A hybrid intelligent path planning method and system for enteromorpha tracking monitoring

Abstract

The application provides a mixed intelligent path planning method and system for Enteromorpha tracking monitoring, and relates to the technical field of marine ecological environment monitoring. The method comprises the following steps: obtaining monitoring area data, generating an initial path meeting basic constraints by using a Dijkstra algorithm based on a gridding cost model, then taking the path as an initial solution, performing global iterative optimization with a snake optimization algorithm for the purpose of monitoring coverage, device energy consumption and obstacle avoidance safety, and performing smoothing processing on the optimized path; meanwhile, path re-planning is dynamically triggered based on real-time feedback data. The application effectively solves the technical problem that path planning of unmanned monitoring equipment in complex marine environment is difficult to balance initial feasibility, global optimality and dynamic adaptability, and significantly improves the efficiency, accuracy and reliability of Enteromorpha tracking monitoring.

Description

Technical Field

[0001] This invention relates to the field of marine ecological environment monitoring technology, and in particular to a hybrid intelligent path planning method and system for tracking and monitoring seaweed. Background Technology

[0002] As a common large algae in the ocean, large-scale outbreaks of *Ulva prolifera* can severely damage nearshore ecological balance, fisheries, and tourism. Timely and accurate tracking and monitoring are the core prerequisites for *Ulva prolifera* disaster prevention and control. Only by obtaining real-time distribution range, density changes, and diffusion trends of *Ulva prolifera* can we provide scientific data support for subsequent cleanup and disposal work.

[0003] Currently, mainstream methods for monitoring *Ulva prolifera*, such as satellite remote sensing, aerial patrols, and manual shipboard monitoring, have significant limitations in terms of resolution, cost, efficiency, and safety. With the development of intelligent unmanned systems such as drones and unmanned vessels, monitoring technologies based on these devices have become a research hotspot. However, existing path planning methods for unmanned systems (such as greedy algorithms and genetic algorithms) struggle to achieve an effective balance between path feasibility, global optimality, and dynamic adaptability in complex and dynamic marine environments (such as changing water flow, the coexistence of aerial and surface obstacles, and uneven distribution of *Ulva prolifera*). This results in low monitoring efficiency, frequent blind spots, and an inability to meet the precise requirements for real-time tracking and monitoring of *Ulva prolifera*. Summary of the Invention

[0004] To address the key technical problems of existing unmanned Ulva prolifera monitoring path planning methods that are difficult to adapt to dynamic environments and cannot simultaneously achieve global optimization for multiple objectives, this invention provides a hybrid intelligent path planning method and system for Ulva prolifera tracking and monitoring by integrating the Dijkstra algorithm and the Snake Optimization (SO) algorithm and introducing dynamic adjustment and path smoothing mechanisms. This effectively improves the rationality of the path, its global optimization, and its responsiveness to environmental changes.

[0005] To achieve the above objectives, a first aspect of the present invention provides a hybrid intelligent path planning method for tracking and monitoring *Ulva prolifera*, comprising:

[0006] Acquire data on the distribution of *Ulva prolifera*, marine environmental data, and the status of unmanned equipment in the monitored area;

[0007] The data is preprocessed to obtain a standardized input dataset;

[0008] Based on the input dataset, a rasterized cost model is constructed, with the density of seaweed, the distance to obstacles, and the energy consumption of navigation as the raster cost function. The Dijkstra algorithm is used to generate an initial path that satisfies the constraints of covering the core area of ​​seaweed, avoiding obstacles, and controlling basic energy consumption.

[0009] The initial path is used as the initial solution of the snake optimization algorithm. The algorithm uses the monitoring coverage, total energy consumption of the equipment, the matching degree between the path and the spread trend of Ulva prolifera, and obstacle avoidance safety as multi-objective optimization functions to perform iterative optimization and obtain the global optimal path.

[0010] The global optimal path is smoothed; the field data fed back by the unmanned equipment is received in real time, and it is determined whether the replanning conditions are met based on the real-time monitoring data. If the conditions are met, the process of initial path generation and global path optimization is retried, and finally the planned path is converted into control commands and sent to the unmanned equipment.

[0011] Furthermore, the preprocessing step includes:

[0012] Outlier detection methods were used to remove invalid data;

[0013] Standardize the format of data from different sources to unify the geographic coordinate system and numerical accuracy;

[0014] Dimensionality reduction is performed on high-dimensional data to extract key features, and weights are assigned to each feature using a weight allocation method.

[0015] Furthermore, the construction of the rasterization cost model includes: dividing the monitoring area into uniform grids;

[0016] The grid includes at least three types: open ground, seaweed distribution area, and obstacle area; a cost value is calculated for each grid, which is obtained by weighted summation of a cost term based on seaweed density, a cost term based on distance to the obstacle, and a cost term based on navigation energy consumption.

[0017] Furthermore, the iterative optimization process of the snake optimization algorithm includes:

[0018] Initialize the population by using the key nodes of the initial path as the initial particle swarm for the snake optimization algorithm; construct a multi-objective fitness function, which includes at least several indicators such as monitoring coverage, equipment energy consumption, path matching with the spread trend of Ulva prolifera, obstacle avoidance safety, and path smoothness; calculate the fitness of each individual in the population; determine the particle motion mode based on the iterative state parameters, and adjust the particle position according to the corresponding position update formula;

[0019] Update the global optimal solution and the individual optimal solution; when the iteration termination condition is met, output the path with the highest fitness as the global optimal path.

[0020] Furthermore, the step of determining the particle motion mode based on the iterative state parameters includes:

[0021] Based on temperature parameters and food index, particles can switch between a foraging mode that explores new paths, a combat mode that moves towards the globally optimal particle, and a mating mode that combines the advantages of different groups of particles.

[0022] Furthermore, the method also includes:

[0023] Unmanned equipment periodically feeds back on-site data;

[0024] When the density of seaweed is detected to have reached the preset disaster warning threshold, an early warning message is sent to the monitoring center while generating an adjusted path.

[0025] A second aspect of the present invention provides a hybrid intelligent path planning system for tracking and monitoring seaweed, comprising:

[0026] The data acquisition module is used to acquire data on the distribution of *Ulva prolifera*, marine environmental data, and the status of unmanned equipment in the monitoring area.

