Multi-source heterogeneous data space-time alignment method and system

Abstract

The invention relates to the technical field of data processing, in particular to a multi-source heterogeneous data space-time alignment method and system. According to the method, the time feature and the space feature in each standardized data set are extracted, and the space-time association diagram is constructed, so that the space-time relationship between different data can be clearly revealed, the association between the data is improved, the interpolation process is more accurate, and the calculation efficiency is improved. By introducing a joint space-time interpolation algorithm, changes of time and space characteristics are considered in the space-time alignment process, the situation that time interpolation errors are brought into space interpolation in a traditional method is avoided, the error situation in the interpolation process is fed back in time through an error evaluation mechanism, and interpolation parameters are adjusted. Therefore, the problem of error accumulation and amplification is avoided, comprehensive space-time alignment of multi-source heterogeneous data can be realized by integrating a plurality of middle alignment data sets and constructing a global space-time coordinate system, and accurate matching of different data sources under a unified coordinate system is ensured.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and system for spatiotemporal alignment of multi-source heterogeneous data. Background Technology

[0002] Multi-source heterogeneous data refers to data from different sources with varying structures and formats. Spatiotemporal alignment refers to the need to align data in time and space during multi-source heterogeneous data analysis in order to compare, analyze, and integrate data from different sources. Currently, the main method for spatiotemporal alignment of multi-source heterogeneous data is to first map data from different time points to the same time scale using temporal interpolation, and then map data from different spatial locations after temporal interpolation to the same spatial location using spatial interpolation. However, since heterogeneous data from different sources have different temporal and spatial resolutions, errors will occur when aligning data with different temporal resolutions. The errors generated during temporal interpolation will be substituted into spatial interpolation, which will lead to the accumulation and amplification of errors in the spatiotemporal alignment results. Summary of the Invention

[0003] The main objective of this invention is to provide a spatiotemporal alignment method for multi-source heterogeneous data, aiming to solve the technical problems in the prior art.

[0004] This invention proposes a spatiotemporal alignment method for multi-source heterogeneous data, comprising: Obtain the basic attribute information and spatiotemporal reference parameters of multi-source heterogeneous data, and perform classification and standardization processing on the multi-source heterogeneous data according to the basic attribute information and spatiotemporal reference parameters to obtain multiple sets of standardized datasets; Extract the temporal and spatial features of each data sample in each of the standardized datasets, and construct a spatiotemporal correlation graph of the corresponding standardized datasets based on the multiple temporal and spatial features; Based on each of the spatiotemporal correlation graphs, a joint spatiotemporal interpolation algorithm is used to synchronously interpolate the corresponding standardized dataset to obtain the joint spatiotemporal interpolation result, and the temporal error component and spatial error component are obtained in real time during the interpolation process; Based on each of the time error components and spatial error components, obtain the corresponding total error evaluation value, and determine whether the total error evaluation value is greater than a preset threshold. If the total error assessment value is not greater than the preset threshold, the joint spatiotemporal interpolation result will be used as the intermediate aligned dataset. If the total error assessment value is greater than the preset threshold, then the error correction coefficient is obtained according to the time error component and the spatial error component, and the error correction coefficient is fed back to the joint spatiotemporal interpolation algorithm to dynamically adjust the interpolation parameters until the total error assessment value is not greater than the preset threshold, and the joint spatiotemporal interpolation result at this time is used as the intermediate aligned dataset. The multiple intermediate alignment datasets are integrated and summarized to obtain the final alignment dataset, and a global spatiotemporal coordinate system is constructed based on the spatiotemporal feature parameters of the final alignment dataset. All data points in the final aligned dataset are mapped to the global spatiotemporal coordinate system to complete the spatiotemporal alignment of multi-source heterogeneous data.

[0005] This application also provides a spatiotemporal alignment system for multi-source heterogeneous data, including: The classification module is used to obtain the basic attribute information and spatiotemporal reference parameters of multi-source heterogeneous data, and to perform classification and standardization processing on the multi-source heterogeneous data according to the basic attribute information and spatiotemporal reference parameters to obtain multiple sets of standardized datasets. An extraction and construction module is used to extract the temporal and spatial features of each data sample in each of the standardized datasets, and to construct a spatiotemporal correlation graph of the corresponding standardized dataset based on the multiple temporal and spatial features. The interpolation module is used to perform synchronous interpolation processing on the corresponding standardized dataset according to each spatiotemporal correlation graph using a joint spatiotemporal interpolation algorithm to obtain the joint spatiotemporal interpolation result, and to acquire the temporal error component and spatial error component in real time during the interpolation process; The judgment module is used to obtain the corresponding total error evaluation value based on each of the time error component and the spatial error component, and to determine whether the total error evaluation value is greater than a preset threshold. If the total error assessment value is not greater than the preset threshold, the joint spatiotemporal interpolation result will be used as the intermediate aligned dataset. If the total error assessment value is greater than the preset threshold, then the error correction coefficient is obtained according to the time error component and the spatial error component, and the error correction coefficient is fed back to the joint spatiotemporal interpolation algorithm to dynamically adjust the interpolation parameters until the total error assessment value is not greater than the preset threshold, and the joint spatiotemporal interpolation result at this time is used as the intermediate aligned dataset. The aggregation and construction module is used to integrate and aggregate multiple intermediate alignment datasets to obtain a final alignment dataset, and to construct a global spatiotemporal coordinate system based on the spatiotemporal feature parameters of the final alignment dataset. The mapping module is used to map all data points in the final aligned dataset to the global spatiotemporal coordinate system, thereby completing the spatiotemporal alignment of multi-source heterogeneous data.

[0006] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described multi-source heterogeneous data spatiotemporal alignment method.

[0007] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described multi-source heterogeneous data spatiotemporal alignment method.

[0008] The beneficial effects of this invention are as follows: By extracting the temporal and spatial features from each standardized dataset and constructing a spatiotemporal correlation graph, this invention can clearly reveal the spatiotemporal relationships between different data, which not only improves the correlation between data but also makes the interpolation process more accurate. By introducing a joint spatiotemporal interpolation algorithm, the changes in time and space features are considered simultaneously during the spatiotemporal alignment process, avoiding the situation where time interpolation errors are carried over to spatial interpolation in traditional methods. By acquiring and analyzing the temporal and spatial error components in real time, the interpolation parameters can be dynamically adjusted to correct errors when the error exceeds a preset threshold, significantly improving the accuracy of the alignment results. The error evaluation mechanism provides timely feedback on the error situation during the interpolation process and adjusts the interpolation parameters based on the error feedback, thereby avoiding the problem of error accumulation and amplification, ensuring the accuracy and reliability of the final alignment results. By integrating multiple intermediate alignment datasets and constructing a global spatiotemporal coordinate system, comprehensive spatiotemporal alignment of multi-source heterogeneous data can be achieved, ensuring accurate matching of different data sources under a unified coordinate system. Attached Figure Description

[0009] Figure 1 This is a schematic diagram of a method flow according to an embodiment of the present invention.

[0010] Figure 2 This is a schematic diagram of the system structure according to an embodiment of the present invention.

[0011] Figure 3 This is a schematic diagram of the internal structure of a computer device according to an embodiment of this application.

