An internet of things big data marine culture environment monitoring and evaluation method and system

Abstract

This invention discloses a method and system for monitoring and assessing the marine aquaculture environment using IoT big data. The method includes: collecting multi-dimensional environmental parameters such as dissolved oxygen, nitrogen and phosphorus nutrients, and pH value through IoT sensing devices; performing parallel reception and status verification through a high-concurrency marine environmental data processing platform; using a multi-dimensional time-series data adaptive clustering algorithm to perform time-series segmentation and similarity clustering of the parameters; quantitatively assessing water quality status based on the clustering results using a eutrophication assessment model; and then dynamically predicting stress intensity and duration using an environmental stress response prediction model, combined with parameters adapted to the growth cycle of aquaculture organisms. The platform integrates the assessment and prediction results to generate a structured report including parameter change trajectories, level indicators, and predicted trend curves, achieving precise monitoring and assessment of the marine aquaculture environment throughout the entire process, providing reliable technical support for aquaculture environment regulation.

Description

Technical Field

[0001] This invention relates to the field of marine aquaculture environment monitoring and assessment technology, and in particular to a method and system for marine aquaculture environment monitoring and assessment based on Internet of Things big data. Background Technology

[0002] Marine aquaculture, as a crucial component of aquaculture, relies heavily on the stability of its aquaculture environment for both yield and quality. However, issues such as eutrophication and environmental stress have become key factors hindering the industry's high-quality development. With the widespread adoption of IoT and big data technologies, marine aquaculture environmental monitoring is gradually shifting towards intelligent and precise monitoring. However, environmental parameters in aquaculture areas are characterized by multidimensionality, temporal sequence, and high concurrency, making it difficult for traditional monitoring methods to efficiently process and deeply analyze multi-source data. Currently, aquaculture areas require an integrated solution that combines parameter acquisition, data processing, algorithm analysis, model evaluation, and prediction to comprehensively capture the dynamic changes of key parameters such as dissolved oxygen, nitrogen and phosphorus nutrients, and pH, scientifically assess the degree of eutrophication, accurately predict environmental stress responses, and provide reliable support for the scientific regulation of aquaculture activities. Therefore, it is urgent to construct a marine aquaculture environmental monitoring and assessment method and system based on IoT big data.

[0003] Existing technologies have two significant drawbacks: First, traditional monitoring and assessment methods lack adaptive clustering processing for multi-dimensional time-series environmental parameters, often employing fixed-dimensional partitioning methods. This fails to dynamically construct clusters based on the time-series changes in parameters, resulting in input data for eutrophication assessment models lacking specificity and failing to accurately reflect environmental differences across different time periods and regions. Second, existing systems do not achieve deep integration of high-concurrency data processing with assessment and prediction models. Either the data processing platform struggles to handle the high-concurrency data transmission and verification requirements of multiple sensor sources, or the assessment and prediction models operate independently, failing to fully incorporate parameters adapted to the growth cycle of aquaculture organisms. This leads to insufficient dynamism in stress response prediction, hindering timely and accurate decision-making support for aquaculture environment regulation. Summary of the Invention

[0004] In order to overcome the shortcomings and deficiencies of existing technologies, this invention provides a method and system for monitoring and evaluating the marine aquaculture environment using Internet of Things big data.

[0005] The technical solution adopted in this invention is a method for monitoring and evaluating the marine aquaculture environment using IoT big data, comprising the following steps: S1, collecting multi-dimensional environmental parameters of the marine aquaculture area through IoT sensing devices, including dissolved oxygen content, nitrogen and phosphorus nutrient concentration, pH value, water temperature, salinity, turbidity, and plankton community structure-related parameters; S2, transmitting the collected multi-dimensional environmental parameters to a high-concurrency marine environmental data processing platform, and performing parallel reception of multi-source data and data transmission status verification through the platform's distributed data receiving module; S3, calling a multi-dimensional time-series data adaptive clustering algorithm to perform time-series dimension division and similarity clustering processing on the status-verified multi-dimensional environmental parameters, based on the time-series changes of the aquaculture environmental parameters. The process involves: S4, constructing dynamic clusters based on the clustered environmental parameter dataset; S5, running a water eutrophication assessment model to quantitatively assess the eutrophication level of the aquaculture area and generating eutrophication level correlation data; S6, inputting the eutrophication level correlation data and clustered multidimensional environmental parameters into an aquaculture environmental stress response prediction model, and combining it with environmental adaptation parameters related to the growth cycle of aquaculture organisms to dynamically predict the intensity and duration of aquaculture environmental stress; and S7, integrating and structured outputting the eutrophication assessment results and stress response prediction results through a high-concurrency marine environmental data processing platform to generate a marine aquaculture environmental monitoring and assessment report that includes parameter change trajectories, assessment level identifiers, and prediction trend curves.

[0006] Furthermore, the expression for the eutrophication assessment model of the aquaculture area is as follows: ,in, It is a water eutrophication assessment index. For model fit coefficients, 1 is the dissolved oxygen content parameter, TP is the total phosphorus concentration parameter, TN is the total nitrogen concentration parameter, PH is the pH value parameter, TURB is the turbidity parameter, and TEMP is the water temperature parameter.

[0007] Furthermore, the core expression of the multidimensional time-series data adaptive clustering algorithm is: ,in, This is the cluster similarity threshold. For the number of clusters, For the first The and the first Similarity measures of time-series environmental parameters, The time interval for collecting the two parameters. This is the time-series decay coefficient. For a single environmental parameter time series data point, This is the time-series mean of this type of parameter.

[0008] Furthermore, the expression for the prediction model of the aquaculture environmental stress response is: ,in, This represents the intensity value of the environmental stress response. , The stress prediction coefficient, The eutrophication assessment index is SAL, which is the salinity parameter. For the first Community proportion parameters of planktonic organisms For the first Environmental stress sensitivity coefficients corresponding to planktonic organisms This represents the number of categories in the planktonic community.

[0009] Furthermore, the data processing efficiency model of the high-concurrency marine environmental data processing platform is as follows: ,in, For platform data processing speed, For the number of concurrent data transmission channels, This is the data transmission volume per channel. This serves as a baseline for data processing latency. This represents the channel interference coefficient.

[0010] Furthermore, the temporal feature extraction model for the multidimensional environmental parameters is as follows: ,in, For time series feature quantization values, for The variance of environmental parameters at any given time. for Time and Covariance of time parameters These are the feature weight coefficients. for Skewness of time series parameters.

