Ground meteorological station network service management system and method

Abstract

The invention relates to the technical field of information networks, in particular to a service management system and method for a ground meteorological station network, and the system comprises a data collection module, a background service module, and a web front-end module. And the background service module comprises a data preprocessing component, a data storage component, a data analysis component, an anomaly detection component and an alarm notification component which supports multiple detection algorithms and flexibly configures alarm rules. The temperature, humidity, pressure and rainfall data of a ground meteorological station are monitored in real time, preprocessing, analysis and anomaly detection are performed on acquired data through a background service module, and during anomaly detection, an adaptive space-time correlation model is constructed and deep fusion of multi-dimensional feature data is combined, so that accurate detection of data anomaly is realized. The interruption or loss of the data in the operation process is reduced, and the accuracy of data acquisition is improved.

Description

Technical Field

[0001] The invention belongs to the technical field of information networks, and in particular relates to a ground meteorological station network business management system and method. Background Art

[0002] The ground meteorological station network covers meteorological observation and data collection, data transmission and processing, meteorological forecast and warning, climate research and resource assessment, equipment maintenance and calibration, business management and personnel training. The development of these businesses is of great significance to improving the accuracy of meteorological forecasts, protecting the lives and property of the public, and promoting climate research and resource assessment.

[0003] At present, the business management method of ground meteorological station network is usually to regularly transmit the collected data from different meteorological stations to the superior meteorological department, analyze and process the data at the same time, and provide timely feedback on relevant meteorological information;

[0004] However, ground-based meteorological observation data are usually transmitted to data centers for processing and analysis via radio, the Internet, etc. In actual operations, data transmission may be interrupted or lost due to network failures, equipment failures, etc., thus affecting the accuracy of data collection. Therefore, we need to propose a ground meteorological station network business management system and method to solve the above-mentioned problems, so that it can monitor abnormal data in real time, provide an efficient early warning mechanism, and improve the accuracy and reliability of data collection. Summary of the invention

[0005] In view of the above problems, the present invention provides a ground weather station network business management system, comprising: a data acquisition module, which is responsible for collecting multi-dimensional data of temperature, pressure, humidity and rainfall from multiple ground weather stations in real time;

[0006] A backend service module, which includes a data preprocessing component, a data storage component, a data analysis component, an anomaly detection component, and an alarm notification component that supports multiple detection algorithms and flexibly configures alarm rules;

[0007] The data preprocessing component is used to perform preprocessing operations such as cleaning, format conversion and verification on the collected data to ensure data quality. The data preprocessing component can expand preprocessing rules and support plug-in rule addition;

[0008] The data storage component is used to store the processed data in a time series database, a relational database and a cache database;

[0009] The data analysis component is used to perform statistical analysis, trend analysis and data mining on the data to provide decision support. The data analysis component also supports batch processing and streaming processing and provides a flexible analysis model;

[0010] The anomaly detection component achieves accurate detection of data anomalies by building an adaptive spatiotemporal correlation model and combining deep fusion of multi-dimensional features;

[0011] The alarm notification component is responsible for sending alarm notifications when an abnormality is detected, and supports multiple notification methods;

[0012] A web front-end module, which is used to provide a user-friendly interface and display data visualization results, including map display, data reports and alarm information;

[0013] The backend service module is electrically connected to the data acquisition module and the web front-end module respectively.

[0014] Furthermore, the data acquisition module includes a sensor unit and a data transmission unit. The sensor unit includes a temperature sensor, a humidity sensor, a meteorological pressure sensor and a rainfall sensor. The temperature sensor is used to sense the temperature data of the meteorological station in real time. The humidity sensor is used to collect the humidity data of the meteorological station in real time. The meteorological pressure sensor is used to sense the pressure of the atmosphere on the unit area of ​​the earth's surface. The rainfall sensor is used to collect rainfall data.

[0015] The data transmission unit converts the received sensor data into digital signals, and then transmits the converted digital signals to the background service module via wireless or limited transmission for data processing.

[0016] Furthermore, the process of the data preprocessing component performing preprocessing operations on the data is as follows:

[0017] B11. Process missing values, abnormal values, and duplicate data in the received data; the missing values ​​are processed by deleting them or filling them with the mean; duplicate data are directly deleted;

[0018] There are three situations for outlier processing:

[0019] 1) If the number of outliers is small and does not affect the overall distribution of the data, delete them directly;

[0020] 2) If the outlier is due to data entry error or measurement error, correct it to the correct value;

[0021] 3) If the outlier contains important information, it will be retained;

[0022] B12. Convert data in different formats into internal standard formats;

[0023] B13. Verify the integrity and rationality of the data to filter out invalid data.

