Bay area ecological quality assessment and early warning method based on machine learning

Abstract

The invention discloses a bay area ecological quality assessment and early warning method based on machine learning, and relates to the technical field of bay ecological assessment, and the method comprises the steps: collecting data related to the ecological quality of a target bay area based on multiple data sources, including a water quality parameter, an air quality parameter and an environment pressure parameter; feature analysis is conducted on the data related to the ecological quality of the bay area, the ecological quality change trend of the bay area is determined, and the ecological quality change trend comprises the water quality deviation trend, the air quality deviation trend and the environment pressure negative trend. By collecting and processing water quality, air quality and environment pressure data and using the trained abnormal ecological quality identification model for dynamic evaluation, real-time monitoring of the ecological quality of a bay area can be realized, and once an abnormal condition is detected, an early warning notification can be immediately sent out to help a manager to take action quickly, so that the management efficiency is improved. And further deterioration of the ecological system is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of bay ecological assessment, and in particular to a bay area ecological quality assessment and early warning method based on machine learning. Background Art

[0002] The Gulf region is an important part of the marine ecosystem, with rich biodiversity and important ecological functions. With the continuous increase of human activities, the Gulf ecosystem faces serious threats, such as pollution, overfishing, habitat destruction, etc. These threats not only affect the ecological balance of the Gulf region, but may also have a negative impact on the economic development and lives of residents in the surrounding areas. Therefore, it is particularly important to evaluate and warn of the ecological quality of the Gulf region in order to carry out targeted protection and management.

[0003] In the existing technology, the ecological quality of the bay area is affected by many factors, which affects the accuracy of the ecological quality assessment and early warning of the bay area. Therefore, how to capture the dynamic changes of the ecosystem in the bay area and improve the accuracy of the ecological quality assessment and early warning in order to match the corresponding governance measures is the problem we need to solve. To this end, a method for ecological quality assessment and early warning of the bay area based on machine learning is proposed. Summary of the invention

[0004] The purpose of the present invention is to provide a bay area ecological quality assessment and early warning method based on machine learning to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is: A method for early warning of ecological quality assessment in the bay area based on machine learning includes the following steps: S1. Collect data related to the ecological quality of the target bay area based on multiple data sources, including water quality parameters, air quality parameters, and environmental pressure parameters; S2. Analyze the characteristics of data related to the ecological quality of the Bay Area and clarify the changing trend of the ecological quality of the Bay Area, including the deviation trend of water quality, the deviation trend of air quality and the negative trend of environmental pressure; S3. Use historical data of known ecological quality status to train an abnormal ecological quality identification model to identify abnormal conditions in the ecological quality of the target bay area; S4. Combine the identification results of the abnormal ecological quality identification model with the trend of ecological quality changes to conduct a dynamic assessment of the ecological quality of the bay area; S5. Set ecological quality thresholds, conduct early warning analysis on dynamic assessment results, and implement corresponding governance measures based on the early warning results.

[0006] A further improvement of the technical solution of the present invention is that in S1, the process of collecting ecological quality data of the target bay area is: S11. Select and determine the data sources for ecological quality assessment in the target bay area, which are water quality monitoring stations, air quality monitoring stations and environmental pressure data sources. Among them, select and determine multiple water quality monitoring stations to cover different areas of the target bay, including key locations such as nearshore, deep sea, and estuary. Determine multiple air quality monitoring stations to reflect the air quality conditions around the target bay. Use satellite remote sensing technology to monitor environmental information in a large geographical area, and combine pollutant emission reports and meteorological stations to comprehensively obtain environmental pressure data; S12. According to the characteristics of the target bay area, clarify the data types of the bay area ecological quality assessment, including water quality parameters, air quality parameters and environmental pressure parameters; S13. For water quality parameter data, use automatic water quality monitoring instruments to regularly collect data on pH, suspended particulate matter, dissolved oxygen, chemical oxygen demand, and total nitrogen and phosphorus content; for air quality parameter data, install sensors at selected air quality monitoring sites to monitor and obtain data on sulfur dioxide concentration, nitrogen oxide concentration, and inhalable particulate matter (PM2.5 and PM10) concentration; for environmental pressure parameters, obtain data on pollutant emissions and temperature change rates based on satellite remote sensing monitoring, pollutant emission reports, and meteorological stations; S14. Build a data warehouse, perform data cleaning and data standardization preprocessing operations on the acquired data on water quality parameters, air quality parameters and environmental pressure parameters, preprocess the original data to remove noise, fill in missing values, correct erroneous data, etc. Through data standardization, unify the format of data from different sources to facilitate subsequent analysis, and integrate data from different data sources into the data warehouse for storage.

