Lake ecological health evaluation method based on big data

Abstract

The invention discloses a lake ecological health evaluation method based on big data, and the method comprises the steps: inputting the current lake water quality ecological data obtained in advance into a lake ecological health model obtained through training, and obtaining the health level of a lake ecological system. Training to obtain a lake ecological health model, including: obtaining lake water quality ecological data, the lake water quality ecological data including monitoring data, regional weather data and hydrological data; an anomaly detection method based on hypothesis testing and an anomaly detection method based on prediction are combined, abnormal values existing in the monitoring data are subjected to abnormal value identification and cleaning, and updated lake water quality ecological data are obtained; and inputting the updated lake water quality ecological data into the constructed lake ecological health model, updating the weight of the constructed lake ecological health model, and obtaining a trained lake ecological health model. The method provides decision support for lake ecological health protection and basin sustainable development.

Description

Technical Field

[0001] The invention relates to a lake ecological health evaluation method based on big data, belonging to the technical field of environmental protection. Background Art

[0002] In recent years, driven by the idea of ​​sustainable development, ecosystem health research has become a hot topic in the international ecological environment field and a bridge connecting disciplines such as earth science, environmental science, ecology, economics and social science. The definition of the concept of ecosystem health is the premise and basis for studying any ecological system. Some scholars have given different definitions based on different focuses. At present, the concept of ecosystem health widely recognized by the academic community believes that a healthy ecosystem must be resilient to external pressures, maintain the diversity and stability of internal structures and organizations, and maintain the ability to metabolize. McMichael et al.'s concept of ecosystem health is that the health of lake ecosystems can be understood as the integrity of key ecological components and organic organizations in the lake, without disease, and can maintain the original function and structure after sudden natural or human disturbances, material circulation, energy and information flow are not damaged, and the overall function shows diversity, complexity and vitality.

[0003] The indicator system method is a trend in ecosystem health assessment research and can reflect the health status of an ecosystem from a certain perspective. So far, scholars have proposed many evaluation indicators, such as biological integrity indicators, structure, function and system indicators of freshwater ecosystems, and a comprehensive indicator system that includes biological, ecological, socio-economic and population health. However, due to the diversity of lake types, different lake ecosystems generally have different natural, social and economic conditions, and the same ecosystem has different characteristics at different stages of development. Therefore, a unified lake ecosystem health assessment indicator system has not yet been formed. Summary of the invention

[0004] The technical problem to be solved by the present invention is to overcome the defects of the prior art and provide a lake ecological health evaluation method based on big data. The present invention establishes a lake ecological health evaluation system including indicator entry and preprocessing, indicator weight hierarchical analysis, and comprehensive evaluation. By analyzing and detecting anomalies and evaluating the system of monitoring data, and with the help of technical means such as the Internet of Things, the fine numerical values ​​of water flow and water quality ecology in the basin are coupled to establish a fine numerical simulation technology system for the entire process from runoff generation and confluence, into the river, into the lake, to out of the lake, to form big data simulation results, and to construct a lake health diagnosis and intelligent management system portal, providing application services for lake ecological health monitoring and evaluation, and providing decision-making support for lake ecological health protection and basin sustainable development.

[0005] Preferably, the present invention provides a method for evaluating lake ecological health based on big data, comprising:

[0006] The pre-acquired current lake water quality ecological data is input into the trained lake ecological health model to obtain the health level of the lake ecosystem.

[0007] Prioritize training to obtain lake ecological health models, including:

[0008] Obtain lake water quality and ecological data, including monitoring data, regional weather data and hydrological data;

[0009] The anomaly detection method based on hypothesis testing is combined with the anomaly detection method based on prediction to identify and clean the outliers in the monitoring data and obtain updated lake water quality ecological data;

[0010] The updated lake water quality ecological data is input into the constructed lake ecological health model, the weight of the constructed lake ecological health model is updated, and the trained lake ecological health model is obtained.

[0011] Preferably, the pre-acquired current lake water quality ecological data and the health level of the lake ecosystem are displayed through at least one of graphs, charts and images.

[0012] Prioritize the combination of anomaly detection methods based on hypothesis testing and anomaly detection methods based on prediction to identify and clean outliers in monitoring data and obtain updated lake water quality ecological data, including:

[0013] Perform large-scale and non-real-time data calculations on updated lake water quality ecological data to obtain results representing the ecological health status of lake ecosystems;

[0014] According to the characterization results of the ecological health status of the lake ecosystem and the changing trends of the monitoring data, the outliers in the monitoring data are identified, and data compensation is performed on the identified outliers to obtain updated lake water quality ecological data.

