Marine environment data processing method, device, electronic device and readable storage medium

Abstract

This invention provides a method, apparatus, electronic device, and readable storage medium for marine environmental data processing, relating to the field of data processing technology. The method includes introducing a global spatiotemporal grid partitioning standard; constructing a spatiotemporal grid coding mapping rule table for marine environmental elements based on the global spatiotemporal grid partitioning standard; determining the target spatiotemporal grid display range based on the spatiotemporal grid coding mapping rule table, the user-selected spatiotemporal grid level, and the spatiotemporal grid display range; querying marine environmental element data corresponding to the target spatiotemporal grid display range in a background database; extracting features from the marine environmental element data to obtain marine environmental element feature vectors; extracting features from the marine environmental element feature vectors; and identifying marine environmental phenomena information corresponding to the marine environmental element data based on the extracted features. This effectively improves the problems of difficulty in aligning and integrating multi-source heterogeneous data in traditional methods, enhancing the comprehensive modeling and information expression capabilities for complex marine environments.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method, apparatus, electronic device, and readable storage medium for processing marine environmental data. Background Technology

[0002] Marine environmental data and phenomena exhibit distinct spatiotemporal attributes, generally summarized as the "4Vs": Volume, Variety, Velocity, and Value. They also possess the "5Hs": High Correlation, High Coupling, High Diversification, Hierarchy, and High Regularity. These characteristics indicate that marine environmental data and phenomena are not only vast in quantity but also inherently complex and variable, with low information density, necessitating effective utilization methods.

[0003] Traditional methods for utilizing marine environmental data primarily rely on converting it into remote sensing data or images, using computer vision techniques for feature extraction, image recognition, and semantic segmentation. However, due to the low information density of marine environmental data, the correlation between marine environmental elements and phenomena is not well reflected in the data. This leads to problems such as difficulty in alignment and integration, and narrow application scope when dealing with multi-source heterogeneous data. Data processing efficiency is low, resource utilization is insufficient, and the comprehensive modeling and information representation of complex environments are relatively lagging, making it difficult to meet current marine-based operational needs.

[0004] Therefore, how to ensure that the information obtained fully reflects the complexity and dynamic changes of the marine environment, while maximizing the acquisition of effective information and reducing redundancy, remains a pressing problem to be solved. Summary of the Invention

[0005] This invention provides a marine environmental data processing method, apparatus, electronic device, and readable storage medium to address the problems of existing marine environmental data utilization methods, such as difficulty in alignment and integration when dealing with multi-source heterogeneous data, low data processing efficiency, insufficient resource utilization, relatively backward comprehensive modeling and information expression effects for complex environments, and difficulty in adapting to current marine-based business needs. It achieves the goal of ensuring that the obtained information can fully reflect the complexity and dynamic characteristics of the marine environment, while maximizing the acquisition of effective information and reducing redundancy.

[0006] This invention provides a marine environmental data processing method, comprising: constructing a spatiotemporal grid coding mapping rule table for marine environmental elements based on a global spatiotemporal grid partitioning standard; the spatiotemporal grid coding mapping rule table containing the correspondence between spatiotemporal grids and marine environmental element data and marine environmental phenomenon information of the corresponding regions;

[0007] The target spatiotemporal grid display range is determined based on the spatiotemporal grid encoding mapping rule table, the spatiotemporal grid level selected by the user, and the spatiotemporal grid display range selected by the user.

[0008] Query the marine environmental element data corresponding to the target spatiotemporal grid display range in the background database;

[0009] Feature extraction is performed on the marine environmental element data to obtain marine environmental element feature vectors;

[0010] Feature extraction is performed on the feature vectors of the marine environmental elements, and marine environmental phenomena corresponding to the marine environmental element data are identified based on the extracted features.

[0011] Optionally, determining the target spatiotemporal grid display range based on the spatiotemporal grid encoding mapping rule table, the spatiotemporal grid level selected by the user, and the spatiotemporal grid display range selected by the user includes:

[0012] After receiving the spatiotemporal grid information sign loading request on the front-end page, the grid identifier, the spatiotemporal grid level, and the spatiotemporal grid display range are determined based on the user's selected operation.

[0013] The target spatiotemporal grid display range is determined based on the grid identifier, the spatiotemporal grid level, and the spatiotemporal grid display range.

[0014] Optionally, after determining the target spatiotemporal grid display range based on the grid identifier, the spatiotemporal grid level, and the spatiotemporal grid display range, the process includes:

[0015] In response to the request for grid coding of marine environmental elements, the grid data information contained in the target spatiotemporal grid is searched from the spatiotemporal grid coding mapping rule table based on the grid identifier;

[0016] Calculate the statistical characteristics of marine environmental elements in the grid data information contained in the target spatiotemporal grid; the statistical characteristics of marine environmental elements include the maximum value, minimum value, average value, variance, and standard deviation of the grid data information;

[0017] The corresponding statistical characteristics of the marine environmental elements are attached to each of the target spatiotemporal grids.

[0018] Optionally, querying the marine environmental element data corresponding to the target spatiotemporal grid display range in the background database includes:

[0019] Query the marine environmental phenomenon data information corresponding to the target spatiotemporal grid display range in the background database; the marine environmental phenomenon data information includes the basic attributes of marine environmental phenomena and the dynamic symbolic representation of marine environmental phenomena.

[0020] Based on the marine environmental phenomenon data, a related spatiotemporal grid is determined; the related spatiotemporal grid is a spatiotemporal grid that intersects with the range of occurrence of the marine environmental phenomenon;

[0021] Based on the spatiotemporal grid coding mapping rule table, the marine environmental element data corresponding to the associated spatiotemporal grid is determined, and used as the marine environmental element data corresponding to the target spatiotemporal grid display range.

