Autonomous water quality monitoring unmanned ship navigation control method and system based on multi-mode communication

Abstract

This invention relates to the field of unmanned surface vessel (USV) path planning, and more particularly to a navigation and control method and system for autonomous water quality monitoring USVs based on multi-mode communication. The method includes the following steps: collecting regional water quality monitoring parameters based on onboard water quality monitoring sensors and performing spatial pollution distribution mining to construct a regional pollution concentration distribution map; acquiring a regional satellite remote sensing map and performing water flow distribution dynamics evolution to construct a water flow path distribution network; performing pollutant migration logic evolution and multi-regional gradient diffusion prediction on the water flow path distribution network based on the regional pollution concentration distribution map to construct a water pollution diffusion prediction trend map; identifying the highest diffusion gradient based on the water pollution diffusion prediction trend map and analyzing the optimal navigation monitoring sequence to construct an optimal navigation monitoring sequence. This invention improves the accuracy and efficiency of water quality monitoring through dynamic and real-time adjustment of the USV's water quality cruise path.

Description

Technical Field

[0001] This invention relates to the field of unmanned vessel path planning, and in particular to a navigation and control method and system for an autonomous water quality monitoring unmanned vessel based on multi-mode communication. Background Technology

[0002] With the increasing severity of global environmental pollution, water pollution has become a significant factor restricting the ecological environment and human health. Water pollution not only affects the safety of drinking water sources but also has a profound impact on the survival of aquatic organisms, the balance of ecosystems, and the sustainable development of economic activities. Traditional water quality monitoring methods mainly rely on manual sampling and on-site analysis. These methods suffer from problems such as long sampling cycles, limited coverage, and slow response times, making it difficult to meet the needs of large-scale, real-time, and high-frequency water quality monitoring. With continuous technological advancements, especially the rapid development of unmanned and intelligent control technologies, autonomous water quality monitoring technology is gradually becoming an important means of solving the challenges of water quality monitoring.

[0003] Autonomous unmanned surface vessels (USVs) for water quality monitoring, as a novel type of water quality monitoring equipment, possess the characteristics of high efficiency, accuracy, and flexibility. They can navigate autonomously in waterways and are equipped with water quality sensors to monitor water conditions in real time, avoiding the limitations of traditional manual monitoring and enabling continuous and comprehensive monitoring of water pollution. However, the complexity of water pollution and the variability of the aquatic environment place higher demands on the navigation and control systems of USVs. USVs need to be able to navigate autonomously in complex aquatic environments and perform precise navigation control based on the distribution of water pollution concentrations to achieve efficient and accurate water pollution tracking and monitoring.

[0004] Currently, research on unmanned surface vessel (USV) navigation and control methods largely focuses on technologies such as autonomous obstacle avoidance and path planning. However, research on USV navigation and control methods based on water pollution concentration tracking is still in its early stages. In practical applications, due to the extremely uneven spatial distribution of water pollution concentration and the influence of environmental factors such as water flow and wind, intelligently guiding USVs based on the concentration distribution of water pollution and optimizing navigation paths while monitoring in real time has become a pressing problem. Furthermore, given the real-time changes in water pollution concentration, how to adjust navigation paths promptly through precise navigation control to ensure that USVs can effectively cover polluted areas, thereby improving monitoring efficiency and accuracy, is also a key direction for technological research. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention proposes an autonomous water quality monitoring unmanned vessel navigation and control method and system based on multi-mode communication, thereby solving at least one of the aforementioned technical problems.

[0006] To achieve the above objectives, this invention provides a navigation and control method for an autonomous water quality monitoring unmanned surface vessel (USV) based on multi-mode communication. The USV device includes an onboard water quality monitoring sensor and a multi-mode communication module; the method comprises the following steps:

[0007] Step S1: Collect regional water quality monitoring parameters based on shipborne water quality monitoring sensors, and conduct spatial pollution distribution mining to construct a regional pollution concentration distribution map;

[0008] Step S2: Obtain regional satellite remote sensing maps, perform dynamic evolution of water flow distribution, and construct a water flow path distribution network;

[0009] Step S3: Based on the regional pollution concentration distribution map, perform pollutant migration logic evolution and multi-regional gradient diffusion prediction on the water flow path distribution network to construct a water pollution diffusion prediction trend map;

[0010] Step S4: Identify the highest diffusion gradient based on the water pollution diffusion prediction trend map, analyze the optimal navigation monitoring sequence, and construct the optimal navigation monitoring sequence;

[0011] Step S5: Plan the pollution source coverage cruise route based on the optimal navigation monitoring sequence, conduct dynamic navigation cruise monitoring, and collect full-process cruise monitoring data;

[0012] Step S6: Based on the full-process cruise monitoring data, perform real-time cruise concentration distribution change analysis and dynamic path adjustment to construct an intelligent pollution source tracing and positioning model.

[0013] This invention utilizes shipborne water quality monitoring sensors to collect various water quality parameters (such as dissolved oxygen, ammonia nitrogen, and pH) in real time, providing a precise data foundation for water quality assessment. Through data mining and spatial analysis techniques, the unmanned surface vessel (USV) can monitor and identify the spatial distribution of pollution in real time. This distribution map clearly shows pollution hotspots, helping to determine the location and diffusion direction of pollution sources in real time. Using the USV's multi-mode communication system, water quality data is uploaded to the cloud in real time. Different communication networks (such as 4G / 5G and satellite) ensure stable data transmission in various environments, guaranteeing the accuracy and timeliness of information. The integration of high-resolution images from satellite remote sensing with data from water quality monitoring sensors provides comprehensive environmental information about the water area, including important dynamic factors such as water flow direction and velocity. Through hydrodynamic models, the evolution of water flow can be simulated, accurately depicting the water flow path network. This allows the USV to better understand the flow direction and velocity of the water area, providing a basis for subsequent path planning and pollution prediction. Constructing a water flow path distribution network provides a comprehensive understanding of the water flow situation, optimizes the USV's navigation path in complex waters, and ensures the accuracy of pollutant monitoring. Based on water flow paths and pollution concentration distribution maps, unmanned surface vessels (USVs) can predict pollutant diffusion trends in real time. The system accurately predicts the future location and concentration of pollutants by dynamically analyzing their migration with the water flow. It simulates pollution gradients in multiple areas within the water body, analyzing regional differences in pollution diffusion. This diffusion pattern helps assess the potential impact of pollutants on various water bodies, ensuring comprehensive monitoring and prevention. By predicting pollution diffusion, the system identifies the potential impact range of pollutants in advance, preventing further escalation of pollution events. Multimode communication ensures that diffusion prediction results can be fed back to the command center or other USVs in real time, forming an efficient decision-making and emergency response mechanism. Through analysis of the diffusion pattern map, the system can identify the areas with the largest gradients in pollution diffusion, prioritizing monitoring of areas with the fastest and most severe pollutant migration. This analysis ensures efficient and targeted monitoring. Based on pollution diffusion and water flow paths, the system optimizes the USV's monitoring route using intelligent algorithms (such as reinforcement learning and genetic algorithms) to ensure that pollution sources are covered as early as possible, maximizing the monitoring efficiency of the cruise. Multimode communication technology ensures real-time data synchronization between unmanned surface vessels (USVs) and other monitoring platforms. Adjustments to different paths can be instantly fed back to the cloud, coordinating the actions of other USVs in the same waters and avoiding duplicate monitoring or missing pollution hotspots. Based on the optimal navigation monitoring sequence, USVs can accurately cover the entire pollution source and its diffusion path, avoiding omissions of polluted areas. USVs dynamically navigate according to path planning and adaptively adjust their paths based on real-time water quality data. For example, when a rapid change in pollution concentration is detected in a certain area, the route can be adjusted in real time to obtain more accurate pollution data. Through continuous real-time monitoring, USVs can collect high-frequency water quality data and upload it to the cloud, ensuring the comprehensiveness and integrity of the data. Multimode communication technology guarantees efficient and stable data transmission.Based on real-time monitoring data, the system can analyze trends in pollution concentration and promptly identify pollution hotspots or areas of sudden change. It reacts instantly to the patterns of change, optimizing pathways and preventing further pollution spread. Through source tracing analysis, the system can quickly locate pollution sources and their diffusion origins, forming accurate pollution source tracing models and providing precise data support for pollution source control. Based on pollution source location data and concentration change trends, the system can intelligently adjust the path of the unmanned vessel, making monitoring more refined and maximizing its timeliness and accuracy. Dynamic path adjustment is particularly important in responding to sudden pollution events.

[0014] This specification provides a navigation and control system for an autonomous water quality monitoring unmanned surface vessel (USV) based on multimode communication, used to execute the navigation and control method for the autonomous water quality monitoring USV based on multimode communication as described above, including:

[0015] The spatial pollution distribution module is used to collect regional water quality monitoring parameters based on shipborne water quality monitoring sensors, and to mine the spatial pollution distribution to construct a regional pollution concentration distribution map.

[0016] The water flow distribution evolution module is used to acquire regional satellite remote sensing maps, perform dynamic evolution of water flow distribution, and construct a water flow path distribution network;

[0017] The diffusion prediction module is used to perform pollutant migration logic evolution and multi-regional gradient diffusion prediction on the water flow path distribution network based on the regional pollution concentration distribution map, and to construct a water pollution diffusion prediction trend map.

[0018] The navigation sequence module is used to identify the highest diffusion gradient based on the water pollution diffusion prediction trend map, analyze the optimal navigation monitoring sequence, and construct the optimal navigation monitoring sequence.

[0019] The path planning module is used to plan the pollution source coverage cruise path based on the optimal navigation monitoring sequence, perform dynamic navigation cruise monitoring, and collect full-process cruise monitoring data.

[0020] The path adjustment module is used to perform real-time analysis of changes in concentration distribution during the entire process of cruise monitoring and to dynamically adjust the path, thereby building an intelligent pollution source tracing and positioning model.

[0021] This invention collects various water quality data in real time using shipborne sensors, enabling rapid location of pollution hotspots and ensuring timely detection of pollution problems in waterways. The construction of spatial pollution distribution maps clearly displays pollution concentrations in different areas, providing crucial data support for subsequent water flow and pollution diffusion analysis. Data mining techniques effectively extract pollution information from large amounts of data, achieving efficient pollution monitoring and control. Through satellite remote sensing maps and water flow evolution models, unmanned vessels can obtain dynamic evolution information on water flow, including velocity and direction, providing a foundation for path planning and pollution diffusion prediction. The establishment of a water flow path distribution network helps understand the propagation path of water in water bodies, making subsequent path planning more accurate and avoiding monitoring errors caused by changes in water flow direction. Pollutant migration and evolution prediction allows for the early identification of pollution diffusion paths and speeds, helping to formulate timely countermeasures. Multi-regional gradient diffusion prediction assesses the pollution diffusion situation in different areas, helping to determine the potential impact range of pollution and providing accurate data for subsequent monitoring and remediation. Based on the diffusion prediction trend map, early warning can be achieved, preventing pollution spread and reducing the difficulty and cost of pollution remediation. By analyzing the optimal navigation monitoring sequence, unmanned surface vessels (USVs) can prioritize monitoring in areas with the most severe pollution spread, improving the overall efficiency of water quality monitoring. Optimizing the navigation sequence avoids duplicate monitoring or missed detection of polluted areas, improving the coverage and accuracy of water quality monitoring. Multiple USVs can work collaboratively according to the optimized navigation monitoring sequence, avoiding conflicts and improving the efficiency of the entire monitoring system. Through the path planning module, USVs can accurately cover pollution sources, ensuring comprehensive monitoring and data collection. Path planning allows USVs to avoid unnecessary path duplication and time waste, thereby improving monitoring efficiency and saving energy. Dynamic navigation allows USVs to flexibly adjust their navigation paths according to real-time conditions (such as changes in water flow and sudden pollution), ensuring efficient completion of monitoring tasks. By analyzing cruise concentration changes in real time, the system can quickly respond to sudden changes in water quality and adjust the path in a timely manner, ensuring effective tracking and location of pollution sources. Dynamic path adjustment can optimize the monitoring path of USVs in real time according to changes in pollution concentration, thereby achieving accurate pollution source tracing and location. Through path adjustment, USVs can adjust their cruise paths according to changes in the environment and pollutants, achieving continuous and efficient monitoring and avoiding limitations in path design. Attached Figure Description

[0022] Figure 1 This is a flowchart illustrating the steps of a multi-mode communication-based autonomous water quality monitoring unmanned vessel navigation and control method according to the present invention.

