Marine antibiotic pollution migration analysis method and system based on deep learning

Abstract

The present invention discloses a method and system for analyzing the migration of marine antibiotic pollution based on deep learning, including: obtaining a water body image of a target marine area to form a first data set; obtaining the water flow characteristics of the target marine area according to the first data set, and analyzing the water flow law of the target marine area using a SARIMA model; obtaining the driving factors that affect the flow change of marine water bodies, using a Logistic regression model to perform driving force analysis, constructing a state transfer rule for water body flow and a pollution migration prediction model; obtaining water quality pollution detection information of each sampling point, performing pollution diffusion simulation based on the water body flow law to obtain the pollution diffusion coefficient, performing pollution migration prediction, and obtaining pollution migration prediction information; analyzing the antibiotic pollution status of the target marine area according to the pollution migration prediction information, analyzing the area that needs to be treated with antibiotic pollution, and generating an antibiotic pollution control plan. Improving the prediction accuracy of pollutant migration provides a scientific basis for pollution prevention and control.

Description

Technical Field

[0001] The present invention relates to the technical field of pollution migration analysis, and in particular to a method and system for analyzing marine antibiotic pollution migration based on deep learning. Background Art

[0002] Antibiotics enter water bodies through wastewater, agricultural runoff and other pathways, and spread in rivers, lakes and oceans, causing a series of environmental problems. Antibiotics in the marine environment may not only cause the spread of antibiotic-resistant bacteria, but also affect the growth, reproduction and ecosystem health of marine organisms.

[0003] Current research methods mainly rely on traditional hydrological and hydrodynamic models. Although these models can describe the macroscopic diffusion and migration paths of pollutants, they have limitations in dealing with multi-dimensional and complex nonlinear data. The pollution migration process in the marine environment is affected by many factors, such as ocean currents, wind fields, tides, water temperature, etc. The interaction of these factors makes the migration of pollutants show highly nonlinear spatiotemporal variation characteristics. Traditional models have difficulty in dealing with these complex dynamic changes, especially in long-term predictions and capturing subtle changes. Therefore, there is an urgent need for an antibiotic pollution migration analysis method that can integrate multi-source data and handle complex nonlinear dynamics, improve the prediction accuracy of pollutant migration, provide a scientific basis for environmental management and pollution prevention, and help formulate effective governance strategies. Summary of the invention

[0004] The present invention overcomes the defects of the prior art and provides a method and system for analyzing the migration of marine antibiotic pollution based on deep learning. Its important purpose is to improve the prediction accuracy of pollutant migration, provide a scientific basis for environmental management and pollution prevention, and help formulate effective governance strategies.

[0005] To achieve the above objectives, the present invention provides a method for analyzing the migration of marine antibiotic pollution based on deep learning, comprising:

[0006] Obtain historical remote sensing monitoring image information and perform image preprocessing, extract water body images of the target ocean area and perform time series processing to form a first data set;

[0007] Acquire the water flow characteristics of the target ocean area according to the first data set to construct a water flow characteristic map, rasterize the water flow characteristic map and perform interpolation processing, use the SARIMA model to analyze the water flow law of the target ocean area, and obtain water flow law analysis information;

[0008] Obtain the driving factors that affect the flow of ocean water bodies, use the Logistic regression model to analyze the driving force, calculate the influence weights of each driving factor based on the driving force analysis results, generate the state transition rules of the water flow state, and build a pollution migration prediction model;

[0009] Acquire water quality pollution detection information of each sampling point, perform pollution diffusion simulation based on the water flow law analysis information using finite element analysis to obtain pollution diffusion coefficient, and perform pollution migration prediction using a pollution migration prediction model to obtain pollution migration prediction information;

[0010] The antibiotic pollution status of the target marine area is analyzed based on the pollution migration prediction information, the area that needs antibiotic pollution control is analyzed, and an antibiotic pollution control plan is generated.

[0011] In this solution, the acquisition of historical remote sensing monitoring image information and image preprocessing, the extraction of water body images of the target ocean area and time series processing to form the first data set specifically include:

[0012] Based on big data retrieval, historical remote sensing monitoring image information of the target ocean area is collected, and image preprocessing is performed on the historical remote sensing monitoring image information, including radiation correction, atmospheric correction and geometric correction, to obtain preprocessed image information;

[0013] Calculate the NDWI index of the target ocean area according to the preprocessed image information, extract water body features according to the calculated NDWI index, and obtain a water body image of the target ocean area;

[0014] A timestamp is preset, and based on the monitoring time characteristics corresponding to the water body image of the target ocean area, the multi-temporal image combination is performed on the water body image of the target ocean area according to the preset timestamp to construct a first data set.

[0015] In this solution, the water flow characteristics of the target ocean area are obtained according to the first data set to construct a water flow characteristic map, the water flow characteristic map is rasterized and interpolated, and the water flow law of the target ocean area is analyzed using the SARIMA model to obtain water flow law analysis information, which specifically includes:

[0016] Acquire a first data set, extract water flow characteristics of a target ocean area according to the first data set, and generate a water flow characteristic map of the target ocean area based on the extracted water flow characteristics;

[0017] Preset the grid unit size, rasterize the water flow characteristic map of the target ocean area to obtain a water flow characteristic grid map, and use the spline function interpolation method to interpolate the unknown characteristic grid;

[0018] According to the coordinates of the unknown grid cell, the water flow characteristics of the neighboring grid cell are obtained, the interpolation interval of the unknown grid cell is formed, the spline equation of the interpolation interval is established and solved to obtain the spline coefficient;

[0019] Interpolate the unknown characteristic grid according to the spline coefficient of the interpolation interval to generate an interpolated water flow characteristic grid map, preset a time period, and combine the interpolated water flow characteristic grid map to form a water flow characteristic time series;

[0020] The constructed water flow characteristic time series is used as the model input, and the SARIMA model is used to analyze the water flow law in the target ocean area to obtain the water flow law analysis information.