[0024] Furthermore, the statistical analysis in the data analysis component is to calculate the basic statistical indicators of average, maximum, minimum and standard deviation; trend analysis is to analyze the changing trend of data and predict future changes; data mining is to use machine learning algorithms to discover hidden patterns and associations in data; the data analysis component generates daily, weekly and monthly reports based on the results of statistical analysis, trend analysis and data mining, which is convenient for data visualization, and provides flexible analysis models for batch processing and streaming processing.

[0025] Furthermore, the process of the anomaly detection component performing data anomaly detection on meteorological data is as follows:

[0026] B21. Extract multi-dimensional features from meteorological data, including temperature, humidity, air pressure and wind speed data;

[0027] B22. Use deep neural networks to perform nonlinear fusion of multi-dimensional features to obtain high-level feature representation;

[0028] B23, dynamically divide the time and space windows and select the site data with the strongest correlation with the target site;

[0029] B24, establish an adaptive spatiotemporal parallel model between the target site and the associated sites;

[0030] B25, calculating the difference between the target site data and the predicted value of the adaptive spatiotemporal correlation model to obtain anomaly scores;

[0031] The calculation formula for the anomaly score is as follows:

[0032] , where S is the anomaly score and y is the actual observed value. is the predicted value of the adaptive spatiotemporal correlation model;

[0033] B26. Use dynamic thresholds to adjust abnormality judgment criteria based on real-time data characteristics;

[0034] The dynamic threshold is to set the confidence level based on the real-time distribution of anomaly scores. , and then calculate the dynamic threshold. The calculation formula of the dynamic threshold is: ,in, is the mean of the anomaly scores, is the standard deviation of the anomaly score, is the quantile of the standard normal distribution, T is the dynamic threshold;

[0035] B27. Adjust the parameters and thresholds of the adaptive spatiotemporal correlation model in real time according to the detection results to continuously optimize the detection performance.

[0036] Furthermore, the steps of nonlinearly fusing the multi-dimensional features are as follows:

[0037] B221, receive multi-dimensional meteorological characteristic data through the input layer, data table ;

[0038] B222, multi-layer nonlinear mapping is performed through hidden layers, and ReLU is used as the activation function to extract high-level features;

[0039] B223, output the fused feature F through the output layer;

[0040] B224. Dynamically adjust network weights based on the importance of features so that the adaptive spatiotemporal correlation model pays more attention to key features;

[0041] B225, use a combination of L1 and L2 regularization to prevent overfitting.

[0042] Furthermore, the steps for establishing the adaptive spatiotemporal correlation model are as follows:

[0043] B241, dynamically adjust the time window size according to the time variation characteristics of the target site data;

[0044] B242. Dynamically adjust the spatial window range according to the distribution density and relevance of the sites around the target site to ensure that the most relevant site is selected;

[0045] B243. Calculate the data correlation coefficient between the target site and the associated site within the time window, and select the site with high correlation as the associated site. The calculation formula is as follows: , where N is the number of data points in the time window, For the The observed value of the target site at a certain moment, is the average value of the target site at the moment, For associated sites The observed value of the target site at a certain moment, is the average value of the associated site at the moment, is the data correlation coefficient between the target site and the associated site within the time window.

[0046] Furthermore, the web front-end module includes a geographic visualization component, a user management component and an interface management component. The geographic visualization component is used to provide map rendering and geographic information display functions, interact with the web front-end module to present visualization results, and support dynamic display of real-time data;

[0047] The user management component is used to manage system users, roles and permissions, provide authentication and authorization services, and support multiple authentication methods;

[0048] The interface management component is used to provide an external API interface and support data collection and message integration functions.

[0049] Furthermore, the geographic visualization component includes a Leaflet map rendering engine, a Tiandi Map data service, and visualization elements. The Leaflet map rendering engine is used to realize the visualization display of geographic information on the web front end. The geographic information includes the location, status, and data of meteorological stations. The Tiandi Map data service integrates Tiandi Map's services through Leaflet plug-ins or custom layers to provide high-quality and diversified map data. The map data includes vector maps, image maps, and topographic maps to enrich the map display effect.

