Intelligent Profiling Sensing Method and System for Underwater Algal Communities

Abstract

This invention relates to the field of underwater ecological environment monitoring technology, specifically disclosing an intelligent profile sensing method and system for underwater algal communities. The invention first dynamically divides non-uniform sensing node clusters based on the correlation of multi-source data; then, it uses a double-layer nested judgment to screen highly reliable data, first verifying the consistency within the cluster, and then comparing it with physical model predictions; next, it uses time-series data to construct metabolic relationship chains, and optimizes them in two stages to make their structure simple and functionally robust; finally, it maps the metabolic chains back to space, generates the optimal sensing path under five-dimensional causal constraints, and executes it after collaborative verification at three levels: spatial, data, and ecological. This invention achieves intelligent, accurate, and robust profile sensing and inference of the spatial distribution, metabolic state, and ecological function of underwater algal communities.

Description

Technical Field

[0001] This invention relates to the field of underwater ecological environment monitoring technology, specifically to an intelligent profile sensing method and system for underwater algal communities. Background Technology

[0002] The distribution and metabolic activity of underwater algal communities are important indicators reflecting the ecological health of aquatic bodies. Accurate profiling and sensing of these communities helps assess water quality changes, predict algal blooms, and formulate ecological restoration strategies. Traditional underwater algal monitoring mainly relies on fixed-point sampling and laboratory analysis. This method is time-consuming, labor-intensive, has limited spatial coverage, and struggles to achieve real-time, continuous, and large-scale vertical profiling monitoring. In recent years, monitoring methods based on sensor networks and remote sensing technologies have been gradually promoted, but problems remain, such as limited data sources, poor environmental adaptability, and insufficient model interpretability. Especially in complex underwater environments, strong data heterogeneity and significant noise interference lead to low reliability of the sensing results.

[0003] In existing technologies, common underwater sensing methods often employ uniform data collection points or fixed profile paths, neglecting the spatiotemporal heterogeneity of the underwater environment and failing to adaptively adjust the layout of sensing nodes, resulting in insufficient coverage of key ecological areas. Furthermore, most methods rely on a single algorithm or model for data processing, lacking multi-model collaborative verification and conflict resolution mechanisms, making them prone to outputting erroneous conclusions due to model bias or environmental interference. Regarding metabolic relationship inference, existing methods often overlook the coupling effect between algal physiological mechanisms and the physical environment, leading to unstable and poorly robust metabolic chain topologies that are unsuitable for supporting long-term ecological prediction and pathway planning.

[0004] Therefore, there is an urgent need in this field for an underwater algal community profile sensing method that can adapt to dynamic changes in the underwater environment, integrate multi-source heterogeneous data, and has the ability to perform model cross-validation and intelligent optimization, so as to achieve high-precision and high-reliability vertical profile sensing and ecological status assessment. Summary of the Invention

[0005] To overcome the problems of existing technologies, such as difficulty in accurately and automatically distinguishing and quantifying the vertical distribution of different algal phyla in water bodies, susceptibility to environmental interference, and poor model universality, this invention provides an intelligent profile perception method and system for underwater algal communities. Through technical solutions such as non-uniform perception node cluster division, double-layer nested judgment mechanism, construction and optimization of metabolic relationship chains driven by time-series biochemical dynamics, generation of five-dimensional causal constraint paths, and multi-layer collaborative verification, it achieves intelligent, accurate, and robust profile perception and inference of the spatial distribution, metabolic state, and ecological function of underwater algal communities.

[0006] The technical solution of this application specifically includes: One aspect of this application provides a method for intelligent profile sensing of underwater algal communities, comprising: Based on the correlation of multi-source heterogeneous data of the original underwater environment, the underwater vertical profile space is divided into non-uniform sensing node clusters. Sensing data is collected through the cluster of sensing nodes, and the sensing data is input in parallel into the cluster model composed of clustering algorithms and the dissident model constructed based on physical and physiological mechanisms, and a double-layer nested judgment is performed. Using the sensory data judged through double-layer nesting, a primary metabolic relationship chain is constructed by a multi-source data coupling inference method driven by time-series biochemical dynamics. The initial metabolic relationship chain is then optimized by a two-stage optimization model to obtain the final metabolic relationship chain. The optimal profile perception path is generated by solving the final metabolic relationship chain through five-dimensional causal constraints. Before output, the spatial layer, data layer and ecological layer collaborative verification mechanism is forcibly activated. The path is only unlocked and executed after the collaborative verification mechanism is passed.

[0007] As a further option of the method of the present invention, the sensing node cluster partitioning step includes: For any two spatial nodes, calculate the temporal correlation coefficient within the sliding time window for each type of observation data; Based on the temporal correlation coefficient, the comprehensive correlation similarity between nodes is calculated. The comprehensive correlation similarity is obtained by weighting and summing the temporal correlation coefficients of multiple observation data according to ecological importance weights, and then inputting them into the hyperbolic tangent function for calculation. A symmetric comprehensive similarity matrix is ​​constructed based on the comprehensive correlation similarity among all node pairs; Based on the comprehensive similarity matrix, construct the degree matrix and the normalized Laplacian matrix, solve for the eigenvectors corresponding to the smallest eigenvalues ​​of the Laplacian matrix, and form the eigenvector matrix. The three-dimensional spatial coordinate vectors of the nodes are concatenated to the corresponding rows of the feature vector matrix with preset weights to form extended feature vectors; The extended feature vectors are divided using the K-means++ clustering algorithm to obtain non-uniform clusters of sensing nodes.

[0008] As a further option of the method of the present invention, the first-level judgment logic in the double-layer nested judgment is as follows: Trigger the sensing node cluster to collect data and extract the core observation data of algal biomass collected by all nodes in the sensing node cluster at the current moment; Calculate the arithmetic mean of the core algal biomass observation data, and calculate the discrete value representing the consistency of data within the cluster according to the discrete value calculation formula; wherein, the discrete value calculation formula is: ;in, For sensing node clusters Discrete value at the current moment , For sensing node clusters Number of nodes internally transmitting effective biomass data. For sensing node clusters All effective biomass data within, , For all effective biomass data arithmetic mean ; The discrete value is compared with a preset consistency threshold. If the discrete value is less than the consistency threshold, the data within the sensing node cluster is determined to be highly consistent, and the second-level judgment logic is entered. If the discrete value is greater than or equal to the consistency threshold, the sensing node cluster is marked as a highly discrete cluster, and the sensing node cluster re-partitioning process is triggered.

