Mobile underway analysis visualization system

Abstract

The invention relates to the technical field of underway systems, in particular to a mobile underway analysis visualization system. According to the technical scheme, the system comprises a data acquisition layer, a data processing and analysis layer, a software integration layer, a function application layer and a user interaction layer, the data acquisition layer is used for acquiring key data from a professional data source, and meteorological data are acquired by means of a professional meteorological API interface, a platform vehicle-mounted ultrasonic meteorological sensor and a vehicle-mounted three-dimensional ultrasonic anemograph; gas concentration data of greenhouse gas is collected by a Picarro analyzer and a PRI analyzer in real time, GPS data is obtained through a vehicle-mounted high-precision RTK multi-frequency Beidou module, and map data is called from a map service API. According to the invention, meteorological and gas concentration multi-source data can be comprehensively collected and integrated, the advanced technology is adopted to efficiently process and analyze the data for real-time collection, real-time analysis and real-time display, the real-time performance is ensured, accurate prediction is realized, various requirements are met, and the operation is convenient; and an advanced architecture and a stable mechanism ensure reliable operation of the system.

Description

Technical Field

[0001] The present invention relates to the technical field of navigation systems, and in particular to a mobile navigation analysis visualization system. Background Art

[0002] With the rapid development of industrialization and urbanization, environmental pollution is a growing concern. Monitoring atmospheric emissions of various gases (such as methane, carbon dioxide, and carbon monoxide) and understanding meteorological conditions in real time have become crucial. Traditional fixed-point monitoring methods have limited coverage and cannot fully reflect the distribution and dynamic changes of pollutants in a region. To more accurately understand atmospheric environmental quality, a mobile monitoring technology is needed.

[0003] Mobile navigation generates massive amounts of meteorological data, including gas and water vapor concentration data, and GPS data. This data is diverse, real-time, and massive. Effectively cleaning, converting, and deeply analyzing this data to extract valuable information is a major challenge. Traditional data processing methods often struggle to meet real-time and accuracy requirements, and are unable to provide timely support for decision-making.

[0004] In the field of environmental monitoring, data visualization and user interaction are crucial for understanding and applying data. However, some existing monitoring systems lack visualization capabilities, lack user-friendly interfaces, and lack rich functionality, making it difficult to intuitively display data trends and spatial distribution. This makes data querying and analysis cumbersome for users, making it difficult to quickly obtain the information they need.

[0005] A complete mobile tracking analysis system involves multiple data sources, multiple functional modules, and diverse technical architectures. Effectively integrating these components and enabling data exchange and collaboration between modules is a complex issue. Data formats and interface standards may differ across data sources, making data integration and sharing difficult. Furthermore, system stability and reliability must be guaranteed to ensure proper operation in complex mobile monitoring environments.

[0006] In summary, this application proposes a mobile navigation analysis visualization system. Summary of the Invention

[0007] The purpose of the present invention is to address the problem in the background technology that the existing navigation analysis system does not effectively integrate multiple data sources, multiple functional modules and different technical architectures, and to propose a mobile navigation analysis visualization system with real-time collection, real-time analysis and real-time display.

[0008] The technical solution of the present invention is a mobile cruise analysis and visualization system, which includes a data acquisition layer: key data is obtained from professional data sources, including meteorological data obtained by using a professional meteorological API interface and a platform-mounted ultrasonic meteorological sensor and a vehicle-mounted three-dimensional ultrasonic anemometer. Greenhouse gas concentration data is collected in real time by a Picarro analyzer and a PRI analyzer. GPS data is obtained through a vehicle-mounted high-precision RTK multi-frequency Beidou module. Map data is called from a map service API.

[0009] Data processing and analysis layer: This layer performs multi-dimensional processing and in-depth analysis of collected data. It first uses data cleaning algorithms to remove noise, outliers, and duplicate data, then performs format unification and standardization conversion. Finally, it conducts in-depth mining and analysis based on mathematical models and intelligent algorithms, while also employing parallel computing and asynchronous processing technologies.

[0010] Software integration layer: adopts a front-end and back-end separation architecture. The front-end is built based on the front-end framework and realizes real-time data communication with the back-end through WebSocket. The back-end is built with a server-side framework and provides RESTful API services. It is responsible for data storage, management, processing and push, and performs data integration and interaction.

[0011] Functional application layer: Integrates multiple special function modules, including toolkit module, comprehensive weather forecast module, multi-gas concentration time series monitoring module, wind speed-concentration correlation analysis module, multi-gas real-time synchronous monitoring module, methane carbon isotope characteristic analysis module, GPS and water vapor data integrated monitoring module, payload platform information module, multi-gas collaborative analysis module, and cruise real-time analysis module;

[0012] User interaction layer: Design a user-friendly interface layout and distribute various functional modules in different areas, including real-time data viewing, historical data query, parameter setting, and report generation operations. The interface can be dynamically updated in real time according to user operations.

[0013] Optionally, the toolkit module includes:

[0014] Map multi-dimensional display switching unit: provides switching between plane, 3D, and satellite map modes. Users can quickly switch map perspectives with one click to understand the geographic information of the monitored area. The map multi-dimensional display switching unit uses the WebGL 3D rendering engine and the Cesium geospatial framework to achieve terrain elevation mapping, and dynamically loads map tiles of different precisions through the quadtree LOD algorithm.

[0015] Intelligent Graphic Display Switching Unit: This unit supports intelligent switching of graphical display modes such as concentration heat maps, trajectory roadmaps, and time series graphs. Users can select the display format to analyze data based on their needs. The intelligent graphic display switching unit has a built-in image recognition module based on convolutional neural networks, which automatically identifies the user's gesture trajectory and matches the optimal visualization format.

[0016] Dynamic threshold customization unit: Users can dynamically set multiple gas concentration threshold intervals based on actual conditions and adjust the color identification of each interval to provide personalized visualization of different concentration areas. The dynamic threshold customization unit integrates a fuzzy PID control algorithm, automatically corrects the threshold interval boundary values ​​based on historical concentration fluctuations, and generates a color gradient dynamic mapping function.

[0017] Flexible data source switching unit: users can switch between data provided by Picarro and PRI and compare monitoring results from different data sources;

[0018] Query mode multi-switch unit: provides real-time data viewing, historical data query by time range, and specific event data filtering query;

[0019] External resource quick link unit: integrates external resource links of on-site picture libraries and satellite image libraries related to the monitoring area.