[0012] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0013] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0014] like Figure 1 As shown, this application provides a method for spatiotemporal alignment of multi-source heterogeneous data, including: S1. Obtain the basic attribute information and spatiotemporal reference parameters of multi-source heterogeneous data, and classify and standardize the multi-source heterogeneous data according to the basic attribute information and spatiotemporal reference parameters to obtain multiple sets of standardized datasets. S2. Extract the temporal and spatial features of each data sample in each standardized dataset, and construct the spatiotemporal correlation graph of the corresponding standardized dataset based on multiple temporal and spatial features; S3. Based on each spatiotemporal correlation graph, a joint spatiotemporal interpolation algorithm is used to synchronously interpolate the corresponding standardized dataset to obtain the joint spatiotemporal interpolation result, and the time error component and spatial error component are obtained in real time during the interpolation process. S4. Obtain the corresponding total error assessment value based on each time error component and spatial error component, and determine whether the total error assessment value is greater than the preset threshold. If the total error assessment value is not greater than the preset threshold, the joint spatiotemporal interpolation result will be used as the intermediate aligned dataset. If the total error assessment value is greater than the preset threshold, the error correction coefficient is obtained based on the time error component and the spatial error component, and the error correction coefficient is fed back to the joint spatiotemporal interpolation algorithm to dynamically adjust the interpolation parameters until the total error assessment value is not greater than the preset threshold, and the joint spatiotemporal interpolation result at this time is used as the intermediate aligned dataset. S5. Integrate and summarize multiple intermediate aligned datasets to obtain the final aligned dataset, and construct a global spatiotemporal coordinate system based on the spatiotemporal feature parameters of the final aligned dataset; S6. Map all data points in the final aligned dataset to the global spatiotemporal coordinate system to complete the spatiotemporal alignment of multi-source heterogeneous data.

[0015] As described in steps S1-S6 above, multi-source heterogeneous data often originates from different sensors, exhibiting significant differences in data format, temporal resolution, and spatial resolution. By acquiring basic attribute information and spatiotemporal reference parameters—including data source type, temporal resolution, and spatial resolution, and a unified time reference system and a unified spatial coordinate system—initial classification can be performed according to data source type. Then, secondary classification is performed based on temporal resolution, followed by standardization using the unified time reference system. Finally, a final secondary classification is performed based on spatial resolution, again standardized using the unified spatial reference system. This results in multiple standardized datasets, providing a standardized foundation for subsequent data classification and ensuring a unified understanding of data from different sources, avoiding inconsistencies caused by the inherent heterogeneity of the data. By extracting the temporal and spatial characteristics of each data sample in each standardized dataset, a more comprehensive understanding of the data distribution and variation patterns can be achieved, thereby enabling… Establishing clear spatiotemporal relationships between data facilitates precise positioning in subsequent interpolation and alignment processes, avoiding errors in alignment results caused by insufficient consideration of spatiotemporal features. Constructing a spatiotemporal relationship graph provides clear spatiotemporal connections for subsequent interpolation operations, ensuring data consistency and comparability in time and space. By independently extracting temporal and spatial features, potential sources of error during data alignment can be more clearly identified. When performing interpolation operations at different time and spatial scales, time and spatial errors can be analyzed relatively independently, avoiding the increase in errors caused by simple superposition. By using a joint spatiotemporal interpolation algorithm, the interpolation process can be considered simultaneously in both time and spatial dimensions, improving the overall accuracy of the interpolation results. Real-time acquisition of temporal and spatial error components allows for monitoring and correction of errors during the interpolation process, reducing the possibility of error propagation and improving the accuracy and reliability of spatiotemporal alignment results.

[0016] The steps for real-time acquisition of temporal and spatial error components during the interpolation process include: extracting the theoretical timestamp and actual sampling timestamp corresponding to each interpolation moment during the interpolation process, obtaining the corresponding instantaneous time error based on each theoretical timestamp and actual sampling timestamp, and obtaining the temporal error component based on multiple instantaneous time errors; extracting the theoretical spatial coordinates and actual sampling spatial coordinates corresponding to each interpolation position during the interpolation process, obtaining the corresponding instantaneous spatial error based on each theoretical spatial coordinate and actual sampling spatial coordinate, and obtaining the spatial error component based on multiple instantaneous spatial errors; and adjusting the temporal interpolation window size, spatial interpolation radius, and interpolation kernel function parameters by inputting the error correction coefficients into the joint spatiotemporal interpolation algorithm and according to the preset mapping relationship between the error correction coefficients and the target interpolation parameters.

[0017] By evaluating errors in real time and dynamically adjusting interpolation parameters, the results can be continuously optimized during the interpolation process, ensuring that the final interpolation result meets the preset error requirements. This effectively avoids inaccurate alignment results caused by errors exceeding the tolerance range. A dynamic error correction mechanism continuously monitors and adjusts errors, making the final alignment result more accurate. Continuous optimization of interpolation parameters prevents error accumulation and propagation, ensuring that the alignment result of each data point is within the error tolerance range, thereby improving the reliability and stability of the entire system. By integrating multiple intermediate alignment datasets and constructing a global spatiotemporal coordinate system, the spatiotemporal relationships between different data sources can be unified, forming a unified framework. This improves the usability and accuracy of the integrated data. By mapping all data points to the global spatiotemporal coordinate system, it ensures that data from all data sources are compared and analyzed within the same spatiotemporal framework, enabling seamless data integration and ensuring the accuracy and consistency of the integrated data. This ensures that multi-source heterogeneous data can be analyzed within a unified framework, increasing the reliability and versatility of the data.

[0018] In one embodiment, step S1, which involves classifying and standardizing multi-source heterogeneous data based on fundamental attribute information and spatiotemporal reference parameters to obtain multiple sets of standardized datasets, includes: S11. Obtain basic attribute information and spatiotemporal reference parameters. The basic attribute information includes data source type, time resolution and spatial resolution. The spatiotemporal reference parameters include unified time reference system and unified spatial coordinate system. S12. Divide the multi-source heterogeneous data into multiple initial datasets according to the data source type, and divide the initial datasets according to the time resolution of the data samples in each initial dataset to obtain multiple time-resolved datasets. S13. Obtain the original timestamp information of the data samples in each time-resolved dataset, and obtain the corresponding time deviation value and time granularity difference value based on each original timestamp information and the unified time reference system. S14. Based on each time deviation value and time granularity difference value, perform time calibration and time granularity normalization on the original timestamp information of the data samples in each time-resolved dataset to obtain the corresponding standardized time-resolved dataset. S15. Divide the standardized time-resolved dataset according to the spatial resolution of the data samples in each standardized time-resolved dataset to obtain multiple spatially resolved datasets. S16. Obtain the original spatial coordinate information of the data samples in each spatial resolution dataset, and obtain the corresponding spatial coordinate transformation parameters and spatial accuracy deviation values ​​based on each original spatial coordinate information and the unified spatial coordinate system. S17. Based on each spatial coordinate transformation parameter and spatial accuracy deviation value, perform coordinate transformation and spatial accuracy correction on the original spatial coordinate information of the data samples in each spatially resolved dataset to obtain the corresponding standardized dataset.