[0011] Further, S3 includes the following sub-steps: S31, constructing a time-series coordinate system based on the collection timestamps of multi-dimensional environmental parameters, setting the time window length according to the environmental change pattern within the breeding cycle, and dividing the complete time-series data into several continuous time-series segments; S32, calculating the rate of change and fluctuation amplitude of each environmental parameter within each time-series segment, establishing parameter change feature vectors, and eliminating the dimensional differences of different parameters through vector standardization; S33, using an adaptive distance metric method to calculate the similarity between feature vectors of each time-series segment, and dynamically adjusting the cluster size threshold according to the similarity distribution; S34, clustering the time-series segments according to the adjusted threshold, merging time-series segments with similarity higher than the threshold to form clusters, and recording the calibration parameter characteristics and time-series interval range of each cluster.

[0012] Further, step S4 includes the following sub-steps: S41, extracting core assessment parameters of total nitrogen, total phosphorus, and dissolved oxygen from the clustered environmental parameter dataset, and performing data screening and association mapping according to the parameter requirements of the eutrophication assessment model for aquaculture areas; S42, determining the α, β, and γ adaptation coefficients corresponding to the current aquaculture area through the parameter adaptation module built into the model, and substituting the screened parameter data into the model for index calculation; S43, mapping the calculated assessment index E to the corresponding eutrophication level according to the preset eutrophication level classification standard, and generating a level identifier and index association table; S44, performing data verification on the level association table, removing abnormal mapping data, and ensuring the consistency and accuracy of the eutrophication assessment results.

[0013] Further, S5 includes the following sub-steps: S51, integrating eutrophication level correlation data and clustered multidimensional environmental parameters to construct an input dataset for stress response prediction, and clarifying the weight ratio of each parameter in the dataset; S52, importing environmental adaptation parameters corresponding to the growth cycle of aquaculture organisms, and determining the δ and ε stress prediction coefficients in the model in combination with the input dataset; S53, substituting the input dataset and prediction coefficients into the aquaculture environmental stress response prediction model to calculate the environmental stress response intensity S, and classifying the stress level based on the intensity value; S54, predicting the stress duration range according to the stress level and the time-series change trend of the parameters, and generating dynamic prediction results including intensity value, stress level and duration.

[0014] A marine aquaculture environment monitoring and assessment system based on Internet of Things (IoT) big data includes: an IoT multi-dimensional environmental parameter sensing and acquisition unit, used to collect parameters related to dissolved oxygen, nitrogen and phosphorus nutrients, pH value, water temperature, salinity, turbidity, and plankton community structure through distributed sensor devices, and establish a bidirectional data transmission connection with a high-concurrency marine environmental data processing unit; a high-concurrency marine environmental data processing unit, used to receive multi-source data transmitted by the sensing and acquisition unit, perform parallel reception and status verification through a distributed receiving module, and establish data interaction channels with a multi-dimensional time-series data adaptive clustering processing unit, an aquaculture environment assessment and prediction unit, and an assessment result integration and output unit; and a multi-dimensional time-series data adaptive clustering processing unit, used to call clustering... The algorithm performs time-series partitioning and similarity clustering on the validated parameters, and transmits the clustering results to the aquaculture environment assessment and prediction unit. The aquaculture area water quality eutrophication assessment unit is used to run the eutrophication assessment model based on the clustering results, generate eutrophication level association data, and send it to the aquaculture environment stress response prediction unit. The aquaculture environment stress response prediction unit is used to receive the eutrophication level association data and clustering parameters, combine them with biological growth cycle adaptation parameters to predict the stress intensity and duration, and transmit the prediction results to the assessment result integration and output unit. The assessment result integration and output unit is used to receive the assessment results and prediction results, perform structured integration, generate an assessment report including parameter trajectories, level identifiers, and trend curves, and output it.

[0015] The present invention has the following beneficial effects:

[0016] Key environmental parameters of marine aquaculture areas are comprehensively collected through IoT sensing devices. A high-concurrency marine environmental data processing platform handles parallel reception and verification of multi-source data, solving the problem of traditional systems struggling to handle high-concurrency data transmission. A multi-dimensional time-series adaptive clustering algorithm dynamically constructs clusters based on the temporal variation characteristics of parameters, replacing fixed-dimensional partitioning methods. This makes the data input to the evaluation model more targeted and accurately compensates for the inability of traditional methods to reflect spatiotemporal environmental differences. Furthermore, a quantitative assessment is conducted by integrating a eutrophication assessment model for aquaculture areas with a prediction model for aquaculture environmental stress response, combined with parameters adapted to the growth cycle of aquaculture organisms. With dynamic prediction, the system achieves a deep integration of assessment and prediction, enhancing the scientific rigor and dynamism of the results. Through the coordinated operation of six functional units, the system completes a closed-loop process from parameter acquisition, data processing, cluster analysis, assessment and prediction to result output, generating a structured report that includes parameter change trajectories, level indicators, and trend curves. This not only provides comprehensive and accurate decision-making basis for aquaculture environment control but also, through the organic combination of technologies in each stage, completely solves the core shortcomings of existing technologies, such as low data processing efficiency, lack of targeted assessment, and insufficient predictive dynamism, thus promoting the development of marine aquaculture environment monitoring towards intelligence and precision. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the overall process of the method of the present invention.

[0018] Figure 2 This is a flowchart of method step S3 of the present invention;

[0019] Figure 3 This is a flowchart of method step S4 of the present invention;

[0020] Figure 4 This is a flowchart of step S5 of the method of the present invention;

[0021] Figure 5 This is a diagram showing the system unit composition of the present invention. Detailed Implementation

[0022] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. The application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0023] like Figure 1 As shown, a method for monitoring and assessing the marine aquaculture environment using IoT big data includes the following steps:

[0024] S1. Collect multi-dimensional environmental parameters of the marine aquaculture area through Internet of Things (IoT) sensing devices. The multi-dimensional environmental parameters include dissolved oxygen content, nitrogen and phosphorus nutrient concentration, pH value, water temperature, salinity, turbidity, and plankton community structure related parameters.

