Machine learning-based comprehensive culture and proliferation sea area space division method

Abstract

The invention discloses a comprehensive augmented culture sea area space division method based on machine learning, and relates to the technical field of comprehensive augmented culture sea area space division. The method comprises the following steps: for each augmented culture species, quantifying the influence weight of a plurality of environment variables on the distribution of the augmented culture species, and eliminating redundant environment variables in the plurality of environment variables according to the influence weight to obtain a target environment variable; inputting the target environment variable into the maximum entropy model, and respectively predicting suitable growth area distribution raster data of the proliferation and culture species; calculating a landscape pattern index of the suitable growth area, positioning a multi-species ecological niche overlapping area, and generating a conflict thermodynamic diagram based on the landscape pattern index and the multi-species ecological niche overlapping area; and in a geographic information system platform, stacking the suitable growth area distribution raster data of all the proliferation and culture species according to the weight, and determining a preferential culture area, a limited culture area and a prohibited culture area according to the stacking result and the conflict thermodynamic diagram. The method can improve the utilization rate of sea area resources.

Description

Technical Field

[0001] This invention relates to the fields of marine resource management and aquaculture, and in particular to a method for integrated aquaculture marine spatial zoning based on machine learning. Background Technology

[0002] With the development of the marine economy, the rational planning and utilization of aquaculture areas has become an important aspect of marine resource development.

[0003] Traditional aquaculture area planning relies heavily on experience and lacks scientific quantitative analysis methods, making it difficult to fully consider the complexity and diversity of the marine ecological environment. Furthermore, current technologies for zoning aquaculture areas mostly focus on single species or single environmental factors, resulting in low utilization rates of marine resources. Summary of the Invention

[0004] Therefore, it is necessary to provide a machine learning-based method for the spatial zoning of aquaculture areas to address the aforementioned technical problems.

[0005] The present invention adopts the following technical solution: This invention provides a comprehensive spatial zoning method for aquaculture areas based on machine learning, comprising: Acquire multi-source data for the marine spatial area to be planned; multi-source data includes marine environmental element data, marine ecological data, and aquaculture species data; For each aquaculture species within the planned marine spatial area, the influence weights of multiple environmental variables in the marine environmental data on the distribution of aquaculture species are quantified using a random forest model. Redundant environmental variables are then eliminated based on their influence weights to obtain the target environmental variables. These environmental variables include temperature, salinity, and current velocity. The target environmental variable is input into the maximum entropy model to predict the raster data of suitable habitat distribution for aquaculture species; the value of each raster in the suitable habitat distribution raster data represents the suitability probability of the aquaculture species within the raster. Based on marine ecological data, the landscape pattern index of suitable habitats is calculated and the overlapping areas of multiple species niches are located. A conflict heat map is generated based on the landscape pattern index and the overlapping areas of multiple species niches. The landscape pattern index includes patch density, assemblage, and separation. The suitable habitat distribution raster data of all aquaculture species are superimposed according to weights, and priority aquaculture areas, restricted aquaculture areas, and prohibited aquaculture areas are determined based on the superposition results and conflict heat map.

[0006] Preferably, priority aquaculture zones, restricted aquaculture zones, and prohibited aquaculture zones are determined based on the overlay results and conflict heatmaps, specifically including: When the overlay result corresponding to the grid is greater than or equal to the first preset threshold and the grid is marked as having no ecological conflict at the corresponding position in the conflict heatmap, the grid will be marked as a priority aquaculture area. When the overlay result corresponding to the grid is less than the first preset threshold and greater than the second preset threshold, and the grid is marked as requiring manual intervention at the corresponding position in the conflict heatmap, the grid will be marked as a restricted aquaculture area. When the overlay result corresponding to the grid is less than the second preset threshold and the grid is marked as severely conflicted at the corresponding position in the conflict heatmap, the area corresponding to the overlay result will be marked as a prohibited breeding area.

[0007] Preferably, the target environmental variables are input into the maximum entropy model to predict the raster data of suitable habitat distribution for aquaculture species, specifically including: The constraints on the distribution of aquaculture species are determined based on the characteristics of the target environmental variables at the occurrence points of aquaculture species; the characteristics of the target environmental variables include the spatial distribution characteristics and numerical range characteristics of the target environmental variables; Estimate the distribution of MaxEnt under constraints, and determine the raster data of suitable habitat distribution for aquaculture species based on the distribution.

