A multi-source data fusion marine fishery environment early warning method and system

Abstract

The application discloses a kind of multi-source data fusion's marine fishery environment early warning method and system, belong to marine fishery technical field, this method includes: based on big data acquisition technology, by satellite remote sensing real-time acquisition target monitoring sea area's marine remote sensing data, by buoy observation platform real-time collection target monitoring sea area's each sea area calibration point measured water temperature data;The data collected are grid division pre-processing, spatial dimension and time dimension division standardization processing such as big data processing, obtain the time dimension consistent and spatial dimension consistent multi-source marine environment data set;Through environmental data analysis model, the multi-source marine environment data set is processed, and the environmental abnormal change data of target monitoring sea area is obtained, and the seawater breeding strategy is warned and adjusted. Through the implementation of the application, the problem of not being able to synchronize time dimension data and spatial dimension data in the complex marine environment can be solved, and the comprehensiveness of the guidance to the seawater breeding industry is improved.

Description

Technical Field

[0001] This application belongs to the field of marine fisheries technology, specifically a marine fisheries environmental early warning method and system based on multi-source data fusion. Background Technology

[0002] Marine fisheries environmental early warning research is an important field to ensure the sustainable use of fishery resources and marine ecological security. Its core lies in providing timely and effective guidance for mariculture by monitoring and predicting changes in the marine environment.

[0003] Existing marine fisheries early warning methods primarily rely on monitoring single-dimensional or single-dimensional information to provide effective guidance for mariculture production. Taking single-dimensional information as an example, current technologies typically monitor water temperature changes in a target area and compare them to suitable temperature thresholds for cultured organisms to determine if the conditions are suitable for the growth of shellfish and other cultured organisms. If unsuitable, adjustments are made to current aquaculture strategies to provide effective guidance for mariculture. However, this method of monitoring and issuing warnings from a single time or space dimension has significant limitations when dealing with complex marine environmental data, exhibiting considerable bias. The guidance derived from this approach is also highly biased. Therefore, there is an urgent need for a marine fisheries environmental early warning method that integrates both time and space dimensions to improve the comprehensiveness of guidance for mariculture in complex marine environments. Summary of the Invention

[0004] To address the above issues, this application provides a marine fisheries environmental early warning method and system based on multi-source data fusion. This method solves the problem that mariculture in target monitoring areas cannot synchronize temporal and spatial data under complex marine environments, thereby improving the comprehensiveness of guidance for mariculture in complex marine environments.

[0005] To achieve the above objectives, the technical solution adopted in this application is a marine fisheries environment early warning method based on multi-source data fusion, comprising: The ocean remote sensing data of the target monitoring sea area is acquired in real time by satellite remote sensing, and the measured water temperature data of each calibration point in the target monitoring sea area is collected in real time by buoy observation platform; wherein, the ocean remote sensing data includes: sea surface water temperature remote sensing data and phytoplankton concentration remote sensing data; The marine remote sensing data and the measured water temperature data of each sea area calibration point are preprocessed by grid division to ensure that the marine remote sensing data and the measured water temperature data of each sea area calibration point are within the same grid division accuracy, thereby generating the first multi-source marine environment data set. The first multi-source marine environment data set is subjected to spatial dimension partitioning and time dimension partitioning and standardization processing to obtain a second multi-source marine environment data set that is consistent in both time and spatial dimensions. The second multi-source marine environmental data set is input into the environmental data analysis model so that the environmental data analysis model outputs environmental anomaly change data of the target monitoring sea area based on the second multi-source marine environmental data set; wherein, the environmental anomaly change data includes: water temperature anomaly change data and phytoplankton concentration anomaly change data, sea areas with anomaly change in water temperature and sea areas with anomaly change in phytoplankton concentration; Based on the data on abnormal environmental changes, early warnings and adjustments are made to the marine aquaculture strategies in the target monitoring sea area.

[0006] Further, the preprocessing of the marine remote sensing data and the measured water temperature data from each sea area calibration point involves grid division to ensure that the marine remote sensing data and the measured water temperature data from each sea area calibration point are within the same grid division accuracy, thereby generating a first multi-source marine environmental data set, including: Acquire the latitude and longitude data of the target monitoring sea area, and determine the grid boundary of the target monitoring sea area based on the latitude and longitude data; A grid to be filled is generated based on a preset grid division resolution and the grid boundary of the target monitoring sea area; wherein, the grid to be filled contains several sub-grids to be filled, and each sub-grid to be filled has the same area; The first multi-source marine environmental data set is generated based on the grid to be filled, the marine remote sensing data, and the measured water temperature data of each sea area calibration point.

[0007] Further, the step of performing spatial dimension partitioning and standardization processing on the first multi-source marine environmental data set to obtain a second multi-source marine environmental data set consistent in both time and spatial dimensions includes: The first multi-source marine environment data set is subjected to spatial dimension division and standardization processing to obtain a third multi-source marine environment data set with consistent spatial dimensions. The third multi-source marine environmental data set is subjected to time dimension division and standardization processing to obtain a second multi-source marine environmental data set that is consistent in both time and spatial dimensions.

[0008] Furthermore, the step of performing spatial dimension partitioning and standardization processing on the first multi-source marine environmental data set to obtain a third multi-source marine environmental data set with consistent spatial dimensions includes: Based on each sub-grid to be filled in the first multi-source marine environmental data set, the marine remote sensing data and the measured water temperature data of each sea area calibration point are divided into spatial dimensions to obtain the marine remote sensing data and the measured water temperature data of each sea area calibration point of each sub-grid to be filled. Based on the marine remote sensing data of each subgrid to be filled and the measured water temperature data of the sea area calibration points of each subgrid to be filled, data filling is performed on each subgrid to be filled, generating a third multi-source marine environmental data set containing several filled data subgrids and several unfilled data subgrids.

[0009] Furthermore, the third multi-source marine environmental data set undergoes time-dimension standardization processing to obtain a second multi-source marine environmental data set consistent in both time and spatial dimensions, including: Obtain the spatial distance and temporal identifier between each unfilled data subgrid and each filled data subgrid in the third multi-source marine environmental dataset; Based on the time stamp, the marine remote sensing data and the measured water temperature data of the sea area calibration point of each filled data subgrid are standardized in terms of time dimension to generate each filled data subgrid with consistent time dimension; For each unfilled data subgrid, determine the spatial distance weights of each adjacent subgrid of the current unfilled data subgrid and each filled data subgrid based on the spatial distance between the unfilled data subgrid and each filled data subgrid; Obtain the marine remote sensing data and measured sea temperature data of the sea area calibration point corresponding to each adjacent subgrid of the current unfilled data subgrid. Based on the marine remote sensing data, measured sea temperature data of the sea area calibration point corresponding to the adjacent subgrid of the current unfilled data subgrid and the spatial distance weight of each adjacent subgrid, generate the marine remote sensing data and measured sea temperature data of the sea area calibration point corresponding to the current unfilled data subgrid. Then, perform data interpolation on the current unfilled data subgrid to generate the interpolated data subgrid. A third multi-source marine environmental data set with consistent temporal and spatial dimensions is generated based on each interpolated data subgrid and each filled data subgrid.

[0010] Further, the step of generating marine remote sensing data and measured sea temperature data corresponding to the current unfilled data subgrid based on the marine remote sensing data, measured sea temperature data at sea calibration points, and spatial distance weights of each adjacent subgrid includes: Based on the spatial distance weights between the currently unfilled data subgrid and each adjacent subgrid, determine the weights of the sea surface temperature remote sensing data, the phytoplankton concentration remote sensing data, and the measured sea surface temperature data at the sea area calibration point for each adjacent subgrid. The sea surface temperature remote sensing data of the current unfilled subgrid is obtained by weighting the weights of the sea surface temperature remote sensing data of each adjacent subgrid. The phytoplankton concentration remote sensing data of each adjacent subgrid is weighted and averaged to obtain the phytoplankton concentration remote sensing data of the current unfilled subgrid. The measured water temperature data of the sea area calibration points of each adjacent subgrid are weighted and averaged to obtain the measured water temperature data of the sea area calibration points of the current unfilled data subgrid.