[0025] Specifically, step S1 involves the comprehensive collection of multi-dimensional environmental parameters in the marine aquaculture area using distributed IoT sensing devices. These devices include dissolved oxygen sensors, nitrogen and phosphorus nutrient analyzers, pH sensors, water temperature sensors, salinity sensors, turbidity sensors, and plankton community structure analyzers. The devices are deployed at a standard density of one monitoring node per 500 square meters to ensure coverage of the entire aquaculture area. The collected parameters include dissolved oxygen content, total nitrogen concentration, total phosphorus concentration, pH value, water temperature, salinity, turbidity, and related parameters such as the proportion of various plankton communities and biomass. The accuracy of dissolved oxygen content collection is ±0.01 mg / L, nitrogen and phosphorus nutrient concentration collection is ±0.001 mg / L, pH value collection is ±0.01 mg / L, water temperature collection is ±0.1°C, salinity collection is ±0.1‰, and turbidity collection is ±1 NTU. The data collection frequency is set to once every 15 minutes, and automatically adjusted to once every 5 minutes during special weather conditions (such as before and after heavy rain or typhoons). During the collection process, the data is calibrated in real time through the sensor's built-in calibration module to ensure the accuracy of the collected data. This step provides comprehensive and accurate raw data support for subsequent data processing, cluster analysis, and assessment and prediction. It is the foundation of the entire monitoring and assessment process, and its implementation quality directly affects the reliability of the results of all subsequent links. By comprehensively covering the key parameters affecting the aquaculture environment, it avoids assessment and prediction deviations caused by missing parameters.

[0026] S2 transmits the collected multi-dimensional environmental parameters to the marine environmental data high-concurrency processing platform, and performs parallel reception of multi-source data and data transmission status verification through the platform's distributed data receiving module;

[0027] Specifically, step S2 transmits the multi-dimensional environmental parameters collected in S1 to the high-concurrency marine environmental data processing platform via a hybrid communication network of 4G, 5G, and LoRa. Encrypted transmission protocols are used to ensure data security during transmission. The high-concurrency marine environmental data processing platform is equipped with 16 distributed data receiving modules, each supporting up to 1000 simultaneous data streams, enabling parallel reception of multi-source data. The platform's overall peak data reception rate reaches 10Gbps, meeting the high-concurrency data transmission needs of large-scale IoT sensing devices. During data reception, the platform performs real-time verification of the transmission status of each data stream. Verification includes data integrity, data format compliance, and data transmission delay. Data integrity verification is achieved through checksum comparison to ensure consistency between received and transmitted data. Data format compliance verification follows a preset unified data format standard; data that does not meet the format requirements is marked and temporarily stored. Data transmission delay verification sets a delay threshold of 500 milliseconds; data with delays exceeding this threshold triggers a retransmission mechanism. This step, through efficient parallel reception and rigorous status verification, ensures the rapid transmission and data quality of multi-source environmental parameters, providing reliable data input for subsequent clustering and model evaluation. It avoids distortion of subsequent processing results due to data transmission loss, format errors, or excessive latency. At the same time, the platform's high concurrency processing capability ensures the stability and timeliness of data transmission when large-scale deployment of sensing devices in aquaculture areas.

[0028] S3 calls a multi-dimensional time-series adaptive clustering algorithm to perform time-series dimension division and similarity clustering on the multi-dimensional environmental parameters that have passed state verification, and constructs dynamic clusters based on the time-series change characteristics of the aquaculture environment parameters;

[0029] Specifically, step S3 calls a multi-dimensional time-series adaptive clustering algorithm to process the state-verified multi-dimensional environmental parameters. First, based on the timestamp information of parameter collection, the time-series dimension is divided. The division uses the breeding cycle (e.g., 30 days) as the overall time range, and breaks down the complete time-series data into several continuous time segments using 24-hour periods as the basic time unit. Each time segment includes all parameter data generated according to the collection frequency within that time unit. Then, the variation characteristics of each environmental parameter within each time segment are calculated, including the parameter change rate, fluctuation amplitude, and extreme value distribution. Based on these characteristics, a parameter change feature vector for each time segment is constructed. Vector standardization eliminates the influence of differences in the dimensions of different parameters. An adaptive distance metric method is used to calculate the similarity between the feature vectors of each time segment. The similarity calculation is based on the consistency of parameter change trends and the degree of agreement on numerical fluctuations. The threshold for the number of clusters is dynamically adjusted according to the similarity distribution among all time segments, with the threshold adjustment range being 3-10 clusters to ensure that the clustering results accurately reflect the variation patterns of environmental parameters in different time periods. Finally, based on the adjusted threshold, all time series segments are clustered. Time series segments with similarity higher than the threshold are merged into a cluster. At the same time, the core parameter features (including average value, range of variation, etc.) and time series interval range of each cluster are recorded. This step achieves effective classification of multidimensional time series environmental parameters through dynamic clustering, enabling subsequent evaluation models to accurately analyze parameter clusters with different features, thereby improving the relevance and accuracy of the evaluation results.

[0030] S4. Based on the clustered environmental parameter dataset, run the eutrophication assessment model of aquaculture area water quality to quantitatively assess the eutrophication level of aquaculture area water quality and generate eutrophication level correlation data.

[0031] Specifically, step S4, based on the environmental parameter dataset processed by clustering in S3, runs the eutrophication assessment model for the aquaculture area to quantitatively assess the eutrophication level of the aquaculture area. First, core assessment parameters such as total nitrogen concentration, total phosphorus concentration, dissolved oxygen content, pH value, turbidity, and water temperature are extracted from the environmental parameter dataset corresponding to each cluster. During extraction, outliers are removed from the dataset (outliers are defined as data deviating from the mean of the cluster's parameters by ±3 standard deviations) to ensure the reliability of the parameters input to the model. During model execution, the corresponding model adaptation coefficients are called according to the type of the current aquaculture area (e.g., fish farming area, shellfish farming area, algae farming area). These adaptation coefficients are obtained through training with a large amount of historical monitoring data and are optimized and adjusted for the water quality requirements of different aquaculture types. The model comprehensively calculates core assessment parameters to generate a eutrophication assessment index for each cluster within its corresponding time interval. Based on the index range, eutrophication levels are divided into five grades: below 0.3 (oligotrophic), 0.3-0.5 (mesotrophic), 0.5-0.7 (slightly eutrophic), 0.7-0.9 (moderately eutrophic), and above 0.9 (severely eutrophic). Eutrophication level correlation data is generated based on the classification results. This data includes the time interval, mean of core parameters, assessment index, and corresponding eutrophication level for each cluster. This step, through targeted parameter extraction and grading assessment, achieves accurate quantification of eutrophication levels in aquaculture areas, providing crucial water quality data support for subsequent stress response prediction and helping aquaculture managers promptly grasp water quality changes.

[0032] S5. The eutrophication level correlation data and the clustered multidimensional environmental parameters are input into the aquaculture environmental stress response prediction model. Combined with the environmental adaptation parameters related to the growth cycle of aquaculture organisms, the intensity and duration of aquaculture environmental stress are dynamically predicted.