[0023] Figure 2 This is a detailed flowchart illustrating the implementation steps of step S1.

[0024] Figure 3 This is a detailed flowchart illustrating the implementation steps of step S2;

[0025] Figure 4 This is a flowchart illustrating the detailed implementation steps of step S3. Detailed Implementation

[0026] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0027] This application provides a navigation and control method and system for autonomous water quality monitoring unmanned surface vessels (USVs) using multi-mode communication. The execution entities of the multi-mode communication-based autonomous water quality monitoring USV navigation and control method and system include, but are not limited to, mechanical equipment, data processing platforms, cloud server nodes, and network upload devices mounted on the system, which can be considered as general-purpose computing nodes in this application. The data processing platform includes, but is not limited to, at least one of an audio-visual management system, an information management system, and a cloud-based data management system.

[0028] Please see Figures 1 to 4 This invention provides a navigation and control method for an autonomous water quality monitoring unmanned surface vessel based on multimode communication. The method includes the following steps:

[0029] Step S1: Collect regional water quality monitoring parameters based on shipborne water quality monitoring sensors, and conduct spatial pollution distribution mining to construct a regional pollution concentration distribution map;

[0030] Step S2: Obtain regional satellite remote sensing maps, perform dynamic evolution of water flow distribution, and construct a water flow path distribution network;

[0031] Step S3: Based on the regional pollution concentration distribution map, perform pollutant migration logic evolution and multi-regional gradient diffusion prediction on the water flow path distribution network to construct a water pollution diffusion prediction trend map;

[0032] Step S4: Identify the highest diffusion gradient based on the water pollution diffusion prediction trend map, analyze the optimal navigation monitoring sequence, and construct the optimal navigation monitoring sequence;

[0033] Step S5: Plan the pollution source coverage cruise route based on the optimal navigation monitoring sequence, conduct dynamic navigation cruise monitoring, and collect full-process cruise monitoring data;

[0034] Step S6: Based on the full-process cruise monitoring data, perform real-time cruise concentration distribution change analysis and dynamic path adjustment to construct an intelligent pollution source tracing and positioning model.

[0035] This invention utilizes shipborne water quality monitoring sensors to collect various water quality parameters (such as dissolved oxygen, ammonia nitrogen, and pH) in real time, providing a precise data foundation for water quality assessment. Through data mining and spatial analysis techniques, the unmanned surface vessel (USV) can monitor and identify the spatial distribution of pollution in real time. This distribution map clearly shows pollution hotspots, helping to determine the location and diffusion direction of pollution sources in real time. Using the USV's multi-mode communication system, water quality data is uploaded to the cloud in real time. Different communication networks (such as 4G / 5G and satellite) ensure stable data transmission in various environments, guaranteeing the accuracy and timeliness of information. The integration of high-resolution images from satellite remote sensing with data from water quality monitoring sensors provides comprehensive environmental information about the water area, including important dynamic factors such as water flow direction and velocity. Through hydrodynamic models, the evolution of water flow can be simulated, accurately depicting the water flow path network. This allows the USV to better understand the flow direction and velocity of the water area, providing a basis for subsequent path planning and pollution prediction. Constructing a water flow path distribution network provides a comprehensive understanding of the water flow situation, optimizes the USV's navigation path in complex waters, and ensures the accuracy of pollutant monitoring. Based on water flow paths and pollution concentration distribution maps, unmanned surface vessels (USVs) can predict pollutant diffusion trends in real time. The system accurately predicts the future location and concentration of pollutants by dynamically analyzing their migration with the water flow. It simulates pollution gradients in multiple areas within the water body, analyzing regional differences in pollution diffusion. This diffusion pattern helps assess the potential impact of pollutants on various water bodies, ensuring comprehensive monitoring and prevention. By predicting pollution diffusion, the system identifies the potential impact range of pollutants in advance, preventing further escalation of pollution events. Multimode communication ensures that diffusion prediction results can be fed back to the command center or other USVs in real time, forming an efficient decision-making and emergency response mechanism. Through analysis of the diffusion pattern map, the system can identify the areas with the largest gradients in pollution diffusion, prioritizing monitoring of areas with the fastest and most severe pollutant migration. This analysis ensures efficient and targeted monitoring. Based on pollution diffusion and water flow paths, the system optimizes the USV's monitoring route using intelligent algorithms (such as reinforcement learning and genetic algorithms) to ensure that pollution sources are covered as early as possible, maximizing the monitoring efficiency of the cruise. Multimode communication technology ensures real-time data synchronization between unmanned surface vessels (USVs) and other monitoring platforms. Adjustments to different paths can be instantly fed back to the cloud, coordinating the actions of other USVs in the same waters and avoiding duplicate monitoring or missing pollution hotspots. Based on the optimal navigation monitoring sequence, USVs can accurately cover the entire pollution source and its diffusion path, avoiding omissions of polluted areas. USVs dynamically navigate according to path planning and adaptively adjust their paths based on real-time water quality data. For example, when a rapid change in pollution concentration is detected in a certain area, the route can be adjusted in real time to obtain more accurate pollution data. Through continuous real-time monitoring, USVs can collect high-frequency water quality data and upload it to the cloud, ensuring the comprehensiveness and integrity of the data. Multimode communication technology guarantees efficient and stable data transmission.Based on real-time monitoring data, the system can analyze trends in pollution concentration and promptly identify pollution hotspots or areas of sudden change. It reacts instantly to the patterns of change, optimizing pathways and preventing further pollution spread. Through source tracing analysis, the system can quickly locate pollution sources and their diffusion origins, forming accurate pollution source tracing models and providing precise data support for pollution source control. Based on pollution source location data and concentration change trends, the system can intelligently adjust the path of the unmanned vessel, making monitoring more refined and maximizing its timeliness and accuracy. Dynamic path adjustment is particularly important in responding to sudden pollution events.

[0036] In the embodiments of the present invention, see Figure 1 This is a flowchart illustrating the steps of an autonomous water quality monitoring unmanned vessel navigation and control method based on multi-mode communication according to the present invention. In this example, the steps of the autonomous water quality monitoring unmanned vessel navigation and control method based on multi-mode communication include:

[0037] Step S1: Collect regional water quality monitoring parameters based on shipborne water quality monitoring sensors, and conduct spatial pollution distribution mining to construct a regional pollution concentration distribution map;

[0038] In this embodiment, firstly, the configuration and calibration of the shipborne water quality monitoring sensors are ensured to be in good working order. A combination of multiple sensors is selected, such as pH sensors, turbidity sensors, dissolved oxygen sensors, and ammonia nitrogen sensors, to comprehensively monitor water quality parameters. The selection of each sensor should be based on the pollution characteristics of the target area. For example, if nitrogen pollution is known in the area, particular attention is paid to the accuracy of the ammonia nitrogen sensor. Before deployment, each sensor is calibrated to ensure its measurement accuracy. The calibration process includes comparison using standard solutions and adjusting the sensor output to ensure that its measured values ​​are consistent with the standard values. The initial calibration parameters are recorded, and the stability and accuracy of the sensors are guaranteed throughout the monitoring process. When the unmanned surface vessel (USV) is performing navigation missions, the data acquisition frequency and sample points are set. Generally, it is recommended to record water quality parameters every 5 minutes and collect samples every 10 meters during navigation to ensure the timeliness and comprehensiveness of the spatial distribution of the data. The geographical location of each sample point is recorded using the shipborne GPS system, and the data collected by the sensors (such as pH value, turbidity, dissolved oxygen concentration, etc.) is combined with their corresponding latitude and longitude to form a dataset with spatial coordinates. For example, during navigation, if the sensor detects a pH of 7.5, turbidity of 15 NTU, and dissolved oxygen of 8 mg / L at a certain point, its coordinates (x, y) are recorded as (120.5, 30.2) for subsequent analysis. After collecting the monitoring data, the data is organized and preprocessed. First, the collected data is screened to remove obvious outliers and interfering data (such as sensor malfunctions or external influences). For example, if the turbidity data at a certain measuring point suddenly jumps to 100 NTU, while the values ​​at surrounding measuring points are all around 15 NTU, this data can be considered an anomaly. Second, missing values ​​are handled by using interpolation methods (such as linear interpolation or Kriging interpolation) to fill in the missing values ​​to ensure the integrity of the dataset. Finally, the data format is standardized, and all monitoring parameters are organized into a structured data table, including timestamps, coordinates, and corresponding water quality parameter values. The preprocessed data is then used to mine the spatial distribution of pollution. Choosing a suitable spatial analysis method, Geographic Information System (GIS) technology and interpolation algorithms (such as inverse distance weighted interpolation or Kriging interpolation) are typically used to construct pollution concentration distribution maps. These methods can effectively transform discrete monitoring point data into continuous regional distribution maps. In practice, the prepared dataset is first imported into the GIS software, and an interpolation algorithm is used to generate a regional pollution concentration distribution map. By setting appropriate parameters (such as neighborhood range and interpolation method), the generated distribution map will show the changes in pollution concentration in different areas. For example, after interpolation, the ammonia nitrogen concentration in a certain area shows a gradual change from point A (concentration 20 mg / L) to point B (concentration 40 mg / L), forming a clear concentration gradient. Finally, the generated regional pollution concentration distribution map is visualized using GIS software, applying appropriate color gradients and legends to intuitively display different concentration areas.By combining concentration distribution maps with base maps, key water body features and potential pollution sources are marked, facilitating subsequent analysis and decision support. During visualization, high-pollution concentration points and flow paths within the area are recorded; this information provides crucial data for subsequent pollution source tracing and remediation strategies. For example, analyzing the distribution map might reveal that ammonia nitrogen concentrations in a particular body of water are significantly higher than in the surrounding area, potentially indicating a potential pollution source and thus guiding subsequent water quality monitoring.