[0021] In this scheme, the driving factors that affect the flow of ocean water bodies are obtained, the driving force analysis is performed using the Logistic regression model, the influence weights of each driving factor are calculated according to the driving force analysis results to generate the state transition rules of the water body flow state, and the pollution migration prediction model is constructed, which specifically includes:

[0022] Based on data retrieval, the driving factors of regional ocean water flow changes are obtained, the Pearson correlation coefficient between each driving factor and water flow changes is calculated, and the driving factors related to water flow are screened by judging with the preset threshold;

[0023] According to the selected driving factors, the historical data corresponding to the driving factors are obtained, the water flow change is taken as the dependent variable, the driving factors are taken as the independent variables, and the driving force analysis is performed using the Logistic regression model to obtain the driving force analysis information;

[0024] Acquire water flow law analysis information, and construct a state transfer matrix of water flow state in the target ocean area based on a Markov algorithm combined with the water flow law analysis information to form an initial state transfer rule;

[0025] Calculate the influence weight of each driving factor according to the driving force analysis information, and correct the initial state transition rule by the calculated influence weight of each driving factor to obtain a corrected state transition rule;

[0026] A pollution migration prediction model is constructed based on LSTM, the pollution migration prediction rule is defined using the corrected state transition rule, and a training data set is constructed to train the pollution migration prediction model.

[0027] In this solution, the water pollution detection information of each sampling point is obtained, based on the water flow law analysis information, the pollution diffusion simulation is performed using the finite element analysis method to obtain the pollution diffusion coefficient, and the pollution migration prediction model is used to predict the pollution migration to obtain the pollution migration prediction information, which specifically includes:

[0028] Conduct water pollution detection on the target marine area to obtain water pollution detection information of each sampling point, compare the water pollution detection information with the preset threshold, evaluate the antibiotic pollution status of the area corresponding to each sampling point, and obtain antibiotic pollution status information;

[0029] Acquire water flow law analysis information, and model the target ocean area using finite element analysis method based on the water flow law analysis information to form a regional water flow simulation model;

[0030] Gridding the regional water flow simulation model, defining an initial contaminated area based on the antibiotic contamination status information, and setting regional water flow rules and boundary conditions according to the corrected state transition rules;

[0031] Obtain real-time meteorological information and input it into the initialized regional water flow simulation model to perform pollution diffusion simulation, and calculate the pollution diffusion coefficient based on the simulation results;

[0032] The human activity data of the target ocean area is obtained and the pollution migration driving weight is set in combination with the real-time meteorological information. The pollution diffusion coefficient is used to define the pollution migration change rate and the antibiotic pollution status information is input into the pollution migration prediction model to perform pollution migration prediction and obtain pollution migration prediction information.

[0033] In this solution, the antibiotic pollution status of the target marine area is analyzed according to the pollution migration prediction information, the area that needs antibiotic pollution control is analyzed and an antibiotic pollution control plan is generated, which specifically includes:

[0034] Generate a pollution migration path map of the target ocean area according to the pollution migration prediction information, calculate the migration pollution concentration and migration pollution time of the antibiotic pollution in the target ocean area based on the pollution migration path map combined with the pollution diffusion coefficient, and obtain first analysis information;

[0035] Divide the target ocean area by pollution concentration according to the first analysis information, form an antibiotic pollution change map of the target ocean area in the future time period, and extract pollution concentration characteristics and regional location characteristics of different concentration areas;

[0036] Based on the regional location characteristics of each pollution concentration, the ecological data and hydrological data of the corresponding area are obtained, and the environmental carrying capacity of each area at the current moment is evaluated in combination with the antibiotic pollution status information to obtain the second analysis information;

[0037] The second analysis information is judged with the extracted pollution concentration characteristics of different concentration areas, and the marine area to be treated is defined according to the judgment result to obtain the information of the area to be treated;

[0038] Based on the first analysis information, the time point when the antibiotic pollution in the area to be treated exceeds the environmental carrying capacity is calculated, the areas to be treated are sorted, and the treatment priority of each area to be treated is set according to the sorting result to obtain the third analysis information;

[0039] According to the third analysis information, data retrieval is performed to obtain an antibiotic pollution control plan and provide antibiotic pollution control recommendations.

[0040] The second aspect of the present invention provides a marine antibiotic pollution migration analysis system based on deep learning, the system comprising: a memory, a processor, the memory containing a marine antibiotic pollution migration analysis method program based on deep learning, the marine antibiotic pollution migration analysis method program based on deep learning when executed by the processor to implement the following steps:

[0041] Obtain historical remote sensing monitoring image information and perform image preprocessing, extract water body images of the target ocean area and perform time series processing to form a first data set;

[0042] Acquire the water flow characteristics of the target ocean area according to the first data set to construct a water flow characteristic map, rasterize the water flow characteristic map and perform interpolation processing, use the SARIMA model to analyze the water flow law of the target ocean area, and obtain water flow law analysis information;

[0043] Obtain the driving factors that affect the flow of ocean water bodies, use the Logistic regression model to analyze the driving force, calculate the influence weights of each driving factor based on the driving force analysis results, generate the state transition rules of the water flow state, and build a pollution migration prediction model;

[0044] Acquire water quality pollution detection information of each sampling point, perform pollution diffusion simulation based on the water flow law analysis information using finite element analysis to obtain pollution diffusion coefficient, and perform pollution migration prediction using a pollution migration prediction model to obtain pollution migration prediction information;

[0045] The antibiotic pollution status of the target marine area is analyzed based on the pollution migration prediction information, the area that needs antibiotic pollution control is analyzed, and an antibiotic pollution control plan is generated.

