A Dynamic Visualization Method and System Based on Intelligent Overlay of Multiple Meteorological Remote Sensing Layers
By using a multi-source heterogeneous access engine and dynamic preloading technology, the problem of poor multi-source data processing in meteorological disaster monitoring has been solved, achieving efficient dynamic visualization and interactive simulation, and improving the efficiency of disaster analysis and decision-making.
Patent Information
- Application Number
- CN202611568220.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-12
- Publication Date
- 2026-11-10
AI Technical Summary
Existing technologies are insufficient for efficient collaborative processing and dynamic visualization of multi-source data in meteorological disaster monitoring and early warning, resulting in low efficiency in disaster analysis and decision-making, especially in the case of rapidly changing disasters, where issues such as screen lag and loading delays often occur.
The system receives cross-platform meteorological observation data through a multi-source heterogeneous access engine, segments it into multi-source collaborative observation slices based on the consistency of meteorological attributes, performs confidence verification of meteorological elements within the group, and then fuses and overlays the slices for rendering. It also constructs a dynamic meteorological slice index and tracks user interaction behavior for pre-loading and caching, thereby optimizing the collaborative processing and interactive response of multi-source data.
It has achieved high-precision dynamic visualization and interactive simulation of the evolution process of meteorological disasters, enhanced the real-time response capability of user interaction operations, and optimized the collaborative processing efficiency of multi-source heterogeneous data.
Smart Images

Figure CN123068670A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of image visualization technology, and in particular to a dynamic visualization method and system based on intelligent overlay of multiple layers of meteorological remote sensing. Background Technology
[0002] In the field of meteorological disaster monitoring and early warning, existing technologies usually adopt a visualization method that uses a single data source to statically present data or simply overlay multiple data sources, which makes it difficult to accurately reflect the dynamic evolution process of the disaster system.
[0003] Meanwhile, traditional time-series stitching methods lack real-time interactive response capabilities, making it impossible for users to smoothly review historical states or predict evolution trends. Especially when dealing with rapidly changing disasters such as typhoons and rainstorms, issues such as screen lag and loading delays often occur, seriously affecting the efficiency of disaster analysis and decision-making.
[0004] In summary, existing technologies have significant technical bottlenecks in areas such as multi-source data collaborative processing, dynamic visualization rendering, and interactive optimization. Summary of the Invention
[0005] This invention provides a dynamic visualization method and system based on intelligent overlay of multiple layers of meteorological remote sensing, which addresses the technical problems of poor dynamic visualization and reproduction of the evolution process of meteorological disasters in existing technologies and the inconvenience of interactive operation.
[0006] In view of the above problems, the present invention provides a dynamic visualization method and system based on intelligent overlay of multiple layers of meteorological remote sensing.
[0007] A first aspect of the present invention provides a dynamic visualization method based on intelligent overlay of multiple layers of meteorological remote sensing data, the method comprising:
[0008] The system receives cross-platform meteorological observation data via a multi-source heterogeneous access engine. This cross-platform meteorological observation data includes multiple meteorological observation data streams from multiple meteorological observation data sources. Based on the consistency of meteorological attributes, the multiple meteorological observation data streams are divided into P groups of multi-source collaborative observation slices across the platform. After performing confidence verification of meteorological elements within each group, the P groups of multi-source collaborative observation slices are fused and overlaid with multiple image layers to obtain multiple visualized dynamic meteorological slices. After constructing frame-level slice indexes for the multiple visualized dynamic meteorological slices, the multiple visualized dynamic meteorological slices are sequentially connected to output a dynamic meteorological disaster projection map. The system tracks the user's spatiotemporal interaction behavior trajectory within the dynamic meteorological disaster projection map and outputs meteorological preloading features. The meteorological preloading features are used to traverse the dynamic meteorological disaster projection map for hierarchical cache preloading.
