Marine pollution traceability big data analysis method based on space-time diagram neural network

Abstract

The invention discloses a marine pollution traceability big data analysis method based on a space-time diagram neural network, which is applied to the field of marine monitoring, and comprises the following steps: extracting seawater pollution diffusion characteristics according to periodic water quality data of a target sea area and a marine flow field spectrum, and processing the seawater pollution diffusion characteristics based on the diagram neural network, and reconstructing the pollution diffusion network through ocean surface remote sensing data, and adjusting nodes and edges of the pollution diffusion network according to a bidirectional propagation analysis result to obtain a target pollution diffusion network. According to the method, the diffusion path of pollutants can be simulated more accurately, so that the positioning precision of the pollution source is improved, the method better adapts to the actual marine pollution monitoring condition, and the reliability of the traceability result is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of marine monitoring, and particularly to a big data analysis method for marine pollution source tracing based on a spatio-temporal graph neural network. Background Art

[0002] The existing methods for marine pollution source tracing mainly rely on manual on-site investigations and the analysis of single water quality monitoring data. These methods are time-consuming and laborious and difficult to cover vast sea areas, resulting in limitations in data space. Therefore, it is impossible to trace the source of water quality pollution, and only the marine water quality at the monitoring sites can be monitored and managed. When water quality pollution is detected, purification treatment is carried out, which can only solve the pollution problems in local areas, making it difficult to detect and effectively control pollution sources in a timely manner. The pollution sources cannot be eliminated fundamentally, and pollution may occur again, making it difficult to fundamentally solve the problem of marine pollution.

[0003] Therefore, it has become an urgent technical problem for those skilled in the art to optimize the method for tracing the source of marine water quality pollution in order to accurately trace the source of marine pollution and thus restore the marine ecological environment. Summary of the Invention

[0004] The present invention provides a big data analysis method for marine pollution source tracing based on a spatio-temporal graph neural network to accurately trace the source of marine pollution.

[0005] To solve the above technical problems, an embodiment of the present invention provides a big data analysis method for marine pollution source tracing based on a spatio-temporal graph neural network, including: In actual marine pollution source tracing analysis, the seawater data collected in real time is input into a target pollution diffusion network, and a marine pollution source tracing report of the target sea area is generated according to the output result of the target pollution diffusion network. Among them, the construction process of the target pollution diffusion network includes: Perform water quality fingerprint analysis on the water body collection data of the target sea area obtained to obtain the periodic water quality data of the target sea area.

[0006] Determine the marine flow field map of the target sea area, where the marine flow field map is obtained by analyzing the first marine dynamics data of the target sea area.

[0007] Extract the seawater pollution diffusion characteristics according to the periodic water quality data and the marine flow field map, and process the seawater pollution diffusion characteristics based on a graph neural network to obtain a first pollution diffusion network of pollutants in the target sea area.

[0008] Perform identification and analysis on the obtained marine surface remote sensing data, and reconstruct the first pollution diffusion network according to the analysis result to obtain a second pollution diffusion network.

[0009] Input the periodic water quality data into the second pollution diffusion network for two-way propagation analysis to obtain a forward propagation path and a reverse propagation path. Adjust the nodes of the second pollution diffusion network according to the segment offset between the forward propagation path and the reverse propagation path to obtain a target pollution diffusion network.

[0010] Further, performing water quality fingerprint analysis on the collected water body data of the target sea area to obtain the periodic water quality data of the target sea area, including: Determine a preset collection interval according to the monitoring requirements of the target sea area and the sea current speed.

[0011] Collect seawater samples at each monitoring point in the target sea area at the preset collection interval.

[0012] Perform analysis on key water quality indicators for each of the seawater samples to obtain corresponding water quality fingerprint characteristic data.

[0013] Generate random Poisson noise according to the historical pollution statistics data of the target sea area, and use a noise replacement value to replace the value of the noise seed of the random Poisson noise, where the noise replacement value is determined by the typical pollution data of the target sea area. λ value for replacement, where the noise replacement value is determined by the typical pollution data of the target sea area.

[0014] Add the random Poisson noise to the water quality fingerprint characteristic data to obtain the periodic water quality data of the target sea area.

[0015] Further, performing analysis on key water quality indicators for each of the seawater samples to obtain corresponding water quality fingerprint characteristic data, including: Perform analysis on key water quality indicators for each of the seawater samples to obtain the characteristic index of each key water quality indicator of the seawater sample.

