Multi-dimensional spatio-temporal data analysis and visualization method, system and equipment

Abstract

The invention relates to the technical field of big data analysis and disaster management, and discloses a multi-dimensional spatio-temporal data analysis and visualization method, system and device. The method comprises the steps of obtaining multi-source spatio-temporal data and performing standardization processing to obtain processed spatio-temporal data; performing real-time analysis and trend prediction according to the processed spatio-temporal data based on a deep learning algorithm to obtain an analysis prediction result; performing three-dimensional modeling through a GIS platform according to the processed spatio-temporal data and an analysis prediction result to obtain a three-dimensional dynamic simulation model and performing visual display; and generating a decision strategy according to an analysis prediction result, and optimizing the decision strategy based on a reinforcement learning mechanism. According to the method, multi-source spatio-temporal data are efficiently integrated, real-time processing of large-scale data is achieved, and the data response speed is remarkably increased. Prediction is carried out based on multi-source data fusion and a deep learning algorithm, and the automation and accuracy of analysis are improved. Real-time dynamic display helps a user to understand a data trend more visually, and decision scientificity and timeliness are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of big data analysis and disaster management technology, and in particular to a multidimensional spatiotemporal data analysis and visualization method, system and device. Background Art

[0002] With the frequent occurrence of natural disasters, the complexity of disaster management and the difficulty of emergency response are increasing. Existing disaster prediction and emergency management technologies rely on traditional rule systems and static models, which cannot effectively deal with complex and changing disaster scenarios. Traditional rule systems and static models usually have the following shortcomings:

[0003] Limited data processing capabilities: Disaster scenarios require the integration of data from multiple data sources (such as meteorological data, satellite images, sensor data, etc.). These data types are numerous and large-scale, but traditional methods fail to fully utilize these heterogeneous data, resulting in slow system processing speed and low accuracy.

[0004] Low accuracy of disaster scenario prediction: Existing disaster prediction technologies are unable to efficiently integrate multi-source data (such as remote sensing images, meteorological data, and social media data), resulting in low prediction accuracy. Especially in the early stages of a disaster, the prediction results are often inaccurate, making it difficult to provide timely decision-making support for emergency management.

[0005] Failure to capture the nonlinear characteristics of disaster evolution: Disasters (such as typhoons, earthquakes, floods, etc.) usually have complex dynamic evolution processes, which are often highly nonlinear. Existing models cannot fully capture the nonlinear characteristics of disaster evolution, resulting in large errors in prediction results. Disasters are dynamic and complex, and traditional methods often rely on linear models or static models, which makes it impossible to accurately simulate the nonlinear evolution of disasters. For example, the spread of floods and the expansion of damage after earthquakes are highly nonlinear, and traditional models cannot effectively deal with these complex situations.

[0006] Lack of flexible emergency decision support: After a disaster occurs, how to efficiently allocate resources and formulate emergency response plans is a complex multi-objective optimization problem. Existing decision models usually only focus on a single goal (such as completing rescue in the shortest time), while ignoring the balance between multiple goals (such as resource optimization, personnel safety, post-disaster recovery, etc.). Disaster emergency response involves multiple goals (such as shortest response time, maximum number of rescuers, optimal resource allocation, etc.). Existing systems usually cannot balance between multiple goals, resulting in waste of resources or poor rescue results. Summary of the invention

[0007] The main purpose of the present invention is to provide a multidimensional spatiotemporal data analysis and visualization method, system and device, aiming to solve at least one of the above-mentioned technical problems.

[0008] To achieve the above object, the present invention provides a multidimensional spatiotemporal data analysis and visualization method, comprising:

[0009] Acquire multi-source spatiotemporal data, and perform standardization processing on the multi-source spatiotemporal data to obtain processed spatiotemporal data;

[0010] Based on the deep learning algorithm, real-time analysis and trend prediction are performed according to the processed spatiotemporal data to obtain analysis and prediction results;

[0011] Performing three-dimensional modeling based on the processed spatiotemporal data and analysis and prediction results through a GIS platform to obtain a three-dimensional dynamic simulation model and perform visual display;

[0012] A decision strategy is generated according to the analysis and prediction results, and the decision strategy is optimized based on a reinforcement learning mechanism.

[0013] In some embodiments, the acquiring multi-source spatiotemporal data and performing standardization processing on the multi-source spatiotemporal data to obtain processed spatiotemporal data includes:

[0014] Acquire satellite remote sensing data, meteorological data, ground sensor data, and media data from multiple heterogeneous data sources based on a stream processing architecture;

[0015] The satellite remote sensing data, meteorological data, ground sensor data and media data are uniformly processed by using ETL tools to obtain multi-source spatiotemporal data;

[0016] Data standardization is performed on the multi-source spatiotemporal data to obtain processed spatiotemporal data.

[0017] In some embodiments, performing data standardization processing on the multi-source spatiotemporal data to obtain processed spatiotemporal data includes:

[0018] Performing image denoising and data cleaning on the multi-source spatiotemporal data to obtain preprocessed data;

[0019] Normalizing pixel values ​​of image data in the preprocessed data to obtain image normalized data;

[0020] Performing z-score standardization on the time series data in the preprocessed data to obtain time series normalized data;

[0021] The processed spatiotemporal data is obtained according to the image normalization data and the time series normalization data.

[0022] In some embodiments, the deep learning algorithm is used to perform real-time analysis and trend prediction based on the processed spatiotemporal data to obtain analysis and prediction results, including:

[0023] Based on the graph neural network, data fusion is performed on the processed spatiotemporal data to obtain disaster scenario prediction features;

[0024] Based on the spatiotemporal convolutional network, the processed spatiotemporal data is subjected to spatiotemporal joint modeling to obtain spatiotemporal comprehensive features;

[0025] Real-time analysis and trend prediction are performed based on the disaster scenario prediction characteristics and spatiotemporal comprehensive characteristics to obtain analysis and prediction results.

[0026] In some embodiments, the graph neural network-based data fusion is performed on the processed spatiotemporal data to obtain disaster scenario prediction features, including:

[0027] Representing the processed spatiotemporal data as a graph structure;

[0028] Performing feature learning on each node of the graph structure through a graph convolutional network to construct an adjacency matrix;

[0029] In each layer of the graph convolutional network, the node features are updated through the information transmission mechanism;

[0030] In the last layer of the graph convolutional network, a graph pooling operation is performed to obtain a global feature representation of the graph structure;

[0031] The global feature representation is used as a disaster scenario prediction feature.