[0022] Optionally, the marine environmental element data can be classified;

[0023] The classification of the marine environmental element data includes:

[0024] Marine environmental phenomena corresponding to marine environmental element data involving one of the aforementioned spatiotemporal grids are considered as small-scale phenomena in the current spatiotemporal grid representation sense;

[0025] Marine environmental phenomena corresponding to marine environmental element data involving two or more of the aforementioned associated spatiotemporal grids are considered as large-scale phenomena in the current spatiotemporal grid representation sense.

[0026] The small-scale phenomena are identified on a single corresponding spatiotemporal grid, and the large-scale phenomena are identified on multiple corresponding spatiotemporal grids.

[0027] Optionally, the step of extracting features from the marine environmental element data to obtain marine environmental element feature vectors includes:

[0028] The spectrum was obtained by performing a Fourier spatial transform on the marine environmental element data.

[0029] The spectrum is convolved by a neural network to obtain a single-grid, single-element marine environmental element feature vector; the single-grid, single-element marine environmental element feature vector is the feature vector of a specific marine environmental element in a single spatiotemporal grid unit.

[0030] The feature vectors of the marine environmental elements are placed into the spatiotemporal grid coding mapping rule table.

[0031] Optionally, the step of extracting features from the feature vector of the marine environmental elements and identifying the marine environment corresponding to the marine environmental element data based on the extracted features includes:

[0032] In response to the request to enhance the characteristics of marine environmental phenomena, a weighted average is performed on the feature vectors of each marine environmental element to obtain a single spatiotemporal grid comprehensive feature vector; the single spatiotemporal grid comprehensive feature vector is the comprehensive feature vector of all elements corresponding to a single spatiotemporal grid;

[0033] In the context of large-scale phenomena within the current spatiotemporal grid, multiple spatiotemporal grids are included. A weighted average of the comprehensive feature vectors of all the single spatiotemporal grids in the multiple spatiotemporal grids is performed using learnable weighting parameters to obtain the feature vector associated with the marine environmental phenomenon. The feature vector associated with the marine environmental phenomenon is the comprehensive feature vector of all spatiotemporal grids corresponding to the marine environmental phenomenon.

[0034] The feature vectors associated with the marine environmental phenomena are classified using a classifier to identify the marine environmental phenomena corresponding to the marine environmental element data.

[0035] Optionally, the global spatiotemporal grid subdivision standard is the BeiDou grid location code or the H3 hexagonal standard grid.

[0036] The present invention also provides a marine environmental data processing device, comprising the following modules:

[0037] The module constructs a spatiotemporal grid coding mapping rule table for marine environmental elements based on the global spatiotemporal grid subdivision standard; the spatiotemporal grid coding mapping rule table contains the correspondence between spatiotemporal grids and marine environmental element data of the corresponding regions;

[0038] The grid determination module determines the target spatiotemporal grid display range based on the spatiotemporal grid encoding mapping rule table, the spatiotemporal grid level selected by the user, and the spatiotemporal grid display range selected by the user.

[0039] The query module retrieves marine environmental element data corresponding to the target spatiotemporal grid display range from the background database;

[0040] The feature extraction module extracts features from the marine environmental element data to obtain marine environmental element feature vectors.

[0041] The feature processing module extracts features from the feature vectors of the marine environmental elements and identifies the marine environmental phenomena corresponding to the marine environmental element data based on the extracted features.

[0042] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the marine environmental data processing method as described above.

[0043] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the marine environmental data processing method as described above.

[0044] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the marine environmental data processing method as described above.

[0045] The marine environmental data processing method, apparatus, electronic device, and readable storage medium provided by this invention establish a systematic framework for marine environmental data processing by introducing a global spatiotemporal grid partitioning standard and constructing a spatiotemporal grid coding mapping rule table. This ensures that subsequent data processing and analysis are based on a unified and standardized standard, enhancing the universality and comparability of the data and effectively improving the problems of difficulty in aligning and integrating multi-source heterogeneous data in traditional methods. By extracting feature vectors from marine environmental element data, it is possible to deeply explore the intrinsic relationship between marine environmental element data and marine environmental phenomena, making up for the shortcomings of traditional methods in reflecting the insufficient correlation between marine environmental element data and phenomena. This allows for more comprehensive and accurate acquisition of marine environmental phenomenon information, improving the comprehensive modeling and information expression capabilities for complex marine environments, and adapting to diverse marine-based business needs. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0047] Figure 1 A flowchart of a marine environmental data processing method provided by the present invention.

[0048] Figure 2 This is a schematic diagram of a spatiotemporal grid partitioning standard provided by the present invention.

[0049] Figure 3 This is a flowchart of a sub-step of S2 provided by the present invention.

[0050] Figure 4 This is a flowchart of a sub-step S3 provided by the present invention.

[0051] Figure 5 A flowchart of a sub-step S4 provided by the present invention.

[0052] Figure 6 This is a schematic diagram illustrating feature extraction and classification of marine environmental element data provided by the present invention.

[0053] Figure 7 A flowchart of a sub-step S5 provided by the present invention.

[0054] Figure 8 This is a schematic diagram of the structure of a marine environmental data processing device provided by the present invention.

[0055] Figure 9 A schematic diagram of the physical structure of an electronic device is provided.

[0056] Figure label:

[0057] Marine environmental data processing device 80; construction module 801; grid determination module 802; query module 803; feature extraction module 804; feature calculation module 805; processor 910; communication interface 920; memory 930; communication bus 940. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0059] Unless otherwise defined, 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 invention pertains. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0060] Before providing a detailed description of the present invention, let's first introduce the technical terms involved in the present invention.

[0061] Global Spatiotemporal Grid Subdivision Standard: A standardized method developed for the spatial division and management of the Earth's surface.