[0046] The present invention discloses a method and system for analyzing the migration of marine antibiotic pollution based on deep learning, including: obtaining a water body image of a target marine area to form a first data set; obtaining the water flow characteristics of the target marine area according to the first data set, and analyzing the water flow law of the target marine area using a SARIMA model; obtaining the driving factors that affect the flow change of marine water bodies, using a Logistic regression model to perform driving force analysis, constructing a state transfer rule for water body flow and a pollution migration prediction model; obtaining water quality pollution detection information of each sampling point, performing pollution diffusion simulation based on the water body flow law to obtain the pollution diffusion coefficient, performing pollution migration prediction, and obtaining pollution migration prediction information; analyzing the antibiotic pollution status of the target marine area according to the pollution migration prediction information, analyzing the area that needs to be treated with antibiotic pollution, and generating an antibiotic pollution control plan. Improving the prediction accuracy of pollutant migration provides a scientific basis for pollution prevention and control. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments or exemplary embodiments of the present invention, the drawings required for use in the embodiments or exemplary descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained according to the drawings without paying creative work.

[0048] Figure 1 A flow chart of a method for analyzing marine antibiotic pollution migration based on deep learning provided by one embodiment of the present invention;

[0049] Figure 2 A flow chart of a method for predicting antibiotic contamination migration provided by an embodiment of the present invention;

[0050] Figure 3 A block diagram of a marine antibiotic pollution migration analysis system based on deep learning provided by an embodiment of the present invention;

[0051] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0052] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0053] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0054] Figure 1 A flow chart of a method for analyzing marine antibiotic pollution migration based on deep learning provided by one embodiment of the present invention;

[0055] like Figure 1 As shown, the present invention provides a flow chart of a method for analyzing the migration of marine antibiotic pollution based on deep learning, including:

[0056] S102, acquiring historical remote sensing monitoring image information and performing image preprocessing, extracting water body images of the target ocean area and performing time series processing to form a first data set;

[0057] S104, acquiring the water flow characteristics of the target ocean area according to the first data set to construct a water flow characteristic map, rasterizing and interpolating the water flow characteristic map, and analyzing the water flow law of the target ocean area using the SARIMA model to obtain water flow law analysis information;

[0058] S106, obtaining driving factors that affect the flow of ocean water bodies, performing driving force analysis using a Logistic regression model, calculating the influence weights of each driving factor based on the driving force analysis results, generating state transition rules for the flow state of the water body, and constructing a pollution migration prediction model;

[0059] S108, obtaining water pollution detection information of each sampling point, performing pollution diffusion simulation using finite element analysis method based on the water flow law analysis information to obtain pollution diffusion coefficient, and performing pollution migration prediction using pollution migration prediction model to obtain pollution migration prediction information;

[0060] S110, analyzing the antibiotic pollution status of the target marine area according to the pollution migration prediction information, analyzing the area that needs antibiotic pollution control and generating an antibiotic pollution control plan.

[0061] It should be noted that the present invention provides a method and system for analyzing the migration of marine antibiotic pollution based on deep learning, which uses historical remote sensing monitoring image information to obtain water images of the target marine area and form a first data set; extracts the water flow characteristics of the target marine area according to the first data set to construct a water flow characteristic map, rasterizes the water flow characteristic map, and further enriches the water flow characteristics of the target marine area using the spline function interpolation method. The SARIMA model is used to analyze the water flow law of the target marine area to obtain water flow law analysis information. Next, the driving factors affecting the flow change of marine water bodies are analyzed, and the driving force analysis is performed using the Logistic regression model. According to the driving force analysis results, the influence weights of each driving factor are calculated to generate the state transition rules of the water flow state, and a pollution migration prediction model is constructed; based on the water quality pollution detection information of each sampling point, the finite element analysis method is used to perform pollution diffusion simulation to obtain the pollution diffusion coefficient, and then the pollution migration prediction model is used to consider the weight of the driving factors of the water flow change to perform pollution migration prediction, and obtain pollution migration prediction information. According to the pollution migration prediction information, the antibiotic pollution status of the target marine area is analyzed, and the area that needs to be treated with antibiotics is analyzed and an antibiotic pollution control plan is generated. Improve the prediction accuracy of pollutant migration, provide a scientific basis for environmental management and pollution prevention, and help formulate effective governance strategies.

[0062] Further, in a preferred embodiment of the present invention, the acquisition of historical remote sensing monitoring image information and image preprocessing, the extraction of target ocean area water body images and time series processing to form a first data set specifically include:

[0063] Based on big data retrieval, historical remote sensing monitoring image information of the target ocean area is collected, and image preprocessing is performed on the historical remote sensing monitoring image information, including radiation correction, atmospheric correction and geometric correction, to obtain preprocessed image information;

[0064] Calculate the NDWI index of the target ocean area according to the preprocessed image information, extract water body features according to the calculated NDWI index, and obtain a water body image of the target ocean area;

[0065] A timestamp is preset, and based on the monitoring time characteristics corresponding to the water body image of the target ocean area, the multi-temporal image combination is performed on the water body image of the target ocean area according to the preset timestamp to construct a first data set.