[0039] Step S2: Obtain regional satellite remote sensing maps, perform dynamic evolution of water flow distribution, and construct a water flow path distribution network;

[0040] In this embodiment, high-resolution satellite images of the target area are acquired from open data platforms (such as NASA, ESA, or Google Earth Engine) using satellite remote sensing technology. The spatial resolution of the images is ensured to be sufficiently high (e.g., 10 meters) to clearly identify water bodies and surface features. The acquisition time, resolution, sensor type, and other metadata of the acquired images are recorded for subsequent data analysis. For example, the acquired image comes from the Sentinel-2 satellite, has a resolution of 10 meters, and was acquired on March 1, 2023. Necessary preprocessing is performed on the acquired satellite remote sensing images, including geometric correction, radiometric correction, and atmospheric correction. The geometric accuracy and radiometric characteristics of the images are ensured to meet the analysis requirements. For example, atmospheric correction algorithms (such as the 6S model) are used to eliminate atmospheric effects, ensuring that the spectral characteristics of the images accurately reflect surface features. The processed satellite images are then classified to identify surface features such as water bodies, soil, and vegetation. Supervised or unsupervised classification methods can be used, such as K-means clustering or random forest classification algorithms. For example, the random forest algorithm is used to classify the images, successfully identifying water areas and generating classified images. Geometric features of water bodies, including shape, area, and watershed boundaries, are extracted from classified satellite images. These features are used for subsequent hydrodynamic analysis. Relevant parameters of the water bodies, such as water area (e.g., 1000 square meters) and surrounding topographic elevation, are recorded to calculate flow velocity and direction. A suitable hydrodynamic model is selected, typically a physics-based hydrological model (e.g., SWMM or HEC-RAS), to analyze the distribution and variation of water flow within the region. Input parameters for the model are determined, including rainfall, evaporation, watershed area, and soil type. Based on the selected flow model, relevant parameters are input and simulations are performed, recording the dynamic changes of water flow under different conditions. For example, the flow velocity and direction after a rainfall event are simulated. For instance, if the simulation results show a flow velocity of 0.5 m / s for a certain water body under 50 mm of rainfall, this result is recorded for subsequent analysis. A suitable path extraction method is selected, typically using watershed analysis-based algorithms (e.g., DEM analysis), utilizing a digital elevation model (DEM) to identify water flow paths within the watershed. Ensure that elevation data and satellite imagery are geospatially aligned for accurate extraction of water flow paths. Use watershed analysis tools to identify water flow paths within the region, extracting multiple paths and recording their characteristics, including path length, direction, and velocity. For example, identify three main water flow paths within the region, recording the characteristics of path 1 (northward), path 2 (southeastward), and path 3 (southwestward). Based on the extracted water flow paths, construct a water flow path distribution network. Use graph theory methods to treat the water flow paths as nodes and edges in the network, establishing connections between paths. Record the velocity, flow rate, and connections for each path, forming a complete water flow path distribution network dataset.

[0041] Step S3: Based on the regional pollution concentration distribution map, perform pollutant migration logic evolution and multi-regional gradient diffusion prediction on the water flow path distribution network to construct a water pollution diffusion prediction trend map;

[0042] In this embodiment, pollutant concentration data within the region is acquired from a water quality monitoring system, typically in mg / L. The data ensures coverage of the entire monitoring area, including concentration changes at different time points. For example, during a certain monitoring period, the pollutant concentration in area A is recorded as 40 mg / L, in area B as 60 mg / L, and in area C as 30 mg / L. GIS software is used to generate a pollution concentration distribution map from the collected data. Blank areas within the region are filled using interpolation methods (such as Kriging interpolation or inverse distance weighting) to obtain a continuous concentration distribution. For example, the generated pollution concentration distribution map clearly shows the spatial distribution hotspots of pollutant concentration, facilitating subsequent analysis. Time series analysis is performed on the pollution concentration distribution map to identify the changing trends of pollutant concentrations in different areas. This process helps to understand the diffusion of pollutants. A suitable pollutant migration model is selected, typically using a fluid dynamics or hydrological model (such as the Advancement-Diffusion equation) to analyze the migration patterns of pollutants in the water flow path. The input parameters of the model are determined, including water flow velocity, pollutant concentration, and diffusion coefficient. Based on the selected migration model, relevant parameters are input for simulation, and the migration of pollutants under different flow velocities and concentrations is recorded. How pollutants propagate along the water flow path is monitored. For example, in the simulation, the water flow velocity is assumed to be 0.5 m / s and the diffusion coefficient to be 0.1 m. 2The system records the changes in pollutant concentration over time in each watershed, per second. The simulated pollutant migration results are recorded to form a pollutant migration pattern report. A detailed description of the concentration changes and spatiotemporal distribution of pollutants along the water flow path is provided. For example, at a certain time point, the pollutant concentration is recorded as 50 mg / L in path 1, 30 mg / L in path 2, and 10 mg / L in path 3, forming a concentration change trend graph. Based on the pollutant migration patterns, a multi-region gradient diffusion prediction model is constructed. Mathematical models (such as diffusion equations) or machine learning models are typically used to predict pollutant concentration changes in multiple regions. The model's input parameters are determined, including initial concentration, flow velocity, diffusion coefficient, and environmental factors (such as temperature and humidity). The model predicts pollutant diffusion within the region, recording the pollutant concentration changes in each region at a future time point. For example, the model predicts that within 48 hours, the pollutant concentration in region A will diffuse from 40 mg / L to 25 mg / L, and in region B from 60 mg / L to 35 mg / L. The pollution concentration distribution, migration logic evolution, and diffusion prediction results are integrated to prepare a water pollution diffusion prediction trend map. Ensure data is formatted consistently for easy visualization. Record pollution concentrations and trends for each area for later display. Use GIS software or data visualization tools to construct a water pollution diffusion prediction map. Correlate pollutant concentrations with geographic locations to generate heat maps or distribution maps. For example, the generated map can display pollution concentrations in different areas, and combined with water flow paths, it facilitates understanding of pollutant migration and diffusion trends.

[0043] Step S4: Identify the highest diffusion gradient based on the water pollution diffusion prediction trend map, analyze the optimal navigation monitoring sequence, and construct the optimal navigation monitoring sequence;

[0044] In this embodiment, pollutant concentration data is extracted from the water pollution diffusion prediction map to ensure data coverage of the entire monitoring area. The data is organized into a grid, with each grid representing the pollutant concentration value (in mg / L) at a specific location. For example, assuming the area is divided into a 10m × 10m grid, the pollutant concentration of each grid is recorded to ensure data integrity for subsequent analysis. A suitable diffusion gradient calculation method is selected, typically using numerical differentiation or gradient calculation algorithms, to calculate the concentration change between adjacent grids. The Sobel operator or a simple difference method can be used to identify the location of the maximum concentration change. The diffusion gradient is defined as the concentration difference between adjacent grids, and the calculation formula is set as: gradient = |C1 - C2| / d, where C1 and C2 are the concentration values ​​of adjacent grids, and d is the grid spacing. The diffusion gradient is calculated for each pair of adjacent grids, and the highest gradient value is identified. This process helps to determine the location where pollutant diffusion is most significant. For example, if the concentration in grid A is 50 mg / L and the concentration in grid B is 80 mg / L, the calculated gradient is (80-50) / 10m = 3 mg / L / m. The gradient values ​​of all grids are recorded to identify the highest gradient. The optimal navigation monitoring sequence is defined as the planning of the monitoring path by the unmanned surface vessel (USV) during water quality monitoring, based on the priority of pollutant concentration and diffusion gradient. The goal is to ensure coverage of areas with high pollution concentrations. Monitoring priorities are set, such as prioritizing monitoring areas with the highest diffusion gradient, followed by grids with higher concentrations in adjacent areas. A suitable path planning method is selected, typically a greedy algorithm or shortest path algorithm, to calculate the optimal navigation path of the USV between pollution sources. The path planning must consider not only concentration but also flow velocity and path feasibility. The starting and ending points of the path planning are determined, the initial position of the USV is set, and the path is ensured to effectively connect all monitoring points. Based on the identified highest diffusion gradient and its location, the optimal navigation monitoring sequence of the USV is calculated. For example, the highest gradient position is prioritized as the first monitoring point, and adjacent grids with higher concentrations are monitored sequentially. The planned path is recorded, including the order, location, and corresponding pollutant concentration of each monitoring point, to form the final monitoring sequence. For example, the monitoring point order might be point 1 (x1, y1), point 2 (x2, y2), and point 3 (x3, y3). The previously identified monitoring points and path order are then integrated to form a complete navigation monitoring sequence. The location, concentration, and gradient information of each monitoring point are clearly recorded to facilitate subsequent navigation and data collection. For example, the coordinates, pollutant concentration, diffusion gradient, and estimated monitoring time for each monitoring point are recorded to form a detailed navigation plan. During the unmanned surface vessel's navigation, the navigation monitoring sequence is dynamically adjusted based on real-time monitoring data. For example, if the concentration in a certain area suddenly increases, the navigation path is prioritized to monitor that area. The changes in the monitoring path before and after dynamic adjustments are recorded to analyze their impact on pollution monitoring effectiveness.

[0045] Step S5: Plan the pollution source coverage cruise route based on the optimal navigation monitoring sequence, conduct dynamic navigation cruise monitoring, and collect full-process cruise monitoring data;

[0046] In this embodiment, based on the previously identified optimal navigation monitoring sequence, a suitable path planning algorithm is selected, typically a heuristic algorithm (such as the A* algorithm) or a shortest path algorithm, to achieve coverage of all potential pollution sources and minimize navigation time. The start and end points of the monitoring path are determined, the initial position of the unmanned surface vessel (USV) is set, and the path is ensured to effectively connect all monitoring points. Based on the optimal navigation monitoring sequence, the coverage cruise path of the USV between pollution sources is calculated. Areas with higher pollutant concentrations are prioritized for monitoring, ensuring the path covers all key monitoring points. For example, if the monitoring sequence is point 1 (x1, y1), point 2 (x2, y2), and point 3 (x3, y3), the USV navigates in this order, recording the actual navigation path for each monitoring point. The USV's navigation system is configured to ensure it can dynamically adjust according to the planned coverage cruise path. The system should be able to receive real-time water flow velocity, concentration data, and other environmental information to adjust the navigation strategy. The operating parameters of the navigation system, such as GPS positioning accuracy and navigation algorithm selection, are recorded to ensure accurate navigation of the USV. Based on the coverage cruise path, the USV is controlled to dynamically navigate along the planned path. Real-time monitoring of water quality parameters is crucial to ensure the recording of data such as pollutant concentration, temperature, and pH during the cruise. For example, the unmanned surface vessel (USV) collects water quality data every 5 minutes during its voyage and records changes in relevant parameters for subsequent analysis. During the voyage, the monitoring system should be able to provide real-time feedback on water quality data and dynamically adjust the navigation path based on the monitoring results. For instance, if the pollutant concentration in a certain area increases, the navigation path should be adjusted to prioritize monitoring that area in greater detail. The dynamically adjusted navigation path should be recorded to analyze its impact on pollution monitoring effectiveness. The USV's water quality monitoring equipment should be configured to accurately collect various water quality parameters, including pollutant concentration, temperature, pH, and dissolved oxygen. The calibration and proper functioning of the equipment must be ensured. For example, using multi-parameter water quality monitoring instruments can collect multiple water quality parameters at the same time point, ensuring the timeliness and accuracy of the data. During the cruise, the USV monitors according to a preset monitoring path and records the collected water quality data in real time. The completeness and accuracy of the data must be ensured. For example, water quality data at each monitoring point should be recorded, including timestamps, location coordinates, and pollutant concentrations, for subsequent analysis and use.

[0047] Step S6: Based on the full-process cruise monitoring data, perform real-time cruise concentration distribution change analysis and dynamic path adjustment to construct an intelligent pollution source tracing and positioning model.