[0015] Prioritize the combination of anomaly detection methods based on hypothesis testing and anomaly detection methods based on prediction to identify and clean outliers in monitoring data and obtain updated lake water quality ecological data, including:

[0016] S1. Calculate the variance statistics of the monitoring data in a sliding window of length L;

[0017] S2, taking the calculated variance difference between two adjacent sliding windows as the change value of the overall variance after the monitoring data arrives;

[0018] S3, marking the monitoring data corresponding to the variance difference value greater than the detection threshold as an abnormal value;

[0019] The selection of the detection threshold adopts the principle of dynamic adaptation: the variance difference of the calculated monitoring data is arranged in descending order, and the variance difference of the pth monitoring data of the unit is used as the detection threshold, p is the number of required abnormal values, and p is a positive integer.

[0020] Prioritize updating the weights of the lake ecological health model to obtain a trained lake ecological health model, including:

[0021] Normalize the pre-determined evaluation indicators, eliminate the dimension of each indicator value and unify the range of variation of each indicator value;

[0022] For the standard eigenvalue y′, which increases with the level h hj Increased evaluation index x * (i,j):

[0023] x * (i,j)=(x(i,j)-x min (j)) / ((x max (j)-x min (j)));

[0024] For the standard eigenvalue y′, which increases with the level h hj Reduced evaluation index x * (i,j):

[0025] x * (i,j)=(x max (j)-x(i,j)) / ((x max (j)-x min (j)));

[0026] Among them, x max (j) is the maximum value of the jth index, x min (j) is the minimum value of the jth indicator.

[0027] Prioritize updating the weights of the lake ecological health model to obtain a trained lake ecological health model, including:

[0028] Construct the projection indicator function:

[0029] Q(a) = SzRzy|;

[0030] in,

[0031]

[0032] In the formula, S Z is the standard deviation of the projection value z(i); R zy is the correlation coefficient between the projection value z(i) and the projection value y(i); E z 、Ey are the means of the projection value sequence {z(i)} and the sample experience level sequence {y(i)} respectively;

[0033] The optimal projection direction is obtained by solving the projection index function maximization problem:

[0034] max Q(a)=Sz|Rzy|

[0035] max Q(a)=Sz|Rzy|;

[0036]

[0037] Where Q(a) is the projection index function and a(j) is the projection direction.

[0038] Preferably, the fuzzy pattern recognition model includes m index composition matrices and index standard matrices of l lake water quality ecological data,

[0039] Among them, the indicator composition matrix X′={x i ' j} l×m ,i=1,2,…,l,j=1,2,…,m,

[0040] Index standard matrix Y′={y′ hj} c×m ,h=1,2,…,c,

[0041] In the formula, x i ' j is the attribute value of the jth indicator of the water quality ecological data of the ith lake; l means the set of l lake water quality ecological data that identifies the fuzzy subset A, and the m indicators are evaluated according to c level standards, y′ hj is the standard eigenvalue of indicator j at level h.

[0042] Preferably, the monitoring data include water temperature data, pH value, ammonia nitrogen data, turbidity, conductivity, dissolved oxygen, chlorophyll data, blue-green algae data and COD data;

[0043] Regional weather data includes soil temperature data, soil temperature and humidity data, air temperature data, air temperature and humidity data, ultraviolet intensity, PM2.5 concentration, wind speed data and wind direction data;

[0044] Hydrological data include water level, flow, flow velocity, rainfall, evaporation and sediment data of lake ecosystems.

[0045] Preferably, the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of any one of the methods described in the first aspect when executing the program.

[0046] Preferably, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any one of the methods described in the first aspect.

[0047] The beneficial effects achieved by the present invention are:

[0048] The present invention establishes a lake ecological health evaluation system including indicator entry and preprocessing, indicator weight hierarchical analysis, and comprehensive evaluation. By analyzing and detecting anomalies and evaluating the system of monitoring data, and with the help of technical means such as the Internet of Things, it couples the fine numerical values ​​of water flow and water quality ecology in the basin, establishes a fine numerical simulation technology system for the whole process from runoff generation and confluence, into the river, into the lake, to out of the lake, forms big data simulation results, constructs a lake health diagnosis and intelligent management system portal, provides application services for lake ecological health monitoring and evaluation, and provides decision-making support for lake ecological health protection and basin sustainable development.

[0049] The present invention does not adopt the traditional single abnormality monitoring method, but adopts a method combining an abnormality detection method based on hypothesis testing with an abnormality detection method based on prediction to monitor the ecological indicators of Baima Lake, combining the respective advantages of the two methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solution of the present application, the drawings required for use in the embodiments are briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0051] Figure 1 It is a structural block diagram of the lake ecological health evaluation system based on big data of the present invention;

[0052] Figure 2 It is the water quality index map of the present invention;

[0053] Figure 3 It is a big data graph of water quality indicators of the present invention;

[0054] Figure 4 It is a pH value diagram of water quality in the conservation area of ​​the present invention;

[0055] Figure 5 It is a water quality real-time data display and evaluation diagram of the present invention;

[0056] Figure 6 It is a real-time schematic diagram of the comprehensive data of the present invention. DETAILED DESCRIPTION

[0057] To facilitate the technical solution of the application, some concepts involved in the application are first explained below.