[0008] Preferably, the method further includes: After determining the priority breeding areas, restricted breeding areas, and prohibited breeding areas, a dynamic zoning map is generated; Based on the raster data of suitable habitat distribution for all aquaculture species, the area of ​​suitable habitat is marked on the dynamic zoning map; Based on the conflict heatmap, the locations of conflict areas are marked on the dynamic zoning map; Mark the dominant environmental factors on the dynamic zoning map.

[0009] Preferably, the dominant environmental factors are calculated using a niche model to determine the niche overlap among multiple aquaculture species, and then identified using a cut-off method.

[0010] This invention provides a machine learning-based integrated aquaculture marine spatial zoning device, comprising: The acquisition module is used to acquire multi-source data of the marine spatial area to be planned; the multi-source data includes marine environmental element data, marine ecological data, and aquaculture species data. The filtering module is used to quantify the influence weights of multiple environmental variables in the marine environmental data on the distribution of each aquaculture species in the planned marine spatial area using a random forest model, and to remove redundant environmental variables based on the influence weights to obtain the target environmental variables; the environmental variables include temperature, salinity and current velocity. The prediction module is used to input the target environmental variables into the maximum entropy model to predict the raster data of suitable habitat distribution for aquaculture species; the value of each raster in the suitable habitat distribution raster data represents the suitability probability of the aquaculture species within the raster. The generation module is used to calculate the landscape pattern index of suitable habitats and locate the overlapping areas of multiple species niches based on marine ecological data, and generate a conflict heat map based on the landscape pattern index and the overlapping areas of multiple species niches; the landscape pattern index includes patch density, assemblage and separation. The segmentation module is used to overlay the raster data of suitable habitat distribution of all aquaculture species according to weights, and determine priority aquaculture areas, restricted aquaculture areas, and prohibited aquaculture areas based on the overlay results and conflict heatmap.

[0011] The present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-mentioned method for integrated aquaculture marine spatial zoning based on machine learning.

[0012] This invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the aforementioned machine-based... A comprehensive method for spatial zoning of aquaculture sea areas based on instrument learning.

[0013] The above-mentioned at least one technical solution adopted in this invention can achieve the following beneficial effects: Multi-source data of the planned marine area is acquired. Multi-source data eliminates data silos and avoids planning biases caused by missing information. For each aquaculture species within the planned marine area, the influence weights of multiple environmental variables on the species distribution are quantified using a random forest model. Redundant environmental variables are then removed based on their influence weights to obtain the target environmental variable. Quantifying the influence weights of environmental variables and removing redundant variables reduces model computational complexity while focusing on key ecological factors, accurately reflecting the species-environment relationship. The target environmental variable is then input into a maximum entropy model to predict the aquaculture species distribution. Raster data on the suitable habitat distribution of aquaculture species, generated based on target environmental variables, can identify potentially highly suitable marine areas for species and uncover hidden high-quality aquaculture spaces. It calculates landscape pattern indices for suitable habitats and locates overlapping ecological niches of multiple species. Based on these indices and overlapping areas, a conflict heatmap is generated, visually representing areas of resource competition and providing a scientific basis for zoning management. The raster data on the suitable habitat distribution of all aquaculture species are weighted and overlaid. Based on the overlay results and the conflict heatmap, priority aquaculture areas, restricted aquaculture areas, and prohibited aquaculture areas are determined. This method improves the utilization rate of marine resources. Attached Figure Description

[0014] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this invention, illustrate exemplary embodiments of the invention and are used to explain the invention, but do not constitute an undue limitation of the invention. In the drawings:

[0015] Figure 1 A schematic diagram of a machine learning-based integrated aquaculture marine spatial zoning method provided by the present invention; Figure 2 The flowchart of marine spatial zoning provided by this invention; Figure 3 A schematic diagram of a machine learning-based integrated aquaculture marine spatial zoning device provided by the present invention; Figure 4 A schematic diagram of a computer device for implementing a machine learning-based integrated aquaculture marine spatial zoning method provided by the present invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0017] Devices such as desktop computers, servers, and laptops are capable of executing the solutions of this invention. For ease of explanation, the following description will focus on servers as the executing entity.

[0018] The technical solutions provided by the various embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0019] Figure 1 This is a schematic diagram of a machine learning-based integrated aquaculture marine spatial zoning method according to the present invention, which specifically includes the following steps: S101: Obtain multi-source data for the marine spatial area to be planned; multi-source data includes marine environmental element data, marine ecological data, and aquaculture species data.

