Wetland ecological risk control method, system and equipment based on multimodal early warning

Abstract

The present invention discloses a wetland ecological risk control method, system, and equipment based on multimodal early warning, which relates to the field of water quality parameter abnormality alarm. The method includes: real-time collection of physical and chemical parameters of wetland water bodies to generate a dynamic water quality data set; detection and optimization of the dynamic water quality data set based on the frequency of historical abnormal data to determine a water quality detection sequence; multi-level risk assessment of the target wetland area according to the water quality detection sequence to obtain multiple risk scores; triggering a multimodal early warning mechanism based on the multiple risk scores, generating an adaptive purification plan and pushing it to the management terminal. This solves the technical problems of existing wetland water quality monitoring in terms of insufficient early warning timeliness, monitoring comprehensiveness, and data accuracy, and achieves the technical effect of improving the early warning timeliness, monitoring comprehensiveness, and data accuracy of wetland water quality monitoring.

Description

Technical Field

[0001] The present application relates to the field related to abnormal alarm of water quality parameters, and in particular to wetland ecological risk control methods, systems and equipment based on multimodal early warning. Background Art

[0002] Effective monitoring and management of wetland water quality are crucial for maintaining wetland ecological balance, ensuring water resource security, and promoting regional sustainable development. Currently, wetland water quality monitoring primarily relies on traditional manual sampling and laboratory analysis. This method is not only inefficient, making it difficult to issue early warning signals when water quality anomalies occur, but also suffers from a limited monitoring scope and insufficient data continuity, resulting in monitoring results that fail to fully and accurately reflect the dynamic changes in the overall wetland water quality. This lagging and one-sided monitoring situation makes it difficult for management departments to obtain timely and effective early warning information, which in turn seriously hinders the scientific formulation and rapid response of wetland protection and management decisions.

[0003] Among the current related technologies, wetland water quality monitoring has technical problems such as insufficient warning timeliness, comprehensive monitoring and data accuracy. Summary of the Invention

[0004] This application solves the technical problems of existing wetland water quality monitoring such as insufficient warning timeliness, comprehensive monitoring and data accuracy in terms of warning timeliness, comprehensive monitoring and data accuracy by providing a wetland ecological risk control method, system and equipment based on multimodal early warning, and adopts real-time collection of physical and chemical parameters of wetland water bodies to generate a dynamic data set, optimizes the data set according to the frequency of historical abnormal data to determine the water quality detection sequence, and then conducts multi-level risk assessment on the target wetland area according to this sequence to obtain multiple risk scores. Finally, the multimodal early warning mechanism is triggered based on the score and an adaptive purification plan is generated and pushed to the management terminal. The technical means achieves the technical effect of improving the warning timeliness, comprehensive monitoring and data accuracy of wetland water quality monitoring.

[0005] The present application provides a wetland ecological risk control method based on multimodal early warning, including: real-time collection of physical and chemical parameters of wetland water bodies to generate a dynamic water quality data set; detection and optimization of the dynamic water quality data set based on the frequency of historical abnormal data to determine a water quality detection sequence; multi-level risk assessment of the target wetland area according to the water quality detection sequence to obtain multiple risk scores, triggering a multimodal early warning mechanism based on the multiple risk scores, generating an adaptive purification plan and pushing it to the management terminal.

[0006] In a possible implementation, the dynamic water quality data set is detected and optimized according to the frequency of historical abnormal data, and a water quality detection sequence is determined, and the following processing is performed: an abnormality analysis is performed on the multi-source water quality parameters of the target wetland area according to the historical monitoring period, and parameter abnormal event data is determined, and a spatiotemporal distribution feature matrix is ​​constructed according to the parameter abnormal event data; an abnormal correlation calculation is performed on the multi-source water quality parameters based on the spatiotemporal distribution feature matrix to generate a parameter detection priority weight matrix; the multi-source sensor network is detected and sorted according to the parameter detection priority weight matrix to construct the water quality detection sequence.

[0007] In a possible implementation, an abnormal correlation calculation is performed on the multi-source water quality parameters based on the spatiotemporal distribution characteristic matrix to generate a parameter detection priority weight matrix, and the following processing is performed: an abnormal analysis is performed on the multi-source water quality parameters based on the spatiotemporal distribution characteristic matrix to determine multiple abnormal events, and the multiple abnormal events include multiple abnormal occurrence probabilities; spatial correlation analysis is performed on the multiple abnormal events according to the multiple abnormal occurrence probabilities to obtain spatial aggregation characteristics of the multiple abnormal events; weights are assigned to adjacent monitoring nodes by traversing the target wetland area according to the spatial aggregation characteristics to construct the parameter detection priority weight matrix.

[0008] In a possible implementation, the multi-source sensor network is sorted for detection according to the parameter detection priority weight matrix, the water quality detection sequence is constructed, and the following processing is performed: the target wetland area is traversed to perform water quality exceeding standard analysis, and water quality exceeding standard events are determined, and the water quality exceeding standard events include spatiotemporal distribution characteristics, and the spatiotemporal distribution characteristics correspond to the water quality exceeding standard events; the parameter detection priority weight matrix is ​​mapped to the multi-source sensor network for detection identification, and multiple sensor priority identifications are determined; the detection tasks are allocated in descending order according to the multiple sensor priority identifications to construct the water quality detection sequence.

[0009] In a possible implementation, a multi-level risk assessment is performed on the target wetland area according to the water quality detection sequence to obtain multiple risk scores, and the following processing is performed: the target wetland area is traversed according to the water quality detection sequence to determine the first pollution point; a spatial coordinate system of the target wetland area is constructed, and the first spatial coordinate data of the first pollution point is extracted; pollution analysis is performed according to the spatiotemporal distribution feature matrix combined with the first spatial coordinate data to determine a pollution impact data set, and the pollution impact data set includes pollutant concentration extreme value data and pollutant impact depth data; a local pollution risk assessment is performed on the target wetland area based on the first spatial coordinate data combined with the pollutant concentration extreme value data to generate a first risk score; an overall ecological risk assessment is performed on the target wetland area based on the first spatial coordinate data combined with the pollutant impact depth data to generate a second risk score; the first risk score and the second risk score are associated and integrated with the first pollution point to obtain the multiple risk scores.

[0010] In a possible implementation, a multimodal early warning mechanism is triggered based on the multiple risk scores, and the following processing is performed: a local diffusion analysis is performed on the first pollution point based on the first risk score, and a local pollution diffusion trend map is drawn; an ecological risk analysis is performed on the first pollution point based on the second risk score, and the overall ecological risk level of the wetland is set; pollution diffusion calculation is performed based on the local pollution diffusion trend map to generate a pollution diffusion rate; wetland self-purification calculation is performed based on the overall ecological risk level of the wetland to generate a wetland self-purification rate; a balance analysis is performed based on the pollution diffusion rate and the wetland self-purification rate to generate a dynamic balance coefficient; and the multimodal early warning mechanism is triggered based on the dynamic balance coefficient.

[0011] In a possible implementation, the multimodal early warning mechanism is triggered according to the dynamic balance coefficient, and the following processing is performed: the abnormal critical value is defined according to the parameter abnormal event data, and the expected balance numerical range is set according to the abnormal critical value; whether the dynamic balance coefficient is in the expected balance numerical range is determined, when the dynamic balance coefficient is not in the expected balance numerical range, the type of the exceeded parameter is determined according to the parameter abnormal event data, and the spatial distribution characteristics are determined according to the spatiotemporal distribution characteristics; the multimodal early warning mechanism is triggered based on the exceeded parameter type and the spatial distribution characteristics.