[0009] A second aspect of the present invention provides a dynamic visualization system based on intelligent overlay of multiple layers of meteorological remote sensing data, the system comprising:
[0010] The system includes a data receiving unit for receiving cross-platform meteorological observation data via a multi-source heterogeneous access engine, wherein the cross-platform meteorological observation data includes multiple meteorological observation data streams from multiple meteorological observation data sources; a data slicing unit for dividing the multiple meteorological observation data streams into P groups of multi-source collaborative observation slices across platforms based on the consistency of meteorological attributes; a fusion rendering unit for performing multi-image layer fusion and overlay rendering on the P groups of multi-source collaborative observation slices after performing confidence verification of meteorological elements within the group, thereby obtaining multiple visualized dynamic meteorological slices; a map construction unit for constructing frame-level slice indexes for the multiple visualized dynamic meteorological slices, sequentially connecting the multiple visualized dynamic meteorological slices, and outputting a dynamic meteorological disaster prediction map; a preloading unit for tracking the user's spatiotemporal interaction behavior trajectory in the dynamic meteorological disaster prediction map and outputting meteorological preloading features; and a map caching unit for using the meteorological preloading features to traverse the dynamic meteorological disaster prediction map and perform hierarchical caching preloading.
[0011] One or more technical solutions provided in this invention have at least the following technical effects or advantages:
[0012] The method provided in this invention receives cross-platform meteorological observation data via a multi-source heterogeneous access engine, and then, based on the consistency of meteorological attributes, divides the multiple meteorological observation data streams into P groups of multi-source collaborative observation slices across the platform. After performing confidence verification of meteorological elements within each group, the P groups of multi-source collaborative observation slices are fused and overlaid with multiple image layers to obtain multiple visualized dynamic meteorological slices. After constructing frame-level slice indexes for the multiple visualized dynamic meteorological slices, the multiple visualized dynamic meteorological slices are sequentially connected to output a dynamic meteorological disaster prediction map. The spatiotemporal interaction trajectory of the user in the dynamic meteorological disaster prediction map is tracked, and meteorological preloading features are output. The meteorological preloading features are used to traverse the dynamic meteorological disaster prediction map for hierarchical caching and preloading. This achieves the technical effect of high-precision dynamic visualization and interactive prediction of the evolution process of meteorological disasters, while optimizing the collaborative processing efficiency of multi-source heterogeneous data and enhancing the real-time response capability of user interactive operations. Attached Figure Description
[0013] Figure 1 A schematic diagram of the dynamic visualization method based on intelligent overlay of multiple layers of meteorological remote sensing provided by the present invention is shown.
[0014] Figure 2 A schematic diagram of the structure of the dynamic visualization system based on intelligent overlay of multiple layers of meteorological remote sensing provided by the present invention is shown.
[0015] Figure labeling: Data receiving unit 1, data slicing unit 2, fusion rendering unit 3, map construction unit 4, preloading unit 5, map caching unit 6. Detailed Implementation
[0016] This invention provides a dynamic visualization method and system based on intelligent overlay of multiple layers of meteorological remote sensing, which addresses the technical problems of poor dynamic visualization and reproduction of the evolution process of meteorological disasters in existing technologies and the inconvenience of interactive operation.
[0017] The technical solutions of the present invention will now be clearly and completely described with reference to the accompanying drawings. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention. It should also be noted that, for ease of description, only the parts related to the present invention are shown in the drawings, not all of them.
[0018] Example 1: A flowchart of the dynamic visualization method based on intelligent overlay of multiple meteorological remote sensing layers provided in this embodiment of the invention. (See attached diagram) Figure 1 The method includes:
[0019] Step A100: Receive cross-platform meteorological observation data via a multi-source heterogeneous access engine, wherein the cross-platform meteorological observation data includes multiple meteorological observation data streams from multiple meteorological observation data sources.
[0020] In one implementation, the observation data stream type of the cross-platform meteorological observation data is a heterogeneous data stream including but not limited to satellite remote sensing, ground-based observation and numerical model output.
[0021] Specifically, the multi-source heterogeneous access engine refers to a real-time data receiving and processing entity designed specifically for cross-platform integration of meteorological data. Its core function is to connect meteorological observation data sources with different protocols, formats, and transmission mechanisms.
[0022] The cross-platform meteorological observation data specifically refers to a diverse set of data originating from satellite remote sensing systems (such as the Fengyun series meteorological satellites), ground-based observation networks (including meteorological radar stations and automatic weather stations), and numerical weather prediction models (such as the output of WRF and ECMWF models). These data are input in the form of continuous streams, each following different industry standard data formats, forming observation data streams with varying spatiotemporal coverage and resolutions.