[0032] In some embodiments, the processing spatiotemporal data is subjected to spatiotemporal joint modeling based on a spatiotemporal convolutional network to obtain spatiotemporal comprehensive features, including:

[0033] Based on a convolutional neural network, spatial feature modeling is performed on the processed spatiotemporal data to obtain spatial features of the disaster area;

[0034] Based on the long short-term memory network, time series modeling is performed on the processed spatiotemporal data to obtain the time characteristics of disaster evolution;

[0035] Based on the spatial characteristics of the disaster area and the temporal characteristics of the disaster evolution, spatiotemporal joint modeling is performed to obtain spatiotemporal comprehensive characteristics.

[0036] In some embodiments, the three-dimensional modeling is performed based on the processed spatiotemporal data and the analysis and prediction results through the GIS platform to obtain a three-dimensional dynamic simulation model and perform a visual display, including:

[0037] Performing three-dimensional modeling on the analysis and prediction results through Cesium's three-dimensional rendering engine to obtain an analysis result simulation model;

[0038] Dynamically displaying three-dimensional data of the analysis result simulation model based on a particle system;

[0039] Constructing a dynamic heat map and trend curve based on the processed spatiotemporal data through a GIS platform;

[0040] The dynamic heat map and trend curve are visualized by combining graphic transformation and smooth transition effect technology.

[0041] In some embodiments, generating a decision strategy according to the analysis and prediction results, and optimizing the decision strategy based on a reinforcement learning mechanism, includes:

[0042] Construct a multi-objective decision-making model for disaster emergency response;

[0043] Generate a decision strategy based on the multi-objective decision model according to the analysis and prediction results;

[0044] Optimizing the multi-objective decision-making model based on a reinforcement learning algorithm to obtain an updated multi-objective decision-making model;

[0045] An optimized decision strategy is obtained based on the updated multi-objective decision model and the analysis and prediction results.

[0046] In addition, to achieve the above purpose, the present invention also proposes a multidimensional spatiotemporal data analysis and visualization system, comprising:

[0047] A data access module is used to obtain multi-source spatiotemporal data and perform standardization processing on the multi-source spatiotemporal data to obtain processed spatiotemporal data;

[0048] The analysis and prediction module is used to perform real-time analysis and trend prediction based on the processed spatiotemporal data based on the deep learning algorithm to obtain the analysis and prediction results;

[0049] A visualization display module is used to perform three-dimensional modeling based on the processed spatiotemporal data and analysis and prediction results through a GIS platform to obtain a three-dimensional dynamic simulation model and perform visualization display;

[0050] The reinforcement learning module is used to generate a decision strategy according to the analysis and prediction results, and optimize the decision strategy based on the reinforcement learning mechanism.

[0051] In addition, to achieve the above-mentioned purpose, the present invention also proposes an electronic device, which includes: a memory, a processor, and a multidimensional spatiotemporal data analysis and visualization program stored in the memory and runnable on the processor, wherein the multidimensional spatiotemporal data analysis and visualization program is configured to implement the multidimensional spatiotemporal data analysis and visualization method as described above.

[0052] The present invention provides a multi-dimensional spatiotemporal data analysis and visualization method, including: acquiring multi-source spatiotemporal data, and performing standardization processing on the multi-source spatiotemporal data to obtain processed spatiotemporal data; performing real-time analysis and trend prediction based on the processed spatiotemporal data based on a deep learning algorithm to obtain analysis and prediction results; performing three-dimensional modeling based on the processed spatiotemporal data and the analysis and prediction results through a GIS platform to obtain a three-dimensional dynamic simulation model and perform visual display; generating a decision strategy based on the analysis and prediction results, and optimizing the decision strategy based on a reinforcement learning mechanism. In the present invention, multi-source spatiotemporal data is efficiently integrated to achieve real-time processing of large-scale data, significantly improve data response speed, and reduce data lag. Prediction is performed based on multi-source data fusion and deep learning algorithms to improve the automation and accuracy of analysis. Visual real-time dynamic display helps users understand data and trends more intuitively, enhances the scientificity and timeliness of decision-making, and solves problems such as low disaster prediction accuracy, difficulty in capturing nonlinear evolution processes, and lack of multi-objective optimization of emergency decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 A schematic diagram of the structure of an electronic device in a hardware operating environment involved in an embodiment of the present invention;

[0054] Figure 2 It is a flowchart of an embodiment of a multi-dimensional spatiotemporal data analysis and visualization method of the present invention;

[0055] Figure 3 It is a technical flow chart involved in the embodiment of the present invention;

[0056] Figure 4 It is a structural block diagram of an embodiment of the multi-dimensional spatiotemporal data analysis and visualization system of the present invention.

[0057] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0059] It should be noted that all directional indications in the embodiments of the present invention (such as up, down, left, right, front, back, etc.) are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0060] In addition, the descriptions of "first", "second", etc. in the present invention are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but it must be based on the ability of ordinary technicians in the field to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by the present invention. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0061] Reference Figure 1 , Figure 1 The figure is a schematic diagram of the structure of an electronic device of the hardware operating environment involved in the embodiment of the present invention.

[0062] like Figure 1 As shown, the electronic device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (Wireless-Fidelity, Wi-Fi) interface). The memory 1005 may be a high-speed random access memory (Random Access Memory, RAM memory) or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk memory. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0063] Those skilled in the art will understand that Figure 1 The structure shown in the figure does not constitute a limitation on the electronic device, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0064] like Figure 1 As shown, the memory 1005 as a storage medium may include an operating system, a network communication module, a user interface module, and a multi-dimensional spatiotemporal data analysis and visualization program.

[0065] exist Figure 1 In the electronic device shown, the network interface 1004 is mainly used for data communication with a network server; the user interface 1003 is mainly used for data interaction with a user; the processor 1001 and the memory 1005 in the electronic device of the present invention can be set in the electronic device, and the electronic device calls the multidimensional spatiotemporal data analysis and visualization program stored in the memory 1005 through the processor 1001, and executes the multidimensional spatiotemporal data analysis and visualization method provided by an embodiment of the present invention.

[0066] Geographic Information System (GIS) technology can provide users with accurate spatial data and visual representation of its spatiotemporal changes through the collection, processing and analysis of geographic data. At present, GIS technology is widely used in many fields such as urban planning, traffic management, and environmental monitoring. With the advancement of computer graphics and data visualization technology, three-dimensional geographic information visualization has become an important development direction of GIS. With the development of remote sensing technology and global positioning system (GPS), the collection and application of spatiotemporal data has become a hot topic of research. How to efficiently process a large amount of spatiotemporal data and conduct intelligent analysis and visual display is a key issue in the current GIS field. When facing massive and dynamic spatiotemporal data, traditional GIS systems often find it difficult to achieve real-time processing and dynamic display, especially the lack of the ability to fuse multi-source data and make effective predictions. At present, the sources of spatiotemporal data are becoming more and more extensive, including satellite remote sensing data, meteorological data, geological data, traffic data, etc. These data have different formats, scales and timeliness. How to effectively fuse these multi-source data to provide more accurate and comprehensive decision support is a challenge in the field of GIS and AI. Traditional systems rely more on static data, while the present invention emphasizes the efficient integration and analysis of real-time streaming data.

