A method for evaluating the effectiveness of green tide harvesting of *Ulva prolifera* considering its drifting growth.

Abstract

This invention relates to the field of *Ulva prolifera* harvesting assessment technology, and particularly to a method for assessing the effectiveness of *Ulva prolifera* green tide harvesting considering its drifting growth. The method includes the following steps: acquiring temporal multispectral remote sensing images of the target sea area, extracting the *Ulva prolifera* distribution vector boundary and the spatiotemporal trajectory data of harvesting vessels; identifying the core area of ​​*Ulva prolifera* aggregation, generating a spatial distribution heat map of harvesting intensity; generating floating growth bias data based on random sampling analysis of *Ulva prolifera* growth characteristics and correlation with hydrological environmental characteristics; simulating the spatial growth evolution of *Ulva prolifera* using the floating growth bias data, calculating the harvesting efficiency coefficient by combining harvesting intensity and growth prediction, and finally scientifically evaluating the harvesting volume. This invention, by integrating temporal multispectral remote sensing, vessel trajectory, spatial growth prediction, and harvesting assessment, achieves precise dynamic monitoring of *Ulva prolifera* green tides and quantitative assessment of harvesting effectiveness, effectively improving the efficiency of floating *Ulva prolifera* harvesting.

Description

Technical Field

[0001] This invention relates to the field of evaluation technology for seaweed harvesting, and in particular to a method for evaluating the effectiveness of seaweed green tide harvesting that takes into account the drifting and growth of seaweed. Background Technology

[0002] Vessel salvage is the primary technical means for controlling green tides of *Ulva prolifera*, commonly including methods such as strafe net salvage and mechanized salvage, requiring a large number of salvage vessels and incurring significant salvage and management costs. As current control efforts shift towards precision and technological advancements, developing targeted methods for evaluating the effectiveness of *Ulva prolifera* drift and growth in green tide salvage is a crucial issue that urgently needs to be addressed to improve the efficiency of current green tide salvage and reduce salvage and transportation costs. Currently, the evaluation of green tide salvage effectiveness mainly relies on manual experience and on-site visual inspection, with single evaluation indicators and a lack of quantitative analysis, making it difficult to comprehensively reflect the salvage results. Subsequently, with the development of remote sensing and water quality monitoring technologies, the use of remote sensing imagery and water quality indicators for quantitative monitoring of *Ulva prolifera* coverage area and water nutrient status has improved the accuracy of salvage effectiveness assessment. In recent years, the combination of Geographic Information Systems (GIS), UAV aerial photography, and automated sampling technology has enabled large-scale, multi-temporal, and high-precision dynamic monitoring, providing strong data support for evaluating the effectiveness of green tide salvage. However, current technologies often rely on manual experience or simple clustering methods to determine the core areas of vegetation accumulation during harvesting, leading to large errors. Furthermore, harvesting data is often disconnected from vegetation ecological parameters, making it impossible to assess the true impact of the external environment on vegetation harvesting operations. Summary of the Invention

[0003] Therefore, it is necessary to provide a method for evaluating the effectiveness of green tide harvesting of Ulva prolifera that takes into account the drifting and growth of Ulva prolifera, in order to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a method for evaluating the effectiveness of *Ulva prolifera* (seaweed) green tide harvesting considering its drifting and growth is provided. The method includes the following steps:

[0005] Step S1: Acquire temporal multispectral remote sensing images of the target sea area; perform spectral analysis of the Ulva prolifera characteristics on the temporal multispectral remote sensing images to generate the Ulva prolifera distribution vector boundary; obtain the spatiotemporal trajectory data of the salvage vessel and the statistics of salvage operation volume based on the Ulva prolifera distribution vector boundary;

[0006] Step S2: Extract the visual morphological features of the distribution vector boundary of *Ulva prolifera* to identify the core area of ​​*Ulva prolifera* aggregation; conduct a radiation analysis of the salvage impact on the core area of ​​*Ulva prolifera* aggregation using the spatiotemporal trajectory data of salvage vessels, and calculate the spatial distribution heat map of salvage intensity for each time period based on the analysis results.

[0007] Step S3: Randomly sample the green tide of *Ulva prolifera* based on the spatial distribution heat map of salvage intensity to obtain *Ulva prolifera* sample data; analyze the growth characteristics of *Ulva prolifera* in the sample data and correlate the growth characteristics of *Ulva prolifera* with hydrological environmental characteristics to generate *Ulva prolifera* floating path data; use the *Ulva prolifera* floating path data to conduct spatial growth evolution of the *Ulva prolifera* aggregation core area to generate *Ulva prolifera* growth prediction data.

[0008] Step S4: Calculate the salvage efficiency coefficient based on the spatial distribution heat map of salvage intensity and the predicted growth data of Ulva prolifera, and evaluate the salvage effect of the statistical salvage operation volume based on the salvage efficiency coefficient, and generate a Ulva prolifera green tide salvage effect evaluation report.

[0009] This invention identifies the distribution boundaries of *Ulva prolifera* from temporal remote sensing imagery and combines this with salvage vessel trajectories and operational data to achieve full-process coverage from macroscopic monitoring and salvage behavior modeling to microscopic biological characteristic analysis and effect evaluation, significantly improving the scientific and systematic nature of green tide management. By performing spectral analysis of *Ulva prolifera* characteristics and extracting the visual morphology of the aggregation core area from remote sensing images, combined with growth characteristics and environmental features, a multi-scale, multi-source data fusion *Ulva prolifera* identification system is formed, improving the accuracy of *Ulva prolifera* identification and growth prediction. Radiation analysis of the impact range of *Ulva prolifera* aggregation areas is performed using salvage trajectory data, and the salvage efficiency coefficient is calculated using salvage intensity heatmaps and growth prediction data, thereby quantitatively evaluating the salvage workload and assisting in the scientific formulation and adjustment of salvage strategies. Spatial growth evolution modeling based on *Ulva prolifera* floating path data can effectively predict the spread and growth trend of *Ulva prolifera*, providing a basis for early deployment and precise management, and reducing the risk of green tide outbreaks. This invention utilizes a heat map of salvage intensity to guide random sampling locations, ensuring that the sampling area covers zones affected by different intensities, thereby improving the representativeness of the collected *Ulva prolifera* samples and the comprehensiveness of the detection results. This method integrates remote sensing imagery, vector boundaries, ship trajectories, salvage data, growth characteristics, and environmental features to establish a unified data flow and logical closed loop, contributing to the construction of an intelligent, data-driven marine green tide management system. Therefore, by fusing temporal multispectral remote sensing, morphological analysis, and ship trajectories, this invention achieves precise dynamic monitoring of *Ulva prolifera* green tides, core area identification, growth prediction, and quantitative evaluation of salvage effectiveness, improving monitoring accuracy and the scientific rigor of salvage efficiency assessment.

[0010] Preferably, step S1 includes the following steps:

[0011] Step S11: Acquire a set of remote sensing images of the target sea area;

[0012] Step S12: Reconstruct the time series of remote sensing images of the target sea area to generate time series multispectral remote sensing image data; extract the spectral features of the time series multispectral remote sensing image data to obtain the characteristic spectral data of Ulva prolifera.

[0013] Step S13: Extract the boundary from the characteristic spectral data of *Ulva prolifera* to generate boundary data of the *Ulva prolifera* distribution vector;

[0014] Step S14: Spatial overlay of the distribution vector boundary data of *Ulva prolifera* to generate mask data of suspected *Ulva prolifera* salvage areas; ship trajectory screening of the mask data of suspected *Ulva prolifera* salvage areas to generate spatiotemporal trajectory data of salvage vessels.

[0015] Step S15: Calculate the salvage operation volume based on the spatiotemporal trajectory data of the salvage vessel, and generate salvage operation volume statistics.

[0016] This invention reconstructs a time-series of remote sensing images to form temporal multispectral remote sensing image data, enabling the complete capture of the changing trends of the spectral characteristics of *Ulva prolifera*, avoiding the limitations of static images and effectively improving the accuracy and timeliness of *Ulva prolifera* identification. Extracting the *Ulva prolifera* boundaries and combining this with the spatial overlay analysis in step S14 accurately generates mask data for suspected *Ulva prolifera* harvesting areas, helping to pinpoint key areas for remediation and providing precise spatial reference for subsequent trajectory screening and operational assessment. The mask data guides the screening of salvage vessel trajectories, effectively eliminating irrelevant vessel behavior and ensuring that only vessel trajectory data related to green tide management is extracted, improving the relevance and reliability of subsequent salvage statistical analysis. Through the correspondence between trajectory data and area masks, the amount of salvage operations is statistically analyzed, ensuring that the salvage intensity analysis is based on real, spatially defined operational data, providing support for salvage efficiency evaluation and resource optimization scheduling.

[0017] Preferably, step S2 involves extracting the visual morphological features of the distribution vector boundary of *Ulva prolifera* to identify the core area of ​​*Ulva prolifera* aggregation, including:

[0018] The vector boundary data of the distribution vector boundary of Ulva prolifera is rasterized to generate a raster image of the Ulva prolifera boundary.

[0019] Binarize the Ulva prolifera boundary raster image and then refine the morphology of the binarized Ulva prolifera boundary raster image to generate a refined visual morphology initial image.

[0020] Noisy pixels are removed from the initial image data to refine the visual morphology, resulting in a cleaned visual morphology image.

[0021] Visual morphology pixel connectivity analysis is performed on purified visual morphology images to generate visual morphology connectivity structure data.

[0022] Extract key nodes from the visual morphology connected structure data, calculate the node density of the visual morphology node set data, and generate a visual morphology node density distribution map.

[0023] Density clustering is performed on the visual morphology node density distribution map to generate high-density visual morphology cluster center data;

[0024] By extracting stable clustering regions from the visual morphology node density distribution map using high-density visual morphology cluster center data, the core area of ​​Ulva prolifera aggregation can be obtained.

[0025] This invention transforms vector boundary data into raster images and performs binarization and thinning operations to accurately extract morphological visual lines from the distribution area of ​​*Ulva prolifera*, avoiding recognition errors caused by complex boundaries or irregular shapes, and enhancing the integrity and resolution of spatial structure representation. Through noise pixel removal and visual morphological connectivity analysis, discrete or pseudo-structured pixels are filtered out, ensuring clear structure and good connectivity in the visual morphological image. This provides a high-quality data foundation for subsequent node density analysis and spatial clustering, improving recognition robustness. The visual morphological node density distribution map reveals spatial aggregation trends. By calculating and extracting high-density cluster centers, the core areas of concentrated *Ulva prolifera* distribution can be effectively identified, overcoming the sensitivity of traditional methods to region size and shape, and enhancing adaptability and generalization ability. The visual morphological structure, as the "central axis" of *Ulva prolifera* spatial distribution, can truly reflect its spatial aggregation path and core evolution direction, providing a clear structural foundation for subsequent salvage intensity analysis and aggregation trend evolution modeling, improving the efficiency of dynamic monitoring and control.