[0062] The BeiDou grid location code and the H3 hexagonal standard grid are two common global spatiotemporal grid subdivision standards.

[0063] BeiDou Grid Location Code: This is a global spatial location coding standard proposed by the BeiDou Navigation Satellite System (BDS). Based on the latitude and longitude grid division of the Earth's surface, it adopts a hierarchical coding method, dividing the Earth's surface into grid units of different levels, with each unit corresponding to a unique code.

[0064] H3 hexagonal standard grid: It is a global hexagonal grid subdivision standard designed to provide an efficient gridding method for geospatial data analysis and services. Based on the icosahedral projection, it divides the Earth's surface into multiple hexagonal grid units, supporting 16 layers of subdivision, with the size of each grid unit gradually decreasing.

[0065] Marine environmental data refers to data describing various physical (temperature, salinity, etc.), chemical (dissolved oxygen, pH, etc.), biological (plankton, benthic organisms, etc.), and geological parameters (sediments, seabed topography, etc.) of the marine environment. These data are of great significance for marine scientific research, resource management, environmental protection, and climate prediction.

[0066] Marine environmental phenomena refer to various natural phenomena occurring in the ocean, such as mesoscale eddies, ocean circulation, upwelling and downwelling currents, and tides, encompassing multiple aspects including physics, chemistry, biology, and geology. Large-scale phenomena such as internal solitary waves, ocean fronts, storm surges, and tides have a wide range of impacts, long durations, and can affect the activities of large ships and other equipment. Small-scale phenomena such as shore crashing waves and rift currents have limited impacts, short durations, and generally only affect the activities of people or small vessels and equipment. These phenomena not only affect marine ecosystems but also have a profound impact on global climate, weather patterns, and human activities.

[0067] Information about marine environmental phenomena can usually be captured and generated from marine environmental element data based on the phenomenon dynamics generation mechanism. Therefore, deep learning can be used to discover and distinguish the correlation between marine environmental element data and marine environmental phenomenon information.

[0068] Figure 1 A flowchart of a marine environmental data processing method provided by the present invention is shown below. Figure 1 As shown, this marine environmental data processing method is used in devices such as servers, desktop computers, and laptops, and includes the following steps:

[0069] In S1, a spatiotemporal grid coding mapping rule table for marine environmental elements is constructed based on the global spatiotemporal grid partitioning standard.

[0070] The global spatiotemporal grid subdivision standard can be a BeiDou grid location code or an H3 hexagonal standard grid. The spatiotemporal grid coding mapping rule table contains the correspondence between the spatiotemporal grid and the corresponding marine environmental element data and marine environmental phenomenon information. For example, the global spatiotemporal grid subdivision standard can be as follows: Figure 2 The spatiotemporal grid partition shown is as follows: Figure 2This invention provides a schematic diagram of a spatiotemporal grid partitioning standard, which divides the digital earth ocean region into spatiotemporal grids of equal area at a fixed hierarchy. The spatiotemporal grid coding mapping rule table is used to characterize the correspondence between grid identifiers, latitude and longitude ranges, feature data ranges, feature data statistical characteristics, and feature features; for example, the spatiotemporal grid coding mapping rule table can be shown in the following table:

[0071]

[0072] Among them, the grid identifier is the unique identifier of the spatiotemporal grid, such as the network ID in the table above; the latitude and longitude range is the set of spatiotemporal grid boundary lines, used to characterize the latitude and longitude range and grid display range contained in the spatiotemporal grid; the feature data range is the range of marine environmental feature data, and the feature data index is set in the spatiotemporal grid coding mapping rule table; the feature data statistical characteristics are the maximum value (max), minimum value (min), mean value (mean), variance (Var), and standard deviation (std) of the grid point data information of the spatiotemporal grid; the feature features are the feature vectors of the marine environmental feature data.

[0073] It should be noted that the table above only shows four grid data points as an example. In actual applications, there is more data involved, which will not be elaborated here.

[0074] In one implementation, feature data within the latitude and longitude range contained in the spatiotemporal grid are obtained. When the resolution of a certain feature data is low, and it is not present or only a small amount of data exists in the grid, and is below a certain threshold, it is insufficient to form a basis for statistical characteristic analysis. In this case, the assignment of the spatiotemporal grid is set to null, and the weighting of the corresponding feature data is reset to 0, and it is not learnable.

[0075] Add a grid inclusion field to the spatiotemporal grid coding mapping rule table. This field, or attribute, indicates whether a grid cell contains valid data. It identifies which grid cells are "valid," meaning they contain actual observed or calculated marine environmental element data, thus avoiding unnecessary queries and calculations on invalid or empty data grids. For grids with valid data, add the latitude and longitude coordinates of the grid points. ,index Sum .

[0076] By introducing a global spatiotemporal grid partitioning standard and constructing a spatiotemporal grid coding mapping rule table, a systematic framework for marine environmental data processing is established. This ensures that subsequent data processing and analysis are based on a unified and standardized standard, enhancing the universality and comparability of the data and effectively improving the problems of difficulty in aligning and integrating multi-source heterogeneous data in traditional methods.

[0077] In S2, the target spatiotemporal grid display range is determined based on the spatiotemporal grid encoding mapping rule table, the spatiotemporal grid level selected by the user, and the spatiotemporal grid display range selected by the user.

[0078] Among them, the spatiotemporal grid level selected by the user is the level selected by the user on the front-end page; the spatiotemporal grid display range selected by the user is the spatiotemporal grid display range selected by the user on the front-end page; the target spatiotemporal grid display range is the grid display range determined based on the spatiotemporal grid encoding mapping rule table, the spatiotemporal grid level selected by the user, and the spatiotemporal grid display range selected by the user on the front-end page, including specific latitude and longitude ranges, spatiotemporal grid level, grid ID, and other information; the target spatiotemporal grid display range is the spatiotemporal grid used to display marine environmental element data and / or marine environmental phenomenon information.