[0009] As a further option of the method of the present invention, the second-level judgment logic in the double-nested judgment is as follows: The multidimensional observation data of the perception node cluster that passes through the first layer of judgment logic is input into the dissenter model constructed based on physical and physiological mechanisms. The dissenter model estimates the biomass baseline value predicted by the physical model based on the environmental driving factors of depth, temperature and photosynthetically effective radiation intensity, and the algal growth simulation based on light limitation and temperature regulation. The difference between the actual observed biomass of each node in the sensing node cluster and the biomass baseline value predicted by the physical model is calculated, and the normalized root mean square error is calculated as the physical prediction value based on the difference. The physical prediction value is compared with a preset conflict threshold. If the physical prediction value is less than the conflict threshold, the observation data of the sensing node cluster is determined to be high-reliability data and output. If the physical prediction value is greater than or equal to the conflict threshold, the sensing node cluster is marked as a conflict cluster and the sensing node cluster is re-divided.

[0010] As a further option of the method of the present invention, the primary metabolic relationship chain construction step includes: The cluster of perception nodes judged through double-layer nesting is regarded as an ideal mixing reaction unit, and the state vector of metabolite concentration of the reaction unit at time t is defined. Define a set of possible biochemical reactions and establish a kinetic rate equation for each reaction that depends on the current concentration of metabolites and environmental driving factors; For each metabolite, establish its stoichiometric coefficient in each reaction; Using concentration data observed at multiple consecutive time points, the kinetic parameters, stoichiometric coefficients, and transport term parameters of all reactions are estimated through a global optimization algorithm. Reactions with non-zero significant fluxes and their interconnections form a primary metabolic relationship chain.

[0011] As a further option of the method of the present invention, the first stage optimization step of the two-stage optimization model includes: The primary metabolic relationship chain is represented as a directed weighted graph, and the topological entropy of the network is calculated based on the normalized edge weights. Construct and solve an optimization problem with the objective function of minimizing topological entropy and the constraint that the fitting error of the optimized network to the observed data does not exceed a first preset threshold. A heuristic search algorithm is used to solve the problem and output the intermediate network that satisfies the fitting error constraint and has the lowest topological entropy. The intermediate network has fewer connections than the primary metabolic relation chain.

[0012] As a further option of the method of the present invention, the second-stage optimization step of the two-stage optimization model includes: Define a functional robustness metric for the network, which is evaluated by simulating the average size of the largest connected subgraph of the remaining network after a predetermined proportion of edges are randomly removed from the network. An optimization problem is constructed and solved with the objective function of maximizing the functional robustness index and the constraint that the network is obtained by adjusting the local structure of the intermediate network and the fitting error does not exceed the second preset threshold. The local structure adjustment includes adding alternative conversion paths between key metabolic nodes, enhancing the feedback regulation connection in the existing metabolic cycle, or adjusting the edge weight ratio. An evolutionary algorithm is used to solve the problem and output the final metabolic relationship chain with the highest functional robustness.

[0013] As a further option of the method of the present invention, the step of generating the optimal profile sensing path by solving the final metabolic relationship chain through five-dimensional causal constraints includes: The final metabolic relationship chain is mapped back to the three-dimensional water body profile space, metabolically active areas are identified, and a set of candidate paths from the start point to the end point is generated in the three-dimensional spatial grid. For each path in the candidate path set, an objective function for a multi-objective optimization problem is constructed, which includes maximizing spatial coverage, maximizing data information gain, maximizing metabolic flux continuity, minimizing energy consumption, and minimizing ecological disturbance. The Pareto optimal solution set is obtained by solving the objective function of the multi-objective optimization problem using a multi-objective evolutionary algorithm. The path is selected from the Pareto optimal solution set as the optimal profile sensing path.

[0014] As a further option of the method of the present invention, the forced activation of the spatial layer-data layer-ecological layer collaborative verification mechanism includes: Spatial layer verification: Calculate the minimum distance from all points on the optimal profile sensing path to the geometric center of each sensing node cluster or the position of all nodes. If the minimum distance corresponding to a sensing node cluster is greater than the preset spatial coverage radius threshold, the spatial layer verification is deemed to have failed. Data layer verification: Based on the dissenter model, the virtual sampling point data along the optimal profile sensing path is predicted, the virtual sampling point data is allocated to the sensing node cluster, and the double-layer nested judgment logic is simulated. If the proportion of data passing the double-layer nested judgment logic is lower than the preset threshold, the data layer verification is judged to have failed. Ecosystem layer verification: Calculate the morphological similarity between the depth-attribute profile of the optimal profile sensing path and the typical metabolic profile inferred from the final metabolic relationship chain. If the morphological similarity is lower than the threshold, the ecosystem layer verification is deemed to have failed. The optimal profile sensing path is unlocked only after all three layers of verification—spatial layer, data layer, and ecosystem layer—have passed.

[0015] Another aspect of this application provides an intelligent underwater algal community profile sensing system, comprising: The node cluster partitioning module is used to divide the underwater vertical profile space into non-uniform sensing node clusters based on the correlation of multi-source heterogeneous data of the original underwater environment. The double-nested judgment module is used to collect sensing data through the sensing node cluster and input the sensing data in parallel into the cluster model composed of clustering algorithm and the dissident model constructed based on physical and physiological mechanisms to perform double-nested judgment. The metabolic relationship chain construction and optimization module is used to construct a primary metabolic relationship chain using the sensory data judged through double-layer nesting, and adopt a multi-source data coupling inference method driven by time-series biochemical dynamics. The primary metabolic relationship chain is then optimized using a two-stage optimization model to obtain the final metabolic relationship chain. The path generation and verification module is used to generate the optimal profile sensing path by solving the final metabolic relationship chain through five-dimensional causal constraints. Before output, the spatial layer-data layer-ecological layer collaborative verification mechanism is forcibly started. The optimal profile sensing path is only unlocked and executed after the collaborative verification mechanism passes.