[0048] In this embodiment, real-time water quality data is collected from the entire process of cruise monitoring, including timestamps, location information, pollutant concentrations (in mg / L), temperature, and pH values. The integrity and accuracy of the data are ensured for subsequent analysis. For example, water quality data for each monitoring point is recorded, assuming a pollutant concentration of 45 mg / L, a temperature of 22°C, and a pH of 7.5 at a certain monitoring point, and ensuring these data are arranged chronologically. Statistical analysis methods are used to calculate the pollutant concentration changes between monitoring points. By comparing data from different time points, areas with significant concentration changes are identified. For example, by comparing two monitoring data points, if the pollutant concentration in a certain area increases from 30 mg / L to 50 mg / L, the concentration change is calculated to be 20 mg / L, this change is recorded, and the area is marked as a pollution concern. Using GIS software or data visualization tools, a concentration distribution map is generated from the collected water quality data. Interpolation methods (such as Kriging interpolation) are used to fill in the gaps in the areas to obtain a continuous concentration distribution map. For example, the generated concentration distribution map can clearly show the pollutant concentration in each area, helping to identify pollution hotspots and providing a basis for subsequent dynamic path adjustments. Based on the concentration change analysis results, dynamic path adjustment rules for the unmanned surface vessel (USV) are established. For example, if the pollutant concentration in a certain area rises to a certain threshold (e.g., 50 mg / L), the navigation path is preferentially adjusted to that area for more detailed monitoring. Adjustment rules are recorded, such as "If the concentration at a monitoring point exceeds 40 mg / L, immediately adjust the path to that point." During the cruise, a real-time monitoring and feedback mechanism is established to ensure that the USV can dynamically adjust its navigation path according to changes in water quality data. When a concentration change is detected, the system can automatically send a feedback signal, instructing the USV to adjust its path. For example, if the USV detects a pollutant concentration of 60 mg / L at a certain monitoring point, the system will automatically generate a new navigation path, guiding the USV to that area for detailed investigation. Based on the established path adjustment rules and real-time feedback information, the USV is controlled to navigate along the new path. It is ensured that the USV can effectively cover areas with high pollutant concentrations along the adjusted path. For example, the adjusted path may point to a new monitoring point (x2, y2). The adjusted monitoring point and its concentration data are recorded to ensure data continuity and integrity. Choosing a suitable intelligent pollution source tracing and location model typically involves using machine learning algorithms (such as random forests and support vector machines) to analyze pollutant sources. The model's input parameters are then defined, including environmental factors such as pollutant concentration, flow velocity, and wind speed. The required feature data for the model is determined, such as historical concentration data from monitoring points, watershed topography, and water flow direction. A training dataset is constructed based on historical monitoring data and real-time cruise monitoring data. Concentration changes, environmental conditions, and pollution source information from different monitoring points are integrated into a single dataset to facilitate subsequent model training.For example, monitoring data from the past few months is collected to construct a training set containing concentration changes, flow rates, and monitoring point characteristics, ensuring the diversity and representativeness of the dataset. The selected intelligent model is trained using this dataset and cross-validated to ensure its accuracy and generalization ability. Model parameters are adjusted to improve prediction performance. For instance, by training the model, it is possible to predict the possible sources and diffusion paths of pollutants under different environmental conditions, and the model's accuracy and recall are recorded to evaluate its performance. The trained intelligent pollution source tracing model is used to analyze real-time monitoring data to identify the possible locations of pollution sources. Based on the model's predictions, monitoring strategies are adjusted, prioritizing areas with potential pollution sources. For example, if the model predicts that a certain area is a major source of pollutants, the characteristic data of that area is recorded, and a corresponding monitoring plan is developed.

[0049] In this embodiment, see Figure 2 The diagram below illustrates the detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include:

[0050] The system collects regional water quality monitoring parameters based on shipborne water quality monitoring sensors and uploads them to the cloud for data optimization to obtain optimized regional water quality monitoring parameters.

[0051] Water pollution is identified by optimizing regional water quality monitoring parameters, and regional water pollution characteristics are generated.

[0052] Based on the characteristics of regional water pollution, the sources of water pollution are inferred to obtain the regional water pollution sources.

[0053] Pollution concentrations are calculated based on optimized regional water quality monitoring parameters to generate regional pollution concentration values;

[0054] Based on the regional water pollution sources, spatial pollution distribution is mined to construct a regional pollution concentration distribution map.

[0055] In this embodiment, various water quality monitoring sensors are installed on the ship, including pH sensors, turbidity sensors, dissolved oxygen sensors, and conductivity sensors. These sensors should have real-time data acquisition capabilities and be able to operate stably under different water conditions. The installation locations of the sensors are determined to ensure uniform coverage of the water area and to avoid interference from the ship's hull or other obstacles. For example, a location slightly forward of the ship's center is chosen to improve data representativeness. A data acquisition system is configured, with a sampling frequency set to once per minute to obtain dynamic information on water quality changes. The timestamp, sensor type, and corresponding monitoring parameters for each data collection are recorded. For example, at a certain time point, the pH value is recorded as 7.2, turbidity as 5 NTU, dissolved oxygen as 8 mg / L, and conductivity as 500 μS / cm, ensuring data integrity and consistent format.

[0056] Collected water quality monitoring data is uploaded to a cloud database via wireless network to ensure real-time data availability and accessibility. Data transmission is performed using MQTT or HTTP protocols to ensure reliability and security. In the cloud database, data is categorized and stored according to timestamps and sensor types for easy subsequent data optimization and analysis. The uploaded water quality monitoring data undergoes cleaning, including removing missing values, outliers, and duplicate data. Statistical analysis methods, such as the Z-score method, are used to identify outliers to ensure data accuracy. For example, if the dissolved oxygen value at a certain time point is abnormally higher than the normal range (e.g., exceeding 15 mg / L), it is marked as an outlier and removed. Data optimization algorithms (such as weighted average or moving average methods) are used to smooth the cleaned data to eliminate noise and obtain more accurate water quality monitoring parameters. For example, a 10-minute time window is set, and the average dissolved oxygen value within this time window is calculated to optimize the final monitoring parameters. Regional averages for each water quality monitoring parameter are generated based on the optimized data, and corresponding reports are produced. These parameters will be used for subsequent water pollution identification and analysis. For example, the optimized pH value is 7.1, the turbidity is 4 NTU, and the dissolved oxygen is 7.5 mg / L. These values ​​will be used as the baseline for subsequent analyses.

[0057] Based on national or regional water quality standards, set thresholds for water pollution identification. For example, a pH value below 6.5 or above 8.5 may indicate water pollution. Record all relevant threshold standards for subsequent pollution identification and data analysis. Identify pollution in the optimized water quality monitoring parameters to determine if they exceed the set pollution thresholds. Compare each monitoring parameter with its corresponding standard to identify the type of pollution. For example, if turbidity is 10 NTU at a certain time, exceeding the set threshold of 5 NTU, the area is marked as a "polluted area." Based on the identification results, generate corresponding water pollution characteristics, including pollution type, pollution level, and frequency. These characteristics will be used for subsequent pollution source inference. For example, if an area records turbidity exceeding the standard multiple times within a week, the pollution characteristic of that area can be recorded as "frequent high turbidity." Based on the water pollution characteristics, construct a pollution source inference model. Statistical models (such as linear regression) or machine learning models (such as decision trees) can be used to analyze the relationship between pollution sources and water quality parameters. Define the input variables for the model, including historical water quality monitoring parameters, pollution characteristics, and their corresponding time and location. Train the inference model using historical data and perform cross-validation to evaluate its accuracy. Record the model's performance metrics, such as root mean square error (RMSE) or coefficient of determination (R²). 2 For example, if the model's R... 2A value of 0.85 indicates that the model can explain the relationship between water quality parameters and pollution sources relatively well. The trained model is used to infer pollution sources based on the current water pollution situation and identify potential pollution sources. For example, if the model output shows a high correlation between wastewater discharge from an industrial area and pollution parameters, then the industrial area is inferred to be a pollution source. The inference results are recorded, including the name and type of the pollution source and its potential impact range.

[0058] Based on water quality monitoring parameters, select an appropriate method for calculating pollution concentration. Standardized methods can be used to convert monitoring parameters into pollution concentration values. For example, the concentration of a pollutant can be calculated using the formula: Concentration (mg / L) = Monitored value × Dilution factor. Calculate pollution concentrations for the optimized water quality monitoring parameters, recording the pollution concentration values ​​for each time period and region. Ensure that the recorded concentration values ​​are consistent with the actual monitoring data. For example, if the calculated pollution concentration for a certain region is 50 mg / L, this value will be used for subsequent spatial pollution distribution analysis. Based on the calculated regional pollution concentration values, prepare visualization data. Ensure the data format is clear and easy to generate pollution concentration distribution maps. Record the coordinates, pollution concentration values, and related information for each region to facilitate subsequent visualization processing.

[0059] Using Geographic Information System (GIS) software or data visualization tools, construct a regional pollution concentration distribution map. Correlate pollution concentration values ​​with geographic locations to generate heatmaps or distribution maps. For example, use heatmaps to display the pollution concentration distribution across different areas, with color intensity representing concentration levels, facilitating intuitive identification of severely polluted areas. Analyze the generated pollution concentration distribution maps to identify areas of concentrated pollution and generate corresponding reports, detailing the pollution situation and its potential impacts. For instance, if the pollution concentration in a certain area is significantly higher than in other areas, record the characteristics of that area and possible remediation recommendations, providing a basis for subsequent water quality management.

[0060] In this embodiment, the specific steps for collecting regional water quality monitoring parameters based on shipborne water quality monitoring sensors and uploading them to the cloud for data optimization to obtain optimized regional water quality monitoring parameters are as follows:

[0061] The multimodal communication module identifies multiple available networks within the area;

[0062] The communication distance, bandwidth, and anti-interference capability of the multiple available networks are calculated, and a comprehensive network status assessment is performed to obtain the real-time signal status of each available network.

[0063] Real-time optimal network evaluation is performed based on the real-time signal status of each available network, and the real-time optimal available network is marked.

[0064] Determine the dynamic change characteristics of the real-time signal state of each available network;

[0065] Based on the dynamic change characteristics, the real-time optimal available network is dynamically switched to construct a multimodal intelligent switching strategy.

[0066] Water quality monitoring parameters of the area are collected based on shipborne water quality monitoring sensors;

[0067] Based on a multimodal intelligent switching strategy, the regional water quality monitoring parameters are uploaded to the cloud for cloud data optimization to obtain optimized regional water quality monitoring parameters.

[0068] The cloud data optimization specifically refers to:

[0069] Sensor error anomalies in regional water quality monitoring parameters are identified, and sensor error parameters are extracted.

[0070] The sensor error parameters are processed by data removal to obtain optimized error monitoring parameters;

[0071] Spatiotemporal missing detection is performed on error optimization monitoring parameters to identify the locations of spatiotemporal missing points;

[0072] Calculate the average value of the prime number error optimization monitoring parameters;

[0073] Based on the average value, interpolation is performed to fill in the missing locations in time and space to optimize the water quality monitoring parameters for the optimized region.