[0007] A further improvement of the technical solution of the present invention is that in S2, the analysis process of the trend of ecological quality changes in the bay area is: S21, traversing the data warehouse to retrieve pre-processed water quality parameter, air quality parameter and environmental pressure parameter data, and performing feature analysis on each parameter data to extract indicators reflecting the ecological quality of the bay area; S22. Based on the pre-treated water quality data, extract the evaluation indicators of pH, suspended particles, dissolved oxygen, chemical oxygen demand and total nitrogen and phosphorus content, and determine the baseline value of each water quality evaluation indicator in combination with the ecological quality requirements and historical database of the target area, and then calculate the water quality deviation index, and analyze the change trend of water quality parameters over time during the evaluation period; S23. Based on the pre-processed air quality data, extract the evaluation indicators of sulfur dioxide concentration, nitrogen oxide concentration and inhalable particulate matter (PM2.5 and PM10) concentration, and determine the baseline value of each air quality evaluation indicator in combination with the ecological quality requirements and historical database of the target area, and then calculate the air quality deviation index, and analyze the change trend of air quality parameters over time during the evaluation period; S24. Based on the pre-processed environmental pressure parameter data, the evaluation indicators of pollutant emissions and temperature change rate are extracted, and the baseline value of each environmental pressure evaluation indicator is determined in combination with the ecological quality requirements and historical database of the target area, and then the environmental pressure trend index is calculated to analyze the negative trend of environmental pressure.

[0008] A further improvement of the technical solution of the present invention is that the calculation formula of the water quality deviation index is as follows: ; Where W is the water quality deviation index, is the actual measured value of the i-th water quality parameter evaluation index, is the benchmark value of the i-th water quality parameter assessment indicator, which is the standard value determined based on the ecological quality requirements of the target area and the historical database. is the weight of the ith water quality parameter evaluation index, which is used to reflect its importance in the comprehensive evaluation. p is the adjustment factor, which is used to control the sensitivity of the exponential function and is set to 2 by default. n is the number of water quality parameter evaluation indicators. i is the index of the water quality parameter evaluation indicator, i=1,2,3,4,5, corresponding to pH, suspended particulate matter, dissolved oxygen, chemical oxygen demand and total nitrogen and phosphorus content, respectively. The value range of W is between 0 and 1. When all water quality parameter evaluation indicators are completely consistent with the benchmark value, W is 0. When one or more water quality parameter evaluation indicators deviate from the benchmark value, W is closer to 1. When W approaches 0, it means that the water quality parameters are close to the benchmark value and the water quality is good. When W approaches 1, it means that the water quality parameters deviate seriously from the benchmark value and the water quality is poor. The calculation formula of the air quality deviation index is as follows: ; Where R is the air quality deviation index, is the actual measured value of the j-th air quality parameter evaluation index, is the benchmark value of the jth air quality parameter evaluation index, which is the standard value determined based on the ecological quality requirements of the target area and the historical database. is the weight of the jth air quality parameter evaluation index, which is used to reflect its importance in the comprehensive evaluation. m is the number of air quality parameter evaluation indicators. j is the index of the air quality parameter evaluation indicator. j=1,2,3, corresponding to the sulfur dioxide concentration, nitrogen oxide concentration and inhalable particulate matter concentration, respectively. The value range of R is between 0 and 1. When all air quality parameter evaluation indicators are completely consistent with the benchmark value, R is 0. When one or more air quality parameter evaluation indicators deviate from the benchmark value, R is closer to 1. When R approaches 0, it means that the air quality parameter is close to the benchmark value and the air quality is good. When R approaches 1, it means that the air quality parameter deviates seriously from the benchmark value and the air quality is poor. The calculation formula of the environmental pressure trend index is as follows: ; Where E is the environmental pressure trend index, is the actual measured value of the kth environmental pressure parameter evaluation index, is the benchmark value of the kth environmental pressure parameter assessment indicator, which is the standard value determined based on the ecological quality requirements of the target area and the historical database. is the weight of the kth environmental pressure parameter evaluation indicator, which is used to reflect its importance in the comprehensive evaluation. h is the number of environmental pressure parameter evaluation indicators. k is the index of the environmental pressure parameter evaluation indicator. k=1,2, corresponding to pollutant emissions and temperature change rate respectively. The value range of E is between 0 and 1. When all environmental pressure parameter evaluation indicators are completely consistent with the benchmark values, E is 0. When one or more environmental pressure parameter evaluation indicators deviate from the benchmark value, E is closer to 1. When E approaches 0, it means that the environmental pressure parameter is close to the benchmark value and the environmental condition is good. When E approaches 1, it means that the environmental pressure parameter deviates seriously from the benchmark value and the environmental condition is poor.

[0009] A further improvement of the technical solution of the present invention is that in S3, the abnormal situation identification process is: S31. Collect historical data containing known ecological quality status in the target bay area, and mark abnormal water quality parameters, abnormal air quality parameters, and abnormal environmental pressure parameters in the historical data; S32, integrating various evaluation indicators of water quality parameters, air quality parameters and environmental pressure parameters to obtain a comprehensive data set, and dividing it into a training set and a test set, ensuring that each sample contains all relevant parameters and their corresponding normal or abnormal labels; S33. Based on the evaluation index data of the training set combined with the neural network model, the water quality parameters, air quality parameters and environmental pressure parameters in the training set are used as input features, and the normal or abnormal labels are used as outputs. The abnormal ecological quality recognition model is trained, and the model parameters are adjusted to enable it to recognize abnormal ecological quality. The performance of the model is evaluated by calculating the accuracy, recall rate, F1-score, AUC and other indicators in combination with the test set data, and then the model is optimized according to the evaluation results; S34. Apply the trained model to the real-time ecological quality data of the target bay area, calculate the anomaly recognition index based on the probability of the model output, and set the anomaly threshold Y to determine whether there is an anomaly.