[0050] Based on the above-described ground weather station network business management system, the present invention also provides a ground weather station network business management method, comprising the following steps:

[0051] S1, receiving multi-dimensional data of temperature, pressure, humidity and rainfall from the weather station in real time through the data acquisition module;

[0052] S2. Use data preprocessing components to clean, convert and verify the collected data to ensure data quality;

[0053] S3, storing the preprocessed data in the time series database, relational database and cache database;

[0054] S4. Establish data index and backup mechanism to ensure data security and accessibility;

[0055] S5. Use data analysis components to perform statistical analysis, trend analysis and data mining on the data to reveal the inherent laws and trends of meteorological data;

[0056] S6. Build an adaptive spatiotemporal correlation model through the anomaly detection component, and accurately detect data anomalies by combining the deep fusion of multi-dimensional features; when an anomaly is detected, trigger the alarm notification component to promptly notify relevant personnel in a variety of ways;

[0057] S7. Use the web front-end module to provide users with a friendly interface and display the visualization results of the data.

[0058] The beneficial effects of the present invention are:

[0059] 1. The present invention cooperates with a data acquisition module, a background service module and a web front-end module, monitors the temperature, humidity, pressure and rainfall data of a ground meteorological station in real time, and pre-processes, analyzes and detects anomalies of the collected data through the background service module. During anomaly detection, an adaptive spatiotemporal correlation model is constructed, combined with the deep fusion of multi-dimensional feature data, to achieve accurate detection of data anomalies, reduce data interruption or loss during operation, and improve the accuracy of data collection.

[0060] 2. The present invention integrates adaptive spatiotemporal parallel modeling, multi-dimensional feature fusion and dynamic threshold adjustment technology through the cooperation of data preprocessing components, data storage components, data analysis components, anomaly detection components and alarm notification components. It can monitor and identify abnormal data in real time, dynamically adjust the parameters of the spatiotemporal correlation model, adapt to the meteorological data characteristics of different regions and time periods, use deep learning technology to integrate various meteorological feature data, capture complex nonlinear relationships, and automatically adjust the anomaly detection threshold according to the real-time data distribution characteristics to improve detection accuracy. It is highly innovative and practical.

[0061] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0063] Figure 1 A block diagram of a processing system according to an embodiment of the present invention is shown;

[0064] Figure 2 A flowchart showing a data preprocessing operation performed by a data preprocessing component according to an embodiment of the present invention is shown;

[0065] Figure 3 A flowchart showing data anomaly detection by an anomaly detection component on meteorological data according to an embodiment of the present invention is shown;

[0066] Figure 4 A flowchart of nonlinear fusion of multi-dimensional features according to an embodiment of the present invention is shown;

[0067] Figure 5A flowchart of a processing method according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0068] 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.

[0069] The embodiment of the present invention provides a ground weather station network service management system, such as Figure 1-4 As shown, it includes a data acquisition module, a background service module and a web front-end module, and the background service module is electrically connected to the data acquisition module and the web front-end module respectively; the data acquisition module is responsible for collecting multi-dimensional data of temperature, pressure, humidity and rainfall from multiple ground meteorological stations in real time;

[0070] The data acquisition module includes a sensor unit and a data transmission unit. The sensor unit includes a temperature sensor, a humidity sensor, a meteorological pressure sensor and a rainfall sensor. The temperature sensor is installed in a well-ventilated place of the weather station and can represent the ambient temperature, and is used to sense the temperature data of the weather station in real time; the humidity sensor is used to collect the humidity data of the weather station in real time, the meteorological pressure sensor is used to sense the pressure of the atmosphere on the unit area of ​​the earth's surface, and the rainfall sensor is used to collect rainfall data;

[0071] The data transmission unit converts the received sensor data into digital signals, and then transmits the converted digital signals to the background service module via wireless or limited transmission for data processing.

