A highland multi-source meteorological data intelligent fusion method and system

Abstract

This invention relates to the field of meteorological information processing and intelligent analysis technology, specifically to an intelligent fusion method and system for multi-source meteorological data from plateau regions. The method includes: using environmental constraints as the core, synchronously binding meteorological observations with environmental parameters during the data acquisition phase to form environmentally perceptive data units; completing temporal alignment based on environmental similarity and achieving spatial mapping through environmental correction to construct a spatiotemporally consistent observation structure; performing reliability stratification and dynamic hierarchical adjustment of observation units through local environmental consistency assessment, and then reconstructing features and dynamically adjusting weights using an environmental-driven mechanism to achieve collaborative fusion of multi-source data under temporal, spatial, and environmental constraints; and forming a highly adaptable and interpretable fusion result through environmental evolution prediction and feedback correction. This invention achieves a fully intelligent closed-loop process from data acquisition, processing, fusion to optimized output.

Description

Technical Field

[0001] This invention relates to the field of meteorological information processing and intelligent analysis technology, specifically to a method and system for intelligent fusion of multi-source meteorological data from plateau regions. Background Technology

[0002] Due to its unique topography and complex underlying surface conditions, the plateau region exhibits strong spatiotemporal variability in meteorological elements, making it difficult for traditional meteorological observations and single data sources to comprehensively and accurately depict its true state. With the increasing abundance of multi-source meteorological data, such as satellite remote sensing, automatic weather stations, reanalysis data, and mobile observations, effectively integrating these heterogeneous, heterogeneous, and spatiotemporally and spatially varying data to form a reliable, continuous, and physically consistent plateau meteorological dataset has become a key technical challenge for improving plateau meteorological monitoring and forecasting capabilities.

[0003] Chinese invention patent CN119961876B discloses a method for assimilating and fusing multi-source heterogeneous meteorological data, which includes the following steps: synchronously acquiring meteorological data from multiple data sources through a distributed data interface and preprocessing the data sources, including satellite remote sensing data, ground meteorological station data, and meteorological radar data; after completing the preprocessing of the meteorological data from multiple data sources, fusing the data from multiple data sources into meteorological element values ​​for each meteorological element based on the weighted least squares method; and assimilating the fused meteorological element values ​​with a numerical weather prediction model.

[0004] Existing data fusion methods are mostly based on spatiotemporal interpolation, statistical assimilation, or simple weighting. They often fail to fully consider the strong modulating effects of key factors such as altitude, radiation, and underlying surface on meteorological elements in the plateau environment. This leads to physical distortion or over-smoothing of the fusion results in areas with abrupt environmental changes or complex terrain, making it difficult to meet the actual needs of refined analysis and forecasting of plateau meteorology. Therefore, there is an urgent need to develop an intelligent fusion method that can deeply embed plateau environmental constraints into the fusion process and achieve synergy between data, environment, and mechanisms. This would enhance the environmental adaptability, spatiotemporal continuity, and physical interpretability of the fusion products, providing high-quality data support for plateau meteorological scientific research and operational applications. Summary of the Invention

[0005] The purpose of this invention is to address the problems existing in the background technology by proposing an intelligent fusion method and system for multi-source meteorological data in plateau regions.

[0006] The technical solution of this invention: an intelligent fusion method for multi-source meteorological data in plateau regions, comprising the following specific implementation steps:

[0007] S1. Synchronously collect plateau environmental status information corresponding to multi-source meteorological data to form environmentally correlated meteorological data units; complete time reference alignment based on environmental status similarity, and introduce an environmental correction mechanism to realize spatial location mapping, thus constructing an environmentally perceptive spatiotemporally consistent observation unit.

[0008] S2. Perform local environmental consistency assessment on each environmentally conscious spatiotemporally consistent observation unit and generate an environmental consistency index. Based on this index, divide the environmentally conscious spatiotemporally consistent observation unit into different environmental effectiveness levels and dynamically adjust its level based on the time evolution of the environmental state to form a hierarchical observation unit with environmental effectiveness constraints.

[0009] S3. Based on the environmental effectiveness level of the hierarchical observation unit, determine its weight scope for fusion; combined with the environmental state changes of the hierarchical observation unit, dynamically generate its fusion weight through the environmental driving mechanism, and carry out multi-source collaborative adjustment to form a dynamic fusion weight unit.

[0010] S4. Perform time-scale fusion within the same plateau environment stability range, and perform spatial-scale fusion by combining neighborhood points and environmental correction factors; at the same time, calculate the environmental constraint deviation and dynamically adjust the fusion weights to suppress non-environmental consistency anomalies, and output the multi-scale environmental constraint fusion results.

[0011] S5. Based on the environmental state changes of the fusion results, construct an environment-driven prediction model to obtain the predicted values; compare the predicted values ​​with short-term observations to obtain the error, and perform adaptive correction in combination with environmental consistency; perform neighborhood weighted smoothing on the correction results, and finally encapsulate and output the plateau adaptive fusion dataset.

[0012] Preferably, the plateau environmental status information corresponding to the multi-source meteorological data collected in step S1 includes:

[0013] Altitude range parameters used to reflect the vertical gradient characteristics of meteorological elements;

[0014] Parameters of underlying surface type used to describe the influence of surface properties on local meteorological conditions;

[0015] Pressure level characteristic parameters used to distinguish atmospheric conditions at different altitudes;

[0016] Radiation environment state parameters used to characterize the impact of strong solar radiation on observation results on plateaus.

[0017] Preferably, step S1, which involves completing time reference alignment based on environmental state similarity, specifically includes:

[0018] Construct an environmental difference metric between environmental state vectors, and calculate an environmental similarity index based on the difference metric and a preset environmental difference scale parameter;

[0019] When the environmental similarity index is higher than the preset threshold, it is determined that the corresponding observation data are in a similar plateau environment, and they are included in the same effective time reference window to complete the time series alignment.