[0020] Optionally, the comprehensive weather forecast module includes:

[0021] Real-time weather station information unit: displays detailed weather parameters provided by the on-board weather station in real time, including temperature, air pressure, humidity, wind direction, and wind speed, providing meteorological background information for mobile navigation monitoring;

[0022] Hourly Weather Forecast Unit: Based on a professional weather forecast model, it provides hourly weather forecast information for the next 24 hours, including weather conditions, temperature changes, and wind speed and direction trends, helping users plan monitoring tasks in advance. The hourly weather forecast unit integrates the WRF mesoscale numerical forecast model with the LSTM time series prediction algorithm, and performs multi-source data assimilation through Kalman filtering.

[0023] Multi-day weather trend forecast unit: displays the weather trend forecast for the next 7 days, including temperature range, precipitation probability, and weather type. The multi-day weather trend forecast unit adopts ensemble forecast technology, integrates the output results of the three major global models ECMWF, GFS, and CMA, and uses the Monte Carlo method to generate probability distribution cloud maps.

[0024] Optionally, the data cleaning algorithm of the data processing and analysis layer adopts an anomaly detection model based on machine learning, combined with statistical analysis methods, to automatically identify and remove outliers and noise in the data; data standardization conversion adopts international standard units and time formats; mathematical models combine fluid mechanics and chemical kinetics to simulate the process of gas diffusion and chemical reaction;

[0025] The anomaly detection model adopts an improved IsolationForest algorithm, introduces a time sliding window mechanism and spatial density clustering features, and constructs a three-dimensional spatiotemporal anomaly detection framework. The fluid mechanics model adopts the large eddy simulation (LES) method, coupled with the OpenFOAM solver to calculate pollutant diffusion trajectories in real time. The chemical kinetics module integrates the gas-phase chemical mechanism in the CMAQ air quality model.

[0026] Optionally, the front-end framework of the software integration layer adopts virtual DOM technology and component development mode; WebSocket communication adopts encrypted transmission protocol, and the back-end server framework adopts distributed storage system and cache technology to optimize data storage and access performance;

[0027] The virtual DOM adopts a difference algorithm optimization strategy and combines WebAssembly to accelerate the DOM tree comparison process; the encrypted transmission protocol adopts the national secret SM4 algorithm and TLS1.3 dual-channel encryption mechanism; the distributed storage system is based on the Ceph object storage architecture design, and adopts the CRUSH algorithm to achieve automatic data sharding and load balancing. The cache layer adopts the Redis cluster architecture accelerated by RDMA.

[0028] Optionally, the top navigation bar of the user interaction layer interface integrates common function entrances, the sidebar displays toolkits and query options, the middle main area is a map and data display area, and the bottom status bar displays system status and key data indicators in real time; the interactive components are designed with icons and buttons, and support mouse hover prompts and shortcut key operations;

[0029] The interface layout adopts a responsive grid system and uses the Bezier curve algorithm to achieve smooth deformation transitions of visual elements; shortcut key operations support user-defined key combinations and integrate the Levenshtein distance algorithm to achieve fault-tolerant command recognition; the status bar uses Microcharts micro-visualization technology and dynamically renders data indicators through a radial gradient shading algorithm.

[0030] Optionally, communications between system layers and with external data sources utilize HTTP / 2 protocol and MQTT message transmission protocol, including but not limited to TLS encryption. The system has a network fault detection and automatic recovery mechanism, which can quickly switch to an alternative network or retry the connection when a network anomaly occurs.

[0031] The network fault detection mechanism combines bidirectional heartbeat detection packets with BGP route monitoring, and automatically enables MPTCP multipath transmission when the packet loss rate is greater than 15%. The retry connection strategy adopts an exponential backoff algorithm and a reinforcement learning strategy optimizer to dynamically adjust the retry interval. The backup network switching uses software-defined radio (SDR) technology to achieve seamless switching of 4G / 5G / WiFi multimodes.

[0032] Optionally, the system is built on a hybrid cloud architecture, combining public and private clouds for data storage and processing; blockchain technology is used to encrypt and store key data and trace its origin; at the same time, the system has a disaster recovery backup mechanism and regularly conducts data backup and recovery drills;

[0033] The hybrid cloud architecture adopts Kubernetes federated cluster management, public cloud deploys stateless computing nodes, and private cloud deploys sensitive data storage; the blockchain adopts a sharded consortium chain structure and uses zero-knowledge proof to make data available but invisible; disaster recovery backup adopts erasure coding technology and designs a (10,6) Reed-Solomon coding scheme to achieve distributed redundant storage.

[0034] Optional features include intelligent decision-making support, which automatically generates gas pollution warnings, monitoring route optimization suggestions, and equipment maintenance reminders through in-depth analysis of historical and real-time data, combined with preset business rules and machine learning models.

[0035] The monitoring route optimization suggestion adopts an improved ant colony algorithm, introduces a meteorological factor cost function and a real-time road condition weight matrix; the equipment maintenance reminder module integrates vibration spectrum analysis and infrared thermal imaging feature extraction, and constructs a deep residual network to achieve fault prediction; the business rule engine is optimized using the Rete algorithm.

[0036] Optionally, the gas pollution warning in the intelligent decision-making assistance function is constructed using a deep learning model based on a spatiotemporal convolutional neural network. The model uses historical and real-time gas concentration data, meteorological data, and geographic information data as input, and achieves accurate prediction and warning of gas pollution conditions by extracting and analyzing the features of the data in spatial and temporal dimensions;

[0037] The spatiotemporal convolutional network adopts a three-dimensional dilated convolution structure, designs a causal convolution layer in the time dimension, and embeds a graph attention mechanism in the spatial dimension; the model training adopts a curriculum learning strategy, first pre-training ERA5 and then migrating to local monitoring data; the input data preprocessing uses the Kriging spatial interpolation method to fill in missing values, and the wavelet denoising algorithm is used to eliminate high-frequency interference.

[0038] Compared with the prior art, this application has at least one of the following beneficial technical effects:

[0039] It can obtain meteorological, gas concentration, GPS, map and other data from multiple professional sources, integrate them in a unified manner, provide a comprehensive and accurate basis for analytical applications, and facilitate comprehensive analysis of the relationship between different factors.