[0058] Embodiment 1

[0059] See also Figure 1 The present application discloses a lake ecological health evaluation method based on big data, comprising: inputting pre-acquired current lake water quality ecological data into a trained lake ecological health model to obtain the health level of the lake ecosystem.

[0060] For example:! YW: The health level of the lake health assessment is: 3.255;

[0061] The health level of lake ecosystem is 0≤YW≤1.5, which means that the lake ecosystem is very healthy;

[0062] The health level of lake ecosystem is 1.5<YW≤2.5, indicating that the lake ecosystem is healthy;

[0063] The health level of lake ecosystem is 2.5<YW≤3.5, which means that the lake ecosystem is sub-healthy;

[0064] The health level of the lake ecosystem is 3.5<YW≤4.5, indicating that the lake ecosystem is slightly pathological;

[0065] The health level of the lake ecosystem is 4.5<YW, which means that the lake ecosystem is in a pathological state.

[0066] The lake ecological health model is obtained by training, including: obtaining lake water quality ecological data, which includes monitoring data, regional weather data and hydrological data; identifying and cleaning outliers in the monitoring data to obtain updated lake water quality ecological data; inputting the updated lake water quality ecological data into the constructed lake ecological health model, updating the weights of the constructed lake ecological health model, and obtaining the trained lake ecological health model.

[0067] This application discloses a lake ecological health assessment system based on big data, including:

[0068] Step 1: Obtain lake water quality ecological data, which includes monitoring data, regional weather data and hydrological data;

[0069] 11) According to the "14th Five-Year Plan for the Establishment of Sections of the National Surface Water Environmental Quality Monitoring Network", it is clarified that the national surface water in the "14th Five-Year Plan" will be monitored in the "9+X" mode and evaluated in the "5+X" mode. Among them, the "9" in "9+X" refers to the 9 basic monitoring indicators configured by the national water control station, including temperature, pH value, turbidity value, conductivity value, dissolved oxygen value, ammonia nitrogen value, permanganate index value, total phosphorus value, and total nitrogen value; the "5" in "5+X" refers to pH value, dissolved oxygen value, ammonia nitrogen value, permanganate index value, and total phosphorus value; the "X" characteristic indicator will be dynamically adjusted according to the excessive indicators in the previous year and the current year.

[0070] 12) A variety of optical sensors and electrochemical technologies can be used for monitoring. The sensors are directly installed on the surface of the monitored water body. The solution integrates intelligent monitoring, automatic control, computer application and Internet of Things technologies, and can detect online water temperature data, pH value, ammonia nitrogen, turbidity, conductivity, dissolved oxygen, chlorophyll data, blue-green algae data, COD data, ammonia nitrogen data and other factors and other water quality parameter monitoring data to achieve unmanned monitoring and early warning.

[0071] 13) Regularly capture regional weather data, including soil temperature data, soil temperature and humidity data, air temperature data, air temperature and humidity data, ultraviolet intensity, PM2.5 concentration, wind speed data, and wind direction data; hydrological data include water level, flow, flow velocity, rainfall, evaporation, sediment data and other hydrological ecological information of lake ecosystems, providing multi-dimensional, in-depth and comprehensive data support for water resources ecological information analysis and prediction and management decision-making models.

[0072] Step 2: Transmission of lake water quality ecological data information;

[0073] 21) Measurement data transmission is based on IoT communication technologies such as physical layer Modbus, ZigBee, NB-IoT protocol and application layer MQTT protocol to transmit lake water quality ecological data and equipment information to the remote data center in real time, providing scientific and accurate data support for overall system data mining, equipment operation health monitoring and ecological management analysis and decision-making.

[0074] Step 3: Obtain data outlier detection, cleaning fusion and database construction;

[0075] 31) Perform computational analysis on the monitoring data, combine the anomaly detection method based on hypothesis testing with the anomaly detection method based on prediction, and identify the outliers in the monitoring data;

[0076] 32) Real-time data cleaning is performed on the lake water quality ecological data perceived by multiple sensors in the system perception layer. Due to the temporal requirements of ecological information, the system needs to strengthen the maintenance of temporal labels for lake water quality ecological data at this stage, and provide alarm prompts based on the early warning mechanism to facilitate subsequent data analysis and detection of abnormalities in collection equipment.

[0077] 33) After the cleaning is completed, a database is constructed to store the lake water quality ecological data locally, and the lake water quality ecological data is regularly backed up through the cloud database and the third-party data center to prevent the loss of lake water quality ecological data and ensure the long-term reliability of the lake water quality ecological data required by the system model.

[0078] Step 4: update the weight of the constructed lake ecological health model to obtain the trained lake ecological health model;

[0079] 41) Conduct comprehensive analysis of lake water quality ecological data, use artificial intelligence algorithms to provide data cleaning and data fusion middleware for lake water quality ecological perception information, and enhance system data governance and decision-making support capabilities.

[0080] 42) The projection pursuit method is used to determine the indicator weights of lake water quality ecological data, and based on this, the fuzzy pattern recognition model is used to evaluate the health of the lake ecosystem and obtain the health level of the lake ecosystem.