[0016] Perform time-domain analysis on each of the characteristic indexes to obtain the change trend curve of the corresponding key water quality indicator in the time series, and obtain the time-domain characteristic data of the seawater sample by non-linearly fitting each of the change trend curves.

[0017] Obtain the frequency spectrum data of each of the characteristic indexes through Fourier transform, calculate the power spectral density of the corresponding characteristic index according to each of the frequency spectrum data to obtain the frequency component data of the characteristic index, and obtain the frequency-domain characteristic data of the seawater sample according to each of the frequency component data.

[0018] Generate the water quality fingerprint characteristic data of the seawater sample based on the time-domain characteristic data and the frequency-domain characteristic data by using a feature fusion algorithm.

[0019] Further, extracting the seawater pollution diffusion characteristics based on the periodic water quality data and the ocean current field atlas includes: Performing point-by-point analysis on the periodic water quality data to obtain statistical characteristic data of each of the key water quality indicators.

[0020] Setting a normal threshold range for each of the key water quality indicators according to the historical pollution statistical data and the environmental standards of the target sea area.

[0021] Comparatively analyzing the statistical characteristic data and the normal threshold range. If there is a statistical characteristic data that exceeds the corresponding normal threshold range, mark the corresponding point as a potential pollution point, and extract the seawater pollution diffusion characteristics of the potential pollution point.

[0022] Further, determining the ocean current field atlas of the target sea area includes: Obtaining the first ocean dynamic data of the target sea area, and performing time alignment on the first ocean dynamic data to obtain the second ocean dynamic data.

[0023] Performing feature extraction on the second ocean dynamic data to obtain ocean movement characteristics.

[0024] Constructing ocean movement grid data of the target sea area according to the ocean movement characteristics.

[0025] Constructing the ocean current field atlas of the target sea area according to the ocean movement grid data and the geographic information system data of the target sea area.

[0026] Further, the ocean surface remote sensing data at least includes sea surface temperature data, chlorophyll concentration data, suspended matter concentration data, and sea surface oil film distribution data.

[0027] The identifying and analyzing the obtained ocean surface remote sensing data includes: Obtaining the ocean surface remote sensing data of the target sea area monitored by the satellite remote sensing system, and preprocessing the ocean surface remote sensing data.

[0028] Performing image recognition and analysis on the preprocessed ocean surface remote sensing data to obtain the distribution characteristics of ocean surface pollutants.

[0029] Further, reconstructing the first pollution diffusion network according to the analysis result to obtain the second pollution diffusion network includes: Determining the surface diffusion path and surface distribution area of pollutants in the target sea area according to the distribution characteristics.

[0030] Adjust the nodes and connection paths of the first pollution diffusion network according to the surface diffusion path and the surface distribution area.

[0031] Reconstruct the first pollution diffusion network according to the adjusted nodes and connection paths to obtain a second pollution diffusion network.

[0032] Further, the step of inputting the periodic water quality data into the second pollution diffusion network for two-way propagation analysis to obtain a forward propagation path and a reverse propagation path includes: Input the periodic water quality data into the second pollution diffusion network to simulate the forward diffusion process of pollutants from the pollution source to the affected area, and obtain the forward propagation path.

[0033] Input the result of the forward diffusion process into the second pollution diffusion network to trace the source of pollutants, and obtain the reverse propagation path.

[0034] Further, the step of adjusting the nodes of the second pollution diffusion network according to the segment offset between the forward propagation path and the reverse propagation path to obtain a target pollution diffusion network includes: Calculate the spatial offset and time offset between the forward propagation path and the reverse propagation path.

[0035] Obtain the segmented spatial offset and segmented time offset respectively according to the distribution characteristics of the spatial offset and the time offset.

[0036] Adjust the node positions of the second pollution diffusion network according to the segmented spatial offset, and adjust the connection path weights of the second pollution diffusion network according to the segmented time offset to obtain a target pollution diffusion network.

[0037] Further, the step of processing the seawater pollution diffusion characteristics based on a graph neural network to obtain a first pollution diffusion network in the target sea area includes: Construct a graph neural network model, input the seawater pollution diffusion characteristics into the graph neural network model, and simulate the dynamic diffusion process of pollutants in the target sea area through graph convolution operations and time convolution operations.

[0038] Generate a first pollution diffusion network in the target sea area according to the simulation result of the graph neural network model.

