; inputting the hidden variable vector, the observable state information and the external force field code into a Transform encoder, and outputting future sea ice center position distribution; constructing a confidence collision avoidance domain of the time index according to future sea ice center position distribution, wherein the confidence collision domain is updated in a rolling manner along with time; taking observable state information, hidden variable vectors and transformation embedding of a confidence collision avoidance domain as reinforcement learning states, and adopting near-end strategy optimization to train a sea ice dynamic model; and obtaining a ship ice region navigation decision model based on a joint termination criterion of the verification set collision rate, the arrival rate and the strategy divergence. The method improves the strategy convergence stability and the ice region safety passing rate, and is suitable for autonomous navigation and simulation training of polar regions and high-latitude sea areas.

The invention discloses a polar region water surface unmanned ship path planning method based on an improved A * algorithm. Aiming at the problems that interference information exists in a radar map in a polar region environment, a sea ice area is difficult to accurately extract and the like, firstly, a radar induction area is extracted by utilizing a Hough circle detection technology, and a multi-threshold image segmentation method based on a Kapur information entropy and emperor penguin optimization algorithm is adopted to filter out the interference information such as latitude and longitude marks and radar echoes; and an accurate sea ice distribution diagram is obtained. And then, an improved cost evaluation function fusing credibility, threat information potential energy and pheromone concentration is constructed, a TI-PC-A * path planning algorithm is proposed, and safe path planning of the sea ice area is realized. The method improves the accuracy of polar region sea ice image processing and the intelligence of path planning, and is suitable for an autonomous navigation task in a polar region environment.

The invention provides a training method and device of a north pole sea fog model, a storage medium and electronic equipment. The electronic equipment obtains the sea fog visibility of the research area at a plurality of historical time points by using a pre-constructed PWRF model; for each historical time point, weighting the original meteorological data, the original sea surface temperature data and the original sea ice data of the plurality of meteorological models at the historical time point to obtain sample meteorological data, sample sea surface temperature data and sample sea ice data at the historical time point; and training the to-be-trained model through the synthesized data to obtain the north pole sea fog model. Therefore, the problem that sea fog visibility data with long time sequence, large range area and high spatial resolution cannot be obtained at present is solved, and input data formed by fusing multiple models has high confidence, so that the trained north pole sea fog model can effectively capture complex mechanisms such as humidity threshold effect, interaction of wind speed and sea ice and the like; and the sea fog prediction precision is improved.

The invention discloses an ice water remote sensing classification method based on an LAU-Net model, and belongs to the technical field of remote sensing image intelligent processing. The method comprises the following steps: constructing an ice water classification sample set; based on a standard U-Net network architecture, constructing an improved model LAU-Net fused with a local attention module; training the LAU-Net model by using the ice water classification sample set, and dynamically adjusting a learning rate, a batch size, an optimizer weight attenuation factor and regularization strength through an automatic hyper-parameter optimization module; and adopting the trained LAU-Net model to carry out ice water classification on the input image. According to the method, the local attention mechanism and network structure optimization are introduced, so that the classification precision of the ice water boundary in the remote sensing image and the model operation efficiency are improved, and the technical requirements of dynamic sea ice monitoring are met.

The invention discloses a sea ice melting pool detection method based on ICESat-2 laser radar point cloud data. The method is used for solving the problem of double-surface precise separation caused by low photon signal-to-noise ratio of a sea ice melting pool area in summer. According to the method, ICESat-2 photon data are segmented through data preprocessing, and a sliding window is generated to position the sea ice and melting pool height range; gaussian fitting is adopted to divide an on-ice / under-ice photon set, and noise elimination is carried out in combination with dynamic height difference threshold iteration and terrain adaptive ellipse density dual thresholds; and performing double-peak detection and saddle point analysis based on the smooth height histogram of the denoised photons, and dynamically dividing the height layer range of the top and the bottom of the fusion pool. According to the method, through multi-stage dynamic threshold optimization and terrain adaptive ellipse density calculation, full-automatic precise separation of the double surfaces of the sea ice melting pool under the low signal-to-noise ratio environment is achieved, and reliable technical support is provided for global polar region environment monitoring and climate change research.

This invention discloses a method for predicting changes in Antarctic krill fishery grounds based on a coupled ocean circulation-sea ice-species distribution model. This method, which relates to the field of marine science and technology, collects historical observational data on the Antarctic krill fishery, satellite remote sensing data, and reanalysis data, integrates Antarctic krill resource and environmental data, and constructs a polar resource and environmental database covering all physical, biological, and environmental factors. Based on the FVCOM model, a coupled ocean circulation-sea ice model is constructed for the waters surrounding the Antarctic Peninsula to simulate and verify the physical environmental characteristics of the target sea area. By constructing a coupled ocean circulation-sea ice-species distribution model, integrating multi-source data, and analyzing the synergistic driving mechanisms of krill resource changes through multi-scale physical processes, this method can more accurately simulate and predict changes in Antarctic krill fishery grounds than traditional methods, effectively addressing the problem of inaccurate predictions caused by the limited spatiotemporal coverage of observational data.

The invention belongs to the field of sea ice evolution monitoring, and provides a sea ice evolution prediction method based on a space-time Swin long-short-term memory network, and the method specifically comprises the steps: S1, obtaining long-time-sequence remote sensing image data, and making a long-time-sequence sea ice data set and a label; s2, processing the long-time-sequence sea ice data set by adopting a loop iteration Swin long-short-term memory network model as a basic framework; s3, integrating the Swin Transform model and the convolutional long-short-term memory network, capturing a time dependency relationship and a global spatial dependency relationship based on an entry gating mechanism, constructing a learning network of a space-time prediction task, and performing loop iteration; and S4, predicting a sea ice evolution result at the last layer of loop iteration of the Swin long-short-term memory network model. According to the method, the global space information and the dynamic time dependency relationship in the long-time-sequence sea ice data are effectively generated, the prediction network with the efficient space-time modeling capability is generated, and the sea ice evolution prediction performance is improved.