[0066] It should be noted that the historical remote sensing monitoring image information of the target ocean area covering multiple time periods is first collected. By comprehensively analyzing the remote sensing monitoring images of different periods, the changes in the marine environment can be effectively evaluated. Image preprocessing usually includes three parts: radiation correction, atmospheric correction and geometric correction. Radiation correction is used to eliminate the inconsistency of image brightness caused by changes in sensor performance or differences in lighting conditions, so that images taken at different times and locations are comparable in radiation intensity. Atmospheric correction is to correct the effects of suspended particles, gas molecules, etc. in the atmosphere on light propagation to remove the interference of the atmosphere on image quality. Geometric correction is used to adjust the geometric distortion in the image so that the spatial relationship of the image is consistent with the actual geographical location, thereby obtaining more accurate geographic information. Next, based on these corrected preprocessed image information, the water body index of the target ocean area is calculated. The NDWI index is calculated by the spectral reflectance of a specific band, which is used to enhance the signal of the water body and suppress the interference of vegetation, soil and other objects. The calculated NDWI index can effectively distinguish water from other objects and realize the extraction of water body characteristics. By presetting timestamps, the water body images corresponding to different time points are combined in multiple temporal phases to construct the first data set reflecting the changing trend of water bodies. The multi-temporal combination process can not only capture the dynamic changes of water bodies, but also reflect seasonal fluctuations and long-term trends, providing an important data basis for subsequent analysis and prediction of water body change laws.

[0067] Further, in a preferred embodiment of the present invention, the water flow characteristics of the target ocean area are obtained according to the first data set to construct a water flow characteristic map, the water flow characteristic map is rasterized and interpolated, and the SARIMA model is used to analyze the water flow law of the target ocean area to obtain water flow law analysis information, which specifically includes:

[0068] Acquire a first data set, extract water flow characteristics of a target ocean area according to the first data set, and generate a water flow characteristic map of the target ocean area based on the extracted water flow characteristics;

[0069] Preset the grid unit size, rasterize the water flow characteristic map of the target ocean area to obtain a water flow characteristic grid map, and use the spline function interpolation method to interpolate the unknown characteristic grid;

[0070] According to the coordinates of the unknown grid cell, the water flow characteristics of the neighboring grid cell are obtained, the interpolation interval of the unknown grid cell is formed, the spline equation of the interpolation interval is established and solved to obtain the spline coefficient;

[0071] Interpolate the unknown characteristic grid according to the spline coefficient of the interpolation interval to generate an interpolated water flow characteristic grid map, preset a time period, and combine the interpolated water flow characteristic grid map to form a water flow characteristic time series;

[0072] The constructed water flow characteristic time series is used as the model input, and the SARIMA model is used to analyze the water flow law in the target ocean area to obtain the water flow law analysis information.

[0073] It should be noted that the water flow characteristics of the target ocean area are extracted through the first data set. These characteristics may include dynamic information such as water temperature, current vector, chlorophyll concentration, etc. Based on these extracted water flow characteristics, a water flow characteristic map of the target ocean area is generated to display the spatial distribution and changes of water flow in a visual way. Then, the size of the grid unit is preset, and the water flow characteristic map is rasterized to convert the continuous spatial data into discrete grid units. However, after the rasterization process, there are some unknown grid units whose water flow characteristics are not yet clear. The spline function interpolation method is used to interpolate these unknown characteristic grid units to fill the data gaps and enrich the water characteristics of the target area. Finally, an interpolated water flow characteristic grid map is generated. This grid map not only fills the data gaps, but also can accurately reflect the spatial changes of water flow characteristics in the target area. The preset time period is combined with the interpolated water flow characteristic grid map to form a water flow characteristic time series. The constructed water flow characteristic time series is input into the SARIMA model, and the time series analysis capability of the SARIMA model is used to conduct an in-depth analysis of the water flow law in the target ocean area. Through the calculation and analysis of the model, analytical information on the water flow law is obtained, thereby understanding the water flow trend and changes in the target area.

[0074] Further, in a preferred real-time example of the present invention, the driving factors affecting the flow change of the ocean water body are obtained, the driving force analysis is performed using the Logistic regression model, the influence weights of each driving factor are calculated according to the driving force analysis results to generate the state transition rules of the water body flow state, and the pollution migration prediction model is constructed, which specifically includes:

[0075] Based on data retrieval, the driving factors of regional ocean water flow changes are obtained, the Pearson correlation coefficient between each driving factor and water flow changes is calculated, and the driving factors related to water flow are screened by judging with the preset threshold;

[0076] According to the selected driving factors, the historical data corresponding to the driving factors are obtained, the water flow change is taken as the dependent variable, the driving factors are taken as the independent variables, and the driving force analysis is performed using the Logistic regression model to obtain the driving force analysis information;

[0077] Acquire water flow law analysis information, and construct a state transfer matrix of water flow state in the target ocean area based on a Markov algorithm combined with the water flow law analysis information to form an initial state transfer rule;

[0078] Calculate the influence weight of each driving factor according to the driving force analysis information, and correct the initial state transition rule by the calculated influence weight of each driving factor to obtain a corrected state transition rule;

[0079] A pollution migration prediction model is constructed based on LSTM, the pollution migration prediction rule is defined using the corrected state transition rule, and a training data set is constructed to train the pollution migration prediction model.