[0026] Preferably, the extraction of stable clustering regions from the visual morphology node density distribution map using high-density visual morphology clustering center data includes:

[0027] A region is identified as having an abnormal visual morphological node density when any of the following conditions are met, and abnormal visual morphological node density data is obtained: the visual morphological node density per unit area exceeds 150 nodes / km² or is less than 30 nodes / km²; the average distance between visual morphological nodes is less than 5m or greater than 50m; and the standard deviation of the density distribution is higher than 25.

[0028] A region is identified as having unstable cluster center stability when all of the following conditions are met, and cluster center deviation data are obtained: the distance the cluster center moves is greater than 10m in five consecutive time-series analyses; the cluster profile coefficient is less than 0.5; and the coefficient of variation of node density within the corresponding region exceeds 0.3.

[0029] An unstable clustering region is identified and its data is obtained when the following conditions occur simultaneously: the number of visual morphological nodes decreases by more than 20% within 30 minutes; the average node density decreases to below 40 nodes / km²; and the local maximum density point moves more than 15m within the spatial range, and this change continues to occur within 60 minutes.

[0030] By extracting stable clustering regions from the visual morphological node density distribution map using abnormal node density data, cluster center deviation data, and unstable clustering region data, the core clustering area of ​​*Ulva prolifera* can be obtained.

[0031] This invention effectively identifies visually morphological regions that are too dense, too sparse, or unevenly distributed by using a multi-indicator joint threshold judgment based on node density per unit area, average distance between nodes, and standard deviation of density distribution. This filters out invalid or misleading clustering trends, improving the accuracy and scientific validity of subsequent extraction results. By utilizing indicators such as cluster center movement distance, silhouette coefficient, and node density variation coefficient, it achieves dynamic monitoring and evaluation of the changing trends of core cluster positions, enabling timely capture of unstable fluctuations or drifting behavior in *Ulva prolifera* clusters and enhancing the ability to model complex dynamic processes. By setting joint judgment conditions for a rapid decrease in the number of visually morphological nodes, a significant weakening of density, and continuous drift of local maximum density points, it can accurately identify temporary or declining clustering regions that occur during spatiotemporal evolution, ensuring the stability and practical value of the core clustering area identification results. By comprehensively applying density anomaly data, clustering deviation data, and unstable region data to the extraction process of stable clustering areas, a robust spatial filtering mechanism is formed, ultimately identifying the true core clustering area of ​​*Ulva prolifera* that is structurally coherent, dynamically stable, and significantly dense. The identification results of the steady-state aggregation core area can directly support key applications such as subsequent salvage path optimization, salvage resource scheduling, and marine risk early warning, improve the targeting, timeliness, and resource utilization of salvage operations, and significantly enhance the intelligence and refinement of green tide governance.

[0032] Preferably, step S2 involves conducting a radiation analysis of the impact of salvage on the core area of ​​*Ulva prolifera* accumulation using the spatiotemporal trajectory data of the salvage vessel, including:

[0033] The spatiotemporal trajectory data of the salvaged vessel is resampled for time series of trajectory points to generate time-uniformed data of the vessel trajectory;

[0034] By calculating the positional thermodynamic intensity kernel density of the spatiotemporal trajectory data of the salvaged vessel using the time homogenization data of the vessel trajectory, a spatial distribution map of salvage intensity is obtained.

[0035] Based on a preset frequency threshold, the high-frequency salvage boundary is extracted from the spatial distribution map of salvage intensity to obtain the boundary data of the high-frequency salvage area;

[0036] Spatial overlap analysis was performed on the boundary data of the high-frequency dredging area and the core area of ​​seaweed aggregation to generate a core area dredging interaction map.

[0037] Calculate the salvage disturbance coefficient of the core area salvage interaction map, and perform spatial radiation impact modeling of the core area of ​​seaweed aggregation based on the salvage disturbance coefficient to generate spatial modeling data of salvage impact;

[0038] The temporal dimension of the spatial radiation model data affected by salvage was simulated to obtain the analysis results of the radiation impact of salvage.

[0039] This invention addresses the uneven temporal distribution of original trajectories by resampling the spatiotemporal trajectory data of salvage vessels over time, ensuring consistency and representativeness of subsequent kernel density calculations in the temporal dimension. This lays the foundation for high-precision modeling of spatial salvage intensity. By using kernel density estimation to model the homogenized trajectory data, a spatial distribution map of salvage intensity is generated, objectively reflecting frequent areas and spatial intensity gradients of actual salvage activities, enabling visualization and quantitative expression of the spatial characteristics of salvage behavior. High-frequency salvage area boundaries are extracted based on frequency thresholds, effectively identifying high-intensity salvage areas and providing a reference for stability analysis of aggregation areas and ecological disturbance assessment, improving the sensitivity and accuracy of salvage impact identification. Spatial overlay analysis of high-frequency salvage boundary data with the core area of ​​*Ulva prolifera* aggregation yields a real interactive area map, revealing the potential impact range of salvage activities on specific core areas and providing a data basis for disturbance assessment. The "salvage disturbance coefficient" index is introduced and modeled in conjunction with the interactive area map, innovatively transforming human interference behavior into a spatial radiation effect, scientifically quantifying the spatial response and diffusion path of *Ulva prolifera* aggregation areas to salvage behavior. By simulating the evolution of the spatial model of salvage impacts over time, the changing trends of disturbance effects can be dynamically tracked, providing crucial time-series support for adjusting salvage strategies, optimizing operational sequences, and assessing ecological restoration periods. The analysis results can serve as a theoretical basis for constructing a salvage disturbance classification mechanism (e.g., high-disturbance, medium-disturbance, and low-disturbance zones) and setting regional intervention intensity thresholds, thus promoting a shift in Ulva prolifera salvage from passive emergency response to proactive regulation, and from overall coverage to precise local control.

[0040] Preferably, random sampling of the *Ulva prolifera* green tide based on the spatial distribution heat map of salvage intensity includes:

[0041] Use manual or mechanical nets to collect surface seaweed samples at designated sampling points. Collect at least 1 square meter of seaweed at each sampling point and record the sampling time, location coordinates and harvesting environmental conditions to obtain seaweed sample data.

[0042] Calculate the sample area of ​​the *Ulva prolifera* sample data to obtain the initial biomass of the *Ulva prolifera* sample;

[0043] Based on the initial biomass of the Ulva prolifera samples, seawater temperature, light intensity, light duration, and nutrient concentration were monitored to obtain environmental monitoring data for Ulva prolifera.

[0044] This invention utilizes a thermal map of the spatial distribution of *Ulva prolifera* (a type of seaweed) ripples for random sampling, combined with surface *Ulva prolifera* samples collected manually or mechanically. This enables representative acquisition of *Ulva prolifera* biomass in areas with different ripple intensities, effectively improving the spatial distribution balance of samples and the accuracy of environmental response analysis. By precisely recording the sampling area, biomass, and related environmental factors at each sampling point, a coupling relationship between ripple intensity, *Ulva prolifera* biomass, and environmental factors is constructed. This provides a scientific basis for subsequent ripple prediction, ripple scheduling optimization, and ecological response assessment, offering advantages such as strong data representativeness, high environmental adaptability, and strong practicality.

[0045] Preferably, the growth characteristics of *Ulva prolifera* in step S3, when analyzing the *Ulva prolifera* sample data, include:

[0046] A comprehensive growth model for *Ulva prolifera* was constructed based on environmental monitoring data. The formula for the comprehensive growth model is shown below:

[0047]

[0048]

[0049]

[0050]

[0051] In the formula, For a moment The biomass of *Ulva prolifera* This is the initial biomass. The maximum growth coefficient, Temperature is a factor that affects the environment. The light intensity influencing factor, As a factor affecting sunshine duration, Factors affecting nutrient concentration For growth time, For the potential maximum growth rate, The total loss rate, The baseline proliferation rate at the reference temperature. For activation energy, The gas constant is... This is the actual water temperature. As respiratory activation energy, To maintain the consumption rate, To decompose the loss rate, This refers to the rate of loss due to disease or stress.

[0052] A comprehensive growth model for Ulva prolifera was used to predict the initial biomass of Ulva prolifera samples, generating predicted biomass data.

[0053] The growth rate of the initial biomass of *Ulva prolifera* samples was calculated based on the predicted biomass data, thus obtaining the growth characteristics of *Ulva prolifera*.

[0054] This invention constructs a comprehensive growth model for *Ulva prolifera* that integrates key environmental factors such as temperature, light intensity, sunshine duration, and nutrient concentration. This model can accurately simulate the dynamic changes in biomass of *Ulva prolifera* under different environmental conditions. By using this model to predict biomass and calculate growth rate of initial samples, it not only improves the quantitative expression of *Ulva prolifera* growth trends but also enables dynamic assessment of *Ulva prolifera* growth characteristics. This helps to scientifically reveal the laws governing the occurrence and development of green tides and provides high-precision data support and decision-making basis for predicting the spatiotemporal distribution of *Ulva prolifera*, disaster early warning, and salvage scheduling. It has the advantages of high modeling accuracy, strong environmental adaptability, and broad application prospects.

[0055] Preferably, the association of Ulva prolifera growth characteristics with environmental features and the spatial growth evolution of Ulva prolifera aggregation core areas through Ulva prolifera floating path data include:

[0056] Acquire wind field data and water flow velocity field data for the target area;

[0057] Vortex structure is extracted from wind field data in the target area to generate wind field vortex feature data.

[0058] A non-uniform flow field model is constructed based on the velocity field data of the water body flow field, and the local shear stress distribution of the velocity field data of the water body flow field is calculated based on the non-uniform flow field model to obtain the shear characteristic data of the flow channel.

[0059] Acquire wave observation data of the core area of ​​seaweed aggregation; analyze the wind-wave coupling intensity and main wave direction characteristics of the core area of ​​seaweed aggregation based on wind field vortex characteristic data, channel shear characteristic data and wave observation data, and conduct lateral migration analysis of seaweed in the core area of ​​seaweed aggregation to generate seaweed migration bias data.

[0060] Velocity gradient stripping analysis was performed on the flow channel shear characteristic data to generate data on the distribution stability of Ulva prolifera.

[0061] The floating potential energy surface is comprehensively quantified by combining the migration bias data and the distribution stability data of *Ulva prolifera* to generate *Ulva prolifera* floating path data. The formula for the comprehensive quantification of the floating potential energy surface is shown below:

[0062]

[0063] in, To synthesize the buoyancy potential energy distribution function, Representing the horizontal and vertical coordinates of space, respectively. For a moment, For wind field in Local wind speed value at a point For the water flow field in Local velocity value at a point The weighting coefficient for the influence of wind speed on the floating trend. The weighting coefficient for the influence of flow velocity on the floating trend;

[0064] A time-series frequency analysis of the floating path data of *Ulva prolifera* was performed to generate *Ulva prolifera* drift trend data; the drift trend data of *Ulva prolifera* was coupled and fitted with the spatial location of the core area of ​​*Ulva prolifera* aggregation to generate spatial mapping data of drift trend.