[0040] By using advanced technologies to quickly clean, convert and mine data, and relying on customized models and algorithms to accurately simulate and predict, parallel and asynchronous processing ensures real-time performance and provides timely analysis results.

[0041] By integrating multiple functional modules, it meets diverse needs and has good module collaboration. The user-friendly interface and simple interactive components support multiple operations and are updated in real time, improving convenience and visualization.

[0042] The front-end and back-end separation architecture enables efficient integrated interaction, with excellent scalability and compatibility. It uses secure communication protocols and has fault detection, recovery, and disaster recovery mechanisms to ensure stable system operation.

[0043] It has real-time collection, real-time analysis and real-time visualization functions, can efficiently process and present data, and provide users with instant and accurate information support.

[0044] The present invention can comprehensively collect and integrate multi-source data on meteorology and gas concentrations, laying a solid foundation for analysis; it uses advanced technology to efficiently process and analyze data, ensuring real-time and accurate predictions; it has rich functional modules and a friendly interactive interface to meet diverse needs and is easy to operate; its advanced architecture and stability mechanism ensure reliable system operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 This is a principle block diagram of a mobile navigation analysis and visualization system. DETAILED DESCRIPTION

[0046] The technical solution of the present invention is further described below with reference to the accompanying drawings and specific embodiments.

[0047] Example 1

[0048] like Figure 1 As shown, the present invention proposes a mobile navigation analysis visualization system, which is described in detail below.

[0049] 1. Overview of the overall system architecture

[0050] This mobile navigation analysis and visualization system primarily consists of a data acquisition layer, a data processing and analysis layer, a software integration layer, a functional application layer, and a user interaction layer. Each layer works closely together to achieve comprehensive processing and visualization of mobile navigation data. The following details the implementation of each layer.

[0051] 2. Data Collection Layer

[0052] (1) Data source and collection method

[0053] Meteorological Data: Meteorological data is obtained by calling professional meteorological API interfaces and the platform's onboard ultrasonic meteorological sensors and three-dimensional ultrasonic anemometers. We collaborate with authoritative domestic meteorological data service providers and follow their prescribed API call rules. The system's backend server periodically sends requests to the API. Request parameters include the geographic location (latitude and longitude) and time range of the monitored area to obtain real-time meteorological data for that area, such as temperature, pressure, humidity, wind direction, and wind speed.

[0054] Gas and water vapor concentration data: High-precision Picarro and PRI analyzers are used for real-time data collection. Installed on a mobile device, these analyzers measure the concentrations of various gases (such as CH4, CO2, C2H6, CO, N2O, etc.) and water vapor in the surrounding environment at a preset sampling frequency (e.g., once per second). The data is then transmitted to a data acquisition server via serial communication or a network interface.

[0055] GPS Data: Utilizes a high-precision, multi-frequency RTK BeiDou module mounted on a vehicle to obtain real-time geographic location information for mobile devices. The GPS device uses satellite positioning technology to determine the current latitude and longitude, altitude, and speed, and transmits this data to a data collection server via Bluetooth or a serial port.

[0056] Map data: Call map data from the map service API. For example, using the API of a well-known domestic map service provider, call the basic base map data, road data, terrain data, etc. according to the system requirements to display the geographical location and route information of the mobile navigation in the system.

[0057] (2) Data collection frequency and synchronization

[0058] To ensure data accuracy and consistency, the collection frequency of each data source must be appropriately set. Meteorological and map data can be collected regularly based on their update frequencies, such as every 10 minutes for meteorological data and daily for map data. Gas and water vapor concentration data and GPS data must be collected in real time to reflect real-time changes during navigation. The system also timestamps the collected data to ensure synchronization across different data sources, facilitating subsequent processing and analysis.

[0059] 3. Data Processing and Analysis Layer

[0060] (1) Data cleaning

[0061] Data cleaning is performed using a machine learning-based anomaly detection model combined with statistical analysis methods. For gas vapor concentration data and GPS data, the standard deviation method is first used to determine the normal fluctuation range of the data. Data outside this range is marked as outliers. The data is then further screened using an anomaly detection model based on the isolation forest algorithm. This model can automatically identify abnormal patterns in the data and improve the accuracy of outlier detection. For duplicate data, key feature fields (such as timestamps, longitude and latitude, and gas concentration values) are compared to determine if the data is consistent and duplicates are deleted.

[0062] (2) Data standardization conversion

[0063] Convert collected data to international standard units and time formats. Convert gas concentration data to ppm (parts per million), temperature data to degrees Celsius, and time formats to ISO 8601 standards. By writing data conversion scripts, batch processing of data from different data sources ensures data consistency and comparability.

[0064] (3) In-depth analysis

[0065] In-depth analysis is performed based on customized mathematical models and intelligent algorithms. For gas diffusion analysis, a gas diffusion model is developed by combining fluid mechanics and chemical kinetics. This model considers factors such as meteorological conditions (wind speed, direction, and temperature), topography, and the physical and chemical properties of gases. Numerical simulations are used to predict the diffusion range and concentration distribution of gases under different environments. Machine learning algorithms (neural networks and decision trees) are also used to classify and predict gas concentration data, analyzing trends and influencing factors.

[0066] (4) Real-time processing technology

[0067] Efficient parallel computing and asynchronous processing technologies are used to ensure real-time data processing. Multi-threaded programming technology is used to assign data processing tasks to multiple threads for parallel execution, thereby increasing the speed of data processing. Tasks such as data cleaning, standardization conversion, and in-depth analysis are assigned to different threads for processing. At the same time, asynchronous I / O technology is used to avoid thread blocking during data acquisition and storage, thereby improving the system's concurrent processing capabilities. The anomaly detection model uses an improved IsolationForest algorithm, introduces a time sliding window mechanism and spatial density clustering features, and constructs a three-dimensional spatiotemporal anomaly detection framework. The fluid mechanics model uses the large eddy simulation (LES) method, coupled with the OpenFOAM solver to calculate pollutant diffusion trajectories in real time. The chemical kinetics module integrates the gas-phase chemical mechanism in the CMAQ air quality model.

[0068] 4. Software Integration Layer

[0069] (1) Front-end and back-end separation architecture

[0070] The front-end is built using a front-end framework (such as Vue.js), leveraging its virtual DOM technology and component-based development model to improve interface rendering efficiency and maintainability. The front-end page establishes a real-time connection with the back-end via WebSocket, enabling real-time data push and interaction. The back-end is built using a server-side framework (such as Django), providing a rich set of RESTful API services for data storage, management, processing, and push.