[0011] Furthermore, after obtaining the second multi-source marine environmental dataset, it also includes: For each filled data subgrid, the sea surface temperature remote sensing data, phytoplankton concentration remote sensing data and measured sea area calibration point water temperature data of the current filled data subgrid are weighted and fused according to the preset weighting weight to obtain the comprehensive environmental index of the current filled data subgrid. For each filled data subgrid, spatial error assessment is performed to generate the numerical error and positional error of the current filled data subgrid. Based on the current sea surface water temperature remote sensing data and the measured water temperature data of the sea area calibration point, the comprehensive water temperature data of the current filled data subgrid is determined. Pearson correlation coefficient analysis is used to determine the correlation between the comprehensive water temperature data and the phytoplankton concentration remote sensing data, and the correlation coefficient of the current filled data subgrid is generated. For each interpolated data subgrid, the sea surface temperature remote sensing data, phytoplankton concentration remote sensing data, and measured sea area calibration point water temperature data of the current interpolated data subgrid are weighted and fused according to the preset weighting weights to obtain the comprehensive environmental index of the current interpolated data subgrid. For each interpolated data subgrid, spatial error assessment is performed to generate the numerical error and positional error of the current interpolated data subgrid. Based on the current sea surface temperature remote sensing data and the measured water temperature data of the sea area calibration point, the comprehensive water temperature data of the current interpolated data subgrid is determined. Pearson correlation coefficient analysis is used to determine the correlation between the comprehensive water temperature data and the phytoplankton concentration remote sensing data, and the correlation coefficient of the current interpolated data subgrid is generated. Based on the grid identifiers of each filled data subgrid, remote sensing data of sea surface temperature, remote sensing data of phytoplankton concentration, measured water temperature data of sea area calibration points, comprehensive environmental indicators, numerical errors, location errors, and correlation coefficients, and the grid identifiers of each interpolated data subgrid, remote sensing data of sea surface temperature, remote sensing data of phytoplankton concentration, measured water temperature data of sea area calibration points, comprehensive environmental indicators, numerical errors, location errors, and correlation coefficients, a multi-source marine environmental data matrix corresponding to the second multi-source marine environmental data set is generated. The step of inputting the second multi-source marine environmental data set into the environmental data analysis model, so that the environmental data analysis model outputs dynamic environmental anomaly change data of the target monitoring sea area based on the second multi-source marine environmental data set, includes: The multi-source marine environmental data matrix corresponding to the second multi-source marine environmental data set is input into the environmental data analysis model so that the environmental data analysis model outputs environmental anomaly change data of the target monitoring sea area based on the multi-source marine environmental data matrix corresponding to the second multi-source marine environmental data set.

[0012] Further, the step of inputting the multi-source marine environmental data matrix corresponding to the second multi-source marine environmental data set into the environmental data analysis model, so that the environmental data analysis model outputs environmental anomaly change data of the target monitoring sea area based on the multi-source marine environmental data matrix corresponding to the second multi-source marine environmental data set, includes: The multi-source marine environmental data matrix is ​​input into the environmental data analysis model so that the environmental data analysis model can determine the initial abnormal water temperature change data and the initial abnormal water temperature change area based on the sea surface temperature remote sensing data and the measured water temperature data of the sea area calibration point in the multi-source marine environmental data matrix, and determine the initial abnormal phytoplankton concentration change data and the initial abnormal phytoplankton concentration change area based on the phytoplankton concentration remote sensing data in the multi-source marine environmental data matrix. The initial water temperature anomaly data, initial water temperature anomaly data, initial phytoplankton concentration anomaly data, and initial phytoplankton concentration anomaly data are corrected using the comprehensive environmental indicators, numerical errors, location errors, and correlation coefficients in the multi-source marine environmental data matrix. The output data on water temperature anomaly data, phytoplankton concentration anomaly data, and the target monitoring sea area are then output.

[0013] Furthermore, the step of providing early warning and adjusting the marine aquaculture strategy for the target monitoring sea area based on the abnormal environmental change data includes: Based on the data of abnormal water temperature changes, the locations of abnormal water temperature changes are determined. Based on the locations of abnormal water temperature changes and the sea areas of abnormal water temperature changes, the sub-regions of abnormal water temperature changes corresponding to each location are determined. The sub-regions of abnormal water temperature changes are marked as areas of risk of sudden water temperature changes. The remaining areas of the sea areas of abnormal water temperature changes, except for the areas of risk of sudden water temperature changes, are marked as areas of potential risk of sudden water temperature changes. Based on the data on abnormal changes in phytoplankton concentration, the locations of abnormal changes in phytoplankton concentration are determined. Based on the locations of abnormal changes in phytoplankton concentration and the sea areas where abnormal changes in phytoplankton concentration occur, sub-regions of abnormal changes in phytoplankton concentration are determined for each location of abnormal changes in phytoplankton concentration. The sub-regions of abnormal changes in phytoplankton concentration are marked as areas of risk of sudden changes in phytoplankton concentration. The remaining areas of the sea areas where abnormal changes in phytoplankton concentration occur, except for the areas of risk of sudden changes in phytoplankton concentration, are marked as areas of potential risk of sudden changes in phytoplankton concentration. The overlapping areas are determined based on the risk areas of sudden changes in water temperature and the risk areas of sudden changes in phytoplankton concentration, and the overlapping areas are marked as high-risk areas. The remaining target monitoring areas, except for the sea areas with abnormal changes in water temperature and the sea areas with abnormal changes in phytoplankton concentration, are marked as risk-free areas. Based on high-risk areas, areas at risk of sudden changes in water temperature, areas at risk of sudden changes in phytoplankton concentration, areas at risk of potential sudden changes in water temperature, and areas at risk of potential sudden changes in phytoplankton concentration, tiered early warning information is generated. The scale of mariculture in high-risk areas, areas at risk of sudden changes in water temperature, and areas at risk of sudden changes in phytoplankton concentration is reduced. The mariculture yield in areas at risk of potential sudden changes in water temperature and areas at risk of potential sudden changes in phytoplankton concentration is monitored in real time, and the scale of mariculture in risk-free areas is increased.

[0014] Based on the above method embodiments, this application provides corresponding system embodiments; One embodiment of this application provides a marine fisheries environment early warning system based on multi-source data fusion. The system includes: a multi-source data acquisition module, a spatiotemporal data synchronization module, an anomaly data analysis module, and an early warning module. The multi-source data acquisition module is used to acquire marine remote sensing data of the target monitoring sea area in real time through satellite remote sensing, and to acquire measured water temperature data of each calibration point in the target monitoring sea area in real time through the buoy observation platform; wherein, the marine remote sensing data includes: sea surface water temperature remote sensing data and phytoplankton concentration remote sensing data; The spatiotemporal data synchronization module is used to perform grid division preprocessing on the marine remote sensing data and the measured water temperature data of each sea area calibration point, so that the marine remote sensing data and the measured water temperature data of each sea area calibration point are in the same grid division accuracy, generating a first multi-source marine environment data set; and to perform spatial dimension division standardization processing and temporal dimension division standardization processing on the first multi-source marine environment data set to obtain a second multi-source marine environment data set with consistent temporal and spatial dimensions. The abnormal data analysis module is used to input the second multi-source marine environmental data set into the environmental data analysis model, so that the environmental data analysis model outputs environmental anomaly change data of the target monitoring sea area based on the second multi-source marine environmental data set; wherein, the environmental anomaly change data includes: abnormal water temperature change data and abnormal phytoplankton concentration change data, sea areas with abnormal water temperature change and sea areas with abnormal phytoplankton concentration change. The early warning module is used to provide early warnings and adjustments to the marine aquaculture strategy in the target monitoring sea area based on the abnormal environmental change data.

[0015] The implementation of this application has the following beneficial effects: This application provides a method and system for early warning of marine fisheries environment based on multi-source data fusion. The method acquires real-time marine remote sensing data of the target monitoring sea area via satellite remote sensing, and collects real-time measured water temperature data from various calibration points in the target monitoring sea area via a buoy observation platform. After combining the marine remote sensing data acquired by satellite remote sensing and the measured water temperature data from the calibration points acquired by the buoy observation platform through grid partitioning to achieve a first multi-source marine environmental data set with unified accuracy, further spatial and temporal dimension partitioning and standardization are performed to obtain a second multi-source marine environmental data set with consistent temporal and spatial dimensions. This is achieved through grid partitioning and the time dimension based on the grid partitioned data. The processing and spatial dimension processing solves the problem that existing technologies cannot synchronize temporal and spatial dimension data when guiding mariculture in complex marine environments, enabling comprehensive analysis of temporal and spatial dimension data in complex marine environments. Furthermore, after solving the problem of the inability to integrate and comprehensively evaluate temporal and spatial dimension data, based on a second multi-source marine environmental data set with unified spatiotemporal dimensions and an environmental data analysis model, it outputs data on abnormal changes in water temperature and phytoplankton concentration in the target monitoring sea area, as well as environmental anomaly data such as sea areas with abnormal changes in water temperature and phytoplankton concentration, enabling comprehensive early warning and refined adjustment of mariculture strategies in the target monitoring sea area. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating a marine fisheries environmental early warning method based on multi-source data fusion, as provided in an embodiment of this application.