[0010] A further improvement of the technical solution of the present invention is that the calculation formula of the abnormality identification index is as follows: ; Where T is the anomaly recognition index, is the actual measured value of the tth real-time ecological quality parameter, is the weight of the tth parameter, which is used to reflect its importance in the comprehensive assessment. N is the number of ecological quality parameters, that is, the sum of the evaluation indicators of water quality parameters, air quality parameters and environmental pressure parameters. is an adjustment factor used to control the influence of the parameter average on the threshold. It is an adjustment factor used to introduce randomness so that the threshold has a certain fluctuation range. Represents a randomly generated number in the interval [0,1]. If , it is considered to be abnormal.

[0011] A further improvement of the technical solution of the present invention is that in S4, the process of dynamic evaluation of ecological quality of the bay area is: S41, inputting the parameter data of water quality, air quality and environmental pressure of the target area into the trained abnormal ecological quality recognition model, performing abnormal detection on the target area, the model outputs the probability of normal or abnormal, and obtains the recognition result of the abnormal state according to the set abnormal threshold; S42. Calculate the water quality deviation index, air quality deviation index and environmental pressure trend index of the target area, analyze the changing trend of each parameter over time, and assign corresponding weights according to the importance of each parameter in the ecosystem; S43. Based on the weights assigned to the various parameters, combined with the calculation results of the water quality deviation index, air quality deviation index and environmental pressure trend index of the target area, the ecological quality assessment coefficient is comprehensively calculated, and according to the size of the comprehensive ecological quality assessment coefficient, the ecological quality status of the target bay area is analyzed.

[0012] A further improvement of the technical solution of the present invention is that the calculation formula of the ecological quality assessment coefficient is as follows: ; In the formula, SC is the ecological quality assessment coefficient, W is the water quality deviation index, R is the air quality deviation index, and E is the environmental pressure trend index. is the weight of the water quality deviation index, is the weight of the air quality deviation index, is the weight of the environmental pressure trend index. It should be noted that SC approaching 1 indicates that the water quality, air quality and environmental pressure parameters are close to the benchmark values, and the ecological quality is in good condition.

[0013] A further improvement of the technical solution of the present invention is that in S5, the process of early warning analysis is: S51. Based on the requirements of the ecological quality of the bay area and the historical ecological quality data of the bay area, the ecological quality of the bay area is graded into high-quality ecological quality grade, good ecological quality grade and poor ecological quality grade; S52. Combine the calculation results of the ecological quality assessment coefficients, match them with each ecological quality level, and set corresponding ecological quality thresholds for each ecological quality level; S53. Based on the value of the ecological quality assessment coefficient, determine the current ecological quality level and mark it with different status colors, mark the high-quality ecological quality level with a green status color, mark the good ecological quality level with a yellow status color, and mark the poor ecological quality level with a red status color; S54. When the ecological quality level is good, an early warning notice is issued, suggesting preventive measures. When the ecological quality level is poor, an emergency early warning notice is issued, the emergency response mechanism is immediately activated, and the reasons for the early warning are analyzed, including water quality deterioration, air pollution aggravation, and environmental pressure increase; S55. Based on the results of early warning analysis, formulate targeted governance plans and implement corresponding governance measures to strengthen water quality monitoring and governance, improve air quality, and reduce environmental pressure. At the same time, supervise the implementation of governance measures to ensure that the measures are effectively implemented, regularly evaluate the governance effects, and adjust the governance plans based on the evaluation results.

[0014] A further improvement of the technical solution of the present invention is that: the multiple ecological quality levels correspond to multiple ecological quality thresholds, wherein the ecological quality thresholds include an upper threshold and a lower threshold; The multiple ecological quality levels and the multiple ecological quality thresholds satisfy the following relationship: The ecological quality thresholds of the high-quality ecological quality level are: ; The ecological quality thresholds of the good ecological quality level are: ; The ecological quality thresholds of the poor ecological quality level are: ; Among them, SC is the ecological quality assessment coefficient, is the lower threshold corresponding to the high-quality ecological quality level and the upper threshold corresponding to the good ecological quality level, is the lower threshold corresponding to the good ecological quality level and the upper threshold corresponding to the poor ecological quality level, , .

[0015] Due to the adoption of the above technical solution, the present invention has the following technical advances compared with the prior art: The present invention provides a method for early warning of ecological quality assessment in a bay area based on machine learning. By collecting and processing water quality, air quality and environmental pressure data and using a trained abnormal ecological quality recognition model for dynamic assessment, it is possible to achieve real-time monitoring of the ecological quality of the bay area. Once an abnormal situation is detected, an early warning notification can be issued immediately to help managers take prompt action to avoid further deterioration of the ecosystem.

