Intelligent acquisition methods, systems and storage media for multi-dimensional marine information

Abstract

This invention relates to the field of marine environmental monitoring technology, and discloses an intelligent method, system, and storage medium for acquiring multi-dimensional marine information. The method includes: simultaneously acquiring environmental parameters at multiple monitoring points and depth layers to construct a hydrological feature dataset; calculating spatial correlation and dividing regions into high-variability and low-variability areas; constructing a device spacing model based on information density classification results and generating a deployment scheme using an optimization algorithm; identifying significant changes and dynamically updating deployment parameters based on real-time monitoring data; and issuing commands to devices via primary and backup wireless channels to complete deployment. This invention improves the adaptability and deployment efficiency of marine environmental sensing, and is suitable for intelligent monitoring applications in complex and variable sea areas.

Description

Technical Field

[0001] This invention relates to the field of marine environmental monitoring and information acquisition technology, specifically to an intelligent acquisition method, system and storage medium for multi-dimensional marine information based on dynamic analysis of spatial correlation and information density, belonging to the interdisciplinary technical fields of marine information perception, environmental monitoring, spatial optimization deployment and intelligent communication control. Background Technology

[0002] With the increasing demand for marine resource development, marine disaster early warning, and climate change research, higher requirements are being placed on the real-time and accurate collection of multi-dimensional marine environmental information. Traditional marine information collection mainly relies on fixed measuring equipment, deployed vessels, or remote sensing systems, primarily collecting hydrological parameters such as temperature, salinity, current velocity, and water depth. While these methods have played an important role in basic marine observation, they have gradually revealed problems such as rigid deployment, insufficient sampling resolution, and response lag in large-scale, multi-dimensional, and high-frequency dynamic monitoring.

[0003] In recent years, with the development of wireless communication, spatial data modeling, and optimization algorithms, academia and engineering have attempted to introduce technologies such as sensor networks, adaptive control, and machine learning into marine environmental monitoring systems. One type of research focuses on improving data acquisition frequency and node deployment flexibility, such as dynamic deployment methods based on mobile observation platforms (e.g., AUVs, buoys); the other type focuses on establishing spatial optimization models, attempting to identify key areas and perform fine-grained deployment through methods such as information entropy, kriging interpolation, and autocorrelation coefficients.

[0004] However, a complete methodology that can simultaneously address spatial correlation analysis, information density modeling, and resource optimization is currently lacking. Existing systems often suffer from the following main technical problems:

[0005] 1. The deployment scheme is too static and generalized: the collection points are often divided based on experience or rules, which makes it difficult to adapt to the characteristics of the marine hydrological environment that changes drastically with time and space, resulting in insufficient monitoring in key areas and waste of resources in redundant areas.

[0006] 2. Lack of dynamic adaptive adjustment mechanism: Most existing deployments are one-time deployments, lacking the ability to update and optimize based on real-time data feedback;

[0007] Large number of devices and low deployment efficiency: Under the requirement of balancing coverage accuracy and monitoring range, a large number of data collection nodes are often required, which increases deployment costs and system maintenance pressure;

[0008] 4. Separation of data acquisition, control and communication systems, lack of intelligent linkage: Existing solutions mostly rely on manual intervention and adjustment, making it difficult to achieve real-time response and closed-loop control.

[0009] Therefore, there is an urgent need for a dynamic optimization deployment technology based on real-time data, which can adaptively adjust the deployment mode of the acquisition equipment by combining key environmental characteristics such as spatial correlation and information density distribution, so as to reduce the number of devices, improve deployment efficiency and system intelligence while ensuring monitoring accuracy. This is the core technical problem that this invention aims to solve. Summary of the Invention

[0010] To address the problems of rigid deployment, lack of dynamic adaptability, and low deployment efficiency in existing marine information acquisition methods, this invention provides an intelligent acquisition method, system, and storage medium for multi-dimensional marine information. It can combine the spatial variation characteristics of the marine environment and the differences in information density to achieve intelligent optimization and dynamic adjustment of equipment deployment while ensuring monitoring accuracy.

[0011] In a first aspect, the present invention provides an intelligent method for acquiring multi-dimensional marine information, the method comprising the following steps:

[0012] S1. Simultaneously collect environmental parameters at multiple measuring points and multiple depth layers in the sea area. The environmental parameters include temperature, salinity, current velocity and water depth, forming a hydrological feature dataset with time series and spatial coordinates.

[0013] S2. Based on the hydrological feature dataset, calculate the spatial autocorrelation coefficient between each measuring point, and generate a spatial correlation matrix using a spatial interpolation algorithm;

[0014] S3. Based on the spatial correlation matrix, the sea area is divided into high-variability regions and low-variability regions. When the correlation coefficient between each measuring point in the sea area and multiple other measuring points in the spatial neighborhood of the measuring point is lower than a preset threshold, the area where the measuring point is located is determined to be a high-variability region; otherwise, it is determined to be a low-variability region.

[0015] S4. Calculate the information density index of the high-variability region and the low-variability region, and classify them into information-dense and sparse regions.

[0016] S5. Construct an inverse proportional relationship model between device spacing and information density, and set a smaller device spacing for areas with high variability and dense information.

[0017] S6. Use optimization algorithms to optimize the number of devices and their layout coordinates while meeting the monitoring coverage requirements, and generate a layout plan.

[0018] S7. Acquire real-time marine environmental monitoring data. If significant changes in hydrological characteristics are detected, recalculate the device spacing and update the deployment plan.

[0019] S8. Send updated working parameters and location instructions to the acquisition device through the wireless communication system to complete the dynamic optimization deployment.

[0020] In conjunction with the first aspect, in the first implementation of the first aspect of this application, S1 includes a data anomaly detection mechanism. When a parameter of a certain measuring point continuously exceeds a preset threshold in the time or depth dimension, high-frequency sampling of the measuring point is triggered, and the sampling results are used to update the hydrological feature dataset.

[0021] In S2, the spatial correlation calculation uses the Moran index or Geary-C index as an evaluation method to measure the degree of synchronous change of hydrological parameters between adjacent measuring points.

[0022] The spatial interpolation algorithm is the Kriging interpolation method. It performs local smoothing in areas with large parameter gradient changes. Based on the known parameter values ​​of the measuring points and their spatial coordinates, the Kriging interpolation method is used to estimate the unmeasured areas. Based on the known measuring point weights and spatial variability function, the Kriging interpolation method calculates the predicted hydrological parameters at the target location, thereby realizing the spatial reconstruction of the parameter field.

[0023] In conjunction with the first aspect, in the second implementation of the first aspect of this application, in step S4, based on the division results of the high-variability region and the low-variability region, a two-dimensional grid structure covering the entire monitoring sea area is constructed. Each grid cell is used as a calculation unit, and hydrological parameter data in any grid cell is extracted. The statistical characteristics of the hydrological parameter data are used as input, and the information entropy calculation method is used to evaluate the data distribution uncertainty of any grid cell. The data distribution uncertainty is used as the information density index of any grid cell.