[0012] In a possible implementation, the process of generating an adaptive purification plan and pushing it to the management terminal performs the following processing: matching the multi-level warning mode based on the type of the exceeded parameter and the spatial distribution characteristics, and generating a multimodal warning linkage instruction according to the matching result; adaptively adjusting the target wetland area through the multimodal warning linkage instruction to generate an adaptive purification plan; pushing the adaptive purification plan to the management terminal through the blockchain encryption channel, and synchronously updating the control log of the target wetland area.

[0013] The present application also provides a wetland ecological risk control system based on multimodal early warning, including: a dynamic water quality data set generation module, which is used to collect the physical and chemical parameters of wetland water bodies in real time to generate a dynamic water quality data set; a water quality detection sequence determination module, which is used to detect and optimize the dynamic water quality data set according to the frequency of historical abnormal data and determine the water quality detection sequence; an adaptive purification scheme generation module, which is used to perform multi-level risk assessment on the target wetland area according to the water quality detection sequence, obtain multiple risk scores, trigger a multimodal early warning mechanism based on the multiple risk scores, generate an adaptive purification scheme and push it to the management terminal.

[0014] The present application also provides an electronic device, comprising: a memory for storing executable instructions; and a processor for implementing a wetland ecological risk control method based on multimodal early warning when executing the executable instructions stored in the memory.

[0015] The proposed wetland ecological risk control method, system, and equipment based on multimodal early warning will first collect the physical and chemical parameters of wetland water bodies in real time to generate a dynamic water quality data set. This data set will then be tested and optimized based on the frequency of historical abnormal data to determine a water quality detection sequence. Finally, a multi-level risk assessment will be conducted on the target wetland area according to the water quality detection sequence to obtain multiple risk scores. Based on these multiple risk scores, a multimodal early warning mechanism will be triggered to generate an adaptive purification plan and push it to the management terminal. This will achieve the technical effect of improving the timeliness of early warning, comprehensiveness of monitoring, and data accuracy of wetland water quality monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention are briefly introduced below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in precise order. Instead, various steps may be processed in reverse order or simultaneously as needed. Furthermore, other operations may be added to these processes, or one or more operations may be removed from these processes.

[0017] Figure 1A flow chart of a wetland ecological risk control method based on multimodal early warning provided in an embodiment of the present application.

[0018] Figure 2 A schematic diagram of the structure of a wetland ecological risk control system based on multimodal early warning provided in an embodiment of the present application.

[0019] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0020] Description of the accompanying drawings: dynamic water quality data set generation module 10, water quality detection sequence determination module 20, adaptive purification scheme generation module 30, input device 301, processor 302, memory 303, output device 304. DETAILED DESCRIPTION

[0021] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0022] In order to make the purpose, technical solutions and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limiting this application. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0023] In the following description, reference is made to “some embodiments”, which describes a subset of all possible embodiments, but it will be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms “first\second” involved are merely to distinguish similar objects and do not represent a specific ordering of the objects. The terms “including” and “having” and any variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application belongs. The terms used herein are for the purpose of describing the embodiments of this application only.

[0024] The present application embodiment provides a wetland ecological risk control method based on multimodal early warning, such as Figure 1 As shown, the method includes:

[0025] Step S100 , collecting physical and chemical parameters of wetland water bodies in real time to generate a dynamic water quality data set.

[0026] Specifically, various types of sensors are strategically placed throughout the wetland water, including dissolved oxygen sensors, ORP (oxidation-reduction potential) sensors, pH sensors, and turbidity sensors. These sensors transmit collected data in real time to a data acquisition terminal via wireless communication modules (such as LoRa and NB-IoT). Dissolved oxygen sensors use the fluorescence method to determine dissolved oxygen concentration by measuring changes in fluorescence intensity caused by the interaction between a fluorescent substance on a fluorescent probe and dissolved oxygen molecules in the water. ORP sensors use electrochemical principles to measure the redox state of the water by measuring the potential difference between an electrode and the water. pH sensors, based on the glass electrode method, determine pH by measuring the activity of hydrogen ions in the water. Turbidity sensors use the scattered light method to determine turbidity by measuring changes in light intensity after incident light is scattered in the water.

[0027] After receiving data from the sensors, the data acquisition terminal performs preliminary preprocessing, including data format conversion and filtering (such as using the Kalman filter algorithm to remove noise interference) to ensure data accuracy and integrity. The processed data is then stored in a local database, and a dynamic water quality dataset is generated based on a time series. This dynamic water quality dataset is a collection of water quality data that is continuously updated over time and can reflect the changing water quality of the wetland water body at different points in time.

[0028] For example, in a river within a wetland, sensor nodes are deployed every 50 meters, each containing the four sensors mentioned above. The sensors collect data every 10 minutes and transmit it to a nearby data collection terminal via a LoRa module. Upon receiving the data, the terminal first filters it to remove noise caused by current fluctuations and other factors. The data is then stored in a local database, forming a dynamic water quality dataset. Each record in the dataset includes information such as the acquisition time, sensor number, dissolved oxygen concentration, ORP value, pH value, and turbidity.

[0029] Step S200 , performing detection optimization on the dynamic water quality data set according to the frequency of historical abnormal data, and determining a water quality detection sequence.

[0030] Specifically, historical water quality data is extracted from the database, and the frequency of abnormalities (exceeding the normal range) for each water quality parameter is counted. For example, a SQL query can be written to count the number of times dissolved oxygen concentrations fell below the normal range (e.g., less than 3 mg / L) or the number of times ORP values ​​rose above the normal range (e.g., greater than 700 mV) over the past year. Based on the historical frequency of abnormalities, a priority queue algorithm (such as a heap-based priority queue) is used to determine the water quality testing sequence. Sensor testing tasks corresponding to water quality parameters with a high frequency of abnormalities are assigned a higher priority, allowing them to be executed first. Furthermore, a dynamic sensor sampling frequency adjustment algorithm is used to dynamically adjust the sensor sampling frequency based on real-time test results and historical data. For example, if a water quality parameter has normal test results for multiple consecutive times, the sampling frequency is appropriately reduced; if an abnormality is detected, the sampling frequency is increased for detailed monitoring. This algorithm can employ an adaptive sampling frequency adjustment algorithm based on feedback control to calculate a new sampling frequency based on the degree of deviation between the current test results and historical data.

[0031] For example, statistics show that in the past six months, the frequency of abnormal pH values ​​in the wetland water body was the highest, accounting for 40% of the total abnormal times; followed by turbidity, accounting for 30%; the frequency of abnormal dissolved oxygen and ORP values ​​was relatively low. Therefore, when determining the detection sequence, the sensor detection tasks corresponding to pH and turbidity are set to the highest priority. In the actual detection process, the initial sampling frequency of the pH sensor and turbidity sensor is set to once every 5 minutes, and the sampling frequency of the dissolved oxygen and ORP sensors is once every 10 minutes. When the pH sensor has normal test results for 5 consecutive times, its sampling frequency is adjusted to once every 10 minutes; if an abnormal pH value is detected, the sampling frequency is immediately restored to once every 5 minutes, and other related sensors are notified for collaborative monitoring.