[0033] Specifically, step S5 inputs the eutrophication level association data generated in S4 and the multidimensional environmental parameters after clustering in S3 into the aquaculture environmental stress response prediction model. Simultaneously, it imports environmental adaptation parameters associated with the growth cycle of the current farmed organism (e.g., specific fish or shellfish). These environmental adaptation parameters include the suitable range, tolerance threshold, and sensitivity coefficient of each environmental parameter for different growth stages of the farmed organism, such as the seedling stage, growth stage, and maturity stage. These parameters are determined based on the physiological characteristics of the farmed organism and historical aquaculture data. During model execution, the eutrophication level association data is first converted into corresponding stress factor weights, with a weight of 0.8 for severe eutrophication, 0.6 for moderate eutrophication, 0.4 for mild eutrophication, 0.2 for mesotrophication, and 0.1 for oligotrophication. Simultaneously, the basic stress intensity is calculated by combining the differences between the clustered multidimensional environmental parameters and the environmental adaptation parameters. By comprehensively analyzing the temporal trends of basic stress intensity, stress factor weights, and environmental parameters through model analysis, the intensity and duration of environmental stress in aquaculture are dynamically predicted. Stress intensity is categorized into three levels: mild, moderate, and severe. Stress duration prediction is based on a fitting analysis of parameter trends, with prediction accuracy controlled within ±2 hours. During the prediction process, the model updates the input data hourly and dynamically corrects the prediction results to ensure that they reflect the latest changes in environmental parameters in a timely manner. This step, by combining eutrophication assessment results with the characteristics of aquaculture organisms, achieves accurate prediction of environmental stress, providing a time window and scientific basis for aquaculture managers to take targeted control measures and reduce the adverse effects of environmental stress on the growth of aquaculture organisms.

[0034] S6 integrates and structures the eutrophication assessment results and stress response prediction results through a high-concurrency marine environmental data processing platform, generating a marine aquaculture environmental monitoring and assessment report that includes parameter change trajectories, assessment level identifiers, and prediction trend curves.

[0035] Specifically, step S6 uses a high-concurrency marine environmental data processing platform to integrate and structure the eutrophication assessment results from S4 and the stress response prediction results from S5. During data integration, the platform first correlates and matches the assessment and prediction results according to time intervals, ensuring that the eutrophication level, assessment index, stress intensity, and stress duration data for each time interval correspond one-to-one. Simultaneously, it integrates the raw parameter data collected in S1 with the clustering results from S3 to form a complete data chain. For structured output, a marine aquaculture environmental monitoring and assessment report is generated according to a preset report template. The report includes three core parts: parameter change trajectory, assessment level identifier, and prediction trend curve. The parameter change trajectory presents the numerical changes of each core environmental parameter throughout the monitoring period in a line graph, marking the parameter values ​​and change rates at key time points. The assessment level identifier lists the time interval, eutrophication level, and mean core parameter for each cluster in a table format. The prediction trend curve presents the changing trends of eutrophication level and stress intensity over the next 72 hours in a curve format, marking the possible peak stress time and intensity values. The report output format supports three forms: PDF, Excel, and web-based visualization. At the same time, the platform will push the report to the mobile and computer terminals of aquaculture managers. This step provides aquaculture managers with intuitive and comprehensive monitoring and evaluation information through comprehensive data integration and clear structured output, which makes it easier for them to quickly grasp the status of the aquaculture environment and make scientific decisions, thereby achieving refined management of the marine aquaculture environment.

[0036] Preferably, the expression for the eutrophication assessment model of the aquaculture area is: ,in, It is a water eutrophication assessment index. For model fit coefficients, 1 is the dissolved oxygen content parameter, TP is the total phosphorus concentration parameter, TN is the total nitrogen concentration parameter, PH is the pH value parameter, TURB is the turbidity parameter, and TEMP is the water temperature parameter.

[0037] Specifically, the eutrophication assessment model for aquaculture areas is designed for the water quality characteristics of marine aquaculture environments, achieving precise quantification of eutrophication levels by integrating key water quality parameters. The model selects dissolved oxygen content, total phosphorus concentration, total nitrogen concentration, pH value, turbidity, and water temperature as core input parameters. These parameters are key factors affecting eutrophication in aquaculture areas, covering multiple dimensions such as nutrient levels, dissolved oxygen status, and physical environmental characteristics. The model's adaptation coefficients are differentiated according to different aquaculture area types (e.g., nearshore cage aquaculture areas, intertidal aquaculture areas, and deep-water aquaculture areas), with values ​​ranging from 0.1 to 0.9. These coefficients are obtained through extensive historical monitoring data and corresponding analysis of actual eutrophication conditions, ensuring coefficient adaptability. During implementation, the mean or eigenvalues ​​of each core parameter are extracted from the clustered environmental parameter dataset. These are then comprehensively calculated according to the model's predefined computational logic to obtain the eutrophication assessment index. The numerical range of this index corresponds to different eutrophication levels: below 0.3 corresponds to oligotrophic state, 0.3 to 0.5 to mesotrophic state, 0.5 to 0.7 to slightly eutrophic state, 0.7 to 0.9 to moderately eutrophic state, and above 0.9 to severely eutrophic state. By applying this model, multi-dimensional water quality parameters can be transformed into intuitive quantitative indices, accurately reflecting the actual degree of eutrophication in aquaculture areas. This provides reliable basic water quality data for subsequent environmental stress response prediction. The implementation process strictly follows the logical flow of parameter extraction, coefficient matching, index calculation, and level mapping to ensure the scientific validity and consistency of the assessment results.

[0038] Preferably, the core expression of the multidimensional time-series data adaptive clustering algorithm is: ,in, This is the cluster similarity threshold. For the number of clusters, For the first The and the first Similarity measures of time-series environmental parameters, The time interval for collecting the two parameters. This is the time-series decay coefficient. For a single environmental parameter time series data point, This is the time-series mean of this type of parameter.

[0039] Specifically, a multidimensional time-series adaptive clustering algorithm enables dynamic and efficient clustering of multidimensional time-series parameters of marine aquaculture environments, providing a structured dataset with time-series characteristics for subsequent assessment and prediction. The core logic of this algorithm revolves around calculating the cluster similarity threshold. The number of clusters is dynamically adjusted based on the complexity of the time-series changes in aquaculture environmental parameters, ranging from 3 to 10, ensuring that the clustering results reflect the differences in parameter changes without data fragmentation due to an excessive number of clusters. In implementation, the temporal position of each data point is first determined based on the collection timestamp of the environmental parameters. The similarity metric between the i-th and j-th time-series environmental parameters is calculated. This metric comprehensively considers the degree of agreement between parameter values ​​and the consistency of change trends, ranging from 0 to 1, with values ​​closer to 1 indicating higher similarity. Simultaneously, the collection time interval of the two parameters is introduced as a time-series weighting factor; the shorter the time interval, the greater the weight, ensuring that recent data has a more significant impact on the clustering results. The time-series decay coefficient ranges from 0.01 to 0.1, and is used to adjust the impact of the deviation of a single environmental parameter time-series data point from the time-series mean of that type of parameter on the similarity calculation, avoiding interference from outlier data points in the clustering effect. The algorithm obtains a cluster similarity threshold by performing a weighted summation and averaging of the similarities of all time-series data points, and then divides the time-series segments into different clusters based on this threshold. This algorithm does not require manual pre-setting of fixed clustering criteria, can adapt to the time-series changes in aquaculture environmental parameters, and forms clusters with clear time-series patterns. This allows subsequent evaluation models to accurately analyze parameter characteristics at different time periods, improving the targeting and efficiency of the entire monitoring and evaluation process.