[0017] Figure 2 This is a schematic diagram of the structure of a marine fisheries environmental early warning system based on multi-source data fusion, provided as an embodiment of this application. Detailed Implementation

[0018] To enable those skilled in the art to better understand the technical solution, the present application will be described in detail below with reference to the embodiments. The description in this section is only exemplary and explanatory, and should not be used to limit the scope of protection of the present application in any way.

[0019] To address the issue of inconsistent temporal and spatial data in mariculture guidance under complex marine environments, and to improve the comprehensiveness of mariculture guidance in such environments, such as... Figure 1 The image shown is an embodiment of a marine fisheries environmental early warning method based on multi-source data fusion provided in this application, comprising: Step S1: Obtain marine remote sensing data of the target monitoring sea area in real time through satellite remote sensing, and collect measured water temperature data of each calibration point in the target monitoring sea area in real time through the buoy observation platform; wherein, the marine remote sensing data includes: sea surface water temperature remote sensing data and phytoplankton concentration remote sensing data; Step S2: Perform grid division preprocessing on the marine remote sensing data and the measured water temperature data of each sea area calibration point to ensure that the marine remote sensing data and the measured water temperature data of each sea area calibration point are in the same grid division accuracy, and generate the first multi-source marine environment data set; Step S3: Perform spatial dimension partitioning and standardization processing on the first multi-source marine environment data set to obtain a second multi-source marine environment data set that is consistent in both time and spatial dimensions; Step S4: Input the second multi-source marine environmental data set into the environmental data analysis model, so that the environmental data analysis model outputs environmental anomaly change data of the target monitoring sea area based on the second multi-source marine environmental data set; wherein, the environmental anomaly change data includes: water temperature anomaly change data and phytoplankton concentration anomaly change data, sea areas with anomaly changes in water temperature and sea areas with anomaly changes in phytoplankton concentration; Step S5: Based on the abnormal environmental change data, issue early warnings and make adjustments to the marine aquaculture strategy in the target monitoring sea area.

[0020] For step S1, after determining the target monitoring sea area, marine remote sensing data of the target monitoring sea area is acquired in real time using satellite remote sensing technology based on big data acquisition. This marine remote sensing data includes sea surface temperature remote sensing data and phytoplankton concentration remote sensing data obtained from aerial images of the target monitoring sea area. Since there will be some deviation between the sea surface temperature data acquired by satellite remote sensing and the actual seawater temperature data, to make the water temperature data monitoring more accurate, this application also sets multiple sea area calibration points in the target monitoring sea area and collects the measured water temperature data of each sea area calibration point in real time through a buoy observation platform. By monitoring the sea surface temperature remote sensing data, the overall water temperature of a large area of ​​the sea can be obtained, and the overall water temperature environmental trend can be determined. By monitoring the measured water temperature data of each sea area calibration point, the overall sea area water temperature data of the large area of ​​the sea obtained by satellite remote sensing can be calibrated and filled in, so as to obtain a more accurate water temperature of the aquaculture area. Water temperature directly controls the growth rate, temperature intensity, and survival rate of fish, shrimp, and shellfish in marine aquaculture areas. Sudden changes in water temperature can also cause large-scale mortality of these organisms, directly impacting aquaculture output. Therefore, water temperature is a crucial monitoring factor for early warning of marine aquaculture strategies in this application. Furthermore, phytoplankton concentration directly reflects the abundance of food, water quality, and algae blooms in the marine environment. Phytoplankton concentration directly indicates whether there is sufficient food for fish, shrimp, and shellfish in the marine aquaculture area. When food availability is moderate (i.e., when phytoplankton concentration is moderate), farmed animals grow quickly and are less prone to disease and death. Excessively high phytoplankton concentrations can lead to red tides, green tides, and other disturbances, causing animal mortality. Conversely, low phytoplankton concentrations result in insufficient food, affecting the growth rate and yield of farmed fish, shrimp, and shellfish. Therefore, monitoring phytoplankton concentration is also a very important aspect of marine aquaculture monitoring.

[0021] It should be noted that, in order to ensure sufficient data collection to guarantee the accuracy of subsequent analysis and processing, the big data acquisition technology in this application is used throughout the entire process of multi-source marine environmental data collection. The focus is on collecting remote sensing data of sea surface water temperature obtained by satellite remote sensing, remote sensing data of phytoplankton concentration, and measured water temperature data of sea area calibration points collected by buoy observation platforms. These data obtained based on big data collection lay the foundation for subsequent data processing and analysis.

[0022] For step S2, in order to ensure that the marine remote sensing data and the measured water temperature data of each sea area calibration point are within the same grid division accuracy, after completing the data acquisition step of step S1, big data processing operations are performed on the acquired marine remote sensing data and the measured water temperature data of each sea area calibration point, etc. This big data processing operation includes the processing in steps S2 and S3. In step S2, grid division preprocessing is mainly performed to ensure that the marine remote sensing data and the measured water temperature data of each sea area calibration point are within the same grid division accuracy, and the first multi-source marine environment data set is generated based on the divided data.

[0023] In a preferred embodiment, the preprocessing of the marine remote sensing data and the measured water temperature data of each sea area calibration point, to ensure that the marine remote sensing data and the measured water temperature data of each sea area calibration point are within the same grid division accuracy, and generating a first multi-source marine environmental data set, includes: acquiring the latitude and longitude data of the target monitoring sea area; determining the grid boundary of the target monitoring sea area based on the latitude and longitude data of the target monitoring sea area; generating a grid to be filled based on a preset grid division resolution and the grid boundary of the target monitoring sea area; wherein the grid to be filled contains several sub-grids to be filled, and each sub-grid to be filled has the same area; and generating the first multi-source marine environmental data set based on the grid to be filled, the marine remote sensing data, and the measured water temperature data of each sea area calibration point.

[0024] Specifically, the purpose of grid division is to place ocean remote sensing data acquired by satellite remote sensing and measured water temperature data from calibration points in various sea areas acquired by buoy observation platforms within the same grid's specified precision, ensuring that both are within the same calculation and comparison range. In this embodiment, this step is achieved by acquiring the latitude and longitude data of the target monitoring sea area, specifically the latitude and longitude data corresponding to the four sides of the smallest quadrilateral that can encompass the target monitoring sea area. The size of the grid required to encompass the target monitoring sea area is determined using this latitude and longitude data, and the grid boundary is also determined using the same latitude and longitude data. After determining the grid boundary, the grid size is determined. The grid required to encompass the target monitoring sea area is then divided according to a preset grid division resolution, resulting in a grid to be filled containing several sub-grids of equal area. A first multi-source ocean environmental data set is generated based on the grid to be filled, the ocean remote sensing data, and the measured water temperature data from calibration points in each sea area.

[0025] It should be noted that the preset grid division resolution can be set according to the requirements. In this embodiment, the preset grid division resolution is set to 5 meters by 5 meters, that is, each sub-grid to be filled is a square grid with a length of 5 meters and a width of 5 meters.

[0026] For step S3, after completing the generation of the grid to be filled and data collection, it is necessary to further fill the grid with data while ensuring that the filled grid data is consistent in both time and spatial dimensions, so as to integrate the data in both time and spatial dimensions into the grid corresponding to the target monitoring sea area.

[0027] In a preferred embodiment, the step of performing spatial dimension partitioning and standardization processing on the first multi-source marine environmental data set to obtain a second multi-source marine environmental data set consistent in both time and space includes: performing spatial dimension partitioning and standardization processing on the first multi-source marine environmental data set to obtain a third multi-source marine environmental data set consistent in both space; and performing time dimension partitioning and standardization processing on the third multi-source marine environmental data set to obtain a second multi-source marine environmental data set consistent in both time and space.