[0039] Compared with the prior art, the beneficial effects of the embodiments of the present invention are at least one of the following: Combined with water quality fingerprint analysis, ocean current field atlas, and graph neural network, it is possible to more accurately simulate the diffusion path of pollutants, thereby improving the accuracy of pollution source localization; introducing random Poisson noise into the water quality fingerprint feature data and replacing the eigenvalue of the noise seed through typical pollution data enhances the adaptability of data noise to ocean data and can better adapt to the actual situation of ocean pollution monitoring; using the graph neural network to process the diffusion characteristics of seawater pollution can quickly simulate the dynamic diffusion process of pollutants and improve the analysis efficiency; reconstructing the pollution diffusion network through remote sensing data and using two-way propagation analysis and segmented offset adjustment to further optimize the pollution diffusion network, reduce errors, and improve the reliability of the tracing results. Brief Description of the Drawings

[0040] Figure 1 It is a flowchart of the steps of the big data analysis method for ocean pollution tracing based on spatio-temporal graph neural network in one embodiment of the present invention. Detailed Embodiments

[0041] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0042] In the description of this application, the terms "first", "second", "third", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first", "second", "third", etc. may explicitly or implicitly include one or more of such features. In the description of this application, unless otherwise specified, the meaning of "a plurality" is two or more.

[0043] In the description of the present application, it should be noted that unless otherwise clearly specified and defined, the terms "installation", "connection", and "coupling" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, and it can be the communication inside two components. The terms "vertical", "horizontal", "left", "right", "upper", "lower" and similar expressions used herein are only for the purpose of illustration, rather than indicating or implying that the device or component referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus cannot be construed as a limitation to the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific circumstances.

[0044] In the description of the present application, it should be noted that unless otherwise defined, all technical and scientific terms used in the present invention have the same meanings as those commonly understood by those skilled in the technical field to which the present invention belongs. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific circumstances.

[0045] The sources of marine pollution are extensive and complex, including industrial wastewater, domestic sewage, agricultural non-point source pollution, ship oil spills, marine litter, etc. These pollutants enter the ocean through various channels such as rivers, atmospheric deposition, and maritime transportation, making it extremely difficult to locate and trace the pollution sources. Moreover, the ocean is a complex fluid environment, and the diffusion of pollutants in the ocean is affected by various marine dynamic factors such as ocean currents, tides, and wind fields. Monitoring and treating only the seawater at the polluted area can hardly fundamentally solve the problem of seawater pollution. Therefore, this embodiment proposes a big data analysis method for marine pollution source tracing based on a spatio-temporal graph neural network. The specific implementation process is as follows: In the actual marine pollution source tracing analysis, the seawater data collected in real time is input into the target pollution diffusion network, and a marine pollution source tracing report for the target sea area is generated according to the output result of the target pollution diffusion network.

[0046] Among them, the construction process of the target pollution diffusion network is shown in Figure 1 , Figure 1 Fig. shows the flow chart of the steps of the big data analysis method for marine pollution source tracing based on a spatio-temporal graph neural network in one embodiment of the present invention, including steps S11 - S15: S11. Perform water quality fingerprint analysis on the water body collection data of the target sea area obtained, and obtain the periodic water quality data of the target sea area.

[0047] Before collecting water body data, it is necessary to determine an appropriate water body collection interval. In this embodiment, the preset collection interval is determined according to the sea current velocity of each monitoring point in the target sea area and the monitoring requirements of the target sea area. The faster the seawater flow velocity, the shorter the collection interval of the current monitoring point should be, because a faster sea current velocity will cause pollutants to spread to a farther distance in a short time. In order to be able to capture the changes of pollutants in time, it is necessary to collect samples more frequently in order to more accurately reflect the dynamic changes of water quality.

[0048] Moreover, since the environmental requirement standards of each sea area are different, the monitoring requirements are obtained according to the standards for marine pollution purification in the target sea area, so as to adjust the preset collection interval, and then collect seawater samples of each monitoring point in the target sea area at the preset collection interval.

[0049] Water quality fingerprint data refers to a unique data set that can reflect the characteristics of a water body obtained by analyzing a series of key water quality indicators of a water body sample. Water quality fingerprints are similar to the uniqueness of human fingerprints and can provide an identification for the water quality status of a specific water body. The acquisition of water quality fingerprint data requires the monitoring of multiple key water quality indicators of the sample, including but not limited to chemical indicators such as chemical oxygen demand, heavy metal content, dissolved oxygen content, ammonia nitrogen content, and petroleum pollutant concentration, as well as physical indicators such as water temperature and turbidity, and also includes some biological indicators such as plankton species, algal content, and bacterial content.