[0032] In one possible implementation, the dynamic water quality dataset is optimized for detection based on the frequency of historical abnormal data to determine a water quality detection sequence. Step S200 further includes step S210, where anomaly analysis is performed on multi-source water quality parameters in the target wetland area according to historical monitoring cycles to determine parameter abnormality event data, and a spatiotemporal distribution feature matrix is ​​constructed based on the parameter abnormality event data. Specifically, a monitoring buffer zone with a radius of 5 km can be defined centered on the target wetland. This target wetland area serves as the specific implementation scope of the monitoring activity. Geographic Information System (GIS) technology is used to determine the boundaries of the buffer zone, and monitoring points are arranged within this zone. For example, using GIS software such as ArcGIS, a circular zone with a radius of 5 km is drawn on a map centered on the wetland, and the locations of all monitoring points within this zone are marked. Hourly water quality monitoring data for the past three years is extracted from a database, including multi-source water quality parameters such as dissolved oxygen, ORP, pH, and turbidity. For example, the database stores hourly monitoring data, as shown in Table 1.

[0033] Table 1: Example of historical monitoring data of water quality parameters from multiple sources

[0034]

[0035] The K-means clustering algorithm is used to analyze historical monitoring data and identify the spatiotemporal cluster centers of abnormal events. Each data point in the monitoring data (including time, location, and water quality parameter values) is used as input. The clustering algorithm aggregates abnormal data points together to determine the spatiotemporal distribution center of the abnormal event.

[0036] Calculate the frequency and duration of each parameter exceeding the standard. For example, for dissolved oxygen, set the normal range to 6-10 mg / L, and values ​​below 6 mg / L are considered exceeded. Count the number of dissolved oxygen exceeding the standard at each monitoring point in the hourly data, recording the frequency and duration of each exceedance.

[0037] A three-dimensional space-time cube model is constructed, where the three dimensions are longitude, latitude, and time. In this model, the density values ​​of abnormal events are plotted along the longitude, latitude, and time axes. The three-dimensional space-time cube model is the spatiotemporal distribution feature matrix described above, which is used to visually represent the distribution of abnormal events in time and space. For example, the three-dimensional space-time cube model can be represented as a three-dimensional array, where each element corresponds to the density value of abnormal events at a specific time and location. The density value can be quantified by the number of abnormal events or the degree of abnormality.

[0038] For example, suppose there are five monitoring points within a 5km radius of a target wetland, labeled A, B, C, D, and E. After extracting hourly water quality monitoring data from the past three years, a K-means clustering algorithm is used to identify the spatiotemporal cluster centers of abnormal events. For example, if the dissolved oxygen level at monitoring point C remained below 6 mg / L between 2:00 PM and 4:00 PM on July 15, 2024, this would form an abnormal event cluster center. In the 3D space-time cube model, the density value of this abnormal event is annotated as a higher value at the corresponding time and location.

[0039] Step S220, based on the spatiotemporal distribution characteristic matrix, the multi-source water quality parameters are subjected to abnormal correlation calculation to generate a parameter detection priority weight matrix. Specifically, the probability of abnormal occurrence of each water quality parameter is calculated. For example, for the dissolved oxygen parameter, the proportion of the number of abnormal occurrences in the historical data to the total number of monitoring times is counted. The spatial correlation between the water quality parameters is calculated. The spatial autocorrelation analysis method (such as Moran's I index) or the association rule mining algorithm based on machine learning is used to analyze the spatial correlation of abnormal events of different water quality parameters. For example, the spatial correlation between dissolved oxygen anomalies and pH value anomalies is calculated. If it is found that areas with dissolved oxygen anomalies are often accompanied by pH value anomalies, it is considered that the two parameters have a strong correlation in space.

[0040] Each water quality parameter is assigned a priority weight based on its anomaly probability and spatial correlation. Priority weights can be calculated using a weighted summation method: for example, weight = anomaly probability × α + spatial correlation × β, where α and β are weight coefficients that can be adjusted based on actual needs. The priority weights for each water quality parameter are organized into a matrix, with the rows and columns representing different water quality parameters, and the matrix elements representing the priority weights between the corresponding parameters.

[0041] Step S230, the multi-source sensor network is detected and sorted according to the parameter detection priority weight matrix to construct the water quality detection sequence. Specifically, the sensors in the multi-source sensor network are detected and sorted according to the parameter detection priority weight matrix. The water quality parameters corresponding to the sensors with high priority weights are detected first. For example, for the four parameters of dissolved oxygen, ORP value, pH value and turbidity, according to the priority weight matrix, the pH value has the highest priority, followed by dissolved oxygen, then ORP value, and finally turbidity. Therefore, the order of sensor detection is pH sensor, dissolved oxygen sensor, ORP sensor and turbidity sensor. The sorted sensor detection tasks are arranged in chronological order to form a water quality detection sequence. In the actual monitoring process, the detection tasks of each sensor are executed in sequence according to this sequence. For example, the detection sequence can be expressed as: pH sensor → dissolved oxygen sensor → ORP sensor → turbidity sensor. Each sensor can be tested at a set interval (e.g., every 10 minutes). For example, at minute 1, pH is tested; at minute 11, dissolved oxygen is tested; at minute 21, ORP is tested; at minute 31, turbidity is tested; and at minute 41, pH is tested again, repeating the process. This sequencing ensures that high-priority parameters (such as pH) are monitored more frequently, allowing potential water quality issues to be identified promptly.

[0042] In one possible implementation, anomaly correlation calculations are performed on the multi-source water quality parameters based on the spatiotemporal distribution feature matrix to generate a parameter detection priority weight matrix. Step S220 further includes step S221: performing anomaly analysis on the multi-source water quality parameters based on the spatiotemporal distribution feature matrix to determine multiple abnormal events, each of which includes multiple abnormal occurrence probabilities. Specifically, the anomaly analysis of the multi-source water quality parameters is performed using a spatiotemporal distribution feature matrix (i.e., a three-dimensional spatiotemporal cube model). This matrix contains the spatiotemporal distribution information of abnormal events in historical monitoring data, annotating the density of abnormal events along longitude, latitude, and time axes. Abnormal events are identified using a threshold judgment method. For each water quality parameter, a normal range is set (for example, the normal range for dissolved oxygen is 6-10 mg / L). When monitoring data exceeds this range, it is marked as an abnormal event. The abnormal occurrence probability of each abnormal event is calculated. The abnormal occurrence probability refers to the frequency of occurrence of the abnormal event in the historical monitoring data. For example, the abnormal occurrence probability can be calculated by counting the proportion of abnormal dissolved oxygen events in the total number of monitoring times. The probability of abnormal occurrence reflects the frequency of abnormality of the parameter.

[0043] Step S222, performing spatial correlation analysis on the multiple abnormal events according to the multiple abnormal occurrence probabilities, and obtaining spatial aggregation characteristics of the multiple abnormal events. Specifically, for each abnormal event, calculate the spatial correlation between it and other abnormal events. A spatial autocorrelation analysis method (such as Moran's I index) or an association rule mining algorithm based on machine learning can be used. The spatial correlation reflects the degree of spatial aggregation of abnormal events. The spatial aggregation characteristics are used to describe the spatial distribution pattern of abnormal events, reflecting whether abnormal events tend to appear in certain specific areas. For example, if two abnormal events often appear at the same time at adjacent monitoring points, their spatial correlation is high. Sort the abnormal events according to the probability of abnormal occurrence, and give priority to analyzing the spatial correlation of high-probability abnormal events. For example, the spatial correlation of pH abnormal events is analyzed first because it has the highest probability of abnormal occurrence.