[0024] Clustering algorithms are used to cluster all information density indicators. Based on the clustering results, all grid cells are divided into different information density categories to obtain preliminary classification results.

[0025] The information density index difference and category consistency relationship between each grid cell and its multiple directly adjacent grid cells are detected. If the information density index of any grid cell is close to that of any adjacent grid cell, but the categories are inconsistent, then any grid cell is reclassified and merged into the main category of the adjacent grid cells.

[0026] In conjunction with the first aspect, in the third implementation of the first aspect of this application, in step S5, the inverse proportional relationship model between the device spacing and the information density satisfies d = k / ρ, where d is the device spacing, ρ is the information density index, and k is a proportional constant set according to the target coverage.

[0027] In conjunction with the first aspect, in the fourth implementation of the first aspect of this application, the highly variable and information-dense region is identified, and the corresponding device spacing is compared with a preset minimum deployment spacing threshold. If the device spacing is less than the minimum deployment spacing threshold, a parameter optimization mechanism is triggered, and the proportional constant of the inverse proportional relationship model is automatically iteratively updated with the current information density index value as input.

[0028] In conjunction with the first aspect, in the fifth implementation of the first aspect of this application, in step S6, the results of the sea area division, the equipment spacing parameters, the information density weights, and the boundary constraints are input into the optimization algorithm to search for the coordinate set corresponding to the minimum number of equipment within the deployment space. Specifically, this includes:

[0029] A genetic algorithm is used as a global search tool. Multiple deployment combinations are generated by initializing the population. The fitness of each combination is evaluated based on the total number of devices, the integrity of the coverage area, the rationality of the spacing, and the connectivity. The deployment combination with the fewest devices and the best coverage is selected.

[0030] Based on the aforementioned layout combination, the layout structure is iteratively evolved through crossover and mutation operations. In each iteration, the local optimal solution is dynamically updated until the global optimal solution or the convergence condition is met. The final output solution is then used as the layout scheme.

[0031] In conjunction with the first aspect, in the sixth implementation of the first aspect of this application, in step S7, any hydrological parameter is extracted, its time series data is obtained, the parameter change trend and multiple historical change trends are extracted through a sliding window mechanism, and a change threshold is generated based on the average value and standard deviation of the multiple historical change trends. If the parameter change trend exceeds the change threshold, it is determined that the hydrological characteristics have changed significantly.

[0032] If the monitoring data of any region shows significant changes in hydrological characteristics over multiple consecutive periods, the region is marked as a high-dynamic state.

[0033] Priority parameters are assigned to each region based on its environmental sensitivity, hydrological change frequency, and monitoring target weight. For high-dynamic regions where the priority parameters are greater than preset values, a release point update process is initiated.

[0034] In conjunction with the first aspect, in the seventh implementation of the first aspect of this application, in step S8, the wireless communication system includes a main communication channel and a backup communication channel. If the main communication channel fails to transmit parameters, it automatically switches to the backup communication channel for retransmission until the device confirms successful reception.

[0035] Secondly, this application provides an intelligent system for acquiring multi-dimensional marine information, the system comprising:

[0036] The hydrological sensor network module is used to collect hydrological parameters at multiple depth levels and perform anomaly detection and data updates.

[0037] The data processing and analysis module is used to construct hydrological feature datasets, generate spatial correlation matrices and information density classification maps;

[0038] The deployment optimization module is used to output the optimal equipment deployment scheme based on the inverse proportional relationship model and optimization algorithm.

[0039] The real-time monitoring and adaptive update module is used to analyze hydrological change trends and dynamically trigger the update of the release points;

[0040] The communication management module is used to send updated parameters and location information to the data acquisition device and complete the receipt confirmation.

[0041] A third aspect of this application provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the aforementioned intelligent acquisition method for multidimensional marine information.

[0042] Compared with the prior art, the beneficial effects of the technical solution of this application are at least as follows:

[0043] 1. Dynamic adaptive sampling point deployment enhances monitoring accuracy and responsiveness: This invention constructs a marine environmental spatial feature model based on spatial autocorrelation and real-time changes in information density, enabling the sampling point deployment to be flexibly adjusted according to changes in sea conditions, significantly improving the monitoring coverage of key areas and the overall system response capability.

[0044] 2. Optimize equipment resources and reduce deployment costs: By establishing an inverse proportional relationship model between equipment spacing and information density, and combining optimization algorithms to control the number and location of data points, the total number of data collection devices is reduced while ensuring coverage, thereby reducing equipment investment, operation and maintenance costs and energy consumption.

[0045] 3. Intelligent data acquisition strategy to achieve closed-loop control throughout the entire lifecycle: Combining real-time data monitoring and anomaly detection mechanisms, this invention can automatically trigger the update of the installation scheme and parameter reconfiguration, and realize remote dynamic control of the acquisition nodes with the help of the communication system, thereby improving the intelligence level and operational stability of the entire system.

[0046] 4. Versatility and scalability: The proposed technical framework has good versatility and can be used for fixed platforms and buoy systems, as well as for the deployment planning of mobile platforms such as AUVs. It also has the ability to be extended to multiple types of marine monitoring systems. Attached Figure Description

[0047] Figure 1 This is a flowchart of an intelligent method for acquiring multi-dimensional marine information according to the present invention;

[0048] Figure 2 This is a schematic diagram of an intelligent marine multi-dimensional information acquisition system according to the present invention. Detailed Implementation

[0049] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings and examples. However, those skilled in the art should understand that these embodiments are only used to illustrate the present invention and do not constitute a limitation on the scope of protection of the present invention.

[0050] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 In this embodiment, a method for intelligent acquisition of multi-dimensional marine information may specifically include:

[0051] Step S1: Simultaneously collect environmental parameters at multiple measuring points and multiple depth layers in the sea area. The environmental parameters include temperature, salinity, current velocity and water depth, forming a hydrological feature dataset with time series and spatial coordinates.

[0052] In one specific embodiment, S1 includes a data anomaly detection mechanism. When a parameter of a certain measuring point continuously exceeds a preset threshold in the time or depth dimension, high-frequency sampling of the measuring point is triggered, and the sampling results are used to update the hydrological feature dataset.

[0053] Specifically, in one embodiment of the present invention, a sensor network is deployed at multiple measuring points and multiple depth layers in the sea area. Each node is used to synchronously collect environmental parameters such as temperature, salinity, current velocity, and water depth. The collected data is recorded in time series form and accompanied by three-dimensional spatial coordinate information, forming a structured hydrological feature dataset. Each data record includes a measuring point number, spatial location, timestamp, and corresponding parameter value.