[0027] A data preprocessing module is used to preprocess the data to obtain a standardized input dataset;

[0028] The initial path generation module is used to construct a rasterized cost model based on the input dataset, using the density of seaweed, the distance to obstacles, and the navigation energy consumption as the raster cost function, and using the Dijkstra algorithm to generate an initial path that satisfies the constraints of covering the core area of ​​seaweed, avoiding obstacles, and controlling basic energy consumption.

[0029] The global path optimization module is used to take the initial path as the initial solution of the snake optimization algorithm, and to perform iterative optimization with monitoring coverage, total device energy consumption, path matching degree with the spread trend of Ulva prolifera and obstacle avoidance safety as multi-objective optimization functions to obtain the global optimal path.

[0030] The path output control module is used to smooth the globally optimal path; it receives field data from the unmanned equipment in real time, determines whether the replanning conditions are met based on the real-time monitoring data, and re-triggers the initial path generation and global path optimization process when the conditions are met, and finally converts the planned path into control commands and sends them to the unmanned equipment.

[0031] A third aspect of the present invention provides an electronic device including a memory, a processor, and a program stored in the memory and running on the processor, wherein the processor executes the program to implement the steps of the hybrid intelligent path planning method for Ulva prolifera tracking and monitoring as described in the first aspect of the present invention.

[0032] A fourth aspect of the present invention provides a computer-readable storage medium having a program stored thereon that, when executed by a processor, implements the steps of the hybrid intelligent path planning method for Ulva prolifera tracking and monitoring as described in the first aspect of the present invention.

[0033] A fifth aspect of the present invention provides a computer program product comprising software code, wherein the program in the software code performs the steps of the hybrid intelligent path planning method for tracking and monitoring seaweed as described in the first aspect of the present invention.

[0034] Compared with existing technologies, the hybrid intelligent path planning method and system for tracking and monitoring Ulva prolifera provided by this invention has the following beneficial effects:

[0035] (1) This invention uses Dijkstra's algorithm to quickly construct a reliable path framework that satisfies basic constraints, and then uses the powerful global search capability of SO algorithm to perform refined multi-objective optimization, which solves the problem that a single algorithm cannot take into account both local rationality and global optimality, and effectively balances monitoring coverage, equipment energy consumption and path security.

[0036] (2) The present invention constructs a closed-loop control system of “planning-execution-feedback-adjustment”, which has the ability to provide real-time data feedback and dynamic path adjustment. It can respond in a timely manner to changes in the spread of seaweed and sudden situations in the marine environment, avoid monitoring delays or path interruptions caused by static paths, ensure that unmanned equipment can always accurately track the trajectory of seaweed, and improve the continuity and reliability of monitoring tasks.

[0037] (3) The present invention uses smoothing algorithms such as Bézier curves to post-process the path, making the final path more consistent with the kinematic model of the unmanned equipment, reducing energy loss and execution difficulty during turning, and improving the execution efficiency of the path and the stability of equipment operation. Attached Figure Description

[0038] The accompanying drawings, which form part of this disclosure, are used to provide a further understanding of this disclosure. The illustrative embodiments of this disclosure and their descriptions are used to explain this disclosure and do not constitute an undue limitation of this disclosure.

[0039] Figure 1 This is a flowchart of a hybrid intelligent path planning method for tracking and monitoring Ulva prolifera provided in Embodiment 1 of the present invention;

[0040] Figure 2 This is an overall framework diagram of the hybrid intelligent path planning method for tracking and monitoring Ulva prolifera provided in Embodiment 1 of the present invention;

[0041] Figure 3 A rasterized schematic diagram of the Dijkstra algorithm generating the initial path provided in Embodiment 1 of the present invention;

[0042] Figure 4 This is a flowchart of the SO algorithm provided in Embodiment 1 of the present invention;

[0043] Figure 5 The following is a logical diagram of the composition of the multi-objective fitness function provided in Embodiment 1 of the present invention;

[0044] Figure 6 This is a schematic diagram of the triggering logic for dynamic path adjustment provided in Embodiment 1 of the present invention;

[0045] Figure 7 This is a comparison diagram of the smoothing effect of the Bézier curve path provided in Embodiment 1 of the present invention;

[0046] Figure 8 This is an architecture diagram of a hybrid intelligent path planning system for tracking and monitoring seaweed, provided in Embodiment 2 of the present invention. Detailed Implementation

[0047] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0048] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, unless the context clearly indicates otherwise, the singular form is intended to include the plural form as well. Furthermore, it should be understood that the terms “comprising” and “having”, and any variations thereof, are intended to cover non-exclusive inclusion, for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0049] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0050] All data acquisition in this embodiment is carried out in accordance with laws and regulations and with user consent, and the data is used legally.

[0051] Before describing the embodiments of the present invention in detail, the technical terms involved in the present invention are explained and defined as follows:

[0052] 1. Gridded Cost Model: This refers to a mathematical model that discretizes a continuous physical monitoring area into a uniform set of grid cells and assigns a comprehensive cost value to each grid cell. This model is the spatial basis of path planning. Its core lies in quantifying the "passage cost" of each grid cell in the Ulva prolifera monitoring task through a mathematical expression called the "grid cost function." This cost comprehensively reflects the monitoring benefits (such as Ulva prolifera density) and execution costs (such as obstacle threats and navigation energy consumption) brought about by crossing the grid cell.

[0053] 2. Dijkstra's Algorithm (also known as the Shortest Path Algorithm): This is a classic graph search algorithm used to find the shortest (minimum cost) path from a single source node to all other nodes in a graph or raster with non-negative weights. In this invention, this algorithm is applied to the rasterized cost model. Its core function is to quickly and reliably search for an initial feasible path from the starting point to the target point based on the comprehensive cost value, avoiding high-cost areas (such as obstacles) and prioritizing the passage through low-cost areas (such as areas with high density of seaweed), thus providing a high-quality solution for subsequent global optimization.