[0044] For example, suppose spatial correlation analysis reveals that the spatial correlation between abnormal dissolved oxygen events and abnormal pH events is 0.7, indicating that the two events are highly spatially correlated. The spatial correlation between abnormal ORP events and abnormal turbidity events is 0.5, indicating that the two events are also spatially related. The spatial correlation between abnormal pH events and other abnormal events is generally high, indicating that abnormal pH events have strong spatial clustering characteristics.

[0045] Step S223, traverse the target wetland area according to the spatial aggregation characteristics to assign weights to adjacent monitoring nodes, and construct the parameter detection priority weight matrix. Specifically, according to the spatial aggregation characteristics of abnormal events, weights are assigned to adjacent monitoring nodes in the target wetland area to reflect the priority of each monitoring node when detecting abnormal events. For abnormal events with high spatial correlation, higher weights are assigned. For example, if the dissolved oxygen abnormal event and the pH abnormal event are highly correlated in space, these two parameters will be given a higher weight when allocating weights. Traverse all monitoring nodes in the target wetland area, and calculate the weight value of each node based on the abnormal event characteristics and spatial correlation of each node. The weight values ​​of each monitoring node are sorted into a matrix, the rows and columns of the matrix represent different water quality parameters, and the matrix elements represent the priority weights between the corresponding parameters.

[0046] In one possible implementation, the multi-source sensor network is sorted according to the parameter detection priority weight matrix to construct the water quality detection sequence. Step S230 further includes step S231, traversing the target wetland area to perform water quality exceedance analysis and determine water quality exceedance events. The water quality exceedance events include spatiotemporal distribution characteristics, and the spatiotemporal distribution characteristics correspond to the water quality exceedance events. Specifically, all monitoring nodes in the target wetland area are traversed, and the real-time water quality data of each node is analyzed for exceedances. The exceedance analysis is based on preset water quality parameter thresholds (for example, the normal range of dissolved oxygen is 6-10 mg / L, and the normal range of pH is 6.5-8.5). For each monitoring node, check whether its water quality parameters exceed the normal range. If so, record the event as a water quality exceedance event. Record the spatiotemporal distribution characteristics of each water quality exceedance event, including the time, location (longitude and latitude) and duration of occurrence.

[0047] Step S232: Mapping the parameter detection priority weight matrix to the multi-source sensor network for detection identification determines multiple sensor priority identifiers. Specifically, the parameter detection priority weight matrix is ​​mapped to the multi-source sensor network. Each sensor corresponds to a water quality parameter, and a priority identifier is assigned to each sensor based on the weight matrix. The priority identifier reflects the priority of the sensor when detecting water quality exceeding the standard. The higher the priority identifier value, the higher the priority of the sensor in the detection.

[0048] Step S233 : Assigning detection tasks in descending order based on the multiple sensor priority identifiers to construct the water quality detection sequence. Specifically, the detection tasks are assigned in descending order based on the priority identifier of each sensor. The sensor with the highest priority identifier is given priority for detection. This water quality detection sequence ensures that the highest-priority sensors are tested first.

[0049] In step S300 , a multi-level risk assessment is performed on the target wetland area according to the water quality detection sequence to obtain multiple risk scores. A multimodal early warning mechanism is triggered based on the multiple risk scores to generate an adaptive purification plan and push it to the management terminal.

[0050] Specifically, the spread trend of local pollution is assessed based on diffusion equations (such as the Gaussian diffusion model) and real-time data from water quality monitoring sequences. For example, with the pollution source as the center, the Gaussian diffusion model is used to calculate the concentration distribution of pollutants in the surrounding area over a certain period of time based on parameters such as the pollutant concentration gradient and water flow velocity. This determines the spread range and speed of the local pollution and generates a risk score. This score can be expressed as a percentage, with higher values ​​indicating a greater risk of local pollution spread.

[0051] The overall ecological risk level of wetlands is assessed by comprehensively considering various factors of wetland ecosystems, including water quality parameters, biodiversity, and vegetation coverage. For example, an assessment model based on the Analytic Hierarchy Process (AHP) is established, using water quality parameters (such as dissolved oxygen and pH value), biodiversity (such as the number of fish species and the coverage of aquatic plants), and vegetation coverage as assessment indicators. The weight of each indicator is determined through expert scoring, and then the score of each indicator is calculated based on real-time monitoring data. The final risk score for the overall ecological risk level of the wetland is obtained, which is also based on a percentage system. The higher the value, the greater the overall ecological risk of the wetland.

[0052] Different warning thresholds are set based on multiple risk scores. When the risk score exceeds the corresponding threshold, a multimodal warning is triggered. For example, when the local pollution risk score exceeds 70 points, a local pollution spread warning is triggered. Warning information is sent to management personnel via SMS, email, and push notifications on the management terminal app, and the warning information is displayed on the display screen at the wetland site. When the overall wetland ecological risk score exceeds 80 points, an overall wetland ecological risk warning is triggered. In addition to the above warning methods, the sound and light alarm devices in the wetland protection area can also be activated to alert nearby personnel.

[0053] Based on the risk score and specific pollution conditions, an expert system and machine learning algorithms are used to generate adaptive purification plans. For example, based on historical purification cases and real-time monitoring data, a machine learning algorithm (such as a decision tree algorithm) is used to build a purification plan recommendation model. Based on the risk score and pollution type (such as organic pollution, heavy metal pollution), appropriate purification measures (such as biological purification, chemical purification, etc.) are recommended. The generated purification plan is then pushed to the management terminal via the network, where the manager can view the detailed purification plan content, including purification measures, implementation steps, required equipment and materials, etc.

[0054] For example, during a particular monitoring session, the local pollution spread trend assessment model calculated a risk score of 75, while the overall wetland ecological risk level assessment model calculated a risk score of 82. Based on the warning threshold settings, both the local pollution spread warning and the overall wetland ecological risk warning were triggered. Upon receiving the warning information, the manager reviewed the adaptive purification plan, which recommended biological purification measures, including the deployment of a certain amount of purifying algae and microbial agents. It also recommended local isolation of the contaminated area to prevent further spread of the contamination. The manager then arranged for relevant personnel to carry out on-site treatment based on the plan. This embodiment of the present application utilizes real-time acquisition of wetland water physicochemical parameters to generate a dynamic dataset, optimizes this dataset based on historical abnormal data frequency to determine a water quality testing sequence, then conducts a multi-level risk assessment of the target wetland area based on this sequence to obtain multiple risk scores. Finally, based on the scores, a multimodal warning mechanism is triggered and an adaptive purification plan is generated and delivered to the management terminal. These technical approaches address the technical issues of existing wetland water quality monitoring, which suffer from insufficient warning timeliness, monitoring comprehensiveness, and data accuracy. This approach improves the timeliness, comprehensiveness, and data accuracy of wetland water quality monitoring.

[0055] In one possible implementation, a multi-level risk assessment is performed on the target wetland area according to the water quality detection sequence to obtain multiple risk scores. Step S300 further includes step S310, traversing the target wetland area according to the water quality detection sequence to determine the first pollution point. Specifically, according to the water quality detection sequence, the real-time water quality data of each monitoring node is checked in turn. The water quality parameters of each monitoring node are analyzed for exceeding the standard to determine whether they exceed the preset threshold (for example, the normal range of dissolved oxygen is 6-10 mg / L, and the normal range of pH value is 6.5-8.5). The monitoring node of the first detected water quality exceeding standard event is recorded and marked as the first pollution point.