[0054] The hydrological feature dataset is dynamically updated based on a pre-defined anomaly detection mechanism. The system establishes judgment logic based on time-varying trends and depth differences for various collected parameters, setting preset anomaly thresholds. When an environmental parameter at a measuring point continuously exceeds this threshold over time, or when the parameter difference between its depth layer and adjacent depth layers at the same time point exceeds a set threshold, the system determines that the measuring point is in an abnormal state. When the triggering conditions are met, the control module adjusts the sampling frequency for that measuring point, performing high-frequency sampling to acquire multiple sets of continuous data within a short period.

[0055] The newly added data after the trigger is merged with the original dataset to form an updated dataset containing data from the anomaly phase. This updated dataset is directly available for use by the spatial correlation calculation module and the interpolation module to improve the accuracy of continuity assessment for anomaly regions in the spatial model. A closed-loop control link is formed between the anomaly detection module and the sampling control module, enabling adaptive adjustment of sampling behavior driven by dynamic data changes.

[0056] In this implementation, the anomaly detection mechanism and high-frequency sampling strategy functionally form an integrated data feedback control mechanism with the hydrological data acquisition system. It dynamically adjusts the sampling frequency of target monitoring points through data change identification, enabling the system to prioritize sampling accuracy and temporal density in highly variable areas under resource constraints. The high-frequency data sampling results support subsequent site optimization and information density model construction, constituting a core means of addressing insufficient monitoring and uneven data distribution in key areas.

[0057] Step S2: Based on the hydrological feature dataset, calculate the spatial autocorrelation coefficient between each measuring point, and generate a spatial correlation matrix using a spatial interpolation algorithm.

[0058] In one specific embodiment, in S2, the spatial correlation calculation uses the Moran index or the Geary-C index as an evaluation method to measure the degree of synchronous change of hydrological parameters between adjacent measuring points;

[0059] The spatial interpolation algorithm is the Kriging interpolation method. It performs local smoothing in areas with large parameter gradient changes. Based on the known parameter values ​​of the measuring points and their spatial coordinates, the Kriging interpolation method is used to estimate the unmeasured areas. Based on the known measuring point weights and spatial variability function, the Kriging interpolation method calculates the predicted hydrological parameters at the target location, thereby realizing the spatial reconstruction of the parameter field.

[0060] Specifically, based on the constructed hydrological feature dataset, the spatial structural relationships between various measuring points are analyzed to obtain a spatial correlation matrix reflecting the continuity of marine hydrological parameters and the variability of the regional environment. The hydrological feature dataset is a structured collection containing parameter values ​​such as temperature, salinity, current velocity, and water depth at multiple measuring points at different times and depths, along with corresponding timestamps and spatial coordinates. The system extracts data fields with spatial correlation attributes from this dataset and calculates spatial autocorrelation based on the geographical relationships between measuring points and the trends in parameter value changes. The spatial autocorrelation calculation uses the Moran index or the Geary-C index as evaluation methods to measure the degree of synchronous change of hydrological parameters between adjacent measuring points. The Moran index reflects global correlation characteristics, while the Geary-C index highlights local differences. Either index can be used for analysis depending on the selected parameter type and analysis objectives.

[0061] During autocorrelation calculation, the system establishes adjacency relationships between measuring points in a gridded manner and calculates similarity indices based on parameter value differences. By statistically analyzing the similarity levels between measuring points in different regions, the system identifies the spatial distribution pattern of correlation strength. A preliminary spatial structure representation matrix is ​​generated based on the correlation coefficient results to express the degree of spatial coupling of hydrological characteristics between different measuring point regions. To further obtain the continuous parameter distribution across locations within the sea area, the system uses Kriging interpolation to estimate parameters in unmeasured areas based on known parameter values ​​and their spatial coordinates. Kriging interpolation calculates predicted hydrological parameters at target locations based on known measuring point weights and spatial variability functions, thereby achieving spatial reconstruction of the parameter field.

[0062] During interpolation, the system sets the interpolation window size and covariance function type based on a spatial variability model to ensure that the interpolation process fully reflects the correlation structure of surrounding measurement points and reduces local error interference. During interpolation, the system simultaneously calculates spatial gradient changes, performs gradient scanning on the parameter field formed by the interpolation results, and marks regions with gradient changes exceeding a preset range as high-variability areas. To reduce the impact of numerical abrupt changes in these regions on continuity judgment, the system implements local smoothing processing for these regions. This is achieved by introducing edge point weighting and smoothing kernel convolution to adjust the interpolation results, maintaining a balance between spatial continuity and anomaly sensitivity.

[0063] The entire spatial correlation calculation and interpolation reconstruction process is interconnected with the data acquisition module and the point-based optimization module within the system architecture. By analyzing the spatial correlation matrix, the system can identify drastically changing areas in the sea area and assign them higher point density weights in subsequent steps, providing a basis for deployment strategies. The combination of spatial interpolation algorithms and data structures ensures both parameter estimation accuracy and regional representativeness. The Kriging method's adaptive response to the distribution structure of measurement points and parameter variance meets the processing requirements of the rapid, multi-source, and multi-scale characteristics of marine environmental changes. The spatial autocorrelation calculation results can also be directly input into the information density assessment model to support regional clustering and the construction of dynamic allocation strategies for data acquisition resources.

[0064] Taking nearshore waters during typhoon impacts as an example, the system received high-frequency sampling data for several consecutive hours. At the depth level, it found a significant increase in the consistency of salinity change trends at certain measuring points. After calculation using the Moran index, it was determined that the regional correlation was increased, and the interpolation results formed a high-gradient strip structure. The system categorized these areas as key collection regions and, based on this, triggered adjustments to reduce the spacing between measuring points. This process demonstrates the mutual support and synergy between algorithmic and technical features at the functional level. The system comprehensively solves problems in existing technologies, such as insufficient accuracy of spatial models, lag in measuring point placement in variable regions, and the impact of local jumps on interpolation stability, through the aforementioned technical means.

[0065] Step S3: Based on the spatial correlation matrix, the sea area is divided into high-variability regions and low-variability regions. When the correlation coefficient between each measuring point in the sea area and multiple other measuring points in the spatial neighborhood of the measuring point is lower than a preset threshold, the area where the measuring point is located is determined to be a high-variability region; otherwise, it is determined to be a low-variability region.

[0066] Specifically, based on the spatial correlation matrix constructed in step S2, a sea area division operation is performed, dividing the target monitoring sea area into high-variability and low-variability regions to guide the differentiated implementation of subsequent site optimization strategies. The spatial correlation matrix is ​​a structured data result constructed based on the collected monitoring points and their spatial location relationships and the consistency of changes in hydrological characteristic parameters. Each element in the matrix represents the degree of spatial similarity of hydrological parameters between any two monitoring points. By analyzing the distribution of correlation coefficients in this matrix, the system extracts the parameter response characteristics of each monitoring point in the spatial dimension, thereby establishing a correspondence between monitoring points and regional variability.