[0054] 3. Snake Optimization Algorithm (SO): This is a swarm intelligence optimization algorithm that simulates the foraging and reproductive behaviors of snakes. In this invention, this algorithm is used to globally optimize the initial path. It treats the path as "snake particles" and dynamically updates the particle positions by simulating three modes: "foraging," "fighting," and "mating," thereby guiding the path to efficiently explore and utilize the solution space, ultimately finding a globally optimal or near-optimal path that satisfies multi-objective constraints.

[0055] 4. Multi-objective optimization function: A mathematical expression that considers multiple mutually constrained optimization objectives (such as "maximum monitoring coverage", "lowest possible equipment energy depletion", and "maximum obstacle avoidance safety"). By assigning appropriate weights to each objective, the multi-objective problem is transformed into a computable single-objective function, ensuring that the final path achieves balance across multiple dimensions.

[0056] 5. Path Smoothing: This is a post-processing step that performs curve fitting (using Bézier curves in this invention) on the planned discrete path (e.g., a polygonal line). The aim is to make the path more consistent with the kinematic characteristics of the unmanned equipment (drone / unmanned vessel), reduce energy consumption and equipment wear caused by frequent turning, and enable the equipment to perform monitoring tasks smoothly.

[0057] 6. Grubbs' Criterion: A statistical method for identifying and removing outlier data by setting a significance level (in this invention). =0.05), to determine whether the data is invalid due to sensor failure or environmental interference (such as wave reflection or radar interference), and to ensure the reliability of the input data.

[0058] 7. Principal Component Analysis (PCA): A high-dimensional data dimensionality reduction algorithm that can extract a few unrelated "key features" (3 in this invention) from multiple related original data indicators (such as seaweed density, water flow velocity, obstacle type, etc.), simplifying the model's computational complexity while retaining core information.

[0059] 8. Obstacle Avoidance Penalty Term: A mathematical term in the multi-objective fitness function specifically designed to enhance obstacle avoidance safety. When the distance between a path node and an obstacle is less than a preset safety threshold, this penalty term reduces the fitness value of the path, forcing the algorithm to adjust the position of the path node to ensure that the device maintains a safe distance from obstacles.

[0060] Example 1

[0061] like Figure 1 This embodiment provides a hybrid intelligent path planning method for tracking and monitoring seaweed, including:

[0062] Acquire data on the distribution of *Ulva prolifera*, marine environmental data, and the status of unmanned equipment in the monitored area;

[0063] The data is preprocessed to obtain a standardized input dataset;

[0064] Based on the input dataset, a rasterized cost model is constructed, with the density of seaweed, the distance to obstacles, and the energy consumption of navigation as the raster cost function. The Dijkstra algorithm is used to generate an initial path that satisfies the constraints of covering the core area of ​​seaweed, avoiding obstacles, and controlling basic energy consumption.

[0065] The initial path is used as the initial solution of the snake optimization algorithm. The algorithm uses the monitoring coverage, total energy consumption of the equipment, the matching degree between the path and the spread trend of Ulva prolifera, and obstacle avoidance safety as multi-objective optimization functions to perform iterative optimization and obtain the global optimal path.

[0066] The global optimal path is smoothed; the field data fed back by the unmanned equipment is received in real time, and it is determined whether the replanning conditions are met based on the real-time monitoring data. If the conditions are met, the process of initial path generation and global path optimization is retried, and finally the planned path is converted into control commands and sent to the unmanned equipment.

[0067] This invention first acquires and perceives the environment through data acquisition. Then, leveraging the deterministic search characteristics of Dijkstra's algorithm, it quickly finds a safe (obstacle avoidance), effective (covering the core area), and feasible (energy-controllable) initial path, solving the problems of "creating a path from scratch" and "basic reliability." Next, this high-quality initial path is "refined" by the SO algorithm. The SO algorithm simulates the intelligent behavior of a snake swarm, performing multi-objective iterative search within the neighborhood of the initial path, ultimately outputting a global path that achieves a better balance across multiple dimensions such as coverage, energy consumption, matching degree, and safety. Finally, through real-time feedback and dynamic replanning mechanisms, it ensures that the entire system can continuously adapt to changes in the external environment. This forms a complete technical solution with both static optimization and dynamic response capabilities, fundamentally improving the intelligence level and practicality of unmanned system monitoring.

[0068] Specifically, the preprocessing steps include:

[0069] Outlier detection methods were used to remove invalid data;

[0070] Standardize the format of data from different sources to unify the geographic coordinate system and numerical accuracy;

[0071] Dimensionality reduction is performed on high-dimensional data to extract key features, and weights are assigned to each feature using a weight allocation method.

[0072] Raw, multi-source data (such as image coordinates and environmental parameters) often suffers from noise and inconsistent dimensions, leading to model distortion when directly used for planning. By removing outliers, standardizing the format, and weighting features through dimensionality reduction, the chaotic raw data is transformed into a clean, uniform, and feature-rich standardized dataset. This significantly improves the accuracy and reliability of subsequent cost model construction, laying a solid data foundation for generating reasonable initial paths and ensuring the quality of the final path planning from the outset.

[0073] Specifically, the construction of the rasterization cost model includes: dividing the monitoring area into uniform grids;

[0074] The grid includes at least three types: open ground, seaweed distribution area, and obstacle area; a cost value is calculated for each grid, which is obtained by weighted summation of a cost term based on seaweed density, a cost term based on distance to the obstacle, and a cost term based on navigation energy consumption.

[0075] By discretizing a continuous sea area into a uniform grid, a structured space is provided for path search. Different values ​​are assigned to grids with different attributes (open areas, areas with seaweed, areas with obstacles), quantifying abstract objectives such as "covering more seaweed," "avoiding obstacles," and "saving energy" into computable costs. Dijkstra's algorithm searches on this quantified cost map to find the path with the minimum cumulative cost from the starting point to the destination. This transforms a complex multi-objective optimization problem into a single-objective (minimum cost) search problem in a structured space, making path planning computable and executable.

[0076] Specifically, the iterative optimization process of the snake optimization algorithm includes:

[0077] Initialize the population by using the key nodes of the initial path as the initial particle swarm for the snake optimization algorithm; construct a multi-objective fitness function, which includes at least several indicators such as monitoring coverage, equipment energy consumption, path matching with the spread trend of Ulva prolifera, obstacle avoidance safety, and path smoothness; calculate the fitness of each individual in the population; determine the particle motion mode based on the iterative state parameters, and adjust the particle position according to the corresponding position update formula;

[0078] Update the global optimal solution and the individual optimal solution; when the iteration termination condition is met, output the path with the highest fitness as the global optimal path.