[0065] Spatiotemporal growth evolution modeling is performed on the drift trend spatial mapping data to generate Ulva prolifera growth prediction data.

[0066] This invention, through deep fusion analysis of wind vortex characteristics and water flow field shear characteristics, constructs a non-uniform flow field model and quantifies the buoyancy potential energy distribution during the floating process of *Ulva prolifera*, effectively revealing the dynamic behavior mechanism of *Ulva prolifera* aggregation and migration under the synergistic effect of wind and flow. Further, by using lateral transport analysis guided by wind and wave intensity and stability assessment under velocity gradient stripping, it achieves refined modeling and prediction of the *Ulva prolifera* floating path. The introduction of a comprehensive buoyancy potential energy surface quantification formula gives the model high resolution in both spatial distribution and temporal evolution, helping to accurately locate high-risk areas of green tides and their evolutionary trends. This provides strong theoretical support and technical assurance for early warning and scientific management of *Ulva prolifera* disasters, and has advantages such as high environmental coupling, high path extrapolation accuracy, and significant practical value.

[0067] Preferably, the spatiotemporal growth and evolution modeling processing of the drift trend spatial mapping data includes:

[0068] Multi-temporal sample segmentation is performed on the drift trend spatial mapping data to generate a time-series segmented growth sample set; the characteristics of the boundary of Ulva prolifera patch changing with time are extracted from the time-series segmented growth sample set to obtain time-driven boundary deformation feature data.

[0069] The growth boundary velocity is fitted to the time-driven boundary deformation feature data to generate spatial boundary growth rate data; the dominant growth direction of the spatial boundary growth rate data is identified.

[0070] Based on the dominant growth direction, the growth rate data of the spatial boundary is used to derive the growth rules of the rasterized region, generating raster growth transfer data; the set of raster growth transfer functions is subjected to spatiotemporal iterative simulation to generate dynamic process data of Ulva growth evolution;

[0071] The dynamic process data of Ulva prolifera growth and evolution were validated and processed to generate Ulva prolifera growth prediction data.

[0072] This invention, through multi-temporal sample segmentation and boundary deformation feature extraction, can reflect the expansion and contraction trends of *Ulva prolifera* patches at different time scales in real time, significantly improving the resolution and accuracy of growth prediction. Growth boundary velocity fitting and dominant direction identification can quantitatively characterize the dominant diffusion channels of *Ulva prolifera* within the core area, providing a reliable basis for subsequent derivation of growth rules in gridded areas. Through spatiotemporal iterative simulation of the gridded growth transfer function set, it achieves dynamic reproduction of the entire evolution process of *Ulva prolifera* communities, which can be used for short-term emergency salvage decisions and support long-term ecological evolution research. Comparison and verification of simulation results with actual monitoring data not only continuously corrects growth model parameters but also generates highly reliable *Ulva prolifera* growth prediction data, providing a scientific basis for salvage efficiency assessment and resource allocation. Through refined growth evolution prediction, it enables precise salvage and early warning of *Ulva prolifera* aggregation core areas, reducing manpower and vessel deployment costs, effectively minimizing environmental disturbance, and improving overall governance benefits.

[0073] Preferably, step S4 includes the following steps:

[0074] Step S41: Perform spatial correspondence analysis between the heat map of spatial distribution of salvage intensity and the predicted data of seaweed growth to generate salvage coverage consistency data; calculate the salvage effectiveness per unit area based on the salvage coverage consistency data to generate salvage efficiency coefficient data.

[0075] Step S42: Perform correlation efficiency analysis on the salvage efficiency coefficient data and the salvage operation volume statistics to generate salvage effect evaluation index data; use the salvage effect evaluation index data to evaluate and visualize the statistical salvage operation volume, and generate a seaweed green tide salvage effect evaluation report.

[0076] This invention, through spatial correspondence analysis of heat maps and growth prediction data, can accurately quantify the consistency of salvage coverage in each operational area, avoiding the omission of high-density clusters or the waste of resources in low-density areas. The salvage efficiency coefficient generated by the calculation of salvage effectiveness per unit area can reveal the differences in operational effectiveness of different plots under the same input, providing a basis for optimizing vessel deployment and manpower allocation. By correlating the efficiency coefficient with salvage operation volume statistics, operational efficiency can be evaluated from both overall and local perspectives, helping managers identify efficient and inefficient operational modes. The use of evaluation index data to graphically and map-wise display the operation volume makes salvage performance clear at a glance, facilitating the rapid detection of anomalies, adjustment of strategies, and intuitive reporting to decision-makers. The generated salvage effect evaluation report not only archives the quantitative results of each operation but also serves as a scientific basis for subsequent operational plan adjustments, performance evaluations, and long-term green tide governance decisions, thereby improving the overall governance level. Attached Figure Description

[0077] Figure 1A schematic diagram of the steps in an evaluation method for the effectiveness of green tide harvesting of Ulva prolifera that takes into account the drifting and growth of Ulva prolifera.

[0078] Figure 2 for Figure 1 A detailed flowchart illustrating the implementation steps of step S1.

[0079] Figure 3 for Figure 1 A detailed flowchart illustrating the implementation steps of step S4.

[0080] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0081] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0082] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0083] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0084] To achieve the above objectives, please refer to Figures 1 to 3 A method for evaluating the effectiveness of *Ulva prolifera* (seaweed) green tide harvesting considering its drifting and growth, the method comprising the following steps:

[0085] Step S1: Acquire temporal multispectral remote sensing images of the target sea area; perform spectral analysis of the Ulva prolifera characteristics on the temporal multispectral remote sensing images to generate the Ulva prolifera distribution vector boundary; obtain the spatiotemporal trajectory data of the salvage vessel and the statistics of salvage operation volume based on the Ulva prolifera distribution vector boundary;

[0086] Step S2: Extract the visual morphological features of the distribution vector boundary of *Ulva prolifera* to identify the core area of ​​*Ulva prolifera* aggregation; conduct a radiation analysis of the salvage impact on the core area of ​​*Ulva prolifera* aggregation using the spatiotemporal trajectory data of salvage vessels, and calculate the spatial distribution heat map of salvage intensity for each time period based on the analysis results.

[0087] Step S3: Randomly sample the green tide of *Ulva prolifera* based on the spatial distribution heat map of salvage intensity to obtain *Ulva prolifera* sample data; analyze the growth characteristics of *Ulva prolifera* in the sample data and correlate the growth characteristics of *Ulva prolifera* with hydrological environmental characteristics to generate *Ulva prolifera* floating path data; use the *Ulva prolifera* floating path data to conduct spatial growth evolution of the *Ulva prolifera* aggregation core area to generate *Ulva prolifera* growth prediction data.

[0088] Step S4: Calculate the salvage efficiency coefficient based on the spatial distribution heat map of salvage intensity and the predicted growth data of Ulva prolifera, and evaluate the salvage effect of the statistical salvage operation volume based on the salvage efficiency coefficient, and generate a Ulva prolifera green tide salvage effect evaluation report.

[0089] This invention identifies the distribution boundaries of *Ulva prolifera* from temporal remote sensing imagery and combines this with salvage vessel trajectories and operational data to achieve full-process coverage from macroscopic monitoring and salvage behavior modeling to microscopic biological characteristic analysis and effect evaluation, significantly improving the scientific and systematic nature of green tide management. By performing spectral analysis of *Ulva prolifera* characteristics and extracting the visual morphology of the aggregation core area from remote sensing images, and combining this with *Ulva prolifera* growth characteristics and environmental features, a multi-scale, multi-source data fusion *Ulva prolifera* identification system is formed, improving the accuracy of *Ulva prolifera* identification and growth prediction. Radiation analysis of the impact range of *Ulva prolifera* aggregation areas is performed using salvage trajectory data, and the salvage efficiency coefficient is calculated using salvage intensity heatmaps and growth prediction data, thereby quantitatively evaluating the salvage workload and assisting in the scientific formulation and adjustment of salvage strategies. Spatial growth evolution modeling based on *Ulva prolifera* floating path data can effectively predict the spread and growth trend of *Ulva prolifera*, providing a basis for early deployment and precise management, and reducing the risk of green tide outbreaks. This invention utilizes a heat map of salvage intensity to guide random sampling locations, ensuring that the sampling area covers zones affected by different intensities, thereby improving the representativeness of the collected *Ulva prolifera* samples and the comprehensiveness of the detection results. This method integrates remote sensing imagery, vector boundaries, ship trajectories, salvage data, growth characteristics, and environmental features to establish a unified data flow and logical closed loop, contributing to the construction of an intelligent, data-driven marine green tide management system. Therefore, by fusing temporal multispectral remote sensing, morphological analysis, and ship trajectories, this invention achieves precise dynamic monitoring of *Ulva prolifera* green tides, core area identification, growth prediction, and quantitative evaluation of salvage effectiveness, improving monitoring accuracy and the scientific rigor of salvage efficiency assessment.

[0090] In this embodiment of the invention, reference Figure 1The diagram shows a flowchart illustrating the steps of a method for evaluating the effectiveness of *Ulva prolifera* (seaweed) green tide harvesting that considers the drifting growth of *Ulva prolifera* according to the present invention. In this example, the method includes the following steps:

[0091] Step S1: Acquire temporal multispectral remote sensing images of the target sea area; perform spectral analysis of the Ulva prolifera characteristics on the temporal multispectral remote sensing images to generate the Ulva prolifera distribution vector boundary; obtain the spatiotemporal trajectory data of the salvage vessel and the statistics of salvage operation volume based on the Ulva prolifera distribution vector boundary;

[0092] Step S2: Extract the visual morphological features of the distribution vector boundary of *Ulva prolifera* to identify the core area of ​​*Ulva prolifera* aggregation; conduct a radiation analysis of the salvage impact on the core area of ​​*Ulva prolifera* aggregation using the spatiotemporal trajectory data of salvage vessels, and calculate the spatial distribution heat map of salvage intensity for each time period based on the analysis results.

[0093] Step S3: Randomly sample the green tide of *Ulva prolifera* based on the spatial distribution heat map of salvage intensity to obtain *Ulva prolifera* sample data; analyze the growth characteristics of *Ulva prolifera* in the sample data and correlate the growth characteristics of *Ulva prolifera* with hydrological environmental characteristics to generate *Ulva prolifera* floating path data; use the *Ulva prolifera* floating path data to conduct spatial growth evolution of the *Ulva prolifera* aggregation core area to generate *Ulva prolifera* growth prediction data.