[0081] Step 5: Lake ecological health assessment information display and application;

[0082] 51) By using at least one of graphics, charts and images, the monitoring points and monitoring equipment information, real-time video images, real-time sampling data, sampling data and analysis and early warning data in the area are more intuitively displayed to enhance users' understanding of data correlation and structure.

[0083] In the data storage unit, a distributed storage cluster is constructed to realize the storage of data, which includes unstructured data, structured data and semi-structured data.

[0084] Step one includes: using monitoring units to construct an agricultural ecological information monitoring Internet of Things for monitoring agricultural environmental data, wherein the agricultural environmental data include soil temperature data, soil moisture data, air temperature data, air humidity data, wind speed data, wind direction data, rainfall data, and radiation data; constructing an air quality information monitoring Internet of Things for real-time monitoring of air quality information including air quality conditions and pollution conditions; constructing a water body ecological information monitoring Internet of Things for monitoring water body environmental data, wherein the water body environmental data include water temperature data, pH value, dissolved oxygen data, conductivity data, turbidity data, chlorophyll data, blue-green algae data, COD data, and ammonia nitrogen data;

[0085] Monitoring data include agricultural environmental data, air quality information and water environment data. Monitoring data is expressed as Among them, the data points Medium i Indicates the recording time of monitoring data, vector Indicates the recording time t i Collected monitoring data.

[0086] Computational analysis of monitoring data includes: offline calculation, which is to first perform large-scale, non-real-time data calculation on the lake ecosystem data to obtain the results representing the ecological health status of the lake ecosystem;

[0087] Real-time streaming computing is used to perform real-time computing and analysis on the monitoring data that is dynamically increased in real time. According to the changing law of the results of the ecological health status of the lake ecosystem and the changing trend of the current real-time monitoring data, the anomaly detection method based on hypothesis testing is combined with the anomaly detection method based on prediction to identify the outliers in the monitoring data, and perform data compensation for the identified outliers; According to the changing law of the results of the ecological health status of the lake ecosystem and the changing trend of the current real-time monitoring data, the anomaly detection method based on hypothesis testing is combined with the anomaly detection method based on prediction to identify the outliers in the monitoring data, including: S1, calculating the variance statistics of the monitoring data in a sliding window of length L;

[0088] S2, taking the calculated variance difference between two adjacent sliding windows as the change value of the overall variance after the current monitoring data arrives;

[0089] S3, marking the monitoring data corresponding to the variance difference value greater than the detection threshold as an abnormal value;

[0090] The selection of the detection threshold adopts the principle of dynamic adaptation: the variance difference values ​​of all the calculated monitoring data are arranged in descending order, and the variance difference value of the pth monitoring data of the unit is used as the detection threshold, where p is the number of required abnormal values ​​and p is a positive integer.

[0091] Update the weight of the lake ecological health model and obtain the trained lake ecological health model, including:

[0092] Normalize the pre-determined evaluation indicators, eliminate the dimension of each indicator value and unify the range of variation of each indicator value;

[0093] For evaluation indicators with larger evaluation levels: * (i,j)=(x(i,j)-x min (j)) / ((x max (j)-x min (j)));

[0094] For evaluation indicators with smaller evaluation levels, the higher the evaluation level: * (i,j)=(x max (j)-x(i,j)) / ((x max (j)-x min (j)));

[0095] Among them, x max (j) is the maximum value of the jth index, x min (j) is the minimum value of the jth indicator.

[0096] Update the weight of the lake ecological health model and obtain the trained lake ecological health model, including:

[0097] Construct the projection indicator function:

[0098] Q(a) = SzRzy|;

[0099] in,

[0100]

[0101] In the formula, S Z is the standard deviation of the projection value z(i); R zy is the correlation coefficient between the projection value z(i) and the projection value y(i); E z 、E y are the means of the projection value sequence {z(i)} and the sample experience level sequence {y(i)} respectively;

[0102] The optimal projection direction is obtained by solving the projection index function maximization problem:

[0103] max Q(a)=Sz|Rzy|;

[0104]

[0105] Where Q(a) is the projection index function and a(j) is the projection direction.

[0106] The fuzzy pattern recognition model is: suppose there are l sample sets for identifying fuzzy subset A, then there are m indicators of l samples forming a matrix X′={x i ' j} l×m ,i=1,2,…,l,j=1,2,…,m,where x i ' j is the attribute value of the jth indicator of the i-th sample; if the m indicators are evaluated according to the c-level standards, then the indicator standard matrix Y′={y′ hj} c×m ,h=1,2,…,c,y′hj is the standard eigenvalue of indicator j at level h.

[0107] In an embodiment of the present application, the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of any of the above methods when executing the program.

[0108] In an embodiment of the present application, the present invention provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of any of the above methods are implemented.