[0016] The beneficial effects of this application are as follows: This invention significantly improves the targeting and reliability of data acquisition in complex underwater environments through a dynamically adaptive non-uniform sensing node layout and a dual verification mechanism, increasing the confidence level of measured data by more than 30%. Based on a metabolic network inference technology using temporal dynamic coupling and dual optimization, it constructs an ecological model with both high interpretability and strong stability, reducing network topology complexity by about 40% and enhancing anti-interference capability by 25%. The final generated sensing path deeply integrates ecological mechanisms, and through multi-objective optimization and collaborative verification, it can improve information acquisition efficiency by about 20% while ensuring comprehensive coverage of key ecological areas and minimizing ecological disturbance, thus achieving adaptive and highly interpretable continuous monitoring of underwater algal community profiles. Attached Figure Description

[0017] Figure 1 A schematic diagram of the intelligent profile sensing method for underwater algal communities; Figure 2 S100 flowchart of intelligent profile sensing method for underwater algal communities; Figure 3 S200 flowchart of intelligent profile sensing method for underwater algal communities; Figure 4 Flowchart of the S300 intelligent profile sensing method for underwater algal communities; Figure 5 Flowchart of the S400 intelligent profile sensing method for underwater algal communities. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] Profiling underwater algal communities is a core task in aquatic ecological environment monitoring, water quality assessment, and ecological model construction. Traditional methods rely on discrete point sampling or sensors with uniform grid deployment, which struggle to effectively capture the non-uniform distribution characteristics of algae in vertical profiles driven by light, temperature, and nutrient gradients. Fusion of multi-source heterogeneous data typically employs simple weighted averaging or statistical interpolation, lacking coupled modeling of the interaction between algal physiological and ecological mechanisms and the physical environment. Existing methods for inferring metabolic relationships are mostly based on static correlation analysis, failing to reflect the temporal dynamics of biochemical reactions, and the sensing path planning is often disconnected from the inferred ecological processes, resulting in low sensing efficiency and poor ecological interpretability.

[0020] The theoretical foundation of this invention is built upon four pillars: non-uniform sensing topology theory of underwater vertical profile space, double-layer nested judgment theory based on physical-physiological coupling, metabolic network optimization theory driven by temporal biochemical dynamics, and path generation and collaborative verification theory under causal constraints. By constructing a dynamic non-uniform sensing node cluster, performing double-layer nested judgment to filter high-quality sensing data, driving the construction and optimization of metabolic relationship chains, and finally generating verifiable optimal sensing paths under multiple causal constraints, intelligent, accurate, and robust profile sensing of the structure, function, and dynamics of underwater algal communities is achieved.

[0021] The derivation of the core theoretical formula is as follows: It has There are 1 initial sensing node, each node At any moment Collected Heterogeneous observation data constitutes the observation vector. To partition non-uniform sensing node clusters, nodes are defined. and Comprehensive correlation and similarity between : ;in, For the first The preset weights of the observation data satisfy ; For the first Class data in time window The temporal correlation coefficient within; This refers to the Sigmoid function. Based on the similarity matrix. The space is divided into spectral clustering algorithms. A cluster of non-uniform sensing nodes .

[0022] In the double-nested judgment mechanism, the first-level judgment calculation cluster Discrete values ​​of algal biomass observation data Suppose that the cluster contains Each node, with biomass data as follows: Its mean is Then discrete value for: Set a consistency threshold .like If so, proceed to the second level of judgment.

[0023] The second layer of judgment is based on the dissenter model to calculate the physical prediction value. The dissenter model integrates physical mechanisms to predict theoretical biomass distribution. Physical prediction value Defined as the normalized root mean square error between observed and predicted values. A conflict threshold is set. .like If the data is deemed reliable, it is output; otherwise, the perception node cluster is re-divided.

[0024] In the metabolic relation chain construction and optimization stage, an initial metabolic relation chain is first constructed based on time-series data. Then, a two-stage optimization is performed. The first stage aims to minimize the network topology entropy. The network is defined... The adjacency matrix is Its elements Indicates from matter arrive Connection strength. Normalized edge weights. Network structure entropy for: ;in, This represents the total number of metabolic substances. The first stage of optimization solution... The constraints are The fitting error of the observed data does not exceed the threshold. To obtain the intermediate network .

[0025] The second stage focuses on the robustness of metabolic substance cycling. Maximize robustness. The evaluation is performed by simulating the average size of the largest connected subgraph in the remaining network after randomly removing a certain proportion of edges from the network. The second stage of optimization is then solved. The constraints are Depend on It was obtained through local structural adjustments, and the fitting error does not exceed the threshold. Ultimately, the final metabolic relationship chain is obtained. .

[0026] During the path generation and verification phase, Mapping back to space, we define a multi-objective optimization problem under five-dimensional causal constraints to generate the optimal profile-aware path. Objective function for: ;in, For spatial coverage, For data information gain, For the continuity of metabolic flux. For energy consumption, This constitutes ecological disturbance. , , , , These are the weighting coefficients. They are obtained by solving a multi-objective evolutionary algorithm. .path Before execution, mandatory verification must be performed through three layers: spatial layer, data layer, and ecological layer.

[0027] The specific embodiments of the present invention will be described in detail below.

[0028] Example 1: Please see Figure 1 The diagram illustrates a flowchart of an intelligent underwater algal community profiling sensing method according to an embodiment of the present invention, the method comprising: S100: Based on the correlation of multi-source heterogeneous data of the original underwater environment, the underwater vertical profile space is divided into non-uniform sensing node clusters.

[0029] S200: Sensing data is collected through the cluster of sensing nodes. The sensing data is input in parallel into the cluster model composed of clustering algorithms and the dissident model constructed based on physical and physiological mechanisms, and a double-layer nested judgment is performed.

[0030] S300: Using the sensory data judged through double-layer nesting, a primary metabolic relationship chain is constructed by a multi-source data coupling inference method driven by time-series biochemical dynamics. The initial metabolic relationship chain is then optimized by a two-stage optimization model to obtain the final metabolic relationship chain.

[0031] S400: The final metabolic relationship chain is solved by five-dimensional causal constraints to generate the optimal profile perception path, and the spatial layer-data layer-ecological layer collaborative verification mechanism is forcibly started before output. The path is only unlocked and executed after the collaborative verification mechanism is passed.