[0094] Step S4: Calculate the salvage efficiency coefficient based on the spatial distribution heat map of salvage intensity and the predicted growth data of Ulva prolifera, and evaluate the salvage effect of the statistical salvage operation volume based on the salvage efficiency coefficient, and generate a Ulva prolifera green tide salvage effect evaluation report.

[0095] In this embodiment of the invention, Sentinel-2 or PlanetScope temporal multispectral remote sensing images are downloaded, and surface reflectance is obtained using the Sen2Cor atmospheric correction module. Then, feature indices such as GREI and NDVI705 are calculated. A binary mask for the distribution of *Ulva prolifera* is generated using SVM or random forest supervised classification. After morphological opening and closing operations are used for denoising, the boundaries are simplified using the Douglas-Peucker algorithm and converted into vector polygons. Furthermore, mooring tracks are cleaned from commercial or publicly available AIS data, and vessel operation segments (speeds between 0.5 and 5 knots) and their salvage quality information are extracted. Subsequently, the QGIS "centerline" plugin or ArcGIS Polyline is used. Skeleton extracts the visual morphology of the *Ulva prolifera* boundary, removes branches shorter than 100m, and defines the core area as the region with the highest node density (5%). Multi-level buffer zones of 500m, 1km, and 2km are established around the core area. Operational track density is calculated using ArcGIS kernel density estimation (1km bandwidth) to generate a 100m resolution heatmap. Then, based on historical *Ulva prolifera* harvesting operation data, remote sensing observation data, and UAV patrol data, a spatial distribution heatmap of *Ulva prolifera* harvesting intensity is constructed. This heatmap calculates the harvesting frequency and volume per unit area per unit time and uses spatial interpolation to form a heat distribution map, providing spatial guidance for subsequent sampling. Typical areas representing different harvesting intensity levels (high, medium, and low) are selected from the heatmap. A stratified random sampling method is used to set sampling points within each level area. *Ulva prolifera* samples are collected manually, by unmanned surface vessels, or by UAVs equipped with sampling devices. Environmental parameters such as spatial coordinates, water depth, water temperature, salinity, and transparency of the corresponding sampling points are recorded to generate a *Ulva prolifera* sample dataset. Using laboratory analysis and on-site sensor monitoring, a comprehensive analysis of *Ulva prolifera* samples and their aquatic environment was conducted to extract growth characteristic parameters, including but not limited to: nitrogen and phosphorus concentrations, dissolved oxygen content, pH value, chlorophyll a concentration, and light transmittance. Combined with an ecological model, response curves between growth factors and *Ulva prolifera* biomass were established to identify key factors promoting or inhibiting growth. High-frequency multibeam or side-scan sonar equipment was deployed in the sample collection area to acquire reflected signals from the *Ulva prolifera*-covered area at the bottom of the water body. Sonar inversion was performed through the following processes: waveform analysis and echo intensity calculation of the sonar data; a biomass inversion model was constructed based on the aforementioned growth characteristic parameters and measured biomass data; machine learning algorithms (such as Support Vector Regression (SVR) or Convolutional Neural Network (CNN)) were applied to train and predict the sonar data to estimate the *Ulva prolifera* bed biomass distribution in different areas. Combining the inverted biomass distribution data with water quality growth characteristic data, multidimensional data labels that stably represent different growth states, distribution densities, and biological characteristics were extracted to form *Ulva prolifera* floating path data.Finally, based on the measured population growth rate *r* and diffusion coefficient *D*, the spatial evolution of *Ulva prolifera* was simulated using a cellular automata or reaction-diffusion model with a 10m grid in a Python environment. The predicted distribution at each time step was output, and the efficiency coefficient *E* = *Mrecovery* / (*Acore area* × *T*) was calculated based on the actual recovery mass, core area, and operation duration. A comprehensive salvage effectiveness evaluation report, including remote sensing maps, heat maps, evolution prediction results, and efficiency analysis, was automatically generated using LaTeX or Word. Therefore, this invention, by integrating temporal multispectral remote sensing, morphological analysis, ship trajectory, and sonar inversion, achieves precise dynamic monitoring, core area identification, growth prediction, and quantitative evaluation of salvage effectiveness for *Ulva prolifera* green tides, significantly improving monitoring accuracy and the scientific rigor of salvage effectiveness evaluation.

[0096] As an example of the present invention, reference is made to Figure 2 As shown, in this example, step S1 includes:

[0097] Step S11: Acquire a set of remote sensing images of the target sea area;

[0098] Step S12: Reconstruct the time series of remote sensing images of the target sea area to generate time series multispectral remote sensing image data; extract the spectral features of the time series multispectral remote sensing image data to obtain the characteristic spectral data of Ulva prolifera.

[0099] Step S13: Extract the boundary from the characteristic spectral data of *Ulva prolifera* to generate boundary data of the *Ulva prolifera* distribution vector;

[0100] Step S14: Spatial overlay of the distribution vector boundary data of *Ulva prolifera* to generate mask data of suspected *Ulva prolifera* salvage areas; ship trajectory screening of the mask data of suspected *Ulva prolifera* salvage areas to generate spatiotemporal trajectory data of salvage vessels.

[0101] Step S15: Calculate the salvage operation volume based on the spatiotemporal trajectory data of the salvage vessel, and generate salvage operation volume statistics.

[0102] In this embodiment of the invention, multispectral remote sensing images of the target sea area taken within the required time period are downloaded from publicly available remote sensing data platforms (such as the Sentinel and Landsat official websites). Clear images with minimal cloud cover are selected to ensure the images cover the entire sea area. After downloading, the images are organized chronologically for later use. Multiple images taken at the same location at different times are grouped together chronologically to form a "time series." If images at certain times are obscured by clouds, they can be easily replaced with images from nearby times to ensure temporal continuity. The color changes of the sea surface in each image are observed to identify the typical colors of the Ulva prolifera (e.g., areas of deeper green). Color data for these areas is extracted manually or using simple tools (such as the color analysis function of Photoshop or QGIS). This color data represents the characteristic spectrum of the Ulva prolifera. Based on the color range extracted in the previous step, areas clearly representing the Ulva prolifera are circled on a map using drawing tools (such as QGIS or ArcGIS). These areas are converted into vector graphics (polygon boundaries) using the "boundary extraction" or "polygon generation" functions in the software. These vector boundary files are exported for subsequent spatial analysis. Overlay the polygonal boundary of the seaweed onto a map of the sea area to mark the distribution range of the seaweed; this is the suspected salvage area. Obtain GPS or AIS trajectory data of the salvage vessels (usually including time and location). Filter the data of vessels entering or passing through the suspected seaweed area on the map. Only retain these trajectories as the spatiotemporal trajectory data of the salvage vessels. Determine the time period for salvage operations based on simple indicators such as the time the vessels spend in the suspected salvage area and changes in their speed. Count the number of vessels and the total operation time within these time periods. Summarize the salvage operation time by date or region to obtain statistical data on the salvage volume. Display the data in tables or simple charts for subsequent analysis and management decisions.

[0103] Preferably, step S2 involves extracting the visual morphological features of the distribution vector boundary of *Ulva prolifera* to identify the core area of ​​*Ulva prolifera* aggregation, including:

[0104] The vector boundary data of the distribution vector boundary of Ulva prolifera is rasterized to generate a raster image of the Ulva prolifera boundary.

[0105] Binarize the Ulva prolifera boundary raster image and then refine the morphology of the binarized Ulva prolifera boundary raster image to generate a refined visual morphology initial image.

[0106] Noisy pixels are removed from the initial image data to refine the visual morphology, resulting in a cleaned visual morphology image.

[0107] Visual morphology pixel connectivity analysis is performed on purified visual morphology images to generate visual morphology connectivity structure data.

[0108] Extract key nodes from the visual morphology connected structure data, calculate the node density of the visual morphology node set data, and generate a visual morphology node density distribution map.

[0109] Density clustering is performed on the visual morphology node density distribution map to generate high-density visual morphology cluster center data;

[0110] By extracting stable clustering regions from the visual morphology node density distribution map using high-density visual morphology cluster center data, the core area of ​​Ulva prolifera aggregation can be obtained.

[0111] In this embodiment of the invention, the existing vector boundary (polygonal lines) of *Ulva prolifera* is converted into a grid-like image (rasterization). Using the "vector to raster" function in map software or GIS tools (such as QGIS), an appropriate resolution is set to ensure the boundary lines are clearly displayed in the raster image. The pixels containing the *Ulva prolifera* boundary in the raster image are set to 1 (white), and all others to 0 (black), resulting in a binary image. A morphological "thinning" operation (thinning algorithms are available in most image processing software, such as Photoshop, ImageJ, and OpenCV) is used to transform the boundary lines into a single-pixel-width "visual morphology," preserving the boundary structure but making it thinner. The thinned visual morphology image is examined, and isolated small points or broken short line segments (noise) are deleted. Simple connected component analysis methods can be used to filter out connected regions with areas smaller than a threshold. Connected pixel segments in the visual morphology are identified, and each continuous visual morphology line is marked. This step helps to understand the overall structure of the visual morphology network, such as which parts of the visual morphology are the backbone and which are branches. Key nodes such as bifurcation points and endpoints in the visual morphology are identified. The number of nodes within a certain range of each node is counted, and the node density is calculated to obtain a "density distribution map" showing the degree of node clustering. Using simple clustering methods (such as threshold-based region partitioning), areas with particularly high node density are identified, which are visually dense "hot spots." These hot spots represent the visual clustering cores of the *Ulva prolifera* distribution. Using the cluster centers as the core, stable and dense surrounding areas are delineated as the core clustering areas of *Ulva prolifera*. These core areas represent the most densely and concentrated areas of *Ulva prolifera* growth, facilitating subsequent focused management or harvesting.

[0112] Preferably, the extraction of stable clustering regions from the visual morphology node density distribution map using high-density visual morphology clustering center data includes:

[0113] A region is identified as having an abnormal visual morphological node density when any of the following conditions are met, and abnormal visual morphological node density data is obtained: the visual morphological node density per unit area exceeds 150 nodes / km² or is less than 30 nodes / km²; the average distance between visual morphological nodes is less than 5m or greater than 50m; and the standard deviation of the density distribution is higher than 25.

[0114] A region is identified as having unstable cluster center stability when all of the following conditions are met, and cluster center deviation data are obtained: the distance the cluster center moves is greater than 10m in five consecutive time-series analyses; the cluster profile coefficient is less than 0.5; and the coefficient of variation of node density within the corresponding region exceeds 0.3.

[0115] An unstable clustering region is identified and its data is obtained when the following conditions occur simultaneously: the number of visual morphological nodes decreases by more than 20% within 30 minutes; the average node density decreases to below 40 nodes / km²; and the local maximum density point moves more than 15m within the spatial range, and this change continues to occur within 60 minutes.

[0116] By extracting stable clustering regions from the visual morphological node density distribution map using abnormal node density data, cluster center deviation data, and unstable clustering region data, the core clustering area of ​​*Ulva prolifera* can be obtained.