[0019] As described in steps S11-S17 above, the data source types include three initial categories: sensor data, remote sensing image data, and log data. Therefore, the multiple initial datasets are sensor datasets, remote sensing image datasets, and log datasets, respectively. Temporal resolution datasets include high temporal resolution datasets, medium temporal resolution datasets, and low temporal resolution datasets. The corresponding time sampling interval is obtained based on the temporal resolution, and the initial datasets are divided according to the time sampling interval. For example, a sampling interval ≤ 1 minute is classified as a high temporal resolution dataset, 1 minute < sampling interval ≤ 1 hour is classified as a medium temporal resolution dataset, and a sampling interval > 1 hour is classified as a low temporal resolution dataset. Spatial resolution datasets include high spatial resolution datasets, medium spatial resolution datasets, and low spatial resolution datasets. The cell side length of the spatial sampling grid is obtained based on the spatial resolution, and the standardized temporal resolution datasets are divided according to the cell side length. For example, a grid side length ≤ 10 meters is classified as a high spatial resolution dataset, 10 meters < grid side length ≤ 100 meters is classified as a medium spatial resolution dataset, and a grid side length > 100 meters is classified as a low spatial resolution dataset.

[0020] A unified time reference system (UTS) is a pre-defined, globally unique time standard used to eliminate time discrepancies between different data sources and ensure that the time dimension of all data is comparable. This is achieved by converting the original timestamp information into a time format identical to the UTS, and then calculating the time difference between the converted original timestamps and the UTS. In essence, through format normalization and time zone calibration, inconsistent original times are converted into a unified standard time. The difference between the original time and the standard time is then quantified numerically. Based on the time deviation value, the original timestamp information is compensated for the deviation, converting the original timestamps to timestamps under the UTS, thus completing time calibration. The time granularity difference value is obtained by calculating the difference between the time sampling interval of each data sample in the time-resolved dataset and the preset standard time granularity of the UTS. If the time granularity difference value is not zero, linear interpolation or a sliding window averaging method is used to supplement the missing time points on the calibrated timestamps, making the time granularity of the time-resolved dataset consistent with the standard time granularity, thus obtaining a standardized time-resolved dataset.

[0021] The coordinate system type of the original spatial coordinate information in each standardized time-resolved dataset is determined. The projection parameters and coordinate offsets in the transformation formula between this coordinate system and the unified spatial coordinate system are used as spatial coordinate transformation parameters. Then, the original spatial coordinate information is substituted into the transformation formula corresponding to the spatial coordinate transformation parameters to complete the coordinate system transformation, thus obtaining the preliminary transformed spatial coordinates. By selecting feature points with known precise locations in the standardized time-resolved dataset, the difference between the original spatial coordinates of these feature points and the standard coordinates under the unified spatial coordinate system is calculated. The average value of these differences is taken as the spatial accuracy deviation value. The preliminary transformed spatial coordinates are corrected according to the spatial accuracy deviation value to obtain a standardized dataset that meets the accuracy requirements of the unified spatial coordinate system.

[0022] Traditional methods often overlook differences in temporal granularity, leading to the accumulation of errors during interpolation and even affecting the accuracy of spatial interpolation. This invention categorizes multi-source heterogeneous data based on data source type, ensuring effective classification and organization of data from different sources in the initial stage. By classifying different data sources, targeted optimization can be performed in subsequent processing, allowing each type of data to undergo specialized preprocessing according to its characteristics, thereby improving the accuracy and efficiency of data processing. By acquiring the original timestamp information of each data sample and performing time calibration according to a unified time reference system, time differences in different data sources can be accurately aligned. By calculating time deviation and time granularity differences, accurate alignment of data samples on the same time axis is ensured, eliminating errors caused by inconsistent time definitions from data sources and effectively improving the quality of data spatiotemporal alignment.

[0023] Data from different sources may have different spatial resolutions. Directly merging them together for subsequent processing can easily lead to information loss or error propagation. By dividing the dataset according to spatial resolution and classifying data samples according to spatial precision, we can ensure that data with different spatial resolutions are properly processed and refined. Through spatial resolution-based division, each dataset can be processed more finely, making the alignment process of data under different spatial resolutions more accurate. By obtaining the original spatial coordinate information and calculating the spatial coordinate transformation parameters, we can ensure the consistency of spatial coordinate systems between data sources, thereby avoiding errors caused by differences in coordinate systems. By correcting the spatial coordinates of data samples based on spatial coordinate transformation parameters and spatial precision deviation values, we finally obtain a standardized dataset. By uniformly transforming data from different spatial coordinate systems to the standard coordinate system and correcting the spatial precision, we can ensure accurate spatial alignment of data, avoid the accumulation of interpolation errors, and improve the consistency of multi-source data in spatial dimensions.

[0024] In one embodiment, step S2, which constructs a spatiotemporal correlation graph of a corresponding standardized dataset based on multiple temporal and spatial features, includes: S21. Obtain the timestamp, time interval, and time series trend for each time feature, and establish a corresponding time feature vector based on each timestamp, time interval, and time series trend; S22. Obtain the spatial coordinates, spatial topological relationships, and spatial distribution density information of each spatial feature, and establish the corresponding spatial feature vector based on each spatial coordinate, spatial topological relationship, and spatial distribution density information; S23. Obtain the corresponding association weight coefficients based on each time feature vector and spatial feature vector, and determine the corresponding coupling relationship based on each association weight coefficient; S24. Treat each data sample as a node, and obtain the temporal similarity and spatial similarity based on the temporal feature vector and spatial feature vector of the two adjacent nodes, respectively. S25. Obtain the comprehensive correlation degree between two adjacent nodes by weighted summation based on each time similarity and spatial similarity, and determine the edge connection relationship and edge weight between the two adjacent nodes based on each comprehensive correlation degree. S26. Construct an initial spatiotemporal correlation graph based on the coupling relationships, edge connection relationships, and edge weights of multiple nodes.

[0025] As described in steps S21-S26 above, the step of determining the corresponding coupling relationship based on the correlation weight coefficient is to analyze the relationship between the correlation weight coefficient and the threshold interval. If the correlation weight coefficient is less than the lower limit of the threshold interval, the coupling relationship is weak. If the correlation weight coefficient is within the threshold interval, the coupling relationship is medium. If the correlation weight coefficient is greater than the upper limit of the threshold interval, the coupling relationship is strong. Temporal similarity can be calculated by subtracting the temporal feature contribution from 1, and spatial similarity can be calculated by subtracting the spatial feature contribution from 1.

[0026] The steps for determining the edge connection relationship and edge weight between adjacent nodes based on each comprehensive correlation degree include: judging the relationship between each comprehensive correlation degree and the correlation degree judgment threshold; if the comprehensive correlation degree is not less than the correlation degree judgment threshold, it is determined that the adjacent nodes meet the edge connection condition, and the edge connection relationship between the adjacent nodes is determined; if the comprehensive correlation degree is less than the correlation degree judgment threshold, it is determined that the adjacent nodes do not meet the edge connection condition, and no edge connection is established between the nodes; determining the initial edge weight of the corresponding edge connection based on each comprehensive correlation degree and the preset edge weight mapping rule; obtaining the spatiotemporal decay coefficient between nodes based on the spatiotemporal dynamic characteristics of the standardized dataset; and determining the spatiotemporal decay coefficient based on the spatiotemporal decay coefficient. The coefficients are used to correct the initial edge weights to obtain the edge weights. The spatiotemporal decay coefficient is determined based on the time interval, spatial distance, and data update frequency between nodes. The longer the time interval and the greater the spatial distance, the smaller the spatiotemporal decay coefficient. The preset correlation threshold can be set according to the data density, spatiotemporal correlation requirements, and application scenarios of the standardized dataset. The preset edge weight mapping rules can divide the weight levels according to the range of the comprehensive correlation. For example, the weight level is 1.0 when the comprehensive correlation is in the range of [0.8, 1.0], the weight level is 0.8 when it is in the range of [0.6, 0.8), and the weight level is 0.6 when it is in the range of [0.4, 0.6].