[0079] The SO algorithm treats key nodes in the initial path as "snake-like particles." During iteration, the algorithm dynamically adjusts the particle behavior patterns using temperature parameters and a food index: in "foraging mode," particles conduct large-scale random exploration to discover potentially better regions, enhancing the algorithm's ability to escape local optima; in "combat mode," particles learn from the current globally optimal particle, accelerating convergence; in "mating mode," the path characteristics of particles with different advantages (such as high-coverage particles and low-energy-consumption particles) are fused, potentially producing offspring with multiple advantages. Through this collaborative "exploration-utilization-fusion" mechanism, the SO algorithm can efficiently search for Pareto optimal solutions in a complex solution space that simultaneously satisfy multiple competing objectives such as high coverage, low energy consumption, and strong security, overcoming the limitations of traditional single-objective algorithms.

[0080] Specifically, determining the particle motion mode based on the iterative state parameters includes:

[0081] Based on temperature parameters and food index, particles can switch between a foraging mode that explores new paths, a combat mode that moves towards the globally optimal particle, and a mating mode that combines the advantages of different groups of particles.

[0082] Building upon dynamic adjustments, an independent disaster early warning threshold has been added. When the real-time monitored density of *Ulva prolifera* exceeds this threshold, the system will proactively send an early warning signal to the monitoring center while performing path replanning. This upgrades the simple path planning function into a decision support system integrating "monitoring-planning-early warning," providing valuable emergency response time for back-end personnel and greatly enhancing the invention's value in practical *Ulva prolifera* disaster prevention and control.

[0083] Specifically, the method further includes:

[0084] Unmanned equipment periodically feeds back on-site data;

[0085] When the density of seaweed is detected to have reached the preset disaster warning threshold, an early warning message is sent to the monitoring center while generating an adjusted path.

[0086] Through modular design, the complete path planning process is decomposed into independent yet collaborative functional modules, including data acquisition, preprocessing, initial planning, global optimization, smooth output, and dynamic adjustment. Each module is responsible for a specific task and interacts with data through clear interfaces. This achieves high cohesion and low coupling of functions, resulting in a clear system structure that is easy to maintain, expand, and implement, providing a clear software architecture blueprint for deploying this invention in a practical unmanned system control platform.

[0087] In one specific embodiment, the hybrid intelligent path planning method for tracking and monitoring Ulva prolifera includes: acquiring basic data of the monitoring area;

[0088] The collected data is preprocessed to obtain a standardized input dataset;

[0089] Based on the preprocessed dataset, a cost model for the monitoring path of *Ulva prolifera* is constructed: the monitoring area is divided into a gridded space, and the grid cost function is based on the *Ulva prolifera* density within the grid (higher density, higher weight), the distance from the grid to the obstacle (closer distance, higher cost), and the equipment's navigation energy consumption (energy consumption per unit distance calculated based on water flow and wind direction). Based on Dijkstra's algorithm, the path with the minimum cost is searched, with the starting position of the unmanned equipment as the starting point and the preset monitoring endpoint (or the starting point of the cruise loop) as the target point. This initial path must meet the constraints of covering the core distribution area of ​​*Ulva prolifera*, avoiding high-threat obstacles, and having controllable basic energy consumption, to ensure the initial rationality and feasibility of the path.

[0090] The initial path generated by Dijkstra's algorithm is used as the initial solution of the SO algorithm. The multi-objective optimization function is to achieve the highest global monitoring coverage, the lowest total energy consumption of the equipment, the best matching degree between the path and the spread trend of Ulva prolifera, and the strongest obstacle avoidance safety. The SO algorithm is started for iterative optimization to obtain the globally optimal path.

[0091] The system receives real-time feedback data from unmanned equipment and compares the current execution path with the actual situation on site. When the spread of seaweed exceeds the coverage of the current path by more than 10%, or when new high-threat obstacles appear, the system re-triggers the Dijkstra initial path generation → SO global optimization process to generate the adjusted optimal path.

[0092] The overall approach of the scheme described in this embodiment is as follows: A hybrid intelligent path planning method for tracking and monitoring *Ulva prolifera* is proposed. This method addresses the imbalance between path rationality and optimality, and insufficient dynamic adaptability of traditional single algorithms in complex marine environments, through a full-process design encompassing data preprocessing, initial path construction, global path optimization, and dynamic adjustment. The core design concept integrates the reliable path search capability of Dijkstra's algorithm with the multi-objective optimization advantages of the SO algorithm. First, an initial path framework satisfying basic constraints is built using Dijkstra's algorithm. Then, the SO algorithm is used for refined iteration of the path. Simultaneously, a real-time feedback mechanism is introduced to ensure that the path can be flexibly adjusted according to the spread of *Ulva prolifera* and environmental changes, ultimately achieving a multi-dimensional balance between monitoring efficiency, energy consumption control, environmental adaptability, and operational safety for unmanned equipment.

[0093] This embodiment proposes a complete path planning technology framework. This framework, with cost modeling, dual-algorithm hybrid intelligent optimization, and dynamic calibration as its core logic, addresses the feasibility, optimality, and adaptability of paths, respectively. The cost model quantifies the density of *Ulva prolifera*, obstacle threats, and navigation energy consumption, providing a scientific evaluation basis for path planning and avoiding path defects caused by a single-goal orientation. The Dijkstra algorithm quickly generates a collision-free initial path covering the core area, utilizing its greedy search characteristics to ensure the basic reliability of the path. The SO algorithm, starting from the initial path, achieves global optimization of the path in terms of coverage, energy cost, diffusion adaptability, and obstacle avoidance safety through an iterative mechanism simulating serpentine motion. The dynamic adjustment mechanism triggers path replanning by collecting real-time field data, addressing path failure issues caused by dynamic changes in the marine environment and *Ulva prolifera* distribution. The key idea of ​​the entire framework is to avoid the inherent defects of a single algorithm through algorithmic collaboration and a closed-loop design throughout the entire process, ensuring that the path adapts to the actual needs of *Ulva prolifera* monitoring throughout its entire lifecycle, from planning to execution.