[0067] At present, geographic information system (GIS) and artificial intelligence (AI) technologies have been widely used in many fields, such as urban management, environmental monitoring and disaster warning. However, existing technologies have some limitations in processing large-scale, real-time spatiotemporal data, which are mainly reflected in the following aspects: 1. Limitations of GIS systems: Although traditional GIS platforms such as Cesium and ArcGIS can effectively display spatial data, they have low processing efficiency when facing large-scale real-time data and cannot quickly respond to dynamic changes in data. Most current GIS systems rely on static data analysis and lack the ability to update and display in real time, resulting in insufficient and inaccurate decision support. 2. Insufficient artificial intelligence: Although AI technology has been applied in spatiotemporal data analysis, existing models often cannot effectively integrate multi-source data from meteorology, transportation, remote sensing and other fields, resulting in insufficient accuracy and real-time performance of analysis results. Traditional AI systems have limited processing capabilities for real-time data and are difficult to meet the high timeliness requirements in fields such as disaster warning and smart city management. 3. Challenges of data fusion and visualization: It is difficult to integrate multi-source data, especially the fusion of cross-domain data (such as meteorology, transportation, geology, etc.), which limits the comprehensiveness of analysis. The GIS platform lacks dynamic visualization capabilities and cannot display prediction results or data change trends in real time, which affects users' decision-making effects.

[0068] Existing GIS systems are inefficient in processing large-scale spatiotemporal data, especially in processing real-time data streams, and are difficult to meet the needs of rapid response. Batch processing is usually used to update data, resulting in delayed data updates and failure to reflect actual changes in real time. Most existing systems rely on manual analysis and empirical judgment, and lack automated intelligent analysis and prediction capabilities. Especially when faced with complex spatiotemporal data, traditional systems cannot fully utilize artificial intelligence technology for trend prediction, anomaly detection, and intelligent decision-making, which reduces the intelligence level of the system. Many GIS platforms focus on static data display and lack real-time update and visualization of dynamic data. Especially for complex spatiotemporal data, such as disaster simulation and traffic flow changes, existing systems cannot provide intuitive and interactive dynamic display effects, which affects users' understanding and decision-making of data. Existing technologies have great difficulties in integrating multi-source data from different fields (such as meteorology, geology, transportation, etc.). The differences in the format, update frequency, and data quality of data in different fields make the analysis effect after cross-domain data fusion poor, and it is difficult to provide comprehensive and accurate decision support. The current system lacks a rapid feedback mechanism for real-time data changes, and cannot adaptively adjust analysis and prediction strategies. When data sources change or new trends emerge, existing systems are often unable to adjust models or optimize results in a timely manner, resulting in large prediction errors and affecting the reliability of decision-making.

[0069] In view of this, the present invention proposes a multi-dimensional spatiotemporal data analysis and visualization method, system and device.

[0070] The embodiment of the present invention provides a multidimensional spatiotemporal data analysis and visualization method, referring to Figure 2 , Figure 2 It is a flowchart of an embodiment of a multi-dimensional spatiotemporal data analysis and visualization method of the present invention.

[0071] like Figure 2 As shown, the multidimensional spatiotemporal data analysis and visualization method includes:

[0072] Step S100: acquiring multi-source spatiotemporal data, and performing standardization processing on the multi-source spatiotemporal data to obtain processed spatiotemporal data;

[0073] Step S200: performing real-time analysis and trend prediction based on the processed spatiotemporal data based on a deep learning algorithm to obtain analysis and prediction results;

[0074] Step S300: performing three-dimensional modeling based on the processed spatiotemporal data and the analysis and prediction results through the GIS platform to obtain a three-dimensional dynamic simulation model and perform visual display;

[0075] Step S400: generating a decision strategy according to the analysis and prediction results, and optimizing the decision strategy based on a reinforcement learning mechanism.

[0076] It should be noted that the execution subject in this embodiment may be an electronic device, which may be a computer device with data processing functions, or other devices that can achieve the same or similar functions. This embodiment does not limit this. In this embodiment, a computer device is used as an example for explanation.

[0077] It is understandable that this embodiment is an innovative application in spatiotemporal data analysis, real-time prediction and dynamic visualization. By combining the three-dimensional visualization capabilities of the geographic information system (GIS) and the analysis and prediction functions of artificial intelligence (AI), it aims to efficiently process and analyze large-scale spatiotemporal data and provide real-time dynamic decision support. This embodiment is a multi-dimensional spatiotemporal data analysis and visualization solution based on a GIS platform (such as Cesium) and artificial intelligence algorithms, aiming to solve the challenges faced in the current fields of smart city management, natural disaster warning, environmental monitoring and resource optimization.

[0078] Specifically, this embodiment is explained by applying it to smart city management and disaster warning as an example. Smart city management: By integrating GIS and AI technologies, it can help governments and enterprises better manage cities, such as in areas such as traffic flow management, environmental monitoring, and resource allocation. Especially in the context of rapid urban development, real-time, intelligent management platforms can help decision makers respond quickly and improve the operating efficiency of cities. Natural disaster and environmental monitoring: It can monitor environmental changes in real time and identify potential disaster risks by analyzing multi-dimensional data such as meteorology, geology, and transportation, and improve disaster prevention capabilities and decision-making efficiency through intelligent early warning and emergency response systems.

[0079] In one embodiment, multi-source spatiotemporal data are acquired, and the multi-source spatiotemporal data are standardized to obtain processed spatiotemporal data, including: acquiring satellite remote sensing data, meteorological data, ground sensor data, and media data from multiple heterogeneous data sources based on a stream processing architecture; uniformly processing the satellite remote sensing data, meteorological data, ground sensor data, and media data through an ETL tool to obtain multi-source spatiotemporal data; and performing data standardization on the multi-source spatiotemporal data to obtain processed spatiotemporal data.

[0080] It should be noted that multi-source data fusion: the sources of current spatiotemporal data are becoming more and more extensive, including satellite remote sensing data, meteorological data, geological data, traffic data, etc. These data have different formats, scales and timeliness. How to effectively fuse these multi-source data to provide more accurate and comprehensive decision support is a challenge in the field of GIS and AI. Traditional systems mostly rely on static data, while this embodiment emphasizes the efficient integration and analysis of real-time streaming data.