[0023] Furthermore, this embodiment clarifies the types of observation data streams covered by cross-platform meteorological observation data, and specifically lists three typical data sources: satellite remote sensing data streams (including Earth observation data in the visible light, infrared and microwave bands), ground-based observation data streams (real-time monitoring data of elements such as temperature, humidity, wind speed, and precipitation collected by ground meteorological facilities), and numerical model output data streams (forecast products such as wind fields and temperature fields generated by numerical weather prediction models).
[0024] Step A200: Based on the consistency of meteorological attributes, the multiple meteorological observation data streams are divided into P groups of multi-source collaborative observation slices across the platform.
[0025] In one implementation, based on the consistency of meteorological attributes, the multiple meteorological observation data streams are divided into P groups of multi-source collaborative observation slices across platforms. The method step A200 provided by this invention includes:
[0026] Step A210: Based on meteorological attribute identification, the multiple meteorological observation data streams are divided into multiple sets of meteorological observation slices.
[0027] Step A220: Locate the global slice segmentation node by aligning the multiple sets of meteorological observation slices in time and space.
[0028] Step A230: Based on the global slice segmentation node, perform cross-platform secondary segmentation of the multiple sets of meteorological observation slices after spatiotemporal alignment to obtain the P sets of multi-source collaborative observation slices.
[0029] In one implementation, based on meteorological attribute identification, the multiple meteorological observation data streams are segmented into multiple sets of meteorological observation slices. Step A210 of the method provided by this invention includes:
[0030] Step A211: Based on the custom segmentation granularity, divide the first meteorological observation data stream into N first-granularity data streams.
[0031] Step A212: Extract key meteorological element features from the N first-granularity data streams to obtain N multi-dimensional meteorological element features.
[0032] Step A213: By loading the N multidimensional meteorological element features into the pre-constructed meteorological attribute particle space, the probability distribution of the N meteorological elements is matched and output.
[0033] Step A214: Calculate the meteorological attribute transition probability for the probability distribution of the N meteorological elements and output N meteorological attribute features.
[0034] Step A215: Based on the consistency of meteorological attributes, the N first-granularity data streams are spliced together according to the N meteorological attribute features to output the first set of meteorological observation slices.
[0035] Specifically, this embodiment first defines custom segmentation granularity parameters that match the target meteorological phenomenon, including spatial resolution and temporal window size. Based on the custom segmentation granularity parameters, the first meteorological observation data stream (such as satellite infrared brightness temperature data stream) is divided into several primary data units, each covering a specific spatiotemporal range, ensuring that the variation amplitude of meteorological elements within the unit does not exceed a preset threshold, and outputting the N first-granularity data streams.
[0036] For the N first-granularity data streams generated by the segmentation, a key meteorological element feature extraction operation is performed. Specifically, feature indicators strongly correlated with the target attribute (such as the central pressure gradient, cloud top temperature extreme value, wind field curl, etc. in the typhoon scenario) are extracted from the multimodal data of each granularity data stream. A multidimensional feature vector representing the meteorological state of the granularity data stream is constructed to obtain N multidimensional meteorological element features. The obtained N multidimensional meteorological element features are used as input for subsequent attribute recognition.
[0037] By loading the N multidimensional meteorological element features into a pre-constructed meteorological attribute particle space, N meteorological element probability distributions are matched and output. The meteorological element probability distribution is the probability distribution of an initial data unit belonging to different meteorological types (such as typhoon eye area, outer rainband). The technical implementation of this embodiment will be described in detail in the following description.
[0038] Meteorological attribute transfer probability calculation is performed on the probability distribution of the N meteorological elements to correct the attribute probability distribution of isolated abnormal units, enhance the continuity of the segmentation results in the spatiotemporal dimension, and output N meteorological attribute features. The technical implementation of this embodiment will be described in detail in the following description.
[0039] Based on the corrected meteorological attribute characteristics, a merging operation is performed on the first-granularity data streams that have consistent attribute labels and are spatiotemporally adjacent. A boundary optimization algorithm is used to eliminate the jagged edges remaining from the cutting process, generating meteorological observation slices with smooth boundaries. This ensures the uniformity and physical rationality of meteorological attributes within the slices, and the first set of meteorological observation slices is output.
[0040] Similarly, based on meteorological attribute identification, the multiple meteorological observation data streams are divided into multiple sets of meteorological observation slices.