[0050] In addition, since the monitoring object of this embodiment is a marine polluted water source, a considerable part of the sources of marine pollution come from the pollution brought by the production of surrounding chemical plants. Therefore, the key water quality indicators of this embodiment can also include heavy metal content (such as mercury, lead, cadmium, etc.) and organic pollutants (such as polycyclic aromatic hydrocarbons, pesticide residues, etc.). Specifically, the types of pollutants in seawater are shown in Table 1, and the detection results of these indicators together constitute the basic water quality fingerprint characteristic data of the target sea area.

[0051] Table 1 Marine Pollutant Information Table

[0052] The change of water quality indicators is a dynamic process and is affected by various factors, such as the emission law of pollution sources, marine dynamic conditions (such as sea currents, tides), etc. In order to better understand the change law of water quality indicators, it is necessary to perform time-domain analysis on the characteristic index. The specific analysis process is as follows: Arrange the characteristic indices of seawater samples collected at the same monitoring point at different times in chronological order to form time series data.

[0053] By plotting the change trend curve of the time series data, the change of key water quality indicators over time can be visually observed.

[0054] Since the changes in water quality indicators are usually non - linear, non - linear fitting methods need to be used to more accurately describe the change trend. Non - linear fitting can use polynomial fitting, exponential fitting or other suitable mathematical models, which will not be elaborated here.

[0055] Through non - linear fitting, a smoother and more accurate change trend curve can be obtained, so as to extract the time - domain characteristic data that can reflect the change law of water quality indicators in the time dimension.

[0056] In addition to time - domain analysis, frequency - domain analysis is also an important method for studying the changes in water quality indicators. Frequency - domain analysis can reflect the periodicity and frequency characteristics of the changes in water quality indicators, and help to identify pollution sources and understand the diffusion law of pollutants. The specific frequency - domain analysis process is as follows: Performing Fourier transform on the characteristic indices of each key water quality indicator can decompose the time - series data into a combination of sine waves and cosine waves with different frequencies, thus obtaining the spectrum data.

[0057] Calculating the power spectral density that reflects the energy distribution of different frequency components in the signal from the spectrum data. By calculating the power spectral density, the main frequency - component data of the changes in water quality indicators can be identified. According to the frequency - component data corresponding to each characteristic index, the frequency - domain characteristic data of the seawater sample can be obtained, which is used to provide additional information about the periodicity and regularity of water quality changes.

[0058] Inputting the time - domain characteristic data and the frequency - domain characteristic data into the feature fusion algorithm. Through the processing of the algorithm, these data are fused into a comprehensive feature vector. The selection of the feature fusion algorithm can be linear discriminant analysis or other machine - learning algorithms. In this embodiment, the principal component analysis method is used to fuse the two to obtain the water - quality fingerprint characteristic data.

[0059] Since seawater is constantly changing, when performing water - quality fingerprint analysis, in order to more realistically simulate the dynamic changes of the water quality in the target sea area, the random fluctuations in the water - quality data need to be considered. In this embodiment, random Poisson noise is generated to simulate this random fluctuation and added to the water - quality fingerprint characteristic data, so as to obtain data closer to the actual situation.

[0060] When conducting water quality monitoring in the target sea area, historical pollution statistical data is an important reference basis. The historical pollution statistical data records the pollution situation in the target sea area over a past period, including information such as the types, concentrations, and occurrence frequencies of pollutants. By analyzing these historical data, it can be found that there are certain random fluctuation characteristics in the water quality data. Poisson noise is a typical random noise, and its probability distribution conforms to the Poisson distribution, which can better simulate this random fluctuation. Therefore, in this embodiment, the historical pollution statistical data of the target sea area is used to generate random Poisson noise to simulate the random changes in the water quality data.

[0061] The generation of Poisson noise depends on the λ value of the noise seed, which determines the intensity and distribution characteristics of the noise. However, there are differences in the pollution characteristics of different sea areas, and the Poisson noise generated directly using the default λ value may not accurately reflect the actual water quality fluctuations in the target sea area. To make the generated Poisson noise more conform to the pollution characteristics of the target sea area, this embodiment introduces a noise replacement value. The noise replacement value is determined based on the typical pollution data of the target sea area, and these typical pollution data reflect the main characteristics and common patterns of the pollution in the target sea area. By replacing the λ value of the noise seed with the noise replacement value, the distribution characteristics of the Poisson noise can be adjusted to make it closer to the random fluctuations in the actual water quality data of the target sea area.