[0079] S2 includes the following steps S21-S25, see below. Figure 3 , Figure 3 This is a flowchart of a sub-step of S2 provided by the present invention. Any parts not described in detail in this step will be described in the sub-step.

[0080] In S21, after the front-end page receives the request to load the spatiotemporal grid information sign, it determines the grid identifier, spatiotemporal grid level, and spatiotemporal grid display range based on the user's selected operation.

[0081] For example, when a user selects a region or clicks on a spatiotemporal grid cell on the map interface of the front-end page, the front-end page will receive a request to load spatiotemporal grid information labels. The request typically includes the following information: the user's current map view range, such as latitude and longitude boundaries; the spatiotemporal grid level selected by the user, such as H3 level 7; the grid identifier selected by the user, such as grid ID; the spatiotemporal grid display range selected by the user; and the data type of marine environmental elements to be displayed, such as temperature and salinity.

[0082] In S22, the target spatiotemporal grid display range is determined based on the grid identifier, spatiotemporal grid level, and spatiotemporal grid display range.

[0083] In S23, in response to the request for grid coding of marine environmental elements, the grid data information contained in the target spatiotemporal grid is searched from the spatiotemporal grid coding mapping rule table based on the grid identifier.

[0084] For example, when a user clicks on the front-end page to issue a request for marine environmental element grid coding, the system responds by searching the spatiotemporal grid coding mapping rule table for grid point data information contained in the target spatiotemporal grid based on the grid ID. The grid point data information includes, but is not limited to, latitude and longitude. ,index ,value Information such as marine environmental elements (e.g., temperature, salinity, current velocity) and timestamps.

[0085] In S24, the statistical characteristics of marine environmental elements are calculated from the grid data information contained in the target spatiotemporal grid.

[0086] Among them, the statistical characteristics of marine environmental elements include the maximum, minimum, average, variance, and standard deviation of grid data information.

[0087] Calculate the maximum value of grid data information Minimum value ,average value ,variance Standard deviation Statistical characteristics are displayed on a spatiotemporal grid in the form of labels, and the grid data is recorded as follows: , This refers to the non-zero-padded portion of the corresponding grid point value in the matrix. The zero-padded portion is as follows: Figure 2 As shown, for the grid-covered portion, the calculation formula for the statistical characteristics of marine environmental elements is as follows:

[0088]

[0089] in, The maximum value of the grid data information. For gridded data, This represents the minimum value of the grid data information. This represents the average value of the grid data. The score of the target spatiotemporal grid. The variance of the grid data information. This represents the standard deviation of the grid data.

[0090] In S25, corresponding marine environmental element statistical characteristics are attached to each target spatiotemporal grid.

[0091] After calculating the maximum, minimum, average, variance, and standard deviation of the grid data information contained in the target spatiotemporal grid, the corresponding maximum, minimum, average, variance, and standard deviation are displayed in the form of labels on each target spatiotemporal grid.

[0092] The target spatiotemporal grid display range is determined based on the spatiotemporal grid level and display range selected by the user. This can accurately locate the marine environmental element data required by the user, avoid interference from invalid data, maximize the acquisition of effective information, reduce redundancy, improve data query and processing efficiency, and meet the needs of different users for data at different spatiotemporal scales.

[0093] In S3, query the marine environmental element data corresponding to the target spatiotemporal grid display range in the background database.

[0094] S3 includes the following steps S31-S33, see below. Figure 4 , Figure 4 This is a flowchart of a sub-step S3 provided by the present invention. Any parts not described in detail in this step will be described in the sub-step.

[0095] In S31, the marine environmental phenomenon data information corresponding to the target spatiotemporal grid display range is queried in the background database.

[0096] Among them, the data information on marine environmental phenomena includes the basic attributes of marine environmental phenomena and the dynamic symbolic representation of marine environmental phenomena. The basic attributes of marine environmental phenomena include the range, location, name, time of occurrence, duration and scale of marine environmental phenomena. The dynamic symbolic representation of marine environmental phenomena includes symbol type, dynamic changes, time series, spatial distribution and interactive functions.

[0097] For example, based on the grid ID, the database can be queried to obtain information such as the basic attributes of marine environmental phenomena corresponding to the target spatiotemporal grid display range, the range of occurrence of marine environmental phenomena, and the latitude and longitude range occupied by marine environmental phenomena.

[0098] In S32, the associated spatiotemporal grid is determined based on marine environmental phenomenon data.

[0099] Among them, the associated spatiotemporal grid is a spatiotemporal grid that intersects with the range of marine environmental phenomena.

[0100] For example, the grid ID of the associated spatiotemporal grid can be searched in the spatiotemporal grid coding mapping rule table of marine environmental elements based on the latitude and longitude range occupied by the marine environmental phenomenon.

[0101] In S33, the marine environmental element data corresponding to the associated spatiotemporal grid is determined based on the spatiotemporal grid coding mapping rule table, and used as the marine environmental element data corresponding to the target spatiotemporal grid display range.

[0102] In one implementation, after determining the marine environmental element data corresponding to the target spatiotemporal grid display range, the marine environmental element data is classified. Classifying the marine environmental element data includes: treating marine environmental phenomena corresponding to marine environmental element data involving one associated spatiotemporal grid as small-scale phenomena in the current spatiotemporal grid's representational sense; treating marine environmental phenomena corresponding to marine environmental element data involving two or more associated spatiotemporal grids as large-scale phenomena in the current spatiotemporal grid's representational sense; identifying small-scale phenomena on a single corresponding spatiotemporal grid, and identifying large-scale phenomena on multiple corresponding spatiotemporal grids.