[0071] (2) Front-end implementation details

[0072] In front-end development, we used the VueCLI tool to quickly build the project structure and encapsulate different functional modules into independent components, such as map components, chart components, and toolkit components. We implemented routing management with VueRouter, switching between different page views based on user actions. We used the ECharts library to create various charts, such as line charts, bar charts, and heat maps, to demonstrate gas concentration trends and distribution. We also used the WebSocket protocol for real-time communication with the backend, monitoring data update messages sent by the backend and promptly updating the front-end interface display.

[0073] (3) Backend implementation details

[0074] The backend uses the ORM (object-relational mapping) function of the Django framework to interact with the database to implement data storage and query. The DjangoRESTframework is used to build a RESTfulAPI to provide a unified data interface for the front end. For real-time data push, DjangoChannels is used in combination with WebSocket to implement asynchronous message processing. In terms of data storage, a distributed storage system (such as HBase or MongoDB) is used to store large amounts of historical data, and Redis is used as a cache database to improve data access speed. The virtual DOM adopts a difference algorithm optimization strategy, combined with WebAssembly to accelerate the DOM tree comparison process; the encrypted transmission protocol adopts the national secret SM4 algorithm and TLS1.3 dual-channel encryption mechanism; the distributed storage system is based on the Ceph object storage architecture design, and the CRUSH algorithm is used to achieve automatic data sharding and load balancing. The cache layer adopts the RDMA-accelerated Redis cluster architecture.

[0075] 5. Functional Application Layer

[0076] (1) Toolkit module

[0077] Multi-dimensional Map Display Switching Unit: A switch button is set up on the front-end page. When the user clicks the button, the map service API calls are used to switch between 2D, 3D, and satellite map modes. When the user clicks the "3D Map" button, the front-end sends a request to the map service API to obtain 3D map data and display it on the page. The Multi-dimensional Map Display Switching Unit uses the WebGL 3D rendering engine and the Cesium geospatial framework to achieve terrain elevation mapping, and dynamically loads map tiles of varying precision using a quadtree LOD algorithm.

[0078] Intelligent Graphic Display Switching Unit: This unit provides a switch button for various graphical display modes. When a user clicks a button, the front-end retrieves the corresponding data from the back-end and uses the ECharts library to draw different types of charts based on the selected display mode. When the user selects "Concentration Heat Map," the front-end sends a request to the back-end for gas concentration data, then uses ECharts' heat map component to draw a concentration heat map. The Intelligent Graphic Display Switching Unit includes a built-in image recognition module based on a convolutional neural network, which automatically identifies the user's gesture trajectory and matches it with the optimal visualization format.

[0079] Dynamic Threshold Customization Unit: This unit provides a threshold setting interface on the front-end. Users can set multiple gas concentration threshold intervals using input boxes or sliders, and select color codes for each interval. The front-end sends the user-defined threshold information to the back-end, which processes the gas concentration data based on these thresholds and displays different concentration areas on a map using different colors. The Dynamic Threshold Customization Unit integrates a fuzzy PID control algorithm to automatically adjust the threshold interval boundaries based on historical concentration fluctuations and generate a dynamic color gradient mapping function.

[0080] Flexible data source switching unit: Set up a data source switching button. When the user clicks the button, the front-end sends a request to the back-end to switch the data source. Based on the user's request, the back-end obtains data from the corresponding data source (such as a Picarro analyzer or PRI analyzer) and returns the data to the front-end for display.

[0081] Query Mode Multi-Switch Unit: This unit provides buttons for switching between various query modes and a query condition input box. After the user selects a query mode and enters query conditions, the front-end sends the query request to the back-end. The back-end then filters the database for matching data based on the query conditions and returns the data to the front-end for display. For example, if the user selects the "Historical Data by Time Range" mode and enters a time range for the query, the back-end will query the database for the corresponding gas concentration and GPS data based on that time range.

[0082] External Resource Quick Link Unit: Set up an external resource link button on the front-end page. When the user clicks the button, the front-end opens the corresponding external resource link through the browser window object. For example, when the user clicks the "On-site Photo Gallery" button, the front-end opens the link to the webpage storing the on-site photos.

[0083] (2) Comprehensive weather forecast module

[0084] Real-time Weather Station Information Submodule: The backend periodically retrieves real-time weather data from the weather station and stores it in a database. The frontend calls the backend API to retrieve real-time weather data and displays it on the page. A specific area on the page displays the current temperature, pressure, humidity, wind direction, and wind speed.

[0085] Hourly Weather Forecast Submodule: The backend uses a specialized weather forecast model to generate hourly weather forecasts for the next 24 hours and stores the results in a database. The frontend accesses hourly forecast data by calling the backend's API and uses the ECharts library to draw line charts showing temperature trends, wind speed, and wind direction. This hourly forecast unit integrates the WRF mesoscale numerical forecast model with the LSTM time series prediction algorithm, using Kalman filtering for multi-source data assimilation.

[0086] Multi-Day Weather Trend Forecast Submodule: The backend obtains weather trend forecast data for the next seven days or longer and stores it in a database. The frontend accesses this data by calling the backend API and displays temperature range, precipitation probability, and weather type information on the page in a table or chart. The multi-day weather trend forecast unit uses ensemble forecasting technology, integrating the output of the three major global models: ECMWF, GFS, and CMA, and applying Monte Carlo methods to generate probability distribution cloud maps.

[0087] (3) Other functional modules

[0088] The multi-gas concentration time-series monitoring module, wind speed-concentration correlation analysis module, multi-gas real-time synchronous monitoring module, methane carbon isotope characteristic analysis module, and GPS and water vapor data integrated monitoring module are implemented in a similar manner. The backend is responsible for data processing and analysis, while the frontend is responsible for data display. In the multi-gas concentration time-series monitoring module, the backend performs time series analysis on the concentration data of different gases, calculates statistical indicators such as the rate of change and average value of the concentration, and stores the analysis results in the database. The frontend obtains the analysis results by calling the backend API and uses the ECharts library to draw time series graphs to display the changing trends of gas concentrations.