[0081] Specifically, this embodiment introduces ETL (extraction, transformation, loading) tools to uniformly process data of different formats and sources to ensure data compatibility and consistency. It can automatically integrate multi-source data from satellite remote sensing, meteorological data, geographic information, traffic flow, etc., and standardize these data. The Cesium platform is used for three-dimensional modeling to process the coordinate transformation of spatial data and the standardization of time dimension data, so that data from different sources can be displayed and analyzed in a unified three-dimensional environment. Combined with distributed computing frameworks (such as Hadoop and Spark), large-scale data is processed in parallel to ensure efficient data processing capabilities. The computing process uses a cluster architecture to improve performance to meet the needs of big data processing. In order to ensure the security and privacy protection of data, encryption technology and permission control mechanisms can be used during data transmission and processing to ensure the security and compliance of sensitive data.

[0082] Specifically, Figure 3As shown, big data and real-time stream processing: With the popularization of sensors, satellite remote sensing technology and smart devices, a large amount of spatiotemporal data is generated. The real-time processing and analysis of these data puts higher requirements on existing traditional technologies. In order to cope with the processing of massive data streams, this embodiment introduces big data architecture and stream data processing technology, such as distributed computing frameworks such as Apache Kafka and Apache Spark. These frameworks can realize efficient data collection, processing and storage, and are particularly suitable for real-time analysis of large-scale spatiotemporal data.

[0083] For example, multi-source data integration: supports access to a wide range of data sources, including satellite remote sensing data lakes, meteorological data, geographic information data, traffic flow data, and social media data. In this embodiment, multi-format data can be collected and integrated through specially designed data interfaces. By introducing ETL (Extract-Transform-Load) tools, the data extraction, transformation, and loading processes are automatically managed to ensure efficient data access and integration.

[0084] Exemplarily, data standardization and model building: normalize multi-source data. Spatial data needs to be converted and georeferenced to meet the needs of 3D modeling. Time dimension data needs to have a unified time base and necessary processing such as time zone conversion. In this embodiment, Cesium's 3D Tiles format can be used to build a dynamic model, so that the data has higher compatibility and expressiveness in a 3D environment.

[0085] For example, real-time data stream processing: In this embodiment, real-time access to data streams is achieved based on a stream processing architecture (such as Apache Kafka). A distributed messaging system is used to ensure rapid processing and response to real-time data, thereby meeting the needs of high timeliness.

[0086] In one embodiment, data standardization is performed on the multi-source spatiotemporal data to obtain processed spatiotemporal data, including: image denoising and data cleaning are performed on the multi-source spatiotemporal data to obtain preprocessed data; pixel value normalization is performed on the image data in the preprocessed data to obtain image normalized data; z-score normalization is performed on the time series data in the preprocessed data to obtain time series normalized data; and processed spatiotemporal data is obtained based on the image normalized data and the time series normalized data.

[0087] Specifically, data collection and preprocessing: In order to achieve accurate disaster prediction, this embodiment first needs to collect data from multiple heterogeneous data sources, including but not limited to: Remote sensing images: Satellite remote sensing data (such as Landsat, Sentinel satellites) provide surface images, which can be used to monitor the spatial distribution of disasters (such as floods, fires, storms, etc.). Meteorological data: Real-time weather data from meteorological stations and meteorological satellites (such as precipitation, wind speed, temperature, etc.) is used to capture the environmental conditions of disasters. Ground sensor data: including earthquake monitoring sensors, flood monitoring equipment, etc., to provide accurate disaster activity data. Social media data: Capture disaster-related text information from social platforms to capture disaster dynamics, public feedback and evacuation situations.

[0088] Specifically, data preprocessing: The original data (multi-source spatiotemporal data) may have problems such as noise and missing values, so all multi-source spatiotemporal data are first cleaned and normalized.

[0089] Exemplarily, the data preprocessing steps include: 1. Denoising and data cleaning: All data sources need to be denoised, including image denoising (for example, using convolutional neural network denoising) and missing value filling for time series data (using interpolation, mean filling, etc.). 2. Data normalization: All data are normalized to eliminate scale differences between data sources. For image data, pixel value normalization is used, and for time series data, z-score normalization is used.

[0090] In one embodiment, real-time analysis and trend prediction are performed on the processed spatiotemporal data based on a deep learning algorithm to obtain analysis and prediction results, including: data fusion is performed on the processed spatiotemporal data based on a graph neural network to obtain disaster scenario prediction features; spatiotemporal joint modeling is performed on the processed spatiotemporal data based on a spatiotemporal convolutional network to obtain spatiotemporal comprehensive features; real-time analysis and trend prediction are performed based on the disaster scenario prediction features and spatiotemporal comprehensive features to obtain analysis and prediction results.

[0091] It is understandable that artificial intelligence, especially deep learning technology, has achieved remarkable results in image recognition, time series analysis, anomaly detection, etc., and is widely used in the analysis and prediction of spatiotemporal data, especially in the real-time processing and trend prediction of dynamic data, which can provide accurate prediction and automated analysis that traditional methods cannot achieve. Technologies such as deep learning, long short-term memory network (LSTM), and convolutional neural network (CNN) have shown great potential in the fields of traffic flow prediction, disaster warning, and environmental monitoring.

[0092] Specifically, this embodiment integrates data from different sources (such as remote sensing images, meteorological data, ground sensor data, social media text, etc.) and uses graph neural networks (GNN) for data fusion to build a unified disaster scenario prediction feature representation. This section will focus on how to efficiently process and fuse different types of data and extract useful features.

[0093] For example, data preprocessing: the original data may have problems such as noise and missing values, so all data are first cleaned and normalized. Graph neural network fusion: Graph neural network GNN is used to fuse multi-source data to capture the spatial relationship and mutual influence between data sources. Graph neural network can model the interdependence between data through adjacency matrix, thereby improving the accuracy of disaster scenario prediction.

[0094] In one embodiment, the processed spatiotemporal data is fused based on a graph neural network to obtain disaster scenario prediction features, including: representing the processed spatiotemporal data as a graph structure; performing feature learning on each node of the graph structure through a graph convolutional network to construct an adjacency matrix; at each layer in the graph convolutional network, updating the features of the nodes through an information passing mechanism; at the last layer of the graph convolutional network, performing a graph pooling operation to obtain a global feature representation of the graph structure; and using the global feature representation as a disaster scenario prediction feature.