[0056] Step S320 constructs a spatial coordinate system for the target wetland area and extracts first spatial coordinate data for the first contamination point. Specifically, the spatial coordinate system for the target wetland area is constructed using Geographic Information System (GIS) technology. The location of each monitoring node is represented by its longitude and latitude coordinates. The longitude and latitude coordinates of the first contamination point are extracted as the first spatial coordinate data.

[0057] Step S330 performs pollution analysis based on the spatiotemporal distribution feature matrix and the first spatial coordinate data to determine a pollution impact dataset. This dataset includes data on extreme pollutant concentrations and pollutant impact depths. Specifically, historical and real-time water quality data for the first pollution point and its surrounding area are analyzed using the spatiotemporal distribution feature matrix (a three-dimensional spatiotemporal cube model). For each water quality parameter (such as dissolved oxygen and pH), the minimum and maximum values ​​within a specified time range are calculated. These extreme values ​​reflect the severity of the pollution event. For example, suppose that at 00:00 on April 1, 2024, the pH value at the first pollution point C was 6.2 (lower than the normal range of 6.5-8.5) and the dissolved oxygen value was 5.8 mg / L (lower than the normal range of 6-10 mg / L). Using the spatiotemporal distribution feature matrix, data from the past 24 hours for this point and its surrounding area are extracted to calculate the lowest dissolved oxygen value of 4.5 mg / L and the lowest pH value of 6.0.

[0058] Use pollutant diffusion models (such as Gaussian diffusion models or numerical simulation methods) to calculate the spread and depth of pollutants. These models are based on principles of fluid mechanics and chemical transport, taking into account factors such as water velocity, direction, and the amount of pollutant released. Historical and real-time water quality data for the first pollution point and its surrounding area are extracted from the spatiotemporal distribution matrix to determine the extent of the pollution spread. Combined with topographic and water depth data, the depth of the pollutant's impact is calculated. GIS technology can be used to analyze topographic and water depth information. For example, assuming the pollution spread range of the first pollution point C is 1 km in radius, GIS analysis of the water depth data for this area can determine the impact depth to be 0.5 m.

[0059] Step S340: Perform a local pollution risk assessment on the target wetland area based on the first spatial coordinate data combined with the pollutant concentration extreme value data to generate a first risk score. Specifically, a preset risk assessment model is used to calculate the local pollution risk by combining the spatial coordinate data of the first pollution point and the pollutant concentration extreme value data. The risk assessment model can output a risk score based on an expert system or a machine learning algorithm. An example of a risk assessment model is as follows: First Risk Score = ∑(Pollutant Concentration Extreme Value Score × Weight), where Dissolved Oxygen Score = max(0, 6 − Dissolved Oxygen Extreme Value), pH Score = max(0, 6.5 − pH Extreme Value, pH Extreme Value − 8.5), ORP Score = max(0, 200 − ORP Extreme Value, ORP Extreme Value − 300), and Turbidity Score = max(0, Turbidity − 10). Based on the example in step S330, assuming that the dissolved oxygen weight is 0.6, the pH value weight is 0.4, the lowest dissolved oxygen value at the first pollution point C is 4.5 mg / L, and the lowest pH value is 6.0, then: First Risk Score = Dissolved Oxygen Score × 0.6 + pH Score × 0.4 = 1.5 × 0.6 + 0.5 × 0.4 = 0.9 + 0.2 = 1.1.

[0060] Step S350, conduct an overall ecological risk assessment of the target wetland area based on the first spatial coordinate data combined with the pollutant impact depth data, and generate a second risk score. Specifically, use a preset risk assessment model to combine the spatial coordinate data and pollutant impact depth data of the first pollution point to calculate the overall ecological risk. The risk assessment model can output a risk score based on an expert system or a machine learning algorithm. An example of a risk assessment model is as follows: Second risk score = ∑(pollutant impact depth score × weight). Combined with the example in step S330, the pollutant impact depth data of the first pollution point C is as follows: dissolved oxygen impact depth: 0.5m, pH value impact depth: 0.5m. Assuming that the impact depth weights of each parameter are as follows: dissolved oxygen weight: 0.6, pH value weight: 0.4, then: Second risk score = 0.5×0.6+0.5×0.4=0.5.

[0061] Step S360: Associate the first risk score and the second risk score with the first contamination point to obtain the multiple risk scores. Specifically, the first risk score and the second risk score are associated with the first contamination point to form a comprehensive risk assessment result. The comprehensive risk assessment result is used to trigger a multimodal early warning mechanism and generate an adaptive decontamination plan.

[0062] In one possible implementation, a multimodal early warning mechanism is triggered based on the multiple risk scores. Step S300 further includes step S370: performing a local diffusion analysis of the first contamination point based on the first risk score, plotting a local pollution diffusion trend graph; performing an ecological risk analysis of the first contamination point based on the second risk score, and setting an overall ecological risk level for the wetland. Specifically, a pollutant diffusion model (such as a Gaussian diffusion model or numerical simulation method) is used to simulate the local pollution diffusion at the first contamination point. Based on the first risk score, initial pollution diffusion conditions (such as pollutant concentration and diffusion range) are determined. A local pollution diffusion trend graph is plotted to display how pollution changes over time and space. Based on the second risk score, the impact of the first contamination point on the overall wetland ecology is assessed, and an overall ecological risk level for the wetland is set: low, medium, and high. Specific thresholds are set based on historical data and expert experience. For example, the first risk score for contamination point C is 1.1, and the second risk score is 0.5. Using the Gaussian diffusion model, assuming a pollution diffusion range of 1 km in radius and an impact depth of 0.5 m, a local pollution diffusion trend graph is plotted to display how pollution changes over time and space. For example, if pollution spreads to a 1km radius within an hour, the concentration gradually decreases. Based on the second risk score of 0.5, the overall ecological risk level of the wetland is set as medium (risk level classification criteria: low risk: 0-0.3; medium risk: 0.3-0.7; high risk: 0.7-1.0).

[0063] Step S380: Perform pollution diffusion calculations based on the local pollution diffusion trend map to generate a pollution diffusion rate. Specifically, extract key data from the local pollution diffusion trend map, such as the distance and time of pollution diffusion. Calculate the pollution diffusion rate using a mathematical model (such as linear regression or nonlinear fitting). Assume the following data is extracted from the local pollution diffusion trend map: pollution diffuses to a radius of 1 km within 1 hour. Calculate the pollution diffusion rate: Pollution diffusion rate = diffusion area ÷ time = π km 2 ÷1hour=πkm 2 / hour.

[0064] Step S390, performs wetland self-purification calculation based on the overall ecological risk level of the wetland to generate the wetland self-purification rate. Specifically, a wetland ecological model (such as an ecosystem model or a biogeochemical model) is used to evaluate the self-purification capacity of the wetland. For example, the ecosystem model takes into account parameters such as biodiversity, vegetation coverage, water flow velocity and water depth. The biogeochemical model takes into account parameters such as biodegradation rate and plant absorption rate. According to the overall ecological risk level of the wetland, the calculation parameters of the self-purification rate are adjusted. The specific adjustment method can adjust the key parameters in the model, such as biodegradation rate, plant absorption rate, etc., according to different risk levels. For example, if the overall ecological risk level of the wetland is medium, the biodegradation rate and plant absorption rate can be appropriately reduced to reflect the decline in the self-purification capacity of the wetland.