[0020] Preferably, the spatial location mapping achieved by introducing an environmental correction mechanism in step S1 is specifically as follows:

[0021] Based on the original spatial coordinates, an environmental correction term is added, consisting of the product of the elevation difference correction amount and its corresponding elevation correction weight coefficient, and the product of the underlying surface difference correction amount and its corresponding underlying surface correction weight coefficient, thereby obtaining the effective spatial location after environmental correction.

[0022] Preferably, the environmental effectiveness levels in step S2 include at least: a high environmental consistency observation layer, a medium environmental consistency observation layer, and a low environmental consistency observation layer;

[0023] That is, the environmental consistency index is compared with the preset environmental consistency classification threshold. Environmentally perceptual spatiotemporal consistency observation units that are higher than the high consistency lower threshold are classified into the high environmental consistency observation layer, environmentally perceptual spatiotemporal consistency observation units that are between the medium consistency lower threshold and the high consistency lower threshold are classified into the medium environmental consistency observation layer, and environmentally perceptual spatiotemporal consistency observation units that are lower than the medium consistency lower threshold are classified into the low environmental consistency observation layer.

[0024] Preferably, the dynamic adjustment of the hierarchy based on the time evolution of the environmental state in step S2 is specifically as follows:

[0025] Calculate the change in environmental state of the environment-aware spatiotemporally consistent observation unit within a set environmental evolution time interval, and adjust the mapping relationship according to the change and the preset hierarchy;

[0026] The environmental effectiveness hierarchy of the environmentally perceptive spatiotemporally consistent observation unit is reassessed and adjusted.

[0027] The hierarchical adjustment mapping relationship responds to the continuous evolution trend of the environmental state and suppresses frequent hierarchical switching caused by short-term environmental fluctuations.

[0028] Preferably, the dynamic generation of the fusion weights through the environment-driven mechanism in step S3 is specifically as follows:

[0029] The changes in the plateau environmental state of the stratified observation units over time are input into the environmental response function;

[0030] This environmental response function is used to map changes in environmental state into standardized adjustment factors.

[0031] The initial dynamic fusion weight is calculated using this adjustment factor within the fusion weight scope corresponding to the hierarchical observation unit.

[0032] Preferably, the multi-source coordinated regulation in step S3 specifically involves:

[0033] The initial dynamic fusion weights of all hierarchical observation units participating in the fusion are processed using a weight adjustment function based on the environmental consistency index.

[0034] This weight adjustment function tends to maintain the existing weight configuration when the environment is highly consistent;

[0035] When a significant shift in the environment is detected, the weights are smoothly guided to tilt towards information sources that are more reliable for the current environment. Finally, the final fusion weights of all hierarchical observation units are obtained through normalization.

[0036] Preferably, the calculation of environmental constraint deviation in step S4 is specifically as follows:

[0037] First, a reference fusion value representing the ideal value under environmental consistency is obtained through neighborhood averaging, trend prediction, and environmental state weighting.

[0038] Then, the difference between the fusion result of the hierarchical observation unit and the reference fusion value is calculated, and the final environmental constraint deviation metric is obtained by combining the variance of the neighborhood environment vector of the hierarchical observation unit.

[0039] The technical solution of this invention: A plateau multi-source meteorological data intelligent fusion system, used to execute the above-mentioned plateau multi-source meteorological data intelligent fusion method, comprising:

[0040] The environmental correlation data acquisition and spatiotemporal alignment module is used to collect and bind meteorological data and plateau environmental status information, perform temporal alignment based on environmental similarity and spatial mapping based on environmental correction, and output environmentally perceptive spatiotemporal consistent observation units.

[0041] The environmental relevance modeling and feature extraction module is used to calculate the local environmental consistency index of the environmentally perceptive spatiotemporally consistent observation unit, and based on this, to perform environmental effectiveness stratification and realize dynamic adjustment of the stratification, and output the stratified observation unit with environmental effectiveness constraints.

[0042] The time series fusion and dynamic weight adjustment module is used to determine the weight scope of the hierarchical observation unit based on the hierarchical results, dynamically generate and coordinately adjust the multi-source fusion weight in combination with environmental changes, and output the dynamic fusion weight unit.

[0043] The spatial scale collaborative fusion and anomaly suppression module is used to perform temporal fusion within the environmental stability range, combine environmental correction factors for spatial fusion, and adjust weights based on environmental constraint deviation to suppress anomalies, outputting multi-scale environmental constraint fusion results.

[0044] The plateau adaptation feedback correction and final output module is used for prediction based on environmental evolution, adaptive correction and local smoothing through error feedback, and finally encapsulates and outputs the plateau adaptation fusion dataset.

[0045] Compared with the prior art, the above-mentioned technical solution of the present invention has the following beneficial technical effects:

[0046] This invention designs an intelligent fusion method and system for multi-source meteorological data from plateau regions. By deeply embedding plateau environmental constraints into the entire meteorological data fusion process, it effectively improves the physical consistency, environmental adaptability, and spatiotemporal continuity of the fusion results of multi-source meteorological data from plateau regions. Firstly, by synchronously binding environmental parameters such as altitude, underlying surface, air pressure, and radiation at the data acquisition source, an environmentally aware spatiotemporally consistent observation unit is constructed, fundamentally enhancing the interpretability of the data and providing a realistic and reliable environmental background for subsequent processing. Secondly, a time-series matching mechanism based on environmental similarity rather than simple timestamp alignment, and a spatial mapping method incorporating environmental correction, significantly reduce fusion bias caused by abrupt changes in the plateau environment or spatial heterogeneity, improving the physical rationality and reliability of spatiotemporal alignment. Furthermore, through dynamic layering of environmental consistency and adaptive weight adjustment, the system can intelligently identify and prioritize the use of highly reliable data, while preserving and reasonably suppressing potential anomalies or transitional states. This ensures the robustness of the fusion while avoiding false smoothing of real extreme meteorological signals. In addition, multi-scale collaborative fusion and plateau adaptability feedback correction mechanisms further enhance the responsiveness of the fusion results to complex terrain and local climate, and the output data has both overall continuity and local detail authenticity. Finally, the system forms a complete environment-driven, closed-loop optimized fusion process, which significantly enhances the usability, stability, and regional adaptability of plateau meteorological data products, providing a high-quality and highly interpretable data foundation for plateau meteorological monitoring, forecasting, and related disaster early warning. Attached Figure Description

[0047] Figure 1 This is a flowchart of a method for intelligent fusion of multi-source meteorological data in high-altitude areas proposed in this invention.