[0095] Specifically, Graph Neural Network (GNN) data fusion: In order to effectively process heterogeneous data, this embodiment uses a graph neural network (GNN) for data fusion. Heterogeneous data sources are represented by a graph structure, where nodes represent different types of data sources and edges represent the spatial or temporal dependencies between them. Exemplarily, the design of a graph neural network includes the following steps:

[0096] ① Node representation learning: For each node (data source), feature learning is performed through a graph convolutional network (GCN). The initial features of the node are the original features of the data source (such as the pixel values ​​of remote sensing images, the statistical features of meteorological data, etc.).

[0097] ② Adjacency matrix construction: The adjacency matrix A represents the spatial or temporal relationship between data sources. For example, the relationship between a weather station and a ground sensor can be defined by the distance of their geographical locations, and the relationship between social media texts and remote sensing images can be defined by the similarity of disaster types.

[0098] ③ Graph convolution operation: Each layer of graph convolution updates the features of nodes through the information transmission mechanism:

[0099]

[0100] Among them, H (l)represents the node features of the lth layer, A ij represents the adjacency matrix, W (l) represents the weight matrix of the graph convolutional layer, and σ represents the activation function.

[0101] ④ Graph pooling: In the last layer of the graph neural network, a graph pooling operation is performed to aggregate the features of all nodes to obtain a global feature representation of the entire graph. Through global pooling operations, such as maximum pooling or average pooling, local features are converted into global features (disaster scenario prediction features).

[0102] Through these steps, the graph neural network can learn node features, use the adjacency matrix to represent the relationship between nodes, and extract the global features of the entire graph through graph convolution and pooling operations.

[0103] In one embodiment, the processed spatiotemporal data are subjected to spatiotemporal joint modeling based on a spatiotemporal convolutional network to obtain spatiotemporal comprehensive characteristics, including: performing spatial feature modeling on the processed spatiotemporal data based on a convolutional neural network to obtain spatial characteristics of the disaster area; performing time series modeling on the processed spatiotemporal data based on a long short-term memory network to obtain temporal characteristics of disaster evolution; performing spatiotemporal joint modeling based on the spatial characteristics of the disaster area and the temporal characteristics of disaster evolution to obtain spatiotemporal comprehensive characteristics.

[0104] It should be noted that spatiotemporal joint modeling: In order to improve the accuracy of disaster prediction, this embodiment combines the spatiotemporal convolutional network (ST-GCN) to model the spatiotemporal evolution of disasters. The spatiotemporal convolutional network ST-GCN jointly models spatial and temporal information, and can simultaneously capture the spatial distribution characteristics and temporal change trends of disasters.

[0105] Specifically, spatial feature modeling: use convolutional neural networks (CNN) to extract the features of spatial data, such as using convolutional neural networks CNN to process remote sensing images and extract the spatial features of disaster areas. Time series data processing: use long short-term memory networks (LSTM) to process time series data to capture the temporal trend of disaster evolution. Combination of spatiotemporal convolution: combine spatial convolution with temporal convolution, use spatiotemporal convolution network ST-GCN to comprehensively model the spatiotemporal evolution process of disasters, obtain spatiotemporal comprehensive features, and thus obtain accurate disaster development predictions.

[0106] Exemplarily, spatial feature modeling: In disaster prediction, spatial information (such as remote sensing images) is crucial for the prediction of disaster scenarios. This embodiment uses a convolutional neural network (CNN) to extract spatial features. The specific steps are as follows: Image preprocessing: For remote sensing images, first enhance the image contrast and improve the image quality through standard image processing methods (such as histogram equalization). Convolution operation: Use a convolutional neural network (CNN) to extract features from the image. Here, the specific architecture can choose classic network architectures such as ResNet and VGGNet, and extract spatial features in the image (spatial features of the disaster area) through convolution layers and pooling layers. Here, the spatial feature extraction formula is:

[0107] F spatial =CNN(I)

[0108] Where I represents the input remote sensing image, F spatial Represents the extracted spatial features.

[0109] Exemplarily, temporal feature modeling: Disasters usually show obvious temporal dynamics, such as the intensity of typhoons changing over time, the expansion of floods, etc. In order to capture these temporal features, this embodiment uses a long short-term memory network (LSTM) for time series modeling. The design of LSTM is as follows: Input layer: The time series of meteorological data and social media texts are used as input, and they are input into the LSTM network after preprocessing. Here, the LSTM formula is:

[0110] h t =σ(W h x t +b h )

[0111] Among them, h t represents the hidden state at time step t; x t represents the input feature; W h represents the weight matrix; b h represents the bias; σ represents the activation function.

[0112] Exemplarily, spatiotemporal joint modeling (ST-GCN): This embodiment uses a spatiotemporal convolutional network (ST-GCN) to jointly model spatial and temporal features. The spatiotemporal convolutional network combines the advantages of convolutional neural networks (CNN) and long short-term memory networks (LSTM), and improves the prediction accuracy of the disaster evolution process by jointly modeling spatial and temporal features. The specific steps are as follows: Spatiotemporal convolution operation: In ST-GCN, spatial convolution operations (CNN) and temporal convolution operations (LSTM) are jointly processed. Spatial features and temporal features work together through convolutional layers and temporal recursive layers to finally obtain spatiotemporal comprehensive features. Here, the spatiotemporal convolution formula used is:

[0113] F spatial-time =ST-GCN(F spatial , F time )

[0114] Among them, F spatial-time represents the spatiotemporal features (spatial-temporal comprehensive features) output by the spatiotemporal convolutional network, F spatial and F time Represent the spatial features and temporal features extracted from CNN and LSTM, respectively.

[0115] It is understandable that through AI analysis of time series prediction and anomaly detection, a deep learning model based on LSTM (long short-term memory network) is used in this embodiment to analyze and predict trends of time series data. Combined with convolutional neural network (CNN) to identify and classify spatial images, more targeted analysis results are generated, so that trends can be identified, future changes can be predicted, and potential anomalies can be detected in a timely manner. In addition, hybrid model fusion and optimization: ensemble learning methods (such as random forests and XGBoost) can be used to fuse the analysis results of multiple models to improve the accuracy and stability of the overall prediction. For specific application scenarios (such as traffic forecasting, environmental monitoring, or disaster emergency response, etc.), specific AI models are customized for deeper optimization. Real-time feedback and adaptive learning: Introduce reinforcement learning mechanism, make adaptive adjustments in the constantly updated data stream, and the deep learning model will automatically correct the prediction strategy based on real-time data feedback to improve the accuracy and reliability of future predictions.