[0041] After obtaining multiple sets of meteorological observation slices through initial segmentation, it is necessary to eliminate the benchmark differences between multi-source data through spatiotemporal alignment, project meteorological observation slices from different sources onto a unified spatiotemporal coordinate system, detect the consistency of attributes in overlapping areas, and identify global slice segmentation nodes with cross-platform indicative significance (such as typhoon eye center coordinates and frontal convergence lines) to provide global anchor points for cross-source collaborative segmentation.
[0042] Guided by global segmentation nodes, a two-level collaborative segmentation is performed on the spatiotemporally aligned multi-source slices to generate P groups of multi-source collaborative observation slices that integrate multi-source information, ensuring that the cross-source observation data within each group of slices has interactively verifiable physical consistency.
[0043] Based on the continuous pattern of meteorological elements in spatiotemporal distribution, this embodiment performs collaborative segmentation on raw data streams from different platforms to form P groups of multi-source collaborative observation slices with consistent spatiotemporal attributes and cross-source fusion. This achieves the technical effect of transforming heterogeneous data streams into a set of multi-source collaborative observation slices with unified spatiotemporal attributes, providing standardized data input for subsequent visualization and simulation.
[0044] Step A300: After performing confidence verification of meteorological elements within the P group of multi-source collaborative observation slices, perform multi-image layer fusion and overlay rendering to obtain multiple visualized dynamic meteorological slices.
[0045] In one embodiment, after performing confidence verification of meteorological elements within the P group of multi-source collaborative observation slices, multi-image layer fusion and overlay rendering are performed to obtain multiple visualized dynamic meteorological slices. The method step A300 provided by the present invention includes:
[0046] Step A310: Interact to obtain multiple sets of meteorological confidence weights from the multiple meteorological observation data sources.
[0047] Step A320: Based on the multiple sets of meteorological confidence weights, perform confidence verification of meteorological elements within the P-group multi-source collaborative observation slices to obtain the P-group multi-source confidence features.
[0048] Step A330: Based on the P-group multi-source confidence features, dynamically adjust the transparency of the P-group multi-source collaborative observation slices during the multi-image layer fusion and overlay rendering process to obtain multiple visualized dynamic meteorological slices.
[0049] Specifically, this embodiment establishes a dynamic weight allocation matrix for different meteorological element categories (such as temperature, precipitation, and wind field) and the observation advantages of each data source.
[0050] For example, in the dimension of temperature, the confidence weight of satellite infrared hyperspectral observation temperature data is higher than that of ground-based observations. In the dimension of precipitation intensity, the weight of ground-based radar reflectivity data is higher than that of numerical model output. For wind field, the weight of numerical model data decays exponentially over the forecast lead time. This weight matrix is dynamically updated through historical data verification tests (such as error analysis of temperature comparison between satellites and radiosondes) and real-time quality monitoring (such as radar transmit power stability detection), forming a differentiated confidence assessment system that is associated with meteorological element type, spatiotemporal location, and equipment status.
[0051] Based on this, by decomposing the obtained dynamic weight allocation matrix, multiple sets of meteorological confidence weights are obtained from the multiple meteorological observation data sources.
[0052] Within a unified spatiotemporal grid framework, weighted cross-validation is performed on specific meteorological element categories within each of the P-group multi-source collaborative observation slices. For example, in temperature grid validation, satellite data is assigned a higher weight, and the weighted median is calculated as the baseline value. In precipitation intensity grid validation, radar data is assigned a higher weight. For anomalous observations that deviate from the baseline value by more than three times the weighted average absolute deviation, the removal threshold is dynamically adjusted based on the weight of its data source—observations from high-weight sources require even greater deviations to be marked as anomalous.
[0053] After verification, P sets of multi-source confidence features are output. Each multi-source confidence feature includes the feature value, confidence score, and the identifier of the data source participating in the verification.
[0054] Based on the P-group multi-source confidence features, the transparency of the P-group multi-source collaborative observation slices is dynamically adjusted during the multi-image layer fusion and overlay rendering process to dynamically adjust the layer visibility and obtain multiple visualized dynamic meteorological slices.
[0055] After completing cross-platform segmentation, this embodiment assesses the credibility of meteorological elements within each slice and enhances the visualization of disaster characteristics by dynamically adjusting layer overlay parameters. This achieves the technical effect of ensuring that the final generated visualized slices accurately reflect the real meteorological conditions while highlighting key disaster information.
[0056] Step A400: After constructing the frame-level slice index of the multiple visualized dynamic meteorological slices, the multiple visualized dynamic meteorological slices are sequentially connected to output a dynamic meteorological disaster projection map.