[0016] The present invention provides a bay area ecological quality assessment and early warning method based on machine learning. By integrating the water quality deviation index, air quality deviation index and environmental pressure trend index, the ecological quality of the bay area is comprehensively evaluated from multiple dimensions, which more accurately reflects the overall ecological status and avoids the deviation caused by a single indicator. Moreover, according to the early warning analysis results, targeted governance plans can be formulated to ensure the long-term health and sustainable development of the ecosystem. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0018] Figure 1 is a flow chart of the method of the present invention; Figure 2 The flowchart of the analysis of the changing trend of ecological quality in the bay area of ​​the present invention; Figure 3 This is a flow chart of the dynamic assessment of the ecological quality of the bay area of ​​the present invention. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] Embodiment 1, as Figure 1 , Figure 2 As shown, the present invention provides a method for early warning of ecological quality assessment in a bay area based on machine learning, comprising the following steps: S1. Collect data related to the ecological quality of the target bay area based on multiple data sources, including water quality parameters, air quality parameters and environmental pressure parameters. Water quality parameters include pH, suspended particulate matter, dissolved oxygen, chemical oxygen demand and total nitrogen and phosphorus content. Air quality parameters include sulfur dioxide concentration, nitrogen oxide concentration and inhalable particulate matter (PM2.5 and PM10) concentration. Environmental pressure parameters include pollutant emissions and temperature change rate. Select and determine the data sources for ecological quality assessment of the target bay area, which are water quality monitoring stations, air quality monitoring stations and environmental pressure data sources. Select and determine multiple water quality monitoring stations. Sites, covering different areas of the target bay, including nearshore, deep sea, estuary and other key locations, determine multiple air quality monitoring sites to reflect the air quality conditions around the target bay, use satellite remote sensing technology to monitor environmental information in a large geographical area, and combine pollutant emission reports and meteorological stations to comprehensively obtain environmental pressure data. According to the characteristics of the target bay area, clarify the data types of ecological quality assessment in the bay area, including water quality parameters, air quality parameters and environmental pressure parameters. For water quality parameter data, use automatic water quality monitoring instruments to regularly collect data on pH, suspended particulate matter, dissolved oxygen, chemical oxygen demand and total nitrogen and phosphorus content; For air quality parameter data, sensors are installed at selected air quality monitoring sites to monitor and obtain data on sulfur dioxide concentration, nitrogen oxide concentration and inhalable particulate matter (PM2.5 and PM10) concentration; for environmental pressure parameters, pollutant emission and temperature change rate data are obtained based on satellite remote sensing monitoring, pollutant emission reports and meteorological stations. Among them, pH value reflects the acid-base balance of the water body, suspended particulate matter indicates the turbidity of the water body, dissolved oxygen reflects the self-purification capacity and biological activity of the water body, chemical oxygen demand measures the degree of organic pollution in the water body, total nitrogen and phosphorus content indicates the eutrophication risk of the water body, and sulfur dioxide concentration is One of the harmful gases, nitrogen oxide concentration includes NO and NO2, inhalable particulate matter includes PM2.5 and PM10, pollutant emissions measure the pressure of human activities on the environment, and the temperature change rate reflects the impact of climate change on the ecosystem. Build a data warehouse, perform data cleaning and data standardization preprocessing operations on the acquired water quality parameters, air quality parameters and environmental pressure parameters, preprocess the original data, remove noise, fill in missing values, correct erroneous data, etc. Through data standardization, the data from different sources are unified in format for subsequent analysis, and the data from different data sources are integrated into the data warehouse for storage; S2. Conduct feature analysis on data related to the ecological quality of the bay area to clarify the changing trend of the ecological quality of the bay area, where the changing trend of ecological quality includes the deviation trend of water quality, the deviation trend of air quality and the negative trend of environmental pressure. Traverse the data warehouse to retrieve the pre-processed water quality parameter, air quality parameter and environmental pressure parameter data, and conduct feature analysis on each parameter data to extract indicators reflecting the ecological quality of the bay area. Based on the pre-processed water quality data, extract the evaluation indicators of pH, suspended particulate matter, dissolved oxygen, chemical oxygen demand and total nitrogen and phosphorus content, and combine the ecological quality requirements and historical database of the target area to determine the baseline value of each water quality evaluation indicator, and then calculate the water quality deviation index, and analyze the evaluation results. The changing trend of water quality parameters over time during the period, based on the pre-treated air quality data, the evaluation indicators of sulfur dioxide concentration, nitrogen oxide concentration and inhalable particulate matter (PM2.5 and PM10) concentration are extracted, and the benchmark values ​​of each air quality evaluation indicator are determined in combination with the ecological quality requirements and historical database of the target area, and then the air quality deviation index is calculated, and the changing trend of air quality parameters over time during the evaluation period is analyzed. Based on the pre-treated environmental pressure parameter data, the evaluation indicators of pollutant emissions and temperature change rate are extracted, and the benchmark values ​​of each environmental pressure evaluation indicator are determined in combination with the ecological quality requirements and historical database of the target area, and then the environmental pressure trend index is calculated to analyze the negative trend of environmental pressure; Furthermore, the calculation formula of the water quality deviation index is as follows: ; Where W is the water quality deviation index, is the actual measured value of the i-th water quality parameter evaluation index, is the benchmark value of the