[0117] In this embodiment of the invention, the density of visual morphological nodes per unit area (e.g., how many visual morphological nodes per square kilometer) is calculated. If this density exceeds 150 nodes / km² or is less than 30 nodes / km², the node density of the area is considered abnormal. The average distance between visual morphological nodes is calculated: if it is less than 5 meters, it indicates that the nodes are too concentrated; if it is greater than 50 meters, it indicates that the nodes are too sparse, also considered abnormal. The standard deviation of the node density is calculated; if it is greater than 25, it indicates that the node distribution is uneven and belongs to an abnormal area. Areas that meet any of the above conditions are marked as areas with abnormal visual morphological node density, and corresponding abnormal data records are generated. Through five consecutive time-series data analyses, the changes in the location of the cluster centers are observed. If the distance the cluster center moves each time exceeds 10 meters, it indicates that the cluster location is unstable. The silhouette coefficient of the cluster is calculated; if it is less than 0.5, it indicates poor clustering effect and low stability. The coefficient of variation of the node density in the area is calculated; if it exceeds 0.3, it indicates that the node density fluctuates greatly. Only when all three conditions are met simultaneously is the area determined to be a cluster center stability deviation area, and deviation data records are generated. Observe the changes in the number of visual morphological nodes within a certain time period (e.g., 30 minutes): If the number of nodes decreases by more than 20%, it indicates a significant reduction in the aggregation intensity of the area. If the average node density drops below 40 nodes / km² during the same period, it indicates that the density is below the stable aggregation standard. Monitor the positional changes of the local maximum density point: If it moves more than 15 meters spatially and this change continues within 60 minutes, it indicates that the aggregation location is unstable. If all the above conditions occur simultaneously, the area is determined to be an unstable aggregation area, and relevant data is generated. Overlay and analyze the above three types of data—abnormal visual morphological node density data, cluster center deviation data, and unstable aggregation area data—on the visual morphological node density distribution map to exclude abnormal and unstable areas. The remaining area is the stable core area of ​​Ulva prolifera aggregation.

[0118] Preferably, step S2 involves conducting a radiation analysis of the impact of salvage on the core area of ​​*Ulva prolifera* accumulation using the spatiotemporal trajectory data of the salvage vessel, including:

[0119] The spatiotemporal trajectory data of the salvaged vessel is resampled for time series of trajectory points to generate time-uniformed data of the vessel trajectory;

[0120] By calculating the positional thermodynamic intensity kernel density of the spatiotemporal trajectory data of the salvaged vessel using the time homogenization data of the vessel trajectory, a spatial distribution map of salvage intensity is obtained.

[0121] Based on a preset frequency threshold, the high-frequency salvage boundary is extracted from the spatial distribution map of salvage intensity to obtain the boundary data of the high-frequency salvage area;

[0122] Spatial overlap analysis was performed on the boundary data of the high-frequency dredging area and the core area of ​​seaweed aggregation to generate a core area dredging interaction map.

[0123] Calculate the salvage disturbance coefficient of the core area salvage interaction map, and perform spatial radiation impact modeling of the core area of ​​seaweed aggregation based on the salvage disturbance coefficient to generate spatial modeling data of salvage impact;

[0124] The temporal dimension of the spatial radiation model data affected by salvage was simulated to obtain the analysis results of the radiation impact of salvage.

[0125] In this embodiment of the invention, the trajectory data of salvage vessels is typically collected at uneven time points, such as GPS records that are sometimes dense and sometimes sparse. These trajectory points need to be resampled at fixed time intervals (e.g., every minute or every 5 minutes) to supplement or interpolate missing points, ensuring uniform time points. After this processing, the trajectory time series becomes regular, facilitating subsequent analysis. Using the homogenized trajectory data, the frequency and density of vessels appearing around each location point are calculated. Specifically, kernel density estimation is used to smooth the density around each trajectory point using an "influence range," obtaining a thermal distribution map of salvage intensity in space. This map visually shows where salvage activity is most frequent and where it is less frequent. Based on a pre-set salvage frequency threshold (e.g., areas in the thermal map where the density exceeds a certain value), high-frequency salvage areas are identified. Contour extraction techniques are used to draw the boundary coils of these areas, generating vector boundary data for the high-frequency salvage areas. The boundaries of the high-frequency salvage areas are spatially superimposed with the boundaries of the Ulva prolifera aggregation core area. Identify the overlapping areas, i.e., the interactive areas within the core area of ​​the seaweed that are frequently affected by salvage vessels, and generate a vector layer for this area for analysis and management. Within the core salvage interaction area, calculate the salvage disturbance coefficient to reflect the intensity of the disturbance to the seaweed caused by salvage activities. The salvage disturbance coefficient can be calculated comprehensively based on factors such as vessel salvage frequency, vessel size, and operation duration (e.g., disturbance coefficient = salvage frequency × operation intensity weight). Based on the disturbance coefficient, establish a spatial radiation model to simulate the range and intensity of the salvage activity's influence spreading from the interactive area to the surrounding area. More specifically, the spatial radiation model construction includes using the boundary of the core salvage interaction area and the corresponding salvage disturbance intensity data as input to define the spatial scope of the analysis, typically extending outwards from the core area of ​​seaweed aggregation by a certain distance. Next, divide the analysis area into spatial grids of uniform size, each grid representing a spatial unit, facilitating quantitative analysis of spatial disturbance intensity. Then, assign corresponding disturbance intensities to the grids within the core area containing salvage trajectory points, while the initial disturbance intensity of the outer grids is zero. To simulate the actual attenuation of disturbance effects with distance, a suitable spatial attenuation function is designed, typically employing an exponential attenuation form to ensure the influence intensity decreases exponentially from the core area outwards. Subsequently, the distance from each grid to key points in the core area is calculated, and the core disturbance intensity is distributed to each grid according to the distance attenuation function. If there are multiple high-disturbance points in the core area, their combined influence on the grids is used to form the final disturbance intensity value for each spatial unit. The disturbance intensities of all grids are plotted as a spatial radiation influence map, with color depth reflecting the strength and spatial diffusion range of the salvage disturbance, thus visually demonstrating the impact of salvage on the core area and surrounding regions of the *Ulva prolifera* accumulation.Finally, by comparing measured data, the attenuation parameters were adjusted to optimize the model's accuracy. When necessary, spatiotemporal dynamic simulations were performed using salvage data from different time points to analyze the changing trends of the salvage impact over time, achieving a comprehensive assessment and monitoring of the spatial radiation from salvage disturbances. The salvage disturbance coefficient and spatial radiation model were combined with data from different time periods to dynamically simulate the temporal changes in the salvage impact. By comparing the radiation range and intensity at different time points, the long-term and short-term impacts of salvage activities on the core area of ​​*Ulva prolifera* were analyzed, generating a time-series analysis report of the salvage impact radiation to help decision-makers evaluate and adjust salvage operation plans.

[0126] Of particular importance, the simulation of the temporal changes in the spatial radiation model data of the salvage impact also includes:

[0127] Time series slicing is performed on the spatial radiation model data of salvage impact to generate salvage impact time series slice dataset;

[0128] Extract the inter-frame variation features of the salvage impact time-series slice dataset, and perform salvage impact diffusion speed analysis on the salvage impact time-series slice dataset based on the inter-frame variation features to generate a dynamic feature map of salvage impact diffusion.

[0129] Clustering of key time periods is performed on the dynamic feature map of the impact of salvage, and radiation change map of key time periods is generated;

[0130] Trend regression modeling was performed on the radiation change map during key time periods to generate a trend model of the time evolution of the impact of salvage on salvage.

[0131] Anomaly migration detection processing was performed on the data of the spatial radiation model of salvage impact by using the salvage impact temporal evolution trend model, and salvage impact temporal stability analysis data was generated.

[0132] A full-time fusion and visualization model was performed on the data of the salvage impact time stability analysis to generate the analysis results of the salvage impact radiation.

[0133] In this embodiment of the invention, the acquired spatial radiation model data of the salvage impact is sliced ​​along the time dimension to form a data sequence with continuous time steps (such as days, hours, etc.). The radiation model data at each moment constitutes a "frame," and consecutive frames are combined to form a salvage impact time-series slice dataset. The slicing process can adopt a sliding window or fixed time step slicing strategy to ensure continuity and comparability. Inter-frame variation calculations are performed on the salvage impact time-series slice dataset to extract the change features (such as area change, intensity gradient change) of key radiation intensity change areas. The diffusion rate of the spatial impact area is calculated based on image processing methods such as frame difference maps and edge tracking, thereby generating a dynamic feature map of salvage impact diffusion to represent the spatial distribution of impact rates in different areas. Time series clustering algorithms (such as K-means, DTW clustering, DBSCAN, etc.) are used to perform pattern recognition in the time domain on the diffusion dynamic feature map. The clustering results identify the "key time period" where the changes in the salvage activity's impact are most significant, and the spatial radiation distribution map during this period is extracted to generate a radiation change map of the key time period. We perform regression modeling on indicators such as radiation intensity, impact range, and diffusion direction in the radiation change spectrum during key time periods. Methods such as multinomial regression, exponential smoothing regression, and time series prediction models (e.g., ARIMA, LSTM) can be used to construct a temporal evolution trend model of the salvage impact, reflecting the evolution of the salvage's spatial radiation impact over time. Using the evolution trend model, we compare predicted and actual values ​​across the entire period to identify anomalous offset areas (e.g., areas not covered by salvage, excessively dense areas, and areas with feedback lag). Anomalies are marked using statistical analysis (e.g., Z-score) or residual analysis methods, forming temporal stability analysis data for the salvage impact, and evaluating the stability of model predictions and consistency with actual operations. We then fuse the temporal evolution process throughout the entire time period into a model, combining visualizations such as time axes, geographic coordinate systems, and heat maps to generate a dynamically rendered spatiotemporal fusion visualization. The output results are the analysis results of the salvage impact radiation, including: spatial heat map animation; temporal evolution curves; highlighted anomalous areas; and a comparison chart of predicted trends.

[0134] Preferably, random sampling of the *Ulva prolifera* green tide based on the spatial distribution heat map of salvage intensity includes:

[0135] Use manual or mechanical nets to collect surface seaweed samples at designated sampling points. Collect at least 1 square meter of seaweed at each sampling point and record the sampling time, location coordinates and harvesting environmental conditions to obtain seaweed sample data.

[0136] Calculate the sample area of ​​the *Ulva prolifera* sample data to obtain the initial biomass of the *Ulva prolifera* sample;

[0137] Based on the initial biomass of the Ulva prolifera samples, seawater temperature, light intensity, light duration, and nutrient concentration were monitored to obtain environmental monitoring data for Ulva prolifera.