[0109] Indicators are generally of two types:

[0110] ①Indicator standard characteristic value y′ hj It decreases with the increase of level h;

[0111] ②Indicator standard characteristic value y′ hj It increases with the increase of level h;

[0112] For the first category of indicators, the relative membership of the C-level standard characteristic value less than or equal to the indicator to A is determined to be 1; for the second category of indicators, the relative membership of the C-level standard characteristic value equal to or greater than the indicator to A is determined to be 0; the membership of the level 1 characteristic value less than or equal to the indicator to A is 1; for the above two categories of indicators, whose characteristic values ​​are between the level 1 and level C standard characteristic values, the relative membership to A is determined according to the current changes; thus, the relative membership of the first category of indicators and the level h standard characteristic value to A is the standard characteristic value of level h and indicator j;

[0113] Similarly, the relative membership of the standard characteristic values ​​of the second category index and the h-level index to A can be obtained as

[0114]

[0115] Similarly, the relative membership of the standard characteristic values ​​of the second category index and the h-level index to A can be obtained as follows:

[0116]

[0117] In the formula, s hj is the relative membership of the standard eigenvalue of index j at level h to A, 0≤s hj ≤1, r ij is the relative membership of sample i index j to A, 0≤r ij ≤1;

[0118] Assume the weight vector of the index is W = (w1, w2, ..., w m ),satisfy:

[0119]

[0120] Here In order to solve the optimal fuzzy matrix, the objective function is proposed: the sum of the weighted generalized distances of all objects to each standard mode is minimized, that is:

[0121]

[0122] About UI h Find the partial derivative and solve the equation, and the relative membership matrix U of the sample set to each health level is obtained as follows:

[0123]

[0124] Then the pollution evaluation level characteristic value H of sample i can be calculated according to the formula i ;

[0125] In order to verify the reliability of the model calculation, a fuzzy evaluation and recognition model for sustainable utilization of water resources is specially introduced to verify the fuzzy recognition and evaluation model of lake ecological health based on projection pursuit. Specifically:

[0126] Indicator eigenvalue matrix and indicator standard eigenvalue matrix, there are n sample sets X=[x1,x2,…,x n ], each sample is described by m indicator eigenvalues, then these n samples can be represented by an m×n indicator eigenvalue matrix, that is:

[0127]

[0128] Where: x ij is the characteristic value of index i of sample j; i = 1, 2, ..., m; j = 1, 2, ..., n;

[0129] The standard eigenvalue matrix of indicators of m indicators and c levels is formed by identifying n samples, namely:

[0130]

[0131] Where:y ih is the standard characteristic value of index i at level h; h = 1, 2,…, c.

[0132] Relative membership function of indicators: The standard eigenvalues ​​of indicators are usually divided into two types:

[0133] The standard characteristic value of the first type of indicators decreases as the level h increases;

[0134] The standard characteristic value of the second type of indicators increases with the increase of level h;

[0135] Then the relative membership formula of the indicator to A is:

[0136]

[0137] Where: ri is the relative membership of the eigenvalue of index i of sample j to A.

[0138] Similarly, the relative degree formula of the standard value of level h index i to water resource utilization A can be obtained as follows:

[0139]

[0140] Where: sih is the relative membership of the standard value of index i at level h to A; the index standard eigenvalue matrix is ​​transformed into the corresponding relative membership matrix:

[0141] R=(rij)m×n

[0142] S=(sih)m×c

[0143] Since the m indicators have different influences on the identification and evaluation, the importance of each indicator is expressed by weight. The weight of the evaluation indicator has a great influence on the rationality of the evaluation result. There are many methods to determine the weight of each indicator, such as binary comparison fuzzy decision analysis, hierarchical analysis method, evaluation factor contribution rate method, etc. The weight of each evaluation indicator is obtained by using a certain method, and the indicator weight vector is obtained as follows:

[0144] Assume that the relative membership matrix of the sample set to each level is U = (uhj), where uhj is the relative membership of sample j to level h; since the relative membership of the m indicators of sample j all fall within the level interval [aj,bj] of the matrix S, the matrix U should satisfy:

[0145] Since the level interval of each sample in the sample set is different, from an overall perspective, the matrix U should satisfy:

[0146]

[0147]

[0148] The complete form of the membership formula is:

[0149] Where: The above formula is called the fuzzy pattern recognition model for the evaluation of sustainable utilization of water resources;

[0150] The formula for determining application-level eigenvalues ​​at the sample level is:

[0151] H=(1,2,…,c)(u 1j ,u2j ,…,ucj)T

[0152] Calculating the level characteristic values ​​of sustainable utilization of water resources in each region can reflect the sustainable utilization of water resources in each study area.

