Anomaly detection method for hydrological series data based on cloud computing

Abstract

The invention relates to the field of hydrological monitoring technology, and in particular to a hydrological sequence data anomaly detection method based on cloud computing. The invention classifies each monitoring unit according to altitude and land use type, generates corresponding environmental characteristic parameters, takes into account the influence of building density and geographical location on hydrological data, and improves the pertinence of wetland carbon fixation and sink monitoring and regulation data analysis and processing. The invention combines remote sensing technology and sensor groups with geographic information system technology to obtain preliminary data from each monitoring unit, and solves the problems of difficulty in speed measurement and sand locating and poor accuracy in traditional monitoring methods when monitoring large floods. Through cloud computing and server load analysis, virtual machine scheduling and load balancing, resources are dynamically allocated according to business needs, and computing tasks are allocated to multiple servers for processing, so as to avoid overloading of a single server, reduce idle computing power and waste, and improve overall resource utilization.

Description

Technical Field

[0001] The present invention relates to the technical field of hydrological monitoring, and in particular to a hydrological sequence data anomaly detection method based on cloud computing. Background Art

[0002] Hydrological monitoring technology is a comprehensive discipline that aims to monitor, measure, analyze and warn the temporal and spatial distribution and change patterns of water in nature through scientific methods. The hydrological monitoring system monitors and records the changes and patterns of hydrological processes by collecting, processing, storing and analyzing hydrological data. Hydrological data include parameters such as flow, water level, temperature, humidity, rainfall, etc., which can be collected by sensors, cameras, lidar and other equipment. Hydrological monitoring technology has a wide range of applications, including the monitoring of water bodies such as rivers, lakes, reservoirs, groundwater and oceans. By using various hydrological monitoring equipment, such as water level meters, flow meters, water quality monitors, etc., hydrological data can be collected in real time to dynamically monitor water resources. These data can reflect the physical, chemical and biological characteristics of water bodies, provide information on water cycle, groundwater recharge and water pollution, and provide a scientific basis for water resource management decisions.

[0003] With the integration of modern artificial intelligence and machine learning technologies in hydrological monitoring and the advancement of sensor technology, the collection and analysis of hydrological data has become more intelligent, achieving in-depth analysis and prediction of collected data. The integration of multi-source data, such as satellite remote sensing and geographic information system data, has also further improved the accuracy and versatility of hydrological information.

[0004] However, the existing hydrological monitoring technology still has many shortcomings. For example, some hydrological monitoring stations still use traditional technologies, which can no longer meet the needs of modern hydrological monitoring. For example, some monitoring stations still use traditional tools such as sounding rods and current meters. These methods have problems such as difficulty in measuring speed and positioning sand when monitoring large floods, and poor accuracy. Some hydrological monitoring equipment has errors when collecting data, affecting the accuracy and reliability of the data.

[0005] In addition, due to the limitation of human, material and financial resources, the number of monitoring points and sampling frequency of traditional monitoring methods are often limited. In large water areas or areas with complex geographical conditions, the coverage of traditional methods is even more insufficient.

[0006] Therefore, it is necessary to propose a hydrological series data anomaly detection method based on cloud computing to solve the above problems. Summary of the invention

[0007] In order to overcome the problems existing in the background technology, an embodiment of the present invention provides a hydrological sequence data anomaly detection method based on cloud computing.

[0008] The purpose of the present invention can be achieved by the following technical solution: A hydrological sequence data anomaly detection method based on cloud computing comprises the following steps:

[0009] Step 1: Division of monitoring areas;

[0010] Obtain a topographic map of the hydrological monitoring area and divide the area into A×A grids to obtain a number of monitoring units, where A is the preset monitoring accuracy factor.

[0011] Each monitoring unit is numbered with the symbol i, i=1, 2, 3, ..., n; wherein n is the total number of monitoring units.

[0012] As a preferred embodiment of the present invention, the altitudes hi1, hi2, hi3 and hi4 of the four grid corner points of each monitoring unit i are obtained, and the average altitude Hi of each monitoring unit i is calculated by the formula Hi=(hi1+hi2+hi3+hi4) / 4.

[0013] On this basis, all monitoring units are divided into their corresponding terrain types according to the preset classification rule 1 and marked:

[0014] The monitoring units with an average altitude greater than 500 meters are marked as plateau terrain;

[0015] Monitoring units with an average altitude greater than 200 meters and less than or equal to 500 meters are marked as hilly terrain;

[0016] The monitoring units with an average altitude greater than 0 and less than or equal to 200 meters are marked as plain terrain;

[0017] Monitoring units with an average elevation less than or equal to 0 are marked as basin terrain.