[0138] In this embodiment of the invention, based on existing thermal maps of salvage intensity spatial distribution, high-intensity, medium-intensity, and low-intensity salvage areas are divided. Sampling points are selected from representative areas to ensure uniformity and typicality of spatial distribution. At each sampling point, surface seaweed is quantitatively collected using manual (e.g., dip nets) or mechanical salvage equipment (e.g., drum samplers, float salvage machines). The sampling area at each point is not less than 1 square meter to ensure statistical effectiveness of the data. During the sampling process, the following metadata must be recorded simultaneously to form the seaweed sample dataset: ① Sampling time (year, month, day, and specific time period); ② Latitude and longitude coordinates of the sampling point (using GPS positioning); ③ Salvage environmental conditions (e.g., sea state, wave height, current speed, wind force level, etc.); ④ Sampling tool type and operation method; ⑤ Sample integrity description. All recorded data are entered into the sampling log in a standard format. Next, the collected seaweed samples are processed by calculating the sample area and weighing. After spreading, compacting, and dehydrating, the samples were accurately weighed to obtain the dry or wet weight per unit area, which was then converted into the initial biomass of the *Ulva prolifera* sample, expressed in g / m² or kg / m². Subsequently, based on the initial biomass of the *Ulva prolifera* sample, environmental parameters closely related to the growth status of the *Ulva prolifera* were collected, including: seawater temperature: monitored in real time using an immersion temperature sensor or CTD probe; light intensity and duration: photosynthetically active radiation (PAR) and sunshine duration (in hours) were recorded using a photon sensor; nutrient concentration: water samples were collected from the water body where the *Ulva prolifera* grew, and the concentrations of NO⁻, NH⁺, PO₄³⁻, and SiO₄⁻ were analyzed. 4 Nutrients such as ⁻ were tested in the laboratory or analyzed online to generate nutrient concentration data. Finally, the initial biomass of the *Ulva prolifera* samples was correlated with the environmental monitoring data of the sampling points to form a *Ulva prolifera* environmental monitoring sample dataset with spatial identification, time labeling, and ecological attributes. This provides basic data support for subsequent green tide biomass evolution modeling, ecological driving factor analysis, and predictive monitoring.

[0139] Preferably, the growth characteristics of *Ulva prolifera* in step S3, when analyzing the *Ulva prolifera* sample data, include:

[0140] A comprehensive growth model for *Ulva prolifera* was constructed based on environmental monitoring data. The formula for the comprehensive growth model is shown below:

[0141]

[0142]

[0143]

[0144] In the formula, For a moment The biomass of *Ulva prolifera* This is the initial biomass. The maximum growth coefficient, Temperature is a factor that affects the environment. The light intensity influencing factor, As a factor affecting sunshine duration, Factors affecting nutrient concentration For growth time, For the potential maximum growth rate, The total loss rate, The baseline proliferation rate at the reference temperature. For activation energy, The gas constant is... This is the actual water temperature. As respiratory activation energy, To maintain the consumption rate, To decompose the loss rate, This refers to the rate of loss due to disease or stress.

[0145] A comprehensive growth model for Ulva prolifera was used to predict the initial biomass of Ulva prolifera samples, generating predicted biomass data.

[0146] The growth rate of the initial biomass of *Ulva prolifera* samples was calculated based on the predicted biomass data, thus obtaining the growth characteristics of *Ulva prolifera*.

[0147] In this embodiment of the invention, collected *Ulva prolifera* sample data and environmental monitoring data (including water temperature, light intensity, sunshine duration, and nutrient levels) are structured, and a dynamic comprehensive growth model for *Ulva prolifera* is constructed based on this data. This model aims to predict the biomass changes of *Ulva prolifera* under the influence of different environmental factors, and its core calculation formula is: For a moment The biomass of *Ulva prolifera* This is the initial biomass. The maximum growth coefficient, Temperature is a factor that affects the environment. The light intensity influencing factor, As a factor affecting sunshine duration, Factors affecting nutrient concentration For the growth time, the model internally performs energy level modeling of the growth and loss processes: in For the potential maximum growth rate, To determine the total loss rate, parameters such as water temperature, light intensity, sunshine duration, and nutrient concentration in the current area are extracted from the environmental monitoring module and configured as input factors for the model. , , , Simultaneously, the initial biomass of the sample... Input model. Based on the parameters at the current time, the system uses the above model formula to solve for the exponential, and obtains the result at a specific time. Predicted biomass of Ulva prolifera at any given time Multi-period simulations can yield the temporal evolution trajectory of *Ulva prolifera* biomass, ultimately outputting *Ulva prolifera* biomass prediction data. Derivative calculations are performed on biomass data over consecutive time periods to estimate the growth rate (biomass increase per unit time). Combined with rate change trends, key indicators such as growth peaks, plateau phases, and decline trends in the growth cycle are extracted. These results form a comprehensive set of *Ulva prolifera* growth characteristics, which can be used for further population monitoring, model optimization, or environmental feedback analysis.

[0148] Preferably, the association of Ulva prolifera growth characteristics with environmental features and the spatial growth evolution of Ulva prolifera aggregation core areas through Ulva prolifera floating path data include:

[0149] Acquire wind field data and water flow velocity field data for the target area;

[0150] Vortex structure is extracted from wind field data in the target area to generate wind field vortex feature data.

[0151] A non-uniform flow field model is constructed based on the velocity field data of the water body flow field, and the local shear stress distribution of the velocity field data of the water body flow field is calculated based on the non-uniform flow field model to obtain the shear characteristic data of the flow channel.

[0152] Acquire wave observation data of the core area of ​​seaweed aggregation; analyze the wind-wave coupling intensity and main wave direction characteristics of the core area of ​​seaweed aggregation based on wind field vortex characteristic data, channel shear characteristic data and wave observation data, and conduct lateral migration analysis of seaweed in the core area of ​​seaweed aggregation to generate seaweed migration bias data.

[0153] Velocity gradient stripping analysis was performed on the flow channel shear characteristic data to generate data on the distribution stability of Ulva prolifera.

[0154] The floating potential energy surface is comprehensively quantified by combining the migration bias data and the distribution stability data of *Ulva prolifera* to generate *Ulva prolifera* floating path data. The formula for the comprehensive quantification of the floating potential energy surface is shown below:

[0155]

[0156] in, To synthesize the buoyancy potential energy distribution function, Representing the horizontal and vertical coordinates of space, respectively. For a moment, For wind field in Local wind speed value at a point For the water flow field in Local velocity value at a point The weighting coefficient for the influence of wind speed on the floating trend. The weighting coefficient for the influence of flow velocity on the floating trend;

[0157] A time-series frequency analysis of the floating path data of *Ulva prolifera* was performed to generate *Ulva prolifera* drift trend data; the drift trend data of *Ulva prolifera* was coupled and fitted with the spatial location of the core area of ​​*Ulva prolifera* aggregation to generate spatial mapping data of drift trend.

[0158] Spatiotemporal growth evolution modeling is performed on the drift trend spatial mapping data to generate Ulva prolifera growth prediction data.

[0159] In this embodiment of the invention, wind field data and water flow velocity field data of the target sea area are acquired. The wind field data includes wind speed and direction information at different height levels, while the flow field data includes surface and subsurface velocity vector fields, used to construct the initial conditions of the marine hydrodynamic environment. Next, vortex structure extraction and analysis are performed on the wind field data. Using Lagrange vortex identification methods (such as Okubo-Weiss parameter analysis) or vorticity tensor decomposition models, vortex cores, vortex boundaries, and other regions in the wind field are identified and labeled, thereby generating wind field vortex characteristic data to support subsequent wave intensity determination. In the water flow field analysis, a non-uniform flow field model is constructed based on existing velocity field data. This model performs weighted fitting of velocity vectors at different grid nodes to construct a local velocity gradient distribution map. Furthermore, using shear stress tensor calculation, the shear stress distribution in each region of the water flow field is obtained, generating channel shear characteristic data to determine stable floating areas of *Ulva prolifera* and hydraulic trigger points where detachment and re-aggregation may occur. Then, the dominant wave direction information and wind field vortex axis information are extracted from the wave observation data; the angle and co-directional factor between the two are calculated to evaluate the wind-wave coordinated propulsion capability; the wind-wave coupling index (CI) is output, the larger the index, the stronger the combined force of the wind field and wave field on the migration of Ulva prolifera. The centroid of the dominant wave direction is calculated using the directional frequency distribution function to obtain the dominant wave direction vector; the dominant wave direction vector is smoothed by a time sliding window to reduce short-period wave direction disturbances; the stability and persistence of the dominant wave direction are analyzed in conjunction with the wind vortex streamlines to form the wave-dominated propulsion path. The shear deformation tensor field in the channel shear characteristic data is projected onto the dominant wave direction axis; the suppression or enhancement effect of the shear tensor field on the lateral disturbance of Ulva prolifera is analyzed, and the lateral disturbance factor is quantified. After obtaining the three-field collaborative feature data, the system constructs a lateral migration model of *Ulva prolifera*, specifically including: superimposing wind direction offset vector and shear stress disturbance vector with the main wave direction as the central axis to form a two-dimensional resultant force field; using the *Ulva prolifera* floating body response coefficient (considering density, adhesion, etc.) to correct the vector intensity, obtaining the floating body migration velocity field. Multiple spatial initial points are set within the *Ulva prolifera* aggregation core area; pathline integration is performed on the two-dimensional resultant force field to calculate the cumulative displacement vector at each point; the lateral offset distribution of all pathline endpoints on the main wave direction axis is statistically analyzed to obtain the lateral migration offset. The offset vectors of all initial points are interpolated and rasterized; the final output is "*Ulva prolifera* migration bias data," which includes spatial attributes such as offset direction, offset amplitude, and offset variability. Simultaneously, velocity gradient stripping analysis is performed on the channel shear feature data. By comparing the velocity change rate between the upper and lower water layers with the local shear threshold, it is determined whether the *Ulva prolifera* is in a stable floating state, and the stripping probability distribution is extracted to generate *Ulva prolifera* distribution stability data. Based on this, using wind speed and flow velocity fields as variable inputs, a floating potential energy surface distribution function is constructed: in, To synthesize the buoyancy potential energy distribution function, Representing the horizontal and vertical coordinates of space, respectively. For a moment, For wind field in Local wind speed value at a point For the water flow field in Local velocity value at a point The weighting coefficient for the influence of wind speed on the floating trend. The weighting coefficient for the influence of flow velocity on the floating trend is defined. Finally, the data on the migration bias of *Ulva prolifera* and the stability of its distribution are superimposed as parameters onto the potential energy surface function. Continuous simulation of the potential energy surface is performed using spatiotemporal sliding window convolution and Gaussian diffusion kernels to obtain floating path data reflecting the future spatial behavior trend of *Ulva prolifera*, which is used for dynamic prediction and environmental response modeling. Throughout the process, data flow, computational logic, and ecological characteristics are tightly coupled, forming a high-precision mechanism for predicting the environmental response and floating behavior of *Ulva prolifera*. Time series statistical methods (such as sliding window counting and frequency distribution analysis) are used to process the timestamps in the *Ulva prolifera* floating path data to calculate the frequency of target occurrence within different time periods. The statistical results are used to form a *Ulva prolifera* drift trend curve, reflecting the dynamic change in the number of *Ulva prolifera* over time, generating *Ulva prolifera* drift trend data. Spatiotemporally coupled data on the drift trend of *Ulva prolifera* with spatial location data (coordinates of the core area of ​​*Ulva prolifera* aggregation), and fitting is achieved using regression analysis, spatiotemporal kriging interpolation, or machine learning methods (such as spatiotemporal convolutional neural networks) to generate spatial mapping data of the drift trend, describing the spatial distribution changes of *Ulva prolifera* growth intensity within the core area at different time points. A suitable spatiotemporal evolution model for *Ulva prolifera* growth characteristics is selected, such as a diffusion-reaction model, a growth model based on partial differential equations, or a spatiotemporal prediction model based on machine learning. Model parameters are set based on historical growth trends and environmental factors (temperature, light, nutrient concentration, etc.). Using the spatial mapping data of the drift trend generated in step S41 as input, combined with environmental driving force parameters, spatiotemporal dynamic simulation is performed. The output is a prediction of the spatial distribution of *Ulva prolifera* at different future time points, forming *Ulva prolifera* growth prediction data.