[0067] To achieve regional division, the system sets a correlation coefficient threshold based on the requirement of hydrological environmental continuity. This threshold can be set according to historical statistical patterns, water type, and monitoring accuracy requirements. When the correlation coefficient between a monitoring point and multiple monitoring points in its spatial neighborhood is lower than the set threshold, the system determines that the hydrological characteristics within the area where the monitoring point is located are spatially discontinuous, i.e., it is classified as a high-variability area. Conversely, if the correlation between monitoring points within the area is generally higher than the threshold and the changing trends are consistent, the system marks the area as a low-variability area. This judgment process is not based on a single monitoring point, but rather forms a local analysis window centered on the monitoring point, combining the statistical characteristics of its spatial neighborhood to form the basis for regional division. The system performs point-by-point sliding window analysis on the entire monitoring area, forming a preliminary regional classification label matrix, and merges spatially adjacent grids with the same label to generate connected high-variability and low-variability area layers.

[0068] In practical applications, such as during nearshore storms, the rapid and spatially localized changes in salinity and current velocity result in multiple regional blocks of low correlation in the spatial correlation matrix collected by the system. In such cases, the aforementioned regional division method can accurately identify these regions as highly variable areas. The system will then increase sampling density in these areas and trigger adjustments to the distribution strategy. To enhance the stability of the boundary divisions and the clarity of the regional structure, the system introduces boundary connectivity rules based on the initial division results to merge fragmented regions. Principal component analysis is then used to extract the dominant hydrological characteristic variables for each region, thereby determining the rationality of the boundary divisions and the dominant parameter attributes.

[0069] The spatial correlation matrix construction algorithm and the region division logic in the above technical solution are functionally closely integrated. The algorithm quantitatively describes spatial continuity through its features, and then uses technical means to complete the region division operation, achieving structure identification and deployment guidance driven by collected data. This process not only serves as an important reference for setting the deployment density but also provides a spatial structural foundation for subsequent information density modeling and equipment spacing adjustment. The linkage between the spatial correlation matrix and the identification of variable regions is central to the entire technical solution; its ability to identify the dynamic characteristics of the monitoring area directly affects the accuracy and adaptability of the system's deployment response.

[0070] This invention integrates quantitative calculation results with deployment decision-making logic through the above-mentioned method, solving technical problems in the prior art such as the lack of spatial differentiation adjustment capability of acquisition strategies, rigid regional deployment structure, and insufficient monitoring density in key variable areas. It has beneficial technical effects such as enhancing deployment adaptability, improving monitoring accuracy in key areas, and supporting optimal resource allocation, which meets the comprehensive requirements of dynamism, efficiency and controllability in marine environmental monitoring.

[0071] Step S4: Calculate the information density index of the high-variability region and the low-variability region and classify them into information-dense region and sparse region.

[0072] In a specific embodiment, in step S4, based on the division results of the high-variability region and the low-variability region, a two-dimensional grid structure covering the entire monitoring sea area is constructed. Each grid cell is used as a calculation unit, and hydrological parameter data in any grid cell is extracted. The statistical characteristics of the hydrological parameter data are used as input, and the information entropy calculation method is used to evaluate the data distribution uncertainty of any grid cell. The data distribution uncertainty is used as the information density index of any grid cell.

[0073] Clustering algorithms are used to cluster all information density indicators. Based on the clustering results, all grid cells are divided into different information density categories to obtain preliminary classification results.

[0074] The information density index difference and category consistency relationship between each grid cell and its multiple directly adjacent grid cells are detected. If the information density index of any grid cell is close to that of any adjacent grid cell, but the categories are inconsistent, then any grid cell is reclassified and merged into the main category of the adjacent grid cells.

[0075] The information density classification is based on the information entropy calculation results of marine grid units. The K-means clustering algorithm is used to divide each grid into different information density categories, and the cluster boundaries are optimized through neighborhood density similarity verification. Specifically, based on the spatial division results of high-variability and low-variability regions, the system constructs a two-dimensional grid structure covering the entire monitoring marine area. Each grid unit is used as a calculation unit to extract hydrological parameter data for calculating the information density index. The information density index is used to quantify the comprehensive complexity and variability level of each region in multidimensional hydrological parameters, reflecting the importance of the region to environmental monitoring results. The system uses the statistical characteristics of hydrological parameter data such as temperature, salinity, current velocity, and water depth collected within each grid as input, and uses the information entropy calculation method to evaluate the distribution uncertainty of data within each grid unit. The higher the information entropy value, the more complex the parameter changes in the region, and the more information it contains. The information entropy value, as a direct indicator of information density, has a direct correspondence with the priority setting of the unit in the site optimization process.

[0076] After calculating the information density, the system organizes the information entropy values ​​of all grid cells into a density feature vector set, which is then input into the clustering analysis module. The K-means clustering algorithm is used to classify the information density values. Based on the principle of minimum variance, the K-means algorithm constructs several class centers in the sample space. The system classifies all grid cells according to a preset number of clusters, minimizing the difference in information density within the same class and maximizing the difference between different classes. The classification results are represented as density level labels. The system automatically determines the relative density level of each class based on the size relationship of the cluster centers, forming a continuous density partitioning system from information-dense areas to information-sparse areas.

[0077] To further improve the spatial continuity and regional connectivity of clustering results, the system performs neighborhood density similarity verification on the preliminary classification results and optimizes the cluster boundaries. The system detects the density value differences and category consistency relationships between each grid cell and its multiple directly adjacent cells. If a cell's density value is found to be close to its neighboring regions but its category label is inconsistent, the system performs a reclassification operation on that cell, merging it into the adjacent main category to avoid unreasonable spatial segmentation caused by isolated points, boundary perturbations, or the initial cluster center distribution. Furthermore, if a cluster category exhibits a spatially broken or disconnected structure, the system, based on the grid topology and connectivity requirements, separately labels the disconnected parts within that category or merges them into the main region, ensuring the spatial interpretability and continuity of the density classification layer.

[0078] In the marine environmental monitoring application scenarios involved in this invention, such as coastal tidal change zones or frontal junctions, there is often a superposition of local hydrological structural abrupt changes and dense disturbances. Parameter distribution in these areas is complex, spatially variable, and exhibits discontinuous information density distribution. Traditional acquisition systems struggle to differentiate density levels based on parameter change trends, resulting in inadequate regional adaptability of equipment deployment strategies. This invention introduces an information entropy model to measure the information complexity within the sea area and combines cluster analysis and neighborhood verification optimization mechanisms to construct a complete information density level classification system. The system can allocate the number of acquisition devices and their spacing based on the differences in information density in different areas, achieving resource optimization based on information value.

[0079] In this technical approach, the information entropy calculation algorithm and clustering classification technology support each other functionally, forming a joint solution for expressing regional information features and structural modeling. The density index calculation and spatial structure identification results directly serve the subsequent equipment deployment parameter generation module. The algorithm features and deployment structure form a stable data interface and control basis, comprehensively solving the problems of lack of density stratification basis for deployment, rigid regional deployment structure, and mismatch between equipment distribution and information value in existing technologies. This improves the monitoring system's adaptability to spatially heterogeneous environments and enhances the pertinence and effectiveness of the overall deployment scheme.