[0020] Data from multiple sources, including marine environmental elements, marine ecology, and aquaculture species, is collected and integrated into a unified Geographic Information System (GIS) platform to complete spatial coordinate registration and multi-source data fusion. Marine environmental elements include water depth, current velocity, water quality, and seabed sediment; marine ecology data includes plankton, nematodes, and macrobenthic organisms.

[0021] S102: For each aquaculture species within the planned marine spatial area, the influence weights of multiple environmental variables in the marine environmental element data on the distribution of aquaculture species are quantified using a random forest model, and redundant environmental variables are eliminated based on the influence weights to obtain the target environmental variables; the environmental variables include temperature, salinity, and current velocity.

[0022] The input data for the random forest model is marine environmental data, and the output is the weights of the influence of multiple environmental variables on the distribution of aquaculture species. The influence weights are sorted in descending order, and redundant environmental variables with lower rankings are removed to obtain the target environmental variables. The number of redundant environmental variables is determined based on specific engineering practices.

[0023] The Random Forest (RF) model was used to analyze the influence weights of environmental variables (such as temperature, salinity, and flow velocity) on species distribution, and redundant variables (such as factors with a correlation of <5% with species distribution) were removed. The selected variables were then normalized to eliminate dimensional differences.

[0024] The Random Forest model consists of three stages: First, constructing independent classification trees; then, evaluating the model based on statistical error (scoring); finally, the final result is determined by the classification tree with the highest score and its criteria.

[0025] S103: Input the target environmental variable into the maximum entropy model to predict the raster data of suitable habitat distribution for aquaculture species; the value of each raster in the suitable habitat distribution raster data represents the suitability probability of the aquaculture species within the raster.

[0026] In an exemplary embodiment, the target environmental variable is input into the maximum entropy model to predict the raster data of the suitable habitat distribution of the aquaculture species. Specifically, this includes: determining the constraints on the distribution of the aquaculture species based on the characteristics of the target environmental variable at the occurrence points of the aquaculture species; the characteristics of the target environmental variable include the spatial distribution characteristics and numerical range characteristics of the target environmental variable; estimating the distribution of MaxEnt under the constraints, and determining the raster data of the suitable habitat distribution of the aquaculture species based on the distribution.

[0027] This invention is based on the spatial distribution and numerical range characteristics of the target environmental variable. For example, water temperature has different values ​​under different geographical locations, seasons, and water depths. The range, trend, and relationship with other environmental variables of these values ​​constitute its spatial distribution characteristics. Simultaneously, a cultured species may have a suitable temperature range for water; for example, a certain fish species thrives in water temperatures between 15°C and 25°C. This range constitutes the numerical range characteristic of the water temperature variable. These characteristics collectively determine the survival and distribution conditions of the cultured species.

[0028] The maximum entropy model (MaxEnt) is used as input to predict raster data of species' suitable habitat distribution (output is a probability map with values ​​ranging from 0 to 1). Niche overlap (Schoener's D index) among multiple species is calculated using niche models (ENMs), and dominant environmental factors are identified using the Jackknife method.

[0029] S104: Based on marine ecological data, calculate the landscape pattern index of suitable habitats and locate the overlapping areas of multiple species niches, and generate a conflict heat map based on the landscape pattern index and the overlapping areas of multiple species niches; the landscape pattern index includes patch density, integration and separation.

[0030] S105: Overlay the raster data of suitable habitat distribution of all aquaculture species according to weights, and determine priority aquaculture areas, restricted aquaculture areas and prohibited aquaculture areas based on the overlay results and conflict heat map.

[0031] In an exemplary embodiment, prioritization areas, restricted areas, and prohibited areas are determined based on the overlay results and the conflict heatmap. Specifically, this includes: when the overlay result corresponding to a grid is greater than or equal to a first preset threshold and the grid is marked as having no ecological conflict at the corresponding position on the conflict heatmap, the grid is marked as a priority area; when the overlay result corresponding to a grid is less than the first preset threshold but greater than a second preset threshold and the grid is marked as requiring human intervention at the corresponding position on the conflict heatmap, the grid is marked as a restricted area; when the overlay result corresponding to a grid is less than the second preset threshold and the grid is marked as having severe conflict at the corresponding position on the conflict heatmap, the area corresponding to the overlay result is marked as a prohibited area.