[0028] Specifically, when integrating data into data that is consistent in both time and spatial dimensions, it is usually necessary to first perform spatial dimension standardization processing, and then perform time dimension standardization processing.

[0029] In a preferred embodiment, the spatial dimension partitioning and standardization process for the first multi-source marine environmental data set to obtain a third multi-source marine environmental data set with consistent spatial dimensions includes: partitioning the marine remote sensing data and the measured water temperature data of each sea area calibration point according to each sub-grid to be filled in the first multi-source marine environmental data set, to obtain the marine remote sensing data and the measured water temperature data of each sea area calibration point of each sub-grid to be filled; filling each sub-grid to be filled with data according to the marine remote sensing data and the measured water temperature data of each sea area calibration point of each sub-grid to be filled, to generate a third multi-source marine environmental data set containing several filled data sub-grids and several unfilled data sub-grids.

[0030] Specifically, the ocean remote sensing data and the measured sea surface temperature data of each sea area calibration point are divided according to the sub-grids to be filled in the constructed grid to be filled. During the division, firstly, the remote sensing images corresponding to the sea surface temperature data in the ocean remote sensing data are determined. Then, the remote sensing images corresponding to the sea surface temperature data are calibrated with the grid to be filled using latitude and longitude, so that the calibrated grid to be filled can accurately segment the remote sensing images corresponding to the sea surface temperature data based on the latitude and longitude information. Then, the segmented remote sensing images corresponding to the sea surface temperature data are matched with the corresponding sub-grids to be filled, so that each sub-grid to be filled acquires the sea surface temperature data corresponding to the area enclosed by the current sub-grid. Similarly, after completing the division of the sea surface temperature data, the ocean remote sensing data... The remote sensing images corresponding to the phytoplankton concentration data are calibrated with the grid to be filled using latitude and longitude coordinates. This allows the calibrated grid to accurately segment the remote sensing images corresponding to the phytoplankton concentration data based on latitude and longitude information. The segmented phytoplankton concentration data is then matched with the corresponding sub-grid to be filled, so that each sub-grid acquires the phytoplankton concentration data corresponding to the area enclosed by the current sub-grid. For the measured water temperature data from each sea area calibration point, the data is filled into the corresponding sub-grid to be filled based on the latitude and longitude information of each calibration point. After the above processing, several filled data sub-grids and several unfilled data sub-grids are obtained after spatial scale alignment. It should be noted that during the process of filling the grid with marine remote sensing data and measured water temperature data from each calibration point, there may be cases where the latitude and longitude coordinates of the data fall within the boundaries of the sub-grid to be filled. Based on this situation, this embodiment assumes that if the data falls within the left boundary of the grid, it belongs to the grid adjacent to the left of the current grid; if it falls within the upper boundary of the grid, it belongs to the grid adjacent to the upper edge of the current grid.

[0031] In a preferred embodiment, the time-dimension standardization process for the third multi-source marine environmental data set, resulting in a second multi-source marine environmental data set with consistent time and spatial dimensions, includes: acquiring the spatial distance and time identifier between each unfilled data sub-grid and each filled data sub-grid in the third multi-source marine environmental data set; performing time-dimension standardization on the marine remote sensing data and measured water temperature data of sea area calibration points of each filled data sub-grid based on the time identifier, generating each filled data sub-grid with consistent time dimension; and for each unfilled data sub-grid, determining the spatial distance between the unfilled data sub-grid and each filled data sub-grid... Define the spatial distance weights of each adjacent subgrid of the currently unfilled data subgrid; obtain the marine remote sensing data and measured sea surface temperature data of the sea area calibration points corresponding to each adjacent subgrid of the currently unfilled data subgrid; generate the marine remote sensing data and measured sea surface temperature data of the sea area calibration points corresponding to the currently unfilled data subgrid based on the marine remote sensing data, measured sea surface temperature data of the sea area calibration points, and the spatial distance weights of each adjacent subgrid; perform data interpolation on the currently unfilled data subgrid to generate interpolated data subgrids; and generate a third multi-source marine environmental data set with consistent temporal and spatial dimensions based on each interpolated data subgrid and each filled data subgrid.

[0032] Specifically, for the obtained filled and unfilled data subgrids, the filled data subgrids first need to undergo time-division standardization processing. This ensures that the data in each filled data subgrid is divided according to a preset time interval, allowing the data in each filled data subgrid to form a time series based on time identifiers or timestamps. This enables the observation of data changes and fluctuations in the corresponding sea area over time from a single filled data subgrid, while also providing sufficiently refined data for data interpolation in the unfilled data subgrids. The time-division standardization process involves dividing the data in the current filled data subgrid according to a preset time interval, such as hourly or daily. For example, if satellite remote sensing data is collected daily and buoy observation platform data is collected hourly, then using hourly intervals as the dividing standard, the marine remote sensing data collected by satellite needs to be divided. This can be done through linear interpolation to divide the collected marine remote sensing data into hourly data, thus ensuring a clear and continuous time dimension in each filled data subgrid. Using 24-hour intervals as the dividing standard, the data collected by the buoy observation platform needs to be integrated. This can be done through weighted averaging to determine the measured temperature data of each sea area calibration point daily. The selection of the time identifier corresponds to the selection of the time interval. For short-term trend estimation, a smaller time interval, such as hourly, is used; for long-term trend estimation, a larger time interval, such as daily or monthly, is used to determine the monthly and quarterly trends. Through the above processing, the data change trends under time series and time dynamic fluctuations can be determined in each filled data subgrid.

[0033] After processing the filled data subgrids, it is necessary to fill the unfilled data subgrids with data based on the filled data subgrids to ensure the data integrity of the grid corresponding to the target monitoring sea area. Generally speaking, the data obtained from big data collection can basically cover the grid corresponding to the target monitoring sea area. However, during the grid division process, due to the possibility of data falling into the grid boundaries, some unfilled data subgrids will exist after grid division. In this case, it is necessary to fill the data with adjacent filled data subgrids. During data infilling, for each unfilled data subgrid, the spatial distance between each filled data subgrid and the current unfilled data subgrid needs to be obtained. This spatial distance is the relative distance between each filled data subgrid and the current unfilled data subgrid. The adjacent subgrids of the current unfilled data subgrid are determined by the spatial distance, and the spatial distance weights of each adjacent subgrid of the current unfilled data subgrid are determined. Usually, an unfilled data subgrid has at least two adjacent subgrids. Using the filled marine remote sensing data and marine calibration point data in these adjacent subgrids, as well as the spatial distance weights of each adjacent subgrid of the current unfilled data subgrid, the data interpolation of the current unfilled data subgrid is completed by weighted calculation, and the corresponding interpolated data subgrid is generated.

[0034] In a preferred embodiment, the step of generating marine remote sensing data and measured sea surface temperature data corresponding to the current unfilled data subgrid based on the marine remote sensing data, measured sea surface temperature data of the sea area calibration point, and the spatial distance weights of each adjacent subgrid includes: determining the weights of the sea surface temperature remote sensing data, phytoplankton concentration remote sensing data, and measured sea surface temperature data of the sea area calibration point for each adjacent subgrid based on the spatial distance weights between the current unfilled data subgrid and each adjacent subgrid; performing a weighted average of the weights of the sea surface temperature remote sensing data of each adjacent subgrid to obtain the sea surface temperature remote sensing data of the current unfilled data subgrid; performing a weighted average of the weights of the phytoplankton concentration remote sensing data of each adjacent subgrid to obtain the phytoplankton concentration remote sensing data of the current unfilled data subgrid; and performing a weighted average of the weights of the measured sea surface temperature data of the sea area calibration point for each adjacent subgrid to obtain the measured sea surface temperature data of the current unfilled data subgrid.

[0035] Specifically, taking sea surface temperature remote sensing data as an example, the current unfilled data subgrid is set to have four adjacent subgrids with corresponding spatial distance weights of 0.8, 0.9, 0.7, and 0.8. When interpolating the current unfilled data subgrid, the spatial distance weights are first assigned to the sea surface temperature remote sensing data of the four adjacent subgrids. Then, the sea surface temperature remote sensing data of the four adjacent subgrids with assigned spatial distance weights are weighted and averaged to obtain the sea surface temperature remote sensing data with a weight of 0.8. The sea surface temperature remote sensing data with a weight of 0.8 is then used as the sea surface temperature remote sensing data of the current unfilled data subgrid.