[0089] 6. User Interaction Layer

[0090] (1) Interface layout design

[0091] Adopting a modular design concept, the interface is divided into a top navigation bar, a sidebar, a middle main area, and a bottom status bar. The top navigation bar integrates commonly used function entrances, system menus, data queries, and help, and uses drop-down menus or button groups to quickly access functions. The sidebar displays toolkits and query options, and the buttons and controls are reasonably spaced, with intuitive icons and text prompts for their functions. The middle main area is the map and data display area, which is used to display mobile navigation map information, gas concentration data, and charts. The bottom status bar displays the system status and key data indicators, data collection status, network connection status, and current gas concentration values ​​in real time. The interface layout adopts a responsive grid system, and realizes smooth deformation transitions of visual elements based on the Bezier curve algorithm; shortcut key operations support user-defined key combinations, and integrate the Levenshtein distance algorithm to achieve fault-tolerant recognition of instructions; the status bar adopts Microcharts micro-visualization technology, and dynamically renders data indicators through a radial gradient shading algorithm.

[0092] (2) Interactive component design

[0093] Interactive components feature simple and intuitive icons and buttons, supporting mouse hover prompts and shortcut key operations. Map action buttons (zoom in, zoom out, pan) are represented by familiar icons, and hovering the mouse over a button displays a description of its function. Shortcut keys are also set for common actions, allowing map panning to be performed using the keyboard's arrow keys, improving user convenience.

[0094] (3) Real-time update mechanism

[0095] When a user performs an operation (querying data, switching map modes, setting thresholds), the front-end sends the operation information to the back-end via WebSocket or AJAX requests. After receiving the request, the back-end performs the corresponding data processing and query, and returns the processing results to the front-end. The front-end updates the interface content based on the returned data, achieving real-time dynamic updates of the interface. When the user enters query conditions in the query mode multi-switch unit and submits a query request, the front-end sends the query request to the back-end, which filters out the data that meets the conditions from the database and returns it to the front-end. The front-end uses the new data to update the tables or charts on the page.

[0096] 7. System Communication and Security

[0097] (1) Communication Protocol

[0098] An efficient and secure communication protocol stack is used for communication between each system layer and with external data sources. TLS-encrypted HTTP / 2 is used for data transmission between the data acquisition layer and the data processing and analysis layer, ensuring security and efficiency during data transmission. WebSocket is used for real-time communication between the front-end and back-end, while the MQTT messaging protocol is used for message publishing and subscription, improving the system's concurrent processing capabilities and message delivery reliability.

[0099] (2) Network Fault Detection and Recovery

[0100] The system has an intelligent network fault detection and automatic recovery mechanism. During the data transmission process, the network connection status is detected by regularly sending heartbeat packets. When a network failure is detected, the system automatically attempts to reconnect and records the fault information. If the primary network connection fails, the system will quickly switch to the backup network to ensure the continuity of data transmission. The network fault detection mechanism combines two-way heartbeat detection packets with BGP route monitoring. When the packet loss rate is >15%, MPTCP multi-path transmission is automatically enabled; the retry connection strategy uses an exponential backoff algorithm and a reinforcement learning strategy optimizer to dynamically adjust the retry interval; the backup network switching uses software-defined radio (SDR) technology to achieve seamless switching of 4G / 5G / WiFi multi-mode.

[0101] (3) Data security and disaster recovery

[0102] Blockchain technology is used to encrypt, store, and trace key data. Data is hashed and stored on the blockchain network, ensuring immutability and traceability. The system also features a robust disaster recovery mechanism that regularly backs up database data and stores the backup data in a remote data center. In the event of a system failure, data can be quickly restored, ensuring data security and business continuity in even the most challenging situations.

[0103] (4) Multilingual Support and Internationalization

[0104] The system supports a multilingual interface using internationalized resource file management. Each language corresponds to a separate resource file, which contains all the text information on the interface. At system startup, the corresponding resource file is loaded according to the user's language setting, enabling automatic switching of interface text. Data formats adhere to international standards, such as ISO8601 for date formats and automatic conversion of number formats based on regional conventions, ensuring a consistent and comfortable user experience for users from different regions and language backgrounds.

[0105] In this embodiment, the hybrid cloud architecture adopts Kubernetes federated cluster management, the public cloud deploys stateless computing nodes, and the private cloud deploys sensitive data storage; the blockchain adopts a sharded consortium chain structure and uses zero-knowledge proof to make data available but invisible; disaster recovery backup adopts erasure coding technology and designs a (10,6) Reed-Solomon coding scheme to achieve distributed redundant storage.

[0106] Furthermore, the monitoring route optimization suggestion adopts the improved ant colony algorithm, introduces the meteorological factor cost function and the real-time road condition weight matrix; the equipment maintenance reminder module integrates vibration spectrum analysis and infrared thermal imaging feature extraction, and constructs a deep residual network to realize fault prediction; the business rule engine is optimized with the Rete algorithm to support real-time reasoning of tens of millions of rule items.

[0107] It is worth noting that the spatiotemporal convolutional network adopts a three-dimensional dilated convolution structure, designs a causal convolution layer in the time dimension, and embeds a graph attention mechanism in the spatial dimension; the model training adopts a curriculum learning strategy, first pre-training ERA5 and then migrating to local monitoring data; the input data preprocessing uses the Kriging spatial interpolation method to fill in missing values, and uses the wavelet denoising algorithm to eliminate high-frequency interference.

[0108] Example 2

[0109] In this embodiment, based on Example 1, the system also includes an intelligent decision-making assistance function, which automatically generates gas pollution warnings, monitoring route optimization suggestions, and equipment maintenance reminder decision support information through in-depth analysis of historical data and real-time data, combined with preset business rules and machine learning models.

[0110] The gas pollution warning in the intelligent decision-making assistance function is constructed using a deep learning model based on a spatiotemporal convolutional neural network (ST-CNN). The model uses historical and real-time gas concentration data, meteorological data, and geographic information data as input. By extracting and analyzing the features of the data in spatial and temporal dimensions, it can achieve accurate prediction and warning of gas pollution conditions.

[0111] Specifically, in the spatial dimension, the model considers the spatial correlation between different monitoring points and captures the spatial distribution pattern of gas concentration in the area through the convolution layer; in the temporal dimension, a recurrent neural network (RNN) or a long short-term memory network (LSTM) is used to process the time series characteristics of the data to capture the changing trend of gas concentration over time.