[0040] Preferably, the expression for the aquaculture environmental stress response prediction model is: ,in, This represents the intensity value of the environmental stress response. , The stress prediction coefficient, The eutrophication assessment index is SAL, which is the salinity parameter. For the first Community proportion parameters of planktonic organisms For the first Environmental stress sensitivity coefficients corresponding to planktonic organisms This represents the number of categories in the planktonic community.

[0041] Specifically, the aquaculture environmental stress response prediction model combines eutrophication status and multidimensional environmental parameters to dynamically predict the intensity and duration of aquaculture environmental stress. The core inputs of the model include the eutrophication assessment index, salinity parameter, phytoplankton community proportion, and corresponding environmental stress sensitivity coefficients. It also incorporates parameters related to the growth cycle of aquaculture organisms to ensure that the prediction results closely match actual aquaculture needs. The stress prediction coefficients are dynamically adjusted according to the species and growth stage of the aquaculture organisms, with δ ranging from 0.5 to 1.2 and ε ranging from 0.02 to 0.1. These values ​​were determined through extensive aquaculture experimental data and environmental stress case analysis, accurately reflecting the response patterns of different organisms to environmental stress. The number of planktonic community categories is determined based on the ecological environment survey results of the aquaculture area, typically ranging from 5 to 8 categories. These include key planktonic groups sensitive to water quality changes and environmental stresses. The community proportion of each planktonic category is calculated using data collected by a biological community structure analyzer. The environmental stress sensitivity coefficient is determined based on the physiological characteristics and environmental tolerance of that planktonic category, ranging from 0.1 to 0.9. In the implementation process, the eutrophication assessment index and salinity parameters are first normalized, then weighted and calculated using planktonic community correlation parameters. The environmental stress response intensity value is obtained through a combination of exponential and quadratic functions. This intensity value corresponds to three levels: mild stress, moderate stress, and severe stress, with values ​​ranging from 0 to 0.3, 0.3 to 0.7, and 0.7 to 1.0, respectively. Subsequently, the duration of stress is predicted by combining the temporal trends of the parameters with linear fitting and trend extrapolation. The implementation of this model achieves a seamless connection between water quality assessment and stress prediction. The prediction results can provide a precise basis for aquaculture managers to take targeted control measures, effectively reducing the adverse effects of environmental stress on the growth of aquaculture organisms.

[0042] Preferably, the data processing efficiency model of the high-concurrency marine environmental data processing platform is as follows: ,in, For platform data processing speed, For the number of concurrent data transmission channels, This is the data transmission volume per channel. This serves as a baseline for data processing latency. This represents the channel interference coefficient.

[0043] Specifically, the data processing efficiency model of the high-concurrency marine environmental data processing platform quantifies the platform's ability to process multi-source concurrent data, ensuring that the platform can meet the data transmission and processing needs of large-scale IoT sensing devices. The core parameters of the model include the number of concurrent data transmission channels, the data transmission volume per channel, the baseline value for data processing latency, and the channel interference coefficient, comprehensively covering the key factors affecting the platform's processing efficiency. The number of concurrent data transmission channels is determined based on the deployment scale of IoT sensing devices in the aquaculture area, ranging from 100 to 1000 channels. The data transmission volume per channel is calculated based on the type and frequency of the collected parameters, typically 10 to 50 KB per channel per second. The baseline value for data processing latency is the average latency time for the platform to process single-channel data under interference-free conditions, ranging from 10 to 50 milliseconds, determined through platform performance testing. The channel interference coefficient considers the impact of network transmission fluctuations and device signal interference, ranging from 0.01 to 0.08. During implementation, the current number of concurrent data transmission channels and the data transmission volume of each channel are statistically analyzed in real time and substituted into the model to calculate the platform's current data processing rate. This model dynamically reflects the platform's operating status and processing capacity. When the calculated processing rate falls below a preset threshold, the platform automatically activates a load balancing mechanism. This mechanism improves processing efficiency by adding distributed processing nodes and optimizing data transmission paths, ensuring that data processing latency does not exceed a preset limit. The implementation of this model provides a quantitative basis for platform performance optimization and stable operation, ensuring the timeliness and accuracy of multi-source environmental parameter transmission and processing. It also provides efficient data support for subsequent cluster analysis, evaluation, and prediction, ensuring the smooth operation of the entire monitoring and evaluation system.

[0044] Preferably, the temporal feature extraction model for the multidimensional environmental parameters is as follows: ,in, For time series feature quantization values, for The variance of environmental parameters at any given time. for Time and Covariance of time parameters These are the feature weight coefficients. for Skewness of time series parameters.

[0045] Specifically, a time-series feature extraction model for multidimensional environmental parameters is used to quantify the time-series variation characteristics of these parameters, providing accurate feature input for adaptive clustering algorithms of multidimensional time-series data. The model comprehensively captures the fluctuation amplitude, trend correlation, and distribution pattern of parameters in the time dimension by calculating the variance, covariance, and skewness of the parameters, ensuring that the extracted time-series features accurately reflect the dynamic changes of environmental parameters. The feature weight coefficient ζ ranges from 0.3 to 0.8, determined through correlation analysis of the influence of time-series features on the clustering results, highlighting the contribution of key features to the clustering effect. In the implementation process, firstly, continuous time-series data segments are selected, and the variance of environmental parameters at time t is calculated. This variance reflects the fluctuation range of the parameters within this time-series segment; a larger value indicates more drastic parameter changes. Then, the covariance of the parameters at time t and time t-1 is calculated. This covariance reflects the correlation of parameter changes at adjacent times; a positive value indicates a consistent trend, while a negative value indicates an opposite trend. Next, the skewness of the time-series parameter sequence at time t is calculated. This skewness reflects the degree of asymmetry in the time-series distribution of parameters, ranging from -1 to 1, and can reveal whether there is a concentration of extreme values. These three statistics are combined and calculated according to the operational logic set in the model to obtain the quantified value of the time-series features. This quantified value serves as the core input of the multidimensional time-series adaptive clustering algorithm, helping the algorithm accurately identify the feature differences between different time-series segments and ensuring the rationality and effectiveness of the clustering results. The implementation of this model provides scientific and accurate feature support for the clustering algorithm, improves the accuracy and efficiency of time-series data clustering, and lays a solid foundation for subsequent eutrophication assessment and environmental stress prediction.