[0036] It should be added that the spatial distance weight between any adjacent subgrid and the unfilled data subgrid can be determined by obtaining the latitude and longitude coordinates of the center point of the adjacent subgrid and the latitude and longitude coordinates of the centerline point of the unfilled data subgrid, calculating the spherical distance between the two points using the Haversine formula, and then determining the weight through inverse distance calculation. For example, the latitude and longitude coordinates of the center point of the adjacent subgrid are set as (…). , The latitude and longitude coordinates of the line points in the unfilled data subgrid are ( , The spherical distance between two points can be calculated using the Haversine formula. : ; Where R is the Earth's radius, which is generally taken as 6371 km; Calculate the spatial distance weight of adjacent subgrids : ; Where α is the distance attenuation coefficient, which is usually taken as 1 or 2, and is usually taken as 2 in marine environments.

[0037] After completing the above processing, we obtain multi-source marine environmental data, which contains complete data and is consistent in both spatial and temporal dimensions, divided by a grid. The corresponding set is the second multi-source marine environmental data set mentioned above. At this point, a comprehensive assessment of the dynamic changes in environmental data of the target monitoring sea area can be conducted based on the second multi-source marine environmental data set.

[0038] However, after the above grid division and interpolation processing, the resulting second multi-source marine environmental dataset usually still has certain errors, which need to be corrected. At the same time, corresponding analysis and processing are required based on the data in each sub-grid.

[0039] In a preferred embodiment, after obtaining the second multi-source marine environmental dataset, the process further includes: For each filled data subgrid, the sea surface temperature remote sensing data, phytoplankton concentration remote sensing data, and measured sea temperature data from the sea area calibration point are weighted and fused according to preset weighting weights to obtain the comprehensive environmental index of the current filled data subgrid. For each filled data subgrid, spatial error assessment is performed to generate the numerical error and positional error of the current filled data subgrid. Based on the current sea surface temperature remote sensing data and the measured sea temperature data from the sea area calibration point, the comprehensive sea temperature data of the current filled data subgrid is determined. Pearson correlation coefficient analysis is used to determine the correlation between the comprehensive sea temperature data and the phytoplankton concentration remote sensing data, and the correlation coefficient of the current filled data subgrid is generated. For each interpolated data subgrid, the sea surface temperature remote sensing data, phytoplankton concentration remote sensing data, and measured sea temperature data from the sea area calibration point are weighted and fused according to preset weighting weights to obtain the comprehensive environmental index of the current interpolated data subgrid. For each interpolated data subgrid, spatial error assessment is performed to generate the numerical error and positional error of the current interpolated data subgrid. Based on the current sea surface temperature remote sensing data and the measured sea temperature data from the sea area calibration point, the comprehensive sea temperature data of the current interpolated data subgrid is determined. Pearson correlation coefficient analysis is used to determine the correlation between the comprehensive sea temperature data and the phytoplankton concentration remote sensing data, and the correlation coefficient of the current interpolated data subgrid is generated. Based on the grid identifiers of each filled data subgrid, remote sensing data of sea surface temperature, remote sensing data of phytoplankton concentration, measured water temperature data of sea area calibration points, comprehensive environmental indicators, numerical errors, location errors, and correlation coefficients, and the grid identifiers of each interpolated data subgrid, remote sensing data of sea surface temperature, remote sensing data of phytoplankton concentration, measured water temperature data of sea area calibration points, comprehensive environmental indicators, numerical errors, location errors, and correlation coefficients, a multi-source marine environmental data matrix corresponding to the second multi-source marine environmental data set is generated. The step of inputting the second multi-source marine environmental data set into the environmental data analysis model so that the environmental data analysis model outputs dynamic environmental anomaly change data of the target monitoring sea area based on the second multi-source marine environmental data set includes: inputting the multi-source marine environmental data matrix corresponding to the second multi-source marine environmental data set into the environmental data analysis model so that the environmental data analysis model outputs dynamic environmental anomaly change data of the target monitoring sea area based on the multi-source marine environmental data matrix corresponding to the second multi-source marine environmental data set.

[0040] Specifically, for each filled data subgrid, the remote sensing data of sea surface temperature, phytoplankton concentration, and measured water temperature data from the sea area calibration point are weighted and fused according to preset weighting weights to obtain the comprehensive environmental index of the current filled data subgrid. For example, if the weight of the remote sensing data of sea surface temperature is set to 0.3, the weight of the measured water temperature data from the sea area calibration point is set to 0.3, and the weight of the remote sensing data of phytoplankton concentration is set to 0.4, then the remote sensing data of sea surface temperature, phytoplankton concentration, and measured water temperature data from the sea area calibration point are weighted and fused according to the set weights to obtain the comprehensive environmental index of the same time dimension. If an hourly or daily comprehensive environmental index is obtained, this comprehensive environmental index indicates the comprehensive estimate of the phytoplankton concentration and water temperature in the sea area corresponding to the current filled data subgrid, which can be used to assess whether the sea area corresponding to the current filled data subgrid is suitable for mariculture.

[0041] For example, when calculating the comprehensive environmental index of a filled data subgrid, the remote sensing data of sea surface temperature and the measured water temperature data of the sea area calibration point are both set to 25℃, and the phytoplankton concentration is 2 mg / m³. The 25℃ data is then compared with a preset water temperature range. If it falls within the range of 20℃ to 29℃, 25℃ is converted to a water temperature calibration value of 1; if it is below 20℃, it is converted to a water temperature calibration value of 0.5; and if it is above 29℃, it is converted to a water temperature calibration value of 0. This classification standard is based on the temperature suitable for fish and shrimp farming. A water temperature calibration value of 1 indicates that the water temperature is the optimal temperature for growth, a water temperature calibration value of 0.5 indicates that growth is possible but not suitable for long-term farming, and a water temperature calibration value of 0 indicates that growth is not possible. Similarly, the phytoplankton concentration is compared with a preset phytoplankton concentration range. If it falls within the range of 0.8 mg / m³ to 3.5 mg / m³, the phytoplankton concentration is also considered. When the concentration is less than 0.8 mg / m³, 2 mg / m³ is converted to a concentration calibration value of 1; when it is less than 0.8 mg / m³, 2 mg / m³ is converted to a concentration calibration value of 0.5; and when it is greater than 3.5 mg / m³, 2 mg / m³ is converted to a concentration calibration value of 0. This classification standard is based on the range of phytoplankton concentrations suitable for fish and shrimp aquaculture. A concentration calibration value of 1 indicates that the phytoplankton concentration is the optimal concentration for growth; a concentration calibration value of 0.5 indicates that aquaculture is possible but not suitable for long-term aquaculture; and a concentration calibration value of 0 indicates that aquaculture is unsuitable. A weighted fusion calculation yields an environmental comprehensive index of 0.3*1 + 0.3*1 + 0.4*1 = 1. An environmental comprehensive index of 1 indicates that both the water temperature and phytoplankton concentration in this area are optimal. If the environmental comprehensive index is less than 0.7, it indicates that at least one of the two indicators—water temperature or phytoplankton concentration—is unsuitable for animal growth. It should be noted that the above classification range is illustrative; the specific classification standard is determined based on the specific aquaculture category in marine aquaculture. This illustrative range does not specifically limit the scope of protection of this application. In addition, spatial error assessment is required for the filled data subgrids. This assessment can utilize spatial interpolation algorithms such as Kriging to calculate the root mean square error (RMSE) between the predicted and actual acquisition locations of the data in the filled data subgrids. The RMS error is then used to determine the numerical and positional errors of the current filled data subgrid. Furthermore, the correlation between water temperature data and phytoplankton concentration data needs to be analyzed. First, the sea surface water temperature remote sensing data is refined using measured water temperature data from calibration points in various sea areas to obtain comprehensive water temperature data. Then, Pearson correlation coefficient analysis is used to determine the correlation between the comprehensive water temperature data and the phytoplankton concentration remote sensing data. This correlation is used to determine whether there is a positive or negative correlation between water temperature and phytoplankton concentration, thus assisting in subsequent adjustments to marine aquaculture strategies.