[0018] Each terrain type is preset with a corresponding terrain characteristic coefficient α, and after the terrain type of each monitoring unit i is determined, a corresponding terrain characteristic coefficient αi is matched for it.

[0019] As a preferred embodiment of the present invention, the building density percentage Wi of each monitoring unit i is obtained.

[0020] On this basis, all monitoring units are divided into their corresponding land use types according to the preset classification rule 2 and marked:

[0021] The monitoring units with a building density percentage greater than 80% are marked as high-density land;

[0022] Monitoring units with a building density percentage greater than 40% and less than or equal to 80% are marked as medium-density land;

[0023] The monitoring units with a building density percentage greater than 10% and less than or equal to 40% are marked as low-density land;

[0024] Monitoring units with a building density percentage less than or equal to 10% are marked as open land.

[0025] Each land use type is preset with a corresponding land use characteristic coefficient β. After the land use type of each monitoring unit i is determined, a corresponding land use characteristic coefficient βi is matched for it.

[0026] The environmental characteristic parameter Ki of each monitoring unit i is calculated by the formula Ki=αi×βi.

[0027] Step 2: Record precipitation data and hydrogeological data;

[0028] Get the average daily precipitation of each monitoring unit i ; Obtain the concentrated precipitation in each monitoring unit i per hour and record it as precipitation intensity ; Count the number of precipitation events that occur in each monitoring unit each month and record it as the precipitation frequency .

[0029] Get the average groundwater level of each monitoring unit i ;

[0030] Get the average dissolved oxygen concentration of surface runoff in each monitoring unit i , pH value , nitrogen pollutant concentration and phosphate ion concentration .

[0031] Step 3: Remote sensing data capture;

[0032] Retrieve the remote sensing satellite map of the hydrological monitoring area in the GIS geographic information system. Obtain the surface water area contained in each monitoring unit According to the reflection and radiation characteristics of surface water systems in different bands such as visible light, near infrared, mid-far infrared and microwave, the reflected light intensity of surface water systems in the green light band in each detection unit is captured with a central wavelength of 550nm and a wavelength of 40nm. , with a central wavelength of 710nm and a wavelength width of 30nm, capturing the reflected light intensity of the surface water system in the red light band in each detection unit By formula Calculate the water eutrophication parameter Ui of each monitoring unit i.

[0033] As a preferred embodiment of the present invention, for the surface water area Greater than the preset threshold The monitoring unit is used to obtain the maximum cross-sectional flow of the largest lake or river in the surface water system. .

[0034] Step 4: Eigenvalue analysis based on cloud computing;

[0035] The data obtained about each monitoring unit i, including environmental characteristic parameters Ki, average daily precipitation , precipitation intensity , precipitation frequency , Maximum cross-sectional flow , average groundwater level , average dissolved oxygen concentration , pH value , nitrogen pollutant concentration , phosphate ion concentration Surface water area The water eutrophication parameters Ui are uploaded to the cloud computing platform for task allocation and analysis and calculation.

[0036] The specific process of task allocation is as follows: perform virtual machine scheduling and load balancing, number each server in the cloud computing platform, the number symbol is j, i=1, 2, 3, ..., m; where m is the total number of servers. At every preset time interval t, the virtual machine of the central server sends a polling signal to all servers j. After receiving the polling signal, each server immediately starts self-checking and sends its own feedback data to the virtual machine of the central server. The feedback data includes: memory bandwidth occupancy rate 、CPU usage , Disk usage and network interface transmission bandwidth The central server counts the time interval Ti between sending polling signals to each server and receiving feedback data, and uses the formula Calculate the load characteristic value of each server ,in It is the preset weight factor corresponding to each server. Sort the data from high to low and extract the first M0 servers and mark them as idle servers. The acquired data about each monitoring unit i is divided into M0 copies and sent to the idle servers respectively, where M0 is the preset server scheduling index, M0 is a positive integer and M0 is less than m.

[0037] The specific process of analysis and calculation is as follows: every preset time interval t, the formula Calculate the rainfall infiltration characteristic value E1i, runoff characteristic value E2i and eutrophication characteristic value E3i of each monitoring unit i, where , , , , , are respectively the preset weight factors; ki is the environmental characteristic parameter.