[0080] Step S5: Construct an inverse proportional relationship model between device spacing and information density, and set a smaller device spacing for highly variable and dense areas.

[0081] In one specific embodiment, in step S5, the inverse proportional relationship model between the device spacing and information density satisfies d = k / ρ, where d is the device spacing, ρ is the information density index, and k is a proportional constant set according to the target coverage.

[0082] Specifically, after classifying information density levels, the system constructs an inverse proportional relationship model between equipment spacing and information density based on the information density index corresponding to each region. This model is used to determine the spacing parameters for equipment deployment in different regions, thereby achieving spatially differentiated control of equipment deployment. The information density index is a density value calculated for each grid unit based on changes in multidimensional hydrological parameters. It reflects the data carrying capacity and environmental change complexity of that unit within the overall monitoring network. A higher value indicates a more sensitive region and a stronger demand for high-frequency sampling and intensive monitoring.

[0083] To enable quantitative control of equipment deployment strategies, the system sets the equipment spacing as an inverse function of the density index. The spacing parameter 'd' and the density index 'ρ' are defined as having a functional relationship of 'd = k / ρ', where k is a proportionality constant related to target coverage, regional deployment capacity, and system constraints. This proportionality constant can be preset based on engineering parameters such as the total area of ​​the target sea area, the budgeted number of equipment, and the minimum deployment tolerance, and can be dynamically adjusted in subsequent optimization calculations. This inverse proportional relationship logically establishes a functional mapping between information value and resource allocation, enabling the monitoring system to automatically allocate deployment spacing based on the changing characteristics of each area, achieving an adaptive match between monitoring density and environmental changes.

[0084] The system calculates the target equipment spacing for each grid cell across the entire sea area based on this model, and organizes all calculation results into a spacing distribution matrix. During processing, the system applies the model to both high-variability and low-variability regions, and sets a lower limit for spacing in areas with high variability and high density to avoid deployment conflicts caused by abnormal spacing convergence due to excessively high density. In low-variability or sparse regions, the system automatically increases the equipment spacing based on lower density values, thereby reducing ineffective deployments and saving resources. In some boundary transition areas, the system introduces a smoothing function to perform interpolation processing on the spacing changes between continuous cells, ensuring deployment continuity and system stability.

[0085] In typical scenarios such as typhoon leading-edge fronts or areas of abrupt seabed topography changes, hydrological gradients are large and densely distributed, resulting in significantly higher calculated information density values ​​compared to the background sea area. Traditional sampling strategies typically use average spacing between devices, making it difficult to maintain sampling accuracy in critical areas. This invention employs a spacing control strategy based on an inverse proportional model to significantly reduce the spacing between sampling points in such areas, forming a high-density mesh structure. This enhances the system's ability to monitor local abrupt changes. Simultaneously, the spacing is appropriately increased in low-variability areas far from the impact zone, thereby maximizing the overall resource utilization efficiency of the system.

[0086] The inverse proportional model construction and information density index calculation in the above technical solution are functionally tightly coupled, with the former using the latter as a direct input parameter, and the latter's calculation logic maintaining consistency with the spatial feature modeling mechanism. This technical solution embeds the spatial feature evaluation results into the equipment deployment rules, forming a linkage mechanism integrating data analysis, parameter modeling, and control execution. It solves the problems of existing technologies, such as the lack of quantitative expression in deployment strategies, rigid spacing adjustments, and the inability to dynamically match regional characteristics. It possesses clear engineering practicality and system adaptability, demonstrating significant technical effects in improving monitoring efficiency, reducing deployment costs, and enhancing coverage in key areas.

[0087] In a preferred embodiment, the process of executing step S5 may further include the following steps: identifying the highly variable and information-dense region, comparing the corresponding device spacing with a preset minimum deployment spacing threshold, and if the device spacing is less than the minimum deployment spacing threshold, triggering a parameter optimization mechanism to automatically iterate and update the proportional constant of the inverse proportional relationship model with the current information density index value as input.

[0088] If a region is highly variable and information-dense, and the device spacing calculated according to the inverse proportional relationship model is lower than the preset minimum deployment spacing threshold, a parameter optimization mechanism is triggered to readjust the model constants. Specifically, during the execution of the inverse proportional relationship model between device spacing and information density, the system compares the device spacing value calculated for each highly variable and information-dense region with the preset minimum deployment spacing threshold. If the calculated spacing for a region is less than the threshold, the system automatically triggers the parameter optimization mechanism to adjust the proportional constants in the model, thereby avoiding deployment conflicts, coverage redundancy, or hardware deployment infeasibility caused by excessively small device deployment spacing. The inverse proportional relationship model is constructed based on the inverse relationship between device spacing d and information density ρ. In high-density or highly variable regions, the value of ρ increases significantly, resulting in a significant decrease in the output value of d. To ensure the operability of actual deployment matches physical limitations, the system sets a reasonable minimum spacing threshold. This threshold is comprehensively set based on parameters such as device sensing range, communication interference distance, and physical fixed distance, serving as an acceptable lower limit for spacing.

[0089] The system executes this threshold comparison logic synchronously during the point layout parameter generation phase, checking each spacing value calculated based on d = k / ρ. If a spacing is less than the threshold, the system does not use the result directly. Instead, it uses the current density value as input to work backwards to adjust the value of the proportional constant k, regenerates the spacing value, and re-verifies whether the updated spacing meets the constraints. The parameter optimization mechanism automatically iterates and updates k through a recursive adjustment method. While maintaining the relative differences in point layout density, it raises all spacing values ​​to a range not less than the threshold, while controlling the spacing gradient change between high and low density areas to be smooth, avoiding abrupt changes in layout structure or density discontinuities due to local corrections.

[0090] In this process, the algorithm module and the deployment logic are linked. Information density serves as the model input parameter, driving the spacing calculation. The comparison between the spacing and the threshold triggers model parameter updates. The updated model applies uniformly to all areas, forming a unified deployment rule system. This mechanism is suitable for areas in the sea with local abrupt changes, extremely high boundary density, or drastic fluctuations in historical sampling data. It can automatically identify problems with excessively small spacing caused by a single scaling factor in the model and prevent the global model structure from collapsing through local corrections, thereby improving the stability and engineering controllability of the deployment strategy.

[0091] For example, in the tidal shear zone of river estuaries, the hydrological information density reaches extremely high peaks due to the coupling effect of complex topography and current velocity. Directly applying the standard inverse proportional model in this area will produce spacing values ​​far smaller than the equipment deployment capability range. Without parameter correction, this will lead to excessive concentration of equipment in the area, mutual interference, or difficulty in deployment. This invention introduces a linkage scheme between a minimum deployment spacing threshold and a parameter optimization mechanism, enabling the system to automatically improve the spacing calculation results after identifying such areas. This ensures monitoring accuracy in densely populated areas while avoiding unnecessary deployment redundancy, achieving a balance between model sensitivity and deployment feasibility.