[0080] It should be noted that in the actual environment, the change law of water flow is not only caused by natural factors, but also by human activities, such as fishing activities. The driving factors of regional marine water flow changes are obtained through data retrieval, the Pearson correlation coefficient between each driving factor and water flow changes is calculated, and the preset threshold is judged to screen the driving factors related to water flow. The driving force analysis is performed using the Logistic regression model to obtain the driving force analysis information, and the influence of multiple variables on regional water changes is further considered to improve the accuracy of water flow prediction. The state transition matrix of the water flow state in the target ocean area is constructed using the Markov algorithm. The state transition matrix reflects the probability of the ocean water body transitioning between different states and can describe the dynamic changes of water flow. The initial state transition rules are formed through this information. Based on the driving force analysis information, the influence weights of each driving factor are calculated. Different driving factors have different influences on water flow, so it is necessary to correct the initial state transition rules through these influence weights. According to the calculated weights, the initial state transition rules are corrected to improve the prediction accuracy of water flow changes and form more accurate corrected state transition rules. Next, a pollution migration prediction model is constructed based on the long short-term memory network (LSTM). LSTM is a deep learning model commonly used in time series prediction, which can effectively capture long-term dependencies in time series data. Using the corrected state transition rules obtained above, the pollution migration prediction rules are defined, and a training data set containing water flow changes and pollution migration data is constructed. The pollution migration prediction model is trained using this data set to obtain a pollution migration prediction model that meets expectations.

[0081] Further, in a preferred embodiment of the present invention, the water quality pollution detection information of each sampling point is obtained, based on the water flow law analysis information, the pollution diffusion simulation is performed using the finite element analysis method to obtain the pollution diffusion coefficient, and the pollution migration prediction model is used to perform pollution migration prediction to obtain the pollution migration prediction information, which specifically includes:

[0082] Conduct water pollution detection on the target marine area to obtain water pollution detection information of each sampling point, compare the water pollution detection information with the preset threshold, evaluate the antibiotic pollution status of the area corresponding to each sampling point, and obtain antibiotic pollution status information;

[0083] Acquire water flow law analysis information, and model the target ocean area using finite element analysis method based on the water flow law analysis information to form a regional water flow simulation model;

[0084] Gridding the regional water flow simulation model, defining an initial contaminated area based on the antibiotic contamination status information, and setting regional water flow rules and boundary conditions according to the corrected state transition rules;

[0085] Obtain real-time meteorological information and input it into the initialized regional water flow simulation model to perform pollution diffusion simulation, and calculate the pollution diffusion coefficient based on the simulation results;

[0086] The human activity data of the target ocean area is obtained and the pollution migration driving weight is set in combination with the real-time meteorological information. The pollution diffusion coefficient is used to define the pollution migration change rate and the antibiotic pollution status information is input into the pollution migration prediction model to perform pollution migration prediction and obtain pollution migration prediction information.

[0087] It should be noted that water pollution detection is carried out in the target marine area to obtain water pollution detection information. Then, the obtained water pollution detection information is compared and judged with the preset pollution threshold, and the antibiotic pollution status of the area where each sampling point is located is analyzed to determine the real-time antibiotic pollution status.

[0088] Next, based on the obtained water flow law analysis information, the finite element analysis method is used to physically model the target ocean area to simulate the flow of water in the ocean. Subsequently, based on the obtained antibiotic pollution status information, the initial pollution area is defined, that is, which grid cells are set as pollution sources in the simulation model. According to the corrected state transition rules, the water flow rules and boundary conditions in the region are set to ensure that the water flow in the simulation process conforms to the real regional water flow change law. Then, the real-time meteorological information of the target area, such as wind speed, wind direction, rainfall, etc., is obtained, and these real-time meteorological data are input into the initialized regional water flow simulation model to perform simulation calculations of pollution diffusion. Through simulation analysis, the diffusion trend of antibiotic pollution under different time and environmental conditions is judged, and the pollution diffusion coefficient is further calculated. The pollution diffusion coefficient reflects the propagation speed and range of pollutants in the water body, and plays a key role in the accurate prediction of pollution migration. At the same time, the human activities data in the target ocean area, such as port transportation, fishery activities, etc., are obtained, and the driving weight of pollution migration is set in combination with real-time meteorological information. Then, the previously calculated pollution diffusion coefficient is used to define the changing rate of pollution migration, which is combined with the antibiotic pollution status information and sent as input data to the pollution migration prediction model to predict the migration path and diffusion status of pollutants in the target area and obtain pollution migration prediction information.

[0089] Further, in a preferred embodiment of the present invention, the antibiotic pollution status of the target marine area is analyzed according to the pollution migration prediction information, the area requiring antibiotic pollution control is analyzed and an antibiotic pollution control plan is generated, specifically including:

[0090] Generate a pollution migration path map of the target ocean area according to the pollution migration prediction information, calculate the migration pollution concentration and migration pollution time of the antibiotic pollution in the target ocean area based on the pollution migration path map combined with the pollution diffusion coefficient, and obtain first analysis information;

[0091] Divide the target ocean area by pollution concentration according to the first analysis information, form an antibiotic pollution change map of the target ocean area in the future time period, and extract pollution concentration characteristics and regional location characteristics of different concentration areas;

[0092] Based on the regional location characteristics of each pollution concentration, the ecological data and hydrological data of the corresponding area are obtained, and the environmental carrying capacity of each area at the current moment is evaluated in combination with the antibiotic pollution status information to obtain the second analysis information;

[0093] The second analysis information is judged with the extracted pollution concentration characteristics of different concentration areas, and the marine area to be treated is defined according to the judgment result to obtain the information of the area to be treated;

[0094] Based on the first analysis information, the time point when the antibiotic pollution in the area to be treated exceeds the environmental carrying capacity is calculated, the areas to be treated are sorted, and the treatment priority of each area to be treated is set according to the sorting result to obtain the third analysis information;

[0095] According to the third analysis information, data retrieval is performed to obtain an antibiotic pollution control plan and provide antibiotic pollution control recommendations.