[0048] Figure 2 This is a system architecture diagram of a plateau multi-source meteorological data intelligent fusion system proposed in this invention. Detailed Implementation

[0049] Example 1, as Figure 1As shown, the specific implementation steps of the intelligent fusion method for multi-source meteorological data in plateau regions proposed in this invention are as follows:

[0050] S1. Addressing the challenge of directly and uniformly processing plateau meteorological data under complex environmental conditions, this study adopts the core principle of "environment first, data follow," introducing plateau environmental constraints during the data acquisition phase. Through synchronous binding of environmental states, spatiotemporal alignment guided by environmental consistency, and spatial mapping after environmental correction, a highly coupled observational data structure of environment, time, and space is constructed. The specific implementation process is as follows:

[0051] S11. By simultaneously collecting plateau environmental status information such as altitude, underlying surface, pressure level, and radiation state while generating multi-source meteorological data, environmental status is solidified as a component of the data structure. This ensures that each meteorological data point carries a clear environmental background, providing a naturally interpretable basic data unit for subsequent processing. Specifically:

[0052] Within the plateau meteorological monitoring area, meteorological data from different observation systems are collected in parallel. These observation systems include fixed automatic weather stations, satellite remote sensing inversion products, regional reanalysis meteorological data, and mobile observation terminals deployed along transportation routes or special areas.

[0053] Simultaneously with the generation of each meteorological observation data point, information on the plateau environment in which it is located is collected, enabling environmental perception and binding to the data at the source, and constructing environmentally correlated meteorological data units:

[0054] ;

[0055] Define the plateau environment state vector: ;

[0056] in, This represents an environmentally correlated meteorological data unit, that is, a meteorological observation source of type m at time t. i The generated environmentally correlated meteorological data unit; This represents the original meteorological observation value, i.e., the value of the m-th type of observation source at time t. i The meteorological element observation results obtained are used to describe the basic meteorological conditions of the plateau region, which may include single or combined meteorological elements such as temperature, air pressure, humidity, wind speed, wind direction or radiation intensity. Represents the plateau environmental state vector, relative to the original meteorological observations. The synchronously collected plateau environmental status set is used to characterize the macroscopic and local environmental background at the time of observation data generation and is the core input of the environmental constraint mechanism; This parameter represents the altitude range of the observation point within the plateau region, reflecting the vertical gradient characteristics of meteorological elements as they change with altitude. The underlying surface type parameter represents the characteristics of the underlying surface type in the area surrounding the observation point. It is used to describe the influence of surface properties on local meteorological conditions, such as grassland, bare land, snow and ice, water bodies, or built-up areas. It represents the characteristic parameters of the pressure level, the pressure level corresponding to meteorological observation data in the vertical structure of the atmosphere, and is used to distinguish the atmospheric state at different altitudes. It is especially suitable for the identification of multi-layer meteorological elements in remote sensing inversion or reanalysis data. This represents the radiation environment state parameters, with the observation point at time t. i The radiation environment conditions are used to characterize the impact of strong solar radiation in plateau areas on meteorological observation results. This parameter can be obtained from radiation observation data or radiation models.

[0057] S12. Utilize the continuity of the plateau environment to construct an environmental similarity discrimination mechanism, using environmental consistency rather than perfect timestamp consistency as the basis for time series alignment. This achieves time reference alignment of multi-source data while ensuring the physical rationality of meteorological evolution, reducing the risk of bias caused by conventional time interpolation under abrupt changes in plateau meteorological conditions. Specifically:

[0058] To address the differences in sampling frequency and timestamps among different observation sources, an environment similarity function is constructed based on the environment state vector: ;

[0059] when When the data exceeds a preset threshold, the corresponding observation data are considered to be in a similar plateau environment state, and are aligned as data within the same valid time reference window, without requiring the timestamps to be completely identical.

[0060] in, This represents the environmental similarity index, i.e., the similarity between two different times t. i With t j The similarity of meteorological observation data at the plateau environment level is used to determine whether different observation data are in similar environmental conditions, thus serving as a basis for time alignment. It represents the environmental difference metric, that is, the comprehensive difference between two environmental state vectors. It is used to measure the overall changes in altitude, underlying surface, air pressure level and radiation conditions, and is an important basis for calculating environmental similarity. The environmental difference scale parameter is a scale control parameter used to adjust the sensitivity of environmental similarity to environmental changes, and is used to avoid the excessive influence of small environmental fluctuations on time alignment judgment.

[0061] S13. An environmental correction mechanism based on altitude and underlying surface differences is introduced during the spatial alignment process to correct the traditional geographic location mapping under environmental constraints. This ensures that spatial consistency is reflected not only in geometric location but also in the actual impact of plateau meteorological elements on environmental changes, thereby enhancing the physical reliability of the spatial alignment results. Specifically:

[0062] Based on the raw spatial location of multi-source meteorological data, an environmentally corrected spatial mapping function is introduced:

[0063] ;

[0064] This correction process ensures that the spatial alignment results are not only geometrically close, but also consistent in the sense of the plateau environment.

[0065] Where X represents the original spatial location coordinates, that is, the original spatial location corresponding to the meteorological observation data, which is used to describe the initial location of the observation point in geographic space; It represents the effective spatial location after environmental correction, that is, the spatial mapping result after correction based on the original spatial location and incorporating plateau environmental factors, which is used to improve the environmental consistency of spatial alignment. This represents the altitude difference correction, a spatial correction term caused by the difference in altitude of the observation points, used to reflect the impact of changes in the vertical height of the plateau on the spatial distribution of meteorological elements; This represents the correction for underlying surface differences, which is a spatial correction term caused by differences in underlying surface types. It is used to reflect the impact of changes in surface conditions on local meteorological fields. This represents the altitude correction weight coefficient, which is the weight of the altitude factor in the spatial mapping correction process and is used to control the degree of influence of altitude differences on spatial location adjustment. This represents the underlying surface correction weight coefficient, which is the weight of underlying surface factors in the spatial mapping correction process. It is used to control the degree of influence of surface differences on spatial location adjustment.