[0027] This invention fully considers the temporal attributes of data by acquiring the timestamp, time interval, and time series trend of each temporal feature. It captures the temporal variation patterns of the data, thus avoiding the error accumulation problems that may arise from simple time interpolation methods in traditional approaches. By extracting multiple dimensions of information for each temporal feature, the spatiotemporal alignment process becomes more accurate. By meticulously capturing the time interval and trend of each time point, it avoids errors between data with large temporal resolution differences from being directly carried into the next stage of data alignment, thereby improving the accuracy of the final spatiotemporal alignment result. Furthermore, by modeling spatial topological relationships, it can clarify the adjacency or relative relationships between data, further improving spatial alignment. Accuracy is crucial, while spatial distribution density reflects the density of data in space. By combining temporal and spatial feature vectors to calculate correlation weight coefficients and establishing coupling relationships based on them, the relationship between time and space dimensions can be fully considered. This makes spatiotemporal alignment not just simple time or space alignment, but a more comprehensive collaborative alignment. By obtaining the correlation weight coefficients of spatiotemporal features, the relative importance between different features can be accurately modeled, making the coupling relationship more adaptable and accurate. By calculating the weights and coupling relationships between spatiotemporal features, errors can be effectively controlled and adjusted, avoiding error amplification and improving the accuracy of the final spatiotemporal correlation graph.

[0028] By calculating temporal and spatial similarity separately, the increased complexity and mutual interference of errors that may occur when mixing them for overall similarity calculation are avoided, which helps reduce the impact of noise in the calculation process. By processing temporal and spatial similarity separately before weighted summation, the weight of each dimension in the overall spatiotemporal relationship can be adjusted more finely. By progressively decomposing spatiotemporal features, potential errors can be identified and adjusted in time when calculating each dimension, thereby improving the calculation accuracy. The calculation of comprehensive correlation not only reflects temporal and spatial similarity but also considers their relative importance in spatiotemporal data, thus more accurately determining the connection relationship and edge weight of adjacent nodes. The dynamic allocation of edge weights can accurately reflect the actual spatiotemporal correlation strength between data points, making the spatiotemporal correlation graph more realistically reflect the intrinsic connection between data. By integrating the coupling relationship, edge connection relationship, and edge weight of multiple nodes, an initial spatiotemporal correlation graph is constructed, which comprehensively considers the complex relationship between spatiotemporal features, providing a more accurate basic model for subsequent spatiotemporal data analysis, effectively reducing the propagation and accumulation of errors, and making the spatiotemporal correlation graph more realistic and reliable.

[0029] In one embodiment, step S23, which obtains the corresponding association weight coefficients based on each time feature vector and spatial feature vector, includes: S231. Obtain the original timestamp information corresponding to each data sample based on the time feature vector of each data sample, and calculate the timestamp difference based on the difference between the timestamps of two adjacent data samples. S232. Obtain the average value based on multiple timestamp differences, and calculate the deviation of each timestamp difference from the average value based on the average value and each timestamp difference to obtain the time interval fluctuation value; S233. Extract the feature changes of adjacent data samples in the time dimension based on the time series trend information in the time feature vector of each data sample, and calculate the similarity of the trends of two adjacent time series based on the feature changes using the cosine similarity algorithm to obtain the time series trend similarity. S234. After standardizing the time series trend similarity, time interval fluctuation value and timestamp difference, the contribution of time features is obtained by weighted summation. S235. Obtain the spatial coordinates of the target data sample and the spatial coordinates of adjacent data samples based on the spatial feature vector of each data sample, and calculate the Euclidean distance between the target data sample and adjacent data samples using the Euclidean distance formula. S236. Obtain the spatial coordinate set of all data samples based on multiple spatial feature vectors, and construct a spatial topology network based on the spatial coordinate set; S237. Determine whether there is a direct connection between the target node and other nodes based on the spatial topology network. If there is a direct connection, mark it as an adjacent node and count the number of adjacent nodes of the target node. S238. The spatial topological adjacency is calculated using the topological adjacency formula based on the number of adjacent nodes of the target node and the total number of nodes in the spatial topological network. S239. Obtain the number of data samples and the local spatial volume of the target data sample within the local region based on the spatial feature vector of each data sample, and calculate the local spatial distribution density based on the ratio of the number of data samples to the local spatial volume. S2310. Obtain the total number of data samples and the total volume of the space covered by the dataset based on the spatial feature vector of the entire standardized dataset, and calculate the global spatial distribution density of the entire dataset based on the ratio of the total number of data samples to the total volume of the space covered by the dataset. S2311. Calculate the spatial distribution density deviation value using the density deviation formula based on the local spatial distribution density and the global spatial distribution density; S2312. After standardizing the spatial distribution density deviation, spatial topological adjacency, and spatial coordinate Euclidean distance, the spatial feature contribution is calculated by weighted summation. The correlation weight coefficient is obtained by weighted summation based on the spatial feature contribution and the temporal feature contribution.

[0030] As described in steps S231-S2312 above, the target node is the target data sample, and other nodes are other data samples. This invention obtains the original timestamp information of each data sample and calculates the timestamp difference between adjacent samples, which can accurately reflect the relative time interval between data points in the time series, avoiding error amplification in the interpolation process. By calculating the average value of the timestamp difference, the central trend of the overall time difference between data samples can be obtained, which helps to analyze the time stability of the data. By evaluating the degree of fluctuation of the time interval, the instability of the data in the time dimension can be reflected. The trend of the time series reflects the change pattern of the data over time. By calculating the similarity between two adjacent time series using the cosine similarity algorithm, the alignment of data in the time dimension not only depends on the matching of time points, but also captures the gradual change characteristics of the data in time. By calculating the contribution of time features by weighted summation, the role of time features in spatiotemporal alignment can be reflected more comprehensively and accurately.

[0031] Calculating the distance between spatial coordinates using the Euclidean distance formula allows for precise quantification of the relative positions of data points in space. By aggregating the spatial coordinates of all data samples to construct a spatial topological network, the interrelationships between spatial data samples can be graphically represented, facilitating subsequent analysis. This spatial topological network not only focuses on the direct adjacency relationships of data points but also represents the indirect relationships between spatial samples through its network topology, thereby enhancing the flexibility and accuracy of spatial alignment. Determining whether a target node has a direct connection with other nodes reveals the local topological structure of data samples in space, helping to discover spatial clustering characteristics and distribution patterns. Topological connectivity assessment effectively identifies the actual relationships between spatial data samples, avoiding unnecessary errors in spatial interpolation. Calculating spatial topological adjacency using the topological adjacency formula quantifies the relative relationship between a node and its neighboring nodes, further revealing the distribution and structure of data in the spatial topological network. Calculating the ratio of the number of data samples to the spatial volume within a local region reflects the data distribution density within that region. Local spatial distribution density effectively identifies and corrects spatial differences between different regions, thereby improving the spatiotemporal alignment effect.