[0096] It should be noted that a pollution migration path map of the target ocean area is generated based on the obtained pollution migration prediction information. The pollution migration path map shows the diffusion trajectory of pollutants in the ocean over time. Combined with the previously calculated pollution diffusion coefficient, the migration pollution concentration and migration time of antibiotic pollution in the target area are further analyzed, and the concentration changes of antibiotic pollutants at different locations and the time required to reach these locations are analyzed to obtain the first analysis information. Next, based on the first analysis information, the target ocean area is divided into pollution concentrations, and the area is divided into different pollution levels according to the calculated pollution concentration values, and an image showing the changes in antibiotic pollution in the future time period is generated. Then, the pollution concentration characteristics of each pollution concentration area and the geographical location characteristics of the area are extracted. Based on the location information of different pollution concentration areas, the ecological data and hydrological data of the corresponding area are obtained. Combined with the status information of antibiotic pollution, the environmental carrying capacity of each area at the current moment is evaluated. Environmental carrying capacity refers to the pollution load that a region can withstand without suffering serious ecological damage, and the second analysis information is obtained to measure whether each region has approached or exceeded its environmental carrying capacity. Next, the second analysis information is compared and judged with the pollution concentration characteristics of each pollution concentration area previously extracted to identify which areas have or are about to exceed the environmental carrying capacity of the area, thereby determining the marine areas that need to be governed and obtaining information on the areas to be governed. On this basis, the time nodes when the antibiotic pollution in each area to be governed exceeds the environmental carrying capacity are calculated based on the first analysis information. That is, it is analyzed when each area will reach the critical point of pollution. According to these time nodes, all areas to be governed are sorted, and those areas that are about to face the collapse of the environmental carrying capacity are given priority. According to the sorting results, the governance priority of each area to be governed is set to obtain the third analysis information. Finally, based on the third analysis information, the governance plan related to antibiotic pollution is obtained through data retrieval to make recommendations for antibiotic pollution governance.

[0097] Figure 2 A flow chart of a method for predicting antibiotic contamination migration provided by an embodiment of the present invention;

[0098] like Figure 2 As shown, the present invention provides a flow chart of a method for predicting antibiotic pollution migration, comprising:

[0099] S202, performing water pollution detection on the target ocean area to obtain water pollution detection information of each sampling point;

[0100] S204, judging the water pollution detection information and the preset threshold, evaluating the antibiotic pollution status of the area corresponding to each sampling point, and obtaining antibiotic pollution status information;

[0101] S206, acquiring real-time meteorological information and inputting it into the initialized regional water flow simulation model to perform pollution diffusion simulation, and calculating the pollution diffusion coefficient according to the simulation results;

[0102] S208, setting the pollution migration driving weight, using the pollution diffusion coefficient to define the pollution migration change rate, inputting the antibiotic pollution status information into the pollution migration prediction model to perform pollution migration prediction, and obtaining pollution migration prediction information.

[0103] It should be noted that the migration prediction of marine antibiotic pollution not only needs to consider the water flow characteristics of the target area, but also the external factors, including natural meteorology and human activities. By adding migration driving weights to influence the state change rules of migration prediction, the prediction results can be further made to fit the real environmental change laws, thereby improving the accuracy of antibiotic pollution prediction.

[0104] In addition, in a deep learning-based marine antibiotic pollution migration analysis method provided by the present invention, the following steps are also included:

[0105] Obtain water flow law analysis information in the target ocean area, extract water change characteristics under different meteorological conditions based on the water flow law analysis information, and obtain water change characteristic information;

[0106] The Euclidean distance between the meteorological conditions corresponding to the change characteristics of each water body is calculated, and the categories are divided using a clustering algorithm. The water flow law of the target ocean area is divided according to the meteorological conditions to obtain the first division result information;

[0107] Preset a number of pollution locations, perform pollution diffusion simulation using the regional water flow simulation model based on the first division result information, analyze the pollution diffusion coefficient of each pollution location under different meteorological conditions, and obtain pollution diffusion coefficient analysis information;

[0108] Based on combining the first division result information and the pollution diffusion coefficient analysis information, a pollution migration portrait under different meteorological conditions is constructed to form a first database;

[0109] When making the next pollution migration prediction, the similarity between the real-time meteorological conditions and the pollution location and the pollution migration portrait in the first database is calculated, and the corresponding pollution migration portrait is selected according to the similarity;

[0110] The corresponding water body change characteristics and diffusion coefficients are extracted through the selected pollution migration portrait, and the prediction results of the pollution migration prediction model are corrected.

[0111] Figure 3 A deep learning-based marine antibiotic pollution migration analysis system 3 is provided in one embodiment of the present invention. The system includes: a memory 31 and a processor 32. The memory 31 contains a deep learning-based marine antibiotic pollution migration analysis method program. When the deep learning-based marine antibiotic pollution migration analysis method program is executed by the processor 32, the following steps are implemented:

[0112] Obtain historical remote sensing monitoring image information and perform image preprocessing, extract water body images of the target ocean area and perform time series processing to form a first data set;

[0113] Acquire the water flow characteristics of the target ocean area according to the first data set to construct a water flow characteristic map, rasterize the water flow characteristic map and perform interpolation processing, use the SARIMA model to analyze the water flow law of the target ocean area, and obtain water flow law analysis information;

[0114] Obtain the driving factors that affect the flow of ocean water bodies, use the Logistic regression model to analyze the driving force, calculate the influence weights of each driving factor based on the driving force analysis results, generate the state transition rules of the water flow state, and build a pollution migration prediction model;

[0115] Acquire water quality pollution detection information of each sampling point, perform pollution diffusion simulation based on the water flow law analysis information using finite element analysis to obtain pollution diffusion coefficient, and perform pollution migration prediction using a pollution migration prediction model to obtain pollution migration prediction information;

[0116] The antibiotic pollution status of the target marine area is analyzed based on the pollution migration prediction information, the area that needs antibiotic pollution control is analyzed, and an antibiotic pollution control plan is generated.