[0092] Step S6: Use optimization algorithms to optimize the number of devices and their layout coordinates while meeting the monitoring coverage requirements, and generate a precise deployment plan.

[0093] Specifically, based on the previously completed information density level classification results and device spacing configuration model, the system uses an optimization algorithm to jointly solve for the number of data acquisition devices and their corresponding spatial coordinates, generating a deployment scheme that meets coverage constraints. The optimization process aims to improve resource utilization efficiency, ensure monitoring integrity, and enhance deployment feasibility. It inputs the sea area division results, device spacing parameters, information density weights, and boundary constraints into the optimization algorithm, searching for the coordinate set corresponding to the minimum number of devices within the deployment space. This ensures that the monitoring coverage rate in each area meets preset requirements, and that the spatial distribution of devices conforms to the spacing constraints derived from the density model.

[0094] The system introduces a feasible solution space into the deployment model, limiting constraints such as minimum spacing between devices, boundary buffers, and signal interference distances to ensure the physical feasibility of the optimization results. For algorithm selection, the system employs a genetic algorithm as a global search tool. Multiple deployment combinations are generated by initializing the population, and a fitness evaluation is performed on each combination. The fitness function uses the total number of devices, coverage area integrity, spacing rationality, and connectivity as evaluation dimensions, prioritizing deployment combinations with the fewest devices and optimal coverage within the constraints. The system iteratively evolves the deployment structure through crossover and mutation operations, dynamically updating local optima in each iteration until a global optimum or convergence condition is met, outputting the final deployment scheme.

[0095] During the optimization process, the system incorporates the information-dense and sparse distribution layers generated in the previous stage into the population initialization strategy. This increases the candidate point density in information-dense areas and reduces the point search frequency in sparse areas to improve population quality and optimization efficiency, while ensuring that the optimization results maintain the coverage priority for dense areas. In each round of point evaluation, the system maps each candidate scheme to a spatial grid model and calculates its fit to the information density layer. If a point scheme fails to cover all density level areas, its fitness value is reduced. The system thus suppresses the retention rate of unbalanced layout structures, gradually converging to the optimal layout structure with strong spatial continuity and high local density matching.

[0096] In typical application scenarios, such as complex intertidal zones or within the deployment zone of marine environmental observation platforms, irregular spatial structures, rapidly changing hydrodynamic characteristics, and limited measurement point deployment costs can lead to equipment redundancy or gaps in area coverage if fixed-rule deployment is adopted. This invention solves for deployment coordinates through an optimization algorithm, dynamically adjusting the number of devices to meet the deployment needs of areas with varying intensity of variation, while ensuring that monitoring within the area does not overlap, miss, or overload, significantly reducing system deployment costs and maintenance intensity.

[0097] In the above technical solution, the optimization algorithm structure and the information density modeling results are functionally directly related. The former relies on the latter's spatial attribute input to drive the device deployment priority and structural control, while the latter completes the deployment feasibility verification through optimization algorithm feedback. The two constitute a closed-loop deployment strategy scheduling system, solving the problems of insufficient response, low deployment efficiency, and high local redundancy in existing fixed-point methods in heterogeneous environmental regions. This technical solution is suitable for large-scale, variable, and highly structure-sensitive marine environmental monitoring scenarios, improving the system's global coverage performance and deployment cost control capabilities.

[0098] In a preferred embodiment, step S6 involves inputting the sea area division results, equipment spacing parameters, information density weights, and boundary constraints into the optimization algorithm to search for the coordinate set corresponding to the minimum number of equipment within the deployment space. Specifically, this includes:

[0099] (1) Genetic algorithm is used as a global search tool. Multiple deployment combination schemes are generated by initializing the population. The fitness of each scheme is evaluated by the total number of devices, the integrity of the coverage area, the rationality of the spacing and the connectivity. The deployment combination with the fewest devices and the best coverage is selected.

[0100] (2) Based on the layout combination, the layout structure is iteratively evolved through crossover and mutation operations. The local optimal solution is dynamically updated in each iteration until the global optimal solution or the convergence condition is met. The final output solution is taken as the layout scheme.

[0101] The optimization algorithm is a genetic algorithm, including population initialization, fitness evaluation, crossover and mutation operations, and iterative calculation to generate a minimum equipment deployment scheme that meets the coverage requirements. Specifically, based on the genetic evolution mechanism, the algorithm achieves global search and optimization of the solution space through iterative evolution of multiple deployment schemes, thereby solving problems that traditional deployment strategies struggle to address, such as uneven equipment deployment density and complex spatial constraints in high- and low-variability areas. The algorithm's input parameters include the spatial boundary of the target monitoring sea area, the hydrological information density layer, the regional classification results, the minimum equipment spacing constraint, the sensing coverage radius, and the total deployment limit. The output is a set of equipment deployment coordinates and their spatial feasibility markers.

[0102] During the population initialization phase, the system randomly generates several placement schemes in space based on the density distribution of high-variability regions. Each placement scheme represents an individual, containing a set of coordinate points. The population initialization strategy considers the spatial gradient differences in information density distribution, placing points in high-density regions with high probability and in sparse regions with low probability. The feasibility of the initial solution is improved through density weighting, providing an effective starting point for iterative optimization.

[0103] During the fitness evaluation process, the system calculates constraints and monitoring coverage for each deployment point, and uses a fitness function to comprehensively score it across multiple dimensions, including spatial coverage within the target area, total number of devices used, and the rationality of adjacent point spacing. Schemes with insufficient coverage or exceeding the upper limit for the number of devices are assigned lower fitness values. Schemes with uniform deployment structure, meeting spacing requirements, and having the smallest number of devices are prioritized and retained for subsequent crossover and mutation phases.

[0104] Crossover is used to generate new individuals by interchanging parent placement schemes, improving population diversity through partial coordinate set swapping. Mutation is used to randomly replace some placement point coordinates in offspring, introducing perturbations to the solution space to escape local optima. After each round of evolution, the system retains the schemes with the highest fitness to form a new generation of population, repeating the above operations until the fitness function converges or the maximum number of iterations is reached, outputting the placement result with the optimal fitness.

[0105] In practical applications, such as long-term monitoring of hydrological parameters in complex nearshore seabed topography, traditional evenly spaced deployments cannot balance monitoring accuracy and equipment cost due to seabed structure limitations and multi-parameter coupling disturbances. This invention utilizes a genetic algorithm to perform multi-constraint optimization, dynamically finding optimal location combinations within a continuous space. This achieves high-density information coverage in key areas and equipment compression in non-key areas under resource-constrained conditions, effectively alleviating redundant deployment problems.