i-th water quality parameter assessment indicator, which is the standard value determined based on the ecological quality requirements of the target area and the historical database. is the weight of the ith water quality parameter evaluation index, which is used to reflect its importance in the comprehensive evaluation. p is the adjustment factor, which is used to control the sensitivity of the exponential function and is set to 2 by default. n is the number of water quality parameter evaluation indicators. i is the index of the water quality parameter evaluation indicator, i=1,2,3,4,5, corresponding to pH, suspended particulate matter, dissolved oxygen, chemical oxygen demand and total nitrogen and phosphorus content, respectively. The value range of W is between 0 and 1. When all water quality parameter evaluation indicators are completely consistent with the benchmark value, W is 0. When one or more water quality parameter evaluation indicators deviate from the benchmark value, W is closer to 1. When W approaches 0, it means that the water quality parameters are close to the benchmark value and the water quality is good. When W approaches 1, it means that the water quality parameters deviate seriously from the benchmark value and the water quality is poor. The calculation formula of air quality deviation index is as follows: ; Where R is the air quality deviation index, is the actual measured value of the j-th air quality parameter evaluation index, is the benchmark value of the jth air quality parameter evaluation index, which is the standard value determined based on the ecological quality requirements of the target area and the historical database. is the weight of the jth air quality parameter evaluation index, which is used to reflect its importance in the comprehensive evaluation. m is the number of air quality parameter evaluation indicators. j is the index of the air quality parameter evaluation indicator. j=1,2,3, corresponding to the sulfur dioxide concentration, nitrogen oxide concentration and inhalable particulate matter concentration, respectively. The value range of R is between 0 and 1. When all air quality parameter evaluation indicators are completely consistent with the benchmark value, R is 0. When one or more air quality parameter evaluation indicators deviate from the benchmark value, R is closer to 1. When R approaches 0, it means that the air quality parameter is close to the benchmark value and the air quality is good. When R approaches 1, it means that the air quality parameter deviates seriously from the benchmark value and the air quality is poor. The calculation formula of the environmental pressure trend index is as follows: ; Where E is the environmental pressure trend index, is the actual measured value of the kth environmental pressure parameter evaluation index, is the benchmark value of the kth environmental pressure parameter assessment indicator, which is the standard value determined based on the ecological quality requirements of the target area and the historical database. is the weight of the kth environmental pressure parameter evaluation indicator, which is used to reflect its importance in the comprehensive evaluation. h is the number of environmental pressure parameter evaluation indicators. k is the index of environmental pressure parameter evaluation indicators. k=1,2, corresponding to pollutant emissions and temperature change rate, respectively. The value range of E is between 0 and 1. When all environmental pressure parameter evaluation indicators are completely consistent with the benchmark value, E is 0. When one or more environmental pressure parameter evaluation indicators deviate from the benchmark value, E is closer to 1. When E approaches 0, it means that the environmental pressure parameter is close to the benchmark value and the environmental condition is good. When E approaches 1, it means that the environmental pressure parameter deviates seriously from the benchmark value and the environmental condition is poor. S3. Use historical data of known ecological quality status to train an abnormal ecological quality recognition model to identify abnormalities in the ecological quality of the target bay area. Collect historical data of the target bay area containing known ecological quality status, mark abnormal water quality parameters, abnormal air quality parameters and abnormal environmental pressure parameters in the historical data, integrate various evaluation indicators of water quality parameters, air quality parameters and environmental pressure parameters to obtain a comprehensive data set, and divide it into a training set and a test set to ensure that each sample contains all relevant parameters and their corresponding normal or abnormal labels. Based on the evaluation indicator data of the training set combined with the neural network model, the water quality parameters, air quality parameters and environmental pressure parameters in the training set are used as input features and normal or abnormal labels are used as outputs to train the abnormal ecological quality recognition model, adjust the model parameters to enable it to recognize abnormalities in ecological quality, and evaluate the performance of the model by calculating accuracy, recall rate, F1-score, AUC and other indicators in combination with the test set data, and then optimize the model according to the evaluation results. Apply the trained model to the real-time ecological quality data of the target bay area, calculate the anomaly recognition index according to the probability of the model output, and set the anomaly threshold Y to determine whether there is an anomaly; Furthermore, the calculation formula of the anomaly recognition index is as follows: ; Where T is the anomaly recognition index, is the actual measured value of the tth real-time ecological quality parameter, is the weight of the tth parameter, which is used to reflect its importance in the comprehensive assessment. N is the number of ecological quality parameters, that is, the sum of the evaluation indicators of water quality parameters, air quality parameters and environmental pressure parameters. is an adjustment factor used to control the influence of the parameter average on the threshold. It is an adjustment factor used to introduce randomness so that the threshold has a certain fluctuation range. Represents a randomly generated number in the interval [0,1], used to introduce randomness. When the parameter value deviates significantly from the reference value, When the parameter value is very close to the reference value, the threshold value will be close to 0.5 or higher, making the standard for abnormality more stringent. , it is considered to be abnormal; S4. Combine the identification results of the abnormal ecological quality identification model with the trend of ecological quality changes to conduct a dynamic assessment of the ecological quality of the bay area; S5. Set ecological quality thresholds, conduct early warning analysis on dynamic assessment results, and implement corresponding governance measures based on the early warning results.