[0057] This embodiment constructs a frame-level slice index structure by assigning a unique timestamp and spatial range identifier to each visualized dynamic meteorological slice. This index supports fast retrieval based on time order or spatial location.
[0058] After sorting the discrete slices by time axis, an optical flow interpolation algorithm is used to fill in the transition frames between adjacent slices, eliminating the visual discontinuity phenomenon during time-series playback, and finally generating a continuously playable dynamic meteorological disaster projection map. The dynamic meteorological disaster projection map integrates a disaster intensity heat map, a movement path trajectory line, and a disaster impact range vector layer, realizing a full-cycle visualization and reproduction of the disaster evolution process.
[0059] Step A500: Track the user's spatiotemporal interactive behavior trajectory on the meteorological disaster dynamic projection map and output meteorological preload features.
[0060] This embodiment captures user actions (such as dragging the timeline, zooming in and out, and selecting key areas) to extract spatiotemporal feature vectors of the operation trajectory, including movement direction angle, instantaneous speed, dwell time distribution, and historical access hotspots. Using an existing LSTM network model, the trajectory patterns are analyzed to predict potentially accessed spatiotemporal areas (such as the typhoon eye location or heavy rain areas in neighboring provinces in the next instant). The output includes meteorological preloading features containing the predicted spatiotemporal range, resolution requirements, and data priority, providing a basis for subsequent caching decisions.
[0061] Step A600: Use the meteorological preloading features to traverse the dynamic projection map of meteorological disasters and perform hierarchical cache preloading.
[0062] In this embodiment, based on the spatiotemporal access priority indicated by meteorological preloading characteristics, hierarchical caching management is implemented for the data resources covered by the meteorological disaster dynamic projection map, so as to achieve a balance between storage resource optimization and user experience assurance.
[0063] This embodiment achieves the technical effect of high-precision dynamic visualization and interactive simulation of the evolution process of meteorological disasters, while optimizing the collaborative processing efficiency of multi-source heterogeneous data and enhancing the real-time response capability of user interactive operations.
[0064] In one embodiment, by loading the N multidimensional meteorological element features into a pre-constructed meteorological attribute particle space and matching and outputting the probability distributions of the N meteorological elements, step A213 of the method provided by the present invention includes:
[0065] Step A2131: Interact to obtain the probability distribution of multiple sample elements of meteorological element characteristics of multiple samples.
[0066] Step A2132: Locate multiple spatial particles corresponding to the meteorological element features of the multiple samples in the meteorological attribute particle space.
[0067] Step A2133: Fill the multiple spatial particles with the probability distribution of the multiple sample elements to complete the local configuration of the meteorological attribute particle space.
[0068] Step A2134: Locate the first virtual particle of the first multidimensional meteorological element feature in the meteorological attribute particle space.
[0069] Step A2135: Based on the space between the first virtual particle and multiple spatial particle points, the probability distribution of the sample elements is weighted and fused to output the probability distribution of the first meteorological element.
[0070] Specifically, this embodiment integrates a historical meteorological event database and an expert experience knowledge base to obtain multiple sample meteorological element features covering typical weather phenomena (such as typhoons, rainstorms, sandstorms, etc.) and their corresponding weather type probability distribution labels, forming a sample element probability distribution set for use in particle space construction.
[0071] Based on the meteorological element indicators, after constructing a multi-dimensional meteorological attribute particle space, multiple spatial particles corresponding to the meteorological element characteristics of the multiple samples are located in the meteorological attribute particle space.
[0072] The probability distributions of the multiple sample elements are used to fill the multiple spatial particles, completing the local configuration of the meteorological attribute particle space. A first virtual particle representing the first multidimensional meteorological element feature is located within this meteorological attribute particle space. Based on the space between the first virtual particle and the multiple spatial particle points, the probability distributions of the sample elements are weighted and fused to output a first meteorological element probability distribution.
[0073] This embodiment is based on particle space registration to locate the probability distribution of meteorological elements, which achieves the technical effect of improving the accuracy of the obtained probability distribution of meteorological elements.
[0074] In one embodiment, the meteorological attribute transition probability is calculated for the probability distribution of the N meteorological elements, and N meteorological attribute features are output. Step A214 of the method provided by this invention includes:
[0075] Step A2141: Extract meteorological evolution patterns from historical meteorological data, and calculate attribute transition probabilities based on the extraction results to obtain the attribute transition probability matrix.