[0066] S14. Based on the integrated environmental binding results, environmental consistency time reference, and environmental correction spatial location, an environmentally conscious spatiotemporally consistent observation unit is constructed. This unit serves as the smallest effective input unit for subsequent environmental consistency screening and fusion calculations, ensuring that the entire plateau multi-source meteorological data fusion process operates under real environmental constraints. Specifically:

[0067] Based on the effective time reference points obtained through environmental similarity screening and the spatially corrected locations, an environmentally-aware spatiotemporally consistent observation unit is constructed: ;

[0068] Among them, the observation unit serves as the direct input for subsequent environmental consistency screening and dynamic fusion weight generation, ensuring that subsequent processing is always carried out under the constraints of the real plateau environment.

[0069] in, This represents the environment-aware spatiotemporally consistent observation unit, i.e., the k-th observation unit after environmental correlation, temporal alignment, and spatial correction. This represents the environmentally corrected spatial location corresponding to the observation unit, that is, the environmentally corrected spatial coordinates corresponding to the k-th environmentally-aware spatiotemporally consistent observation unit.

[0070] S2. Based on the environment-aware spatiotemporal consistent observation unit output from step S1, through local environmental consistency assessment, environmental effectiveness stratification, dynamic hierarchical adjustment, and structured output, the reliability screening and schedulable hierarchical modeling of multi-source meteorological data in the plateau are realized. This provides environment-driven structured input for subsequent fusion, making the fusion process conform to the characteristics of the complex plateau environment and possessing high interpretability and operability. The specific implementation process is as follows:

[0071] S21. Perform consistency calculations on the environmental state of each observation unit within its spatial neighborhood and time window, quantify the degree of matching of the observation unit in the local plateau environment, and generate an environmental consistency index, specifically:

[0072] For the environment-aware spatiotemporal consistent observation unit output in step S1 In its spatial neighborhood and time neighborhood Within this, all neighboring observation units are collected to form a subset of the local environment. ;

[0073] For each observation unit in this subset, calculate the difference between the observation unit and the central observation unit. The environmental state differences are analyzed, and a local consistency index is obtained through an exponential decay function:

[0074] ;

[0075] in, The environmental consistency index represents the degree of overall consistency of the k-th environmentally perceptive spatiotemporally consistent observation unit under its local plateau environmental background. It is used to measure whether the observation unit is in a meteorological state with continuous environmental evolution and reasonable changes. This represents the total number of observation units in the neighborhood of the k-th observation unit; This represents the environmental state vector of a neighboring observation unit, that is, the plateau environmental state corresponding to the j-th observation unit located in the same local spatiotemporal neighborhood as the k-th observation unit; This represents the environmental consistency adjustment parameter, which is used to control the rate at which differences in environmental conditions decay the consistency index, and determines the system's sensitivity to environmental changes. It represents the measure of environmental state differences, that is, the comprehensive differences between two observation units in plateau environmental conditions, and is used to characterize the overall range of changes in factors such as altitude, underlying surface, air pressure level and radiation conditions.

[0076] S22. The environmental consistency index is mapped to a three-layer validity structure. The high-environment-consistency observation layer L1 participates in the fusion first, the medium-environment-consistency observation layer L2 is used for auxiliary purposes, and the low-environment-consistency observation layer L3 is reserved for backup. Through a layered rather than elimination design, the diversity and reliability of data are balanced, providing an operable hierarchical structure for fusion in the complex environment of the plateau. Specifically:

[0077] Environmental consistency index for each observation unit The mapping is hierarchical, used to determine its reference priority in the fusion process: ;

[0078] Each observation unit is labeled with its level to form a preliminary hierarchical structure, providing an environmentally sensitive decision-making basis for dynamic fusion;

[0079] in, This indicates the environmental validity level of the observation unit, that is, the data validity level of the k-th observation unit under the current plateau environmental conditions; This indicates a high environmental consistency observation layer, which is an observation unit level that is highly consistent with the current plateau environmental state. It usually has a high degree of participation in subsequent fusion. This indicates a moderately consistent environmental observation layer, where the environmental conditions partially match the current state of the observation unit. It has some reference value but is at risk of environmental deviation. This indicates a low-environment consistency observation layer, which is an observation unit level with weak consistency with the current plateau environmental state. It is usually weakened in the fusion process, but is not directly discarded. and These represent the environmental consistency classification thresholds, specifically the high consistency lower limit and the medium consistency lower limit, respectively.

[0080] S23. Dynamically adjust the hierarchy of the observation unit based on the changes in the environmental state over time, allowing lower-level data to be upgraded as the environment changes, and higher-level data to be downgraded as the environment deviates, thus synchronizing the hierarchical structure with the evolution of the plateau environment and enhancing the system's adaptability to sudden and phased environmental changes. Specifically:

[0081] Considering the rapid changes in the plateau environment over time, each observation unit level L k Dynamically adjusts to changes in environmental conditions: ; ;

[0082] If a certain unit originally belonged to the low environmental consistency observation layer However, if its state highly matches the surrounding environment after environmental changes, it can be upgraded to... ,vice versa;

[0083] in, It represents the time interval of environmental evolution, that is, the time span used to assess changes in environmental state, and is used to trigger dynamic adjustments in the hierarchical structure; This represents the change in environmental state, specifically the change in the state of the k-th observation unit from time t to... The degree of change in the plateau environment between them; This represents the environment-driven hierarchical adjustment mapping relationship, that is, the mapping rule that re-determines the observation unit level according to changes in the environmental state; This indicates the sensitivity weight to environmental changes and controls the magnitude of hierarchical adjustments;

[0084] It should be noted that the environment-driven hierarchical adjustment mapping relationship This system is used to dynamically adjust the environmental effectiveness level of an observation unit based on changes in the plateau environment. The mapping relationship comprehensively considers the magnitude, direction, and persistence of environmental changes. When the environmental state remains stable or shows a continuous evolution trend, the original level is maintained. When the environmental state undergoes significant but physically reasonable changes, the system guides the observation unit to gradually adjust its level assignment. When short-term environmental fluctuations and a lack of continuity are detected, frequent level switching is suppressed, thereby ensuring that the layered structure is both responsive and maintains overall stability under the complex meteorological conditions of the plateau.