[0094] like Figure 2The overall process of the solution provided in this embodiment includes: firstly, data is collected through multi-source sensors to collect data on the distribution of seaweed, marine environment, and equipment status in the monitoring area, and a standardized dataset is formed after data preprocessing; based on this dataset, a rasterized cost model is constructed, and the minimum cost path is searched as the initial path through initial path generation (using Dijkstra's algorithm); the initial path is input into global path optimization (SO algorithm) for iterative optimization to generate the globally optimal path; the unmanned equipment executes the monitoring task according to the optimal path and provides real-time feedback of on-site data; this method judges whether the path adjustment conditions are met by real-time feedback data, and if so, the initial path generation and global optimization process is restarted, and the adjusted path is output, forming a closed loop of data collection, data preprocessing, initial path generation, global path optimization, path execution, and dynamic adjustment, ensuring the continuity and accuracy of the monitoring task.

[0095] During the data acquisition and preprocessing stage, the basic data covers three core types of information: Ulva prolifera distribution data (outbreak area boundary coordinates, coverage density gradient), marine environmental data (water flow speed and direction, wind direction, obstacle location and type), and unmanned equipment status data (remaining battery power, endurance, sensor monitoring range and accuracy). The preprocessing process mainly includes three steps: First, outlier removal, using the Grubbs criterion (significance level α=0.05) to identify and remove invalid data caused by sensor malfunctions and environmental interference (such as false detections of seaweed density caused by wave reflection and false obstacle coordinates caused by instantaneous radar interference); second, format standardization, converting data from different sources (image coordinates, numerical environmental parameters) into a unified format, using the WGS84 coordinate system for geographic coordinates, and retaining two decimal places for numerical data to ensure data consistency; third, feature extraction and weight allocation, using principal component analysis (PCA) to reduce the dimensionality of high-dimensional data, extracting three key features: seaweed density, obstacle threat, and equipment energy consumption, and determining the weights of each feature using the analytic hierarchy process (AHP), with seaweed density weight set at 0.4, obstacle threat weight at 0.3, and equipment energy consumption weight at 0.3, providing a quantitative basis for the subsequent cost model construction.

[0096] like Figure 3 In the initial path generation stage, the monitoring area is first divided into uniform grids of standard 10m × 10m size (grayscale indicators in the figure: open space = white, seaweed = light gray, obstacles = black). Each grid is assigned a unique coordinate index (i, j), forming a rasterized map as shown in the figure. The vertical coordinate (Y) is labeled from 1 to 20, and the horizontal coordinate (X) is labeled from 1 to 20, providing an intuitive spatial carrier for Dijkstra's algorithm path search. Based on the preprocessed feature data, a raster cost function is constructed, expressed by the formula:

[0097]

[0098] in, , , The weights of each feature; Grid The cost of the density of seaweed is assigned according to the density level (3 for high density areas, 2 for medium density areas, 1 for low density areas, and 0 for areas without seaweed). The obstacle cost is calculated using an exponential decay model, as shown in the formula:

[0099]

[0100] The distance from the grid to the nearest obstacle. This is the attenuation coefficient, with a value of 0.5; The energy cost of navigation is determined based on the angle between the water flow, wind direction, and navigation direction (1 for downstream, 2 for cross current, and 3 for upstream).

[0101] Starting from the unmanned equipment's initial position (e.g., grid (100, 100)) and setting the target point as the monitoring endpoint (e.g., grid (900, 900)), the Dijkstra algorithm is initiated for path search. A priority queue is used to store the grids to be searched. Initially, only the starting grid has a cost of 0, while the costs of the remaining grids are set to infinity. The adjacent grids of the starting point are traversed sequentially (in the four directions of up, down, left, and right), their costs are calculated, and the priority queue is updated until the target grid is found. The final initial path must meet three constraints: first, it must cover ≥90% of the core distribution area of ​​*Ulva prolifera*; second, the minimum distance to high-threat obstacles must be ≥50m; and third, the total energy consumption cost must be ≤ the maximum allowable value calculated based on the equipment's remaining power, ensuring that the path is feasible for actual execution.

[0102] In the global path optimization phase, the initial path generated by Dijkstra's algorithm is used as the initial solution for the SO algorithm, initiating multi-objective iterative optimization. The execution flow of the SO algorithm is as follows: Figure 4As shown, the specific steps include: After starting, the population is initialized, and the snake positions and parameters are randomly generated; then, the fitness (i.e., the objective function value of each snake) is calculated; then, it is determined whether the maximum number of iterations has been reached. If not, the position and speed are updated, and the snake moves towards the global optimum. The food position and energy level are further calculated, and the movement state (including feeding / crawling / attacking / defending modes) is determined. The position and speed are updated according to the movement state, and the snake moves towards the global optimum. At the same time, position constraints and boundary processing are performed, and the global optimum and individual optimum are updated. The above process is repeated until the maximum number of iterations is reached. Finally, the optimal solution and path are output, and the optimization ends. First, the SO algorithm parameters are initialized: key nodes on the initial path (such as turning points, the center point of the high-density area of ​​seaweed) are used as snake particles, and the number of particles is the same as the number of key nodes; the search range of the particles is limited to the monitoring area, and the position change of adjacent particles does not exceed 20 grids to avoid excessive path distortion; the iteration threshold is set to 100 times, and the fitness improvement threshold is set to 0.01 (the iteration stops when the difference between the fitness values ​​of two adjacent iterations is less than this threshold).