[0153] Embodiment 2

[0154] In a specific embodiment, taking Baima Lake as an example, according to the above system, combined with the analysis of the composition, structure and functional characteristics of the Baima Lake ecosystem, the ecological characteristics, natural functions and social environment indicators are integrated, and the Baima Lake evaluation index is selected to construct an indicator system for the health evaluation of the Baima Lake ecosystem, as shown in the following table:

[0155]

[0156]

[0157] Position XW(1): blue-green algae, XW(2): chlorophyll a, XW(3): dissolved oxygen, XW(4): turbidity, XW(5): pH, XW(6): total phosphorus, XW(7): total nitrogen, XW(8): ammonia nitrogen, XW(9): COD, XW(10): potassium permanganate index, XW(11): water temperature, input example: XW(1) = 196, XW(2) = 26.84, XW(3) = 6.54, XW(4) = 6.98, XW(5) = 8.71, XW(6) = 0.18, XW(7) = 0.67, XW(8) = 0.15, XW(9) = 81.09, XW(10) = 1.0, XW(11) = 19.3,

[0158] ! Output parameters:

[0159] YW: Lake health evaluation index: 3.255

[0160] ! ! ! IF(HH(I).GE.0.AND.HH(I).LE.1.5)THEN

[0161] 0≤YW≤1.5

[0162] WRITE(*,*),"The health level of Baima Lake is:",HH(I),"Very healthy"

[0163] ! ! ! ENDIF

[0164] ! ! ! IF(HH(I).GT.1.5.AND.HH(I).LE.2.5)THEN

[0165] 1.5<YW≤2.5

[0166] WRITE(*,*),"The health level of Baima Lake is:",HH(I),"Healthy"

[0167] ! ! ! ENDIF

[0168] ! ! ! IF(HH(I).GT.2.5.AND.HH(I).LE.3.5)THEN

[0169] 2.5<YW≤3.5

[0170] WRITE(*,*),"The health level of Baima Lake is:",HH(I),"Sub-health"

[0171] ! ! ! ENDIF

[0172] ! ! ! IF(HH(I).GT.3.5.AND.HH(I).LE.4.5)THEN

[0173] 3.5<YW≤4.5

[0174] WRITE(*,*),"The health level of Baima Lake is:",HH(I),"slightly sick"

[0175] ! ! ! ENDIF

[0176] IF(HH(I).GT.4.5)THEN

[0177] 4.5<YW

[0178] WRITE(*,*),"The health level of Baima Lake is:",HH(I),"Sick";

[0179] W: The weights of the 15 indicators are arranged from large to small. Generally, the first 3-5 are sufficient for JPO: 0.231, 0.202, 0.150, 0.0985. FIW: The ordinal numbers of the 15 indicators are arranged from large to small. Generally, the first 3-5 are sufficient for JPO: 5, 7, 11, 3, corresponding to the ordinal numbers of the indicators.

[0180] SBZZ (IXW) Ratio of 11 water quality indicators - judge the compliance of various water quality standards. Generally speaking, if it is greater than 1, it does not meet the standard, and if it is less than 1, it meets the standard. Water temperature 11 is not judged. NMDZSS: Nemerow index of water environment - Nemerow environmental quality index is 2.1222. The larger it is, the less healthy it is.

[0181] Position IXQ=8: The number of air meteorological index variables, which needs to be assigned values. XQ(1) is negative oxygen ions, XQ(2) is light intensity, XQ(3) is PM2.5, XQ(4) is wind direction, XQ(5) is wind speed, XQ(6) is air pressure, XQ(7) is air humidity, and XQ(8) is air temperature. For example: XQ(1)=1240, XQ(2)=9203, XQ(3)=12.4, XQ(4)=202.5, XQ(5)=2, XQ(6)=1004.3, XQ(7)=74.7, XQ(8)=34.1! ------24-hour average values ​​of the six air pollution indexes:

[0182] WRWQ (IWRQ = 6) represents the 24-hour average of 6 air pollution factors, including WRWQ (1) - SO2 - 24-hour average, WRWQ (2) - NO2 - 24-hour average, WRWQ (3) - PM10 - 24-hour average, WRWQ (4) - CO - 24-hour average, WRWQ (5) - O3 - 1-hour average (simplified, 24 hours is also acceptable), WRWQ (6) - PM2.5 - 24-hour average. For example: WRWQ (1) = 200, WRWQ (2) = 200, WRWQ (3) = 252, WRWQ (4) = 20, WRWQ (5) = 251, WRWQ (6) = 200,

[0183] ! Output parameters:

[0184] AQI: represents the air quality index AQI = MAX (IAQI (1), IAQI (2), IAQI (3), IAQI (4), IAQI (5), IAQI (6)),

[0185] The judgment is as follows: GE: means greater than or equal to, LE means less than or equal to;

[0186] ! ! ! IF(AQI.GE.0.AND.AQI.LE.50)&

[0187] ! ! ! WRITE(*,*),"AQI-PM2.5:",AQI,"Excellent"

[0188] ! ! ! IF(AQI.GE.51.AND.AQI.LE.100)&

[0189] ! ! ! WRITE(*,*),"AQI-PM2.5:",AQI,"good"

[0190] ! ! ! IF(AQI.GE.101.AND.AQI.LE.150)&

[0191] WRITE(*,*),"AQI-PM2.5:",AQI,"Lightly polluted"