[0160] Preferably, the spatiotemporal growth and evolution modeling processing of the drift trend spatial mapping data includes:

[0161] Multi-temporal sample segmentation is performed on the drift trend spatial mapping data to generate a time-series segmented growth sample set; the characteristics of the boundary of Ulva prolifera patch changing with time are extracted from the time-series segmented growth sample set to obtain time-driven boundary deformation feature data.

[0162] The growth boundary velocity is fitted to the time-driven boundary deformation feature data to generate spatial boundary growth rate data; the dominant growth direction of the spatial boundary growth rate data is identified.

[0163] Based on the dominant growth direction, the growth rate data of the spatial boundary is used to derive the growth rules of the rasterized region, generating raster growth transfer data; the set of raster growth transfer functions is subjected to spatiotemporal iterative simulation to generate dynamic process data of Ulva growth evolution;

[0164] The dynamic process data of Ulva prolifera growth and evolution were validated and processed to generate Ulva prolifera growth prediction data.

[0165] In this embodiment of the invention, by dividing the drift trend spatial mapping data according to the time series and employing time window sliding or segmentation techniques, the overall time series data is divided into multiple continuous temporal sample segments, forming a temporal segmented growth sample set. Each temporal sample corresponds to the spatial distribution of *Ulva prolifera* patches within a specific time period, ensuring that the temporal segmentation has temporal uniformity and growth continuity. Image segmentation or spatial clustering (such as threshold segmentation, region growing algorithms, or deep learning-based semantic segmentation) is performed on the *Ulva prolifera* patches of each temporal sample to extract the spatial boundary lines of the patches. Morphological change indices of patch boundaries between consecutive temporal phases are calculated, such as boundary displacement vectors, curvature changes, and boundary complexity, generating time-driven boundary deformation feature data. Based on the boundary displacement data obtained in step S421, the boundary motion velocity is fitted and calculated using methods such as least squares, polynomial fitting, or Kalman filtering to generate spatial boundary growth rate data. The velocity data includes the magnitude and direction of the local growth velocity of each boundary line segment. Principal component analysis (PCA), directional statistics, or vector field clustering methods are applied to extract the dominant growth direction from the spatial boundary growth rate data, reflecting the main directional characteristics of the overall expansion of *Ulva prolifera* patches. The identification results guide the setting of directional weights for spatial growth rules in subsequent growth models. The spatial boundary growth rate data is converted into a regular raster data format, defining the growth rate and direction of each raster cell to generate raster growth transfer data. Growth rules may include neighborhood propagation probability, direction-dependent weights, and environmental factor regulation, reflecting the evolutionary pattern of *Ulva prolifera* patches in the raster space. Based on the raster growth transfer function set, spatiotemporal iteration is performed using cellular automata (CA) or Markov process models. By iteratively updating the state of each raster cell, the spatial expansion process of *Ulva prolifera* patches is simulated, generating dynamic process data of *Ulva prolifera* growth and evolution. The simulation process considers complex factors such as environmental constraints and growth competition to ensure the dynamic realism of the model. Historical observational data (such as actual multi-temporal remote sensing imagery and field monitoring data) were used to validate the simulated growth dynamics. Indicators such as spatial overlap (Jaccard index), morphological similarity, and error statistics were employed to quantify model accuracy. Model parameters were adjusted based on the validation results to improve predictive performance. After validation and adjustment, the model was used to simulate future periods, outputting predicted growth data for *Ulva prolifera*, including spatial distribution, expansion rate, and dynamic trends. This predicted data supports subsequent optimization of harvesting strategies and ecological management decisions.

[0166] As an example of the present invention, reference is made to Figure 3 As shown, step S4 in this example includes:

[0167] Step S41: Perform spatial correspondence analysis between the heat map of spatial distribution of salvage intensity and the predicted data of seaweed growth to generate salvage coverage consistency data; calculate the salvage effectiveness per unit area based on the salvage coverage consistency data to generate salvage efficiency coefficient data.

[0168] Step S42: Perform correlation efficiency analysis on the salvage efficiency coefficient data and the salvage operation volume statistics to generate salvage effect evaluation index data; use the salvage effect evaluation index data to evaluate and visualize the statistical salvage operation volume, and generate a seaweed green tide salvage effect evaluation report.

[0169] In this embodiment of the invention, by spatially overlaying a heat map of the spatial distribution of salvage intensity with predicted data of *Ulva prolifera* growth, spatial statistical methods (such as spatial correlation coefficient and overlap index) are used to quantitatively describe the spatial correspondence between salvage areas and *Ulva prolifera* growth areas, generating salvage coverage consistency data. Based on the salvage coverage consistency data, combined with the *Ulva prolifera* growth intensity and salvage intensity per unit area, a salvage effectiveness index per unit area (e.g., the ratio of salvage intensity to predicted growth density) is calculated. All unit area effectiveness indices are summarized to form overall salvage efficiency coefficient data. Combining the salvage efficiency coefficient data with actual salvage operation statistics (including salvage time, number of vessels, operation duration, etc.), multivariate statistical analysis (such as correlation analysis and regression analysis) is used to evaluate salvage operation efficiency. Taking into account both salvage efficiency and operation volume, an evaluation index system (such as salvage rate, coverage rate, resource saving rate, etc.) is constructed to quantify the *Ulva prolifera* green tide salvage effect, forming salvage effect evaluation index data. Based on the evaluation index data, an assessment report on the effectiveness of the seaweed green tide harvesting is generated, including textual descriptions, statistical charts (heat maps, time series graphs, spatial distribution maps), and trend predictions. Interactive displays are achieved using GIS platforms or data visualization tools (such as ArcGIS, Tableau, Python's Matplotlib, and Plotly). More specifically, GPS tracks of the harvesting vessels and daily harvesting volume are collected; the daily harvesting volume is aggregated into a "grid harvesting density" (unit: tons / grid) using a 500 m × 500 m grid; kernel density estimation or inverse distance weighted (IDW) interpolation is performed on the grid density to generate a continuous heat map. For example, based on the micro-growth trend data of the past 4 weeks and environmental factors such as sea temperature, current velocity, and nutrients, a diffusion-response model or time series network (such as LSTM) is used to predict the distribution for the next 1, 2, and 4 weeks; the prediction results are output in the form of grid average density (unit: kg / grid). The heat map of harvesting is overlaid with the distribution map of *Ulva prolifera* at each prediction time point. Within each grid, the "coverage consistency percentage" is calculated, representing the proportion of actual harvested density to predicted density, ranging from 0% (no cover) to 100% (complete cover). Coverage consistency percentages are displayed using identical 500 m grids to visually identify areas where harvesting closely matches growth hotspots. For each grid, its coverage consistency percentage is multiplied by the harvesting density of that grid to obtain a "unit effectiveness score." A higher unit effectiveness score indicates greater harvesting intensity in high-growth areas of *Ulva prolifera*. The score theoretically ranges from 0 to the maximum harvesting density of the grid (e.g., 10 tons / grid); it can be normalized to a percentage, for example, a score of 8 tons / grid corresponds to 80 points. The unit effectiveness scores of all grids are then weighted and averaged according to the "ecological sensitivity weight" of each grid (e.g., 0.6 for aquaculture areas and 0.4 for tourist areas) to obtain a total score, typically expressed as 0–100 points.For example, if the aquaculture area scores 70 points and the tourist area scores 50 points, the weighted average total score is approximately 62 points. The effectiveness of the *Ulva prolifera* green tide harvesting is evaluated based on multiple dimensions, including: Resource utilization rate: the ratio of the actual harvested amount (e.g., 1200 tons) to the predicted harvestable amount (e.g., 1500 tons), expressed as 80%. Harvesting timeliness: the average delay from prediction generation to the start of actual harvesting (e.g., 2 days), compared to the maximum permissible delay (e.g., 7 days), can be normalized to 71%. The final comprehensive score (out of 100 points) is synthesized using a weighted average of 50% (efficiency coefficient), 30% (resource utilization rate), and 20% (timeliness).