[0103] Construct a multi-objective fitness function (its composition logic is as follows) Figure 5 As shown, the function, used as an evaluation criterion for path quality, consists of four weighted terms: monitoring coverage C (weight 0.35), energy consumption index (…). (Weight 0.25, where the energy consumption calculation rules are: downstream E=1, cross-flow E=2, upstream E=3), diffusion matching degree M (weight 0.25), and obstacle avoidance penalty term. (Weight 0.15, penalty term is related to the distance from the node to the obstacle, when the distance is ≥50m, =0, the closer the distance, the larger the value), to enhance the path's ability to avoid obstacles, introduce an obstacle avoidance penalty coefficient, the formula is expressed as:

[0104]

[0105] in:

[0106] To monitor coverage (the ratio of the actual number of *Ulva prolifera* grids to the total number of *Ulva prolifera* grids, with a value range of 0-1);

[0107] The total energy consumption of the equipment (calculated based on path length and unit energy consumption). This represents the maximum allowable energy consumption of the equipment. Convert energy consumption into a positive indicator of 0-1;

[0108] The degree of alignment between the path and the direction of *Ulva prolifera* diffusion (the cosine of the angle between the path extension direction and the direction of *Ulva prolifera* diffusion, where the angle is 0°). =1, at 90° =0);

[0109] New obstacle avoidance penalty items The formula used to quantify the degree to which the safe distance between the path and obstacles meets the standard is as follows:

[0110]

[0111] In the formula, This represents the total number of critical nodes on the path. For the first The actual distance (in meters) from each node to the nearest obstacle. Set a preset safe distance threshold (based on the obstacle threat level, generally 50-100m). This is the penalty coefficient (ranging from 0.05 to 0.1), used to adjust the penalty intensity when the distance is insufficient: when ≥ hour, =1, penalty term has no effect; when When the distance is smaller, the value of this item is larger. The smaller the value, the lower the overall fitness F, thus forcing the path to move away from the obstacle.

[0112] The adjusted weighting retains the priority of core monitoring targets while strengthening path safety through independent penalty items. For example, when a node on the path is only 30m from an obstacle (the safety distance is set at 50m), The item incurs a significant penalty, prompting the SO algorithm to adjust the node's position during iterations until the safe distance requirement is met. The core logic of the SO algorithm for this adjustment is to simulate the "foraging, fighting, and mating" behavior of snake groups in nature. Using the key nodes of the initial path as "snake-like particles," the algorithm dynamically balances the exploration of new paths with the optimization of existing paths during iterations. Through coherent behavioral judgments, position updates, and optimal selection, the path continuously converges towards the goal of high coverage, low energy consumption, and strong obstacle avoidance.

[0113] SO algorithm execution flow (corresponding) Figure 4 The process is as follows: First, initialize the population and randomly generate snake positions and parameters; calculate the fitness (objective function value), food location, and energy level of each snake; determine its movement state (foraging / fighting / mating); update particle positions and velocities according to the corresponding mode, moving towards the global optimum; simultaneously perform position constraints and boundary handling; if the maximum number of iterations (100 times) is not reached, repeat the above steps until the termination condition is met, at which point the optimal path is output. The formula for calculating the temperature parameter is...

[0114]

[0115] in The current iteration number (1≤ ≤100), =100 is the maximum number of iterations. The value range is (0,1]: in the initial stage of iteration When smaller, A value close to 1 corresponds to a "high exploration" state, allowing particles to be adjusted over a wide range to cover areas where the seaweed may spread; later in the iteration... When it approaches 100, decay to This corresponds to the "high utilization" state, where only minor adjustments are made to the particles to optimize path details. The formula for calculating the food index is...

[0116]

[0117] The value range is (0, 0.5]: when <0.25 (early iteration) When 0.25 ≤ 50, the "foraging mode" is triggered, exploring new paths through randomly perturbed particles; when 0.25 ≤ ≤0.5 (late stage of iteration) When the number of particles is ≥50, a "mating mode" is randomly triggered to merge the path segments of the best particles from different groups to balance multiple objectives.

[0118] For different behavioral patterns, the SO algorithm uses a differentiated formula to adjust particle positions, ensuring that the adjustment direction is consistent with the optimization objective of the fitness function. In foraging mode, particles need to explore better areas through directional random perturbations, and the position update formula is:

[0119]

[0120] in For the first The particle in the first The coordinates of the next iteration. =0.8 is the disturbance coefficient. A random number in the interval [0,1). Maximum perturbation step size (unit: grid). The direction sign is determined by the fitness function gradient; for example, the direction towards high-density areas of seaweed or away from obstacles is positive. As mentioned earlier, a particle only 30m away from an obstacle can move 5-10 grid squares away from the obstacle using this formula. In combat mode (late iteration), particles need to move towards the globally optimal particle to quickly improve path quality; the position update formula is:

[0121]

[0122] in, =0.3 is the learning factor. For the first The formula for determining the globally optimal particle coordinates in the next iteration (e.g., particles that have met the safety distance and high coverage) guides particles that have not yet reached the safety distance to adjust towards the optimal particle at a distance of more than 50m from the obstacle. In the mating mode (mid-iteration), the advantages of different groups of particles need to be combined, and the position update formula is:

[0123]

[0124] in, For the male group, it is the locally optimal particle (such as the particle with the highest coverage). For the female group, it is the locally optimal particle (such as the particle with the lowest energy consumption). For the random fusion coefficient in the interval [0,1], when When the value is 0.6, the new particles can inherit 60% of the high coverage characteristics and 40% of the low energy consumption characteristics, taking into account both the core monitoring targets and energy consumption control.

[0125] After each iteration, the SO algorithm needs to select local and global optima to provide a benchmark for the next round of optimization. The selection logic is implemented through two formulas: the formula for selecting local optima is:

[0126]

[0127] Where Group represents a single group of particles (male or female), this formula ensures that each group of particles retains its own advantages, such as the male group retaining high coverage and the female group retaining low energy consumption; the formula for selecting the globally optimal particle is:

[0128]

[0129] Where All represents all particles, this formula ensures that the overall path converges to the global optimum. For example, a particle with a coverage of 0.92, an energy consumption coefficient of 0.45, and an obstacle avoidance safety distance of 50m is taken as the global optimum, guiding other particles to optimize towards it.