[0021] Embodiment 2, as Figure 3As shown, based on Example 1, the present invention provides a technical solution: Preferably, in S4, the process of dynamic evaluation of ecological quality of the bay area is: Input the parameter data of water quality, air quality and environmental pressure of the target area into the trained abnormal ecological quality identification model, perform abnormal detection on the target area, and the model outputs the probability of normal or abnormal. According to the set abnormal threshold, the identification result of the abnormal state is obtained, and the water quality deviation index, air quality deviation index and environmental pressure trend index of the target area are calculated. The change trend of each parameter over time is analyzed, and the corresponding weight is assigned according to the importance of each parameter in the ecosystem. Based on the weight assigned to each parameter and the calculation results of the water quality deviation index, air quality deviation index and environmental pressure trend index of the target area, the ecological quality assessment coefficient is comprehensively calculated, and the ecological quality status of the target bay area is analyzed according to the size of the comprehensive ecological quality assessment coefficient. Furthermore, the calculation formula of the ecological quality assessment coefficient is as follows: ; In the formula, SC is the ecological quality assessment coefficient, W is the water quality deviation index, R is the air quality deviation index, and E is the environmental pressure trend index. is the weight of the water quality deviation index, is the weight of the air quality deviation index, is the weight of the environmental pressure trend index. It should be noted that when all deviation indices are 0, SC is 1. When one or more deviation indices increase, SC approaches 0. SC approaches 1, indicating that water quality, air quality and environmental pressure parameters are close to the benchmark values, and the ecological quality is good. SC approaches 0, indicating that some or all parameters deviate seriously from the benchmark values, and the ecological quality is poor. In S5, the process of early warning analysis is: According to the requirements of the ecological quality of the bay area and the historical ecological quality data of the bay area, the ecological quality of the bay area is graded into high-quality ecological quality grade, good ecological quality grade and poor ecological quality grade. Combined with the calculation results of the ecological quality assessment coefficient, it is matched with each ecological quality grade, and the corresponding ecological quality threshold is set for each ecological quality grade. Based on the value of the ecological quality assessment coefficient, the current ecological quality grade is judged and marked with different status colors. The high-quality ecological quality grade is marked as a green status color, the good ecological quality grade is marked as a yellow status color, and the poor ecological quality grade is marked as a red status color. When it is in the good ecological quality grade, an early warning notice is issued and preventive measures are recommended. When it is in the poor ecological quality grade, an emergency early warning notice is issued, the emergency response mechanism is immediately activated, and the reasons for the early warning are analyzed, including water quality deterioration, increased air pollution and increased environmental pressure. According to the early warning analysis results, a targeted governance plan is formulated and corresponding governance measures are implemented to strengthen water quality monitoring and governance, improve air quality and reduce environmental pressure. At the same time, the implementation of governance measures is supervised to ensure that the measures are effectively implemented, the governance effect is regularly evaluated, and the governance plan is adjusted according to the evaluation results; Furthermore, the multiple ecological quality levels correspond to multiple ecological quality thresholds, wherein the ecological quality thresholds include an upper threshold and a lower threshold; Multiple ecological quality levels and multiple ecological quality thresholds satisfy the following relationship: The ecological quality thresholds for the high-quality ecological quality level are: ; Indicates that the ecological quality is excellent, the ecosystem functions are strong, the biodiversity is rich, the ecosystem is in good health, and no immediate emergency measures are needed; The ecological quality thresholds for good ecological quality levels are: ; Indicates that the ecological quality is good, the ecosystem functions are basically stable, and the biodiversity is relatively rich, but there are certain problems with the health of the ecosystem, which requires close attention and preventive measures; The ecological quality threshold of poor ecological quality level is: ; Indicates that the ecological quality is poor, the ecosystem function is damaged, the biodiversity is reduced, the ecosystem health is poor, and urgent governance measures need to be taken immediately; Among them, SC is the ecological quality assessment coefficient, is the lower threshold corresponding to the high-quality ecological quality level and the upper threshold corresponding to the good ecological quality level, is the lower threshold corresponding to the good ecological quality level and the upper threshold corresponding to the poor ecological quality level, , .