[0038] As a preferred method of the present invention, the rainfall infiltration characteristic value E1i, runoff characteristic value E2i and eutrophication characteristic value E3i generated at each preset time for each monitoring unit i are sorted by time t to generate a characteristic value sequence, and the rainfall infiltration characteristic value-time graph, runoff characteristic value-time graph and eutrophication characteristic value-time graph are plotted respectively. The fitting function E1i=f1i(t) of the rainfall infiltration characteristic value-time graph of each monitoring unit i is obtained; the fitting function E2i=f2i(t) of the runoff characteristic value-time graph of each monitoring unit i is obtained; the fitting function E3i=f3i(t) of the eutrophication-time graph of each monitoring unit i is obtained;

[0039] Step 5: Characteristic value sequence analysis and abnormal recording;

[0040] By formula Calculate the fluctuation curvature K(Epi) of the rainfall infiltration characteristic value-time diagram, runoff characteristic value-time diagram and eutrophication characteristic value-time diagram, where p is the characteristic value number symbol, p=1, 2, 3; where fpi'(t) is the derivative of the function fpi(t), fpi(t)=f1i(t), f2i(t), f3i(t).

[0041] By formula Calculate the fluctuation value R(Epi) of the rainfall infiltration characteristic value-time diagram, runoff characteristic value-time diagram and eutrophication characteristic value-time diagram. Where fpi(t)Max and fpi(t)Min are the maximum and minimum values ​​of the characteristic values, respectively, and the characteristic values ​​include rainfall infiltration characteristic value, runoff characteristic value and eutrophication characteristic value; where fpi"(t) is the second-order derivative of the function fpi(t).

[0042] When it is detected that the fluctuation curvature K (Epi) is greater than the preset maximum threshold or the fluctuation value R (Epi) is greater than the preset maximum threshold, it is judged that the corresponding eigenvalue sequence has abnormal fluctuations and marked as an abnormal eigenvalue sequence;

[0043] Extracting the detection unit number i and the characteristic value number p from the fluctuation curvature K(Epi) or the fluctuation value R(Epi) of the abnormal characteristic value sequence;

[0044] If p = 1, it is judged that the ability of monitoring unit i to convert rainfall into groundwater recharge has degraded or fluctuated abnormally;

[0045] If p = 2, it is determined that the surface runoff of monitoring unit i has abnormal fluctuations;

[0046] If p = 3, it is determined that the eutrophication level of the surface water system in monitoring unit i has abnormal fluctuations.

[0047] Compared with the prior art, the present invention has the following beneficial effects:

[0048] 1. The present invention classifies each monitoring unit according to altitude and land use type, generates corresponding environmental characteristic parameters, substitutes them into the calculation of rainfall infiltration characteristic values, runoff characteristic values ​​and eutrophication characteristic values, considers the influence of building density and geographical location on hydrological data, and uses quantitative indicators to distinguish the environmental characteristics of each monitoring unit, including human activity density and altitude, thereby improving the pertinence of wetland carbon fixation and sink monitoring and regulation data analysis and processing;

[0049] 2. The present invention combines remote sensing technology and sensor groups with geographic information system technology to obtain preliminary data from each monitoring unit, solving the problems of difficulty in measuring speed and positioning sand and poor accuracy in traditional monitoring methods when monitoring large floods, and preventing the problems of insufficient number of monitoring points and limited sampling frequency due to limitations in manpower, material and financial resources, thereby improving the monitoring accuracy of hydrological data, increasing the monitoring coverage, and improving the accuracy and reliability of data; and timely capturing abnormal fluctuations in rainfall infiltration capacity, surface runoff and eutrophication contained in hydrological data through computational analysis of hydrological data sequences, thereby improving the comprehensiveness and reliability of data analysis methods;

[0050] 3. The present invention uses cloud computing and server load analysis, virtual machine scheduling and load balancing to dynamically allocate resources according to business needs, distribute computing tasks to multiple servers for processing, and avoid overloading of a single server. It ensures that each virtual machine can obtain the required computing, storage and network resources, reduces idle computing power and waste, and improves overall resource utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0052] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0053] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only 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.

[0054] See also Figure 1 As shown, the hydrological sequence data anomaly detection method based on cloud computing includes the following steps:

[0055] Step 1: Division of monitoring areas;

[0056] Obtain a topographic map of the hydrological monitoring area and divide the area into A×A grids to obtain a number of monitoring units, where A is the preset monitoring accuracy factor.