[0065] Step S3100, perform a balance analysis based on the pollution diffusion rate and the wetland self-purification rate to generate a dynamic balance coefficient. Specifically, compare the pollution diffusion rate and the wetland self-purification rate to evaluate the equilibrium state of the wetland ecosystem. Use a mathematical model (such as a differential equation or a dynamic system model) to calculate the dynamic balance coefficient. For example, assume that the pollution diffusion rate is πkm² / hour and the wetland self-purification rate is 10km² / year. Convert the pollution diffusion rate and the wetland self-purification rate to the same unit (such as km² / year): Pollution diffusion rate = πkm 2 / hour×24hours / day×365days / year=8760πkm 2 / year, calculate the dynamic balance coefficient: Dynamic balance coefficient = wetland self-purification rate ÷ pollution diffusion rate = 10km² / year ÷ 8760πkm 2 / year≈0.00036.

[0066] Step S3110: Trigger the multimodal early warning mechanism based on the dynamic balance coefficient. Specifically, the multimodal early warning mechanism is triggered based on the dynamic balance coefficient. The multimodal early warning mechanism can include various methods such as text messages, emails, and app push notifications. A warning threshold is set, and an early warning is triggered when the dynamic balance coefficient falls below the set threshold.

[0067] In one possible implementation, the multimodal early warning mechanism is triggered based on the dynamic equilibrium coefficient. Step S3110 further includes step S3111, where an abnormal critical value is defined based on the parameter abnormal event data, and an expected equilibrium numerical range is set based on the abnormal critical value. Specifically, the abnormal critical value for each water quality parameter is determined based on historical monitoring data and parameter abnormal event data. The abnormal critical value refers to the value above which the water quality is considered abnormal. For example, the normal range of dissolved oxygen is 6-10 mg / L, and the abnormal critical value can be set to less than 6 mg / L; the normal range of pH is 6.5-8.5, and the abnormal critical value can be set to less than 6.5 or greater than 8.5. Based on the abnormal critical value and the characteristics of the wetland ecosystem, an expected dynamic equilibrium coefficient numerical range is set. The expected equilibrium numerical range reflects the dynamic equilibrium range of the wetland ecosystem under normal conditions. For example, the expected range of the dynamic equilibrium coefficient is set to 0.0005-0.0010.

[0068] Step S3112 determines whether the dynamic balance coefficient is within the expected equilibrium value range. If the dynamic balance coefficient is not within the expected equilibrium value range, the type of parameter exceeding the standard is determined based on the parameter anomaly event data, and the spatial distribution characteristics are determined based on the spatiotemporal distribution characteristics. Specifically, the calculated dynamic balance coefficient is compared with the expected equilibrium value range to determine whether it is within the range. If the dynamic balance coefficient is below the lower limit of the expected range, it indicates that the pollution diffusion rate is far greater than the wetland's self-purification rate, and the wetland ecosystem may face a high risk. Based on the parameter anomaly event data, it is determined which parameters have exceeded the abnormal critical value. For example, if the dissolved oxygen is below 6 mg / L and the pH is below 6.5, these parameters are considered to be exceeding the standard. The spatial distribution characteristics of the exceeding parameters are determined based on the spatiotemporal distribution characteristic matrix. For example, the distribution of the exceeding parameters in the wetland can be analyzed using GIS technology to determine the concentrated areas of pollution. Suppose the calculated dynamic balance coefficient is 0.00036, which is below the expected equilibrium value range of 0.0005-0.0010. Dissolved oxygen (DO) was detected at 5.8 mg / L (less than 6 mg / L) and pH was 6.2 (less than 6.5), indicating that the parameters exceeding the standard were DO and pH. GIS analysis revealed that the exceeding parameters were primarily concentrated in the southeastern region of the wetland, specifically at monitoring points C and D.

[0069] Step S3113, triggering the multimodal early warning mechanism based on the type of parameters exceeding the standard and the spatial distribution characteristics. Specifically, the multimodal early warning mechanism is triggered according to the type of parameters exceeding the standard and the spatial distribution characteristics. The multimodal early warning mechanism may include multiple methods such as SMS, email, and App push. Set an early warning threshold, and trigger an early warning when the dynamic balance coefficient is lower than the set threshold. Assume that the set early warning threshold is 0.0005, and the dynamic balance coefficient is 0.00036, which is lower than the early warning threshold. Send early warning information to management personnel through SMS, email, App push, etc., including the type of parameters exceeding the standard (dissolved oxygen and pH value) and spatial distribution characteristics (southeast region, monitoring points C and D). Display the early warning information on the display screen at the wetland site to alert surrounding personnel.

[0070] In one possible implementation, an adaptive purification plan is generated and pushed to a management terminal. Step S300 further includes step S3120, in which a multi-level warning pattern is matched based on the type of the parameter exceeding the standard and the spatial distribution characteristics, and a multi-modal warning linkage instruction is generated based on the matching results. Specifically, a matching warning pattern is selected from a preset multi-level warning pattern library based on the type of parameter exceeding the standard and the spatial distribution characteristics. The multi-level warning pattern library contains warning strategies corresponding to different types of parameter exceeding the standard and spatial distribution characteristics. For example, if dissolved oxygen and pH values ​​exceed the standard and are mainly concentrated in the southeast area of ​​the wetland, a corresponding warning pattern is selected. Based on the matched warning pattern, a multi-modal warning linkage instruction is generated. These instructions can activate audible and visual alarms, pop-up windows on the management terminal, and emergency equipment. For example, an instruction is generated to activate the audible and visual alarm devices in the southeast area of ​​the wetland, send a pop-up notification to the management terminal, and start emergency equipment (such as the aeration system).

[0071] Step S3130, adaptively adjust the target wetland area through the multimodal early warning linkage instruction, and generate an adaptive purification plan. Specifically, according to the multimodal early warning linkage instruction, the target wetland area is adaptively adjusted. Adjustment measures include dynamically adjusting the aeration intensity and ecological restoration measures of the wetland purification system. For example, increase the aeration intensity of the aeration system to increase the dissolved oxygen level; introduce biological purifiers to improve the pH value. Generate an adaptive purification plan based on the adjustment measures. The plan includes specific purification measures, implementation steps, required equipment and materials, etc. For example, the plan recommends increasing the aeration intensity of the aeration system in the southeast region from 100m³ per hour to 150m³, and introducing a certain amount of purification algae and microbial agents, including 100kg of purification algae and 50kg of microbial agents.

[0072] In step S3140, the adaptive purification scheme is pushed to the management terminal through the blockchain encrypted channel, and the control log of the target wetland area is updated synchronously. Specifically, the adaptive purification scheme is securely pushed to the management terminal through an encrypted channel using blockchain technology. Blockchain technology ensures the immutability and transparency of data. For example, using the Hyperledger Fabric blockchain framework, the purification scheme is encrypted and sent to the management terminal. The implementation status of the purification scheme is synchronously updated in the control log of the target wetland area. The control log records the implementation time, location, effect and other information of all purification measures. For example, it is recorded at 12:00 on October 1, 2024 that the aeration intensity in the southeast region is increased to 150m³ / h, and 100kg of purification algae and 50kg of microbial agents are added.

[0073] In the above, refer to Figure 1 The wetland ecological risk control method based on multimodal early warning according to the embodiment of the present invention is described in detail. Figure 2 A wetland ecological risk control system based on multimodal early warning according to an embodiment of the present invention is described.