[0032] The specific plan is as follows: In an intelligent underwater algal community profiling and sensing method, S100 dynamically constructs a non-uniform sensing topology by analyzing the spatiotemporal correlation patterns of historical or real-time multi-source heterogeneous data, laying the foundation for refined data acquisition.

[0033] Please refer to Figure 2 The diagram illustrates a flowchart of an exemplary intelligent profile sensing method for underwater algal communities, S100, which includes: S110: In this embodiment, a node array containing various sensors is deployed within the vertical profile space of the water target. The nodes are distributed in three dimensions in space. Each node is equipped with at least the following sensor types: chlorophyll fluorescence sensor, dissolved oxygen sensor, water temperature sensor, pH sensor, and nutrient sensor. All sensors periodically acquire data according to a unified time synchronization protocol, and each data record includes the measured value, three-dimensional spatial coordinates, and timestamp.

[0034] In one possible implementation, the sensor node array is deployed via moored profiling buoys, underwater gliders, or fixed mooring platforms. The deployment depth should cover key water layers from the surface to the bottom of the euphotic zone or below the thermocline. Data acquisition frequency is set according to the timescale of algal community changes.

[0035] S120: In this embodiment, the acquired raw time-series data is preprocessed. The preprocessing operation includes removing outliers and missing values ​​caused by sensor failure or communication interruption, and normalizing the observation data of different dimensions to make the data of each dimension fall within a similar numerical range.

[0036] For any two spatial nodes and For each type of observation data Calculate the sliding time window formed by multiple consecutive sampling times. Pearson correlation coefficient within .

[0037] In one possible implementation, the overall similarity between nodes is... The calculation uses the following formula: ;in, For the first Ecological importance weights for observational data. This is the scaling adjustment parameter. Given the hyperbolic tangent function. Calculate the tangent between all node pairs sequentially. Ultimately forming a Symmetric composite similarity matrix .

[0038] S130: In this embodiment, a comprehensive similarity matrix is ​​used. As input, a spectral clustering algorithm incorporating spatial distance constraints is employed to cluster the entire profile space. Each sensing node is divided into Clusters.

[0039] In one possible implementation, the steps for partitioning a non-uniform sensing node cluster are as follows: Based on the similarity matrix Construct degree matrix Degree matrix It is a diagonal matrix, and its first... diagonal elements Equal to the similarity matrix No. The sum of all elements in the row, i.e. .

[0040] Calculate the normalized Laplacian matrix . The calculation formula is: ,in yes An identity matrix of order 1.

[0041] Solving the Laplace matrix The former The smallest eigenvalue and its corresponding eigenvector. The value of can be determined by observing the inflection point on the characteristic value distribution curve. (The preceding...) The eigenvectors are arranged column-wise to form an eigenvector matrix. .

[0042] For the eigenvector matrix each row Norm normalization yields a new matrix. .

[0043] The 3D spatial coordinate information of the nodes is embedded as a constraint. The normalized spatial coordinate vector is then used as the constraint. By weight spliced ​​into a matrix The After the rows are filled, an extended feature vector is formed. .

[0044] For extended feature vectors The dataset was divided using the K-means++ clustering algorithm, aggregating the nodes into... Cluster .

[0045] After clustering is completed, each cluster of sensing nodes The nodes within the clusters exhibit high similarity in multidimensional data features and are relatively clustered in spatial distribution; however, nodes in different clusters show significant differences in data patterns or spatial locations. The resulting partitioning forms a non-uniform, data-driven sensing topology for the underwater profile space.

[0046] In an intelligent profile sensing method for underwater algal communities, S200 uses the sensing node clusters divided by S100 as the basic unit to synchronously collect data and execute a strict double-layer nested judgment logic to filter out reliable sensing data with high consistency and low cognitive conflict.

[0047] Please refer to Figure 3 The diagram illustrates a flowchart of an exemplary intelligent profile sensing method for underwater algal communities, S200, which includes: S210: In this embodiment, for all the divided clusters of sensing nodes... arrive Send a synchronization acquisition command. Upon receiving the command, all sensor nodes within each cluster will simultaneously... Initiate the data acquisition process to obtain complete observational data, including algal biomass indicators and related environmental parameters. .

[0048] S220: In this embodiment, for each cluster of sensing nodes Extract all nodes at the current time. Core observational data of algal biomass were collected. Discrete value indices characterizing intra-cluster data consistency were calculated based on the core observational data.

[0049] In one possible implementation, the first-layer decision logic targets a cluster of sensing nodes. The execution includes the following steps: Statistical clusters The number of nodes that internally transmit effective biomass data is denoted as A minimum effective node count threshold is preset. .like If the data collected by the sensing node cluster is invalid, it will be determined that the data collected in this instance is invalid.

[0050] Computational clusters All effective biomass data arithmetic mean .

[0051] According to the discrete value calculation formula Computational clusters Discrete value at the current moment .

[0052] The calculated discrete values Consistency threshold with preset Compare them.

[0053] Make a consistent decision: if Then determine the cluster of sensing nodes. The internal data exhibits high consistency, allowing for the entry of the second-level decision-making logic. If If the data dispersion within a cluster is too large, the system will immediately label the cluster. It is a highly discrete cluster, and triggers a process of re-partitioning the sensing node clusters in the spatial region where the cluster is located.

[0054] S230: In this embodiment, for all clusters of sensing nodes that pass the first-layer consistency check, their multidimensional observation datasets are input into a dissenter model pre-constructed based on physical and physiological mechanisms. The dissenter model uses known physical laws and basic physiological responses of algae to simulate and calculate the spatial distribution of algal biomass that should be exhibited under current environmental conditions.

[0055] In one possible implementation, the second-layer judgment logic targets the cluster of perception nodes that have passed the first-layer judgment. The execution includes the following steps: Dissident model reading cluster Environmental driving factor data for all nodes, primarily including the depth of each node. ,temperature and the photosynthetically active radiation intensity at that depth layer .

[0056] The core of the model is based on algal growth simulation under light limitation and temperature regulation, estimating the baseline biomass value predicted by the physical model at each node location. .

[0057] For clusters For each node within the range, calculate its actual observed biomass. Compared with the physical model predictions The differences between them.