[0103] Small-scale phenomena typically refer to phenomena with a small spatial scope and short temporal scale. These phenomena often occur within a local area, have a short duration, but may have high intensity and complexity. Large-scale phenomena typically refer to phenomena with a large spatial scope and long temporal scale. These phenomena often affect a wide area, have a long duration, and have a large spatial and temporal span.

[0104] Based on the spatiotemporal grid coding mapping rule table, users can select specific spatiotemporal grid levels and display ranges to accurately obtain marine environmental element data within the target spatiotemporal range. This can better reflect the spatiotemporal correlation between marine environmental element data and marine environmental phenomena, and help to more accurately reflect the complexity and dynamic characteristics of the marine environment.

[0105] In S4, feature extraction is performed on marine environmental element data to obtain marine environmental element feature vectors.

[0106] S4 includes the following steps S41-S43, see below. Figure 5 , Figure 5 This is a flowchart of a sub-step S4 provided by the present invention. Any parts not described in detail in this step will be described in the sub-step.

[0107] In S41, the Fourier spatial transformation of marine environmental element data is performed to obtain the spectrum.

[0108] Data on marine environmental elements Perform Fourier spatial transform: assuming the spatiotemporal grid contains feature data. It is irregular, see reference. Figure 6 , Figure 6 This is a schematic diagram illustrating feature extraction and classification of marine environmental element data provided by the present invention. First, regions in the matrix without data are padded with zeros. Then, the spectrum is obtained by calculating the two-dimensional Fourier transform formula. The advantage of doing this is that the matrix has effective values ​​everywhere in the next step of convolutional feature extraction, and it can also reflect the spatial distribution of marine environmental element data. The formula for calculating the spectrum of marine environmental element data by performing Fourier spatial transformation is shown below.

[0109]

[0110] in, This is the spectrum obtained by performing a Fourier spatial transform on marine environmental element data. It is marine environmental element data Maximum width after padding with zeros It is marine environmental element data Maximum height after padding with zeros For marine environmental element data, It is an imaginary number. u and v These are coordinates in the frequency domain.

[0111] In S42, the spectrum is convolved by a neural network to obtain the feature vector of a single-grid, single-element marine environmental element.

[0112] Among them, the feature vector of marine environmental elements in a single grid and single element is the feature vector of a specific marine environmental element in a single spatiotemporal grid unit.

[0113] For example, refer to Figure 6 Using feature extraction neural networks For the spectrum Convolution is performed to obtain feature maps of marine environmental element data. Finally Each convolutional kernel corresponds to Flattening and connecting the feature maps yields the feature vectors of marine environmental elements. The calculation formula is as follows:

[0114]

[0115] in, It refers to the number and size of the convolution kernels. It is learnable. Convolution kernels are used to obtain the frequency variation characteristics of marine environmental element spectra from small to large scales. Feature map of marine environmental elements; Flat It is a flattening operation, which flattens the feature map. Zhan Cheng The form; Concat It is a join operation that flattens all feature maps. Connect to form a 1-dimensional feature vector .

[0116] In S43, the feature vectors of marine environmental elements are placed into the spatiotemporal grid coding mapping rule table.

[0117] For example, the obtained single-grid single-feature feature vector is put into the spatiotemporal grid coding mapping rule table. Putting the single-grid single-feature feature vector into the spatiotemporal grid coding mapping rule table means associating the feature vector of a specific marine environmental element (such as temperature, salinity, etc.) of each spatiotemporal grid unit (such as H3 grid unit) with the code of the spatiotemporal grid unit and storing it in a database or data table. The establishment of this mapping relationship can facilitate subsequent query, analysis and visualization operations.

[0118] By extracting features from marine environmental element data and performing feature operations, we can deeply explore the intrinsic relationship between marine environmental element data and marine environmental phenomena. This can overcome the shortcomings of traditional methods in reflecting the relationship between marine environmental element data and phenomena, obtain more comprehensive and accurate information on marine environmental phenomena, improve the ability to comprehensively model and process complex marine environments and express information, and adapt to diverse business needs based on the ocean.

[0119] In S5, feature extraction is performed on the feature vectors of marine environmental elements, and marine environmental phenomena corresponding to the marine environmental element data are identified based on the extracted features.

[0120] S5 includes the following steps S51-S53, see below. Figure 7 , Figure 7 This is a flowchart of a sub-step S5 provided by the present invention. Any parts not described in detail in this step will be described in the sub-step.

[0121] In S51, in response to the request to enhance the characteristics of marine environmental phenomena, a weighted average of the feature vectors of each marine environmental element is performed to obtain a single spatiotemporal grid comprehensive feature vector.

[0122] Among them, the comprehensive feature vector of a single spatiotemporal grid is the comprehensive feature vector of all elements corresponding to a single spatiotemporal grid, or it can be simply referred to as the single grid feature vector.

[0123] For example, refer to Figure 6 Use learnable weighted parameters Feature vectors of marine environmental elements Perform a weighted average, such as Figure 6 In , To obtain the comprehensive feature vector of all elements corresponding to a single spatiotemporal grid. , which is calculated as follows:

[0124]

[0125] in, This is the comprehensive feature vector of all elements corresponding to a single spatiotemporal grid. Marine environmental phenomena The total number of elements with valid data in the associated spatiotemporal grid. For learnable weighted parameters, This represents the feature vector of marine environmental elements.

[0126] In S52, multiple spatiotemporal grids are included in the large-scale phenomenon in the current spatiotemporal grid sense. The weighted average of the comprehensive feature vectors of all single spatiotemporal grids of multiple spatiotemporal grids is performed using learnable weighted weight parameters to obtain the feature vectors associated with marine environmental phenomena.

[0127] Among them, the feature vector associated with marine environmental phenomena is the comprehensive feature vector of all spatiotemporal grids corresponding to marine environmental phenomena.