[0116] In one embodiment, three-dimensional modeling is performed on the processed spatiotemporal data and analysis and prediction results through a GIS platform to obtain a three-dimensional dynamic simulation model and perform a visual display, including: three-dimensional modeling of the analysis and prediction results is performed through Cesium's three-dimensional rendering engine to obtain an analysis result simulation model; three-dimensional data of the analysis result simulation model is dynamically displayed based on a particle system; dynamic heat maps and trend curves are constructed based on the processed spatiotemporal data through a GIS platform; and the dynamic heat maps and trend curves are visualized by combining graphic transformation and smooth transition effect technology.

[0117] It should be noted that if Figure 3As shown, three-dimensional visualization and dynamic display: With the development of computer graphics and virtual reality technology, the visualization technology of three-dimensional geographic information is becoming more and more mature. In this embodiment, a GIS platform such as Cesium is used to provide realistic three-dimensional maps and models, so that users can more intuitively understand the spatiotemporal changes of spatial data, especially when dynamically displaying future trends or real-time changes, which can help decision makers make more scientific judgments. In addition to traditional two-dimensional charts and static maps, you can also interact through voice commands, natural language input, gesture operations, etc., adjust the display perspective, query data, generate reports, etc. Especially when processing spatiotemporal data, the interactive function not only improves the user's convenience of operation, but also enhances the flexibility and real-time nature of data display.

[0118] Specifically, visualization includes but is not limited to: dynamic display of three-dimensional data, dynamic heat map and trend curve display, real-time simulation and interactive animation, and multi-dimensional data overlay display. Dynamic display of three-dimensional data: Display AI analysis results through Cesium's three-dimensional rendering engine. Use particle systems to simulate the flow changes of natural phenomena such as wind and rain to enhance the realism and vividness of the display. Dynamic heat map and trend curve: Visualize the dynamic changes of multi-dimensional data and support users to view the changing trends of any time period in real time. Combined with graphics transformation, smooth transition effects and other technologies, the data display is both beautiful and efficient. Real-time simulation and interactive animation: Realize the simulation of dynamic changes in geographical areas, such as the spread of floods and the evolution of traffic congestion. Through real-time updated animation demonstrations, help users intuitively understand the analysis results and make scientific decisions. Multi-dimensional data overlay display: Support the overlay of different data dimensions, such as the comprehensive display of meteorological, geological and traffic data in the same view, making the relationship between data clearer and helping users quickly discover interactive effects.

[0119] In one embodiment, a decision strategy is generated according to the analysis and prediction results, and the decision strategy is optimized based on a reinforcement learning mechanism, including: constructing a multi-objective decision model for disaster emergency response; generating a decision strategy according to the analysis and prediction results based on the multi-objective decision model; optimizing the multi-objective decision model based on a reinforcement learning algorithm to obtain an updated multi-objective decision model; and obtaining an optimized decision strategy based on the updated multi-objective decision model and the analysis and prediction results.

[0120] Specifically, reinforcement learning-driven multi-objective decision optimization: Combine reinforcement learning technology to optimize the multi-objective decision-making process of disaster emergency response. By defining a reward function to evaluate the pros and cons of different decision-making plans, and using reinforcement learning algorithms (such as Q-learning, deep Q network DQN, etc.) to continuously update the decision-making strategy, ultimately achieving the optimal configuration of post-disaster resources. Specific steps include: Decision modeling: Each decision node corresponds to a possible emergency response action (such as resource allocation, personnel evacuation, rescue route planning, etc.); Multi-objective optimization: By designing a multi-objective reward function, multiple objectives such as time, resources, and personnel safety are comprehensively considered, and optimized through reinforcement learning algorithms.

[0121] Exemplarily, multi-objective decision optimization (reinforcement learning): Decision modeling: The decision-making process of disaster emergency response involves multiple objectives, such as shortest response time, maximizing resource utilization efficiency, ensuring personnel safety, etc. Therefore, it is necessary to design a multi-objective decision model (multi-objective decision model for disaster emergency response) to weigh different objectives; Objective function: Each decision node corresponds to a possible emergency response action (such as allocating resources, planning evacuation routes, etc.). Here, the objective function is designed as the weighted sum of multiple objectives:

[0122] R=α1R1+α2R2+...+α n R n

[0123] Among them, R represents the comprehensive reward function, R1, R2, ..., R n Represents the rewards of different goals, α1, α2, ..., α n Represents the weight of each target.

[0124] Exemplarily, reinforcement learning algorithm: In multi-objective decision-making problems, reinforcement learning (RL) algorithm is used for decision optimization. In this embodiment, Q-learning or deep Q network (DQN) can be used for optimization, and the specific steps are as follows: state space: each state represents the current disaster scenario, including the scale of the disaster, resource availability, etc.; action space: each action corresponds to an emergency response decision, such as resource allocation, personnel evacuation, road clearing, etc.; Q-learning update formula:

[0125]

[0126] Among them, s t Indicates the current state, a t Indicates the action of selection, r t represents the immediate reward, γ represents the discount factor, and α represents the learning rate.

[0127] Exemplarily, deep Q network (DQN): In order to cope with the challenges of large-scale state space and action space, this embodiment also uses a deep Q network (DQN) to handle multi-objective decision optimization. The deep Q network DQN uses a convolutional neural network (CNN) to approximate the Q function, and the specific calculation formula is as follows:

[0128] Q(s t , a t )=CNN(s t )

[0129] Among them, s t Indicates the current state, a t Indicates the action of selection.

[0130] In one example, the analysis and prediction model of the deep learning algorithm is automatically adjusted according to the changes in the multi-source spatiotemporal data input in real time based on the reinforcement learning mechanism, and real-time analysis and trend prediction are performed based on the updated analysis and prediction model.

[0131] Specifically, adaptive learning and real-time feedback mechanism: The reinforcement learning mechanism can automatically adjust the analysis and prediction models according to the changes in the real-time data stream to improve the accuracy and reliability of future predictions. When new data enters, the analysis and prediction models of the deep learning algorithm will be adaptively updated to ensure that the analysis results are consistent with the actual situation. At the same time, it supports real-time feedback mechanism, and users can adjust the display perspective or analysis scope as needed through the interactive interface to obtain the most appropriate decision support information.

[0132] It should be noted that the method described in this embodiment can also realize user interaction. Natural language and voice interaction: Combined with NLP technology, the interaction between voice and natural language is realized. Users can query system analysis results, adjust display perspectives, etc. through voice commands or natural language input. The voice recognition engine will respond in real time to improve the user's convenience and experience. Customized data reports and recommendations: Generate customized data analysis reports based on the user's interaction needs. The reports include visual charts, trend curves and forecast suggestions, and can be exported or shared with one click. Using the AI ​​recommendation algorithm, relevant data display suggestions and decision support will be provided based on the user's historical operation records. Dynamic interactive interface: Users can freely zoom, drag, and rotate three-dimensional models through the interface, and adjust the analysis view in real time. It has fast response speed and smooth operation, bringing excellent user experience. It supports setting personalized parameters (such as display style, timeline, etc.) to meet the needs of different users.