[0057] Each monitoring unit is numbered with the symbol i, i=1, 2, 3, ..., n; wherein n is the total number of monitoring units.

[0058] Furthermore, the altitudes hi1, hi2, hi3 and hi4 of the four grid corner points of each monitoring unit i are obtained, and the average altitude Hi of each monitoring unit i is calculated by the formula Hi=(hi1+hi2+hi3+hi4) / 4.

[0059] On this basis, all monitoring units are divided into their corresponding terrain types according to the preset classification rule 1 and marked:

[0060] The monitoring units with an average altitude greater than 500 meters are marked as plateau terrain;

[0061] Monitoring units with an average altitude greater than 200 meters and less than or equal to 500 meters are marked as hilly terrain;

[0062] The monitoring units with an average altitude greater than 0 and less than or equal to 200 meters are marked as plain terrain;

[0063] Monitoring units with an average elevation less than or equal to 0 are marked as basin terrain.

[0064] Each terrain type is preset with a corresponding terrain characteristic coefficient α, and after the terrain type of each monitoring unit i is determined, a corresponding terrain characteristic coefficient αi is matched for it.

[0065] Furthermore, the building density percentage Wi of each monitoring unit i is obtained.

[0066] On this basis, all monitoring units are divided into their corresponding land use types according to the preset classification rule 2 and marked:

[0067] The monitoring units with a building density percentage greater than 80% are marked as high-density land;

[0068] Monitoring units with a building density percentage greater than 40% and less than or equal to 80% are marked as medium-density land;

[0069] The monitoring units with a building density percentage greater than 10% and less than or equal to 40% are marked as low-density land;

[0070] Monitoring units with a building density percentage less than or equal to 10% are marked as open land.

[0071] Each land use type is preset with a corresponding land use characteristic coefficient β. After the land use type of each monitoring unit i is determined, a corresponding land use characteristic coefficient βi is matched for it.

[0072] The environmental characteristic parameter Ki of each monitoring unit i is calculated by the formula Ki=αi×βi.

[0073] It should be noted that the hydrological data recorded by each monitoring unit does not exist in isolation, but is deeply affected by multiple factors, among which the natural environment and human production activities are the two most significant aspects. Natural environmental factors act on the hydrological cycle in a subtle way, causing dynamic changes in hydrological characteristics such as water level, flow, and water quality. Agricultural irrigation, industrial emissions, and the increase in impervious areas during urbanization also have an impact on hydrological data. Environmental characteristic parameters are intended to distinguish the environmental characteristics of each monitoring unit, including human activity density and altitude, through quantitative indicators.

[0074] Step 2: Record precipitation data and hydrogeological data;

[0075] Get the average daily precipitation of each monitoring unit i ; Obtain the concentrated precipitation in each monitoring unit i per hour and record it as precipitation intensity ; Count the number of precipitation events that occur in each monitoring unit each month and record it as the precipitation frequency .

[0076] Get the average groundwater level of each monitoring unit i ;

[0077] Get the average dissolved oxygen concentration of surface runoff in each monitoring unit i , pH value , nitrogen pollutant concentration and phosphate ion concentration .

[0078] It should be noted that the nitrogen pollutants mentioned above refer to ammonia nitrogen, nitrate, nitrite nitrogen and various types of organic nitrogen including urea and amino acids. These nitrogen pollutants accumulate in water, promote the growth of algae, lead to eutrophication of water bodies, and thus affect water quality and the health of the aquatic ecosystem.

[0079] Step 3: Remote sensing data capture;

[0080] Retrieve the remote sensing satellite map of the hydrological monitoring area in the GIS geographic information system. Obtain the surface water area contained in each monitoring unit According to the reflection and radiation characteristics of surface water systems in different bands such as visible light, near infrared, mid-far infrared and microwave, the reflected light intensity of surface water systems in the green light band in each detection unit is captured with a central wavelength of 550nm and a wavelength of 40nm. , with a central wavelength of 710nm and a wavelength width of 30nm, capturing the reflected light intensity of the surface water system in the red light band in each detection unit By formula Calculate the water eutrophication parameter Ui of each monitoring unit i.

[0081] Furthermore, for the surface water area Greater than the preset threshold The monitoring unit is used to obtain the maximum cross-sectional flow of the largest lake or river in the surface water system. .