[0130] The SO algorithm seamlessly integrates the aforementioned behavior judgment, position update, and optimal selection process, allowing the path to be continuously optimized through iteration: For nodes only 30m away from obstacles, it first uses directional random perturbation in foraging mode to initially move them away from obstacles, then moves them closer to the globally optimal particle in combat mode, and finally adjusts them to a safe position more than 50m away from obstacles. At the same time, it integrates the characteristics of high-coverage particles through mating mode to ensure that the node can still cover the core area of ​​the seaweed. After each iteration, the path fitness value corresponding to each particle is calculated, and the local optimal solution and the global optimal solution are selected. When the number of iterations reaches 100 or the difference in global optimal fitness is less than 0.01, the iteration stops, and the globally optimal path is output. Compared with the initial path, this path has achieved significant improvements in monitoring coverage, energy consumption control, diffusion matching degree, and obstacle avoidance safety.

[0131] During iterative optimization, the particle positions are adjusted using a simulated snake-like undulating mechanism: particles near low-density areas of seaweed move towards higher-density areas; particles with large angles to the water flow direction are slightly adjusted downstream to reduce navigation energy consumption; and particles too close to obstacles are shifted to safer areas to enhance obstacle avoidance. After each iteration, the path fitness value for each particle is calculated, and the local optimum (maximum fitness value of a single path) and the global optimum (maximum fitness value among all paths) are selected. When the iteration meets the termination condition, the global optimum path is output, which significantly improves monitoring coverage, energy consumption control, diffusion matching degree, and obstacle avoidance safety compared to the initial path.

[0132] Dynamic adjustment phase (its triggering logic is as follows) Figure 6 As shown in the image, when the unmanned equipment executes the monitoring task according to the globally optimal path, it provides real-time data feedback (cyclic monitoring), sending on-site data to the control center every 5 minutes. This includes real-time images of the *Ulva prolifera* distribution, the location of newly discovered obstacles, and the remaining battery power of the equipment. The spread of *Ulva prolifera* is analyzed using image recognition algorithms to calculate the undercover ratio, expressed by the formula:

[0133]

[0134] in, This represents the total distribution area of ​​*Ulva prolifera* as monitored in real time. This represents the coverage area of ​​the current path.

[0135] The system dynamically adjusts condition judgments based on real-time monitoring data, mainly including two conditions:

[0136] 1) Condition 1: Missed Coverage Ratio Is it greater than 10% (Example scenario: The newly added cleaning area is not covered by the original path, and the number of covered cleaning grids is less than 90% of the total core grids).

[0137] 2) Condition 2: Whether the radar detects new high-threat obstacles, and the minimum distance between the new obstacle and the current path is <50m (Example scenario: newly discovered passing ships / reefs, exceeding the safe distance threshold).

[0138] When any condition is met, the path replanning process is triggered: This method re-invokes the Dijkstra algorithm to generate a new initial path based on the updated dataset (which must cover ≥90% of the core area and avoid obstacles ≥50m). Then, the SO algorithm is started for iterative optimization (iterative optimization of the total fitness function F), and the adjusted optimal path is quickly output. The unmanned equipment executes the new path. If the adjustment conditions are not met, the original optimal path is continued to be executed, and real-time monitoring and feedback are continuously performed.

[0139] The globally optimal path is smoothed (the effect is similar to...). Figure 7 (As shown). Bézier curves are used to fit the planned discrete path nodes, as shown... Figure 7 The comparison chart of the smoothing effect of the Bézier curve path (30×30 map) shows that the smoothed Bézier path (represented by solid lines in the figure) is more continuous and smoother than the original discrete path (represented by dashed lines and discrete path nodes in the figure), which is more in line with the kinematic characteristics of unmanned equipment and facilitates smooth tracking execution from the starting point to the end point.

[0140] Meanwhile, if the density of seaweed reaches the disaster warning threshold (e.g., 10 kg / m²), an early warning signal will be sent to the monitoring center while the path adjustment is output, thus buying time for emergency response.

[0141] In this embodiment, the above-described scheme innovatively combines the Dijkstra algorithm with the SO algorithm, and optimizes the multi-objective function by introducing an obstacle avoidance penalty term. This ensures the rationality of the initial path and achieves a balance between coverage, energy consumption, adaptability, and safety of the global path, effectively solving the problem that traditional single algorithms cannot take into account multiple dimensions of objectives. Compared with existing path planning methods, the advantages of this scheme are mainly reflected in three aspects: First, through refined cost modeling and obstacle avoidance penalty mechanisms, the path can accurately cover the core area of ​​Ulva prolifera while strictly avoiding obstacles, significantly improving basic reliability; second, the iterative optimization of the SO algorithm achieves a balance of multiple objectives, resulting in higher path execution efficiency; and third, the dynamic adjustment mechanism can respond to environmental changes in real time, avoiding monitoring lag and ensuring the continuity and accuracy of tracking and monitoring tasks.

[0142] In response to the complexity of the marine environment and the dynamic nature of the spread of seaweed, this solution constructs a complete link of data acquisition, planning, execution, feedback, and adjustment through a closed-loop design of the entire process. This not only ensures the scientific nature of the path planning but also enhances the flexibility of practical applications, making it widely adaptable to the intelligent unmanned tracking and monitoring needs of seaweed in different scenarios such as nearshore areas, bays, and estuaries.

[0143] Example 2

[0144] like Figure 8 As shown, this embodiment provides a hybrid intelligent path planning system for tracking and monitoring seaweed, including:

[0145] The data acquisition module is used to acquire data on the distribution of *Ulva prolifera*, marine environmental data, and the status of unmanned equipment in the monitoring area.

[0146] A data preprocessing module is used to preprocess the data to obtain a standardized input dataset;

[0147] The initial path generation module is used to construct a rasterized cost model based on the input dataset, using the density of seaweed, the distance to obstacles, and the navigation energy consumption as the raster cost function, and using the Dijkstra algorithm to generate an initial path that satisfies the constraints of covering the core area of ​​seaweed, avoiding obstacles, and controlling basic energy consumption.

[0148] The global path optimization module is used to take the initial path as the initial solution of the snake optimization algorithm, and to perform iterative optimization with monitoring coverage, total device energy consumption, path matching degree with the spread trend of Ulva prolifera and obstacle avoidance safety as multi-objective optimization functions to obtain the global optimal path.

[0149] The path output control module is used to smooth the globally optimal path; it receives field data from the unmanned equipment in real time, determines whether the replanning conditions are met based on the real-time monitoring data, and re-triggers the initial path generation and global path optimization process when the conditions are met, and finally converts the planned path into control commands and sends them to the unmanned equipment.