[0074] The wetland ecological risk control system based on multimodal early warning according to an embodiment of the present invention is designed to address the technical issues of insufficient early warning timeliness, monitoring comprehensiveness, and data accuracy in existing wetland water quality monitoring, thereby achieving the technical effect of improving the timeliness of early warning, comprehensiveness of monitoring, and data accuracy of wetland water quality monitoring. The wetland ecological risk control system based on multimodal early warning includes: a dynamic water quality dataset generation module 10, a water quality detection sequence determination module 20, and an adaptive purification solution generation module 30.

[0075] The dynamic water quality data set generation module 10 is used to collect the physical and chemical parameters of the wetland water body in real time to generate a dynamic water quality data set; the water quality detection sequence determination module 20 is used to detect and optimize the dynamic water quality data set based on the frequency of historical abnormal data and determine the water quality detection sequence; the adaptive purification scheme generation module 30 is used to perform multi-level risk assessment on the target wetland area according to the water quality detection sequence, obtain multiple risk scores, trigger a multimodal early warning mechanism based on the multiple risk scores, generate an adaptive purification scheme and push it to the management terminal.

[0076] The specific configuration of the water quality detection sequence determination module 20 will be described in detail below. As described above, the dynamic water quality data set is detected and optimized based on the frequency of historical abnormal data to determine the water quality detection sequence. The water quality detection sequence determination module 20 may further include: a spatiotemporal distribution feature matrix construction unit for performing anomaly analysis on the multi-source water quality parameters of the target wetland area according to the historical monitoring period, determining parameter abnormal event data, and constructing a spatiotemporal distribution feature matrix based on the parameter abnormal event data; an anomaly correlation calculation unit for performing anomaly correlation calculation on the multi-source water quality parameters based on the spatiotemporal distribution feature matrix to generate a parameter detection priority weight matrix; and a detection sorting unit for sorting the multi-source sensor network according to the parameter detection priority weight matrix to construct the water quality detection sequence.

[0077] Among them, based on the spatiotemporal distribution characteristic matrix, the multi-source water quality parameters are subjected to abnormal correlation calculation to generate a parameter detection priority weight matrix. The abnormal correlation calculation unit may further include: an abnormal event determination subunit for performing abnormal analysis on the multi-source water quality parameters based on the spatiotemporal distribution characteristic matrix to determine multiple abnormal events, wherein the multiple abnormal events include multiple abnormal occurrence probabilities; a spatial correlation analysis subunit for performing spatial correlation analysis according to the multiple abnormal events according to the multiple abnormal occurrence probabilities to obtain spatial aggregation characteristics of the multiple abnormal events; a weight allocation subunit for traversing the target wetland area according to the spatial aggregation characteristics to perform weight allocation on adjacent monitoring nodes to construct the parameter detection priority weight matrix.

[0078] Among them, the multi-source sensor network is detected and sorted according to the parameter detection priority weight matrix to construct the water quality detection sequence, and the detection sorting unit may further include: a water quality exceeding standard event determination subunit is used to traverse the target wetland area to perform water quality exceeding standard analysis and determine the water quality exceeding standard event, the water quality exceeding standard event includes spatiotemporal distribution characteristics, and the spatiotemporal distribution characteristics correspond to the water quality exceeding standard event; the detection identification subunit is used to map the parameter detection priority weight matrix to the multi-source sensor network for detection identification and determine multiple sensor priority identifications; the detection task descending order allocation subunit is used to allocate detection tasks in descending order according to the multiple sensor priority identifications to construct the water quality detection sequence.

[0079] The specific configuration of the adaptive purification scheme generation module 30 will be described in detail below. As described above, a multi-level risk assessment is performed on the target wetland area according to the water quality detection sequence to obtain multiple risk scores. The adaptive purification scheme generation module 30 may further include: a first pollution point determination unit for traversing the target wetland area according to the water quality detection sequence to determine a first pollution point; a first spatial coordinate data extraction unit for constructing a spatial coordinate system for the target wetland area and extracting first spatial coordinate data of the first pollution point; a pollution impact data set determination unit for performing pollution analysis according to the spatiotemporal distribution feature matrix combined with the first spatial coordinate data to determine a pollution impact data set, wherein the pollution impact data set includes pollutant concentration extreme value data and pollutant impact depth data; a local pollution risk assessment unit for performing a local pollution risk assessment on the target wetland area based on the first spatial coordinate data combined with the pollutant concentration extreme value data to generate a first risk score; an overall ecological risk assessment unit for performing an overall ecological risk assessment on the target wetland area based on the first spatial coordinate data combined with the pollutant impact depth data to generate a second risk score; and an association and integration unit for associating and integrating the first risk score with the second risk score to the first pollution point to obtain the multiple risk scores.

[0080] Among them, based on the triggering of the multimodal early warning mechanism based on the multiple risk scores, the adaptive purification scheme generation module 30 may further include: a first pollution point analysis unit for performing a local diffusion analysis on the first pollution point based on the first risk score, drawing a local pollution diffusion trend map, performing an ecological risk analysis on the first pollution point based on the second risk score, and setting the overall ecological risk level of the wetland; a pollution diffusion calculation unit for performing pollution diffusion calculation based on the local pollution diffusion trend map, and generating a pollution diffusion rate; a wetland self-purification calculation unit for performing wetland self-purification calculation based on the overall ecological risk level of the wetland, and generating a wetland self-purification rate; a balance analysis unit for performing a balance analysis based on the pollution diffusion rate and the wetland self-purification rate, and generating a dynamic balance coefficient; a multimodal early warning mechanism triggering unit for triggering the multimodal early warning mechanism based on the dynamic balance coefficient.

[0081] Among them, the multimodal early warning mechanism is triggered according to the dynamic balance coefficient, and the multimodal early warning mechanism triggering unit may further include: an expected balance numerical interval setting subunit for defining the abnormal critical value according to the parameter abnormal event data, and setting the expected balance numerical interval according to the abnormal critical value; a judgment subunit for judging whether the dynamic balance coefficient is in the expected balance numerical interval, and when the dynamic balance coefficient is not in the expected balance numerical interval, determining the type of the exceeded parameter according to the parameter abnormal event data, and determining the spatial distribution characteristics according to the spatiotemporal distribution characteristics; a multimodal early warning mechanism triggering subunit for triggering the multimodal early warning mechanism based on the exceeded parameter type and the spatial distribution characteristics.

[0082] Among them, an adaptive purification plan is generated and pushed to the management terminal. The adaptive purification plan generation module 30 may further include: a multimodal early warning linkage instruction generation unit is used to match the multi-level early warning mode based on the type of the exceeded parameter and the spatial distribution characteristics, and generate a multimodal early warning linkage instruction according to the matching result; an adaptive adjustment unit is used to adaptively adjust the target wetland area through the multimodal early warning linkage instruction to generate an adaptive purification plan; a plan push unit is used to push the adaptive purification plan to the management terminal through the blockchain encryption channel, and synchronously update the control log of the target wetland area.

[0083] The wetland ecological risk control system based on multimodal early warning provided by the embodiment of the present invention can execute the wetland ecological risk control method based on multimodal early warning provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0084] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other and are not used to limit the scope of protection of the present invention.

[0085] Based on the foregoing embodiments, an embodiment of the present application further provides an electronic device. Figure 3 1 is a schematic structural diagram of an electronic device provided by an embodiment of the present invention, showing a block diagram of an exemplary electronic device suitable for implementing an embodiment of the present invention. Figure 3The electronic device shown is merely an example and should not limit the functionality and scope of use of the embodiments of the present invention. The electronic device is implemented as a general-purpose computing device, and its components may include, but are not limited to, an input device 301, a processor 302, a memory 303, and an output device 304. The processor 302 may be one or more; the memory 303 may include a computer-readable medium and at least one program product, which has a set (at least one) of program modules configured to perform the functions of the various embodiments of the present application.