[0062] Specifically, the typical pollution data adopted in this embodiment is obtained from the historical pollution statistical data. The historical pollution statistical data records the pollution situation in the target sea area over a past period, including information such as the types, concentrations, occurrence frequencies, and temporal and spatial distributions of pollution events of pollutants. By analyzing these historical data, representative pollution events and pollutant characteristics can be screened out, thereby determining the typical pollution data.

[0063] For example, if high concentrations of the heavy metal mercury have been found in a certain sea area during multiple past detections, and its sources are mainly concentrated in a specific area, then the concentration of mercury and its distribution characteristics can be regarded as one of the typical pollution data of this sea area.

[0064] The typical pollution data is also related to the main pollution sources in the target sea area. If the main pollution source in a certain sea area is the discharge from a chemical plant, then the typical pollution data may include the concentrations and distributions of specific organic pollutants (such as polycyclic aromatic hydrocarbons) or heavy metals (such as mercury, lead, cadmium, etc.).

[0065] Adding the random Poisson noise to the water quality fingerprint feature data, the periodic water quality data of the target sea area obtained not only contains basic information such as the water quality fingerprint feature data, but also takes into account the random fluctuations of the water body, and can more realistically reflect the changes in the water quality of the target sea area in different time periods.

[0066] S12. Determine the ocean current field map of the target sea area, where the ocean current field map is obtained by analyzing the first ocean dynamic data of the target sea area.

[0067] Obtain the first ocean dynamic data of the target sea area. To ensure the effective integration of ocean dynamic data with water quality data and other relevant data, it is necessary to perform time alignment on the first ocean dynamic data, adjust data from different sources and with different time resolutions to a unified time benchmark, and obtain the second ocean dynamic data.

[0068] Extract features from the second ocean dynamic data to obtain ocean motion features, and then divide the sea area into multiple grid cells according to the geographical environmental features of the target sea area. The size of the grid can be adjusted according to the resources and requirements in specific actual operations to obtain the ocean motion grid data of the target sea area.

[0069] A geographic information system (GIS) can obtain the geographical information of the target sea area, including information such as the coastline and sea surface state that affect the distribution and flow direction of ocean currents. Combine the ocean motion grid data with GIS data through visualization software to construct an ocean current field map.

[0070] S13. Extract the seawater pollution diffusion characteristics based on the periodic water quality data and the ocean current field map, and process the seawater pollution diffusion characteristics based on a graph neural network to obtain the first pollution diffusion network of pollutants in the target sea area.

[0071] Since the water quality data at each monitoring point may include multiple key water quality indicators, through point-by-point analysis, the statistical characteristic data of each indicator at that point can be extracted.

[0072] To determine whether the water quality is abnormal, it is necessary to set a reasonable normal threshold range based on historical pollution statistical data and the environmental standards of the target sea area. Historical pollution statistical data records the pollution situation in the target sea area over a past period of time, including information such as the types, concentrations, and occurrence frequencies of pollutants. By analyzing these data, the normal fluctuation range of each key water quality indicator can be determined. And environmental standards are an important basis for judging whether the water quality meets the standards. These standards are usually formulated by environmental protection departments and reflect the minimum requirements for protecting the marine ecological environment and human health.

[0073] Compare the statistical characteristic data of each monitoring point with the set normal threshold range. If the statistical characteristic data of a certain key water quality indicator at a certain monitoring point exceeds the normal threshold range, it is considered that there is an abnormality at that point, and the monitoring point exceeding the threshold range is marked as a potential pollution point. Potential pollution points may be the direct discharge points of pollution sources or the key nodes on the pollutant diffusion path.

[0074] Extract the seawater pollution diffusion characteristics of potential pollution sites. By analyzing these characteristics, the diffusion law of pollutants can be initially understood.

[0075] The graph neural network model can learn the spatio-temporal dependence relationships in the data. In the simulation of marine pollution diffusion, the nodes of the model can represent monitoring points, and the edges can represent the diffusion relationships between monitoring points. In this embodiment, by constructing a graph neural network model, the seawater pollution diffusion characteristics are input into the graph neural network model, including the concentration, diffusion direction, diffusion speed, etc. of pollutants.