[0032] Calculating the global spatial distribution density allows us to obtain the overall spatial distribution of the dataset, thereby assessing the spatial uniformity of the data globally. The global spatial distribution density helps avoid the spread of local alignment biases and ensures the accuracy of global alignment. By calculating the deviation between local and global spatial distribution densities, we can finely adjust the spatial distribution of the data to better reflect its spatial characteristics. This fine-tuning can effectively improve the accuracy of spatiotemporal alignment and effectively avoid the accumulation of errors caused by density differences during spatial interpolation.

[0033] In one embodiment, step S3, which involves synchronously interpolating the corresponding standardized dataset using a joint spatiotemporal interpolation algorithm based on each spatiotemporal correlation graph to obtain the joint spatiotemporal interpolation result, includes: S31. Obtain the time feature sequence and spatial feature sequence of the data sample based on the spatiotemporal correlation diagram, and obtain the mean of the time interval, the timestamp distribution density and the rate of change of the time trend based on the time feature sequence; S32. Determine the time interpolation accuracy requirements based on the average time interval, timestamp distribution density, and time trend change rate, and determine the time interpolation window of the joint spatiotemporal interpolation algorithm based on the time interpolation accuracy requirements; S33. Obtain the spatial coordinate dispersion, the number of spatial neighboring points, and the spatial topological association strength based on the spatial feature sequence, and determine the spatial interpolation accuracy requirements based on the spatial coordinate dispersion, the number of spatial neighboring points, and the spatial topological association strength. S34. Determine the spatial interpolation radius of the joint spatiotemporal interpolation algorithm according to the spatial interpolation accuracy requirements, and divide the spatial feature sequence into neighborhoods according to the spatial interpolation radius to obtain multiple spatial neighborhood sequences; S35. The time feature sequence is segmented according to the time interpolation window to obtain multiple time segment sequences. The joint spatiotemporal interpolation algorithm is used to perform time dimension interpolation on the multiple time segment sequences and spatial dimension interpolation on the multiple spatial neighborhood sequences to obtain the joint spatiotemporal interpolation result.

[0034] As described in steps S31-S35 above, the step of determining the spatial interpolation radius of the joint spatiotemporal interpolation algorithm based on the spatial interpolation accuracy requirements includes: calculating the minimum neighborhood sampling distance of the spatial dimension based on the spatial interpolation accuracy requirements and the spatial coverage range, and determining the basic length of the spatial interpolation radius based on the minimum neighborhood sampling distance and the spatiotemporal coupling coefficient; judging whether the number of spatial neighborhood points within the basic length range meets the interpolation calculation standard based on the number of spatial neighborhood points required; if not, expanding the basic length by a preset ratio until the number of neighborhood points within the radius meets the requirements, thus obtaining the final spatial interpolation radius.

[0035] Traditional methods typically use fixed time windows for interpolation, which cannot adapt to the temporal characteristics of different datasets. This often results in insufficient interpolation accuracy for data with varying time intervals and distribution densities, leading to errors. This invention analyzes the temporal and spatial characteristics of data samples to gain a preliminary understanding of the data's distribution patterns in both time and space. By calculating relevant parameters of the temporal feature sequence in detail, it provides a better understanding of the data's temporal distribution and trends. By analyzing the mean of time intervals, timestamp distribution density, and the rate of change of time trends, it determines the required temporal interpolation accuracy and identifies an appropriate time interpolation window. This improves the accuracy of temporal interpolation and effectively avoids interpolation errors caused by inappropriate time windows, contributing to accurate time alignment and ensuring that the final spatiotemporal alignment result is optimized in the temporal dimension. By analyzing spatial feature sequences and calculating spatial coordinate dispersion, the number of spatial neighboring points, and the strength of spatial topological associations, it helps to understand the distribution and structure of data in the spatial dimension, providing a basis for the accuracy requirements of spatial interpolation. It can accurately assess the spatial distribution and correlation of data, thus providing accurate input for subsequent spatial interpolation.

[0036] Based on spatial feature parameters, the required spatial interpolation accuracy is determined, and an appropriate spatial interpolation radius is selected to ensure that the interpolation accuracy in the spatial dimension meets the requirements. By dynamically adjusting the spatial interpolation radius based on spatial features, optimization can be performed on the specific spatial structure of each dataset, significantly improving the accuracy of spatial interpolation and avoiding the accumulation of errors in the spatiotemporal alignment results. By using a joint spatiotemporal interpolation algorithm to perform temporal dimension interpolation on multiple time segmented sequences and spatial dimension interpolation on multiple spatial neighborhood sequences, a joint spatiotemporal interpolation result is finally obtained, thereby achieving comprehensive spatiotemporal alignment. This ensures that the errors generated during temporal interpolation are not carried over to spatial interpolation, improving the accuracy of spatiotemporal alignment and avoiding the accumulation and amplification of errors.

[0037] In one embodiment, step S32, which determines the time interpolation window of the joint spatiotemporal interpolation algorithm based on the time interpolation accuracy requirement, includes: S321. Obtain the effective acquisition duration based on the time span of the time feature sequence, and obtain the minimum data sampling interval of the time dimension based on the time interpolation accuracy requirements and the effective acquisition duration. S322. Obtain the time allowable error threshold and the spatial allowable error threshold respectively according to the time interpolation accuracy requirements and the spatial interpolation accuracy requirements; S323. Obtain the temporal and spatial statistical standard deviations from the historical interpolation data, and calculate the time dimension weights based on the time tolerance threshold and the temporal statistical standard deviations. The calculation formula is as follows: ; Where A represents the weight of the time dimension, a represents the standard deviation of time statistics, and b represents the time tolerance threshold. S324. Obtain the spatial dimension weights based on the spatial statistical standard deviation and the spatial allowable error threshold. Then, calculate the spatiotemporal coupling coefficient by multiplying the weighted sum of the spatial dimension weight coefficients and the time dimension weight coefficients with the preset basic coupling coefficient of the joint spatiotemporal interpolation algorithm. S325. Obtain the spatiotemporal coupling coefficient based on the spatiotemporal dimension weight system, and obtain the time interpolation window based on the spatiotemporal coupling coefficient and the minimum data sampling interval.

[0038] As described in steps S321-S325 above, the spatial dimension weight is calculated in the same way as the time dimension weight. The spatiotemporal dimension weight coefficient includes the time dimension weight coefficient and the spatial dimension weight coefficient. The time dimension weight coefficient is calculated based on the standard normalized spatial dimension weight and the time dimension weight. The calculation formula is as follows: Where C represents the time dimension weight coefficient, A represents the time dimension weight, and B represents the spatial dimension weight; the spatial dimension weight coefficient is calculated based on the standard normalized spatial dimension weight and the time dimension weight, using the following formula: Where D represents the spatial dimension weight coefficient.