[0086] The memory 303 shown in the embodiment of the present invention may adopt any combination of one or more computer-readable media; the computer-readable storage medium may be, but is not limited to, an infrared, semiconductor system, device or component, or any combination of the above, for storing software programs, computer executable programs and modules, such as the program instructions / modules corresponding to the wetland ecological risk control method based on multimodal early warning in the embodiment of the present invention. The processor 302 executes various functional applications and data processing of the computer device by running the software programs, instructions and modules stored in the memory 303, thereby realizing the above-mentioned wetland ecological risk control method based on multimodal early warning.

[0087] The above specific embodiments do not constitute a limitation to the scope of protection of this application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent replacements and improvements made within the spirit and principles of this application should be included in the scope of protection of this application. In some cases, the actions or steps recorded in this application can be performed in an order different from that in the embodiments and can still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A wetland ecological risk control method based on multimodal early warning, characterized by: The method comprises: Real-time collection of physical and chemical parameters of wetland water bodies to generate dynamic water quality data sets; Optimize the detection of the dynamic water quality data set according to the frequency of historical abnormal data to determine the water quality detection sequence; Performing a multi-level risk assessment on the target wetland area according to the water quality detection sequence to obtain multiple risk scores, triggering a multimodal early warning mechanism based on the multiple risk scores, generating an adaptive purification plan, and pushing it to the management terminal; The dynamic water quality data set is tested and optimized according to the frequency of historical abnormal data to determine the water quality test sequence, including: Performing anomaly analysis on multi-source water quality parameters in the target wetland area according to the historical monitoring period, determining parameter anomaly event data, and constructing a spatiotemporal distribution feature matrix based on the parameter anomaly event data; Performing abnormal correlation calculation on the multi-source water quality parameters based on the spatiotemporal distribution characteristic matrix to generate a parameter detection priority weight matrix; Sorting the multi-source sensor network for detection according to the parameter detection priority weight matrix to construct the water quality detection sequence; The abnormal correlation calculation of the multi-source water quality parameters is performed based on the spatiotemporal distribution characteristic matrix to generate a parameter detection priority weight matrix, including: Performing anomaly analysis on the multi-source water quality parameters based on the spatiotemporal distribution characteristic matrix to determine a plurality of abnormal events, wherein the plurality of abnormal events include a plurality of abnormal occurrence probabilities; Performing spatial correlation analysis on the multiple abnormal events according to the multiple abnormal occurrence probabilities to obtain spatial aggregation features of the multiple abnormal events; According to the spatial aggregation characteristics, the target wetland area is traversed to assign weights to adjacent monitoring nodes, and the parameter detection priority weight matrix is ​​constructed; The multi-source sensor network is sorted according to the parameter detection priority weight matrix to construct the water quality detection sequence, including: Traversing the target wetland area to perform water quality exceeding standard analysis and determine water quality exceeding standard events, wherein the water quality exceeding standard events include spatiotemporal distribution characteristics, and the spatiotemporal distribution characteristics correspond to the water quality exceeding standard events; Mapping the parameter detection priority weight matrix to the multi-source sensor network for detection identification, and determining multiple sensing priority identifications; The detection tasks are allocated in descending order according to the multiple sensor priority identifiers to construct the water quality detection sequence.

2. The wetland ecological risk control method based on multimodal early warning according to claim 1, characterized in that: Conduct a multi-level risk assessment of the target wetland area according to the water quality testing sequence to obtain multiple risk scores, including: Traversing the target wetland area according to the water quality detection sequence to determine the first pollution point; Constructing a spatial coordinate system for the target wetland area and extracting first spatial coordinate data of the first pollution point; Performing pollution analysis according to the spatiotemporal distribution feature matrix in combination with the first spatial coordinate data to determine a pollution impact data set, wherein the pollution impact data set includes extreme value data of pollutant concentration and pollutant impact depth data; Performing a local pollution risk assessment on the target wetland area based on the first spatial coordinate data in combination with the pollutant concentration extreme value data to generate a first risk score; Performing an overall ecological risk assessment on the target wetland area based on the first spatial coordinate data combined with the pollutant impact depth data to generate a second risk score; The first risk score and the second risk score are associated and integrated with the first contamination point to obtain the multiple risk scores.

3. The wetland ecological risk control method based on multimodal early warning according to claim 2 is characterized in that: Triggering a multimodal early warning mechanism based on the multiple risk scores, the method includes: Performing a local diffusion analysis on the first pollution point based on the first risk score, drawing a local pollution diffusion trend map, and performing an ecological risk analysis on the first pollution point based on the second risk score, and setting an overall ecological risk level for the wetland; Performing pollution diffusion calculation based on the local pollution diffusion trend map to generate a pollution diffusion rate; Calculate the wetland self-purification rate based on the overall ecological risk level of the wetland; Performing a balance analysis based on the pollution diffusion rate and the wetland self-purification rate to generate a dynamic balance coefficient; The multimodal early warning mechanism is triggered according to the dynamic balance coefficient.

4. The wetland ecological risk control method based on multimodal early warning according to claim 3 is characterized in that: The multimodal early warning mechanism is triggered according to the dynamic balance coefficient, and the method includes: Determine an abnormal critical value based on the parameter abnormal event data, and set an expected balance value range based on the abnormal critical value; Determining whether the dynamic balance coefficient is within the expected balance value range; when the dynamic balance coefficient is not within the expected balance value range, determining the type of the exceeded parameter according to the parameter abnormality event data, and determining the spatial distribution characteristics according to the spatiotemporal distribution characteristics; The multimodal early warning mechanism is triggered based on the type of the exceeded parameter and the spatial distribution characteristics.

5. The wetland ecological risk control method based on multimodal early warning according to claim 4 is characterized in that: The process of generating an adaptive purification plan and pushing it to the management terminal includes: Matching a multi-level warning mode based on the type of the exceeded parameter and the spatial distribution characteristics, and generating a multi-modal warning linkage instruction according to the matching result; Adaptively adjust the target wetland area through the multimodal early warning linkage command to generate an adaptive purification plan; The adaptive purification scheme is pushed to the management terminal through the blockchain encrypted channel, and the regulation log of the target wetland area is updated synchronously.

6. A wetland ecological risk control system based on multimodal early warning is characterized by: The system is used to implement the wetland ecological risk control method based on multimodal early warning according to any one of claims 1 to 5, and the system includes: Dynamic water quality data set generation module, used to collect physical and chemical parameters of wetland water bodies in real time and generate dynamic water quality data sets; A water quality detection sequence determination module is used to perform detection optimization on the dynamic water quality data set according to the frequency of historical abnormal data and determine the water quality detection sequence; The adaptive purification scheme generation module is used to perform a multi-level risk assessment of the target wetland area according to the water quality detection sequence, obtain multiple risk scores, trigger a multimodal early warning mechanism based on the multiple risk scores, generate an adaptive purification scheme and push it to the management terminal.

7. An electronic device, characterized in that: The electronic device comprises: a memory for storing executable instructions; The processor is configured to implement the wetland ecological risk control method based on multimodal early warning as described in any one of claims 1 to 5 when executing the executable instructions stored in the memory.

Images