[0117] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0118] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0119] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0120] A person of ordinary skill in the art can understand that: all or part of the steps of implementing the above method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiment; and the aforementioned storage medium includes: a mobile storage device, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0121] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0122] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for analyzing the migration of marine antibiotic pollution based on deep learning, characterized in that: include: Obtain historical remote sensing monitoring image information and perform image preprocessing, extract water body images of the target ocean area and perform time series processing to form a first data set; Acquire the water flow characteristics of the target ocean area according to the first data set to construct a water flow characteristic map, rasterize the water flow characteristic map and perform interpolation processing, use the SARIMA model to analyze the water flow law of the target ocean area, and obtain water flow law analysis information; Obtain the driving factors that affect the flow of ocean water bodies, use the Logistic regression model to analyze the driving force, calculate the influence weights of each driving factor based on the driving force analysis results, generate the state transition rules of the water flow state, and build a pollution migration prediction model; Acquire water quality pollution detection information of each sampling point, perform pollution diffusion simulation based on the water flow law analysis information using finite element analysis to obtain pollution diffusion coefficient, and perform pollution migration prediction using a pollution migration prediction model to obtain pollution migration prediction information; Analyze the antibiotic pollution status of the target marine area according to the pollution migration prediction information, analyze the area that needs antibiotic pollution control and generate an antibiotic pollution control plan; The driving factors that affect the flow of ocean water bodies are obtained, the driving force analysis is performed using a Logistic regression model, the influence weights of each driving factor are calculated according to the driving force analysis results to generate state transition rules for the flow state of the water body, and a pollution migration prediction model is constructed, specifically including: Based on data retrieval, the driving factors of regional ocean water flow changes are obtained, the Pearson correlation coefficient between each driving factor and water flow changes is calculated, and the driving factors related to water flow are screened by judging with the preset threshold; According to the selected driving factors, the historical data corresponding to the driving factors are obtained, the water flow change is taken as the dependent variable, the driving factors are taken as the independent variables, and the driving force analysis is performed using the Logistic regression model to obtain the driving force analysis information; Acquire water flow law analysis information, and construct a state transfer matrix of water flow state in the target ocean area based on a Markov algorithm combined with the water flow law analysis information to form an initial state transfer rule; Calculate the influence weight of each driving factor according to the driving force analysis information, and correct the initial state transition rule by the calculated influence weight of each driving factor to obtain a corrected state transition rule; Building a pollution migration prediction model based on LSTM, defining pollution migration prediction rules using the corrected state transition rules, and building a training data set to train the pollution migration prediction model; The method of obtaining water pollution detection information of each sampling point, using finite element analysis method to perform pollution diffusion simulation based on the water flow law analysis information to obtain pollution diffusion coefficient, and using pollution migration prediction model to perform pollution migration prediction to obtain pollution migration prediction information specifically includes: Conduct water pollution detection in the target marine area to obtain water pollution detection information of each sampling point, compare the water pollution detection information with the preset threshold, evaluate the antibiotic pollution status of the area corresponding to each sampling point, and obtain antibiotic pollution status information; Acquire water flow law analysis information, and model the target ocean area using finite element analysis method based on the water flow law analysis information to form a regional water flow simulation model; Gridding the regional water flow simulation model, defining an initial contaminated area based on the antibiotic contamination status information, and setting regional water flow rules and boundary conditions according to the corrected state transition rules; Obtain real-time meteorological information and input it into the initialized regional water flow simulation model to perform pollution diffusion simulation, and calculate the pollution diffusion coefficient based on the simulation results; The human activity data of the target ocean area is obtained and the pollution migration driving weight is set in combination with the real-time meteorological information. The pollution diffusion coefficient is used to define the pollution migration change rate and the antibiotic pollution status information is input into the pollution migration prediction model to perform pollution migration prediction and obtain pollution migration prediction information.

2. A method for analyzing marine antibiotic pollution migration based on deep learning according to claim 1, characterized in that: The acquisition of historical remote sensing monitoring image information and image preprocessing, the extraction of target ocean area water body images and time series processing to form a first data set specifically include: Based on big data retrieval, historical remote sensing monitoring image information of the target ocean area is collected, and image preprocessing is performed on the historical remote sensing monitoring image information, including radiation correction, atmospheric correction and geometric correction, to obtain preprocessed image information; Calculate the NDWI index of the target ocean area according to the preprocessed image information, extract water body features according to the calculated NDWI index, and obtain a water body image of the target ocean area; A timestamp is preset, and based on the monitoring time characteristics corresponding to the water body image of the target ocean area, the multi-temporal image combination is performed on the water body image of the target ocean area according to the preset timestamp to construct a first data set.

3. The method for analyzing marine antibiotic pollution migration based on deep learning according to claim 1, characterized in that: The method of acquiring the water flow characteristics of the target ocean area according to the first data set and constructing a water flow characteristic map, rasterizing and interpolating the water flow characteristic map, and analyzing the water flow law of the target ocean area using the SARIMA model to obtain water flow law analysis information specifically includes: Acquire a first data set, extract water flow characteristics of a target ocean area according to the first data set, and generate a water flow characteristic map of the target ocean area based on the extracted water flow characteristics; Preset the grid unit size, rasterize the water flow characteristic map of the target ocean area to obtain a water flow characteristic grid map, and use the spline function interpolation method to interpolate the unknown characteristic grid; According to the coordinates of the unknown grid cell, the water flow characteristics of the neighboring grid cell are obtained, the interpolation interval of the unknown grid cell is formed, the spline equation of the interpolation interval is established and solved to obtain the spline coefficient; Interpolate the unknown characteristic grid according to the spline coefficient of the interpolation interval to generate an interpolated water flow characteristic grid map, preset a time period, and combine the interpolated water flow characteristic grid map to form a water flow characteristic time series; The constructed water flow characteristic time series is used as the model input, and the SARIMA model is used to analyze the water flow law in the target ocean area to obtain the water flow law analysis information.