[0112] At the same time, the system sets different thresholds based on the warning level. When the prediction results exceed the corresponding threshold, the warning information of different levels will be immediately triggered. The warning information will not only be displayed on the system interface with eye-catching colors and icons, but will also be promptly pushed to relevant management personnel via SMS and email.

[0113] To optimize monitoring routes, the system uses a hybrid optimization strategy combining genetic algorithms and ant colony algorithms. First, the genetic algorithm performs a global search, quickly finding a set of possible optimal route solutions through encoding, crossover, and mutation operations on the routes. Then, the ant colony algorithm performs local search and fine-tuning within these solutions, taking into account factors such as gas concentration distribution, traffic conditions, and the importance of monitoring points. The system generates the optimal monitoring route with the goal of minimizing monitoring costs (time and distance) and maximizing monitoring effectiveness.

[0114] The equipment maintenance reminder function is based on the equipment's operating status data and fault prediction model. The system collects the equipment's operating parameters (temperature, pressure, operating time) in real time, and uses fault prediction models (support vector machine, random forest) to predict possible equipment failures in advance. When it is predicted that the probability of a possible equipment failure exceeds the set threshold, the system will automatically generate equipment maintenance reminder information, and provide detailed maintenance recommendations and maintenance plans, including the required maintenance tools and spare parts lists, to ensure the stable operation of the equipment and the accuracy of data collection. In addition, in terms of multilingual interface and international data format, the system adopts an international resource file management method. Each language corresponds to an independent resource file, which facilitates subsequent language expansion and updates. For data formats, it follows internationally accepted standards and specifications. The date format adopts the ISO8601 standard, and the digital format is automatically converted according to the habits of different regions to ensure that users from different regions and language backgrounds can obtain a consistent and comfortable user experience.

[0115] Example 3

[0116] 1. Data collection layer implementation

[0117] Hardware configuration:

[0118] Meteorological data collection: Encrypted meteorological data packets are obtained every hour through the China Meteorological Data Network API (CIMISS), and HTTPS two-way authentication is used to ensure data security;

[0119] Gas concentration monitoring: A Picarro G2401 analyzer (detection limit 0.1 ppb) and a PRI-8800 analyzer (sampling frequency 1 Hz) are deployed onboard the vehicle and connected to an industrial computer via an RS-485 bus.

[0120] GPS positioning: using Trimble BD990 GNSS receiver (positioning accuracy ±10cm), outputting NMEA-0183 protocol data stream;

[0121] Map service: Call the 3D terrain service and satellite image service of Amap API V5.0, and perform permission verification through OAuth2.0.

[0122] Data stream processing:

[0123] The data from each device undergoes protocol conversion and preliminary verification through the edge computing gateway (Huawei Atlas500), generates a unified timestamp (accuracy ±1ms), and is pushed to the message queue using the MQTT protocol.

[0124] 2. Data processing and analysis layer implementation

[0125] Anomaly detection module:

[0126] The improved IsolationForest algorithm is used, the time sliding window is set (the default is 30 minutes), and the DBSCAN spatial clustering (eps = 50 meters, min_samples = 5) is combined to build a three-dimensional spatiotemporal anomaly detection model;

[0127] Dynamic adjustment of anomaly judgment threshold: When the proportion of abnormal points in three consecutive windows is greater than 5%, the threshold recalibration is automatically triggered (formula: T_{new}=0.7T_{old}+0.3MADTnew=0.7Told+0.3MAD).

[0128] Numerical simulation module:

[0129] Fluid dynamics calculations: A large eddy simulation (LES) model was constructed based on OpenFOAM 9.0, with dynamic adjustment of the grid resolution (0.5m to 10m) and real-time solution of the Navier-Stokes equations.

[0130] Chemical kinetic model: Integrates the gas-phase chemical mechanism (CB6r3 mechanism) of CMAQ5.3.2 and updates the photolysis rate constant every hour.

[0131] Data standardization process:

[0132] Unit conversion: gas concentration is unified into μmol / mol, and temperature is converted into Kelvin;

[0133] Time alignment: Cubic spline interpolation is used to align the time series of multiple devices and generate UTC+8 timestamps (ISO8601 format).

[0134] 3. Software integration layer implementation

[0135] Front-end architecture:

[0136] Built on Vue3.0+TypeScript, using Vite build tools to optimize the packaging process;

[0137] Virtual DOM optimization: Integrate the WebAssembly module to accelerate DOMDiff calculation (40% speed increase), and use requestIdleCallback to achieve asynchronous rendering;

[0138] Real-time communication: WebSocket connection enables dual encryption channels of SM4 (key length 256bit) and TLS1.3, with a data frame compression rate of ≥70%.

[0139] Backend architecture:

[0140] Adopting Spring Cloud Alibaba microservice framework, the deployment structure is:

[0141] Data storage layer: Ceph cluster (3 nodes, 3 replicas), using the CRUSH algorithm for automatic sharding;

[0142] Cache layer: Redis 6.2 cluster (8 nodes, RDMA network card acceleration), with a QPS of up to 500,000;

[0143] Computing layer: Kubernetes elastic scheduling (CPU / GPU mixed resource pool), supporting automatic expansion and contraction.

[0144] 4. Functional application layer implementation

[0145] Toolkit modules:

[0146] Map Engine: Builds 3D scenes based on CesiumJS 1.95. Terrain data uses Quantized-Mesh format and is loaded hierarchically through quadtree LOD (loading delay < 200ms).

[0147] Intelligent graphics switching: ResNet-18 model is trained to recognize user gesture trajectories (recognition accuracy > 92%), supporting sliding switching between heat maps and trajectory maps;

[0148] Dynamic threshold control: fuzzy PID controller parameter settings (K_p=0.8,K_i=0.2,K_d=0.1Kp=0.8,Ki=0.2,Kd=0.1) are used to generate the HSV color space gradient mapping function.

[0149] Weather forecast module:

[0150] WRF-LSTM hybrid model: WRF4.2 provides 1km resolution initial field, and LSTM network (hidden layer 512 units) performs time series correction;

[0151] Ensemble forecast: Dynamic weight allocation of ECMWF / GFS / CMA models (calculated by entropy weight method), Monte Carlo simulation times ≥ 1000.