[0082] Step 4: Eigenvalue analysis based on cloud computing;

[0083] The data obtained about each monitoring unit i, including environmental characteristic parameters Ki, average daily precipitation , precipitation intensity , precipitation frequency , Maximum cross-sectional flow , average groundwater level , average dissolved oxygen concentration , pH value , nitrogen pollutant concentration , phosphate ion concentration Surface water area The water eutrophication parameters Ui are uploaded to the cloud computing platform for task allocation and analysis and calculation.

[0084] The specific process of task allocation is as follows: perform virtual machine scheduling and load balancing, number each server in the cloud computing platform, the number symbol is j, i=1, 2, 3, ..., m; where m is the total number of servers. At every preset time interval t, the virtual machine of the central server sends a polling signal to all servers j. After receiving the polling signal, each server immediately starts self-checking and sends its own feedback data to the virtual machine of the central server. The feedback data includes: memory bandwidth occupancy rate 、CPU usage , Disk usage and network interface transmission bandwidth The central server counts the time interval Ti between sending polling signals to each server and receiving feedback data, and uses the formula Calculate the load characteristic value of each server ,in It is the preset weight factor corresponding to each server, and its size is customized by the technicians. Sort the data from high to low and extract the first M0 servers and mark them as idle servers. The acquired data about each monitoring unit i is divided into M0 copies and sent to the idle servers respectively, where M0 is the preset server scheduling index, M0 is a positive integer and M0 is less than m.

[0085] The specific process of analysis and calculation is as follows: every preset time interval t, the formula Calculate the rainfall infiltration characteristic value E1i, runoff characteristic value E2i and eutrophication characteristic value E3i of each monitoring unit i, where , , , , , are respectively the preset weight factors; ki is the environmental characteristic parameter.

[0086] Furthermore, the rainfall infiltration characteristic value E1i, runoff characteristic value E2i and eutrophication characteristic value E3i generated at each preset time for each monitoring unit i are sorted by time t to generate a characteristic value sequence, and the rainfall infiltration characteristic value-time graph, runoff characteristic value-time graph and eutrophication characteristic value-time graph are plotted respectively. The fitting function E1i=f1i(t) of the rainfall infiltration characteristic value-time graph of each monitoring unit i is obtained; the fitting function E2i=f2i(t) of the runoff characteristic value-time graph of each monitoring unit i is obtained; the fitting function E3i=f3i(t) of the eutrophication-time graph of each monitoring unit i is obtained;

[0087] Step 5: Characteristic value sequence analysis and abnormal recording;

[0088] By formula Calculate the fluctuation curvature K(Epi) of the rainfall infiltration characteristic value-time diagram, runoff characteristic value-time diagram and eutrophication characteristic value-time diagram, where p is the characteristic value number symbol, p=1, 2, 3; where fpi'(t) is the derivative of the function fpi(t), fpi(t)=f1i(t), f2i(t), f3i(t).

[0089] By formula Calculate the fluctuation value R(Epi) of the rainfall infiltration characteristic value-time diagram, runoff characteristic value-time diagram and eutrophication characteristic value-time diagram. Where fpi(t)Max and fpi(t)Min are the maximum and minimum values ​​of the characteristic values, respectively, and the characteristic values ​​include rainfall infiltration characteristic value, runoff characteristic value and eutrophication characteristic value; where fpi"(t) is the second-order derivative of the function fpi(t).

[0090] When it is detected that the fluctuation curvature K (Epi) is greater than the preset maximum threshold or the fluctuation value R (Epi) is greater than the preset maximum threshold, it is judged that the corresponding eigenvalue sequence has abnormal fluctuations and marked as an abnormal eigenvalue sequence;

[0091] Extracting the detection unit number i and the characteristic value number p from the fluctuation curvature K(Epi) or the fluctuation value R(Epi) of the abnormal characteristic value sequence;

[0092] If p = 1, it is judged that the ability of monitoring unit i to convert rainfall into groundwater recharge has degraded or fluctuated abnormally;

[0093] If p = 2, it is determined that the surface runoff of monitoring unit i has abnormal fluctuations;

[0094] If p = 3, it is determined that the eutrophication level of the surface water system in monitoring unit i has abnormal fluctuations.