[0072] The backend service module includes a data preprocessing component, a data storage component, a data analysis component, an anomaly detection component, and an alarm notification component that supports multiple detection algorithms and flexibly configures alarm rules. The data preprocessing component is used to perform preprocessing operations such as cleaning, format conversion, and verification on the collected data to ensure data quality. The data preprocessing component can expand preprocessing rules and support plug-in rule addition;

[0073] like Figure 2 As shown, the process of the data preprocessing component performing preprocessing operations on data is as follows:

[0074] B11. Process missing values, abnormal values, and duplicate data in the received data to ensure data quality; missing values ​​are processed by deleting them or filling them with the mean; duplicate data are directly deleted;

[0075] There are three situations for outlier processing:

[0076] 1) If the number of outliers is small and does not affect the overall distribution of the data, delete them directly;

[0077] 2) If the outlier is due to data entry error or measurement error, correct it to the correct value;

[0078] 3) If the outlier contains important information, it is retained as a special case;

[0079] B12. Convert data in different formats into internal standard formats to facilitate subsequent processing;

[0080] B13. Verify the integrity and rationality of the data to filter out invalid data.

[0081] The data storage component is used to store the processed data in a time series database, a relational database, and a cache database, supports high-concurrency writing operations, and ensures data consistency and integrity;

[0082] The time series database stores time series data collected in real time and supports high-concurrency writing and querying;

[0083] The time series database uses the TDEngine time series database, which has high concurrent writing, high compression ratio and high query performance, supports super tables, facilitates the management of large amounts of similar time series data, provides a SQL-like query language, and is easy to use. It also reserves a distributed cluster expansion solution to achieve horizontal expansion of the database to meet the growth needs of data volume and concurrency, distribute the load among multiple nodes, and improve the stability and performance of the system. It uses efficient compression algorithms when compressing data to reduce storage space occupancy, sets a data retention period, archives or deletes expired data, and saves storage resources. In terms of query, it establishes time and label indexes to speed up queries, pre-calculates commonly used statistical indicators, and reduces the overhead of real-time calculations.

[0084] The relational database stores business data, metadata and user information, and supports transaction processing and complex queries. The relational database uses MySQL, which is mature, stable, supports transactions and has rich functions. The business data stored in MySQL includes user information, permission configuration and system configuration, etc. When writing to MySQL, the cache data in Redis is updated. When the data is updated, the cache is automatically cleared or updated to achieve data synchronization between MySQL and Redis.

[0085] The cache database is used to cache hotspot data and session information to increase data access speed.

[0086] The cache database uses Redis, which is memory-based, fast, and supports multiple data structures. This Redis caches hot data, session information, and message queues to achieve fast data access; the Redis cache service includes master-slave architecture and cluster deployment, data cache strategy, and Redis data type application; among them, the master-slave architecture and cluster deployment include master-slave replication, sentinel mechanism, and cluster mode. Master-slave replication is used to configure master and slave nodes to achieve data replication and backup; the sentinel mechanism is used to monitor the status of the master node, automatically perform failover, and ensure high availability; the cluster mode is used to deploy Redis clusters, store data in shards, and improve the scalability and performance of the system;

[0087] Data cache strategies include cache strategy, data expiration and data consistency. The cache strategy is to optimize memory usage based on data access frequency by using cache elimination strategies such as LRU and LFU; data expiration is to set the expiration time of the key, automatically clear expired data and save resources; data consistency is to use the consistent hashing algorithm to ensure the balance of data distribution when the node changes;

[0088] Redis data type applications include strings, hashes, lists, sets, and ordered sets. Strings are used to store simple key-value pairs, such as configuration parameters; hashes are used to store object information, such as user sessions; lists are used to implement message queues, supporting push and subscription; sets and ordered sets are used to store non-duplicate databases, supporting ranking and other functions.

[0089] The data analysis component is used to perform statistical analysis, trend analysis and data mining on the data to provide decision support. The data analysis component also supports batch processing and streaming processing and provides a flexible analysis model;

[0090] Among them, statistical analysis is to calculate the basic statistical indicators of average, maximum, minimum and standard deviation; trend analysis is to analyze the changing trend of data and predict future changes; data mining is to use machine learning algorithms to discover hidden patterns and associations in data; the data analysis component generates daily, weekly and monthly reports based on the results of statistical analysis, trend analysis and data mining, which is convenient for data visualization, and provides flexible analysis models for batch processing and streaming processing.