[0152] 5. User interaction layer implementation

[0153] Interface design:

[0154] Layout engine: Based on Bootstrap 5.0 responsive grid system, breakpoint settings (xs:576px, sm:768px, md:992px);

[0155] Dynamic transition: Bezier curve parameter settings (cubic-bezier(0.25,0.1,0.25,1.0)), animation frame rate locked at 60FPS;

[0156] Micro visualization: Microcharts uses the D3.js rendering engine, radial gradient algorithm parameters:

[0157] javascript

[0158] constgradient=d3.radialGradient()

[0159] .stops([[0,'#FF0000'],[0.5,'#FFFF00'],[1,'#00FF00']])

[0160] .radius(0.5);

[0161] 6. Network and security implementation

[0162] Multipath transmission:

[0163] MPTCP sub-flow management strategy: primary path (5G NSA Band n78) + backup path (4G LTE Band 3), switching delay < 150ms;

[0164] Reinforcement learning retry strategy: The DQN network (ε-greedy strategy, ε = 0.1) dynamically optimizes the retry interval (benchmark interval formula: t = 2^{n}\times100msst = 2n×100ms).

[0165] Blockchain storage:

[0166] Sharding consortium chain architecture:

[0167] Data sharding: based on Shamir secret sharing scheme (k=3, n=5);

[0168] Zero-knowledge proof: Using the zk-SNARKs solution, the proof generation time is less than 500ms;

[0169] Smart Contract: Data audit contract written in Solidity and deployed on Hyperledger Fabric 2.4.

[0170] 7. Implementation of intelligent decision-making module

[0171] Route optimization algorithm:

[0172] Improved ant colony algorithm parameters:

[0173] Python

[0174] params={

[0175] 'ant_num':50,

[0176] 'alpha':1.2,

[0177] 'beta':2.5,

[0178] 'evaporation':0.3,

[0179] 'Q':100,

[0180] 'weather_weight': 0.4#Weather cost coefficient}

[0181] Real-time traffic matrix: integrated with Amap’s traffic situation API (update cycle 30 seconds).

[0182] Fault prediction model:

[0183] Deep residual network structure:

[0184] Input layer: vibration spectrum (0-10kHz, 1024-point FFT) + thermal imaging (256×256 pixels);

[0185] Network architecture: ResNet-50 + bidirectional LSTM (128 units);

[0186] Output layer: Softmax classification (three levels: normal / warning / fault).

[0187] 8. Implementation of gas pollution early warning model

[0188] Spatiotemporal Convolutional Network Architecture:

[0189] Input layer:

[0190] Spatial characteristics: 256×256 gridded concentration distribution (Kriging interpolation bandwidth 50m);

[0191] Time characteristics: 72-hour historical series (sampling interval is 1 hour).

[0192] Network structure:

[0193] Python

[0194] model=Sequential([

[0195] Conv3D(32,(3,5,5),dilation_rate=(2,1,1)),#time dimension expansion convolution

[0196] GraphAttentionLayer(heads=4),#Graph Attention Mechanism

[0197] CasualConv1D(64,kernel_size=5),#causal convolution

[0198] LSTM(128),

[0199] Dense(1,activation='sigmoid')])

[0200] Training strategy: Freeze the first three layers during transfer learning, and use a learning rate decay strategy (initial 0.001, decaying by 5% per epoch).

[0201] This embodiment uses the above-mentioned specific technical solutions to achieve accurate collection, intelligent analysis, and visual presentation of environmental monitoring data. The performance indicators are shown in the following table:

[0202]

[0203] The above specific embodiments are merely several optional embodiments of the present invention. Based on the technical solutions of the present invention and the relevant inspirations of the above embodiments, those skilled in the art may make various alternative improvements and combinations to the above specific embodiments.

Claims

1. A mobile navigation analysis and visualization system, characterized in that: include: Data collection layer: This layer obtains key data from professional data sources. Meteorological data is obtained through the meteorological API interface and the platform's on-board ultrasonic meteorological sensors and three-dimensional ultrasonic anemometers. Greenhouse gas concentration data is collected in real time by Picarro and PRI analyzers. GPS data is obtained through the on-board high-precision RTK multi-frequency Beidou module. Map data is called from the map service API. Data processing and analysis layer: This layer performs multi-dimensional processing and in-depth analysis of collected data. It first uses data cleaning algorithms to remove noise, outliers, and duplicate data, then performs format unification and standardization conversion. Finally, it conducts in-depth mining and analysis based on mathematical models and intelligent algorithms, while also employing parallel computing and asynchronous processing technologies. Software integration layer: adopts a front-end and back-end separation architecture. The front-end is built based on the front-end framework and realizes real-time data communication with the back-end through WebSocket. The back-end is built with a server-side framework and provides RESTful API services. It is responsible for data storage, management, processing and push, and performs data integration and interaction. Functional application layer: Integrates multiple special function modules, including toolkit module, comprehensive weather forecast module, multi-gas concentration time series monitoring module, wind speed-concentration correlation analysis module, multi-gas real-time synchronous monitoring module, methane carbon isotope characteristic analysis module, GPS and water vapor data integrated monitoring module, payload platform information module, multi-gas collaborative analysis module, and cruise real-time analysis module; User interaction layer: Design a user-friendly interface layout and distribute various functional modules in different areas, including real-time data viewing, historical data query, parameter setting, and report generation operations. The interface is dynamically updated in real time based on user operations.

2. The mobile navigation analysis and visualization system according to claim 1, characterized in that: The toolkit modules include: Map multi-dimensional display switching unit: provides switching between plane, 3D, and satellite map modes. Users can quickly switch map perspectives with one click to understand the geographic information of the monitored area. The map multi-dimensional display switching unit uses the WebGL 3D rendering engine and the Cesium geospatial framework to achieve terrain elevation mapping, and dynamically loads map tiles of different precisions through the quadtree LOD algorithm. Intelligent Graphic Display Switching Unit: This unit supports intelligent switching of graphical display modes such as concentration heat maps, trajectory roadmaps, and time series graphs. Users can select the display format to analyze data based on their needs. The intelligent graphic display switching unit has a built-in image recognition module based on convolutional neural networks, which automatically identifies the user's gesture trajectory and matches the optimal visualization format. Dynamic threshold customization unit: Users can dynamically set multiple gas concentration threshold intervals based on actual conditions and adjust the color identification of each interval to provide personalized visualization of different concentration areas. The dynamic threshold customization unit integrates a fuzzy PID control algorithm, automatically corrects the threshold interval boundary values ​​based on historical concentration fluctuations, and generates a color gradient dynamic mapping function. Flexible data source switching unit: users can switch between data provided by Picarro and PRI and compare monitoring results from different data sources; Query mode multi-switch unit: provides real-time data viewing, historical data query by time range, and specific event data filtering query; External resource quick link unit: integrates external resource links of on-site picture libraries and satellite image libraries related to the monitoring area.