[0085] S24. Integrate and encapsulate the observation units, environmental consistency indicators, and stratification results into a stratified observation unit for environmental effectiveness, forming a structured input, specifically:

[0086] Environmental consistency index Environmental effectiveness levels of observation units and the environmental perception observation unit in step S1 Unified packaging:

[0087] ;

[0088] in, The hierarchical observation unit representing environmental effectiveness constraints, i.e. the final structured observation unit before the fusion step, integrates the original observation information, environmental consistency assessment results, and effectiveness hierarchy.

[0089] S3. Based on the spatiotemporally consistent feature set of multi-source meteorological data for the plateau formed in step S2, an information enhancement approach driven by environmental constraints is introduced. This elevates meteorological elements from static numerical descriptions to dynamic expressions coupled with plateau topography, underlying surface, and upper-air environment. Through feature recombination, credibility correction, temporal consistency enhancement, and explicit processing of regional differences, a highly reliable feature expression result that can truly reflect the evolution mechanism of plateau meteorology is constructed. The specific implementation process is as follows:

[0090] S31. Using the unified spatiotemporal network meteorological features output in step S2 as input, and combining the unique altitude gradient, slope aspect, and surface type information of the plateau, reconstruct the features of various meteorological elements under environmental constraints. This ensures that parameters such as temperature, air pressure, and humidity are no longer isolated values, but rather embedded with their corresponding environmental background features, thereby enhancing the ability of features to express the true meteorological state from the source. Specifically:

[0091] The hierarchical observation unit (observation unit at the environmental consistency level) is designed for the environmental validity constraints output in step S2. Based on its environmental effectiveness level For each observation unit, construct the fusion influence boundary interval:

[0092] ;

[0093] in, This represents the fusion weight scope of the k-th observation unit under the current plateau environment state, that is, the weight range that the observation unit is allowed to take when participating in fusion; This represents the minimum weight limit that the k-th observation unit is allowed to participate in during the fusion process; This represents the maximum weight limit that the k-th observation unit is allowed to participate in the fusion under the current environmental conditions;

[0094] It should be noted that observation units with high environmental consistency have a wider weight range, allowing them to play a dominant role in the fusion; observation units with weaker environmental consistency have their weight range actively compressed, but not directly set to zero, in order to retain their potential indicative value for anomalies or transitional states.

[0095] S32. To address the reliability differences that may arise from different data sources in the extreme environment of high-altitude regions, a reliability assessment mechanism based on environmental consistency is introduced. This mechanism dynamically corrects the reliability of features by comparing the consistency of responses from multiple data sources under the same environmental conditions. Specifically:

[0096] For each hierarchical observation unit, an environment-driven adaptive weight growth function is constructed based on the evolution of its environmental state over time: ;

[0097] in, This represents the dynamic fusion weight value corresponding to the k-th observation unit at time t; This represents the change in the plateau environment state vector corresponding to the kth observation unit over time. The environmental response function is used to characterize the degree of response of the fusion weights to changes in the plateau environment. This function is used to map the changes in environmental state into a standardized adjustment factor, so that the weights change smoothly within the scope and avoid drastic fluctuations in weights caused by short-term environmental disturbances.

[0098] It should be noted that the environmental response function It is used to characterize the comprehensive mapping relationship of the impact of changes in external environmental conditions on system state and decision variables. By uniformly modeling the temporal characteristics, spatial correlation and disturbance intensity of environmental elements, it transforms complex, multidimensional and uncertain environmental inputs into response results that can be processed and fed back by the system, thereby reflecting the sensitivity, lag and cumulative effect of environmental changes, and providing a basis for subsequent state correction, risk assessment and adaptive adjustment.

[0099] S33. Given the continuous yet irregular rhythmic nature of plateau meteorological changes, temporal consistency enhancement processing is applied to the reconstructed feature sequences. By analyzing the rationality of changes in features at adjacent time points under environmental constraints, smoothing and correcting abrupt or distorted segments are performed. Specifically:

[0100] Based on the generation of single-observation weights, a multi-source environmental collaborative adjustment mechanism is introduced to balance the weights as a whole:

[0101] ;

[0102] This collaborative adjustment ensures that the weight distribution among multiple sources remains generally reasonable, reflecting the dominant role of the environmentally advantageous source while avoiding dependence on a single data source in the fusion results.

[0103] in, The final fusion weight of the k-th observation unit after multi-source environmental collaborative adjustment is represented; N represents the total number of observation units participating in weight collaborative adjustment at the current fusion time. This parameter is used to normalize the multi-source weights to ensure that the fusion weights maintain a reasonable distribution overall. This represents the weighting adjustment function based on the environmental consistency index; This represents the dynamic fusion weight value of the j-th hierarchical observation unit;

[0104] It should be noted that the weighting adjustment function is based on the environmental consistency index. Used to characterize the degree of matching between the current perceived environment and the historical stable environment, and accordingly to adaptively adjust the fusion weights of multi-source data or multi-model results. This function comprehensively considers the similarity of environmental feature distribution, the continuity of temporal changes, and the degree of satisfaction of key constraints. When the environmental consistency is high, h tends to maintain the existing weight configuration to enhance system stability. When a significant shift or disturbance in the environment is detected, h will smoothly guide the weights to tilt towards information sources that are more sensitive to or more reliable in the current environment, thereby improving the system's adaptability to complex dynamic environments while ensuring robustness. Its core function is to avoid weight mutations and suppress the amplified impact of occasional noise on decision results.