[0091] The anomaly detection component achieves accurate detection of data anomalies by building an adaptive spatiotemporal correlation model and combining deep fusion of multi-dimensional features;

[0092] like Figure 3 As shown, the process of the anomaly detection component performing data anomaly detection on meteorological data is as follows:

[0093] B21. Extract multi-dimensional features from meteorological data, including temperature, humidity, air pressure and wind speed data;

[0094] B22. Use deep neural networks to perform nonlinear fusion of multi-dimensional features to obtain high-level feature representation;

[0095] like Figure 4 As shown, the steps of nonlinearly fusing the multi-dimensional features are as follows:

[0096] B221. Receive multi-dimensional meteorological characteristic data through the input layer. The data expression is:

[0097] ;

[0098] B222, multi-layer nonlinear mapping is performed through hidden layers, and ReLU is used as the activation function to extract high-level features;

[0099] B223, output the fused feature F through the output layer;

[0100] B224. Dynamically adjust network weights based on the importance of features so that the adaptive spatiotemporal correlation model pays more attention to key features;

[0101] B225, L1 and L2 regularization are combined to prevent overfitting and improve the generalization ability of the adaptive spatiotemporal correlation model;

[0102] B23, dynamically divide the time and space windows and select the site data with the strongest correlation with the target site;

[0103] B24, establish an adaptive spatiotemporal parallel model between the target site and the associated sites;

[0104] The steps to establish the adaptive spatiotemporal correlation model are as follows:

[0105] B241. Dynamically adjust the time window size according to the time variation characteristics of the target site data. If the recent data fluctuates greatly, shorten the time window to improve sensitivity.

[0106] B242. Dynamically adjust the spatial window range according to the distribution density and relevance of the sites around the target site to ensure that the most relevant site is selected;

[0107] B243. Calculate the data correlation coefficient between the target site and the associated site within the time window, and select the site with high correlation as the associated site. The calculation formula is as follows:

[0108] , where N is the number of data points in the time window, For the The observed value of the target site at a certain moment, is the average value of the target site at the moment, For associated sites The observed value of the target site at a certain moment, is the average value of the associated site at the moment, is the data correlation coefficient between the target site and the associated site within the time window;

[0109] B25, calculating the difference between the target site data and the predicted value of the adaptive spatiotemporal correlation model to obtain anomaly scores;

[0110] The calculation formula for the anomaly score is as follows:

[0111] , where S is the anomaly score and y is the actual observed value. is the predicted value of the adaptive spatiotemporal correlation model;

[0112] B26. Use dynamic thresholds to adjust abnormality judgment criteria based on real-time data characteristics;

[0113] The dynamic threshold is to set the confidence level based on the real-time distribution of anomaly scores. , and then calculate the dynamic threshold. The calculation formula of the dynamic threshold is: ,in, is the mean of the anomaly scores, is the standard deviation of the anomaly score, is the quantile of the standard normal distribution, T is the dynamic threshold;

[0114] B27, adjust the parameters and thresholds of the adaptive spatiotemporal correlation model in real time according to the detection results, and continuously optimize the detection performance;

[0115] When adjusting the parameters and thresholds of the adaptive spatiotemporal correlation model in real time, the confidence level is dynamically adjusted based on the false alarm rate and missed alarm rate of historical detections. , when S>T, the data is judged to be abnormal and the calculation parameters need to be automatically updated;

[0116] The alarm notification component is responsible for sending alarm notifications when an abnormality is detected, and supports multiple notification methods;

[0117] The alarm notification component includes an alarm rule management unit, a multi-channel notification unit, and an alarm recording and tracking unit. The alarm rule management unit includes a rule configuration process and a rule engine process. The rule configuration process includes threshold alarm, trend alarm, and compound condition alarm. The threshold alarm is used to set the upper and lower limits of the indicator, and an alarm is triggered when the range is exceeded; the trend alarm is used to detect the trend of data changes, such as triggering an alarm when a sudden increase or decrease occurs; the compound condition alarm is to combine multiple conditions or indicators to trigger a complex alarm;

[0118] The rule engine performs real-time rule matching on streaming data and also supports custom rules for easy adjustment and optimization.

[0119] The multi-channel notification unit selects different notification methods and frequencies for warning according to the severity of the alarm. The notification methods include SMS, email and APP push. SMS is suitable for timely sending SMS information in case of emergency alarm, email is for sending detailed alarm information and reports, and APP push is for real-time push notification through mobile application. Users can set the receiving method and time period to avoid message interference;

[0120] The alarm recording and tracking unit is used to record the time, type and detailed information of the alarm, which is convenient for query during conditional retrieval, sorting and statistical analysis. When recording, the alarm recording and tracking unit first manages the alarm status, confirms whether the alarm is in the confirmed status, processing status or resolved status, assigns processing personnel, tracks the processing progress, and finally records the processing results.