[0058] Computational clusters Overall physical prediction values This is used to quantify the degree of conflict between observations and predictions based on physical mechanisms. The normalized root mean square error is used for calculation.

[0059] The calculated physical prediction values conflict threshold with preset Compare them.

[0060] Make conflict resolution: if Then determine the current sensing cluster The observational data largely matched the expectations based on physical and physiological mechanisms, and this cluster of data was marked as high-reliability data, allowing it to be output to the subsequent S300 metabolic relationship chain construction stage. If so, the perceptual region is determined to be in a state of cognitive conflict. The labeled clusters... It forms a conflict cluster and triggers a re-partitioning of the perception node cluster.

[0061] In an intelligent underwater algal community profiling and sensing method, S300 uses highly reliable sensing data screened by S200 to drive time-series biochemical dynamics processes, infer the metabolic interaction network within the algal community, and optimizes it in two stages to make it structurally simple and functionally robust.

[0062] Please refer to Figure 4 The diagram illustrates a flowchart of an exemplary intelligent profile sensing method S300 for underwater algal communities, which includes: S310: In this embodiment, reliable observation data from all sensing node clusters determined through double-layer nesting are integrated over multiple consecutive time steps. The reliable observation data focuses on the time series of predefined key metabolite concentrations, which typically include dissolved oxygen (O2), dissolved inorganic carbon (DIC), nitrate (NO3), phosphate (PO4), etc.

[0063] In one possible implementation, the steps of constructing the primary metabolic relationship chain include: Each cluster of sensing nodes The reaction unit is considered as an ideal mixture. The reaction unit is defined at time [time]. The state vector of metabolite concentration is .

[0064] Definition includes A collection of possible biochemical reactions. Biochemical reactions encompass the main metabolic pathways of algal communities.

[0065] For each reaction Establish the dynamic rate equation Rate equations describe how the reaction rate depends on the current concentration of the metabolite. and environmental driving factors .

[0066] For each metabolic substance Establish its stoichiometric coefficient within the reaction unit. , It is matter In the reaction The stoichiometric coefficients in the equation.

[0067] Utilize each cluster Concentration data observed at multiple time points were used to estimate all reactions using a global optimization algorithm. The kinetic parameters, stoichiometric coefficients, and transport term parameters.

[0068] After the parameter estimation process is completed, the response with non-zero significant flux will be... The connections between these connections constitute the topology of the metabolic network, namely the primary metabolic relation chain. .

[0069] S320: In this embodiment, the primary metabolic relationship chain The network structure may be complex, containing too many connections, which is not conducive to ecological interpretation. The goal of this stage of optimization is to simplify the network structure.

[0070] In one possible implementation, the execution steps of the first-stage optimization are as follows: Primary metabolic relationship chain This is represented as a directed weighted graph. Nodes in the graph represent various metabolic substances. Edges represent the transformation or influence relationships between substances, and the weights of the edges are... From the estimated stoichiometric coefficients The results were derived by combining the typical reaction rates.

[0071] Calculate the current network structural entropy Structural entropy According to the formula calculate.

[0072] Construct and solve a constrained optimization problem where the decision variables are the network connection structure and the objective function is... There are two constraints: First, the new network... It must stem from the understanding of The first modification; the second, using a simplified network. The fitting error of the refitted observation data must not exceed [a certain value]. Fitting error times.

[0073] A heuristic search algorithm is used to solve the problem. When the algorithm reaches the termination condition, the final network structure is output, which is denoted as the intermediate optimized network. . compared to With fewer or clearer connections, structural entropy is significantly reduced while maintaining core data interpretability.

[0074] S330: In this embodiment, a healthy metabolic network should have a certain degree of functional redundancy or robustness, meaning that the overall material cycle can still be maintained when some reaction pathways are inhibited. The goal of this stage of optimization is to enhance the intermediate network. Functional robustness.

[0075] In one possible implementation, the execution steps of the second-stage optimization are as follows: Define and compute networks Functional robustness index The calculation was performed using a random edge attack simulation method.

[0076] Construct and solve the second constrained optimization problem. The decision variables are... Based on this, limited local structural adjustments are made. The objective function is... The constraint is: the adjusted network The fitting error of the observed data must not exceed Fitting error times.

[0077] Permitted structural adjustments are designed to enhance robustness and primarily include: adding alternative conversion pathways between key metabolic nodes; strengthening feedback regulatory connections in existing metabolic cycles; or adjusting the edge weight ratio between core and non-core pathways.

[0078] Evolutionary algorithms, such as genetic algorithms, are used to solve the problem. After the evolution reaches a preset number of generations, a selection is made from the final population. The highest-performing network individual that satisfies the error constraint serves as the final metabolic relation chain. . It can not only reasonably interpret observational data, but also has a simple topological structure and enhanced functional robustness, and is more likely to reflect the intrinsic and stable core metabolic interaction patterns of algal communities.

[0079] In an intelligent profiling method for underwater algal communities, S400 converts the final metabolic relationship chain obtained by S300. By combining spatial information, the optimal profile perception path is generated under multiple causal constraints. It must be executed only after rigorous three-layer collaborative verification to ensure the physical feasibility, information value and ecological rationality of the path.

[0080] Please refer to Figure 5 The diagram illustrates a flowchart of an exemplary underwater algal community intelligent profile sensing method S400, which includes: S410: In this embodiment, the abstract final metabolic relationship chain is... Remapping back to a specific three-dimensional water body profile space. Analysis The main sources, sinks, and conversion hotspots of mass flux in the system were identified and correlated with the spatial patterns observed by sensors, thus identifying spatial regions that are crucial for understanding the metabolic state of the entire system.

[0081] In one possible implementation, the step of generating a candidate path set includes: according to The intensity of the inferred key biochemical processes, combined with the observation data of each sensing node cluster, is used to mark metabolically active regions on a three-dimensional spatial grid.

[0082] Define the starting point of the profile sensing task and the end point The starting and ending points are usually set as a point on the water surface and a point on the bottom of the water.

[0083] The three-dimensional water space is discretized into a regular grid. Each grid cell is assigned a basic movement cost.

[0084] A graph search algorithm is used to generate graphs in the grid space from... arrive Several candidate paths are identified. The initial path search aims solely to minimize the basic movement cost, thereby quickly obtaining a set of paths covering all possible directions from the water surface to the bottom. .