[0046] Preferred, such as Figure 2 As shown, step S3 includes the following sub-steps: S31, constructing a time-series coordinate system based on the collection timestamps of multi-dimensional environmental parameters, setting the time window length according to the environmental change pattern within the breeding cycle, and dividing the complete time-series data into several continuous time-series segments; S32, calculating the rate of change and fluctuation amplitude of each environmental parameter within each time-series segment, establishing parameter change feature vectors, and eliminating the dimensional differences of different parameters through vector standardization; S33, using an adaptive distance metric method to calculate the similarity between feature vectors of each time-series segment, and dynamically adjusting the cluster size threshold according to the similarity distribution; S34, clustering the time-series segments according to the adjusted threshold, merging time-series segments with similarity higher than the threshold to form clusters, and recording the calibration parameter characteristics and time-series interval range of each cluster.

[0047] Specifically, step S3, the multidimensional time-series data adaptive clustering process, achieves accurate segmentation and clustering of time-series data through four sub-steps. S31 constructs a complete time-series coordinate system based on the collection timestamps of multidimensional environmental parameters, with timestamps accurate to the millisecond level. The time window length is set according to the periodicity of environmental changes within the aquaculture cycle. The window length can be dynamically adjusted within a range of 1 to 24 hours based on the type of aquaculture area. The complete time-series data within the entire monitoring cycle is divided into several continuous and non-overlapping time-series segments, each segment including all environmental parameter data within the corresponding time window. S32, for each time-series segment, calculates the rate of change of each environmental parameter within the segment (obtained by the ratio of the difference between parameters at adjacent collection points to the time interval) and the fluctuation amplitude (calculated by the difference between the maximum and minimum values ​​of the parameters within the segment). Based on these indicators, a parameter change feature vector for each time-series segment is constructed. Then, vector standardization is used to convert parameter features of different dimensions into vector values ​​with a unified range, eliminating the influence of dimensional differences on subsequent clustering. S33 employs an adaptive distance metric method, quantifying similarity by calculating the Euclidean distance or cosine similarity between feature vectors of each time series segment. It statistically analyzes the similarity distribution of all vector pairs and dynamically adjusts the cluster size threshold based on the distribution peak and dispersion. The threshold adjustment range is 3 to 10 clusters to ensure the clustering results accurately reflect the temporal variation characteristics. S34, based on the adjusted similarity threshold, merges time series segments with similarity higher than the threshold into a single cluster. Simultaneously, it records the core parameter features of each cluster (including the mean, median, and mean rate of change of each parameter) and the time series interval (start and end timestamps). This step-by-step process ensures the systematic and accurate nature of the clustering process, providing a structured dataset with clear temporal characteristics for subsequent model evaluation.

[0048] Preferred, such as Figure 3 As shown, step S4 includes the following sub-steps: S41, extracting core assessment parameters of total nitrogen, total phosphorus, and dissolved oxygen from the clustered environmental parameter dataset, and performing data screening and association mapping according to the parameter requirements of the eutrophication assessment model for aquaculture areas; S42, determining the α, β, and γ adaptation coefficients corresponding to the current aquaculture area through the parameter adaptation module built into the model, and substituting the screened parameter data into the model for index calculation; S43, mapping the calculated assessment index E to the corresponding eutrophication level according to the preset eutrophication level classification standard, and generating a level identifier and index association table; S44, performing data verification on the level association table, removing abnormal mapping data, and ensuring the consistency and accuracy of the eutrophication assessment results.

[0049] Specifically, step S4, the eutrophication assessment process, ensures the accuracy and reliability of the assessment results. S41 precisely extracts core assessment parameters such as total nitrogen, total phosphorus, dissolved oxygen, pH, turbidity, and water temperature from each environmental parameter dataset after clustering in step S3. During the extraction process, the parameter requirements of the aquaculture area eutrophication assessment model are strictly followed. Data is filtered for each cluster, removing abnormal data that exceeds reasonable value ranges (abnormal data is determined based on industry standards and historical data statistics). A mapping relationship is established between core parameters and model input items to ensure that each parameter accurately corresponds to the model's computational requirements. S42, through its built-in parameter adaptation module, retrieves corresponding adaptation coefficients from a pre-set coefficient database based on information such as the type of aquaculture area (e.g., fish farming, shellfish farming, algae farming), scale of aquaculture, and geographical location. These coefficients are derived through extensive training comparing historical monitoring data with actual eutrophication conditions, and their value ranges are rigorously calibrated. Subsequently, the selected core parameter data are substituted into the model according to the correlation mapping relationship, and comprehensive calculations are performed according to the model's set operational logic to obtain the eutrophication assessment index for each cluster's corresponding time-series interval. S43, based on a pre-set eutrophication level classification standard, maps the calculated assessment index to the corresponding eutrophication level. The levels are divided into five categories: oligotrophic, mesotrophic, slightly eutrophic, moderately eutrophic, and severely eutrophic. The index value range for each level is determined through industry standards and experimental data verification. Based on the mapping results, a level identifier and index correlation table are generated, including the time-series interval, assessment index, and eutrophication level. S44 uses cross-validation to verify the data in the gradation association table. It compares the evaluation index and grade calculated by different parameter combinations within the same cluster, removes abnormal mapping data with differences exceeding the allowable range, and verifies the logical consistency of the grade division to ensure that the eutrophication assessment results of all clusters are accurate and provide reliable water quality basis data for subsequent stress response prediction.

[0050] Preferred, such as Figure 4 As shown, S5 includes the following sub-steps: S51, integrating eutrophication level correlation data and clustered multidimensional environmental parameters to construct an input dataset for stress response prediction, and clarifying the weight ratio of each parameter in the dataset; S52, importing environmental adaptation parameters corresponding to the growth cycle of aquaculture organisms, and determining the δ and ε stress prediction coefficients in the model in combination with the input dataset; S53, substituting the input dataset and prediction coefficients into the aquaculture environmental stress response prediction model, calculating the environmental stress response intensity S, and classifying the stress level based on the intensity value; S54, predicting the stress duration range according to the stress level and the time-series change trend of the parameters, and generating dynamic prediction results including intensity value, stress level and duration.