[0121] Through the coordination of data preprocessing components, data storage components, data analysis components, anomaly detection components and alarm notification components, it integrates adaptive spatiotemporal parallel modeling, multi-dimensional feature fusion and dynamic threshold adjustment technology to monitor and identify abnormal data in real time, dynamically adjust the spatiotemporal correlation model parameters, adapt to the meteorological data characteristics of different regions and time periods, and use deep learning technology to integrate a variety of meteorological feature data to capture complex nonlinear relationships. It can also automatically adjust the anomaly detection threshold according to the real-time data distribution characteristics to improve detection accuracy. It is highly innovative and practical.

[0122] The web front-end module is used to provide users with a friendly interface and display data visualization results, which include map display, data reports and alarm information;

[0123] The web front-end module includes a geographic visualization component, a user management component and an interface management component. The geographic visualization component is used to provide map rendering and geographic information display functions, interact with the web front-end module to present visualization results, and support dynamic display of real-time data;

[0124] The geographic visualization component includes a Leaflet map rendering engine, Tiandi Map data service and visualization elements. The Leaflet map rendering engine is used to realize the visualization display of geographic information on the web front end. The geographic information includes the location, status and data of meteorological stations. The Tiandi Map data service integrates Tiandi Map's services through Leaflet plug-ins or custom layers to provide high-quality and diversified map data. The map data includes vector maps, image maps and topographic maps to enrich the map display effect.

[0125] Visual elements include point markers, lines and polygons, information pop-ups and interactive functions; point markers are used to mark the location of weather stations on the map, using different icons or colors to represent different states; lines and polygons are used to draw contour lines, area boundaries and paths, etc.; information pop-ups pop up detailed information when clicking on a station or area, and the detailed information includes real-time data and historical records; interactive functions are used to support map zooming, panning, layer switching and search positioning, etc.

[0126] The user management component is used to manage system users, roles and permissions, provide authentication and authorization services, and support multiple authentication methods;

[0127] The user management component includes account management and password policy management in the user management of the system, wherein account management is used to create, modify and delete user accounts, and password policy management is used to set password complexity, validity period and lock policy to enhance security; the user management component includes role definition, permission configuration and permission review in the role and permission management, wherein role definition is to define different roles according to positions or functions, and permission configuration is to assign permissions corresponding to roles to control access to system resources; permission review is to regularly review user permissions to prevent permission abuse; the user management component can also be used for single sign-on to support docking with the enterprise's unified authentication system;

[0128] The interface management component is used to provide an external API interface and support data collection and message integration functions.

[0129] The interface management component includes a collection interface and a message integration interface. The collection interface uses the common format of JSON and XML, defines a unified message structure, supports HTTP / HTTPS, MQTT and CoAP protocols, uses API keys and digital certificates to verify the identity of the data sender, and prevents eavesdropping and tampering by encrypting the transmission of sensitive data; the message integration interface includes a message sending interface, a message receiving interface, security and authentication, and interface management, wherein the message sending interface allows a third-party system to call the interface and send message notifications to platform users, supports ESTful API, SOAP, etc. The message receiving interface is used to subscribe to platform messages through a third-party system, receive alarm and notification information, supports WebSocket, MQTT, HTTP long connection, etc., and security and authentication are used for OAuth2.0 authentication to ensure the legality of the identity of the interface caller, and prevent unauthorized calls by configuring the access rights of the interface; interface management is used to set the calling frequency and traffic limit of the interface, so that third-party systems need to register and be approved before they can use the interface to prevent abuse;

[0130] Through the cooperation of data acquisition module, background service module and web front-end module, the temperature, humidity, pressure and rainfall data of ground meteorological station are monitored in real time, and the collected data are pre-processed, analyzed and anomaly detected through the background service module. During anomaly detection, an adaptive spatiotemporal correlation model is constructed, combined with the deep fusion of multi-dimensional feature data, to achieve accurate detection of data anomalies, reduce data interruption or loss during operation, and improve the accuracy of data collection.