[0105] S34. From the perspective of regional differences, the characteristics of different climate zones within the plateau are explicitly identified and structurally organized, enabling the same meteorological elements to form distinguishable characteristic expressions under different regional backgrounds. This provides direct support for subsequent regional analysis and refined applications. Specifically:

[0106] By integrating observation unit information, environmental effectiveness results, and final weights, a dynamic fusion weight structure is constructed:

[0107] ;

[0108] in, This represents k dynamically fused weight units.

[0109] S4. Based on the environment-driven dynamic fusion weights generated in step S3, multi-scale collaborative fusion is performed under the constraints of the plateau environment, taking into account time, space, and environmental constraints. This optimizes the fusion results in terms of physical meaning, environmental adaptability, and spatiotemporal continuity, providing reliable input for subsequent self-calibration. The specific implementation process is as follows:

[0110] S41. By dividing the plateau environment into stable intervals, time-scale fusion is performed on observation units with consistent environments within the same time period. This ensures that the fusion results respond smoothly when the environment is stable, while remaining sensitive during abrupt environmental changes, thus avoiding missmoothing of extreme weather events on the plateau. Specifically:

[0111] Dynamic fusion weight unit for the same spatial location Based on the rate of change of its environmental state, the environment is divided into stable intervals, and time-scale collaborative fusion is performed within the same stable interval:

[0112] ;

[0113] When the environment enters a phase of rapid change, the time coordination window should be narrowed to prioritize maintaining the ability to respond to real mutation processes.

[0114] in, This represents the time-scale fusion meteorological results at time t; This represents the original observation value of the k-th observation unit at time t, i.e., the observation unit acquired in step S1 and layered by environmental constraints in step S2; This represents the set of observation units within the same plateau environment stability interval, divided according to the environmental driving weights and environmental state change rate generated in step S3.

[0115] S42. Based on temporal fusion, a plateau environment correction factor is introduced to perform weighted synergy on the fusion results within the spatial neighborhood. Specifically:

[0116] For a certain target point and the corresponding time-scale fusion results First, select its spatial neighborhood: ;

[0117] For each neighboring point Calculate the environmental difference between it and the target point: ;

[0118] For each neighboring point Mapping environmental variability to a correction factor: ;

[0119] The results of combining environmental correction factors and time scales Spatial weighted fusion is performed:

[0120] ;

[0121] Where S represents the set of neighborhood observation units of the target point; Representing neighborhood points With the target point Geographical distance between them; This represents the target point, i.e., the coordinates of the observation point from which the spatial fusion result needs to be calculated. This represents the spatial neighborhood radius, used to determine the neighborhood range participating in the fusion. It is set according to the complexity of the plateau terrain, the density of observation points, and the fusion requirements, and can be an adjustable parameter. and These are the plateau environmental state vectors for the target point and its neighboring points, respectively, including environmental information such as altitude, plateau pressure level, underlying surface type, terrain slope, and radiation status; Representing neighborhood points With the target point Environmental variability measures the degree of inconsistency in environmental conditions; This represents the plateau environment correction factor, used to adjust the spatial contribution of neighboring points; This parameter represents the environmental sensitivity, controlling the magnitude of the impact of environmental differences on the weight. It can be set or adaptively adjusted according to the complexity of the plateau environment and the drastic changes in terrain. Indicates the target point The final spatial scale fusion result is a combination of neighborhood temporal fusion results and environmental correction factors;

[0122] S43. To address local fusion bias, environmental consistency constraints are introduced to suppress non-environmentally consistent anomalies, while allowing environment-driven anomalies to participate in the fusion process, ensuring that the true extreme weather conditions of the plateau can be reasonably reflected. Specifically:

[0123] Fusion results for each spatial point Calculate the deviation from environmental constraints based on the environmental conditions of the surrounding area and its neighborhood.

[0124] ;

[0125] Based on deviation Dynamically adjust the fusion weights:

[0126] ;

[0127] After adjusting the weights of all spatial points, normalization and co-balancing are performed:

[0128] ;

[0129] in, This represents the reference fusion value of the k-th observation unit under the current environmental conditions. It can be obtained by neighborhood averaging, trend prediction, or environmental state weighting, reflecting the "ideal value under environmental consistency". This represents the environmental constraint deviation of the k-th observation unit; This represents the fusion weights after environmental constraint anomaly suppression; This represents the anomaly suppression adjustment coefficient, an adjustable parameter set based on the sensitivity to plateau meteorological anomalies. This represents the final fusion result after anomaly suppression and synergistic balancing;

[0130] S44. The fusion results of time, space, and environmental constraints are uniformly encapsulated and output as an environmental constraint fused meteorological unit, including fused values, environmental status, and spatiotemporal information, i.e.:

[0131] The fusion results are uniformly encapsulated with environmental status, spatial location, and temporal information:

[0132] ;

[0133] in, This represents the multi-scale environmental constraint fusion output unit; Indicates time information.

[0134] S5. Based on the multi-scale environmental constraint fusion results of step S4, a plateau adaptive feedback closed loop is constructed. Through environmental evolution prediction, error correction, local smoothing, and final output encapsulation, high-precision, stable, and interpretable fusion data is formed, realizing intelligent fusion and adaptive optimization in the complex plateau environment. The specific implementation process is as follows:

[0135] S51. Based on the fusion results of step S4, extract the environmental vector and its changes at the target point, construct an environment-driven nonlinear prediction model, map environmental changes to short-term trends of meteorological elements, and generate predicted values ​​for each point, specifically:

[0136] For each fusion unit Extract its environmental vector changes ;

[0137] Building an environment-driven prediction model: ;

[0138] in, Represents a vector of environmental changes; Indicates the time step, used to calculate environmental changes; This represents the fused meteorological values ​​at the predicted time point; This represents an environment-driven prediction function that maps environmental changes to meteorological changes; it can be implemented using a nonlinear model or a neural network.