[0150] Example 3

[0151] Embodiment 3 of the present invention provides an electronic device.

[0152] An electronic device includes a memory, a processor, and a program stored in the memory and running on the processor. When the processor executes the program, it implements the steps in the hybrid intelligent path planning method for tracking and monitoring seaweed as described in Embodiment 1 of the present invention.

[0153] The detailed steps are the same as those of the hybrid intelligent path planning method for tracking and monitoring seaweed provided in Example 1, and will not be repeated here.

[0154] Example 4

[0155] Embodiment 4 of the present invention provides a computer-readable storage medium.

[0156] A computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps of the hybrid intelligent path planning method for tracking and monitoring seaweed as described in Embodiment 1 of the present invention.

[0157] The detailed steps are the same as those of the hybrid intelligent path planning method for tracking and monitoring seaweed provided in Example 1, and will not be repeated here.

[0158] Example 5

[0159] Embodiment 5 of the present invention provides a computer program product.

[0160] A computer program product includes software code, wherein the program in the software code performs the steps of the hybrid intelligent path planning method for tracking and monitoring seaweed as described in Embodiment 1 of the present invention.

[0161] The detailed steps are the same as those of the hybrid intelligent path planning method for tracking and monitoring seaweed provided in Example 1, and will not be repeated here.

[0162] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of the present invention can be implemented using various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0163] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0164] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0165] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0166] The above description is merely a preferred embodiment of this practice and is not intended to limit the scope of this practice. Various modifications and variations can be made to this practice by those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of this practice should be included within the protection scope of this practice.

Claims

1. A hybrid intelligent path planning method for Enteromorpha tracking monitoring, characterized in that, The method comprises the following steps: acquiring green tide distribution data, marine environment data and unmanned device state data of a monitoring area; preprocessing the data to obtain a standardized input data set; based on the input data set, constructing a rasterization cost model, taking green tide density, obstacle distance and navigation energy consumption as the raster cost function, and using Dijkstra algorithm to generate an initial path that meets the constraints of covering the core area of green tide, avoiding obstacles and controlling basic energy consumption; taking the initial path as the initial solution of the snake optimization algorithm, taking monitoring coverage, total energy consumption of the device, path matching degree with green tide diffusion trend and obstacle avoidance safety as the multi-objective optimization function, and iteratively optimizing to obtain a globally optimal path; smoothing the globally optimal path; real-time receiving of on-site data feedback by the unmanned device, judgment of whether the real-time monitoring data meet the re-planning conditions, and re-triggering of the initial path generation and global path optimization processes when the conditions are met, and finally conversion of the planned path into control instructions and sending to the unmanned device.

2. The method of claim 1, wherein, The preprocessing step comprises: using an outlier detection method to eliminate invalid data; format standardization of data from different sources, unification of geographic coordinate systems and numerical precision; dimensionality reduction of high-dimensional data to extract key features, and weight assignment to each feature.

3. The method of claim 1, wherein, The construction of the rasterization cost model comprises: dividing the monitoring area into uniform grids; 4. The method of claim 1, wherein, the grid at least includes three types of empty land, green tide distribution area and obstacle area; the cost value of each grid is calculated by weighted summation of the cost items based on green tide density, distance to obstacles and navigation energy consumption. The iterative optimization process of the snake optimization algorithm comprises: initializing the population, taking the key nodes of the initial path as the initial particle swarm of the snake optimization algorithm; constructing a multi-objective fitness function, which at least includes any one or more of the following indicators: monitoring coverage, device energy consumption, path matching degree with green tide diffusion trend, obstacle avoidance safety and path smoothness; calculating the fitness of each individual in the population; determining the particle motion mode according to the iteration state parameters, and adjusting the particle position according to the corresponding position update formula; 5. The method of claim 4, wherein, updating the global optimal solution and individual optimal solution; when the iteration termination condition is met, outputting the path with the highest fitness as the global optimal path. The determination of the particle motion mode according to the iteration state parameters comprises:

6. The method of claim 1, wherein, based on the temperature parameter and the food index, the particle switches between the foraging mode of exploring new paths, the combat mode of approaching the global optimal particle, and the mating mode of fusing the advantages of different groups of particles. The method further comprises: periodic feedback of on-site data by the unmanned device; 7. A hybrid intelligent path planning system for Entermorpha tracking monitoring, characterized in that, when the green tide density reaches the preset disaster warning threshold, generating an adjusted path while sending warning information to the monitoring center. The method comprises the following steps: a data acquisition module for acquiring green tide distribution data, marine environment data and unmanned device state data of a monitoring area; a data preprocessing module for preprocessing the data to obtain a standardized input data set; An initial path generation module is configured to construct a rasterized cost model based on the input data set, to take the Enteromorpha density, obstacle distance and navigation energy consumption as the raster cost function, and to generate an initial path that meets the constraints of covering the core area of Enteromorpha, avoiding obstacles and controlling the basic energy consumption by using the Dijkstra algorithm. A global path optimization module is configured to take the initial path as the initial solution of the snake optimization algorithm, to take the coverage rate, total energy consumption of the device, matching degree of the path and the Enteromorpha diffusion trend, and obstacle avoidance safety as the multi-objective optimization function, to perform iterative optimization, and to obtain a globally optimal path. A path output control module is configured to perform smoothing processing on the globally optimal path, to receive real-time field data fed back by the unmanned device, to determine whether the re-planning condition is met based on the real-time monitoring data, and to re-trigger the initial path generation and global path optimization processes when the condition is met, and finally to convert the planned path into a control instruction and send it to the unmanned device.

8. An electronic device comprising a memory, a processor, and a computer program stored on the memory and running on the processor, characterized in that, The processor executes the program to implement the steps of the hybrid intelligent path planning method for Enteromorpha tracking and monitoring according to any one of claims 1 to 6.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the steps of the hybrid intelligent path planning method for Enteromorpha tracking and monitoring according to any one of claims 1 to 6.

10. A computer program product comprising software code, characterized in that, The program in the software code performs the steps of the hybrid intelligent path planning method for Enteromorpha tracking and monitoring according to any one of claims 1 to 6.

Images