[0128] For example, learnable weighted parameters can be used. Synthetic eigenvectors of a single spatiotemporal grid Perform a weighted average, such as Figure 6 In , This yields the comprehensive feature vectors of all spatiotemporal grids corresponding to marine environmental phenomena, i.e., the feature vectors associated with marine environmental phenomena. , which is calculated as follows:

[0129]

[0130] in, This is a feature vector associated with marine environmental phenomena. Marine environmental phenomena The number of all associated spatiotemporal grids, For learnable weighted parameters, It is a single spatiotemporal grid integrated feature vector.

[0131] In S53, a classifier is used to classify the feature vectors associated with marine environmental phenomena, thereby identifying the marine environmental phenomena corresponding to the marine environmental element data.

[0132] For example, refer to Figure 6 Using classifiers Feature vectors associated with all marine environmental phenomena To perform clustering, such as using a K-means clustering classifier, the first step is to determine the class centers based on the differences in the observed phenomena. Then, softmax loss and center loss are calculated to minimize the distance between feature vectors associated with marine environmental phenomena of the same category and maximize the distance between feature vectors associated with marine environmental phenomena of different categories. The loss function is shown below:

[0133]

[0134] in, As the class center, It is a marine environmental phenomenon Quantity, This is the feature vector associated with all marine environmental phenomena. For the Feature vectors associated with each sample The softmax loss function is used. The center loss function is used. It is the last layer of the feature extraction network. Convert the feature vector into a category label; For the last layer of the feature extraction network, the corresponding category The weights; The category corresponding to the last layer of the feature extraction network. The weight, It is Total number of samples in all categories The number of categories identified in this clustering; For the currently calculated number of One sample, For the first The category label corresponding to each sample For the first The feature vector extracted from each sample For the last layer of the feature extraction network Category bias; This is the label for all samples of a specific category in this clustering. ， For all corresponding features of the last layer feature extraction network Bias of class samples; It is a hyperparameter, which can be set to 0.5 in one implementation.

[0135] By performing feature operations on the feature vectors of marine environmental elements, it is possible to effectively process multi-source heterogeneous marine environmental data, overcoming the problems of traditional methods in data alignment, integration, and narrow application scope. Through unified spatiotemporal grid coding and mapping rules, data from different sources and of different types can be effectively integrated, broadening the application scope of data and improving the utilization efficiency of data resources.

[0136] The marine environmental data processing method, apparatus, electronic device, and readable storage medium provided by this invention include: introducing a global spatiotemporal grid partitioning standard; constructing a spatiotemporal grid coding mapping rule table for marine environmental elements based on the global spatiotemporal grid partitioning standard; determining a target spatiotemporal grid display range based on the spatiotemporal grid coding mapping rule table, the user-selected spatiotemporal grid level, and the user-selected spatiotemporal grid display range; querying marine environmental element data corresponding to the target spatiotemporal grid display range in a background database; extracting features from the marine environmental element data to obtain marine environmental element feature vectors; and performing feature operations on the marine environmental element feature vectors to obtain the marine environmental phenomena corresponding to the marine environmental element data.

[0137] By introducing a global spatiotemporal grid partitioning standard and constructing a spatiotemporal grid coding mapping rule table, a systematic framework for marine environmental data processing is established. This ensures that subsequent data processing and analysis are based on a unified and standardized standard, enhancing the universality and comparability of the data. It effectively improves the problems of difficulty in aligning and integrating multi-source heterogeneous data in traditional methods. Feature extraction of marine environmental element data yields feature vectors, enabling in-depth exploration of the intrinsic relationship between marine environmental element data and marine environmental phenomena. This compensates for the shortcomings of traditional methods in reflecting the insufficient correlation between marine environmental element data and phenomena, allowing for a more comprehensive and accurate acquisition of marine environmental phenomena. It also enhances the comprehensive modeling and information expression capabilities for complex marine environments, and can adapt to diverse marine-based business needs.

[0138] The marine environmental data processing apparatus provided by the present invention is described below. The marine environmental data processing apparatus described below can be referred to in correspondence with the marine environmental data processing method described above.

[0139] Figure 8 This is a schematic diagram of a marine environmental data processing device provided by the present invention. This marine environmental data processing device is applied in devices such as servers, desktop computers, and laptops. Figure 8 The marine environmental data processing device 80 includes a construction module 801, a grid determination module 802, a query module 803, a feature extraction module 804, and a feature calculation module 805.

[0140] The construction module 801 is configured to construct a spatiotemporal grid coding mapping rule table for marine environmental elements based on the global spatiotemporal grid subdivision standard; the spatiotemporal grid coding mapping rule table contains the correspondence between the spatiotemporal grid and the marine environmental element data and marine environmental phenomenon information of the corresponding region;

[0141] The grid determination module 802 is configured to determine the target spatiotemporal grid display range based on the spatiotemporal grid encoding mapping rule table, the spatiotemporal grid level selected by the user, and the spatiotemporal grid display range selected by the user.

[0142] The query module 803 is configured to query the marine environmental element data corresponding to the target spatiotemporal grid display range in the background database;

[0143] The feature extraction module 804 is configured to extract features from the marine environmental element data to obtain marine environmental element feature vectors.

[0144] The feature processing module 805 is configured to extract features from the feature vector of the marine environmental elements and identify marine environmental phenomenon information corresponding to the marine environmental element data based on the extracted features.