[0051] Specifically, step S5, the aquaculture environmental stress response prediction process, achieves accurate dynamic prediction of stress intensity and duration. S51 first integrates the eutrophication level correlation data generated in step S4 with the multidimensional environmental parameters clustered in step S3, removing duplicate and invalid data to construct the input dataset for stress response prediction. The weight percentage of each parameter in the dataset is determined based on its impact on environmental stress. This weight percentage is determined through the analytic hierarchy process combined with expert evaluation, with the eutrophication assessment index accounting for 30% to 40%, core environmental parameters accounting for 40% to 50%, and other auxiliary parameters accounting for 10% to 20%. S52 imports environmental adaptation parameters corresponding to the growth cycle of the currently farmed organisms (such as specific fish or shellfish), including the suitable range, tolerance threshold, and sensitivity coefficients of each environmental parameter at different stages such as seedling, growth, and maturity. Combining the parameter characteristics and distribution of the input dataset, the stress prediction coefficients in the aquaculture environmental stress response prediction model are determined through a model parameter optimization algorithm, ensuring that the coefficients are highly adapted to the current aquaculture scenario. S53 inputs the completed input dataset and determined prediction coefficients into the aquaculture environmental stress response prediction model according to the model's required format. The model's computational logic calculates the environmental stress response intensity value. Based on the intensity value range, the stress is divided into three levels: mild, moderate, and severe stress. The intensity value range for each level is verified and determined through extensive aquaculture stress experimental data. S54, based on the classified stress levels and combined with the temporal variation trends of multidimensional environmental parameters in the input dataset, uses a combination of linear fitting and trend extrapolation to predict the stress duration range. The interval prediction accuracy is controlled within ±2 hours. Simultaneously, dynamic prediction results are generated, including stress response intensity values, stress levels, duration ranges, and parameter variation trend descriptions, providing a precise and comprehensive decision-making basis for aquaculture managers to take targeted control measures.

[0052] like Figure 5As shown, an IoT big data-based marine aquaculture environment monitoring and assessment system is applied to an IoT big data-based marine aquaculture environment monitoring and assessment method. The system includes: an IoT multi-dimensional environmental parameter sensing and acquisition unit, used to collect parameters related to dissolved oxygen, nitrogen and phosphorus nutrients, pH value, water temperature, salinity, turbidity, and phytoplankton community structure through distributed sensor devices, and establishes a bidirectional data transmission connection with a high-concurrency marine environmental data processing unit; a high-concurrency marine environmental data processing unit, used to receive multi-source data transmitted by the sensing and acquisition unit, perform parallel reception and status verification through a distributed receiving module, and establish data interaction channels with a multi-dimensional time-series data adaptive clustering processing unit, an aquaculture environment assessment and prediction unit, and an assessment result integration and output unit; and a multi-dimensional time-series data processing unit. The adaptive clustering processing unit is used to call the clustering algorithm to perform time-series partitioning and similarity clustering on the verified parameters, and transmit the clustering results to the aquaculture environment assessment and prediction unit; the aquaculture area water quality eutrophication assessment unit is used to run the eutrophication assessment model based on the clustering results, generate eutrophication level association data and send it to the aquaculture environment stress response prediction unit; the aquaculture environment stress response prediction unit is used to receive the eutrophication level association data and clustering parameters, combine them with biological growth cycle adaptation parameters to predict the stress intensity and duration, and transmit the prediction results to the assessment result integration and output unit; the assessment result integration and output unit is used to receive the assessment results and prediction results, perform structured integration and generate an assessment report including parameter trajectories, level identifiers and trend curves and output it.

[0053] A method and system for monitoring and assessing marine aquaculture environment using IoT big data is presented. This system employs IoT sensing devices to comprehensively collect multi-dimensional environmental parameters, covering key influencing factors of the aquaculture environment. Combined with the distributed reception and verification functions of a high-concurrency marine environmental data processing platform, it significantly improves the efficiency and transmission stability of multi-source data processing. An innovative multi-dimensional time-series data adaptive clustering algorithm is applied, dynamically adapting to the temporal changes in parameters to construct clusters, making data segmentation more closely aligned with actual aquaculture environment differences. A dedicated eutrophication assessment model and environmental stress response prediction model are integrated, combined with parameters adapted to the growth cycle of aquaculture organisms, to achieve deep linkage between assessment and prediction, resulting in more targeted and scientific output. The system's six functional units work in synergy, forming a closed-loop architecture from data acquisition, processing, analysis to result output, ensuring smooth connection and efficient operation throughout the entire monitoring and assessment process.

[0054] This method and system address the problem of insufficient environmental variation reflection caused by the fixed-dimensional segmentation of traditional methods. By leveraging the dynamic clustering capability of multi-dimensional time-series data adaptive clustering algorithms, it flexibly segments data based on parameter change characteristics, making the input data of the assessment model more targeted and accurately capturing the environmental change patterns of different time periods and regions. To address the shortcomings of existing systems such as insufficient data processing capabilities and the disconnect between assessment and prediction, a high-concurrency marine environmental data processing platform is used to handle the high-concurrency transmission and processing needs of multi-source data. At the same time, eutrophication assessment results and clustering parameters are deeply integrated into the stress response prediction model, and biological growth cycle parameters are combined to improve the dynamics of prediction, providing timely and accurate decision support for aquaculture environment regulation and comprehensively solving the core pain points of existing technologies.

[0055] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," "link," and "fix" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0056] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for monitoring and assessing the marine aquaculture environment using Internet of Things (IoT) big data, characterized in that, Includes the following steps: S1. Collect multi-dimensional environmental parameters of the marine aquaculture area through IoT sensing devices. These parameters include dissolved oxygen content, nitrogen and phosphorus nutrient concentration, pH value, water temperature, salinity, turbidity, and phytoplankton community structure-related parameters. S2. Transmit the collected multi-dimensional environmental parameters to a high-concurrency marine environmental data processing platform. The platform's distributed data receiving module performs parallel reception of multi-source data and data transmission status verification. S3. Call a multi-dimensional time-series adaptive clustering algorithm to perform time-series dimension partitioning and similarity clustering on the status-verified multi-dimensional environmental parameters, constructing dynamic clusters based on the time-series change characteristics of the aquaculture environmental parameters. S4. Based on the clustered data... The environmental parameter dataset is used to run the eutrophication assessment model for aquaculture areas to quantitatively assess the eutrophication level of the aquaculture areas and generate eutrophication level correlation data. In step S5, the eutrophication level correlation data and clustered multidimensional environmental parameters are input into the aquaculture environmental stress response prediction model. Combined with environmental adaptation parameters related to the growth cycle of aquaculture organisms, the intensity and duration of aquaculture environmental stress are dynamically predicted. Finally, a high-concurrency marine environmental data processing platform is used to integrate and structure the eutrophication assessment results and stress response prediction results, generating a marine aquaculture environmental monitoring and assessment report that includes parameter change trajectories, assessment level identifiers, and prediction trend curves.