[0039] This invention determines the effective acquisition duration by calculating the time span, ensuring that only valid data intervals are used during interpolation, avoiding errors caused by invalid or unreliable data. Combining the temporal interpolation accuracy requirements and the effective acquisition duration to determine the minimum data sampling interval guarantees the rationality of the data sampling interval, avoiding oversampling or sparse sampling, and ensuring high accuracy of the interpolation results. Optimizing data sampling ensures that each time point in the interpolation process is supported by high-quality data, thereby reducing errors introduced by data discontinuity or data quality issues. Based on the temporal and spatial interpolation accuracy requirements, time and spatial error thresholds are determined separately, ensuring that the interpolation accuracy is neither lower than the required accuracy nor excessively high, avoiding unnecessary computational burden. By independently setting the time and spatial error thresholds, fine-grained and adjustable error control is achieved, thereby improving the reliability and accuracy of spatiotemporal interpolation, reducing error propagation and accumulation. By analyzing the temporal and spatial standard deviations of historical interpolation data, the weight of the time dimension is dynamically calculated, allowing for flexible adjustment of the emphasis on the time dimension based on actual data fluctuations, avoiding the incompatibility caused by static weight settings.

[0040] This invention calculates the spatiotemporal coupling coefficient by combining the weights of spatial and temporal dimensions, taking into account the mutual influence between the two dimensions. This provides more accurate spatiotemporal interpolation results. The coupling coefficient effectively reduces the accumulation of errors between the two dimensions, thereby improving the overall accuracy of the spatiotemporal alignment results and reducing the impact of temporal interpolation errors on spatial interpolation. By combining the spatiotemporal coupling coefficient and the minimum data sampling interval, a suitable temporal interpolation window is calculated. This not only considers the coupling relationship between the spatiotemporal dimensions but also the actual situation of the data sampling interval, ensuring the best balance between interpolation accuracy and efficiency. It can effectively avoid over-interpolation or under-interpolation, thereby improving the accuracy and efficiency of data processing.

[0041] In one embodiment, step S6, which constructs a global spatiotemporal coordinate system based on the spatiotemporal feature parameters of the final aligned dataset, includes: S61. Obtain the spatiotemporal feature parameters of the final aligned dataset, where the spatiotemporal feature parameters include the timestamp extreme values, spatial coordinate extreme values, and spatiotemporal distribution density of all data samples. S62. Determine the global time dimension value range based on multiple timestamp extreme values, and obtain the time unit scale value based on the global time dimension value range; S63. Determine the range of global spatial dimensions based on multiple spatial coordinate extreme values, and obtain the spatial unit scale value based on the range of global spatial dimensions. S64. Determine the basic scale system of the global spatiotemporal coordinate system based on the time unit scale value and the spatial unit scale value, and divide the time sub-interval and spatial sub-region in the basic scale system according to the spatiotemporal distribution density. S65. Obtain the number of first data points and the time interval difference in the time sub-interval, and obtain the distribution density coefficient of the time dimension based on the number of first data points and the time interval difference. S66. Determine the first precision level of the time dimension grid based on the distribution density coefficient; S67. Obtain the number of second data points and spatial coordinate deviation value of the spatial sub-region, and obtain the distribution uniformity coefficient of the spatial dimension based on the number of second data points and spatial coordinate deviation value. S68. Determine the second precision level of the spatial dimension grid based on the distribution uniformity coefficient, and adjust the scale interval of the basic scale system according to the first and second precision levels to obtain the global spatiotemporal coordinate system.

[0042] As described in steps S61-S68 above, this invention comprehensively grasps the spatiotemporal distribution characteristics of the dataset by obtaining the spatiotemporal feature parameters of the final aligned dataset, ensuring that subsequent alignment operations can fully consider the spatiotemporal distribution characteristics of the data. By introducing spatiotemporal distribution density and extreme values, the essence of the dataset can be understood more accurately, avoiding the accumulation of errors. By considering the timestamp extreme values ​​of all data samples to determine the value range of the global time dimension, the temporal distribution characteristics of the data can be more comprehensively covered, allowing the scale value of the time dimension to be flexibly adjusted according to the actual data distribution, avoiding the precision loss caused by fixed time granularity. By analyzing the extreme values ​​of spatial coordinates, the accuracy is determined. By defining the range of spatial dimensions and selecting appropriate scale values ​​for spatial units, it can automatically adapt to different spatial distribution characteristics in scenarios with multi-source heterogeneous data, thereby reducing spatial interpolation errors caused by differences in spatial resolution. By combining time and spatial unit scale values, it divides time sub-intervals and spatial sub-regions in the basic scale system, so that spatiotemporal alignment is not only limited to coarse time and spatial scales, but can also be refined to the specific distribution of each time period and spatial region. This can effectively reduce errors caused by uneven data distribution during time and spatial interpolation, and better adapt to the complex distribution of heterogeneous datasets, thereby reducing the accumulation and amplification of errors.

[0043] By analyzing the number of data points and the time interval difference within each time sub-interval, we can further understand the distribution density of the time dimension and calculate the distribution density coefficient. This helps quantify the uneven distribution of data in the time dimension, thus providing a basis for subsequent time dimension grid precision adjustment. By introducing the distribution density coefficient to dynamically determine the precision level of the time dimension grid, the precision level adjustment can be optimized according to the density and interval differences of the actual data distribution. It can adaptively adjust the grid precision according to the actual distribution of spatiotemporal data, maximizing the accuracy of spatiotemporal alignment and reducing the error generated during time interpolation. By analyzing each space... The number of data points and spatial coordinate deviation values ​​in the inter-region can effectively determine the uniformity of spatial distribution, avoiding the limitations of assuming uniform data distribution in traditional spatial interpolation methods. This can more accurately reflect the non-uniformity of spatial distribution. Adjusting the scale interval of the global spatiotemporal coordinate system according to the accuracy levels of time and space dimensions ensures the accuracy of spatiotemporal alignment results. Furthermore, by adaptively adjusting the accuracy of time and space dimensions, the impact of time interpolation and spatial interpolation errors is reduced. By comprehensively considering the accuracy requirements of both, the scale interval of the global spatiotemporal coordinate system can be maximized to adapt to the data distribution, avoiding the accumulation of errors in multiple dimensions.

[0044] like Figure 2 As shown, this application also provides a spatiotemporal alignment system for multi-source heterogeneous data, comprising: The classification module is used to obtain the basic attribute information and spatiotemporal reference parameters of multi-source heterogeneous data, and to perform classification and standardization processing on the multi-source heterogeneous data according to the basic attribute information and spatiotemporal reference parameters to obtain multiple sets of standardized datasets. An extraction and construction module is used to extract the temporal and spatial features of each data sample in each of the standardized datasets, and to construct a spatiotemporal correlation graph of the corresponding standardized dataset based on the multiple temporal and spatial features. The interpolation module is used to perform synchronous interpolation processing on the corresponding standardized dataset according to each spatiotemporal correlation graph using a joint spatiotemporal interpolation algorithm to obtain the joint spatiotemporal interpolation result, and to acquire the temporal error component and spatial error component in real time during the interpolation process; The judgment module is used to obtain the corresponding total error evaluation value based on each of the time error component and the spatial error component, and to determine whether the total error evaluation value is greater than a preset threshold. If the total error assessment value is not greater than the preset threshold, the joint spatiotemporal interpolation result will be used as the intermediate aligned dataset. If the total error assessment value is greater than the preset threshold, then the error correction coefficient is obtained according to the time error component and the spatial error component, and the error correction coefficient is fed back to the joint spatiotemporal interpolation algorithm to dynamically adjust the interpolation parameters until the total error assessment value is not greater than the preset threshold, and the joint spatiotemporal interpolation result at this time is used as the intermediate aligned dataset. The aggregation and construction module is used to integrate and aggregate multiple intermediate alignment datasets to obtain a final alignment dataset, and to construct a global spatiotemporal coordinate system based on the spatiotemporal feature parameters of the final alignment dataset. The mapping module is used to map all data points in the final aligned dataset to the global spatiotemporal coordinate system, thereby completing the spatiotemporal alignment of multi-source heterogeneous data.