[0076] Step A2142: Based on the attribute transition probability matrix, perform state transition prediction on the probability distribution of the first meteorological element and output the probability distribution of the first transition element.
[0077] Step A2143: Sum the attribute distribution probabilities of the probability distribution of the first meteorological element and the probability distribution of the first transfer element, and extract the first meteorological attribute feature corresponding to the probability extreme value.
[0078] Specifically, by analyzing the weather state evolution sequences recorded in the historical meteorological event database (such as changes in eye pressure in the typhoon's path and the spatiotemporal migration of temperature gradients in the frontal system), the conversion frequency of different meteorological attributes (such as typhoon intensity level and precipitation type) between adjacent spatiotemporal units is statistically analyzed to construct an attribute transfer probability matrix.
[0079] Based on the probability distribution of the first meteorological element currently being processed (e.g., the probability of a certain spatiotemporal unit being identified as a typhoon is 70%, and the probability of a strong convective cloud cluster is 25%), combined with the corresponding state transition rules in the attribute transition probability matrix (e.g., the transition probability of tropical depression → typhoon is 0.6), the probability distribution evolution trend of the next time series or adjacent spatial units is predicted, and the output probability distribution of the first transition element is output.
[0080] The probability distributions of the first meteorological element and the probability distribution of the first transfer element are weighted and summed according to a preset ratio (e.g., 60% for current observation and 40% for transfer prediction) to obtain a joint probability distribution of meteorological attributes that comprehensively reflects the real-time observation characteristics and historical evolution patterns. Then, the first meteorological attribute feature corresponding to the probability extreme value is extracted from the joint probability distribution of meteorological attributes as the main meteorological attribute to be detected.
[0081] This embodiment achieves the technical effect of improving the continuity and consistency of meteorological attribute identification results in the spatiotemporal dimension, reducing the probability of misjudging isolated abnormal units, and enhancing the credibility of disaster evolution process inference.
[0082] Example 2 is based on the same inventive concept as the dynamic visualization method based on intelligent overlay of multiple layers of meteorological remote sensing in the previous examples, such as... Figure 2 As shown, this invention provides a dynamic visualization system based on intelligent overlay of multiple layers of meteorological remote sensing data, wherein the system includes:
[0083] The data receiving unit 1 is used to receive cross-platform meteorological observation data via a multi-source heterogeneous access engine, wherein the cross-platform meteorological observation data includes multiple meteorological observation data streams from multiple meteorological observation data sources.
[0084] Data slicing unit 2 is used to divide the multiple meteorological observation data streams across platforms into P groups of multi-source collaborative observation slices based on the consistency of meteorological attributes.
[0085] The fusion rendering unit 3 is used to perform multi-image layer fusion and overlay rendering on the P-group multi-source collaborative observation slices after verifying the confidence of meteorological elements within the group, so as to obtain multiple visualized dynamic meteorological slices.
[0086] Map construction unit 4 is used to construct frame-level slice indexes for the multiple visualized dynamic meteorological slices, and then sequentially connect the multiple visualized dynamic meteorological slices to output a dynamic meteorological disaster projection map.
[0087] The preloading unit 5 is used to track the user's spatiotemporal interactive behavior trajectory in the dynamic simulation map of meteorological disasters and output meteorological preloading features.
[0088] Map cache unit 6 is used to perform hierarchical cache preloading by traversing the dynamic inference map of meteorological disasters using the meteorological preloading features.
[0089] In one implementation, the observation data stream type of the cross-platform meteorological observation data is a heterogeneous data stream including but not limited to satellite remote sensing, ground-based observation and numerical model output.
[0090] In one embodiment, the data slicing unit 2 is further configured to:
[0091] Based on meteorological attribute identification, the multiple meteorological observation data streams are divided into multiple sets of meteorological observation slices; global slice segmentation nodes are located by spatiotemporal alignment of the multiple sets of meteorological observation slices; based on the global slice segmentation nodes, cross-platform secondary segmentation of the multiple sets of meteorological observation slices after spatiotemporal alignment is performed to obtain the P sets of multi-source collaborative observation slices.