[0076] Through graph convolution operations and temporal convolution operations, use the graph neural network model to simulate the dynamic diffusion process of pollutants in the target sea area, including how pollutants diffuse from one monitoring point to another and the changes over time.

[0077] The simulation results of the graph neural network model include information such as the diffusion path, diffusion range, and diffusion speed of pollutants in the target sea area. According to the above simulation results, the first pollution diffusion network of the target sea area is generated, providing an initial diffusion analysis model for marine pollution source tracing analysis.

[0078] S14. Identify and analyze the obtained marine surface remote sensing data, and reconstruct the first pollution diffusion network according to the analysis results to obtain a second pollution diffusion network.

[0079] In marine pollution source tracing analysis, marine surface data is one of the important data sources, mainly because the surface of the ocean is the location that most intuitively and clearly shows the seawater pollution situation. Therefore, it is necessary to reconstruct the first pollution diffusion network of the target sea area according to the marine surface data.

[0080] Satellite remote sensing systems can provide large-scale and high-resolution marine surface information. Therefore, in this embodiment, the marine surface remote sensing data obtained by the satellite remote sensing system is used to analyze the pollution situation on the marine surface. Specifically, the marine surface remote sensing data at least includes sea surface temperature data, chlorophyll concentration data, suspended matter concentration data, and sea surface oil film distribution data, etc.

[0081] Preprocess the obtained ocean surface remote sensing data, and then use image recognition and analysis methods to obtain the distribution characteristics of ocean surface pollutants. Determine the surface diffusion path and surface distribution area of pollutants in the target sea area according to the distribution characteristics. Adjust the nodes and edges of the first pollution diffusion network according to the surface diffusion path and surface distribution area, where the nodes represent pollution sources or areas severely affected, and the edges are the connection paths of the first pollution diffusion network, representing the pollutant propagation paths. Reconstruct the first pollution diffusion network according to the adjusted nodes and connection paths to obtain a second pollution diffusion network combined with satellite remote sensing data, further improving the accuracy and efficiency of ocean pollution source tracing.

[0082] S15. Input the periodic water quality data into the second pollution diffusion network for two-way propagation analysis to obtain the forward propagation path and the reverse propagation path. Adjust the nodes of the second pollution diffusion network according to the segmented offset between the forward propagation path and the reverse propagation path to obtain the target pollution diffusion network.

[0083] After the construction of the second pollution diffusion network is completed, the second pollution diffusion network can be used to trace the propagation path and source of pollutants. However, the actual ocean environment is complex and variable, and the network obtained by unidirectional simulation of only the collected data may not fully and accurately reflect the true diffusion situation of pollutants. Therefore, in order to further improve the accuracy and reliability of the source tracing analysis, the model needs to be more carefully optimized and adjusted.

[0084] Therefore, this embodiment adopts the method of two-way propagation analysis. First, input the periodic water quality data into the second pollution diffusion network in chronological order according to a certain time interval to simulate the forward diffusion process of pollutants from the pollution source to the affected area and obtain the forward propagation path.

[0085] Then, input the result of the forward diffusion process into the second diffusion network to trace the pollutants and obtain the reverse propagation path. Among them, the reverse propagation path is obtained by inversely analyzing the result of the forward diffusion process to determine the source of the pollutants.

[0086] Through two-way propagation analysis, the consistency of the forward propagation and the reverse propagation can be verified according to the diffusion characteristics of pollutants in the ocean. If the forward and reverse propagation paths can match well, it indicates that the simulation result of the second diffusion network is relatively accurate. If there is a deviation between the two, the reasons for these deviations need to be analyzed and the model needs to be adjusted. Specifically, the adjustment process is as follows: The spatial offset reflects the spatial difference between the forward and reverse propagation paths, while the time offset reflects the temporal difference between the two. Based on the specific data points on the paths, the path offsets of the two are obtained by analyzing the position and timeline information of the corresponding points on the forward and reverse propagation paths, and the spatial offset and time offset between the forward propagation path and the reverse propagation path are calculated.

[0087] By observing the variation characteristics of the offset in different paragraphs or time periods, it is divided into different segments, and the segment spatial offset of the spatial offset on different segments is calculated. If the variation range of the segment spatial offset between adjacent segments always remains within a constant range, it indicates that there is no spatial offset between the forward propagation path and the reverse propagation path, but a time offset occurs. At this time, the structure of the second pollution diffusion network is not adjusted. Otherwise, the positions of the nodes and connection paths of the second pollution diffusion network need to be adjusted according to the segment spatial offset of each segment.