[0074] In this embodiment, a multimodal communication module is integrated into the device, capable of scanning and identifying available wireless networks. These networks may include Wi-Fi, Bluetooth, cellular networks, etc. The module's operating frequency band and scanning mode are configured to ensure coverage of the required communication range. The scanning frequency is set, for example, scanning once every 5 seconds to obtain the latest network status information. The timestamp of each scan and the identified network information, including network name, signal strength, type, etc., are recorded. The multimodal communication module is activated to perform network scanning. The module will automatically identify available networks in the vicinity and record their basic information, such as SSID, MAC address, signal strength (dBm), access point type, etc. For example, assuming three networks are identified in a single scan: network A (signal strength -45dBm), network B (signal strength -60dBm), and network C (signal strength -75dBm), this information is recorded for subsequent analysis. Based on the signal strength of each network, its theoretical communication distance is calculated. The effective signal propagation distance is estimated using the Free Space Propagation Model (FSPL), converted to dBm. For example, if the signal strength of network A is -45dBm, the effective distance is calculated using the formula to be 300 meters (assuming ideal conditions), and this result is recorded. Collect bandwidth information for each network. If the available network supports multiple bandwidth options (e.g., 20MHz, 40MHz), record the optimal bandwidth configuration. Theoretically, the wider the bandwidth, the faster the data transmission speed. For example, if network B supports 40MHz bandwidth, record its maximum transmission rate (e.g., 300Mbps) for subsequent comprehensive evaluation. Evaluate the anti-interference capability of each network, considering network type and environmental factors. Historical data or actual measurement results can be used to record the network's stability in interference environments. For example, assuming network C can maintain a good connection in complex environments, record its anti-interference capability as "high" and generate an evaluation report. Based on the calculated communication distance, bandwidth, and anti-interference capability, construct a comprehensive scoring model. Assign weights to each parameter to comprehensively evaluate the real-time signal status of each network. For example, set the signal strength weight to 0.5, the bandwidth weight to 0.3, and the anti-interference capability weight to 0.2 for comprehensive evaluation. Based on the evaluation model, real-time signal status is evaluated for each available network, and a comprehensive score is calculated. The score and status of each network are recorded. For example, assuming network A scores 85, network B scores 70, and network C scores 60, the comprehensive score will be used for subsequent optimal network labeling. Based on the comprehensive evaluation results, criteria for selecting the real-time optimal network are set. For example, the network with the highest score is selected as the real-time optimal network. The selection criteria are recorded for future reference during network switching. The comprehensive scores of all available networks are compared, and the network with the highest score is labeled as the real-time optimal network.

[0075] Record detailed information about the network, including signal status and evaluation score. For example, if network A scores the highest, it is marked as the optimal network in real time, and its signal strength and bandwidth are recorded. Monitor the signal status of the optimal network in real time, observing the dynamic changes in its signal strength and other relevant parameters to make timely network switching decisions. For example, if the signal strength of network A drops to -70dBm in a short period of time, switching to the suboptimal network should be considered. Analyze the real-time signal status of each network to identify the trend and characteristics of signal strength changes. Use a moving average method to smooth the signal change data to clearly observe dynamic changes. For example, if the signal strength of network A has gradually decreased over the past 5 minutes, its dynamic change characteristic is recorded as a "decreasing trend." Record the dynamic change characteristics of each network, including the speed and direction of signal change. These characteristics will provide a basis for subsequent network switching strategies. For example, if the signal of network B remains stable while the signal of network C fluctuates significantly, network B can be considered more reliable. Store the data on dynamic change characteristics in a database for easy subsequent analysis and querying. Ensure that the data format is clear and easy to extract and analyze. For example, data tables can be created to record the dynamic characteristics and trends of each network, facilitating the implementation of subsequent intelligent switching strategies.

[0076] Based on the dynamic changes in real-time signal status, a dynamic switching strategy is designed. For example, automatically switching to a suboptimal network when the signal strength falls below a threshold. Switching conditions and priorities are recorded to ensure rapid response to signal changes. Dynamic switching is implemented, and the signal status is monitored in real time. Once a deterioration in the signal status of the currently preferred network is detected, switching is immediately initiated to the marked suboptimal network. For example, if the signal strength of network A continues to decline, switching to network B is initiated immediately, and the switching event is recorded. After switching, the signal status and performance of the newly selected network are continuously monitored to evaluate the effectiveness of the switching strategy. Changes in signal status before and after switching are recorded to optimize the strategy. For example, if the signal strength of network B after switching is recorded as -50dBm, which is better than the performance of network A before switching, the switching is evaluated as successful. Water quality monitoring sensors, including pH, turbidity, and dissolved oxygen, are installed on the vessel to monitor water quality parameters in real time. The sensors are ensured to be properly calibrated and able to operate normally under different water conditions. The sampling frequency is set to once per minute, and the timestamp and monitoring parameters of each collection are recorded. The collected water quality parameters are uploaded to a cloud database via a multimodal communication module.

[0077] Ensure stable data transmission and avoid data loss. Record the uploaded data format, including timestamp, sensor type, and corresponding monitoring parameters for subsequent optimization and analysis. In the cloud, identify anomalies in uploaded water quality monitoring parameters using statistical methods (such as the Z-score method). For example, if the pH value is abnormally higher than the standard range for a certain period, it is marked as abnormal data. Remove identified abnormal data to ensure subsequent data analysis is based on accurate monitoring parameters. Record the details of removed data for subsequent analysis and verification. Perform spatiotemporal missing data detection to identify missing data in certain time periods or locations. Use interpolation to estimate the location of missing values ​​to ensure data integrity. For example, if turbidity data is not recorded for a certain period, record that period as missing and mark it as requiring interpolation. Calculate the average value for the removed monitoring parameters to facilitate subsequent interpolation and optimization. For example, if the processed turbidity data is 4 NTU, 5 NTU, and 6 NTU, calculate the average value as 5 NTU. Record the calculated average value in the database for subsequent interpolation processing. For example, record the average turbidity value for a certain period as 5 NTU, which will serve as the basis for subsequent interpolation and filling. Select an appropriate interpolation method (such as linear interpolation or spline interpolation) and fill in the missing values ​​based on the calculated average. Record the selected method for subsequent data validation. Interpolate the identified missing data locations to fill in the missing values, ensuring the integrity and continuity of the data. For example, if turbidity is missing for a certain period, interpolate using data from before and after the missing data, filling in a missing value of 5 NTU. Integrate the data after removal, filling, and optimization to generate the final optimized regional water quality monitoring parameters, ensuring their accuracy and reliability. Record the optimized parameters and generate a data report for subsequent analysis and application.

[0078] In this embodiment, see Figure 3 The diagram below illustrates the detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include:

[0079] Obtain regional satellite remote sensing maps; perform global histogram equalization optimization on the regional satellite remote sensing maps to obtain brightness-optimized remote sensing maps;

[0080] Water flow direction is identified based on brightness-optimized remote sensing maps to obtain the regional water flow direction;

[0081] Identify the runoff distribution in the brightness-optimized remote sensing map and extract multiple water flow paths;

[0082] Based on the regional water flow direction, the water system orientation characteristics of multiple water flow paths are analyzed one by one to generate the water flow path orientation distribution characteristics.

[0083] Real-time water flow velocity is calculated based on regional water quality monitoring parameters;

[0084] Based on the distribution characteristics of real-time water flow velocity and water flow path, the dynamic evolution of water flow distribution is carried out, and a water flow path distribution network is constructed.

[0085] In this embodiment, a remote sensing map of the target area is acquired using satellite remote sensing technology. High-resolution satellite images can be downloaded from open data sources such as NASA and ESA, or using commercial remote sensing services (such as Google Earth Engine). It is ensured that the acquired images cover the target area and have sufficient spatial resolution (e.g., 10 meters or higher) and temporal resolution. Metadata such as the acquisition time, resolution, and sensor type are recorded during image acquisition for subsequent analysis. For example, the acquired image may be from the Sentinel-2 satellite, with a resolution of 10 meters, and acquired on March 1, 2023. The acquired satellite remote sensing images undergo preprocessing, including geometric correction, radiometric correction, and atmospheric correction, to eliminate systematic errors and environmental influences. The geometric accuracy and radiometric characteristics of the images are ensured to meet analytical requirements. For example, atmospheric correction algorithms (such as the 6S model) are used to eliminate atmospheric effects, ensuring that the spectral characteristics of the image accurately reflect surface features. The grayscale histogram of the regional satellite remote sensing map is analyzed to calculate the image's brightness distribution. By analyzing histograms, the uneven distribution of brightness in an image can be identified, determining the reason for the need for histogram equalization. For example, if the histogram shows that most pixels are concentrated in the low brightness range (e.g., 0-100), it indicates that the image is generally dark and requires equalization processing. Applying a global histogram equalization algorithm to satellite remote sensing maps redistributes the grayscale values ​​of the image to enhance its contrast and brightness. This process is achieved by calculating the cumulative distribution function (CDF), mapping the brightness values ​​of pixels to a new range. For example, in the processed image, the brightness range is evenly distributed from 0-255, enhancing image visibility and making features such as water bodies, vegetation, and soil more prominent. A suitable water flow direction recognition method is selected. Commonly used algorithms include edge detection-based methods (such as Canny edge detection) and motion analysis based on optical flow. A suitable algorithm is selected to extract the water flow direction. Parameters are set, such as high and low thresholds in Canny edge detection, to optimize edge detection performance. The water flow direction recognition algorithm is then implemented on the brightness-optimized remote sensing map to extract the flow characteristics of the water bodies. By analyzing edges and directions in the image, the main direction of water flow is identified. For example, edge detection methods are used to identify the edges of multiple water bodies in the image, and by calculating the edge directions, the flow direction is determined to be southeast. The identified flow directions are recorded and saved to form a flow direction dataset. The specific degrees of the flow direction (e.g., 135 degrees corresponds to southeast) and possible velocity variations are recorded. For example, the main flow direction in the area is recorded as 135 degrees, with a velocity range of 1-3 m / s, for subsequent analysis. A suitable runoff distribution analysis method is selected, typically using watershed analysis-based algorithms (e.g., DEM analysis), utilizing a digital elevation model (DEM) to identify the runoff characteristics of the watershed. Elevation data is prepared, ensuring that the elevation data is geospatially aligned with the remote sensing image.Watershed analysis tools were used to identify runoff distribution within the region and extract multiple flow paths. By analyzing elevation changes and flow direction, the convergence paths of the water flows were determined. For example, three main flow paths were identified within the region: path 1 (northeast), path 2 (southeast), and path 3 (southwest). The extracted flow paths were recorded to form a flow path dataset. The start and end points of each path, as well as its characteristic attributes (such as velocity and flow rate), were recorded. For example, path 1 has a start point coordinate of (x1, y1), an end point coordinate of (x2, y2), a velocity of 2 m / s, and a flow rate of 10 m³ / s. 3 / s. Select a suitable method for analyzing the flow path characteristics, typically using geometric or statistical analysis. Determine the parameters for analysis, such as path length, curvature, and flow direction distribution. Perform flow path characteristic analysis on each extracted flow path, calculate the path's geometric features, and record the flow path distribution characteristics. For example, if path 1 has a length of 500 meters and a curvature of 0.2, record its flow path characteristic as "relatively straight." Record the results of the flow path characteristic analysis to form a flow path flow distribution characteristic report. Summarize the characteristics of each path for subsequent hydrodynamic analysis. For example, record path 1 as "straight," path 2 as "curved," and path 3 as "multi-branched." Select a suitable method for calculating flow velocity, typically using velocity sensor data or calculation methods based on flow path and flow rate. Determine the calculation parameters, such as flow rate and catchment area, to calculate the flow velocity. Calculate the real-time flow velocity based on the monitored flow rate and catchment area. Use the formula: Flow velocity = Flow rate / Cross-sectional area. For example, if the flow rate of a certain water flow path is 10m³ 3 / s, cross-sectional area is 5m² 2The calculated water flow velocity is 2 m / s. Based on the water flow velocity and path characteristics, a suitable dynamic model is selected. Hydrodynamic models or fluid dynamics-based models can be used to analyze the evolution of the water flow. The input parameters of the model are determined, including real-time water flow velocity, path direction, and their characteristics. The water flow distribution is simulated and evolved using the dynamic model, analyzing the changes and impacts of water flow under different conditions. Changes in key parameters during the evolution process are recorded. For example, if the simulation results show that the water flow velocity on path 1 will increase to 3 m / s under specific rainfall conditions, this change is recorded for further analysis. The results of the dynamic evolution of the water flow distribution are recorded to form a water flow dynamic evolution report. The changes in water flow velocity, path, and their relationship with environmental factors are summarized. For example, recording that the water flow velocity on path 1 increases after rainfall, path 2 remains stable, and path 3 is blocked forms a complete dynamic evolution analysis report. Based on the extracted water flow paths and velocities, a water flow path distribution network is constructed. Using graph theory methods, the water flow paths are treated as nodes and edges in the network, establishing the connections between paths. Determine the network connectivity rules, such as the flow relationships and directions between paths. Based on the water flow paths and velocities, establish a water flow path distribution network model. Record the velocity, flow rate, and connectivity of each path. For example, path 1 is connected to path 2, while path 3 is independent, forming a complete water flow path distribution network. Record the constructed water flow path distribution network and generate corresponding visualization charts to intuitively display the water flow path distribution. For example, generate a network diagram to show the connectivity of water flow paths, velocity changes, and their relationship with environmental factors, facilitating subsequent decision support.