[0092] In one embodiment, the data slicing unit 2 is further configured to:
[0093] Based on a custom segmentation granularity, the first meteorological observation data stream is divided into N first-granularity data streams; key meteorological element features are extracted from the N first-granularity data streams to obtain N multi-dimensional meteorological element features; the N multi-dimensional meteorological element features are loaded into a pre-constructed meteorological attribute particle space to match and output N meteorological element probability distributions; meteorological attribute transition probabilities are calculated on the N meteorological element probability distributions to output N meteorological attribute features; based on meteorological attribute consistency, the N first-granularity data streams are adjacently spliced according to the N meteorological attribute features to output the first set of meteorological observation slices.
[0094] In one embodiment, the data slicing unit 2 is further configured to:
[0095] The system interactively obtains multiple sample element probability distributions of multiple sample meteorological element features; locates multiple spatial particles corresponding to the multiple sample meteorological element features in the meteorological attribute particle space; fills the multiple spatial particles with the multiple sample element probability distributions to complete the local configuration of the meteorological attribute particle space; locates a first virtual particle of a first multidimensional meteorological element feature in the meteorological attribute particle space; and weights and fuses the sample element probability distributions according to the space between the first virtual particle and the multiple spatial particle points to output a first meteorological element probability distribution.
[0096] In one embodiment, the data slicing unit 2 is further configured to:
[0097] Historical meteorological data is used to extract meteorological evolution patterns, and attribute transition probabilities are calculated based on the extraction results to obtain an attribute transition probability matrix. Based on the attribute transition probability matrix, state transition prediction is performed on the probability distribution of the first meteorological element, and the probability distribution of the first transition element is output. The probability distributions of the first meteorological element and the probability distribution of the first transition element are summed to extract the first meteorological attribute feature corresponding to the probability extreme value.
[0098] In one embodiment, the fusion rendering unit 3 is further configured to:
[0099] Multiple sets of meteorological confidence weights are obtained interactively from the multiple meteorological observation data sources; based on the multiple sets of meteorological confidence weights, the confidence of meteorological elements within the P-group multi-source collaborative observation slices is verified to obtain the P-group multi-source confidence features; based on the P-group multi-source confidence features, the transparency of the P-group multi-source collaborative observation slices is dynamically adjusted in the multi-image layer fusion and overlay rendering process to obtain multiple visualized dynamic meteorological slices.
[0100] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.
Claims
1. A dynamic visualization method based on intelligent overlay of multiple layers of meteorological remote sensing data, characterized in that: include: The system receives cross-platform meteorological observation data via a multi-source heterogeneous access engine, wherein the cross-platform meteorological observation data includes multiple meteorological observation data streams from multiple meteorological observation data sources. Based on the consistency of meteorological attributes, the multiple meteorological observation data streams are divided into P groups of multi-source collaborative observation slices across the platform; After performing confidence verification of meteorological elements within the P group of multi-source collaborative observation slices, multi-image layer fusion and overlay rendering are performed to obtain multiple visualized dynamic meteorological slices. After constructing the frame-level slice index of the multiple visualized dynamic meteorological slices, the multiple visualized dynamic meteorological slices are sequentially connected to output a dynamic meteorological disaster projection map; Track the spatiotemporal interactive behavior trajectory of users in the dynamic simulation map of meteorological disasters, and output meteorological preload features; The meteorological preloading features are used to traverse the dynamic projection map of meteorological disasters for hierarchical caching and preloading.
2. The dynamic visualization method based on intelligent overlay of multiple layers of meteorological remote sensing as described in claim 1, characterized in that, The observation data stream types of the cross-platform meteorological observation data include, but are not limited to, heterogeneous data streams from satellite remote sensing, ground-based observation, and numerical model output.
3. The dynamic visualization method based on intelligent overlay of multiple meteorological remote sensing layers as described in claim 1, characterized in that, Based on the consistency of meteorological attributes, the multiple meteorological observation data streams are divided into P groups of multi-source collaborative observation slices across the platform, including: Based on meteorological attribute identification, the multiple meteorological observation data streams are divided into multiple meteorological observation slices; The global slice segmentation node is located by aligning the multiple sets of meteorological observation slices in time and space. Based on the global slice segmentation node, the multiple sets of meteorological observation slices are spatiotemporally aligned and then cross-platform secondary segmentation is performed to obtain the P sets of multi-source collaborative observation slices.