3. The mobile navigation analysis and visualization system according to claim 1, characterized in that: The comprehensive weather forecast module includes: Real-time weather station information unit: displays detailed weather parameters provided by the on-board weather station in real time, including temperature, air pressure, humidity, wind direction, and wind speed, providing meteorological background information for mobile navigation monitoring; Hourly weather forecast unit: Based on a professional weather forecast model, it provides weather forecast information, including weather conditions, temperature changes, and wind speed and direction trends, helping users plan monitoring tasks in advance. The hourly weather forecast unit integrates the WRF mesoscale numerical forecast model with the LSTM time series prediction algorithm and performs multi-source data assimilation through Kalman filtering. Multi-day weather trend forecast unit: displays weather trend forecasts, including temperature range, precipitation probability, and weather type. The multi-day weather trend forecast unit uses ensemble forecasting technology, integrates the output results of the three major global models ECMWF, GFS, and CMA, and uses the Monte Carlo method to generate probability distribution cloud maps.

4. The mobile navigation analysis and visualization system according to claim 1, characterized in that: The data cleaning algorithm of the data processing and analysis layer uses a machine learning-based anomaly detection model combined with statistical analysis methods to automatically identify and remove outliers and noise in the data; data standardization conversion uses international standard units and time formats; mathematical models combine fluid mechanics and chemical kinetics to simulate the process of gas diffusion and chemical reactions; The anomaly detection model adopts an improved IsolationForest algorithm, introduces a time sliding window mechanism and spatial density clustering features, and constructs a three-dimensional spatiotemporal anomaly detection framework. The fluid mechanics model adopts the large eddy simulation (LES) method, coupled with the OpenFOAM solver to calculate pollutant diffusion trajectories in real time. The chemical kinetics module integrates the gas-phase chemical mechanism in the CMAQ air quality model.

5. The mobile navigation analysis and visualization system according to claim 1, characterized in that: The front-end framework of the software integration layer adopts virtual DOM technology and component development mode; WebSocket communication adopts encrypted transmission protocol, and the back-end server framework adopts distributed storage system and cache technology to optimize data storage and access performance; The virtual DOM adopts a difference algorithm optimization strategy and combines WebAssembly to accelerate the DOM tree comparison process; The encrypted transmission protocol adopts the national secret SM4 algorithm and TLS1.3 dual-channel encryption mechanism; the distributed storage system is designed based on the Ceph object storage architecture, using the CRUSH algorithm to achieve automatic data sharding and load balancing, and the cache layer uses the RDMA-accelerated Redis cluster architecture.

6. The mobile navigation analysis and visualization system according to claim 1, characterized in that: The top navigation bar of the user interaction layer integrates common function entrances, the sidebar displays toolkits and query options, the middle main area is a map and data display area, and the bottom status bar displays system status and key data indicators in real time; the interactive components are designed with icons and buttons, supporting mouse hover prompts and shortcut key operations; The interface layout adopts a responsive grid system and uses the Bezier curve algorithm to achieve smooth deformation transitions of visual elements; shortcut key operations support user-defined key combinations and integrate the Levenshtein distance algorithm to achieve fault-tolerant command recognition; the status bar uses Microcharts micro-visualization technology and dynamically renders data indicators through a radial gradient shading algorithm.

7. The mobile navigation analysis and visualization system according to claim 1, characterized in that: Communications between system layers and with external data sources utilize HTTP / 2 protocol and MQTT message transmission protocol, including but not limited to TLS encryption. The system has a network fault detection and automatic recovery mechanism, which can quickly switch to an alternative network or retry the connection when a network anomaly occurs. The network fault detection mechanism combines bidirectional heartbeat detection packets with BGP route monitoring, and automatically enables MPTCP multipath transmission when the packet loss rate is greater than 15%. The retry connection strategy adopts an exponential backoff algorithm and a reinforcement learning strategy optimizer to dynamically adjust the retry interval.

8. The mobile navigation analysis and visualization system according to claim 1, characterized in that: The system is built on a hybrid cloud architecture, combining public and private clouds for data storage and processing; it uses blockchain technology to encrypt, store, and trace key data; The hybrid cloud architecture uses Kubernetes federated cluster management, with stateless computing nodes deployed in the public cloud and sensitive data storage deployed in the private cloud. The blockchain adopts a sharded alliance chain structure.

9. The mobile navigation analysis and visualization system according to claim 1, characterized in that: It also includes intelligent decision-making assistance functions, which automatically generate gas pollution warnings, monitoring route optimization suggestions, and equipment maintenance reminders through in-depth analysis of historical and real-time data, combined with preset business rules and machine learning models; The monitoring route optimization suggestion adopts an improved ant colony algorithm, introduces a meteorological factor cost function and a real-time road condition weight matrix; the equipment maintenance reminder module integrates vibration spectrum analysis and infrared thermal imaging feature extraction, and constructs a deep residual network to achieve fault prediction; the business rule engine is optimized using the Rete algorithm.

10. The mobile navigation analysis and visualization system according to claim 9, characterized in that: The gas pollution warning in the intelligent decision-making support function is constructed using a deep learning model based on a spatiotemporal convolutional neural network. This model uses historical and real-time gas concentration data, meteorological data, and geographic information data as input. By extracting and analyzing the features of the data in spatial and temporal dimensions, it can achieve accurate prediction and warning of gas pollution conditions. The spatiotemporal convolutional network adopts a three-dimensional dilated convolution structure, designs a causal convolution layer in the time dimension, and embeds a graph attention mechanism in the spatial dimension; the model training adopts a curriculum learning strategy, first pre-training ERA5 and then migrating to local monitoring data; the input data preprocessing uses the Kriging spatial interpolation method to fill in missing values, and the wavelet denoising algorithm is used to eliminate high-frequency interference.

Images