[0133] It is understandable that in this embodiment, an efficient distributed computing framework is adopted: Hadoop and Spark and other distributed computing frameworks are combined to perform parallel processing of large-scale spatiotemporal data, and cluster architecture is used to improve computing efficiency and ensure efficient processing of big data. Security and privacy protection mechanisms are set up: encryption technology and user authority control mechanisms are adopted throughout the entire process of data transmission and processing to ensure the confidentiality and security of data. Sensitive data is managed in a hierarchical manner, and data compliance and industry standards are strictly followed.

[0134] In this embodiment, multi-source spatiotemporal data, including satellite remote sensing, meteorological, geographical, and traffic data, are obtained, and the data is standardized and three-dimensionally modeled through a GIS platform (such as Cesium); based on standardized data, artificial intelligence algorithms are used to perform real-time analysis and trend prediction of spatiotemporal data, and identify dynamic factors such as potential risks, environmental changes, and traffic flow; based on the analysis and prediction results, decision support suggestions are generated, which can be applied to urban management, natural disaster warning, and environmental monitoring and other fields; the analysis results are displayed in a three-dimensional dynamic visualization form through a GIS platform, including real-time data display, dynamic simulation of prediction results, etc.; according to real-time changing data, the prediction results are automatically updated, and the user is allowed to adjust the plan through an interactive interface. The method described in this embodiment can realize intelligent analysis and dynamic display of spatiotemporal data, provide efficient and intuitive decision support, and is suitable for scenarios such as smart city management and disaster warning.

[0135] It can be understood that this embodiment has the following beneficial effects: Improve data processing efficiency: It can efficiently integrate multi-source spatiotemporal data, adopt stream processing architecture and distributed computing framework, realize real-time processing of large-scale data, significantly improve data response speed, reduce data lag, and integrate data from different fields (such as meteorology, transportation, geology, etc.) and conduct unified analysis. Traditional technology is difficult to achieve cross-domain multi-source data integration. The analysis results after data fusion in this embodiment are more comprehensive, providing effective support for complex decision-making problems and improving the comprehensiveness and accuracy of decision-making. Intelligent data analysis: Through artificial intelligence technologies such as deep learning and integrated learning, it can automatically identify trends, rules and anomalies in data and make predictions, greatly reducing manual intervention, improving the automation and accuracy of analysis, helping decision makers to respond faster, and performing intelligent analysis and prediction of real-time data through artificial intelligence algorithms, and dynamically displaying analysis results, improving the timeliness and accuracy of data processing. Compared with traditional technologies, this embodiment can better adapt to rapidly changing environments and needs. Dynamic visualization effects: It can display dynamic spatiotemporal data in real time, including data changes, trend simulation and future predictions, support interactive operations, help users understand data and trends more intuitively, and enhance the scientificity and timeliness of decision-making. Intelligent decision support: The decision support function based on AI analysis can provide accurate early warning, trend prediction and response suggestions, which is particularly suitable for disaster warning, traffic management and other fields, helping users make scientific decisions in complex and sudden situations. Intelligent disaster warning and response: Based on AI analysis capabilities, it provides real-time disaster warning and risk assessment, timely detects potential dangers and proposes response plans, significantly improving the response speed and accuracy of the disaster monitoring system. Intelligent recommendation and automated decision-making: AI-driven intelligent recommendation and automated decision-making functions have greatly improved urban planning and management efficiency, reduced dependence on manual operations, and provided customized decision support to help decision makers make more scientific and accurate judgments.

[0136] This embodiment provides a multi-dimensional spatiotemporal data analysis and visualization method, including: obtaining multi-source spatiotemporal data, and standardizing the multi-source spatiotemporal data to obtain processed spatiotemporal data; performing real-time analysis and trend prediction based on the processed spatiotemporal data based on a deep learning algorithm to obtain analysis and prediction results; performing three-dimensional modeling based on the processed spatiotemporal data and the analysis and prediction results through a GIS platform to obtain a three-dimensional dynamic simulation model and perform visual display; generating a decision strategy based on the analysis and prediction results, and optimizing the decision strategy based on a reinforcement learning mechanism. In this embodiment, multi-source spatiotemporal data is efficiently integrated to achieve real-time processing of large-scale data, significantly improve data response speed, and reduce data lag. Prediction is performed based on multi-source data fusion and deep learning algorithms to improve the automation and accuracy of analysis. Visual real-time dynamic display helps users understand data and trends more intuitively, enhances the scientificity and timeliness of decision-making, and solves problems such as low disaster prediction accuracy, difficulty in capturing nonlinear evolution processes, and lack of multi-objective optimization of emergency decision-making.

[0137] In addition, an embodiment of the present invention further proposes a storage medium, on which a multidimensional spatiotemporal data analysis and visualization program is stored. When the multidimensional spatiotemporal data analysis and visualization program is executed by a processor, the steps of the multidimensional spatiotemporal data analysis and visualization method described above are implemented.

[0138] Reference Figure 4 , Figure 4 It is a structural block diagram of an embodiment of the multi-dimensional spatiotemporal data analysis and visualization system of the present invention.

[0139] like Figure 4 As shown, the multidimensional spatiotemporal data analysis and visualization system comprises:

[0140] The data access module 10 is used to obtain multi-source spatiotemporal data and perform standardization processing on the multi-source spatiotemporal data to obtain processed spatiotemporal data;

[0141] The analysis and prediction module 20 is used to perform real-time analysis and trend prediction based on the processed spatiotemporal data based on a deep learning algorithm to obtain analysis and prediction results;

[0142] A visualization display module 30 is used to perform three-dimensional modeling based on the processed spatiotemporal data and analysis and prediction results through a GIS platform to obtain a three-dimensional dynamic simulation model and perform visualization display;

[0143] The reinforcement learning module 40 is used to generate a decision strategy according to the analysis and prediction results, and optimize the decision strategy based on the reinforcement learning mechanism.

[0144] This embodiment provides a multi-dimensional spatiotemporal data analysis and visualization system. In this embodiment, the system realizes real-time processing of large-scale data by efficiently integrating multi-source spatiotemporal data, significantly improving data response speed and reducing data lag. Predictions are made based on multi-source data fusion and deep learning algorithms to improve the automation and accuracy of analysis. Visual real-time dynamic display helps users understand data and trends more intuitively, enhances the scientificity and timeliness of decision-making, and solves problems such as low disaster prediction accuracy, difficulty in capturing nonlinear evolution processes, and lack of multi-objective optimization of emergency decision-making.