[0145] Figure 9An example of a physical structure diagram of an electronic device is shown below. Figure 9 As shown, the electronic device may include: a processor 910, a communication interface 920, a memory 930, and a communication bus 940, wherein the processor 910, the communication interface 920, and the memory 930 communicate with each other through the communication bus 940. The processor 910 can call logical instructions in the memory 930 to execute a marine environmental data processing method, which includes: constructing a spatiotemporal grid encoding mapping rule table for marine environmental elements based on a global spatiotemporal grid subdivision standard; the spatiotemporal grid encoding mapping rule table containing the correspondence between spatiotemporal grids and marine environmental element data and marine environmental phenomenon information of corresponding regions; determining a target spatiotemporal grid display range based on the spatiotemporal grid encoding mapping rule table, the spatiotemporal grid level selected by the user, and the spatiotemporal grid display range selected by the user; querying the marine environmental element data corresponding to the target spatiotemporal grid display range in a background database; extracting features from the marine environmental element data to obtain marine environmental element feature vectors; extracting features from the marine environmental element feature vectors, and identifying marine environmental phenomenon information corresponding to the marine environmental element data based on the extracted features.

[0146] Furthermore, the logical instructions in the aforementioned memory 930 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0147] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the marine environmental data processing methods provided by the above methods. The method includes: constructing a spatiotemporal grid coding mapping rule table for marine environmental elements based on a global spatiotemporal grid partitioning standard; the spatiotemporal grid coding mapping rule table containing the correspondence between spatiotemporal grids and marine environmental element data and marine environmental phenomenon information of corresponding regions; determining a target spatiotemporal grid display range based on the spatiotemporal grid coding mapping rule table, the spatiotemporal grid level selected by the user, and the spatiotemporal grid display range selected by the user; querying the marine environmental element data corresponding to the target spatiotemporal grid display range in a background database; extracting features from the marine environmental element data to obtain a marine environmental element feature vector; extracting features from the marine environmental element feature vector, and identifying the marine environmental phenomenon information corresponding to the marine environmental element data based on the extracted features.

[0148] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the marine environmental data processing method provided by the above methods. The method includes: constructing a spatiotemporal grid coding mapping rule table for marine environmental elements based on a global spatiotemporal grid partitioning standard; the spatiotemporal grid coding mapping rule table containing the correspondence between spatiotemporal grids and marine environmental element data and marine environmental phenomenon information of corresponding regions; determining a target spatiotemporal grid display range based on the spatiotemporal grid coding mapping rule table, the spatiotemporal grid level selected by the user, and the spatiotemporal grid display range selected by the user; querying the marine environmental element data corresponding to the target spatiotemporal grid display range in a background database; performing feature extraction on the marine environmental element data to obtain a marine environmental element feature vector; performing feature extraction on the marine environmental element feature vector, and identifying the marine environmental phenomenon information corresponding to the marine environmental element data based on the extracted features.

[0149] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0150] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0151] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for processing marine environmental data, characterized in that, include: A spatiotemporal grid coding mapping rule table for marine environmental elements is constructed based on the global spatiotemporal grid subdivision standard; The spatiotemporal grid coding mapping rule table contains the correspondence between spatiotemporal grids and marine environmental element data and marine environmental phenomenon information of the corresponding regions; The target spatiotemporal grid display range is determined based on the spatiotemporal grid encoding mapping rule table, the spatiotemporal grid level selected by the user, and the spatiotemporal grid display range selected by the user. Query the marine environmental element data corresponding to the target spatiotemporal grid display range in the background database; Feature extraction is performed on the marine environmental element data to obtain marine environmental element feature vectors; Feature extraction is performed on the feature vector of the marine environmental elements, and marine environmental phenomena corresponding to the marine environmental element data are identified based on the extracted features; the extracted features are feature vectors associated with marine environmental phenomena, and the feature vectors associated with marine environmental phenomena are comprehensive feature vectors of all spatiotemporal grids corresponding to the marine environmental phenomena. Feature vectors associated with all marine environmental phenomena using a classifier Clustering is performed using the K-means clustering classifier. First, the cluster centers are determined based on the differences in the phenomena observed. Then, softmax loss and center loss are calculated to minimize the distance between feature vectors associated with marine environmental phenomena of the same category and maximize the distance between feature vectors associated with marine environmental phenomena of different categories. The loss function is shown below: ； in, As the class center, It is a marine environmental phenomenon Quantity, This is the feature vector associated with all marine environmental phenomena. For the first Feature vectors associated with each sample The softmax loss function is used. The center loss function is used. It is the last layer of the feature extraction network. Convert the feature vector into a category label; For the last layer of the feature extraction network, the corresponding category The weights; The category corresponding to the last layer of the feature extraction network. The weight, It is the first Total number of samples in all categories The number of categories identified in this clustering; For the currently calculated number of One sample, For the first The category label corresponding to each sample For the first The feature vector extracted from each sample For the last layer of the feature extraction network Category bias; This is the label for all samples of a specific category in this clustering. ， For all corresponding features of the last layer of the feature extraction network Bias of class samples; It is a hyperparameter.

2. The method according to claim 1, characterized in that, The process of determining the target spatiotemporal grid display range based on the spatiotemporal grid encoding mapping rule table, the spatiotemporal grid level selected by the user, and the spatiotemporal grid display range selected by the user includes: After receiving the spatiotemporal grid information sign loading request on the front-end page, the grid identifier, the spatiotemporal grid level, and the spatiotemporal grid display range are determined based on the user's selected operation. The target spatiotemporal grid display range is determined based on the grid identifier, the spatiotemporal grid level, and the spatiotemporal grid display range.

3. The method according to claim 2, characterized in that, After determining the target spatiotemporal grid display range based on the grid identifier, the spatiotemporal grid level, and the spatiotemporal grid display range, the process includes: In response to the request for grid coding of marine environmental elements, the grid data information contained in the target spatiotemporal grid is searched from the spatiotemporal grid coding mapping rule table based on the grid identifier; Calculate the statistical characteristics of marine environmental elements in the grid data information contained in the target spatiotemporal grid; the statistical characteristics of marine environmental elements include the maximum value, minimum value, average value, variance, and standard deviation of the grid data information; The corresponding statistical characteristics of the marine environmental elements are attached to each of the target spatiotemporal grids.