[0095] It should be understood that the terms "include" and "comprising" used in the specification and claims of the present disclosure indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0096] It should also be understood that the terms used in this disclosure are only for the purpose of describing specific embodiments and are not intended to limit the disclosure. As used in this disclosure and claims, the singular forms "a", "an", and "the" are intended to include plural forms unless the context clearly indicates otherwise. It should also be further understood that the term "and / or" used in this disclosure and claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes these combinations;

[0097] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A hydrological sequence data anomaly detection method based on cloud computing, characterized in that: The following steps are involved: Step 1: Division of monitoring area: Obtain a topographic map of the hydrological monitoring area and divide the area into A×A grids to obtain a number of monitoring units, where A is a preset monitoring precision factor; number each monitoring unit, the number symbol is i, i=1, 2, 3, ..., n; Where n is the total number of monitoring units; according to the preset classification rule 1, all monitoring units are divided into their corresponding terrain types, marked and matched to the corresponding terrain characteristic coefficient αi; according to the preset classification rule 2, all monitoring units are divided into their corresponding land use types, marked and matched to the corresponding land use characteristic coefficient βi: the environmental characteristic parameter Ki of each monitoring unit i is calculated by the formula Ki=αi×βi; The classification rule 1 is: Get the altitudes hi1, hi2, hi3 and hi4 of the four corner points of each monitoring unit i; calculate the average altitude Hi of each monitoring unit i by the formula Hi=(hi1+hi2+hi3+hi4) / 4; The monitoring units with an average altitude greater than 500 meters are marked as plateau terrain; Monitoring units with an average altitude greater than 200 meters and less than or equal to 500 meters are marked as hilly terrain; The monitoring units with an average altitude greater than 0 and less than or equal to 200 meters are marked as plain terrain; The monitoring units with average elevation less than or equal to 0 are marked as basin terrain; Each terrain type is pre-set with a corresponding terrain characteristic coefficient α. After determining the terrain type of each monitoring unit i, a corresponding terrain characteristic coefficient αi is matched for it; The second classification rule is: Obtain the building density percentage Wi of each monitoring unit i; The monitoring units with a building density percentage greater than 80% are marked as high-density land; Monitoring units with a building density percentage greater than 40% and less than or equal to 80% are marked as medium-density land; The monitoring units with a building density percentage greater than 10% and less than or equal to 40% are marked as low-density land; Mark monitoring units with a building density percentage less than or equal to 10% as open land; Each land use type is pre-set with a corresponding land use characteristic coefficient β. After determining the land use type of each monitoring unit i, a corresponding land use characteristic coefficient βi is matched for it; Step 2: Record precipitation data and hydrogeological data; obtain multi-dimensional hydrological data of each monitoring unit i; Step 3: Capture remote sensing data: retrieve the remote sensing satellite map of the hydrological monitoring area in the GIS geographic information system, and obtain the multi-dimensional remote sensing data and flow data of each detection unit i respectively; Step 4: Characteristic value analysis based on cloud computing: Upload the acquired data about each monitoring unit i to the cloud computing platform for task allocation and analysis and calculation; Step 5: Eigenvalue sequence analysis and abnormal recording: extract data from the eigenvalue sequence obtained through cloud computing to obtain the fluctuation curvature and fluctuation value. By formula Calculate the fluctuation curvature K(Epi) of the rainfall infiltration characteristic value-time diagram, runoff characteristic value-time diagram and eutrophication characteristic value-time diagram, where p is the characteristic value number, p=1, 2, 3; where fpi'(t) is the derivative of the function fpi(t), fpi(t)=f1i(t), f2i(t), f3i(t); By formula Calculate the fluctuation value R(Epi) of the rainfall infiltration characteristic value-time diagram, the runoff characteristic value-time diagram and the eutrophication characteristic value-time diagram; wherein fpi(t)Max and fpi(t)Min are the maximum and minimum values ​​of the characteristic values, respectively, and the characteristic values ​​include the rainfall infiltration characteristic value, the runoff characteristic value and the eutrophication characteristic value; wherein fpi″(t) is the second-order derivative of the function fpi(t); And according to the numerical value of fluctuation curvature and fluctuation value, judgment is made to detect the corresponding hydrological data anomaly.

2. The hydrological sequence data anomaly detection method based on cloud computing according to claim 1 is characterized in that: The multi-dimensional hydrological data includes: Average daily precipitation ; Obtain the concentrated precipitation in each monitoring unit i per hour and record it as precipitation intensity ; Count the number of precipitation events that occur in each monitoring unit each month and record it as the precipitation frequency ; Average groundwater level of each monitoring unit i ; Average dissolved oxygen concentration of surface runoff in each monitoring unit i , pH value , nitrogen pollutant concentration and phosphate ion concentration .