[0170] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0171] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for evaluating the effectiveness of *Ulva prolifera* (seaweed) green tide harvesting considering its drifting growth, characterized in that... Includes the following steps: Step S1: Acquire temporal multispectral remote sensing images of the target sea area; perform spectral analysis of the Ulva prolifera characteristics on the temporal multispectral remote sensing images to generate the Ulva prolifera distribution vector boundary; obtain the spatiotemporal trajectory data of the salvage vessel and the statistics of salvage operation volume based on the Ulva prolifera distribution vector boundary; Step S2: Extract the visual morphological features of the distribution vector boundary of *Ulva prolifera* to identify the core area of ​​*Ulva prolifera* aggregation; conduct a radiation analysis of the salvage impact on the core area of ​​*Ulva prolifera* aggregation using the spatiotemporal trajectory data of the salvage vessel, and calculate the spatial distribution heat map of salvage intensity for each time period based on the analysis results; wherein, the extraction of the visual morphological features of the distribution vector boundary of *Ulva prolifera* in step S2 to identify the core area of ​​*Ulva prolifera* aggregation includes: The vector boundary data of the distribution vector boundary of Ulva prolifera is rasterized to generate a raster image of the Ulva prolifera boundary. Binarize the Ulva prolifera boundary raster image and then refine the morphology of the binarized Ulva prolifera boundary raster image to generate a refined visual morphology initial image. Noisy pixels are removed from the initial image data to refine the visual morphology, resulting in a cleaned visual morphology image. Visual morphology pixel connectivity analysis is performed on purified visual morphology images to generate visual morphology connectivity structure data. Extract key nodes from the visual morphology connected structure data, calculate the node density of the visual morphology node set data, and generate a visual morphology node density distribution map. Density clustering is performed on the visual morphology node density distribution map to generate high-density visual morphology cluster center data; By extracting stable clustering regions from the visual morphology node density distribution map using high-density visual morphology cluster center data, the core area of ​​*Ulva prolifera* aggregation is obtained. Step S2 involves conducting a salvage impact radiation analysis of the core area of ​​*Ulva prolifera* aggregation using the spatiotemporal trajectory data of the salvage vessel, including: The spatiotemporal trajectory data of the salvaged vessel is resampled for time series of trajectory points to generate time-uniformed data of the vessel trajectory; By calculating the positional thermodynamic intensity kernel density of the spatiotemporal trajectory data of the salvaged vessel using the time homogenization data of the vessel trajectory, a spatial distribution map of salvage intensity is obtained. Based on a preset frequency threshold, the high-frequency salvage boundary is extracted from the spatial distribution map of salvage intensity to obtain the boundary data of the high-frequency salvage area; Spatial overlap analysis was performed on the boundary data of the high-frequency dredging area and the core area of ​​seaweed aggregation to generate a core area dredging interaction map. Calculate the salvage disturbance coefficient of the core area salvage interaction map, and perform spatial radiation impact modeling of the core area of ​​seaweed aggregation based on the salvage disturbance coefficient to generate spatial modeling data of salvage impact; The temporal dimension of the spatial radiation model data of the salvage impact was simulated to obtain the analysis results of the salvage impact radiation. Step S3: Randomly sample the green tide of *Ulva prolifera* based on the spatial distribution heat map of salvage intensity to obtain *Ulva prolifera* sample data; analyze the growth characteristics of *Ulva prolifera* in the sample data and correlate the growth characteristics of *Ulva prolifera* with hydrological environmental characteristics to generate *Ulva prolifera* floating path data; use the *Ulva prolifera* floating path data to conduct spatial growth evolution of the *Ulva prolifera* aggregation core area to generate *Ulva prolifera* growth prediction data. Step S4: Calculate the salvage efficiency coefficient based on the spatial distribution heat map of salvage intensity and the predicted growth data of Ulva prolifera, and evaluate the salvage effect of the statistical salvage operation volume based on the salvage efficiency coefficient, and generate a Ulva prolifera green tide salvage effect evaluation report.

2. The method for evaluating the effectiveness of *Ulva prolifera* green tide harvesting considering the drifting growth of *Ulva prolifera* according to claim 1, characterized in that, Step S1 includes the following steps: Step S11: Acquire a set of remote sensing images of the target sea area; Step S12: Reconstruct the time series of remote sensing images of the target sea area to generate time series multispectral remote sensing image data; extract the spectral features of the time series multispectral remote sensing image data to obtain the characteristic spectral data of Ulva prolifera. Step S13: Extract the boundary from the characteristic spectral data of *Ulva prolifera* to generate boundary data of the *Ulva prolifera* distribution vector; Step S14: Spatial overlay of the distribution vector boundary data of *Ulva prolifera* to generate mask data of suspected *Ulva prolifera* salvage areas; ship trajectory screening of the mask data of suspected *Ulva prolifera* salvage areas to generate spatiotemporal trajectory data of salvage vessels. Step S15: Calculate the salvage operation volume based on the spatiotemporal trajectory data of the salvage vessel, and generate salvage operation volume statistics.

3. The method for evaluating the effectiveness of *Ulva prolifera* green tide harvesting considering the drifting growth of *Ulva prolifera* according to claim 1, characterized in that, Extracting stable clustered regions from the visual morphology node density distribution map using high-density visual morphology cluster center data includes: A region is identified as having an abnormal visual morphological node density when any of the following conditions are met, and abnormal visual morphological node density data is obtained: the visual morphological node density per unit area exceeds 150 nodes / km² or is less than 30 nodes / km²; the average distance between visual morphological nodes is less than 5m or greater than 50m; and the standard deviation of the density distribution is higher than 25. A region is identified as having unstable cluster center stability when all of the following conditions are met, and cluster center deviation data are obtained: the distance the cluster center moves is greater than 10m in five consecutive time-series analyses; the cluster profile coefficient is less than 0.5; and the coefficient of variation of node density within the corresponding region exceeds 0.

3. An unstable clustering region is identified and its data is obtained when the following conditions occur simultaneously: the number of visual morphological nodes decreases by more than 20% within 30 minutes; the average node density decreases to below 40 nodes / km²; and the local maximum density point moves more than 15m within the spatial range, and this change continues to occur within 60 minutes. By extracting stable clustering regions from the visual morphological node density distribution map using abnormal node density data, cluster center deviation data, and unstable clustering region data, the core clustering area of ​​*Ulva prolifera* can be obtained.

4. The method for evaluating the effectiveness of *Ulva prolifera* green tide harvesting considering the drifting growth of *Ulva prolifera* according to claim 1, characterized in that, Step S3, which involves randomly sampling the green tide of *Ulva prolifera* based on the spatial distribution heat map of salvage intensity, includes: Use manual or mechanical nets to collect surface seaweed samples at designated sampling points. Collect at least 1 square meter of seaweed at each sampling point and record the sampling time, location coordinates and harvesting environmental conditions to obtain seaweed sample data. Calculate the sample area of ​​the *Ulva prolifera* sample data to obtain the initial biomass of the *Ulva prolifera* sample; Based on the initial biomass of the Ulva prolifera samples, seawater temperature, light intensity, light duration, and nutrient concentration were monitored to obtain environmental monitoring data for Ulva prolifera.

5. The method for evaluating the effectiveness of *Ulva prolifera* green tide harvesting considering the drifting growth of *Ulva prolifera* according to claim 1, characterized in that, The growth characteristics of *Ulva prolifera* in the *Ulva prolifera* sample data analyzed in step S3 include: A comprehensive growth model for *Ulva prolifera* was constructed based on environmental monitoring data. The formula for the comprehensive growth model is shown below: In the formula, For a moment The biomass of *Ulva prolifera* This is the initial biomass. The maximum growth coefficient, Temperature is a factor that affects the environment. The light intensity influencing factor, As a factor affecting sunshine duration, Factors affecting nutrient concentration. For growth time, For the potential maximum growth rate, The total loss rate, The baseline proliferation rate at the reference temperature. For activation energy, The gas constant is This is the actual water temperature. As respiratory activation energy, To maintain the consumption rate, To decompose the loss rate, This refers to the rate of loss due to disease or stress. A comprehensive growth model for Ulva prolifera was used to predict the initial biomass of Ulva prolifera samples, generating predicted biomass data. The growth rate of the initial biomass of *Ulva prolifera* samples was calculated based on the predicted biomass data, thus obtaining the growth characteristics of *Ulva prolifera*.

6. The method for evaluating the effectiveness of *Ulva prolifera* green tide harvesting considering the drifting growth of *Ulva prolifera* according to claim 1, characterized in that, Step S3 involves associating the growth characteristics of *Ulva prolifera* with environmental features and analyzing the spatial growth evolution of the *Ulva prolifera* aggregation core area using *Ulva prolifera* floating path data, including: Acquire wind field data and water flow velocity field data for the target area; Vortex structure is extracted from wind field data in the target area to generate wind field vortex feature data. A non-uniform flow field model is constructed based on the velocity field data of the water body flow field, and the local shear stress distribution of the velocity field data of the water body flow field is calculated based on the non-uniform flow field model to obtain the shear characteristic data of the flow channel. Acquire wave observation data of the core area of ​​seaweed aggregation; analyze the wind-wave coupling intensity and main wave direction characteristics of the core area of ​​seaweed aggregation based on wind field vortex characteristic data, channel shear characteristic data and wave observation data, and conduct lateral migration analysis of seaweed in the core area of ​​seaweed aggregation to generate seaweed migration bias data. Velocity gradient stripping analysis was performed on the flow channel shear characteristic data to generate data on the distribution stability of Ulva prolifera. The floating potential energy surface is comprehensively quantified by combining the migration bias data and the distribution stability data of *Ulva prolifera* to generate *Ulva prolifera* floating path data. The formula for the comprehensive quantification of the floating potential energy surface is shown below: in, To synthesize the buoyancy potential energy distribution function, Representing the horizontal and vertical coordinates of space, respectively. For a moment, For wind field in Local wind speed value at a point For the water flow field in Local velocity value at a point The weighting coefficient for the influence of wind speed on the floating trend. The weighting coefficient for the influence of flow velocity on the floating trend; A time-series frequency analysis of the floating path data of *Ulva prolifera* was performed to generate *Ulva prolifera* drift trend data; the drift trend data of *Ulva prolifera* was coupled and fitted with the spatial location of the core area of ​​*Ulva prolifera* aggregation to generate spatial mapping data of drift trend. Spatiotemporal growth evolution modeling is performed on the drift trend spatial mapping data to generate Ulva prolifera growth prediction data.

7. The method for evaluating the effectiveness of *Ulva prolifera* green tide harvesting considering the drifting growth of *Ulva prolifera* according to claim 6, characterized in that, Spatiotemporal growth and evolution modeling of drift trend spatial mapping data includes: Multi-temporal sample segmentation is performed on the drift trend spatial mapping data to generate a time-series segmented growth sample set; the characteristics of the boundary of Ulva prolifera patch changing with time are extracted from the time-series segmented growth sample set to obtain time-driven boundary deformation feature data. The growth boundary velocity is fitted to the time-driven boundary deformation feature data to generate spatial boundary growth rate data; the dominant growth direction of the spatial boundary growth rate data is identified. Based on the dominant growth direction, the growth rate data of the spatial boundary is used to derive the growth rules of the rasterized region, generating raster growth transfer data; the set of raster growth transfer functions is subjected to spatiotemporal iterative simulation to generate dynamic process data of Ulva growth evolution; The dynamic process data of Ulva prolifera growth and evolution were validated and processed to generate Ulva prolifera growth prediction data.

8. The method for evaluating the effectiveness of *Ulva prolifera* green tide harvesting considering the drifting growth of *Ulva prolifera* according to claim 1, characterized in that, Step S4 includes the following steps: Step S41: Perform spatial correspondence analysis between the heat map of spatial distribution of salvage intensity and the predicted data of seaweed growth to generate salvage coverage consistency data; calculate the salvage effectiveness per unit area based on the salvage coverage consistency data to generate salvage efficiency coefficient data. Step S42: Perform correlation efficiency analysis on the salvage efficiency coefficient data and the salvage operation volume statistics to generate salvage effect evaluation index data; use the salvage effect evaluation index data to evaluate and visualize the statistical salvage operation volume, and generate a seaweed green tide salvage effect evaluation report.

Images