[0088] After processing the second pollution diffusion network according to the spatial offset, the segment time offset of the second pollution diffusion network on different segments is obtained according to the time offset, and the weights of each connection path in the second pollution diffusion network are adjusted according to the segment time offset to indirectly adjust the diffusion speed of the pollutants simulated by the second pollution diffusion network, so as to obtain the target pollution diffusion network.

[0089] The big data analysis method for tracing the source of marine pollution based on spatio-temporal graph neural network of the present invention combines water quality fingerprint analysis, ocean current field atlas and graph neural network, and can more accurately simulate the diffusion path of pollutants, thereby improving the accuracy of pollution source location; random Poisson noise is introduced into the water quality fingerprint feature data, and the eigenvalue of the noise seed is replaced by typical pollution data, enhancing the adaptability of the data noise to ocean data and being able to better adapt to the actual situation of marine pollution monitoring; the graph neural network is used to process the characteristics of seawater pollution diffusion, and the dynamic diffusion process of pollutants can be quickly simulated, improving the analysis efficiency; the pollution diffusion network is reconstructed by remote sensing data, and the two-way propagation analysis and segment offset adjustment are used to further optimize the pollution diffusion network, reduce errors and improve the reliability of the tracing results.

[0090] The above embodiments only represent several implementation manners of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as a limitation on the scope of the patent of the present invention. It should be pointed out that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the appended claims.

Claims

1. A big data analysis method for tracing the source of marine pollution based on spatio-temporal graph neural network, characterized in that, Including: In actual marine pollution source tracing analysis, the seawater data collected in real time is input into the target pollution diffusion network, and a marine pollution source tracing report of the target sea area is generated according to the output result of the target pollution diffusion network. Among them, the construction process of the target pollution diffusion network includes: Performing water quality fingerprint analysis on the water body collection data of the target sea area obtained to obtain the periodic water quality data of the target sea area; Determining the marine current field map of the target sea area, where the marine current field map is obtained by analyzing the first marine dynamics data of the target sea area; Extracting the seawater pollution diffusion characteristics according to the periodic water quality data and the marine current field map, and processing the seawater pollution diffusion characteristics based on a graph neural network to obtain the first pollution diffusion network of pollutants in the target sea area; Performing identification and analysis on the obtained marine surface remote sensing data, and reconstructing the first pollution diffusion network according to the analysis result to obtain the second pollution diffusion network; Inputting the periodic water quality data into the second pollution diffusion network for two-way propagation analysis to obtain a forward propagation path and a reverse propagation path, and adjusting the nodes of the second pollution diffusion network according to the segment offset between the forward propagation path and the reverse propagation path to obtain the target pollution diffusion network.

2. The method for big data analysis of ocean pollution source tracing based on spatio-temporal graph neural network according to claim 1, wherein The performing water quality fingerprint analysis on the water body collection data of the target sea area obtained to obtain the periodic water quality data of the target sea area includes: Determining a preset collection interval according to the monitoring requirements of the target sea area and the sea current speed; Collecting seawater samples at each monitoring point of the target sea area at the preset collection interval; Performing key water quality index analysis on each of the seawater samples to obtain corresponding water quality fingerprint characteristic data; Generate random Poisson noise based on the historical pollution statistics of the target sea area, and use the noise replacement value to replace the value of the noise seed of the random Poisson noise, where the noise replacement value is determined by the typical pollution data of the target sea area; λ value, wherein the noise replacement value is determined by the typical pollution data of the target sea area; Adding the random Poisson noise to the water quality fingerprint characteristic data to obtain the periodic water quality data of the target sea area.

3. The method for big data analysis of ocean pollution source tracing based on spatio-temporal graph neural network according to claim 2, wherein The performing key water quality index analysis on each of the seawater samples to obtain corresponding water quality fingerprint characteristic data includes: Performing key water quality index analysis on each of the seawater samples to obtain the characteristic index of each key water quality index of the seawater sample; Performing time domain analysis on each of the characteristic indexes to obtain the change trend curve of the corresponding key water quality index in the time series, and obtaining the time domain characteristic data of the seawater sample by non-linearly fitting each of the change trend curves; Obtaining the frequency spectrum data of each of the characteristic indexes through Fourier transform, calculating the power spectral density of the corresponding characteristic index according to each of the frequency spectrum data to obtain the frequency component data of the characteristic index, and obtaining the frequency domain characteristic data of the seawater sample according to each of the frequency component data; Generating the water quality fingerprint characteristic data of the seawater sample based on the feature fusion algorithm according to the time domain characteristic data and the frequency domain characteristic data.