[0045] In one embodiment, the extraction building module includes: The first acquisition unit is used to acquire the timestamp, time interval, and time series trend of each time feature, and to establish a corresponding time feature vector based on each timestamp, time interval, and time series trend. A unit is established to acquire the spatial coordinates, spatial topological relationships, and spatial distribution density information of each spatial feature, and to establish a corresponding spatial feature vector based on each spatial coordinate, spatial topological relationship, and spatial distribution density information. The first determining unit is used to obtain the corresponding association weight coefficient based on each of the time feature vectors and spatial feature vectors, and to determine the corresponding coupling relationship based on each of the association weight coefficients; The second acquisition unit is used to take each data sample as a node and acquire the temporal similarity and spatial similarity respectively based on the temporal feature vector and spatial feature vector of two adjacent nodes. The second determining unit is used to obtain the comprehensive correlation degree between two adjacent nodes based on each of the time similarity and spatial similarity, and to determine the edge connection relationship and edge weight between the two adjacent nodes based on each of the comprehensive correlation degrees. The building unit is used to construct an initial spatiotemporal relation graph based on the coupling relationship, edge connection relationship and edge weight of multiple nodes.

[0046] It should be noted that each module and unit in the multi-source heterogeneous data spatiotemporal alignment system corresponds one-to-one with the steps in the multi-source heterogeneous data spatiotemporal alignment method.

[0047] like Figure 3 As shown, this application also provides a computer device, which can be a server, and its internal structure can be as follows: Figure 3As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores all data required for the process of the multi-source heterogeneous data spatiotemporal alignment method. The network interface is used for communication with external terminals via a network connection. The computer program is executed by the processor to implement the multi-source heterogeneous data spatiotemporal alignment method.

[0048] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer equipment on which the present application is applied.

[0049] An embodiment of this application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements any of the above-described multi-source heterogeneous data spatiotemporal alignment methods.

[0050] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in this application and in the embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual-speed SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0051] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0052] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for spatiotemporal alignment of multi-source heterogeneous data, characterized in that, include: Obtain the basic attribute information and spatiotemporal reference parameters of multi-source heterogeneous data, and perform classification and standardization processing on the multi-source heterogeneous data according to the basic attribute information and spatiotemporal reference parameters to obtain multiple sets of standardized datasets; Extract the temporal and spatial features of each data sample in each of the standardized datasets, and construct a spatiotemporal correlation graph of the corresponding standardized datasets based on the multiple temporal and spatial features; Based on each of the spatiotemporal correlation graphs, a joint spatiotemporal interpolation algorithm is used to synchronously interpolate the corresponding standardized dataset to obtain the joint spatiotemporal interpolation result, and the temporal error component and spatial error component are obtained in real time during the interpolation process; Based on each of the time error components and spatial error components, obtain the corresponding total error evaluation value, and determine whether the total error evaluation value is greater than a preset threshold. If the total error assessment value is not greater than the preset threshold, the joint spatiotemporal interpolation result will be used as the intermediate aligned dataset. If the total error assessment value is greater than the preset threshold, then the error correction coefficient is obtained according to the time error component and the spatial error component, and the error correction coefficient is fed back to the joint spatiotemporal interpolation algorithm to dynamically adjust the interpolation parameters until the total error assessment value is not greater than the preset threshold, and the joint spatiotemporal interpolation result at this time is used as the intermediate aligned dataset. The multiple intermediate alignment datasets are integrated and summarized to obtain the final alignment dataset, and a global spatiotemporal coordinate system is constructed based on the spatiotemporal feature parameters of the final alignment dataset. All data points in the final aligned dataset are mapped to the global spatiotemporal coordinate system to complete the spatiotemporal alignment of multi-source heterogeneous data.

2. The spatiotemporal alignment method for multi-source heterogeneous data according to claim 1, characterized in that, The step of classifying and standardizing multi-source heterogeneous data based on the basic attribute information and spatiotemporal reference parameters to obtain multiple sets of standardized datasets includes: Acquire basic attribute information and spatiotemporal reference parameters, wherein the basic attribute information includes data source type, temporal resolution and spatial resolution, and the spatiotemporal reference parameters include a unified time reference system and a unified spatial coordinate system; Based on the data source type, the multi-source heterogeneous data is divided into multiple initial datasets, and the initial datasets are further divided according to the temporal resolution of the data samples in each initial dataset to obtain multiple time-resolved datasets. Obtain the original timestamp information of the data samples in each time-resolved dataset, and obtain the corresponding time deviation value and time granularity difference value based on each original timestamp information and the unified time reference system; Based on each time deviation value and time granularity difference value, the original timestamp information of the data samples in each time-resolved dataset is subjected to time calibration and time granularity normalization processing to obtain the corresponding standardized time-resolved dataset. The standardized time-resolved dataset is divided according to the spatial resolution of the data samples in each standardized time-resolved dataset to obtain multiple spatially resolved datasets. Obtain the original spatial coordinate information of the data samples in each spatially resolved dataset, and obtain the corresponding spatial coordinate transformation parameters and spatial accuracy deviation values ​​based on each original spatial coordinate information and the unified spatial coordinate system; Based on each of the spatial coordinate transformation parameters and spatial accuracy deviation values, coordinate transformation and spatial accuracy correction are performed on the original spatial coordinate information of the data samples in each spatially resolved dataset to obtain the corresponding standardized dataset.

3. The spatiotemporal alignment method for multi-source heterogeneous data according to claim 1, characterized in that, The step of constructing a spatiotemporal correlation graph of the corresponding standardized dataset based on multiple temporal and spatial features includes: Obtain the timestamp, time interval, and time series trend of each time feature, and establish a corresponding time feature vector based on each timestamp, time interval, and time series trend; Obtain the spatial coordinates, spatial topological relationships, and spatial distribution density information of each spatial feature, and establish a corresponding spatial feature vector based on each spatial coordinate, spatial topological relationship, and spatial distribution density information; Obtain the corresponding association weight coefficient based on each of the said time feature vectors and spatial feature vectors, and determine the corresponding coupling relationship based on each of the said association weight coefficients; Each data sample is treated as a node, and the temporal similarity and spatial similarity are obtained based on the temporal feature vector and spatial feature vector of the two adjacent nodes, respectively. The comprehensive correlation degree between two adjacent nodes is obtained based on each of the time similarity and spatial similarity, and the edge connection relationship and edge weight between the two adjacent nodes are determined based on each of the comprehensive correlation degrees. An initial spatiotemporal relation graph is constructed based on the coupling relationships, edge connection relationships, and edge weights of multiple nodes.