3. The hydrological sequence data anomaly detection method based on cloud computing according to claim 1 is characterized in that: The multi-dimensional remote sensing data and traffic data include: The surface water area contained in each monitoring unit i , water eutrophication parameters are obtained by extracting data and calculating and analyzing the reflection and radiation characteristics of surface water systems in different bands; For surface water area Greater than the preset threshold Monitoring unit i, obtain the maximum cross-sectional flow of the largest lake or river in the surface water system .

4. The hydrological sequence data anomaly detection method based on cloud computing according to claim 3 is characterized in that: The specific process of data extraction and calculation analysis based on the reflection and radiation characteristics of surface water systems in different bands is as follows: The reflected light intensity of the surface water system in the green light band in each detection unit is captured with a central wavelength of 550nm and a wavelength width of 40nm. , with a central wavelength of 710nm and a wavelength width of 30nm, capturing the reflected light intensity of the surface water system in the red light band in each detection unit ; Through the formula Calculate the water eutrophication parameter Ui of each monitoring unit i.

5. The hydrological sequence data anomaly detection method based on cloud computing according to claim 1 is characterized in that: The specific process of task allocation on the cloud computing platform is as follows: Number each server in the cloud computing platform, the number symbol is j, j = 1, 2, 3, ..., m; m is the total number of servers; at every preset time interval t, the virtual machine of the central server sends a polling signal to all servers j, and each server immediately starts self-checking after receiving the polling signal and sends its own feedback data to the virtual machine of the central server; the feedback data includes: memory bandwidth occupancy rate 、CPU usage , Disk usage and network interface transmission bandwidth ; The central server counts the time interval Ti from sending polling signals to each server to receiving feedback data, and uses the formula Calculate the load characteristic value of each server ,in is the preset weight factor corresponding to each server; according to the load characteristic value Sort the servers from high to low and extract the first M0 servers and mark them as idle servers; divide the acquired data about each monitoring unit i into M0 copies and send them to the idle servers respectively, where M is the preset server scheduling index, M0 is a positive integer and M0 is less than m.

6. The hydrological sequence data anomaly detection method based on cloud computing according to claim 1 is characterized in that: The specific process of uploading the acquired data about each monitoring unit to the cloud computing platform for analysis and calculation is as follows: Every preset time interval t, the formula Calculate the rainfall infiltration characteristic value E1i, runoff characteristic value E2i and eutrophication characteristic value E3i of each monitoring unit i, where , , , , , are respectively preset weight factors; ki is the environmental characteristic parameter; The rainfall infiltration characteristic value E1i, runoff characteristic value E2i and eutrophication characteristic value E3i of each monitoring unit i generated at each preset time are sorted according to time t to generate a characteristic value sequence, and the rainfall infiltration characteristic value-time diagram, runoff characteristic value-time diagram and eutrophication characteristic value-time diagram are plotted respectively; the fitting function E1i=f1i(t) of the rainfall infiltration characteristic value-time diagram of each monitoring unit i is obtained; the fitting function E2i=f2i(t) of the runoff characteristic value-time diagram of each monitoring unit i is obtained; the fitting function E3i=f3i(t) of the eutrophication-time diagram of each monitoring unit i is obtained.

7. The hydrological sequence data anomaly detection method based on cloud computing according to claim 1 is characterized in that: According to the fluctuation curvature and fluctuation value, the specific process of detecting the corresponding hydrological data anomaly is as follows: When it is detected that the fluctuation curvature K (Epi) is greater than the preset maximum threshold or the fluctuation value R (Epi) is greater than the preset maximum threshold, it is judged that the corresponding eigenvalue sequence has abnormal fluctuations and marked as an abnormal eigenvalue sequence; Extracting the detection unit number i and the characteristic value number p from the fluctuation curvature K(Epi) or the fluctuation value R(Epi) of the abnormal characteristic value sequence; If p = 1, it is judged that the ability of monitoring unit i to convert rainfall into groundwater recharge has degraded or fluctuated abnormally; If p = 2, it is determined that the surface runoff of monitoring unit i has abnormal fluctuations; If p = 3, it is determined that the eutrophication level of the surface water system in monitoring unit i has abnormal fluctuations.

Images