4. The method for analyzing marine antibiotic pollution migration based on deep learning according to claim 1, characterized in that: The step of analyzing the antibiotic pollution status of the target marine area according to the pollution migration prediction information, analyzing the area that needs antibiotic pollution control and generating an antibiotic pollution control plan specifically includes: Generate a pollution migration path map of the target ocean area according to the pollution migration prediction information, calculate the migration pollution concentration and migration pollution time of the antibiotic pollution in the target ocean area based on the pollution migration path map combined with the pollution diffusion coefficient, and obtain first analysis information; Divide the target ocean area by pollution concentration according to the first analysis information, form an antibiotic pollution change map of the target ocean area in the future time period, and extract pollution concentration characteristics and regional location characteristics of different concentration areas; Based on the regional location characteristics of each pollution concentration, the ecological data and hydrological data of the corresponding area are obtained, and the environmental carrying capacity of each area at the current moment is evaluated in combination with the antibiotic pollution status information to obtain the second analysis information; The second analysis information is judged with the extracted pollution concentration characteristics of different concentration areas, and the marine area to be treated is defined according to the judgment result to obtain the information of the area to be treated; Based on the first analysis information, the time point when the antibiotic pollution in the area to be treated exceeds the environmental carrying capacity is calculated, the areas to be treated are sorted, and the treatment priority of each area to be treated is set according to the sorting result to obtain the third analysis information; According to the third analysis information, data retrieval is performed to obtain an antibiotic pollution control plan and provide antibiotic pollution control recommendations.

5. A deep learning-based marine antibiotic pollution migration analysis system, characterized in that: The system includes: a memory and a processor, wherein the memory contains a marine antibiotic pollution migration analysis method program based on deep learning, and when the marine antibiotic pollution migration analysis method program based on deep learning is executed by the processor, the following steps are implemented: Obtain historical remote sensing monitoring image information and perform image preprocessing, extract water body images of the target ocean area and perform time series processing to form a first data set; Acquire the water flow characteristics of the target ocean area according to the first data set to construct a water flow characteristic map, rasterize the water flow characteristic map and perform interpolation processing, use the SARIMA model to analyze the water flow law of the target ocean area, and obtain water flow law analysis information; Obtain the driving factors that affect the flow of ocean water bodies, use the Logistic regression model to analyze the driving force, calculate the influence weights of each driving factor based on the driving force analysis results, generate the state transition rules of the water flow state, and build a pollution migration prediction model; Acquire water quality pollution detection information of each sampling point, perform pollution diffusion simulation based on the water flow law analysis information using finite element analysis to obtain pollution diffusion coefficient, and perform pollution migration prediction using a pollution migration prediction model to obtain pollution migration prediction information; Analyze the antibiotic pollution status of the target marine area according to the pollution migration prediction information, analyze the area that needs antibiotic pollution control and generate an antibiotic pollution control plan; The driving factors that affect the flow of ocean water bodies are obtained, the driving force analysis is performed using a Logistic regression model, the influence weights of each driving factor are calculated according to the driving force analysis results to generate state transition rules for the flow state of the water body, and a pollution migration prediction model is constructed, specifically including: Based on data retrieval, the driving factors of regional ocean water flow changes are obtained, the Pearson correlation coefficient between each driving factor and water flow changes is calculated, and the driving factors related to water flow are screened by judging with the preset threshold; According to the selected driving factors, the historical data corresponding to the driving factors are obtained, the water flow change is taken as the dependent variable, the driving factors are taken as the independent variables, and the driving force analysis is performed using the Logistic regression model to obtain the driving force analysis information; Acquire water flow law analysis information, and construct a state transfer matrix of water flow state in the target ocean area based on a Markov algorithm combined with the water flow law analysis information to form an initial state transfer rule; Calculate the influence weight of each driving factor according to the driving force analysis information, and correct the initial state transition rule by the calculated influence weight of each driving factor to obtain a corrected state transition rule; Building a pollution migration prediction model based on LSTM, defining pollution migration prediction rules using the corrected state transition rules, and building a training data set to train the pollution migration prediction model; The method of obtaining water pollution detection information of each sampling point, using finite element analysis method to perform pollution diffusion simulation based on the water flow law analysis information to obtain pollution diffusion coefficient, and using pollution migration prediction model to perform pollution migration prediction to obtain pollution migration prediction information specifically includes: Conduct water pollution detection in the target marine area to obtain water pollution detection information of each sampling point, compare the water pollution detection information with the preset threshold, evaluate the antibiotic pollution status of the area corresponding to each sampling point, and obtain antibiotic pollution status information; Acquire water flow law analysis information, and model the target ocean area using finite element analysis method based on the water flow law analysis information to form a regional water flow simulation model; Gridding the regional water flow simulation model, defining an initial contaminated area based on the antibiotic contamination status information, and setting regional water flow rules and boundary conditions according to the corrected state transition rules; Obtain real-time meteorological information and input it into the initialized regional water flow simulation model to perform pollution diffusion simulation, and calculate the pollution diffusion coefficient based on the simulation results; The human activity data of the target ocean area is obtained and the pollution migration driving weight is set in combination with the real-time meteorological information. The pollution diffusion coefficient is used to define the pollution migration change rate and the antibiotic pollution status information is input into the pollution migration prediction model to perform pollution migration prediction and obtain pollution migration prediction information.

Images