[0131] Based on the above-described ground weather station network business management system, the present invention also provides a ground weather station network business management method, such as Figure 5 As shown, the following steps are included:

[0132] S1, receiving multi-dimensional data of temperature, pressure, humidity and rainfall from the weather station in real time through the data acquisition module;

[0133] S2. Use data preprocessing components to clean, convert and verify the collected data to ensure data quality;

[0134] S3, storing the preprocessed data in the time series database, relational database and cache database to facilitate subsequent data analysis and use;

[0135] S4. Establish data index and backup mechanism to ensure data security and accessibility;

[0136] S5. Use data analysis components to perform statistical analysis, trend analysis and data mining on the data to reveal the inherent laws and trends of meteorological data;

[0137] S6. Build an adaptive spatiotemporal correlation model through the anomaly detection component, and accurately detect data anomalies by combining the deep fusion of multi-dimensional features; when an anomaly is detected, trigger the alarm notification component to promptly notify relevant personnel in a variety of ways;

[0138] S7. Use the web front-end module to provide users with a friendly interface to display the visualization results of the data. Users can query, analyze and export data through the interface to improve the convenience and efficiency of data use.

[0139] Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent substitutions for some of the technical features therein; and these modifications or substitutions do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A ground weather station network business management system, characterized in that: include: A data acquisition module, which is responsible for collecting multi-dimensional data of temperature, pressure, humidity and rainfall from multiple ground meteorological stations in real time; A backend service module, which includes a data preprocessing component, a data storage component, a data analysis component, an anomaly detection component, and an alarm notification component that supports multiple detection algorithms and flexibly configures alarm rules; The data preprocessing component is used to perform preprocessing operations such as cleaning, format conversion and verification on the collected data to ensure data quality. The data preprocessing component can expand preprocessing rules and support plug-in rule addition; The data storage component is used to store the processed data in a time series database, a relational database and a cache database; The data analysis component is used to perform statistical analysis, trend analysis and data mining on the data to provide decision support. The data analysis component also supports batch processing and streaming processing and provides a flexible analysis model; The anomaly detection component achieves accurate detection of data anomalies by building an adaptive spatiotemporal correlation model and combining deep fusion of multi-dimensional features; The alarm notification component is responsible for sending alarm notifications when an abnormality is detected, and supports multiple notification methods; A web front-end module, which is used to provide a user-friendly interface and display data visualization results, including map display, data reports and alarm information; The backend service module is electrically connected to the data acquisition module and the web front-end module respectively.

2. A ground weather station network business management system according to claim 1, characterized in that: The data acquisition module includes a sensor unit and a data transmission unit, and the sensor unit includes a temperature sensor, a humidity sensor, a meteorological pressure sensor and a rainfall sensor; The temperature sensor is used to sense the temperature data of the weather station in real time; the humidity sensor is used to collect the humidity data of the weather station in real time, the meteorological pressure sensor is used to sense the pressure of the atmosphere on the unit area of ​​the earth's surface, and the rainfall sensor is used to collect rainfall data; The data transmission unit converts the received sensor data into digital signals, and then transmits the converted digital signals to the background service module via wireless or limited transmission for data processing.

3. A ground weather station network business management system according to claim 2, characterized in that: The process of the data preprocessing component performing preprocessing operations on data is as follows: B11. Process missing values, abnormal values, and duplicate data in the received data; the missing values ​​are processed by deleting them or filling them with the mean; duplicate data are directly deleted; There are three situations for outlier processing: 1) If the number of outliers is small and does not affect the overall distribution of the data, delete them directly; 2) If the outlier is due to data entry error or measurement error, correct it to the correct value; 3) If the outlier contains important information, it will be retained; B12. Convert data in different formats into internal standard formats; B13. Verify the integrity and rationality of the data to filter out invalid data.

4. A ground weather station network business management system according to claim 3, characterized in that: The statistical analysis in the data analysis component is to calculate the basic statistical indicators of average value, maximum value, minimum value and standard deviation; Trend analysis is to analyze the changing trends of data and predict future changes; Data mining is the use of machine learning algorithms to discover hidden patterns and associations in data; The data analysis component generates daily, weekly and monthly reports based on the results of statistical analysis, trend analysis and data mining, which facilitates data visualization and provides flexible analysis models for batch processing and streaming processing.