[0086] In this embodiment, see Figure 4 The diagram below illustrates the detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include:

[0087] Calculate the pollutant concentration gradient based on the regional pollution concentration distribution map;

[0088] The concentration gradient of pollutants is calculated to generate the concentration change range value;

[0089] Based on the magnitude of concentration changes, in-depth analysis of concentration fluctuation trends is performed to generate regional concentration fluctuation trend characteristics.

[0090] The pollutant migration logic evolution is performed on the water flow path distribution network to generate pollutant migration patterns;

[0091] Based on the migration patterns of pollutants, multi-regional gradient diffusion prediction is performed on the regional concentration fluctuation trend characteristics to construct a water pollution diffusion prediction map.

[0092] In this embodiment, relevant pollutant concentration data are extracted from the regional pollution concentration distribution map to ensure data coverage of the entire study area. The data is organized into a grid, with each grid representing the pollutant concentration value at a specific location. For example, assuming the area is divided into a 10m × 10m grid, the pollutant concentration in each grid is recorded (e.g., in mg / L), ensuring data integrity. A suitable concentration gradient calculation method is selected, typically using numerical differentiation or gradient calculation algorithms (such as the Sobel operator) to facilitate the calculation of concentration changes between each grid. The calculation direction is determined, for example, calculating the concentration gradient in the horizontal and vertical directions to comprehensively analyze concentration changes. The selected algorithm is used to calculate the pollutant concentration gradient for each grid within the region. The concentration gradient value for each grid is recorded, typically expressed in mg / L / m. For example, if the pollutant concentration in a grid is 50 mg / L and the concentration in a neighboring grid is 30 mg / L, the calculated concentration gradient is (50-30) / 10m = 2 mg / L / m, and this result is recorded. The concentration change amplitude is defined as the concentration difference between adjacent grids, reflecting the changes in pollutant concentration within the region. Ensure clear definitions for subsequent calculations. For example, define the variation range as |C1-C2|, where C1 and C2 are the concentration values ​​of adjacent grids, respectively. Calculate the concentration difference for each pair of adjacent grids, generating the concentration variation range value, and record the variation range for each grid. For example, if the concentration in grid A is 40 mg / L and the concentration in grid B is 50 mg / L, then the range is |40-50| = 10 mg / L, and this value is recorded. Choose a suitable fluctuation trend analysis method, typically time series analysis or the sliding window method, to analyze the time series characteristics of the concentration variation range. Determine the analysis window size, for example, set it to 5 time points, and observe the trend of the concentration variation range. Analyze the concentration variation range data using the sliding window method, calculating statistical characteristics such as the mean and standard deviation within each window to identify the trend of concentration fluctuations. For example, if the concentration variation ranges at the 5 time points are 10, 15, 12, 8, and 20 mg / L, then the mean within the window is calculated to be 13 mg / L, and this trend characteristic is recorded. Choosing a suitable logical evolution model for pollutant migration, typically employing diffusion equations or fluid dynamics models, involves analyzing the migration patterns of pollutants along a water flow path. The model's input parameters are determined, including water velocity, concentration gradient, and flow direction. The selected model is then used to simulate the pollutant migration along the water flow path, recording changes in pollutant concentration and its distribution along the path. For example, if a pollutant migrates along the path at a speed of 2 m / s, its concentration changes at different locations are recorded. Based on the pollutant migration patterns, a multi-region gradient diffusion prediction model is constructed. Mathematical models (such as diffusion equations) or machine learning models can be used to predict pollutant concentration changes in multiple regions. The model's input parameters are determined, including initial concentration, flow velocity, and diffusion coefficient.The model predicts the spread of pollutants within a region, recording the concentration changes in each area at a future point in time. For example, if the model predicts that the pollutant concentration in a certain area will spread from 10 mg / L to 5 mg / L within 48 hours, this prediction is recorded. All analyzed pollutant concentration data, migration patterns, and spread prediction results are integrated to generate a water pollution spread prediction map. Data format consistency is ensured for easy visualization. The water pollution spread prediction map is constructed using GIS software or data visualization tools. Pollutant concentrations are correlated with geographical locations to generate heat maps or distribution maps. For example, the generated prediction map displays pollution concentrations in different areas, with color intensity representing concentration levels, facilitating rapid identification of severely polluted areas. The generated water pollution spread prediction map is analyzed to identify areas of concentrated pollution, and corresponding reports are generated, detailing the pollution situation and its potential impacts. For example, if the pollution concentration in a certain area is significantly higher than in other areas, corresponding remediation suggestions are proposed to provide a basis for subsequent water quality management.

[0093] In this embodiment, step S4 includes the following steps:

[0094] The highest diffusion gradient is identified and extracted based on the water pollution diffusion prediction trend map.

[0095] Analyze the flow direction of the highest diffusion gradient to generate the flow direction of the highest pollution gradient;

[0096] Based on the direction of water flow with the highest pollution gradient, the potential pollution source directions are located in the water flow path distribution network to obtain the locations of multiple potential pollution sources.

[0097] Identify the current spatial location of the unmanned vessel and calculate the nearest potential pollution source based on the locations of multiple potential pollution sources;

[0098] Calculate the distance difference between the locations of the multiple potential pollution sources;

[0099] Based on the nearest potential pollution source and the distance difference between the pollution sources, an optimal navigation monitoring sequence is constructed through analysis of the optimal navigation monitoring order.

[0100] In this embodiment, pollutant concentration data is extracted from the water pollution diffusion prediction trend map. The data is organized into a grid to ensure that the pollution concentration value of each grid is clear and readable, facilitating subsequent analysis. For example, the area is divided into 10m × 10m grids, and the pollutant concentration (e.g., mg / L) of each grid is recorded. Data integrity is ensured for subsequent gradient calculation. A suitable diffusion gradient calculation method is selected, typically using the concentration gradient calculation formula to calculate the concentration change between adjacent grids. Numerical differentiation methods can be used to identify the location of the greatest change. The diffusion gradient is defined as the concentration difference between adjacent grids, and the calculation formula is set as: gradient = |C1 - C2| / d, where C1 and C2 are the concentration values ​​of adjacent grids, and d is the grid spacing. The diffusion gradient is calculated for each pair of adjacent grids, and the highest gradient value is identified. This process will help determine the location where pollutant diffusion is most significant. For example, if the concentration of grid A is 40 mg / L and the concentration of grid B is 70 mg / L, then the calculated diffusion gradient is (70-40) / 10m = 3 mg / L / m. Record the gradient values ​​of all grids and identify the highest progressive gradient. Prepare flow direction data using the previous flow direction identification results. Ensure the flow direction data matches the concentration data for analysis. Record the flow direction for each grid, ensuring the unit is degrees (0° represents north, clockwise). Select a suitable flow direction analysis method, typically a flow direction map-based method, combined with the concentration gradient identification results, to analyze the flow direction of the highest pollution gradient. Define the flow direction characteristic as "flowing from high concentration to low concentration". Based on the identified highest diffusion progressive gradient, analyze the flow direction of the grid corresponding to that gradient. Record this direction for subsequent potential pollution source location. For example, if the flow direction of the highest diffusion progressive gradient is 135° (southeast), record this flow direction characteristic. Prepare potential pollution source data based on the flow path distribution network. Ensure the coordinates and corresponding concentration values ​​of each potential pollution source are known. Record the geographical location of the potential pollution source and its corresponding pollutant concentration for subsequent analysis. Select a suitable direction location method, typically vector analysis, combining the flow direction and the location of the potential pollution source to determine the orientation of the pollution source. The location of a pollution source is defined as the direction from the current location of the unmanned vessel (UAV) towards the potential pollution source. Based on the direction of the water flow with the highest pollution gradient, the location of each potential pollution source is analyzed, and its azimuth relative to the current spatial location of the UAV is calculated. For example, if the UAV's location is (x0, y0) and the potential pollution source's location is (x1, y1), then the azimuth angle is calculated as atan2(y1-y0, x1-x0), and the azimuths of all potential pollution sources are recorded. The current spatial location of the UAV is obtained using the Global Positioning System (GPS), ensuring the accuracy and reliability of the acquired location information. The current longitude and latitude of the UAV (e.g., longitude X, latitude Y) are recorded, ensuring their format is clear for subsequent calculations.Choose a suitable distance calculation method, typically the Euclidean distance method, to calculate the distance between the unmanned vessel and each potential pollution source. The distance calculation formula is defined as: Distance = √((x1-x0)). 2 +(y1-y0) 2 Let (x0, y0) represent the location of the unmanned surface vessel (USV), and (x1, y1) represent the location of the potential pollution source. Distances to all potential pollution sources are calculated, and the distance between the USV and each potential pollution source is recorded. For example, if the USV's location is (5, 5) and the potential pollution source's location is (10, 10), the calculated distance is √((10-5)² + (10-5)). 2 = 7.07, record this distance. Based on the calculated distance value, identify the nearest potential pollution source and record its location information and distance value. For example, if the distance to potential pollution source A is 6.5 and the distance to potential pollution source B is 7.0, then the nearest pollution source is A. Define the pollution source distance difference as the distance difference between the nearest potential pollution source and other potential pollution sources, reflecting the relative positional relationship of the pollution sources. For example, if the distance to the nearest pollution source A is 6.5 and the distance to another pollution source B is 8.0, then the distance difference is 8.0 - 6.5 = 1.5. Calculate the distance difference for all potential pollution sources and record the distance difference between each potential pollution source and the nearest pollution source. For example, if the distance to potential pollution source C is 9.0, then the calculated distance difference between C and A is 9.0 - 6.5 = 2.5. Record the calculated pollution source distance differences in the dataset to form a pollution source information table containing distance differences. For example, record "Distance to pollution source A: 6.5, Distance to pollution source B: 8.0, Distance to pollution source C: 9.0" and calculate the corresponding distance differences. Select a suitable navigation sequence analysis method, typically a greedy algorithm or shortest path algorithm, to calculate the optimal navigation path of the unmanned surface vessel (USV) among potential pollution sources, determining the analysis objective, such as minimizing navigation distance or time. Based on the distance and orientation of potential pollution sources, calculate the optimal navigation sequence of the USV among different potential pollution sources. For example, prioritize visiting the nearest pollution source, followed by the second closest. Record the optimal navigation sequence to form a navigation path dataset. Record the optimal navigation monitoring sequence and generate corresponding visualization charts to intuitively display the USV's navigation path. For example, generate a navigation path map showing the optimized path of the USV from its current position to each potential pollution source, facilitating subsequent navigation control.