4. The method for big data analysis of ocean pollution source tracing based on spatio-temporal graph neural network according to claim 3, wherein, The extracting the seawater pollution diffusion characteristics according to the periodic water quality data and the marine current field map includes: Performing point-by-point analysis on the periodic water quality data to obtain the statistical characteristic data of each key water quality index; Set a normal threshold range for each of the key water quality indicators according to the historical pollution statistics and the environmental standards of the target sea area; Compare and analyze the statistical characteristic data and the normal threshold range. If there is a piece of statistical characteristic data that exceeds the corresponding normal threshold range, mark the corresponding point as a potential pollution point, and extract the seawater pollution diffusion characteristics of the potential pollution point.

5. The method for big data analysis of marine pollution source tracing based on spatio-temporal graph neural network according to claim 1, wherein The determination of the ocean current field map of the target sea area includes: Obtain the first ocean dynamics data of the target sea area, and perform time alignment on the first ocean dynamics data to obtain the second ocean dynamics data; Extract features from the second ocean dynamics data to obtain ocean motion characteristics; Construct the ocean motion grid data of the target sea area according to the ocean motion characteristics; Construct the ocean current field map of the target sea area according to the ocean motion grid data and the geographic information system data of the target sea area.

6. The method for big data analysis of ocean pollution source tracing based on spatio-temporal graph neural network according to claim 1, wherein The ocean surface remote sensing data at least includes sea surface temperature data, chlorophyll concentration data, suspended matter concentration data, and sea surface oil film distribution data; The identification and analysis of the obtained ocean surface remote sensing data includes: Obtain the ocean surface remote sensing data of the target sea area monitored by the satellite remote sensing system, and preprocess the ocean surface remote sensing data; Perform image recognition and analysis on the preprocessed ocean surface remote sensing data to obtain the distribution characteristics of ocean surface pollutants.

7. The method for big data analysis of ocean pollution source tracing based on spatio-temporal graph neural network according to claim 6, characterized in that The reconstruction of the first pollution diffusion network according to the analysis results to obtain the second pollution diffusion network includes: Determine the surface diffusion path and surface distribution area of pollutants in the target sea area according to the distribution characteristics; Adjust the nodes and connection paths of the first pollution diffusion network according to the surface diffusion path and the surface distribution area; Reconstruct the first pollution diffusion network according to the adjusted nodes and connection paths to obtain the second pollution diffusion network.

8. The method for big data analysis of marine pollution source tracing based on spatio-temporal graph neural network according to claim 1, characterized in that, The input of the periodic water quality data into the second pollution diffusion network for two-way propagation analysis to obtain the forward propagation path and the reverse propagation path includes: Input the periodic water quality data into the second pollution diffusion network to simulate the forward diffusion process of pollutants from the pollution source to the affected area to obtain the forward propagation path; Input the result of the forward diffusion process into the second pollution diffusion network to trace the source of pollutants to obtain the reverse propagation path.

9. The method for big data analysis of ocean pollution source tracing based on spatio-temporal graph neural network according to claim 8, wherein The adjustment of the nodes of the second pollution diffusion network according to the segment offset between the forward propagation path and the reverse propagation path to obtain the target pollution diffusion network includes: Calculate the spatial offset and time offset between the forward propagation path and the reverse propagation path; Obtain the segmented spatial offset and segmented time offset respectively according to the distribution characteristics of the spatial offset and the time offset; Adjust the node positions of the second pollution diffusion network according to the segmented spatial offset, and adjust the connection path weights of the second pollution diffusion network according to the segmented time offset to obtain the target pollution diffusion network.

10. The method for big data analysis of ocean pollution source tracing based on spatio-temporal graph neural network according to claim 1, characterized in that Processing the seawater pollution diffusion characteristics based on a graph neural network to obtain a first pollution diffusion network of pollutants in the target sea area, including: Constructing a graph neural network model, inputting the seawater pollution diffusion characteristics into the graph neural network model, and simulating the dynamic diffusion process of pollutants in the target sea area through graph convolution operations and temporal convolution operations; Generating a first pollution diffusion network of the target sea area according to the simulation results of the graph neural network model.

Images