4. The method according to claim 1, characterized in that, The step of querying the marine environmental element data corresponding to the target spatiotemporal grid display range in the background database includes: Query the marine environmental phenomenon data information corresponding to the target spatiotemporal grid display range in the background database; the marine environmental phenomenon data information includes the basic attributes of marine environmental phenomena and the dynamic symbolic representation of marine environmental phenomena, and the basic attributes of marine environmental phenomena include the occurrence range of marine environmental phenomena. Based on the marine environmental phenomenon data, a related spatiotemporal grid is determined; the related spatiotemporal grid is a spatiotemporal grid that intersects with the range of occurrence of the marine environmental phenomenon; Based on the spatiotemporal grid coding mapping rule table, the marine environmental element data corresponding to the associated spatiotemporal grid is determined, and used as the marine environmental element data corresponding to the target spatiotemporal grid display range.

5. The method according to claim 4, characterized in that, Also includes: The marine environmental element data are classified; The classification of the marine environmental element data includes: Marine environmental phenomena corresponding to marine environmental element data involving one of the aforementioned spatiotemporal grids are considered as small-scale phenomena in the current spatiotemporal grid representation sense; Marine environmental phenomena corresponding to marine environmental element data involving two or more of the aforementioned associated spatiotemporal grids are considered as large-scale phenomena in the current spatiotemporal grid representation sense. The small-scale phenomena are identified on a single corresponding spatiotemporal grid, and the large-scale phenomena are identified on multiple corresponding spatiotemporal grids.

6. The method according to claim 1, characterized in that, The step of extracting features from the marine environmental element data to obtain marine environmental element feature vectors includes: The spectrum was obtained by performing a Fourier spatial transform on the marine environmental element data. The spectrum is convolved by a neural network to obtain a single-grid, single-element marine environmental element feature vector; the single-grid, single-element marine environmental element feature vector is the feature vector of a specific marine environmental element in a single spatiotemporal grid unit. The feature vectors of the marine environmental elements are placed into the spatiotemporal grid coding mapping rule table.

7. The method according to claim 1, characterized in that, The step of extracting features from the feature vectors of the marine environmental elements and identifying the marine environmental phenomena corresponding to the marine environmental element data based on the extracted features includes: In response to the request to enhance the characteristics of marine environmental phenomena, a weighted average is performed on the feature vectors of each marine environmental element to obtain a single spatiotemporal grid comprehensive feature vector; the single spatiotemporal grid comprehensive feature vector is the comprehensive feature vector of all elements corresponding to a single spatiotemporal grid; In the context of large-scale phenomena within the current spatiotemporal grid, multiple spatiotemporal grids are included. A weighted average of the comprehensive feature vectors of all the single spatiotemporal grids of the multiple spatiotemporal grids is performed using learnable weighting parameters to obtain the feature vectors associated with marine environmental phenomena. The feature vectors associated with the marine environmental phenomena are classified using a classifier to identify the marine environmental phenomena corresponding to the marine environmental element data.

8. The method according to claim 1, characterized in that, The global spatiotemporal grid subdivision standard is the BeiDou grid location code or the H3 hexagonal standard grid.

9. A marine environmental data processing device, characterized in that, include: The module constructs a spatiotemporal grid coding mapping rule table for marine environmental elements based on the global spatiotemporal grid subdivision standard; The spatiotemporal grid coding mapping rule table contains the correspondence between spatiotemporal grids and marine environmental element data and marine environmental phenomenon information of the corresponding regions; The grid determination module determines the target spatiotemporal grid display range based on the spatiotemporal grid encoding mapping rule table, the spatiotemporal grid level selected by the user, and the spatiotemporal grid display range selected by the user. The query module retrieves marine environmental element data corresponding to the target spatiotemporal grid display range from the background database; The feature extraction module extracts features from the marine environmental element data to obtain marine environmental element feature vectors. The feature processing module extracts features from the feature vectors of the marine environmental elements and identifies the marine environmental phenomena corresponding to the marine environmental element data based on the extracted features. The extracted features are feature vectors associated with the marine environmental phenomena, and the feature vectors associated with the marine environmental phenomena are the comprehensive feature vectors of all spatiotemporal grids corresponding to the marine environmental phenomena. Using classifiers Feature vectors associated with all marine environmental phenomena Clustering is performed using the K-means clustering classifier. First, the cluster centers are determined based on the differences in the phenomena observed. Then, softmax loss and center loss are calculated to minimize the distance between feature vectors associated with marine environmental phenomena of the same category and maximize the distance between feature vectors associated with marine environmental phenomena of different categories. The loss function is shown below: ； in, As the class center, It is a marine environmental phenomenon Quantity, This is the feature vector associated with all marine environmental phenomena. For the first Feature vectors associated with each sample The softmax loss function is used. The center loss function is used. It is the last layer of the feature extraction network. Convert the feature vector into a category label; For the last layer of the feature extraction network, the corresponding category The weights; The category corresponding to the last layer of the feature extraction network. The weight, It is the first Total number of samples in all categories The number of categories identified in this clustering; For the currently calculated number of One sample, For the first The category label corresponding to each sample For the first The feature vector extracted from each sample For the last layer of the feature extraction network Category bias; This is the label for all samples of a specific category in this clustering. ， For all corresponding features of the last layer of the feature extraction network Bias of class samples; It is a hyperparameter.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the marine environmental data processing method as described in any one of claims 1 to 8.

11. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the marine environmental data processing method as described in any one of claims 1 to 8.

12. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the marine environmental data processing method as described in any one of claims 1 to 8.

Images