[0085] S420: In this embodiment, for each path in the candidate path set The advantages and disadvantages of each path are evaluated from five different dimensions. These five dimensions reflect the various causal constraints that path generation must satisfy. Multi-objective optimization techniques are used to find the path that best balances these constraints.

[0086] In one possible implementation, the evaluation metrics across the five dimensions are calculated as follows: Spatial coverage : Measuring the path The degree of coverage of the spatial range of all current sensing node clusters.

[0087] Data information gain : Predicted path The value of information that can be obtained through sampling. Calculations are based on spatial interpolation models and their uncertainties.

[0088] Continuity of metabolic flux Evaluation Path The direction and The spatial consistency of the revealed major metabolic flux directions.

[0089] Energy consumption Estimate AUV along the path The energy required for navigation.

[0090] Ecological disturbance Evaluation Path Potential disturbances to aquatic ecosystems.

[0091] Constructing a multi-objective optimization problem: finding a path , making , and As large as possible, while making and As small as possible.

[0092] In one possible implementation, a non-dominated sorting genetic algorithm with an elitist strategy (NSGA-II) is used to solve the problem. A path is encoded as a sequence of grid coordinates. The initial population consists of the previously generated set of candidate paths. The process involves calculating the five objective function values ​​for each individual in the population during each generation. Parent individuals are selected based on non-dominated ranking and crowding distance, and offspring are produced through crossover and mutation. This iterative process drives the population towards a Pareto optimal front that balances all five objectives.

[0093] After the algorithm converges, a set of Pareto optimal solutions is obtained. Decision-makers can select a compromise solution from the Pareto solution set as the optimal profile sensing path based on the specific focus of the current task. .

[0094] S430: In this embodiment, in the path Before being finally approved and deployed to AUVs for execution, it must pass through a three-tiered, progressive verification process. Failure at any level will result in the path being abandoned. All will be locked, and the system needs to return to S410 or S420, adjust the constraint weights or regenerate the path until a path that can pass all verifications is found.

[0095] In one possible implementation, the three-layer verification steps of the collaborative verification mechanism are as follows: Spatial layer validation: Validating the optimal profile sensing path Does it physically cover all currently partitioned clusters of sensing nodes? The specific verification criterion is: for each cluster of sensing nodes… Calculation path The minimum distance from all points to the geometric center of the cluster or the location of all nodes. If a cluster exists... This ensures that the minimum distance is greater than the preset spatial coverage radius threshold. If the path does not fully cover the cluster, the spatial layer verification fails.

[0096] Data layer validation: Validating the optimal profile sensing path The question is whether the expected sampled data might satisfy the S200's two-layer nested judgment logic. This is a prospective simulation verification. Based on the currently calibrated dissident model and environmental factor field, the prediction along the path... The algal biomass and other parameter values ​​that may be observed at each planned sampling point are then analyzed. These virtual path sampling points are assigned to the nearest cluster of sensing nodes based on their spatial location. Then, the two-layer judgment logic of S200 is simulated and executed on these virtual data. The simulation judgment requires that more than a preset proportion of path sampling point data simultaneously pass both consistency and conflict checks. If the pass rate is lower than a threshold, the data layer verification fails.

[0097] Ecosystem layer validation: Validating the optimal profile sensing path Relationship with ultimate metabolism Does the revealed ecological process have a strong causal relationship? Does the assessment path traverse... The interface for inferring key metabolic state transitions can be determined by calculating the depth-attribute profile of the path. The morphological similarity between inferred typical metabolic profiles is used for quantification. If the similarity is higher than a threshold, the pathway and metabolic process are considered to have a strong causal coupling, and the ecological layer is verified.

[0098] The optimal profile sensing path will only be determined after the spatial layer, data layer, and ecosystem layer have all passed verification. It was only recently officially unlocked. Its detailed spatial coordinate sequence, navigation speed, and sampling action commands were encapsulated into a mission file and sent to the autonomous underwater vehicle's control system to initiate automated profile sensing operations. The sampling data transmitted in real-time during the AUV's mission can then be used as new input, fed back to the S100 to update the next round of non-uniform sensing node cluster partitioning. This achieves a closed-loop intelligent system of perception-judgment-inference-planning-execution-feedback, continuously optimizing the ability to perceive the dynamics of underwater algal communities.

[0099] Example 2: This invention was field-verified and applied for two weeks in the monitoring of summer cyanobacterial blooms in the Meiliang Bay waters of Taihu Lake. The complete experimental setup, operation process, and data results are as follows.

[0100] 1. Experimental environment and system configuration; The experimental area is located in the northern part of Meiliang Bay in Taihu Lake, with a water area of ​​approximately 1.5 square kilometers and an average depth of 3.5 meters. During the experiment, the weather was mainly sunny to cloudy, with distinct water temperature stratification. The surface water temperature ranged from 28 to 32°C, while the bottom water temperature was around 25°C, and persistent cyanobacterial blooms were observed.

[0101] The system adopts a heterogeneous networking mode of fixed profile monitoring platform + mobile autonomous navigation platform. The specific configuration parameters are shown in the table below: Table 1: Hardware Configuration and Parameters of the Experimental System 2. Algorithm parameter configuration; Based on historical data from the experimental area and the objectives of this experiment, the core parameters of the algorithm were calibrated as follows: Table 2: Core Algorithm Parameter Configuration Table 3. Experimental procedure and data recording; The experiment lasted from August 10th to August 24th, 2023. The system ran automatically in approximately 8-hour cycles. The following are the operation records for three typical cycles: Table 3: Intelligent Sensing Cyclic Operation Record Table 4. Performance evaluation and quantification results; To quantify the advantages of the method of this invention, it is compared with two baseline methods: (A) the traditional uniform grid profiling method (regular zigzag path); and (B) the adaptive sampling method based solely on physical model optimization. Evaluation is based on data and analysis results generated throughout the experimental period.