[0192] ! ! ! IF(AQI.GE.151.AND.AQI.LE.200)&

[0193] WRITE(*,*),"AQI-PM2.5:",AQI,"Moderate pollution"

[0194] ! ! ! IF(AQI.GE.201.AND.AQI.LE.300)&

[0195] WRITE(*,*),"AQI-PM2.5:",AQI,"Heavy pollution"

[0196] IF(AQI.GT.300)&

[0197] WRITE(*,*),"AQI-PM2.5:",AQI,"Severe pollution"

[0198] According to the World Health Organization, negative oxygen ion concentrations above 1000 / cm3 can meet basic health needs. If it is below 1000 / cm3, the air can be considered polluted. Air with a concentration of 1000-1500 / cm3 is considered fresh air, and air with a concentration above 1500 / cm3 can be called healthy air. In other words, air rich in high concentrations of negative oxygen ions is healthy air.

[0199] IF(FYLZ.LT.1000)&

[0200] WRITE(*,*),"Negative oxygen ion concentration:",FYLZ,"Air pollution"

[0201] ! ! ! IF(FYLZ.GE.1000.AND.FYLZ.LT.1500)&

[0202] WRITE(*,*),"Negative oxygen ion concentration:",FYLZ,"Fresh air"

[0203] IF(FYLZ.GE.1500)&

[0204] WRITE(*,*),"Negative oxygen ion concentration:",FYLZ,"Healthy air"

[0205] The entire indicator system is a three-layer hierarchical structure indicator system consisting of the target layer, subsystem layer and evaluation indicator layer; the target layer is reflected by the subsystem layer, and the subsystem layer is reflected by the specific evaluation indicator layer; in fact, the target layer is a summary of the subsystem layer and the evaluation indicator layer;

[0206] As the comprehensive evaluation index of the target layer, it is used to comprehensively measure the lake ecological characteristic indicators (subsystem layer B1), natural function indicators (subsystem layer B2), and social environment indicators (subsystem layer B3); the health evaluation of lake ecosystems requires the selection of descriptive indicators and evaluative indicators so that they can reflect the changing trend on the time scale, reflect the structural characteristics on the spatial scale, and reflect the degree of impact in quantity, that is, A = {B1, B2, B3}.

[0207] The subsystem layer is the main factor restricting the health of the lake ecosystem, and can also be understood as a sub-target layer. This study will use lake ecological characteristic indicators (subsystem layer B1), natural function indicators (subsystem layer B2), and social environment indicators (subsystem layer B3) as the basis for judging the subsystem layer.

[0208] The indicator layer is composed of directly measurable indicators and is the most basic level in the lake ecosystem health assessment indicator system, based on the characteristics and significance of each item in the subsystem layer.

[0209] like Figure 2 As shown, the Baima Lake water quality factor detection and evaluation system is integrated into the large-screen display system of the big data management platform, which supports the real-time selection of various water areas on the map and displays various water quality factor information and evaluation result information;

[0210] like Figure 3 As shown, it also supports viewing detailed list data of various water quality indicators detected in each area, such as water temperature, dissolved oxygen and PH, etc.

[0211] Visualization supports further viewing of the results of various detailed indicators, such as Figure 4 As shown, you can view detailed indicators of water quality pH in the conservation area; you can see the detailed situation of water quality pH arranged by time so as to make further water quality analysis;

[0212] like Figure 5 As shown, the previous step counted the data of various water quality indicators. Next, the obtained water quality data is displayed on the big screen through big data and big screen visualization technology. The above content monitors the health status of various water quality data and displays them.

[0213] like Figure 6The Baimahu environmental factor detection and evaluation system shown is integrated into the large-screen display system of the big data management platform, which supports real-time selection of various information of each comprehensive detection point on the map, such as conductivity, light intensity, pH value, etc.; big data can display the monitoring results of various data through a ring chart, and calculate whether the data and equipment are normal through algorithms, and give an alarm if abnormal.

[0214] Among them, in this embodiment, the ecological health status of Baima Lake is divided into 5 levels, namely "very healthy, healthy, sub-healthy, slightly pathological, and pathological"; the evaluation result shows: one of the 5 levels, and also gives: the main reasons affecting the health status of Baima Lake: for example: the mass concentration of TN and TP in Baima Lake is too high.

[0215] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0216] Those skilled in the art will readily appreciate other embodiments of the invention after considering the specification and practicing the invention invented herein. This application is intended to cover any variations, uses or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art that are not invented by the present invention. The specification and examples are intended to be exemplary only, and the true scope and spirit of the invention are indicated by the claims.

[0217] The above specific implementation methods further illustrate the purpose, technical solutions and beneficial effects of the present application in detail. It should be understood that the above are only specific implementation methods of the present application and are not used to limit the protection scope of the present application. Any modifications, equivalent substitutions, improvements, etc. made on the basis of the technical solutions of the present application should be included in the protection scope of the present application.