[0032] Specifically, the first and second preset thresholds are set according to specific engineering practices. In the GIS platform, the raster data of suitable habitats for each species are superimposed according to weights. Based on the superposition results, three levels of breeding zones are divided: priority breeding zone: suitable probability > 0.8, no ecological conflict; restricted breeding zone: suitable probability 0.5~0.8, requiring artificial intervention (such as bottom sediment improvement); prohibited breeding zone: suitable probability < 0.5 or severe ecological conflict.

[0033] In an exemplary embodiment, the method further includes: generating a dynamic zoning map after determining priority breeding areas, restricted breeding areas, and prohibited breeding areas; marking the area of ​​suitable breeding areas on the dynamic zoning map based on raster data of suitable breeding areas for all aquaculture species; marking the location of conflict areas on the dynamic zoning map based on a conflict heat map; and marking the dominant environmental factors on the dynamic zoning map.

[0034] In one exemplary embodiment, the dominant environmental factor calculates the niche overlap among multiple aquaculture species using a niche model and identifies it using a cut-off method.

[0035] Generate dynamic zoning maps (updated seasonally or according to environmental changes), marking suitable habitat areas, the contribution rate of dominant factors, and the location of conflict zones. Output the mean environmental parameters, species compatibility index, and economic benefit forecasts for each zone, enabling the development of refined aquaculture plans.

[0036] The natural breakpoint method was used to determine the boundary value of the suitable habitat and to count the area of ​​the suitable habitat.

[0037] The landscape pattern index of suitable habitat in the spatial zoning of aquaculture sea areas was calculated using the landscape pattern analysis software Fragstats 4.2. The calculation mainly consists of two indicators: basic indicators and suitable habitat aggregation and dispersion indicators.

[0038] The secondary indicators were weighted and scored using a subjective-objective synthesis method based on DS evidence theory (multi-level analysis and entropy method), and the levels were divided in ArcGIS using the natural breakpoint classification method based on the score distribution. All spatial grids were 10 m × 10 m.

[0039] In one exemplary embodiment, a three-level evaluation system is constructed using the analysis of multiple levels of data (AHP). The weights of each indicator are dynamically adjusted by combining expert scores with real-time monitoring data (e.g., water quality accounts for 30% and sediment type accounts for 20%).

[0040] In one exemplary embodiment, the present invention provides as follows Figure 2 The flowchart of marine spatial zoning shown is as follows: Figure 2 As shown, it includes the following steps: Multi-factor data collection and analysis: Collect multi-source data including marine environmental elements (such as water depth, current velocity, water quality, and bottom sediment), marine ecological data (such as plankton, swimming animals, and macrobenthic organisms), and distribution data of aquaculture species, and integrate them into a unified geographic information system (GIS) platform.

[0041] Establishment of Spatial Zoning Evaluation System: A spatial zoning evaluation system was constructed using a multi-level analysis method, including the target layer, criterion layer, and scheme layer. Water depth, flow velocity, bottom sediment type, water quality, and sediment quality were selected as evaluation indicators and assigned corresponding weights.

[0042] Feature selection and engineering: Feature selection: Random forests are used to screen out the environmental variables that have the greatest impact on species distribution, reducing model complexity and the risk of overfitting.

[0043] Habitat suitability assessment: Combining multi-scale environmental data, the maximum entropy model (MaxEnt) was used to predict the potential suitable habitats for each aquaculture species in the marine spatial zoning. Ecological niche models (ENMs) were used to analyze the niche overlap of key species, and the dominant environmental factors were screened using a cut-off method.

[0044] Optimization of aquaculture spatial zoning: Landscape pattern indices (patch density, separation, integration, etc.) of suitable habitats are calculated using landscape pattern analysis software to assess the rationality of aquaculture spatial zoning and make optimization adjustments.

[0045] Multi-factor spatial overlay analysis: Using a GIS platform and combined with the Analytic Hierarchy Process (AHP), the raster data of the suitable habitats of each aquaculture species are reclassified and overlaid to analyze the niche relationships among the species and determine the final zoning results.

[0046] The core inventive point of the method provided by this invention is: 1. Based on the principles of machine learning models, this study utilizes GIS technology, along with niche models and species distribution models, to conduct the first comprehensive aquaculture spatial zoning technology.