[0102] Table 4: Performance Comparison Evaluation Results 5. Collaborative verification mechanism intercepts records; The collaborative verification mechanism played a crucial quality gate role throughout the experiment. The following are two typical intercept records: Table 5: Interception Record Table of Collaborative Verification Mechanism 6. Conclusion; This field verification fully demonstrates that the intelligent underwater algal community profiling and sensing method and system provided by this invention can effectively integrate multi-source heterogeneous data. Through a series of innovative steps, including non-uniform cluster partitioning, double-layer nested judgment, metabolic network optimization, and causal constraint path planning, it achieves efficient, accurate, and adaptive sensing of the spatial structure, metabolic function, and dynamic changes of algal communities in complex waters. The system not only improves the targeting and intelligence of data acquisition but also elevates the sensing depth from descriptive patterns to analytical processes through metabolic relationship chain inference and collaborative verification mechanisms. This provides a powerful technical tool and data analysis framework for aquatic ecological environment monitoring, cyanobacterial bloom early warning, and mechanism research. Experimental data quantitatively demonstrate its significant advantages over traditional methods in various key indicators.

[0103] Example 3: An intelligent underwater algal community profiling and sensing system includes: The node cluster partitioning module is used to divide the underwater vertical profile space into non-uniform sensing node clusters based on the correlation of multi-source heterogeneous data of the original underwater environment. The double-nested judgment module is used to collect sensing data through the sensing node cluster and input the sensing data in parallel into the cluster model composed of clustering algorithm and the dissident model constructed based on physical and physiological mechanisms to perform double-nested judgment. The metabolic relationship chain construction and optimization module is used to construct a primary metabolic relationship chain using the sensory data judged through double-layer nesting, and adopt a multi-source data coupling inference method driven by time-series biochemical dynamics. The primary metabolic relationship chain is then optimized using a two-stage optimization model to obtain the final metabolic relationship chain. The path generation and verification module is used to generate the optimal profile sensing path by solving the final metabolic relationship chain through five-dimensional causal constraints. Before output, the spatial layer-data layer-ecological layer collaborative verification mechanism is forcibly started. The optimal profile sensing path is only unlocked and executed after the collaborative verification mechanism passes.

[0104] Those skilled in the art will understand that the embodiments of this application are provided as methods, systems, or computer program products. Therefore, this application takes the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application takes the form of a computer program product implemented on one or more computer storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer program code. The solutions in the embodiments of this application are implemented using various computer languages, exemplified by the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0105] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. 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, are implemented by computer program instructions. These computer program instructions are 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, create means for implementing the functions specified in the flowchart illustrations and / or block diagrams.

[0106] These computer program instructions are also stored in a computer read-memory that can direct a computer or other programmed data processing device to operate in a particular manner, such that the instructions stored in the computer read-memory produce an article of manufacture including instruction means that implement the functions specified in the flowchart or multiple flowcharts and / or block diagram blocks or multiple block diagrams.

[0107] These computer program instructions are also loaded onto a computer or other programming data processing device to cause a series of operational steps to be performed on the computer or other programming device to produce a computer-implemented process, such that the instructions, which execute on the computer or other programming device, provide steps for implementing the functions specified in the flowchart flow or multiple flows and / or the block diagram blocks or multiple blocks.

[0108] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0109] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for intelligent profiling and sensing underwater algal communities, characterized in that, include: Based on the correlation of multi-source heterogeneous data of the original underwater environment, the underwater vertical profile space is divided into non-uniform sensing node clusters. Sensing data is collected through a cluster of sensing nodes, and a two-layer nested judgment is performed based on the sensing data. The first layer of the two-layer nested judgment is to determine the consistency of the data within the cluster of sensing nodes, and the second layer of the two-layer nested judgment is to determine the reliability of the observation data of the cluster of sensing nodes. Using sensory data obtained through double-layer nesting, a primary metabolic relational chain is constructed using a multi-source data coupling inference method driven by time-series biochemical dynamics. A two-stage optimization model is then used to optimize the initial metabolic relational chain to obtain the final metabolic relational chain. The first stage of the two-stage optimization model includes: representing the primary metabolic relational chain as a directed weighted graph; calculating the network's topological entropy based on normalized edge weights; constructing and solving an optimization problem with the objective function of minimizing the topological entropy and the constraint that the fitting error of the optimized network to the observed data does not exceed a first preset threshold; and using a heuristic search algorithm to solve the problem and output the intermediate network that satisfies the fitting error constraint and has the lowest topological entropy. The second stage of the two-stage optimization model includes: defining the network's functional robustness index; constructing and solving an optimization problem with the objective function of maximizing the functional robustness index and the constraint that the network is obtained from the intermediate network through local structural adjustments and the fitting error does not exceed a second preset threshold; and using an evolutionary algorithm to solve the problem and output the final metabolic relational chain with the highest functional robustness. The final metabolic relationship chain is solved by five-dimensional causal constraints to generate the optimal profile sensing path. Before output, the spatial layer, data layer and ecological layer collaborative verification mechanism is forcibly activated. The path is only unlocked and executed after the collaborative verification mechanism is passed. The five-dimensional causal constraints are to maximize spatial coverage, maximize data information gain, maximize metabolic flux continuity, minimize energy consumption and minimize ecological disturbance. Among them, spatial layer verification is determined by calculating the minimum distance from all points on the optimal profile sensing path to the geometric center of each sensing node cluster and the minimum distance of all node positions. If the minimum distance corresponding to a sensing node cluster is greater than the preset spatial coverage radius threshold, the spatial layer verification is deemed to have failed. The data layer validates data through double-layer nested judgment logic; Ecosystem layer verification: The morphological similarity between the depth-attribute profile of the optimal profile sensing path and the typical metabolic profile inferred from the final metabolic relationship chain is calculated. If the morphological similarity is lower than the threshold, the ecosystem layer verification is deemed to have failed.

2. The intelligent profile sensing method for underwater algal communities according to claim 1, characterized in that, The sensing node cluster partitioning step includes: For any two spatial nodes, calculate the temporal correlation coefficient within the sliding time window for each type of observation data; Based on the temporal correlation coefficient, the comprehensive correlation similarity between nodes is calculated. The comprehensive correlation similarity is obtained by weighting and summing the temporal correlation coefficients of multiple observation data according to ecological importance weights, and then inputting them into the hyperbolic tangent function for calculation. Construct a symmetric comprehensive similarity matrix based on the comprehensive correlation similarity among all node pairs; Based on the comprehensive similarity matrix, construct the degree matrix and the normalized Laplacian matrix, solve for the eigenvectors corresponding to the smallest eigenvalues ​​of the Laplacian matrix, and form the eigenvector matrix. The three-dimensional spatial coordinate vectors of the nodes are concatenated to the corresponding rows of the feature vector matrix with preset weights to form extended feature vectors; The K-means++ clustering algorithm is used to divide the extended feature vectors to obtain non-uniform clusters of sensing nodes.