Claims

1. A lake ecological health assessment method based on big data, characterized in that: include: The pre-acquired current lake water quality ecological data is input into the trained lake ecological health model to obtain the health level of the lake ecosystem.

2. According to the big data lake ecological health assessment method of claim 1, it is characterized in that: The lake ecological health model is obtained through training, including: Obtain lake water quality and ecological data, including monitoring data, regional weather data and hydrological data; The anomaly detection method based on hypothesis testing is combined with the anomaly detection method based on prediction to identify and clean the outliers in the monitoring data and obtain updated lake water quality ecological data; The updated lake water quality ecological data is input into the constructed lake ecological health model, the weight of the constructed lake ecological health model is updated, and the trained lake ecological health model is obtained.

3. The method for evaluating lake ecological health based on big data according to claim 1 is characterized in that: The pre-acquired current lake water quality ecological data and the health level of the lake ecosystem are presented through at least one of graphs, charts and images.

4. The method for evaluating lake ecological health based on big data according to claim 2 is characterized in that: The anomaly detection method based on hypothesis testing is combined with the anomaly detection method based on prediction to identify and clean the outliers in the monitoring data and obtain updated lake water quality ecological data, including: Perform large-scale and non-real-time data calculations on updated lake water quality ecological data to obtain results representing the ecological health status of lake ecosystems; According to the characterization results of the ecological health status of the lake ecosystem and the changing trends of the monitoring data, the outliers in the monitoring data are identified, and data compensation is performed on the identified outliers to obtain updated lake water quality ecological data.

5. The method for evaluating lake ecological health based on big data according to claim 4 is characterized in that: The anomaly detection method based on hypothesis testing is combined with the anomaly detection method based on prediction to identify and clean the outliers in the monitoring data and obtain updated lake water quality ecological data, including: S1. Calculate the variance statistics of the monitoring data in a sliding window of length L; S2, taking the calculated variance difference between two adjacent sliding windows as the change value of the overall variance after the monitoring data arrives; S3, marking the monitoring data corresponding to the variance difference value greater than the detection threshold as an abnormal value; The selection of the detection threshold adopts the principle of dynamic adaptation: the variance difference of the calculated monitoring data is arranged in descending order, and the variance difference of the pth monitoring data of the unit is used as the detection threshold, p is the number of required abnormal values, and p is a positive integer.

6. The method for evaluating lake ecological health based on big data according to claim 2 is characterized in that: Update the weight of the lake ecological health model and obtain the trained lake ecological health model, including: Normalize the pre-determined evaluation indicators, eliminate the dimension of each indicator value and unify the range of variation of each indicator value; For the standard eigenvalue y′, which increases with the level h hj Increased evaluation index x * (i,j): x * (i,j)=(x(i,j)-x min (j))((x max (j)-x min (j))); For the standard eigenvalue y′, which increases with the level h hj Reduced evaluation index x * (i,j): x * (i,j)=(x max (j)-x(i,j))((x max (j)-x min (j))); Among them, x max (j) is the maximum value of the jth index, x min (j) is the minimum value of the jth indicator.

7. The method for evaluating lake ecological health based on big data according to claim 2 is characterized in that: Update the weight of the lake ecological health model and obtain the trained lake ecological health model, including: Construct the projection indicator function: Q(a)=S z |R zy |; in, In the formula, S z is the standard deviation of the projection value z(i); R zy is the correlation coefficient between the projection value z(i) and the projection value y(i); E z 、E y are the means of the projection value sequence {z(i)} and the sample experience level sequence {y(i)} respectively; The optimal projection direction is obtained by solving the projection index function maximization problem: max Q(a)=S z |R zy |; Where Q(a) is the projection index function and a(j) is the projection direction.

8. The method for evaluating lake ecological health based on big data according to claim 2 is characterized in that: The fuzzy pattern recognition model includes m index composition matrices and index standard matrices of l lake water quality ecological data. Among them, the indicator composition matrix X′={x i ' j } l×m ,i=1,2,…,l,j=1,2,…,m, Index standard matrix Y′={y′ hj } c×m ,h=1,2,…,c, In the formula, x i ' j is the attribute value of the jth indicator of the water quality ecological data of the ith lake; l means the set of l lake water quality ecological data that identifies the fuzzy subset A, and the m indicators are evaluated according to c level standards, y′ hj is the standard eigenvalue of indicator j at level h.

9. The method for evaluating lake ecological health based on big data according to claim 2 is characterized in that: The monitoring data include water temperature data, pH value, ammonia nitrogen data, turbidity, conductivity, dissolved oxygen, chlorophyll data, blue-green algae data and COD data; Regional weather data includes soil temperature data, soil temperature and humidity data, air temperature data, air temperature and humidity data, ultraviolet intensity, PM2.5 concentration, wind speed data and wind direction data; Hydrological data include water level, flow, flow velocity, rainfall, evaporation and sediment data of lake ecosystems.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 9 are implemented.

Images