[0022] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for early warning of ecological quality assessment in the bay area based on machine learning, characterized in that: The following steps are involved: S1. Collect data related to the ecological quality of the target bay area based on multiple data sources, including water quality parameters, air quality parameters, and environmental pressure parameters; S2. Analyze the characteristics of data related to the ecological quality of the Bay Area and clarify the changing trend of the ecological quality of the Bay Area, including the deviation trend of water quality, the deviation trend of air quality and the negative trend of environmental pressure; S3. Use historical data of known ecological quality status to train an abnormal ecological quality identification model to identify abnormal conditions in the ecological quality of the target bay area; S4. Combine the identification results of the abnormal ecological quality identification model with the trend of ecological quality changes to conduct a dynamic assessment of the ecological quality of the bay area; S5. Set ecological quality thresholds, conduct early warning analysis on dynamic assessment results, and implement corresponding governance measures based on the early warning results.

2. The method for early warning of ecological quality assessment in the bay area based on machine learning according to claim 1 is characterized by: In S1, the process of collecting ecological quality data of the target bay area is as follows: S11. Select and determine the data sources for ecological quality assessment in the target bay area, which are water quality monitoring stations, air quality monitoring stations and environmental pressure data sources; S12. According to the characteristics of the target bay area, clarify the data types of the bay area ecological quality assessment, including water quality parameters, air quality parameters and environmental pressure parameters; S13. For water quality parameter data, use automatic water quality monitoring instruments to regularly collect data on pH, suspended particulate matter, dissolved oxygen, chemical oxygen demand, and total nitrogen and phosphorus content; for air quality parameter data, install sensors at selected air quality monitoring sites to monitor and obtain data on sulfur dioxide concentration, nitrogen oxide concentration, and inhalable particulate matter concentration; for environmental pressure parameters, obtain data on pollutant emissions and temperature change rates based on satellite remote sensing monitoring, pollutant emission reports, and meteorological stations; S14. Build a data warehouse, perform preprocessing operations such as data cleaning and data standardization on the acquired data related to water quality parameters, air quality parameters and environmental pressure parameters, and integrate data from different data sources into the data warehouse for storage.

3. The method for early warning of ecological quality assessment in the bay area based on machine learning according to claim 2 is characterized by: In S2, the analysis process of the trend of ecological quality change in the bay area is as follows: S21, traversing the data warehouse to retrieve pre-processed water quality parameter, air quality parameter and environmental pressure parameter data, and performing feature analysis on each parameter data to extract indicators reflecting the ecological quality of the bay area; S22. Based on the pre-treated water quality data, extract the evaluation indicators of pH, suspended particles, dissolved oxygen, chemical oxygen demand and total nitrogen and phosphorus content, and determine the baseline value of each water quality evaluation indicator in combination with the ecological quality requirements and historical database of the target area, and then calculate the water quality deviation index, and analyze the change trend of water quality parameters over time during the evaluation period; S23. Based on the pre-processed air quality data, the evaluation indicators of sulfur dioxide concentration, nitrogen oxide concentration and inhalable particulate matter concentration are extracted, and the baseline value of each air quality evaluation indicator is determined in combination with the ecological quality requirements and historical database of the target area, and then the air quality deviation index is calculated, and the change trend of air quality parameters over time during the evaluation period is analyzed; S24. Based on the pre-processed environmental pressure parameter data, the evaluation indicators of pollutant emissions and temperature change rate are extracted, and the baseline value of each environmental pressure evaluation indicator is determined in combination with the ecological quality requirements and historical database of the target area, and then the environmental pressure trend index is calculated to analyze the negative trend of environmental pressure.

4. The method for early warning of ecological quality assessment in the bay area based on machine learning according to claim 3 is characterized by: The calculation formula of the water quality deviation index is as follows: ； Where W is the water quality deviation index, is the actual measured value of the i-th water quality parameter evaluation index, is the benchmark value of the i-th water quality parameter evaluation index, is the weight of the i-th water quality parameter evaluation index, p is the adjustment factor, n is the number of water quality parameter evaluation indicators, i is the index of the water quality parameter evaluation index, i=1,2,3,4,5, corresponding to pH, suspended particles, dissolved oxygen, chemical oxygen demand and total nitrogen and phosphorus content, respectively. The value range of W is between 0 and 1. W approaches 0, indicating that the water quality parameter is close to the benchmark value and the water quality is good. W approaches 1, indicating that the water quality parameter deviates seriously from the benchmark value and the water quality is poor. The calculation formula of the air quality deviation index is as follows: ； Where R is the air quality deviation index, is the actual measured value of the j-th air quality parameter evaluation index, is the benchmark value of the j-th air quality parameter evaluation index, is the weight of the jth air quality parameter evaluation index, m is the number of air quality parameter evaluation indicators, j is the index of the air quality parameter evaluation index, j=1,2,3, corresponding to sulfur dioxide concentration, nitrogen oxide concentration and inhalable particulate matter concentration respectively, R ranges from 0 to 1, R approaches 0, indicating that the air quality parameter is close to the benchmark value and the air quality is good, and R approaches 1, indicating that the air quality parameter deviates seriously from the benchmark value and the air quality is poor; The calculation formula of the environmental pressure trend index is as follows: ； Where E is the environmental pressure trend index, is the actual measured value of the kth environmental pressure parameter evaluation index, is the benchmark value of the kth environmental pressure parameter evaluation index, is the weight of the kth environmental pressure parameter evaluation index, h is the number of environmental pressure parameter evaluation indicators, k is the index of the environmental pressure parameter evaluation index, k=1,2, corresponding to pollutant emissions and temperature change rate respectively, the value range of E is between 0 and 1, E is close to 0, indicating that the environmental pressure parameter is close to the baseline value and the environmental condition is good, and E is close to 1, indicating that the environmental pressure parameter deviates seriously from the baseline value and the environmental condition is poor.