[0106] In this technical solution, the genetic algorithm's structural design and the marine environmental information density model form a functional closed-loop logic. The algorithm input relies on the density classification results to generate the initial deployment distribution, the optimization objective reflects the relationship between information variation characteristics and equipment deployment, and the evaluation index system is directly supported by weights provided by the front-end model. The two functions support each other and form a feedback loop. This deployment optimization mechanism integrates the algorithm's structural design with the constraints of actual physical deployment capabilities, forming an automatically adjustable deployment parameter control system capable of solving existing problems such as poor regional adaptability, low deployment efficiency, and high redundancy.

[0107] Step S7: Obtain real-time marine environmental monitoring data. If significant changes in hydrological characteristics are detected, recalculate the equipment spacing and update the deployment plan.

[0108] In one specific embodiment, in step S7, any hydrological parameter is extracted, its time series data is obtained, the parameter change trend in the current period and multiple historical change trends are extracted through a sliding window mechanism, and a change threshold is generated based on the average value and standard deviation of the multiple historical change trends. If the parameter change trend exceeds the change threshold, it is determined that the hydrological characteristics have changed significantly.

[0109] If the monitoring data of any region shows significant changes in hydrological characteristics over multiple consecutive periods, the region is marked as a high-dynamic state.

[0110] Priority parameters are assigned to each region based on its environmental sensitivity, hydrological change frequency, and monitoring target weight. For high-dynamic regions where the priority parameters are greater than preset values, a release point update process is initiated.

[0111] In real-time environmental monitoring, trend analysis is performed on changes in hydrological characteristics. The criteria for determining significant changes are whether the magnitude of parameter changes exceeds a dynamic threshold set based on historical change patterns. Furthermore, the system can trigger deployment update processes according to regional risk levels and priorities. Specifically, after deployment, the system continuously acquires real-time monitoring data of the marine environment, collecting key hydrological parameters such as temperature, salinity, current velocity, and water depth. It then identifies trend changes in the monitoring data based on a time-series analysis mechanism. During processing, the system continuously updates the hydrological characteristic database according to a preset monitoring cycle and sampling frequency. It also extracts the current parameter change trend through a sliding window mechanism and compares it with the change patterns within historical cycles to determine whether significant environmental changes have occurred. The criterion for significant changes is that the magnitude of change of one or more hydrological parameters within a specified time period exceeds a change threshold dynamically generated by the system based on historical mean and standard deviation. This threshold is dynamically set in conjunction with parameter type and the heterogeneity of the marine environment, exhibiting both time-varying and regional adaptability.

[0112] The system categorizes trend change results. If monitoring data in a certain area shows fluctuations outside the threshold for multiple consecutive periods, the system automatically marks the area as high-dynamic and initiates a spacing update process. In this process, the system calls a previously constructed inverse-proportional model of device spacing and information density to recalculate the target spacing for the currently high-variability area and, in conjunction with the trigger point optimization module, generates an adjusted deployment suggestion. To avoid resource waste caused by frequent recalculation, the system introduces a risk level mechanism. Priority parameters are assigned based on the environmental sensitivity, hydrological change frequency, and monitoring target weight of each area. When a trend change signal is detected, only the trigger point update process is initiated for high-priority areas; changes are recorded for other areas but no structural update operation is performed.

[0113] For example, in the storm surge front region, flow velocity and salinity parameters change drastically and exhibit rapid spatial migration characteristics. By continuously analyzing the flow velocity gradient changes in real-time monitoring data, the system identifies areas with persistent abnormal changes that exceed dynamic threshold standards, thus determining them as significant environmental disturbance areas. It automatically recalculates the spacing and adjusts the equipment deployment strategy to adapt the deployment results to the changed information distribution structure, thereby improving the system's responsiveness to sudden environmental changes and the robustness of the deployment strategy.

[0114] The parameter trend analysis algorithm and dynamic deployment adjustment mechanism in this technical solution constitute a data-driven control closed loop. The algorithm relies on monitoring data to model parameter fluctuations, while the control module uses the analysis results as trigger signals to dynamically adjust the deployment structure. Functionally, they support each other and form a linked feedback loop, solving the problems of static and unadjustable deployment points and lack of dynamic response to the environment in traditional deployment models. This mechanism is particularly suitable for monitoring tasks in key areas with sudden events, high hydrodynamic variability, or disaster risks, enhancing the adaptability of the deployment scheme and the stability of the system.

[0115] Step S8: Send updated working parameters and location instructions to the acquisition device through the wireless communication system to complete the dynamic optimization deployment.

[0116] In one specific embodiment, in step S8, the wireless communication system includes a main communication channel and a backup communication channel. If the main communication channel fails to transmit parameters, it automatically switches to the backup communication channel for retransmission until the device confirms successful reception.

[0117] Specifically, after dynamically updating the deployment plan, the system uses a wireless communication system to send control information containing the latest operating parameters and location commands to each data acquisition device distributed throughout the sea area, enabling real-time adjustment of the deployment structure and remote command response. The wireless communication system comprises a two-tier architecture: a main communication channel and a backup communication channel. The main channel prioritizes normal communication tasks, while the backup channel serves as a failover path to ensure the continuity and reliability of remote command transmission. Based on real-time deployment optimization results, the control center automatically generates new configuration commands for each data acquisition device. These commands include target operating coordinates, water depth range, sampling frequency, interval duration, and power consumption level, and this information is encoded and sent via the main communication channel to the corresponding device's unique identifier address.

[0118] During communication, the system monitors receipt information in real time and verifies the device's response status within a specified time window. If the device fails to return a successful reception confirmation signal within the response window, the system determines that the main channel communication has failed and automatically activates the backup communication channel to retransmit the command. The backup channel can use a low-power wide-area network, satellite link, or shortwave wireless channel, automatically matching the optimal transmission path based on the current network environment and device receiving capabilities. Multi-frame reconstruction technology enhances reception robustness in environments with poor channel quality. The system continues to monitor the device's receiving status during the retransmission period until a valid confirmation is obtained or the communication link is determined to have completely failed. In the event of communication failure, an anomaly is recorded and sent to the dispatch platform to prompt manual intervention.

[0119] The above describes an intelligent acquisition method for multi-dimensional marine information in the embodiments of this application. The following describes an intelligent acquisition system for multi-dimensional marine information in the embodiments of this application. Please refer to [link / reference]. Figure 2 An intelligent system for acquiring multi-dimensional marine information, as described in this application embodiment, includes:

[0120] The hydrological sensor network module 10 is used to collect hydrological parameters at multiple depth layers and perform anomaly detection and data updates.

[0121] The data processing and analysis module 20 is used to construct hydrological feature datasets, generate spatial correlation matrices and information density classification maps.

[0122] The placement optimization module 30 is used to output the optimal equipment placement scheme based on the inverse proportional relationship model and optimization algorithm.