[0047] 2. Multi-factor comprehensive analysis: This invention comprehensively considers multiple factors such as marine environmental elements, ecological data, and the distribution of aquaculture species, overcoming the limitations of traditional single-factor analysis methods, and can more comprehensively reflect the ecological characteristics of aquaculture areas.

[0048] 3. Multi-model synergy mechanism: For the first time, the MaxEnt model, the niche model and landscape pattern analysis are combined to achieve quantitative prediction of the distribution and ecological suitability of aquaculture species, providing strong support for scientific zoning.

[0049] 4. Integration of GIS technology and models: By combining GIS technology with ecological niche models and species distribution models, the visualization and refined management of aquaculture marine spatial zoning have been achieved.

[0050] 5. Dynamic weight adjustment: Through the analytic hierarchy process and real-time data feedback, the weights of environmental factors are dynamically optimized to adapt to the dynamic needs of marine management; 6. Intelligent identification of conflict areas: Based on GIS overlay analysis and ecological niche overlap quantification, it accurately locates multi-species competition areas, providing data support for conflict mitigation.

[0051] When applying the machine learning-based integrated aquaculture marine spatial zoning method provided by this invention, it is not necessary to consider... Figure 1The steps shown are executed in sequence. The specific execution order of each step can be determined as needed, and this invention does not impose any restrictions on it.

[0052] The above describes a machine learning-based method for spatial zoning of integrated aquaculture areas, provided by one or more embodiments of the present invention. Based on the same idea, the present invention also provides a corresponding machine learning-based device for spatial zoning of integrated aquaculture areas, such as... Figure 3 As shown.

[0053] Figure 3 A schematic diagram of a machine learning-based integrated aquaculture marine spatial zoning device provided by the present invention includes: The acquisition module 301 is used to acquire multi-source data of the sea area to be planned; the multi-source data includes marine environmental element data, marine ecological data and aquaculture species data.

[0054] The screening module 302 is used to quantify the influence weights of multiple environmental variables in the marine environmental element data on the distribution of each aquaculture species in the planned marine spatial area using a random forest model, and to remove redundant environmental variables according to the influence weights to obtain the target environmental variables; the environmental variables include temperature, salinity and current velocity.

[0055] The prediction module 303 is used to input the target environmental variable into the maximum entropy model to predict the suitable habitat distribution grid data of the aquaculture species; the value of each grid in the suitable habitat distribution grid data represents the suitability probability of the aquaculture species within the grid.

[0056] The generation module 304 is used to calculate the landscape pattern index of the suitable habitat area and locate the overlapping area of ​​multiple species niches based on the marine ecological data, and generate a conflict heat map based on the landscape pattern index and the overlapping area of ​​multiple species niches; the landscape pattern index includes patch density, integration degree and separation degree.

[0057] The segmentation module 305 is used to overlay the raster data of suitable habitat distribution of all aquaculture species according to weights, and to determine priority aquaculture areas, restricted aquaculture areas and prohibited aquaculture areas based on the overlay results and the conflict heat map.

[0058] Specific limitations regarding the machine learning-based integrated aquaculture marine spatial zoning device can be found in the above-mentioned limitations on the machine learning-based integrated aquaculture marine spatial zoning method, and will not be repeated here. Each module in the aforementioned machine learning-based integrated aquaculture marine spatial zoning device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0059] The present invention also provides a computer-readable storage medium storing a computer program that can be used to execute the above-described... Figure 1 The provided method is a machine learning-based approach for the spatial zoning of aquaculture areas.

[0060] The present invention also provides Figure 4 The schematic diagram of the computer device shown is as follows: Figure 4 As shown, at the hardware level, this computer device includes a processor, internal bus, network interface, memory, and non-volatile memory, and may also include other hardware required for business operations. The processor reads the corresponding computer program from the non-volatile memory into memory and then executes it to achieve the above. Figure 1 The provided method is a machine learning-based approach for the spatial zoning of aquaculture areas.

[0061] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0062] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this invention.