[0042] Similarly, the same operation is performed on the interpolated data subgrids to obtain the comprehensive environmental index, numerical error, location error, and correlation coefficient of the interpolated data subgrids. To facilitate subsequent data processing by the model, a multi-source marine environmental data matrix corresponding to the second multi-source marine environmental data set is generated based on the grid identifiers of each filled data subgrid, sea surface temperature remote sensing data, phytoplankton concentration remote sensing data, measured sea surface temperature data at sea calibration points, comprehensive environmental indexes, numerical error, location error, and correlation coefficients of each interpolated data subgrid. It should be noted that the grid identifier is used to determine the coordinates of the area enclosed by the grid, and typically includes the latitude and longitude coordinates of the four vertices of the grid.

[0043] For step S4, in a preferred embodiment, inputting the multi-source marine environmental data matrix corresponding to the second multi-source marine environmental data set into the environmental data analysis model, so that the environmental data analysis model outputs environmental anomaly change data of the target monitoring sea area based on the multi-source marine environmental data matrix corresponding to the second multi-source marine environmental data set, includes: The multi-source marine environmental data matrix is ​​input into the environmental data analysis model so that the environmental data analysis model can determine the initial abnormal water temperature change data and the initial abnormal water temperature change area based on the sea surface temperature remote sensing data and the measured water temperature data of the sea area calibration point in the multi-source marine environmental data matrix, and determine the initial abnormal phytoplankton concentration change data and the initial abnormal phytoplankton concentration change area based on the phytoplankton concentration remote sensing data in the multi-source marine environmental data matrix. The initial water temperature anomaly data, initial water temperature anomaly data, initial phytoplankton concentration anomaly data, and initial phytoplankton concentration anomaly data are corrected using the comprehensive environmental indicators, numerical errors, location errors, and correlation coefficients in the multi-source marine environmental data matrix. The output data on water temperature anomaly data, phytoplankton concentration anomaly data, and the target monitoring sea area are then output.

[0044] Specifically, after inputting the multi-source marine environmental data matrix into the environmental data analysis model, the model determines initial anomaly data in sea surface temperature based on remote sensing data and measured sea surface temperature data from calibration points within the matrix. This initial anomaly data is time-series data, used to identify the range, time, and locations of anomalous fluctuations within the anomaly period. The model then identifies the initial sea area experiencing anomaly data using the grid markers corresponding to this data. Similarly, the model determines initial anomaly data in phytoplankton concentration based on remote sensing data from the multi-source marine environmental data matrix. This initial anomaly data is also time-series data, used to identify the range, time, and locations of anomalous fluctuations within the anomaly period. The model then identifies the initial sea area experiencing anomaly data using the grid markers corresponding to this data. The initial water temperature anomaly data, initial water temperature anomaly data, initial phytoplankton concentration anomaly data, and initial phytoplankton concentration anomaly data are corrected using the comprehensive environmental indicators, numerical errors, location errors, and correlation coefficients in the multi-source marine environmental data matrix. The output data on water temperature anomaly data, phytoplankton concentration anomaly data, and the target monitoring sea area are then output.

[0045] Preferably, when correcting initial water temperature anomaly data, the accuracy of the data is ensured primarily by checking for calculation errors using comprehensive environmental indicators and numerical errors. When correcting the initial water temperature anomaly area, location errors are primarily corrected to ensure the accuracy of the obtained location of the anomaly area. When correcting initial phytoplankton concentration anomaly data, corrections are primarily made using comprehensive environmental indicators, numerical errors, and correlation coefficients. If the correlation coefficient is unreasonable, the initial phytoplankton concentration anomaly data needs to be re-optimized and re-output to ensure the correlation between phytoplankton concentration anomaly data and water temperature, and the accuracy of the phytoplankton concentration anomaly data itself.

[0046] For the aforementioned environmental data analysis model, in this embodiment, to ensure the sensitivity and accuracy of processing changes in data over time periods, an environmental data analysis model based on a Long Short-Term Memory (LSTM) network is adopted. When constructing the environmental data analysis model, an initial environmental data analysis model must first be built based on LSTM, and a training set must be constructed. The data in this training set includes pre-processed and verified multi-source marine environmental data matrix samples, as well as environmental anomaly change data samples corresponding to each multi-source marine environmental data matrix sample.

[0047] Supervised learning is used to input multi-source marine environmental data matrix samples into an initial environmental data analysis model, so that the initial environmental data analysis model outputs a predicted environmental anomaly change data sample. This predicted environmental anomaly change data sample is compared with the corresponding environmental anomaly change data sample of the multi-source marine environmental data matrix sample to determine the loss value. The loss function is adjusted based on the loss value and iterated continuously. When the loss function of the initial environmental data analysis model converges, the aforementioned environmental data analysis model is generated.

[0048] For step S5, in a preferred embodiment, the step of providing early warning and adjusting the mariculture strategy for the target monitoring sea area based on the abnormal environmental change data includes: Based on the data of abnormal water temperature changes, the locations of abnormal water temperature changes are determined. Based on the locations of abnormal water temperature changes and the sea areas of abnormal water temperature changes, the sub-regions of abnormal water temperature changes corresponding to each location are determined. The sub-regions of abnormal water temperature changes are marked as areas of risk of sudden water temperature changes. The remaining areas of the sea areas of abnormal water temperature changes, except for the areas of risk of sudden water temperature changes, are marked as areas of potential risk of sudden water temperature changes. Based on the data on abnormal changes in phytoplankton concentration, the locations of abnormal changes in phytoplankton concentration are determined. Based on the locations of abnormal changes in phytoplankton concentration and the sea areas where abnormal changes in phytoplankton concentration occur, sub-regions of abnormal changes in phytoplankton concentration are determined for each location of abnormal changes in phytoplankton concentration. The sub-regions of abnormal changes in phytoplankton concentration are marked as areas of risk of sudden changes in phytoplankton concentration. The remaining areas of the sea areas where abnormal changes in phytoplankton concentration occur, except for the areas of risk of sudden changes in phytoplankton concentration, are marked as areas of potential risk of sudden changes in phytoplankton concentration. The overlapping areas are determined based on the risk areas of sudden changes in water temperature and the risk areas of sudden changes in phytoplankton concentration, and the overlapping areas are marked as high-risk areas. The remaining target monitoring areas, except for the sea areas with abnormal changes in water temperature and the sea areas with abnormal changes in phytoplankton concentration, are marked as risk-free areas. Based on high-risk areas, areas at risk of sudden changes in water temperature, areas at risk of sudden changes in phytoplankton concentration, areas at risk of potential sudden changes in water temperature, and areas at risk of potential sudden changes in phytoplankton concentration, tiered early warning information is generated. The scale of mariculture in high-risk areas, areas at risk of sudden changes in water temperature, and areas at risk of sudden changes in phytoplankton concentration is reduced. The mariculture yield in areas at risk of potential sudden changes in water temperature and areas at risk of potential sudden changes in phytoplankton concentration is monitored in real time, and the scale of mariculture in risk-free areas is increased.

[0049] Specifically, after identifying high-risk areas, areas with sudden water temperature changes, areas with sudden phytoplankton concentration changes, potential areas with sudden water temperature changes, and potential areas with sudden phytoplankton concentration changes, early warning information prohibiting mariculture is generated for high-risk areas, and mariculture strategies are adjusted to reduce mariculture in high-risk areas. For areas with only sudden water temperature changes or only sudden phytoplankton concentration changes, early warning information indicating short-term risks to mariculture is generated, and mariculture strategies are adjusted to reduce the scale of mariculture in these areas in the short term. Simultaneously, mariculture yield per unit time is monitored to monitor the impact of sudden water temperature changes or sudden phytoplankton concentration changes on mariculture yield in real time. For example, by catching the same number of fish in the area daily... Weighing is used to determine if the mutation has affected fish growth. If so, feed supplementation or relocation will be implemented based on the impact. Simultaneously, the number of dead fish and disease monitoring of harvested fish are used to determine if they have affected the health of farmed fish. If so, adjustments are made based on the actual impact. Low-risk warnings are issued for areas with potential water temperature and phytoplankton concentration mutation risks. Water temperature and phytoplankton concentration changes are monitored in real time during marine aquaculture to adjust marine aquaculture strategies in a timely manner before these areas shift towards high-risk, high-risk, or high-risk areas. No warnings are issued for risk-free areas, allowing for the expansion of marine aquaculture scale.