[0101] In this embodiment, step S5 includes the following steps:

[0102] Based on the water flow path distribution network, the navigation direction of the unmanned vessel is determined by the upstream and downstream direction, and the upstream and downstream direction of each path is obtained.

[0103] The navigation speed of each path is calculated based on the real-time water flow velocity and the upstream and downstream directions of each path.

[0104] Based on the navigation speed of each path, the optimal navigation monitoring sequence is used to plan the pollution source coverage cruise path and construct an intelligent coverage cruise path.

[0105] The unmanned surface vessel for water quality monitoring conducts dynamic navigation and patrol monitoring based on the intelligent coverage patrol path, and collects patrol monitoring data throughout the entire process.

[0106] In this embodiment, water flow direction data is extracted from the water flow path distribution network. The water flow direction of each path is ensured to be represented in degrees and matched with the unmanned surface vessel's (USV) navigation system. The directional characteristics of each water flow path are recorded; for example, one path may have a water flow direction of 135° (southeast) and another path 315° (northwest). A suitable method for determining whether the current is downstream or upstream is selected, typically based on a comparison between the water flow direction and the USV's navigation direction. If the water flow direction matches the navigation direction, it is downstream; otherwise, it is upstream. The target direction for the USV's navigation is set, for example, 180° (south). The direction of each water flow path is compared with the USV's navigation direction to determine whether it is downstream or upstream. For example, if the water flow direction is 135° and the navigation direction is 180°, the angle between them is calculated, and it is determined to be "upstream". The downstream / upstream status of each path is recorded to form a data table for subsequent analysis. For example, path A is "downstream" and path B is "upstream". Ensure real-time water flow velocity data is available and matched with the water flow path distribution network. Water flow velocity should be recorded in meters per second (m / s). For example, record the water flow velocity for a certain area as 2 m / s, and ensure this data applies to all paths. Choose a suitable method for calculating the navigation speed, typically using a combination of water flow velocity and navigation direction. If downstream, the navigation speed is the water flow velocity plus the UAV's own speed; if upstream, it is the water flow velocity minus the UAV's own speed. Set the UAV's own navigation speed, for example, to 1 m / s. For each path, calculate the navigation speed based on its downstream / upstream state. If downstream, the speed is water flow velocity + boat speed; if upstream, the speed is water flow velocity - boat speed. For example, if path A is downstream with a water flow velocity of 2 m / s and the UAV's speed is 1 m / s, the navigation speed is 2 + 1 = 3 m / s; if path B is upstream, the navigation speed is 2 - 1 = 1 m / s. Record the sailing speed for each path to create a data table for subsequent cruise path planning. Select a suitable path planning method, typically using heuristic algorithms (such as A* algorithm or genetic algorithm) to optimize pollution source coverage paths. The goal is to ensure the unmanned surface vessel (USV) can cover all potential pollution sources. Determine the start, end, and waypoints of the coverage path to ensure the effectiveness of the path planning. Calculate the optimal coverage cruise path based on the sailing speed of each path and the location of potential pollution sources. For example, prioritize paths with faster sailing speeds that cover multiple pollution sources. Record the planned paths, including the order of each path and its corresponding sailing speed, to form the final coverage cruise path dataset. Configure the USV's navigation system to ensure it can dynamically adjust according to the planned coverage cruise path. The system should be able to receive water flow speed and direction information in real time to adjust the navigation strategy. Record the navigation system's operating parameters, such as GPS positioning accuracy and navigation algorithm selection, to ensure accurate navigation of the USV. Based on the intelligent coverage cruise path, control the USV to dynamically navigate along the planned path.Real-time monitoring of water quality parameters is crucial to ensure the recording of data such as pollutant concentration, temperature, and pH levels during the cruise. For example, the unmanned surface vessel (USV) collects water quality data every 5 minutes during its voyage and records changes in relevant parameters. Appropriate data analysis methods are selected, typically statistical analysis and visualization tools, to analyze the water quality monitoring data collected during the cruise. Key indicators for analysis are identified, such as trends in pollutant concentration changes and regional water quality safety. The collected water quality monitoring data is then analyzed to identify trends in pollutant concentration changes and assess the regional water quality status. For example, time series analysis can be used to identify whether there are significant upward or downward trends in pollutant concentrations, and the results are recorded.

[0107] In this embodiment, step S6 includes the following steps:

[0108] Real-time concentration distribution change analysis was performed on the full-process cruise monitoring data to construct a real-time concentration distribution change curve.

[0109] Concentration abrupt changes are detected based on the real-time concentration distribution change curve, and the location of the concentration abrupt change point is located.

[0110] Secondary pollution source tracing and location are performed based on the location of concentration abrupt changes, thereby obtaining accurate pollution source location coordinates;

[0111] Based on the precise pollution source location coordinates, the intelligent coverage cruise path is dynamically adjusted to obtain the real-time pollution source location adjustment path.

[0112] A smart pollution source tracing and location model is constructed by deep iterative learning and optimization based on real-time pollution source location and adjustment path.

[0113] In this embodiment, water quality monitoring parameters are extracted from the full-process cruise monitoring data, ensuring a unified data format that includes timestamps, location coordinates, and pollutant concentrations (e.g., mg / L). The data is organized by time series for subsequent analysis. For example, water quality monitoring data from the unmanned surface vessel at different time points is recorded to form a dataset, including: time (t1, t2, t3...), location coordinates (x, y), and corresponding pollutant concentrations (C1, C2, C3...). A suitable concentration distribution change analysis method is selected, typically using interpolation (e.g., Kriging interpolation) to process the monitoring data, generating a concentration distribution map and plotting concentration change curves. Interpolation parameters are set, such as selecting an appropriate neighborhood range to ensure accuracy; typically, the five nearest monitoring points are selected for interpolation calculations. Based on the organized data and interpolation results, an instantaneous concentration distribution change curve is constructed. The location coordinates and corresponding concentration values ​​are visualized, and a graph showing the concentration changing over time is plotted. For example, if the concentration values ​​are 10 mg / L, 15 mg / L, and 20 mg / L successively within a specific time period, the generated curve should show a gradual increase in concentration, and the key characteristics of the curve should be recorded. Define the criteria for concentration mutations, typically based on statistical methods to set thresholds. For example, a concentration change exceeding a set value (e.g., 20%) is considered a mutation, or standard deviation can be used to identify mutation points. Record the established mutation detection criteria for subsequent analysis and validation. Analyze the real-time concentration distribution change curve to detect concentration mutation points. Identify mutation locations exceeding the threshold by calculating the magnitude of changes in adjacent concentration values. For example, if at a certain time point, the concentration suddenly rises from 10 mg / L to 30 mg / L, a change of 200%, record this location as a concentration mutation point and mark its specific time and coordinates. Record the detected concentration mutation locations to form a mutation point dataset, including the mutation time, location coordinates, and magnitude of concentration change. For example, record the mutation point location as (x1, y1), the time as t1, and the concentration change magnitude as 200%, for subsequent source tracing analysis. Choosing a suitable pollution source tracing analysis method typically involves using fluid dynamics models or pollutant migration models, combined with the location of abrupt change points, to pinpoint the pollution source. The model's input parameters are determined, including water flow velocity, concentration characteristics, and the spatial distribution of the abrupt change. Based on the concentration abrupt change location, the selected tracing analysis model is used to calculate the pollution source location. The potential pollution source location is inferred based on the concentration at the abrupt change point and its trend. For example, if the abrupt change point is (x1, y1) and the concentration change direction points to a specific area, the pollution source is likely located within that area. The tracing analysis results are recorded to form precise pollution source location coordinate data. The recorded data must include the spatial coordinates of the pollution source and its corresponding concentration characteristics. For example, recording the pollution source location as (x2, y2) with a concentration of 35 mg / L serves as the basis for subsequent path adjustments. Based on the precise pollution source location coordinates, dynamic adjustment standards for the intelligent coverage cruise path are set.For example, if the pollution source is within the unmanned surface vessel's (USV) navigation path, the navigation path needs to be adjusted to ensure coverage of the pollution source. The established adjustment criteria are recorded for subsequent dynamic adjustments. Based on the precise pollution source location coordinates, the intelligent coverage cruise path is dynamically adjusted. The navigation path is recalculated using a path planning algorithm to ensure the USV can effectively cover the newly located pollution source. For example, if the original path is A to B, the adjusted path is A to (x2, y2) and then to B, ensuring that the pollution source location is passed during the process. The adjusted intelligent coverage cruise path is recorded, and corresponding visualization charts are generated to intuitively show the path changes. For example, a map containing the pollution source and the adjusted path is generated to help the USV clearly understand the adjustment target during the cruise. A suitable deep learning model is selected, typically a neural network or reinforcement learning model, to construct an intelligent pollution source tracing and location model. This model will be trained using historical data to improve positioning accuracy. The input features of the model are determined, including historical monitoring data, pollution source location, and its concentration change characteristics. The selected deep learning model is trained using the collected cruise monitoring data to optimize parameters and improve the model's predictive ability. The sufficiency and diversity of the training data are ensured. For example, the model is trained using past monitoring data, and its performance metrics (such as accuracy and loss value) are recorded to ensure model validation. During the patrol, the trained intelligent pollution source tracing and location model is applied in real time to monitor new data and make real-time predictions, dynamically adjusting the route. The model's predictions are compared with actual observations, and the model's performance is analyzed and iteratively optimized to improve future prediction accuracy.

[0114] In this embodiment, a navigation and control system for an autonomous water quality monitoring unmanned surface vessel (USV) based on multimode communication is provided, for executing the navigation and control method for an autonomous water quality monitoring USV based on multimode communication as described above, including:

[0115] The spatial pollution distribution module is used to collect regional water quality monitoring parameters based on shipborne water quality monitoring sensors, and to mine the spatial pollution distribution to construct a regional pollution concentration distribution map.

[0116] The water flow distribution evolution module is used to acquire regional satellite remote sensing maps, perform dynamic evolution of water flow distribution, and construct a water flow path distribution network;

[0117] The diffusion prediction module is used to perform pollutant migration logic evolution and multi-regional gradient diffusion prediction on the water flow path distribution network based on the regional pollution concentration distribution map, and to construct a water pollution diffusion prediction trend map.

[0118] The navigation sequence module is used to identify the highest diffusion gradient based on the water pollution diffusion prediction trend map, analyze the optimal navigation monitoring sequence, and construct the optimal navigation monitoring sequence.

[0119] The path planning module is used to plan the pollution source coverage cruise path based on the optimal navigation monitoring sequence, perform dynamic navigation cruise monitoring, and collect full-process cruise monitoring data.