2. The method for monitoring and assessing marine aquaculture environment using IoT big data according to claim 1, characterized in that, The expression for the eutrophication assessment model of the aquaculture area is as follows: ,in, It is a water eutrophication assessment index. For model fit coefficients, 1 is the dissolved oxygen content parameter, TP is the total phosphorus concentration parameter, TN is the total nitrogen concentration parameter, PH is the pH value parameter, TURB is the turbidity parameter, and TEMP is the water temperature parameter.

3. The method for monitoring and assessing marine aquaculture environment using IoT big data according to claim 1, characterized in that, The core expression of the multidimensional time-series adaptive clustering algorithm is: ,in, This is the cluster similarity threshold. For the number of clusters, For the first The and the first Similarity measures of time-series environmental parameters, The time interval for collecting the two parameters. This is the time-series decay coefficient. For a single environmental parameter time series data point, This is the time-series mean of this type of parameter.

4. The method for monitoring and assessing the marine aquaculture environment using IoT big data according to claim 1, characterized in that, The expression for the prediction model of the aquaculture environment stress response is: ,in, This represents the intensity value of the environmental stress response. , The stress prediction coefficient, The eutrophication assessment index is SAL, which is the salinity parameter. For the first Community proportion parameters of planktonic organisms For the first Environmental stress sensitivity coefficients corresponding to planktonic organisms This represents the number of categories in the planktonic community.

5. The method for monitoring and assessing the marine aquaculture environment using IoT big data according to claim 1, characterized in that, The data processing efficiency model of the high-concurrency marine environmental data processing platform is as follows: ,in, For platform data processing speed, For the number of concurrent data transmission channels, This is the data transmission volume per channel. This serves as a baseline for data processing latency. This represents the channel interference coefficient.

6. The method for monitoring and assessing the marine aquaculture environment using IoT big data according to claim 1, characterized in that, The temporal feature extraction model for the multidimensional environmental parameters is as follows: ,in, For time series feature quantization values, for The variance of environmental parameters at any given time. for Time and Covariance of time parameters These are the feature weight coefficients. for Skewness of time series parameters.

7. The method for monitoring and assessing marine aquaculture environment using IoT big data according to claim 1, characterized in that, S3 includes the following steps: S31, constructing a time-series coordinate system based on the collection timestamps of multi-dimensional environmental parameters, setting the time window length according to the environmental change pattern within the breeding cycle, and dividing the complete time-series data into several continuous time-series segments; S32, calculating the rate of change and fluctuation amplitude of each environmental parameter within each time-series segment, establishing parameter change feature vectors, and eliminating the dimensional differences of different parameters through vector standardization; S33, using an adaptive distance metric method to calculate the similarity between feature vectors of each time-series segment, and dynamically adjusting the cluster size threshold according to the similarity distribution; S34, clustering the time-series segments according to the adjusted threshold, merging time-series segments with similarity higher than the threshold to form clusters, and recording the calibration parameter characteristics and time-series interval range of each cluster.

8. The method for monitoring and assessing the marine aquaculture environment using IoT big data according to claim 1, characterized in that, S4 includes the following steps: S41, extracting core assessment parameters of total nitrogen, total phosphorus, and dissolved oxygen from the clustered environmental parameter dataset, and performing data screening and association mapping according to the parameter requirements of the eutrophication assessment model for aquaculture areas; S42, determining the α, β, and γ adaptation coefficients corresponding to the current aquaculture area through the parameter adaptation module built into the model, and substituting the screened parameter data into the model for index calculation; S43, mapping the calculated assessment index E to the corresponding eutrophication level according to the preset eutrophication level classification standard, and generating a level identifier and index association table; S44, performing data verification on the level association table, removing abnormal mapping data, and ensuring the consistency and accuracy of the eutrophication assessment results.

9. The method for monitoring and evaluating the marine aquaculture environment using IoT big data according to claim 1, characterized in that, S5 includes the following sub-steps: S51, integrating eutrophication level correlation data and clustered multidimensional environmental parameters to construct an input dataset for stress response prediction, and clarifying the weight ratio of each parameter in the dataset; S52, importing environmental adaptation parameters corresponding to the growth cycle of aquaculture organisms, and determining the δ and ε stress prediction coefficients in the model in combination with the input dataset; S53, substituting the input dataset and prediction coefficients into the aquaculture environmental stress response prediction model to calculate the environmental stress response intensity S, and classifying the stress level based on the intensity value; S54, predicting the stress duration range according to the stress level and the time-series change trend of the parameters, and generating dynamic prediction results including intensity value, stress level and duration.

10. The system corresponding to the marine aquaculture environment monitoring and assessment method based on IoT big data as described in any one of claims 1-9, characterized in that, include: The IoT multidimensional environmental parameter sensing and acquisition unit is used to collect parameters related to dissolved oxygen, nitrogen and phosphorus nutrients, pH value, water temperature, salinity, turbidity and plankton community structure through distributed sensor devices, and establish a two-way data transmission connection with the marine environmental data high-concurrency processing unit. The high-concurrency marine environmental data processing unit is used to receive multi-source data transmitted by the sensing and acquisition unit. It performs parallel reception and status verification through a distributed receiving module and establishes data interaction channels with the multi-dimensional time-series data adaptive clustering processing unit, the aquaculture environment assessment and prediction unit, and the assessment result integration and output unit, respectively. The multidimensional time-series data adaptive clustering processing unit is used to call the clustering algorithm to perform time-series partitioning and similarity clustering on the verified parameters, and transmit the clustering results to the aquaculture environment assessment and prediction unit. The eutrophication assessment unit for aquaculture water quality is used to run the eutrophication assessment model based on the clustering results, generate eutrophication level correlation data, and send it to the aquaculture environmental stress response prediction unit. The aquaculture environment stress response prediction unit is used to receive eutrophication level correlation data and clustering parameters, combine them with biological growth cycle adaptation parameters to predict stress intensity and duration, and transmit the prediction results to the evaluation result integration output unit. The evaluation result integration and output unit is used to receive evaluation results and prediction results, perform structured integration, generate and output an evaluation report including parameter trajectory, level label and trend curve.

Images