5. A ground weather station network business management system according to claim 4, characterized in that: The process of the anomaly detection component performing data anomaly detection on meteorological data is as follows: B21. Extract multi-dimensional features from meteorological data, including temperature, humidity, air pressure and wind speed data; B22. Use deep neural networks to perform nonlinear fusion of multi-dimensional features to obtain high-level feature representation; B23, dynamically divide the time and space windows and select the site data with the strongest correlation with the target site; B24, establish an adaptive spatiotemporal parallel model between the target site and the associated sites; B25, calculating the difference between the target site data and the predicted value of the adaptive spatiotemporal correlation model to obtain anomaly scores; The calculation formula of anomaly score is as follows: , where S is the anomaly score and y is the actual observed value. is the predicted value of the adaptive spatiotemporal correlation model; B26. Use dynamic thresholds to adjust abnormality judgment criteria based on real-time data characteristics; The dynamic threshold is to set the confidence level based on the real-time distribution of anomaly scores. , and then calculate the dynamic threshold. The calculation formula of the dynamic threshold is: ,in, is the mean of the anomaly scores, is the standard deviation of the anomaly score, is the quantile of the standard normal distribution, T is the dynamic threshold; B27. Adjust the parameters and thresholds of the adaptive spatiotemporal correlation model in real time according to the detection results to continuously optimize the detection performance.

6. A ground weather station network business management system according to claim 5, characterized in that: The steps of nonlinearly fusing the multi-dimensional features are as follows: B221. Receive multi-dimensional meteorological characteristic data through the input layer. The data expression is: ； B222, multi-layer nonlinear mapping is performed through hidden layers, and ReLU is used as the activation function to extract high-level features; B223, output the fused feature F through the output layer; B224. Dynamically adjust network weights based on the importance of features so that the adaptive spatiotemporal correlation model pays more attention to key features; B225, use a combination of L1 and L2 regularization to prevent overfitting.

7. A ground weather station network business management system according to claim 6, characterized in that: The steps to establish the adaptive spatiotemporal correlation model are as follows: B241, dynamically adjust the time window size according to the time variation characteristics of the target site data; B242. Dynamically adjust the spatial window range according to the distribution density and relevance of the sites around the target site to ensure that the most relevant site is selected; B243. Calculate the data correlation coefficient between the target site and the associated site within the time window, and select the site with high correlation as the associated site. The calculation formula is as follows: , where N is the number of data points in the time window, For the The observed value of the target site at a certain moment, is the average value of the target site at the moment, For associated sites The observed value of the target site at a certain moment, is the average value of the associated site at the moment, is the data correlation coefficient between the target site and the associated site within the time window.

8. A ground weather station network business management system according to claim 7, characterized in that: The web front-end module includes a geographic visualization component, a user management component and an interface management component. The geographic visualization component is used to provide map rendering and geographic information display functions, interact with the web front-end module to present visualization results, and support dynamic display of real-time data; The user management component is used to manage system users, roles and permissions, provide authentication and authorization services, and support multiple authentication methods; The interface management component is used to provide an external API interface and support data collection and message integration functions.

9. A ground meteorological station network business management system according to claim 8, characterized in that: The geographic visualization component includes a Leaflet map rendering engine, a Tiandi map data service, and visualization elements. The Leaflet map rendering engine is used to realize the visualization display of geographic information on the web front end. The geographic information includes the location, status, and data of the meteorological station. The Tiandi Map data service integrates Tiandi Map's services through Leaflet plug-ins or custom layers to provide high-quality and diverse map data, including vector maps, image maps, and topographic maps, enriching the map display effect.

10. A method for managing ground meteorological station network services, based on a ground meteorological station network service management system according to any one of claims 1 to 9, characterized in that: The steps include: S1, receiving multi-dimensional data of temperature, pressure, humidity and rainfall from the weather station in real time through the data acquisition module; S2. Use data preprocessing components to clean, convert and verify the collected data to ensure data quality; S3, storing the preprocessed data in the time series database, relational database and cache database; S4. Establish data index and backup mechanism to ensure data security and accessibility; S5. Use data analysis components to perform statistical analysis, trend analysis and data mining on the data to reveal the inherent laws and trends of meteorological data; S6. Build an adaptive spatiotemporal correlation model through the anomaly detection component, and accurately detect data anomalies by combining the deep fusion of multi-dimensional features; when an anomaly is detected, trigger the alarm notification component to promptly notify relevant personnel in a variety of ways; S7. Use the web front-end module to provide users with a friendly interface and display the visualization results of the data.

Images