5. The method for early warning of ecological quality assessment in the bay area based on machine learning according to claim 4 is characterized by: In S3, the abnormal situation identification process is as follows: S31. Collect historical data containing known ecological quality status in the target bay area, and mark abnormal water quality parameters, abnormal air quality parameters, and abnormal environmental pressure parameters in the historical data; S32, integrating various evaluation indicators of water quality parameters, air quality parameters and environmental pressure parameters to obtain a comprehensive data set, and dividing the data set into a training set and a test set; S33. Based on the evaluation index data of the training set combined with the neural network model, the water quality parameters, air quality parameters and environmental pressure parameters in the training set are used as input features, and the normal or abnormal labels are used as outputs. The abnormal ecological quality recognition model is trained, and the model parameters are adjusted to enable it to recognize abnormal ecological quality. The performance of the model is evaluated in combination with the test set data, and the model is optimized according to the evaluation results. S34. Apply the trained model to the real-time ecological quality data of the target bay area, calculate the anomaly recognition index based on the probability of the model output, and set the anomaly threshold Y to determine whether there is an anomaly.

6. The method for early warning of ecological quality assessment in the bay area based on machine learning according to claim 5 is characterized by: The calculation formula of the abnormal identification index is as follows: ； In the formula, T is the anomaly recognition index, is the actual measured value of the tth real-time ecological quality parameter, is the weight of the tth parameter, N is the number of ecological quality parameters, is the adjustment factor, is the adjustment factor, Represents a randomly generated number in the interval [0,1]. If , it is considered to be abnormal.

7. The method for early warning of ecological quality assessment in the bay area based on machine learning according to claim 6 is characterized by: In S4, the process of dynamic assessment of ecological quality in the bay area is as follows: S41, inputting the parameter data of water quality, air quality and environmental pressure of the target area into the trained abnormal ecological quality recognition model, performing abnormal detection on the target area, the model outputs the probability of normal or abnormal, and obtains the recognition result of the abnormal state according to the set abnormal threshold; S42. Calculate the water quality deviation index, air quality deviation index and environmental pressure trend index of the target area, analyze the changing trend of each parameter over time, and assign corresponding weights according to the importance of each parameter in the ecosystem; S43. Based on the weights assigned to the various parameters, combined with the calculation results of the water quality deviation index, air quality deviation index and environmental pressure trend index of the target area, the ecological quality assessment coefficient is comprehensively calculated, and according to the size of the comprehensive ecological quality assessment coefficient, the ecological quality status of the target bay area is analyzed.

8. The method for early warning of ecological quality assessment in the bay area based on machine learning according to claim 7 is characterized by: The calculation formula of the ecological quality assessment coefficient is as follows: ； In the formula, SC is the ecological quality assessment coefficient, W is the water quality deviation index, R is the air quality deviation index, and E is the environmental pressure trend index. is the weight of the water quality deviation index, is the weight of the air quality deviation index, is the weight of the environmental pressure trend index. It should be noted that SC approaching 1 indicates that the water quality, air quality and environmental pressure parameters are close to the benchmark values, and the ecological quality is in good condition.

9. The method for early warning of ecological quality assessment in the bay area based on machine learning according to claim 8 is characterized by: In S5, the process of early warning analysis is: S51. Based on the requirements of the ecological quality of the bay area and the historical ecological quality data of the bay area, the ecological quality of the bay area is graded into high-quality ecological quality grade, good ecological quality grade and poor ecological quality grade; S52. Combine the calculation results of the ecological quality assessment coefficients, match them with each ecological quality level, and set corresponding ecological quality thresholds for each ecological quality level; S53, based on the value of the ecological quality assessment coefficient, determine the current ecological quality level, and mark it with different status colors, mark the high-quality ecological quality level with a green status color, mark the good ecological quality level with a yellow status color, and mark the poor ecological quality level with a red status color; S54. When the ecological quality level is good, an early warning notice is issued, suggesting preventive measures. When the ecological quality level is poor, an emergency early warning notice is issued, the emergency response mechanism is immediately activated, and the reasons for the early warning are analyzed, including water quality deterioration, air pollution aggravation, and environmental pressure increase; S55. Based on the results of early warning analysis, formulate targeted governance plans and implement corresponding governance measures. At the same time, supervise the implementation of governance measures, regularly evaluate the governance effects, and adjust the governance plans based on the evaluation results.

10. The method for early warning of ecological quality assessment in the bay area based on machine learning according to claim 9, characterized in that: The plurality of ecological quality levels correspond to the plurality of ecological quality thresholds, wherein the ecological quality thresholds include an upper threshold and a lower threshold; The multiple ecological quality levels and the multiple ecological quality thresholds satisfy the following relationship: The ecological quality thresholds of the high-quality ecological quality level are: ; The ecological quality thresholds of the good ecological quality level are: ; The ecological quality thresholds of the poor ecological quality level are: ; Among them, SC is the ecological quality assessment coefficient, is the lower threshold corresponding to the high-quality ecological quality level and the upper threshold corresponding to the good ecological quality level, is the lower threshold corresponding to the good ecological quality level and the upper threshold corresponding to the poor ecological quality level, , .

Images