[0139] S52. Compare the prediction results from step S51 with short-term observation data, calculate the error, and correct the fusion results using environmental consistency indicators and adaptive adjustment coefficients to enhance the adaptability to the plateau environment, while avoiding erroneous corrections and achieving closed-loop feedback correction. Specifically:

[0140] Calculation error ;

[0141] Perform environmental adaptation modifications:

[0142] ;

[0143] ;

[0144] in, Indicates short-term observations; Indicates prediction error; This represents the adaptive adjustment coefficient; Indicates the basic correction factor; This represents the variance of the target point's neighborhood environment vector. This represents the fusion value after feedback correction;

[0145] S53. The corrected results are then subjected to neighborhood-weighted smoothing, with weights adjusted by environmental variability to enhance local continuity while preserving extreme changes driven by the real environment, suppressing noise and inconsistent observations, and forming fused data that is both stable and reflects local anomalies in the plateau environment. Specifically:

[0146] Perform neighborhood-weighted smoothing: ; ;

[0147] It should be noted that the actual changes driven by extreme environments, Large time weight decay allows anomalies to be preserved; adaptive smoothing suppression of noise observations with inconsistent environments improves data stability.

[0148] in, Represents the set of neighborhood units; This represents the fusion value after local stabilization enhancement; This represents the neighborhood weighting coefficient, which adjusts the neighborhood contribution based on environmental differences to ensure that the true plateau environment-driven anomalies are preserved during the smoothing process.

[0149] S54. The fusion results from step S53 are uniformly encapsulated with environmental vectors, spatial coordinates, temporal information, and local uncertainty indices to form a directly applicable plateau adaptability fusion dataset. This ensures data interpretability and traceability, while providing standardized input for downstream prediction, analysis, and risk assessment, achieving closed-loop intelligent plateau fusion. Specifically:

[0150] Construct the final output:

[0151] ;

[0152] ;

[0153] in, This represents the final output vector of the high-altitude adaptation. Indicates the uncertainty index of local fusion; This represents the average fusion value of the neighborhood.

[0154] Example 2, as Figure 2 As shown, the present invention proposes an intelligent fusion system for multi-source meteorological data in plateau regions, which is used to execute the intelligent fusion method for multi-source meteorological data in plateau regions proposed in Embodiment 1. The system includes: an environmental correlation data acquisition and spatiotemporal alignment module, an environmental correlation modeling and feature extraction module, a time series fusion and dynamic weight adjustment module, a spatial scale collaborative fusion and anomaly suppression module, and a plateau adaptive feedback correction and final output module.

[0155] The environmental correlation data acquisition and spatiotemporal alignment module is used to collect data from plateau multi-source meteorological observation equipment, satellite remote sensing, UAVs and historical databases, perform environmental correlation processing and spatiotemporal alignment, match the collected meteorological elements with plateau environmental parameters (such as altitude, slope, surface type, etc.), and unify timestamps and spatial coordinates to provide standardized and complete data input for subsequent fusion;

[0156] The environmental correlation modeling and feature extraction module analyzes the relationship between various meteorological elements and environmental parameters, constructs plateau environmental consistency indicators and multi-source data correlation features, extracts feature vectors of each observation point in the time and environmental dimensions, identifies potential anomalies and trend information, and provides a basis for subsequent fusion and weighting.

[0157] The time series fusion and dynamic weight adjustment module performs sequence fusion of data from various observation points in the time dimension. Through environment-driven dynamic weight adjustment, it achieves short-term prediction and historical trend compensation. It adaptively allocates the weight of data at each time point according to environmental consistency and temporal continuity, ensuring that the fusion result is both smooth and continuous, and can quickly respond to the changing trends of the plateau environment, while providing a robust time series foundation for spatial fusion.

[0158] The spatial scale collaborative fusion and anomaly suppression module performs neighborhood fusion and plateau environment correction on data in the spatial dimension, and performs collaborative fusion on different spatial scales in combination with environmental correction factors; it can identify local anomalies and adjust weights under environmental constraints, suppress non-environmentally consistent biases, and retain anomalies driven by the real plateau environment, realize multi-scale fusion closed loop, and output spatial fusion results with stability and environmental adaptability.

[0159] The plateau adaptive feedback correction and final output module performs environmental evolution trend analysis and adaptive prediction on the fusion results, feeds back the predicted values ​​with real-time observations, performs adaptive correction in combination with environmental consistency, and enhances local continuity through neighborhood weighted smoothing. Finally, the fusion results are uniformly encapsulated with environmental parameters, spatial coordinates, temporal information and uncertainty indicators to form a plateau adaptive fusion dataset, which can be directly used for plateau meteorological analysis, prediction and decision support, and completes intelligent closed-loop output.

[0160] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited thereto. Various changes can be made within the scope of knowledge possessed by those skilled in the art without departing from the spirit of the present invention.