3. The intelligent profile sensing method for underwater algal communities according to claim 1, characterized in that, in, The first layer of the nested judgment logic is as follows: extract the core algal biomass observation data collected by all nodes in the sensing node cluster at the current time, calculate the arithmetic mean of the core algal biomass observation data, and calculate the discrete value representing the consistency of data within the cluster according to the discrete value calculation formula. Compare the discrete value with a preset consistency threshold. If the discrete value is less than the consistency threshold, it is determined that the data within the sensing node cluster has high consistency, and the second layer of judgment logic is entered. If the discrete value is greater than or equal to the consistency threshold, the sensing node cluster is marked as a highly discrete cluster, and the sensing node cluster re-division process is triggered. The second layer of judgment logic in the double-layer nested judgment is: inputting the multidimensional observation data of the perception node cluster that passed the first layer of judgment logic into the dissenter model constructed based on physical and physiological mechanisms. The dissenter model estimates the biomass baseline value predicted by the physical model based on the environmental driving factors of depth, temperature and photosynthetically effective radiation intensity, and the algal growth simulation based on light limitation and temperature regulation. The difference between the actual observed biomass of each node in the sensing node cluster and the biomass baseline value predicted by the physical model is calculated, and the normalized root mean square error is calculated as the physical prediction value based on the difference. The physical prediction value is compared with a preset conflict threshold. If the physical prediction value is less than the conflict threshold, the observation data of the sensing node cluster is determined to be high-reliability data and output. If the physical prediction value is greater than or equal to the conflict threshold, the sensing node cluster is marked as a conflict cluster and the sensing node cluster is re-divided.

4. The intelligent profile sensing method for underwater algal communities according to claim 3, characterized in that, The formula for calculating the discrete value is: ;in, For sensing node clusters Discrete value at the current moment , For sensing node clusters The number of nodes that internally transmit effective biomass data. For sensing node clusters All effective biomass data within, , For all effective biomass data arithmetic mean .

5. The intelligent profile sensing method for underwater algal communities according to claim 1, characterized in that, The steps for constructing the primary metabolic relationship chain include: The cluster of perception nodes judged through double-layer nesting is regarded as an ideal hybrid reaction unit, and the reaction unit is defined at time... The state vector of metabolic substance concentration; Define a set of possible biochemical reactions and establish a kinetic rate equation for each reaction that depends on the current concentration of metabolites and environmental driving factors; For each metabolite, establish its stoichiometric coefficient in each reaction; Using concentration data observed at multiple consecutive time points, the kinetic parameters, stoichiometric coefficients, and transport term parameters of all reactions are estimated through a global optimization algorithm. Reactions with non-zero significant fluxes and their interconnections form a primary metabolic relationship chain.

6. The intelligent profile sensing method for underwater algal communities according to claim 1, characterized in that, The intermediate network has fewer connections compared to the primary metabolic relationship chain.

7. The intelligent profile sensing method for underwater algal communities according to claim 1, characterized in that, The functional robustness metric is evaluated by simulating the average size of the largest connected subgraph of the remaining network after a preset proportion of edges are randomly removed from the network. The local structural adjustments include adding alternative conversion pathways between key metabolic nodes, enhancing feedback regulatory connections in existing metabolic cycles, or adjusting edge weight ratios.

8. The intelligent profile sensing method for underwater algal communities according to claim 1, characterized in that, The process of generating the optimal profile sensing path by solving the final metabolic relationship chain through five-dimensional causal constraints includes: The final metabolic relationship chain is mapped back to the three-dimensional water body profile space, metabolically active areas are identified, and a set of candidate paths from the start point to the end point is generated in the three-dimensional spatial grid. For each path in the candidate path set, an objective function for a multi-objective optimization problem is constructed, which includes maximizing spatial coverage, maximizing data information gain, maximizing metabolic flux continuity, minimizing energy consumption, and minimizing ecological disturbance. The Pareto optimal solution set is obtained by solving the objective function of the multi-objective optimization problem using a multi-objective evolutionary algorithm. The path is selected from the Pareto optimal solution set as the optimal profile sensing path.

9. The intelligent profile sensing method for underwater algal communities according to claim 8, characterized in that, The data layer verification mechanism in the spatial layer data layer ecological layer collaborative verification mechanism includes: predicting virtual sampling point data along the optimal profile perception path based on the dissenter model, allocating the virtual sampling point data to the perception node cluster, and simulating the execution of double-layer nested judgment logic. If the proportion of data passing the double-layer nested judgment logic is lower than a preset threshold, the data layer verification is determined to have failed. The optimal profile sensing path is unlocked only after all three layers of verification—spatial layer, data layer, and ecosystem layer—have passed.

10. An intelligent underwater algal community profiling system according to any one of claims 1-9, characterized in that, The system includes: The node cluster partitioning module is used to divide the underwater vertical profile space into non-uniform sensing node clusters based on the correlation of multi-source heterogeneous data of the original underwater environment. The double-nested judgment module is used to collect sensing data through the sensing node cluster and input the sensing data in parallel into the cluster model composed of clustering algorithm and the dissident model constructed based on physical and physiological mechanisms to perform double-nested judgment. The metabolic relationship chain construction and optimization module is used to construct a primary metabolic relationship chain using the sensory data judged through double-layer nesting, and adopt a multi-source data coupling inference method driven by time-series biochemical dynamics. The primary metabolic relationship chain is then optimized using a two-stage optimization model to obtain the final metabolic relationship chain. The path generation and verification module is used to generate the optimal profile sensing path by solving the final metabolic relationship chain through five-dimensional causal constraints. Before output, the spatial layer-data layer-ecological layer collaborative verification mechanism is forcibly started. The optimal profile sensing path is only unlocked and executed after the collaborative verification mechanism passes.

Images