[0050] By implementing the above methods, this application solves the problem of existing technologies being unable to synchronize temporal and spatial data in guiding mariculture in complex marine environments, thus improving the comprehensiveness of mariculture guidance in complex marine environments. Furthermore, after addressing the integration of temporal and spatial data, this application also combines big data acquisition technology and a time-series analysis model to construct an environmental data analysis model for analyzing anomalies in the integrated multi-source marine environmental data. This identifies anomalous change locations and implements tiered risk management. Based on tiered risk management, it provides comprehensive and accurate guidance strategies for mariculture, significantly improving the practicality and accuracy of mariculture guidance in complex marine environments.

[0051] Based on the above method embodiments, this application provides corresponding system embodiments.

[0052] like Figure 2 As shown, one embodiment of this application provides a marine fisheries environment early warning system based on multi-source data fusion. The system includes: a multi-source data acquisition module, a spatiotemporal data synchronization module, an anomaly data analysis module, and an early warning module. The multi-source data acquisition module is used to acquire marine remote sensing data of the target monitoring sea area in real time through satellite remote sensing, and to acquire measured water temperature data of each calibration point in the target monitoring sea area in real time through the buoy observation platform; wherein, the marine remote sensing data includes: sea surface water temperature remote sensing data and phytoplankton concentration remote sensing data; The spatiotemporal data synchronization module is used to perform grid division preprocessing on the marine remote sensing data and the measured water temperature data of each sea area calibration point, so that the marine remote sensing data and the measured water temperature data of each sea area calibration point are in the same grid division accuracy, generating a first multi-source marine environment data set; and to perform spatial dimension division standardization processing and temporal dimension division standardization processing on the first multi-source marine environment data set to obtain a second multi-source marine environment data set with consistent temporal and spatial dimensions. The abnormal data analysis module is used to input the second multi-source marine environmental data set into the environmental data analysis model, so that the environmental data analysis model outputs environmental anomaly change data of the target monitoring sea area based on the second multi-source marine environmental data set; wherein, the environmental anomaly change data includes: abnormal water temperature change data and abnormal phytoplankton concentration change data, sea areas with abnormal water temperature change and sea areas with abnormal phytoplankton concentration change. The early warning module is used to provide early warnings and adjustments to the marine aquaculture strategy in the target monitoring sea area based on the abnormal environmental change data.

[0053] It should be noted that, in this document, the terms "comprising," "including," and any other variations are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Specific examples have been used in this document to illustrate the principles and implementation methods of the technical solutions of this application. The above examples are only for the purpose of helping to understand the methods and core ideas of this application. The above descriptions are merely preferred embodiments of this application. It should be pointed out that, due to the limitations of written expression and the objective existence of infinite specific structures, those skilled in the art can make several improvements, modifications, or changes without departing from the principles of this application, and can also combine the above technical features in an appropriate manner; these improvements, modifications, changes, or combinations, or the direct application of the concept and technical solutions of this application to other situations without modification, should all be considered within the scope of protection of this application.

Claims

1. A method for early warning of marine fisheries environment using multi-source data fusion, characterized in that, include: Ocean remote sensing data of the target monitoring sea area is acquired in real time by satellite remote sensing, and measured water temperature data of each calibration point in the target monitoring sea area is collected in real time by buoy observation platform; wherein, the ocean remote sensing data includes: sea surface water temperature remote sensing data and phytoplankton concentration remote sensing data; the ocean remote sensing data and the measured water temperature data of each calibration point are preprocessed by grid division to ensure that the ocean remote sensing data and the measured water temperature data of each calibration point are within the same grid division accuracy, thereby generating the first multi-source ocean environmental data set; The first multi-source marine environment data set is subjected to spatial dimension partitioning and time dimension partitioning and standardization processing to obtain a second multi-source marine environment data set that is consistent in both time and spatial dimensions. The second multi-source marine environmental data set is input into the environmental data analysis model so that the environmental data analysis model outputs environmental anomaly change data of the target monitoring sea area based on the second multi-source marine environmental data set; wherein, the environmental anomaly change data includes: water temperature anomaly change data and phytoplankton concentration anomaly change data, sea areas with anomaly change in water temperature and sea areas with anomaly change in phytoplankton concentration; Based on the data on abnormal environmental changes, early warnings and adjustments are made to the marine aquaculture strategies in the target monitoring sea area.

2. The marine fisheries environment early warning method based on multi-source data fusion according to claim 1, characterized in that, The process of performing grid division preprocessing on the marine remote sensing data and the measured water temperature data from each sea area calibration point, so that the marine remote sensing data and the measured water temperature data from each sea area calibration point are within the same grid division accuracy, generates a first multi-source marine environmental data set, including: Acquire the latitude and longitude data of the target monitoring sea area, and determine the grid boundary of the target monitoring sea area based on the latitude and longitude data; A grid to be filled is generated based on a preset grid division resolution and the grid boundary of the target monitoring sea area; wherein, the grid to be filled contains several sub-grids to be filled, and each sub-grid to be filled has the same area; The first multi-source marine environmental data set is generated based on the grid to be filled, the marine remote sensing data, and the measured water temperature data of each sea area calibration point.

3. The marine fisheries environment early warning method based on multi-source data fusion according to claim 2, characterized in that, The step of performing spatial dimension partitioning and standardization processing on the first multi-source marine environmental data set to obtain a second multi-source marine environmental data set consistent in both time and spatial dimensions includes: The first multi-source marine environment data set is subjected to spatial dimension division and standardization processing to obtain a third multi-source marine environment data set with consistent spatial dimensions. The third multi-source marine environmental data set is subjected to time dimension division and standardization processing to obtain a second multi-source marine environmental data set that is consistent in both time and spatial dimensions.

4. The marine fisheries environment early warning method based on multi-source data fusion according to claim 3, characterized in that, The step of performing spatial dimension partitioning and standardization processing on the first multi-source marine environmental data set to obtain a third multi-source marine environmental data set with consistent spatial dimensions includes: Based on each sub-grid to be filled in the first multi-source marine environmental data set, the marine remote sensing data and the measured water temperature data of each sea area calibration point are divided into spatial dimensions to obtain the marine remote sensing data and the measured water temperature data of each sea area calibration point of each sub-grid to be filled. Based on the marine remote sensing data of each subgrid to be filled and the measured water temperature data of the sea area calibration points of each subgrid to be filled, data filling is performed on each subgrid to be filled, generating a third multi-source marine environmental data set containing several filled data subgrids and several unfilled data subgrids.

5. A marine fisheries environmental early warning method based on multi-source data fusion according to claim 4, characterized in that, The process of performing time-dimension standardization on the third multi-source marine environmental data set to obtain a second multi-source marine environmental data set with consistent time and spatial dimensions includes: Obtain the spatial distance and temporal identifier between each unfilled data subgrid and each filled data subgrid in the third multi-source marine environmental dataset; Based on the time stamp, the marine remote sensing data and the measured water temperature data of the sea area calibration point of each filled data subgrid are standardized in terms of time dimension to generate each filled data subgrid with consistent time dimension; For each unfilled data subgrid, determine the spatial distance weights of each adjacent subgrid of the current unfilled data subgrid and each filled data subgrid based on the spatial distance between the unfilled data subgrid and each filled data subgrid; Obtain the marine remote sensing data and measured sea temperature data of the sea area calibration point corresponding to each adjacent subgrid of the current unfilled data subgrid. Based on the marine remote sensing data, measured sea temperature data of the sea area calibration point corresponding to the adjacent subgrid of the current unfilled data subgrid and the spatial distance weight of each adjacent subgrid, generate the marine remote sensing data and measured sea temperature data of the sea area calibration point corresponding to the current unfilled data subgrid. Then, perform data interpolation on the current unfilled data subgrid to generate the interpolated data subgrid. A second multi-source marine environmental data set with consistent temporal and spatial dimensions is generated based on each interpolated data subgrid and each filled data subgrid.

6. The marine fisheries environment early warning method based on multi-source data fusion according to claim 5, characterized in that, The process of generating marine remote sensing data and measured sea temperature data corresponding to the current unfilled data subgrid based on the marine remote sensing data, measured sea temperature data at sea calibration points, and spatial distance weights of each adjacent subgrid includes: Based on the spatial distance weights between the currently unfilled data subgrid and each adjacent subgrid, determine the weights of the sea surface temperature remote sensing data, the phytoplankton concentration remote sensing data, and the measured sea surface temperature data at the sea area calibration point for each adjacent subgrid. The sea surface temperature remote sensing data of the current unfilled subgrid is obtained by weighting the weights of the sea surface temperature remote sensing data of each adjacent subgrid. The phytoplankton concentration remote sensing data of the current unfilled data subgrid is obtained by weighting the weights of the phytoplankton concentration remote sensing data of each adjacent subgrid. The measured water temperature data of the sea area calibration points of each adjacent subgrid are weighted and averaged to obtain the measured water temperature data of the sea area calibration points of the current unfilled data subgrid.