4. The dynamic visualization method based on intelligent overlay of multiple layers of meteorological remote sensing as described in claim 3, characterized in that, Based on meteorological attribute identification, the multiple meteorological observation data streams are segmented into multiple sets of meteorological observation slices, including: Based on a custom segmentation granularity, the first meteorological observation data stream is divided into N first-granularity data streams; Key meteorological element features are extracted from the N first-granularity data streams to obtain N multi-dimensional meteorological element features; By loading the N multidimensional meteorological element features into a pre-constructed meteorological attribute particle space, the probability distributions of the N meteorological elements are matched and output. Calculate the meteorological attribute transition probability for the probability distribution of the N meteorological elements and output N meteorological attribute features; Based on the consistency of meteorological attributes, the N first-granularity data streams are spliced together adjacently according to the N meteorological attribute features to output the first set of meteorological observation slices.
5. The dynamic visualization method based on intelligent overlay of multiple meteorological remote sensing layers as described in claim 4, characterized in that, By loading the N multidimensional meteorological element features into a pre-constructed meteorological attribute particle space, the probability distributions of the N meteorological elements are matched and output, including: Interactively obtain the probability distribution of multiple sample elements of meteorological element characteristics; Locate multiple spatial particles corresponding to the meteorological element features of the multiple samples in the meteorological attribute particle space; The probability distribution of the multiple sample elements is used to fill the multiple spatial particles, thereby completing the local configuration of the meteorological attribute particle space; The first virtual particle is used to locate the first multidimensional meteorological element feature in the meteorological attribute particle space. Based on the space between the first virtual particle and multiple spatial particle points, the probability distribution of the sample elements is weighted and fused to output the probability distribution of the first meteorological element.
6. The dynamic visualization method based on intelligent overlay of multiple layers of meteorological remote sensing as described in claim 5, characterized in that, The meteorological attribute transition probability is calculated for the probability distribution of the N meteorological elements, and N meteorological attribute features are output, including: Interact with historical meteorological data to extract meteorological evolution patterns, and calculate attribute transition probabilities based on the extraction results to obtain an attribute transition probability matrix; Based on the attribute transition probability matrix, state transition prediction is performed on the probability distribution of the first meteorological element, and the probability distribution of the first transition element is output. The probability distributions of the first meteorological element and the first transfer element are summed by attribute distribution probabilities, and the first meteorological attribute feature corresponding to the probability extreme value is extracted.
7. The dynamic visualization method based on intelligent overlay of multiple layers of meteorological remote sensing as described in claim 5, characterized in that, After performing confidence verification of meteorological elements within the P-group multi-source collaborative observation slices, multi-image layer fusion and overlay rendering are performed to obtain multiple visualized dynamic meteorological slices, including: Multiple sets of meteorological confidence weights from the various meteorological observation data sources are obtained interactively. Based on the multiple sets of meteorological confidence weights, the confidence verification of meteorological elements within the P-group multi-source collaborative observation slices is performed to obtain the P-group multi-source confidence features; Based on the P-group multi-source confidence features, the transparency of the P-group multi-source collaborative observation slices is dynamically adjusted during the multi-image layer fusion and overlay rendering process to obtain multiple visualized dynamic meteorological slices.
8. A dynamic visualization system based on intelligent overlay of multiple layers of meteorological remote sensing data, characterized in that: The steps for implementing the method according to any one of claims 1 to 7 include: The data receiving unit is used to receive cross-platform meteorological observation data via a multi-source heterogeneous access engine, wherein the cross-platform meteorological observation data includes multiple meteorological observation data streams from multiple meteorological observation data sources; The data slicing unit is used to divide the multiple meteorological observation data streams into P groups of multi-source collaborative observation slices across the platform based on the consistency of meteorological attributes. The fusion rendering unit is used to perform confidence verification of meteorological elements within the group on the P-group multi-source collaborative observation slices, and then perform multi-image layer fusion and overlay rendering to obtain multiple visualized dynamic meteorological slices. The map construction unit is used to construct the frame-level slice index of the multiple visualized dynamic meteorological slices, and then sequentially connect the multiple visualized dynamic meteorological slices to output a dynamic meteorological disaster projection map. The preloading unit is used to track the spatiotemporal interactive behavior trajectory of the user in the dynamic simulation map of meteorological disasters and output meteorological preloading features; The map cache unit is used to perform hierarchical cache preloading by traversing the dynamic inference map of meteorological disasters using the meteorological preloading features.