Claims

1. A method for spatial zoning of an integrated mariculture sea area based on machine learning, characterized in that, The method comprises the following steps: obtaining multi-source data of a spatial region of a sea area to be planned; the multi-source data comprises marine environmental element data, marine ecological data and aquaculture species data; for each aquaculture species in the spatial region of the sea area to be planned, quantifying the influence weight of multiple environmental variables in the marine environmental element data on the distribution of the aquaculture species by a random forest model, and removing redundant environmental variables according to the influence weight to obtain target environmental variables; the environmental variables comprise temperature, salinity and flow rate; inputting the target environmental variables into a maximum entropy model to predict the distribution grid data of the suitable area of the aquaculture species; the value of each grid in the distribution grid data of the suitable area represents the probability of the aquaculture species in the grid; based on the marine ecological data, calculating the landscape pattern index of the suitable area and positioning the niche overlap area of multiple species, and generating a conflict heat map based on the landscape pattern index and the niche overlap area of multiple species; the landscape pattern index comprises patch density, cohesion and separation; superimposing the distribution grid data of the suitable area of all aquaculture species according to weights, and determining a priority aquaculture area, a restricted aquaculture area and a prohibited aquaculture area according to the superimposed result and the conflict heat map.

2. The method of claim 1, wherein, The determination of the priority aquaculture area, the restricted aquaculture area and the prohibited aquaculture area according to the superimposed result and the conflict heat map comprises the following steps: when the superimposed result of a grid is greater than or equal to a first preset threshold value and the grid is marked as no ecological conflict at the corresponding position of the conflict heat map, the grid is marked as the priority aquaculture area; when the superimposed result of a grid is less than the first preset threshold value and greater than a second preset threshold value and the grid is marked as needing manual intervention at the corresponding position of the conflict heat map, the grid is marked as the restricted aquaculture area; when the superimposed result of a grid is less than the second preset threshold value and the grid is marked as a serious conflict at the corresponding position of the conflict heat map, the superimposed result is marked as the prohibited aquaculture area.

3. The method of claim 1, wherein, The inputting of the target environmental variables into the maximum entropy model to predict the distribution grid data of the suitable area of the aquaculture species comprises the following steps: determining the constraint condition of the distribution of the aquaculture species according to the characteristics of the target environmental variables of the occurrence points of the aquaculture species; the characteristics of the target environmental variables comprise spatial distribution characteristics and numerical range characteristics of the target environmental variables; estimating the distribution of MaxEnt under the constraint condition, and determining the distribution grid data of the suitable area of the aquaculture species according to the distribution.

4. The method of claim 1, wherein, The method further comprises the following steps: generating a dynamic zoning map after the priority aquaculture area, the restricted aquaculture area and the prohibited aquaculture area are determined; based on the distribution grid data of the suitable area of all aquaculture species, marking the area of the suitable area on the dynamic zoning map; based on the conflict heat map, marking the position of the conflict area on the dynamic zoning map; marking the dominant environmental factor on the dynamic zoning map.

5. The method of claim 2, wherein, The dominant environmental factor is calculated by a niche model to calculate the niche overlap degree among multiple aquaculture species, and is identified by a knife-cut method. 6.A device for spatial zoning of an integrated mariculture sea area based on machine learning, characterized in that, The method comprises the following steps: an obtaining module is configured to obtain multi-source data of a spatial region of a sea area to be planned; the multi-source data comprises marine environmental element data, marine ecological data and aquaculture species data; The screening module is configured to, for each mariculture species in a spatial region of a sea area to be planned, quantify the influence weight of a plurality of environmental variables in the marine environmental factor data on the distribution of the mariculture species by a random forest model, and remove redundant environmental variables according to the influence weight to obtain target environmental variables; the environmental variables include temperature, salinity and flow rate; The prediction module is configured to input the target environmental variables into a maximum entropy model to predict the distribution grid data of the suitable area of the mariculture species; the value of each grid in the distribution grid data of the suitable area represents the probability of the mariculture species being suitable for survival in the grid; The generation module is configured to calculate the landscape pattern index of the suitable area and locate the multi-species ecological niche overlap area based on the marine ecological data, and generate a conflict heat map based on the landscape pattern index and the multi-species ecological niche overlap area; the landscape pattern index includes patch density, cohesion and separation; The division module is configured to superimpose the distribution grid data of the suitable area of all mariculture species according to weights, and determine a priority breeding area, a restricted breeding area and a prohibited breeding area according to the superimposed result and the conflict heat map.

7. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, and the computer program is executed by the processor to implement the method of any one of claims 1-5.

8. A computer device, comprising: The computer program is stored in the memory and executable on the processor, and the processor implements the method of any one of claims 1-5 when executing the program.

Images