Claims

1. A highland multi-source meteorological data intelligent fusion method, characterized in that, The specific implementation steps include the following: S1. Synchronously collect plateau environmental status information corresponding to multi-source meteorological data to form environmentally correlated meteorological data units; complete time reference alignment based on environmental status similarity, and introduce an environmental correction mechanism to realize spatial location mapping, thus constructing an environmentally perceptive spatiotemporally consistent observation unit. Among them, the plateau environmental status information corresponding to the synchronous collection of multi-source meteorological data includes: Altitude range parameters used to reflect the vertical gradient characteristics of meteorological elements; Parameters of underlying surface type used to describe the influence of surface properties on local meteorological conditions; Pressure level characteristic parameters used to distinguish atmospheric conditions at different altitudes; Radiation environment state parameters used to characterize the impact of strong solar radiation on observation results on plateaus; Specifically, time reference alignment based on environmental state similarity includes: Construct an environmental difference metric between environmental state vectors, and calculate an environmental similarity index based on the difference metric and a preset environmental difference scale parameter; When the environmental similarity index is higher than the preset threshold, it is determined that the corresponding observation data are in a similar plateau environment, and they are included in the same effective time reference window to complete the time series alignment. Specifically, the spatial location mapping is achieved by introducing an environmental correction mechanism: Based on the original spatial location coordinates, an environmental correction term is added, consisting of the product of the elevation difference correction amount and its corresponding elevation correction weight coefficient, and the product of the underlying surface difference correction amount and its corresponding underlying surface correction weight coefficient, so as to obtain the effective spatial location after environmental correction. S2. Perform local environmental consistency assessment on each environmentally conscious spatiotemporally consistent observation unit and generate an environmental consistency index. Based on this index, divide the environmentally conscious spatiotemporally consistent observation unit into different environmental effectiveness levels and dynamically adjust its level based on the time evolution of the environmental state to form a hierarchical observation unit with environmental effectiveness constraints. S3. Based on the environmental effectiveness level of the hierarchical observation unit, determine its weight scope for fusion; combined with the environmental state changes of the hierarchical observation unit, dynamically generate its fusion weight through the environmental driving mechanism, and carry out multi-source collaborative adjustment to form a dynamic fusion weight unit. S4. Perform time-scale fusion within the same plateau environment stability range, and perform spatial-scale fusion by combining neighborhood points and environmental correction factors; at the same time, calculate the environmental constraint deviation and dynamically adjust the fusion weights to suppress non-environmental consistency anomalies, and output the multi-scale environmental constraint fusion results. S5. Based on the environmental state changes of the fusion results, construct an environment-driven prediction model to obtain the predicted values; compare the predicted values ​​with short-term observations to obtain the error, and perform adaptive correction in combination with environmental consistency; perform neighborhood weighted smoothing on the correction results, and finally encapsulate and output the plateau adaptive fusion dataset.

2. The method according to claim 1, wherein, The environmental validity levels in step S2 include at least: a high environmental consistency observation layer, a medium environmental consistency observation layer, and a low environmental consistency observation layer; That is, the environmental consistency index is compared with the preset environmental consistency classification threshold. Environmentally perceptual spatiotemporal consistency observation units that are higher than the high consistency lower threshold are classified into the high environmental consistency observation layer, environmentally perceptual spatiotemporal consistency observation units that are between the medium consistency lower threshold and the high consistency lower threshold are classified into the medium environmental consistency observation layer, and environmentally perceptual spatiotemporal consistency observation units that are lower than the medium consistency lower threshold are classified into the low environmental consistency observation layer.

3. The method of claim 2, wherein, In step S2, the hierarchy is dynamically adjusted based on the time evolution of the environmental state as follows: Calculate the change in environmental state of the environment-aware spatiotemporally consistent observation unit within a set environmental evolution time interval, and adjust the mapping relationship according to the change and the preset hierarchy; The environmental effectiveness hierarchy of the environmentally perceptive spatiotemporally consistent observation unit is reassessed and adjusted. The hierarchical adjustment mapping relationship responds to the continuous evolution trend of the environmental state and suppresses frequent hierarchical switching caused by short-term environmental fluctuations.

4. The intelligent fusion method for multi-source meteorological data in plateau regions according to claim 3, characterized in that, In step S3, the fusion weights are dynamically generated through an environment-driven mechanism as follows: The changes in the plateau environmental state of the stratified observation units over time are input into the environmental response function; This environmental response function is used to map changes in environmental state into standardized adjustment factors. The initial dynamic fusion weight is calculated using this adjustment factor within the fusion weight scope corresponding to the hierarchical observation unit.

5. The intelligent fusion method for multi-source meteorological data in plateau regions according to claim 4, characterized in that, The multi-source coordinated regulation in step S3 specifically involves: The initial dynamic fusion weights of all hierarchical observation units participating in the fusion are processed using a weight adjustment function based on the environmental consistency index. This weight adjustment function tends to maintain the existing weight configuration when the environment is highly consistent; When a significant shift in the environment is detected, the weights are smoothly guided to tilt towards information sources that are more reliable for the current environment. Finally, the final fusion weights of all hierarchical observation units are obtained through normalization.

6. The intelligent fusion method for multi-source meteorological data in plateau regions according to claim 5, characterized in that, The calculation of environmental constraint deviation in step S4 is as follows: First, a reference fusion value representing the ideal value under environmental consistency is obtained through neighborhood averaging, trend prediction, and environmental state weighting. Then, the difference between the fusion result of the hierarchical observation unit and the reference fusion value is calculated, and the final environmental constraint deviation metric is obtained by combining the variance of the neighborhood environment vector of the hierarchical observation unit.

7. A plateau multi-source meteorological data intelligent fusion system, used to execute the plateau multi-source meteorological data intelligent fusion method according to any one of claims 1 to 6, characterized in that, include: The environmental correlation data acquisition and spatiotemporal alignment module is used to collect and bind meteorological data and plateau environmental status information, perform temporal alignment based on environmental similarity and spatial mapping based on environmental correction, and output environmentally perceptive spatiotemporal consistent observation units. The environmental relevance modeling and feature extraction module is used to calculate the local environmental consistency index of the environmentally perceptive spatiotemporally consistent observation unit, and based on this, to perform environmental effectiveness stratification and realize dynamic adjustment of the stratification, and output the stratified observation unit with environmental effectiveness constraints. The time series fusion and dynamic weight adjustment module is used to determine the weight scope of the hierarchical observation unit based on the hierarchical results, dynamically generate and coordinately adjust the multi-source fusion weight in combination with environmental changes, and output the dynamic fusion weight unit. The spatial scale collaborative fusion and anomaly suppression module is used to perform temporal fusion within the environmental stability range, combine environmental correction factors for spatial fusion, and adjust weights based on environmental constraint deviation to suppress anomalies, outputting multi-scale environmental constraint fusion results. The plateau adaptation feedback correction and final output module is used for prediction based on environmental evolution, adaptive correction and local smoothing through error feedback, and finally encapsulates and outputs the plateau adaptation fusion dataset.

Images