[0145] It should be noted that the technical details that are not described in detail in the embodiment of the multidimensional spatiotemporal data analysis and visualization system can be found in the multidimensional spatiotemporal data analysis and visualization method provided in any embodiment of the present invention as described above, and will not be repeated here.

[0146] It should be understood that the above is only an example and does not constitute any limitation on the technical solution of the present invention. In specific applications, technicians in this field can make settings as needed, and the present invention does not limit this.

[0147] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of the present invention. In practical applications, technicians in this field can select part or all of them according to actual needs to achieve the purpose of the present embodiment, and no limitation is made here.

[0148] In addition, it should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0149] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0150] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory (ROM) / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0151] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A multidimensional spatiotemporal data analysis and visualization method, characterized in that: include: Acquire multi-source spatiotemporal data, and perform standardization processing on the multi-source spatiotemporal data to obtain processed spatiotemporal data; Based on the deep learning algorithm, real-time analysis and trend prediction are performed according to the processed spatiotemporal data to obtain analysis and prediction results; Performing three-dimensional modeling based on the processed spatiotemporal data and analysis and prediction results through a GIS platform to obtain a three-dimensional dynamic simulation model and perform visual display; A decision strategy is generated according to the analysis and prediction results, and the decision strategy is optimized based on a reinforcement learning mechanism.

2. The method according to claim 1, characterized in that The acquiring multi-source spatiotemporal data and performing standardization processing on the multi-source spatiotemporal data to obtain processed spatiotemporal data includes: Acquire satellite remote sensing data, meteorological data, ground sensor data, and media data from multiple heterogeneous data sources based on a stream processing architecture; The satellite remote sensing data, meteorological data, ground sensor data and media data are uniformly processed by using ETL tools to obtain multi-source spatiotemporal data; Data standardization is performed on the multi-source spatiotemporal data to obtain processed spatiotemporal data.

3. The method according to claim 2, characterized in that The step of performing data standardization processing on the multi-source spatiotemporal data to obtain processed spatiotemporal data includes: Performing image denoising and data cleaning on the multi-source spatiotemporal data to obtain preprocessed data; Normalizing pixel values ​​of image data in the preprocessed data to obtain image normalized data; Performing z-score standardization on the time series data in the preprocessed data to obtain time series normalized data; The processed spatiotemporal data is obtained according to the image normalization data and the time series normalization data.

4. The method according to claim 1, characterized in that The deep learning algorithm is used to perform real-time analysis and trend prediction based on the processed spatiotemporal data to obtain analysis and prediction results, including: Based on the graph neural network, data fusion is performed on the processed spatiotemporal data to obtain disaster scenario prediction features; Based on the spatiotemporal convolutional network, the processed spatiotemporal data is subjected to spatiotemporal joint modeling to obtain spatiotemporal comprehensive features; Real-time analysis and trend prediction are performed based on the disaster scenario prediction characteristics and spatiotemporal comprehensive characteristics to obtain analysis and prediction results.

5. The method according to claim 4, characterized in that The processing of the spatiotemporal data based on the graph neural network is performed to fuse the processed spatiotemporal data to obtain disaster scenario prediction features, including: Representing the processed spatiotemporal data as a graph structure; Performing feature learning on each node of the graph structure through a graph convolutional network to construct an adjacency matrix; In each layer of the graph convolutional network, the node features are updated through the information transmission mechanism; In the last layer of the graph convolutional network, a graph pooling operation is performed to obtain a global feature representation of the graph structure; The global feature representation is used as a disaster scenario prediction feature.

6. The method according to claim 4, characterized in that The spatiotemporal convolutional network is used to perform spatiotemporal joint modeling on the processed spatiotemporal data to obtain spatiotemporal comprehensive features, including: Based on a convolutional neural network, spatial feature modeling is performed on the processed spatiotemporal data to obtain spatial features of the disaster area; Based on the long short-term memory network, time series modeling is performed on the processed spatiotemporal data to obtain the time characteristics of disaster evolution; Based on the spatial characteristics of the disaster area and the temporal characteristics of the disaster evolution, spatiotemporal joint modeling is performed to obtain spatiotemporal comprehensive characteristics.

7. The method according to claim 1, characterized in that The three-dimensional modeling is performed based on the processed spatiotemporal data and the analysis and prediction results through the GIS platform to obtain a three-dimensional dynamic simulation model and perform a visual display, including: Performing three-dimensional modeling on the analysis and prediction results through Cesium's three-dimensional rendering engine to obtain an analysis result simulation model; Dynamically displaying three-dimensional data of the analysis result simulation model based on a particle system; Constructing a dynamic heat map and trend curve based on the processed spatiotemporal data through a GIS platform; The dynamic heat map and trend curve are visualized by combining graphic transformation and smooth transition effect technology.

8. The method according to any one of claims 1 to 7, characterized in that Generating a decision strategy according to the analysis and prediction results, and optimizing the decision strategy based on a reinforcement learning mechanism, comprises: Construct a multi-objective decision-making model for disaster emergency response; Generate a decision strategy based on the multi-objective decision model according to the analysis and prediction results; Optimizing the multi-objective decision-making model based on a reinforcement learning algorithm to obtain an updated multi-objective decision-making model; An optimized decision strategy is obtained based on the updated multi-objective decision model and the analysis and prediction results.

9. A multidimensional spatiotemporal data analysis and visualization system, characterized in that: include: A data access module is used to obtain multi-source spatiotemporal data and perform standardization processing on the multi-source spatiotemporal data to obtain processed spatiotemporal data; The analysis and prediction module is used to perform real-time analysis and trend prediction based on the processed spatiotemporal data based on the deep learning algorithm to obtain the analysis and prediction results; A visualization display module is used to perform three-dimensional modeling based on the processed spatiotemporal data and analysis and prediction results through a GIS platform to obtain a three-dimensional dynamic simulation model and perform visualization display; The reinforcement learning module is used to generate a decision strategy according to the analysis and prediction results, and optimize the decision strategy based on the reinforcement learning mechanism.

10. An electronic device, characterized in that: The electronic device includes: a memory, a processor, and a multidimensional spatiotemporal data analysis and visualization program stored in the memory and executable on the processor, wherein the multidimensional spatiotemporal data analysis and visualization program is configured to implement the multidimensional spatiotemporal data analysis and visualization method as described in any one of claims 1 to 8.

Images