7. A marine fisheries environmental early warning method based on multi-source data fusion according to claim 6, characterized in that, After obtaining the second multi-source marine environmental dataset, it also includes: For each filled data subgrid, the sea surface temperature remote sensing data, phytoplankton concentration remote sensing data and measured sea area calibration point water temperature data of the current filled data subgrid are weighted and fused according to the preset weighting weight to obtain the comprehensive environmental index of the current filled data subgrid. For each filled data subgrid, spatial error assessment is performed to generate the numerical error and positional error of the current filled data subgrid. Based on the current sea surface water temperature remote sensing data and the measured water temperature data of the sea area calibration point, the comprehensive water temperature data of the current filled data subgrid is determined. Pearson correlation coefficient analysis is used to determine the correlation between the comprehensive water temperature data and the phytoplankton concentration remote sensing data, and the correlation coefficient of the current filled data subgrid is generated. For each interpolated data subgrid, the sea surface temperature remote sensing data, phytoplankton concentration remote sensing data, and measured sea area calibration point water temperature data of the current interpolated data subgrid are weighted and fused according to the preset weighting weights to obtain the comprehensive environmental index of the current interpolated data subgrid. For each interpolated data subgrid, spatial error assessment is performed to generate the numerical error and positional error of the current interpolated data subgrid. Based on the current sea surface temperature remote sensing data and the measured water temperature data of the sea area calibration point, the comprehensive water temperature data of the current interpolated data subgrid is determined. Pearson correlation coefficient analysis is used to determine the correlation between the comprehensive water temperature data and the phytoplankton concentration remote sensing data, and the correlation coefficient of the current interpolated data subgrid is generated. Based on the grid identifiers of each filled data subgrid, remote sensing data of sea surface temperature, remote sensing data of phytoplankton concentration, measured water temperature data of sea area calibration points, comprehensive environmental indicators, numerical errors, location errors, and correlation coefficients, and the grid identifiers of each interpolated data subgrid, remote sensing data of sea surface temperature, remote sensing data of phytoplankton concentration, measured water temperature data of sea area calibration points, comprehensive environmental indicators, numerical errors, location errors, and correlation coefficients, a multi-source marine environmental data matrix corresponding to the second multi-source marine environmental data set is generated. The step of inputting the second multi-source marine environmental data set into the environmental data analysis model, so that the environmental data analysis model outputs dynamic environmental anomaly change data of the target monitoring sea area based on the second multi-source marine environmental data set, includes: The multi-source marine environmental data matrix corresponding to the second multi-source marine environmental data set is input into the environmental data analysis model so that the environmental data analysis model outputs environmental anomaly change data of the target monitoring sea area based on the multi-source marine environmental data matrix corresponding to the second multi-source marine environmental data set.

8. A marine fisheries environmental early warning method based on multi-source data fusion according to claim 7, characterized in that, The step of inputting the multi-source marine environmental data matrix corresponding to the second multi-source marine environmental data set into the environmental data analysis model, so that the environmental data analysis model outputs environmental anomaly change data of the target monitoring sea area based on the multi-source marine environmental data matrix corresponding to the second multi-source marine environmental data set, includes: The multi-source marine environmental data matrix is ​​input into the environmental data analysis model so that the environmental data analysis model can determine the initial abnormal water temperature change data and the initial abnormal water temperature change area based on the sea surface temperature remote sensing data and the measured water temperature data of the sea area calibration point in the multi-source marine environmental data matrix, and determine the initial abnormal phytoplankton concentration change data and the initial abnormal phytoplankton concentration change area based on the phytoplankton concentration remote sensing data in the multi-source marine environmental data matrix. The initial water temperature anomaly data, initial water temperature anomaly data, initial phytoplankton concentration anomaly data, and initial phytoplankton concentration anomaly data are corrected using the comprehensive environmental indicators, numerical errors, location errors, and correlation coefficients in the multi-source marine environmental data matrix. The output data on water temperature anomaly data, phytoplankton concentration anomaly data, and the target monitoring sea area are then output.

9. A marine fisheries environmental early warning method based on multi-source data fusion according to claim 8, characterized in that, The method of providing early warning and adjusting mariculture strategies for the target monitoring sea area based on the abnormal environmental change data includes: Based on the data of abnormal water temperature changes, the locations of abnormal water temperature changes are determined. Based on the locations of abnormal water temperature changes and the sea areas of abnormal water temperature changes, the sub-regions of abnormal water temperature changes corresponding to each location are determined. The sub-regions of abnormal water temperature changes are marked as areas of risk of sudden water temperature changes. The remaining areas of the sea areas of abnormal water temperature changes, except for the areas of risk of sudden water temperature changes, are marked as areas of potential risk of sudden water temperature changes. Based on the data on abnormal changes in phytoplankton concentration, the locations of abnormal changes in phytoplankton concentration are determined. Based on the locations of abnormal changes in phytoplankton concentration and the sea areas where abnormal changes in phytoplankton concentration occur, sub-regions of abnormal changes in phytoplankton concentration are determined for each location of abnormal changes in phytoplankton concentration. The sub-regions of abnormal changes in phytoplankton concentration are marked as areas of risk of sudden changes in phytoplankton concentration. The remaining areas of the sea areas where abnormal changes in phytoplankton concentration occur, except for the areas of risk of sudden changes in phytoplankton concentration, are marked as areas of potential risk of sudden changes in phytoplankton concentration. The overlapping areas are determined based on the risk areas of sudden changes in water temperature and the risk areas of sudden changes in phytoplankton concentration, and the overlapping areas are marked as high-risk areas. The remaining target monitoring areas, except for the sea areas with abnormal changes in water temperature and the sea areas with abnormal changes in phytoplankton concentration, are marked as risk-free areas. Based on high-risk areas, areas at risk of sudden changes in water temperature, areas at risk of sudden changes in phytoplankton concentration, areas at risk of potential sudden changes in water temperature, and areas at risk of potential sudden changes in phytoplankton concentration, tiered early warning information is generated. The scale of mariculture in high-risk areas, areas at risk of sudden changes in water temperature, and areas at risk of sudden changes in phytoplankton concentration is reduced. The mariculture yield in areas at risk of potential sudden changes in water temperature and areas at risk of potential sudden changes in phytoplankton concentration is monitored in real time, and the scale of mariculture in risk-free areas is increased.

10. A marine fisheries environmental early warning system based on multi-source data fusion, characterized in that, The system includes: a multi-source data acquisition module, a spatiotemporal data synchronization module, an anomaly data analysis module, and an early warning module; The multi-source data acquisition module is used to acquire marine remote sensing data of the target monitoring sea area in real time through satellite remote sensing, and to acquire measured water temperature data of each calibration point in the target monitoring sea area in real time through the buoy observation platform; wherein, the marine remote sensing data includes: sea surface water temperature remote sensing data and phytoplankton concentration remote sensing data; The spatiotemporal data synchronization module is used to perform grid division preprocessing on the marine remote sensing data and the measured water temperature data of each sea area calibration point, so that the marine remote sensing data and the measured water temperature data of each sea area calibration point are in the same grid division accuracy, generating a first multi-source marine environment data set; and to perform spatial dimension division standardization processing and temporal dimension division standardization processing on the first multi-source marine environment data set to obtain a second multi-source marine environment data set with consistent temporal and spatial dimensions. The abnormal data analysis module is used to input the second multi-source marine environmental data set into the environmental data analysis model, so that the environmental data analysis model outputs environmental anomaly change data of the target monitoring sea area based on the second multi-source marine environmental data set; wherein, the environmental anomaly change data includes: abnormal water temperature change data and abnormal phytoplankton concentration change data, sea areas with abnormal water temperature change and sea areas with abnormal phytoplankton concentration change. The early warning module is used to provide early warnings and adjustments to the marine aquaculture strategy in the target monitoring sea area based on the abnormal environmental change data.

Images