[0120] The path adjustment module is used to perform real-time analysis of changes in concentration distribution during the entire process of cruise monitoring and to dynamically adjust the path, thereby building an intelligent pollution source tracing and positioning model.

[0121] This invention collects various water quality data in real time using shipborne sensors, enabling rapid location of pollution hotspots and ensuring timely detection of pollution problems in waterways. The construction of spatial pollution distribution maps clearly displays pollution concentrations in different areas, providing crucial data support for subsequent water flow and pollution diffusion analysis. Data mining techniques effectively extract pollution information from large amounts of data, achieving efficient pollution monitoring and control. Through satellite remote sensing maps and water flow evolution models, unmanned vessels can obtain dynamic evolution information on water flow, including velocity and direction, providing a foundation for path planning and pollution diffusion prediction. The establishment of a water flow path distribution network helps understand the propagation path of water in water bodies, making subsequent path planning more accurate and avoiding monitoring errors caused by changes in water flow direction. Pollutant migration and evolution prediction allows for the early identification of pollution diffusion paths and speeds, helping to formulate timely countermeasures. Multi-regional gradient diffusion prediction assesses the pollution diffusion situation in different areas, helping to determine the potential impact range of pollution and providing accurate data for subsequent monitoring and remediation. Based on the diffusion prediction trend map, early warning can be achieved, preventing pollution spread and reducing the difficulty and cost of pollution remediation. By analyzing the optimal navigation monitoring sequence, unmanned surface vessels (USVs) can prioritize monitoring in areas with the most severe pollution spread, improving the overall efficiency of water quality monitoring. Optimizing the navigation sequence avoids duplicate monitoring or missed detection of polluted areas, improving the coverage and accuracy of water quality monitoring. Multiple USVs can work collaboratively according to the optimized navigation monitoring sequence, avoiding conflicts and improving the efficiency of the entire monitoring system. Through the path planning module, USVs can accurately cover pollution sources, ensuring comprehensive monitoring and data collection. Path planning allows USVs to avoid unnecessary path duplication and time waste, thereby improving monitoring efficiency and saving energy. Dynamic navigation allows USVs to flexibly adjust their navigation paths according to real-time conditions (such as changes in water flow and sudden pollution), ensuring efficient completion of monitoring tasks. By analyzing cruise concentration changes in real time, the system can quickly respond to sudden changes in water quality and adjust the path in a timely manner, ensuring effective tracking and location of pollution sources. Dynamic path adjustment can optimize the monitoring path of USVs in real time according to changes in pollution concentration, thereby achieving accurate pollution source tracing and location. Through path adjustment, USVs can adjust their cruise paths according to changes in the environment and pollutants, achieving continuous and efficient monitoring and avoiding limitations in path design.

[0122] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0123] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement it. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A navigation and control method for an autonomous unmanned surface vessel for water quality monitoring based on multimode communication, characterized in that, The unmanned surface vessel (USV) is equipped with onboard water quality monitoring sensors and a multimodal communication module; the process includes the following steps: Step S1: Collect regional water quality monitoring parameters based on shipborne water quality monitoring sensors, and conduct spatial pollution distribution mining to construct a regional pollution concentration distribution map; Step S2: Obtain regional satellite remote sensing maps, perform dynamic evolution of water flow distribution, and construct a water flow path distribution network; Step S3: Based on the regional pollution concentration distribution map, perform pollutant migration logic evolution and multi-regional gradient diffusion prediction on the water flow path distribution network to construct a water pollution diffusion prediction trend map; Step S4: Identify the highest diffusion gradient based on the water pollution diffusion prediction trend map, analyze the optimal navigation monitoring sequence, and construct the optimal navigation monitoring sequence; Step S5: Plan the pollution source coverage cruise route based on the optimal navigation monitoring sequence, conduct dynamic navigation cruise monitoring, and collect full-process cruise monitoring data; Step S6: Based on the full-process cruise monitoring data, perform real-time cruise concentration distribution change analysis and dynamic path adjustment to construct an intelligent pollution source tracing and positioning model; The specific steps of step S2 are as follows: Obtain regional satellite remote sensing maps; perform global histogram equalization optimization on the regional satellite remote sensing maps to obtain brightness-optimized remote sensing maps; Water flow direction is identified based on brightness-optimized remote sensing maps to obtain the regional water flow direction; Identify the runoff distribution in the brightness-optimized remote sensing map and extract multiple water flow paths; Based on the regional water flow direction, the water system orientation characteristics of multiple water flow paths are analyzed one by one to generate the water flow path orientation distribution characteristics. Real-time water flow velocity is calculated based on regional water quality monitoring parameters; Based on the distribution characteristics of real-time water flow velocity and water flow path, the dynamic evolution of water flow distribution is carried out, and a water flow path distribution network is constructed.

2. The navigation and control method for autonomous water quality monitoring unmanned surface vessel based on multi-mode communication according to claim 1, characterized in that, The specific steps of step S1 are as follows: The system collects regional water quality monitoring parameters using shipborne water quality monitoring sensors and uploads them to the cloud for data optimization to obtain optimized regional water quality monitoring parameters. Water pollution is identified by optimizing regional water quality monitoring parameters, and regional water pollution characteristics are generated. Based on the characteristics of regional water pollution, the sources of water pollution are inferred to obtain the regional water pollution sources. Pollution concentrations are calculated based on optimized regional water quality monitoring parameters to generate regional pollution concentration values; Based on the regional water pollution sources, spatial pollution distribution is mined to construct a regional pollution concentration distribution map.

3. The navigation and control method for autonomous water quality monitoring unmanned surface vessel based on multi-mode communication according to claim 2, characterized in that, The specific steps for collecting regional water quality monitoring parameters based on shipborne water quality monitoring sensors and uploading them to the cloud for data optimization to obtain optimized regional water quality monitoring parameters are as follows: The multimodal communication module identifies multiple available networks within the area; The communication distance, bandwidth, and anti-interference capability of the multiple available networks are calculated, and a comprehensive network status assessment is performed to obtain the real-time signal status of each available network. Real-time optimal network evaluation is performed based on the real-time signal status of each available network, and the real-time optimal available network is marked. Determine the dynamic change characteristics of the real-time signal state of each available network; Based on the dynamic change characteristics, the real-time optimal available network is dynamically switched to construct a multimodal intelligent switching strategy. Water quality monitoring parameters of the area are collected based on shipborne water quality monitoring sensors; Based on a multimodal intelligent switching strategy, the regional water quality monitoring parameters are uploaded to the cloud for cloud data optimization to obtain optimized regional water quality monitoring parameters. The cloud data optimization specifically refers to: Sensor error anomalies in regional water quality monitoring parameters are identified, and sensor error parameters are extracted. The sensor error parameters are processed by data removal to obtain optimized error monitoring parameters; Spatiotemporal missing detection is performed on error optimization monitoring parameters to identify the locations of spatiotemporal missing points; Calculate the average value of the error optimization monitoring parameters; Based on the average value, interpolation is performed to fill in the missing locations in time and space to optimize the water quality monitoring parameters for the optimized region.

4. The navigation and control method for autonomous water quality monitoring unmanned surface vessel based on multi-mode communication according to claim 1, characterized in that, Step S3 is as follows: Calculate the pollutant concentration gradient based on the regional pollution concentration distribution map; The concentration gradient of pollutants is calculated to generate the concentration change range value; Based on the magnitude of concentration changes, in-depth analysis of concentration fluctuation trends is performed to generate regional concentration fluctuation trend characteristics. The pollutant migration logic evolution is performed on the water flow path distribution network to generate pollutant migration patterns; Based on the migration patterns of pollutants, multi-regional gradient diffusion prediction is performed to analyze the regional concentration fluctuation trends, and a water pollution diffusion prediction map is constructed.

5. The navigation and control method for autonomous water quality monitoring unmanned surface vessel based on multi-mode communication according to claim 1, characterized in that, The specific steps of step S4 are as follows: The highest diffusion gradient is identified and extracted based on the water pollution diffusion prediction trend map. Analyze the flow direction of the highest diffusion gradient to generate the flow direction of the highest pollution gradient; Based on the direction of water flow with the highest pollution gradient, the potential pollution source directions are located in the water flow path distribution network to obtain the locations of multiple potential pollution sources. Identify the current spatial location of the unmanned vessel and calculate the nearest potential pollution source based on the locations of multiple potential pollution sources; Calculate the distance difference between the locations of the multiple potential pollution sources; Based on the nearest potential pollution source and the distance difference between the pollution sources, an optimal navigation monitoring sequence is constructed through analysis of the optimal navigation monitoring order.

6. The navigation and control method for autonomous water quality monitoring unmanned surface vessel based on multi-mode communication according to claim 1, characterized in that, The specific steps of step S5 are as follows: Based on the water flow path distribution network, the navigation direction of the unmanned vessel is determined by the upstream and downstream direction, and the upstream and downstream direction of each path is obtained. The navigation speed of each path is calculated based on the real-time water flow velocity and the upstream and downstream directions of each path. Based on the navigation speed of each path, the optimal navigation monitoring sequence is used to plan the pollution source coverage cruise path and construct an intelligent coverage cruise path. The unmanned surface vessel for water quality monitoring conducts dynamic navigation and patrol monitoring based on the intelligent coverage patrol path, and collects patrol monitoring data throughout the entire process.

7. The navigation and control method for autonomous water quality monitoring unmanned surface vessel based on multi-mode communication according to claim 1, characterized in that, The specific steps of step S6 are as follows: Real-time concentration distribution change analysis was performed on the full-process cruise monitoring data to construct a real-time concentration distribution change curve. Concentration abrupt changes are detected based on the real-time concentration distribution change curve, and the location of the concentration abrupt change point is located. Secondary pollution source tracing and location are performed based on the location of concentration abrupt changes, thereby obtaining accurate pollution source location coordinates; Based on the precise pollution source location coordinates, the intelligent coverage cruise path is dynamically adjusted to obtain the real-time pollution source location adjustment path. A smart pollution source tracing and location model is constructed by deep iterative learning and optimization based on real-time pollution source location and adjustment path.

8. A navigation and control system for an autonomous unmanned surface vessel for water quality monitoring based on multimode communication, characterized in that, The method for executing the autonomous water quality monitoring unmanned vessel navigation and control method based on multi-mode communication as described in claim 1 includes: The spatial pollution distribution module is used to collect regional water quality monitoring parameters based on shipborne water quality monitoring sensors, and to mine the spatial pollution distribution to construct a regional pollution concentration distribution map. The water flow distribution evolution module is used to acquire regional satellite remote sensing maps, perform dynamic evolution of water flow distribution, and construct a water flow path distribution network; The diffusion prediction module is used to perform pollutant migration logic evolution and multi-regional gradient diffusion prediction on the water flow path distribution network based on the regional pollution concentration distribution map, and to construct a water pollution diffusion prediction trend map. The navigation sequence module is used to identify the highest diffusion gradient based on the water pollution diffusion prediction trend map, analyze the optimal navigation monitoring sequence, and construct the optimal navigation monitoring sequence. The path planning module is used to plan the pollution source coverage cruise path based on the optimal navigation monitoring sequence, perform dynamic navigation cruise monitoring, and collect full-process cruise monitoring data. The path adjustment module is used to perform real-time analysis of changes in concentration distribution during the entire process of cruise monitoring and to dynamically adjust the path, thereby building an intelligent pollution source tracing and positioning model.

Images