[0123] The real-time monitoring and adaptive update module 40 is used to analyze hydrological change trends and dynamically trigger the update of the release point.

[0124] The communication management module 50 is used to send updated parameters and location information to the acquisition device and complete the receipt confirmation.

[0125] This application also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the intelligent acquisition method for multi-dimensional ocean information.

[0126] The above description is only a preferred embodiment of the present invention. It should be noted that those skilled in the art can make several improvements and additions without departing from the principle of the present invention, and these improvements and additions should also be considered within the scope of protection of the present invention.

Claims

1. A method for intelligent acquisition of multi-dimensional marine information, characterized in that, Includes the following steps: S1. Simultaneously collect environmental parameters at multiple measuring points and multiple depth layers in the sea area. The environmental parameters include temperature, salinity, current velocity and water depth, forming a hydrological feature dataset with time series and spatial coordinates. S2. Based on the hydrological feature dataset, calculate the spatial autocorrelation coefficient between each measuring point, and generate a spatial correlation matrix using a spatial interpolation algorithm; S3. Based on the spatial correlation matrix, the sea area is divided into high-variability regions and low-variability regions. When the correlation coefficient between each measuring point in the sea area and multiple other measuring points in the spatial neighborhood of the measuring point is lower than a preset threshold, the area where the measuring point is located is determined to be a high-variability region; otherwise, it is determined to be a low-variability region. S4. Calculate the information density index of the high-variability region and the low-variability region, and classify them into information-dense and sparse regions. S5. Construct an inverse proportional relationship model between device spacing and information density, and set a smaller device spacing for areas with high variability and dense information. S6. Use optimization algorithms to optimize the number of devices and their layout coordinates while meeting the monitoring coverage requirements, and generate a layout plan. S7. Acquire real-time marine environmental monitoring data. If significant changes in hydrological characteristics are detected, recalculate the device spacing and update the deployment plan. S8. Send updated working parameters and location instructions to the acquisition device through the wireless communication system to complete the dynamic optimization deployment.

2. The method as described in claim 1, characterized in that, S1 includes a data anomaly detection mechanism. When a parameter of a certain measuring point continuously exceeds a preset threshold in the time or depth dimension, high-frequency sampling of that measuring point is triggered, and the sampling results are used to update the hydrological feature dataset. In S2, the spatial correlation calculation uses the Moran index or Geary-C index as an evaluation method to measure the degree of synchronous change of hydrological parameters between adjacent measuring points. The spatial interpolation algorithm is the Kriging interpolation method. It performs local smoothing in areas with large parameter gradient changes. Based on the known parameter values ​​of the measuring points and their spatial coordinates, the Kriging interpolation method is used to estimate the unmeasured areas. Based on the known measuring point weights and spatial variability function, the Kriging interpolation method calculates the predicted hydrological parameters at the target location, thereby realizing the spatial reconstruction of the parameter field.

3. The method as described in claim 1, characterized in that, In step S4, based on the division results of the high-variability region and the low-variability region, a two-dimensional grid structure covering the entire monitoring sea area is constructed. Each grid cell is used as a calculation unit. Hydrological parameter data in any grid cell is extracted. The statistical characteristics of the hydrological parameter data are used as input. The information entropy calculation method is used to evaluate the data distribution uncertainty of any grid cell. The data distribution uncertainty is used as the information density index of any grid cell. Clustering algorithms are used to cluster all information density indicators. Based on the clustering results, all grid cells are divided into different information density categories to obtain preliminary classification results. The information density index difference and category consistency relationship between each grid cell and its multiple directly adjacent grid cells are detected. If the information density index of any grid cell is close to that of any adjacent grid cell, but the categories are inconsistent, then any grid cell is reclassified and merged into the main category of the adjacent grid cells.

4. The method as described in claim 1, characterized in that, In step S5, the inverse proportional relationship model between the device spacing and information density satisfies d = k / ρ, where d is the device spacing, ρ is the information density index, and k is a proportional constant set according to the target coverage.

5. The method as described in claim 4, characterized in that, Identify the highly variable and information-dense region, compare the corresponding device spacing with a preset minimum deployment spacing threshold, and if the device spacing is less than the minimum deployment spacing threshold, trigger a parameter optimization mechanism to automatically iterate and update the proportional constant of the inverse proportional relationship model using the current information density index value as input.

6. The method as described in claim 1, characterized in that, In step S6, the sea area division results, equipment spacing parameters, information density weights, and boundary constraints are input into the optimization algorithm to search for the coordinate set corresponding to the minimum number of equipment within the deployment space. Specifically, this includes: A genetic algorithm is used as a global search tool. Multiple deployment combinations are generated by initializing the population. The fitness of each combination is evaluated based on the total number of devices, the integrity of the coverage area, the rationality of the spacing, and the connectivity. The deployment combination with the fewest devices and the best coverage is selected. Based on the aforementioned layout combination, the layout structure is iteratively evolved through crossover and mutation operations. In each iteration, the local optimal solution is dynamically updated until the global optimal solution or the convergence condition is met. The final output solution is then used as the layout scheme.

7. The method as described in claim 6, characterized in that, In step S7, any hydrological parameter is extracted, its time series data is obtained, and the parameter change trend and multiple historical change trends are extracted in the current period through a sliding window mechanism. A change threshold is generated based on the average value and standard deviation of the multiple historical change trends. If the parameter change trend exceeds the change threshold, it is determined that the hydrological characteristics have changed significantly. If the monitoring data of any region shows significant changes in hydrological characteristics over multiple consecutive periods, the region is marked as a high-dynamic state. Priority parameters are assigned to each region based on its environmental sensitivity, hydrological change frequency, and monitoring target weight. For high-dynamic regions where the priority parameters are greater than preset values, a release point update process is initiated.

8. The method as described in claim 1, characterized in that, In step S8, the wireless communication system includes a main communication channel and a backup communication channel. If the main communication channel fails to transmit parameters, it automatically switches to the backup communication channel for retransmission until the device confirms successful reception.

9. An intelligent acquisition system for implementing the intelligent acquisition method for multi-dimensional marine information as described in any one of claims 1 to 8, characterized in that, include: The hydrological sensor network module is used to collect hydrological parameters at multiple depth levels and perform anomaly detection and data updates. The data processing and analysis module is used to construct hydrological feature datasets, generate spatial correlation matrices and information density classification maps; The deployment optimization module is used to output the optimal equipment deployment scheme based on the inverse proportional relationship model and optimization algorithm. The real-time monitoring and adaptive update module is used to analyze hydrological change trends and dynamically trigger the update of the release points; The communication management module is used to send updated parameters and location information to the data acquisition device and complete the receipt confirmation.

10. A computer-readable storage medium having a computer program stored thereon, the computer program being executed by a processor to perform the steps of the intelligent acquisition method for multidimensional marine information as described in any one of claims 1 to 8.

Images