4. The spatiotemporal alignment method for multi-source heterogeneous data according to claim 1, characterized in that, The step of simultaneously interpolating the corresponding standardized dataset using a joint spatiotemporal interpolation algorithm based on each of the spatiotemporal correlation graphs to obtain the joint spatiotemporal interpolation result includes: The time feature sequence and spatial feature sequence of the data sample are obtained based on the spatiotemporal correlation diagram, and the mean time interval, timestamp distribution density and time trend change rate are obtained based on the time feature sequence. The time interpolation accuracy requirement is determined based on the mean of the time interval, the timestamp distribution density, and the rate of change of the time trend, and the time interpolation window of the joint spatiotemporal interpolation algorithm is determined based on the time interpolation accuracy requirement. The spatial coordinate dispersion, the number of spatial neighboring points, and the spatial topological association strength are obtained based on the spatial feature sequence, and the spatial interpolation accuracy requirements are determined based on the spatial coordinate dispersion, the number of spatial neighboring points, and the spatial topological association strength. The spatial interpolation radius of the joint spatiotemporal interpolation algorithm is determined based on the spatial interpolation accuracy requirements, and the spatial feature sequence is divided into neighborhoods based on the spatial interpolation radius to obtain multiple spatial neighborhood sequences. The time feature sequence is segmented according to the time interpolation window to obtain multiple time segment sequences. The joint spatiotemporal interpolation result is obtained by interpolating the time dimension of the multiple time segment sequences and the spatial dimension of the multiple spatial neighborhood sequences through a joint spatiotemporal interpolation algorithm.

5. The spatiotemporal alignment method for multi-source heterogeneous data according to claim 4, characterized in that, The step of determining the time interpolation window of the joint spatiotemporal interpolation algorithm based on the time interpolation accuracy requirement includes: The effective acquisition duration is obtained based on the time span of the time feature sequence, and the minimum data sampling interval of the time dimension is obtained based on the time interpolation accuracy requirements and the effective acquisition duration. Based on the time interpolation accuracy requirements and spatial interpolation accuracy requirements, the time allowable error threshold and spatial allowable error threshold are obtained respectively. Obtain the time statistical standard deviation and spatial statistical standard deviation from the historical interpolation data, and obtain the time dimension weight based on the time allowable error threshold and the time statistical standard deviation; The spatial dimension weights are obtained based on the spatial statistical standard deviation and the spatial allowable error threshold, and the spatiotemporal dimension weight coefficients are obtained based on the spatial dimension weights and the time dimension weights. The spatiotemporal coupling coefficient is obtained based on the spatiotemporal dimension weight system, and the time interpolation window is obtained based on the spatiotemporal coupling coefficient and the minimum data sampling interval.

6. The spatiotemporal alignment method for multi-source heterogeneous data according to claim 1, characterized in that, The step of constructing a global spatiotemporal coordinate system based on the spatiotemporal feature parameters of the final aligned dataset includes: Obtain the spatiotemporal feature parameters of the final aligned dataset, wherein the spatiotemporal feature parameters include the timestamp extreme values, spatial coordinate extreme values, and spatiotemporal distribution density of all data samples; The global time dimension value range is determined based on multiple timestamp extreme values, and the time unit scale value is obtained based on the global time dimension value range; The global spatial dimension range is determined based on multiple extreme values ​​of the spatial coordinates, and the spatial unit scale value is obtained based on the global spatial dimension range. The basic scale system of the global spatiotemporal coordinate system is determined based on the time unit scale value and the spatial unit scale value, and the time sub-intervals and spatial sub-regions are divided in the basic scale system according to the spatiotemporal distribution density. Obtain the number of first data points and the time interval difference in the time sub-interval, and obtain the distribution density coefficient of the time dimension based on the number of first data points and the time interval difference; The first precision level of the time dimension grid is determined based on the distribution density coefficient. The number of second data points and the spatial coordinate deviation value of the spatial sub-region are obtained, and the distribution uniformity coefficient of the spatial dimension is obtained based on the number of second data points and the spatial coordinate deviation value. The second precision level of the spatial dimension grid is determined based on the distribution uniformity coefficient, and the scale interval of the basic scale system is adjusted according to the first precision level and the second precision level to obtain the global spatiotemporal coordinate system.

7. A spatiotemporal alignment system for multi-source heterogeneous data, characterized in that, include: The classification module is used to obtain the basic attribute information and spatiotemporal reference parameters of multi-source heterogeneous data, and to perform classification and standardization processing on the multi-source heterogeneous data according to the basic attribute information and spatiotemporal reference parameters to obtain multiple sets of standardized datasets. An extraction and construction module is used to extract the temporal and spatial features of each data sample in each of the standardized datasets, and to construct a spatiotemporal correlation graph of the corresponding standardized dataset based on the multiple temporal and spatial features. The interpolation module is used to perform synchronous interpolation processing on the corresponding standardized dataset according to each spatiotemporal correlation graph using a joint spatiotemporal interpolation algorithm to obtain the joint spatiotemporal interpolation result, and to acquire the temporal error component and spatial error component in real time during the interpolation process; The judgment module is used to obtain the corresponding total error evaluation value based on each of the time error component and the spatial error component, and to determine whether the total error evaluation value is greater than a preset threshold. If the total error assessment value is not greater than the preset threshold, the joint spatiotemporal interpolation result will be used as the intermediate aligned dataset. If the total error assessment value is greater than the preset threshold, then the error correction coefficient is obtained according to the time error component and the spatial error component, and the error correction coefficient is fed back to the joint spatiotemporal interpolation algorithm to dynamically adjust the interpolation parameters until the total error assessment value is not greater than the preset threshold, and the joint spatiotemporal interpolation result at this time is used as the intermediate aligned dataset. The aggregation and construction module is used to integrate and aggregate multiple intermediate alignment datasets to obtain a final alignment dataset, and to construct a global spatiotemporal coordinate system based on the spatiotemporal feature parameters of the final alignment dataset. The mapping module is used to map all data points in the final aligned dataset to the global spatiotemporal coordinate system, thereby completing the spatiotemporal alignment of multi-source heterogeneous data.

8. The spatiotemporal alignment system for multi-source heterogeneous data according to claim 7, characterized in that, The extraction construction module includes: The first acquisition unit is used to acquire the timestamp, time interval, and time series trend of each time feature, and to establish a corresponding time feature vector based on each timestamp, time interval, and time series trend. A unit is established to acquire the spatial coordinates, spatial topological relationships, and spatial distribution density information of each spatial feature, and to establish a corresponding spatial feature vector based on each spatial coordinate, spatial topological relationship, and spatial distribution density information. The first determining unit is used to obtain the corresponding association weight coefficient based on each of the time feature vectors and spatial feature vectors, and to determine the corresponding coupling relationship based on each of the association weight coefficients; The second acquisition unit is used to take each data sample as a node and acquire the temporal similarity and spatial similarity respectively based on the temporal feature vector and spatial feature vector of two adjacent nodes. The second determining unit is used to obtain the comprehensive correlation degree between two adjacent nodes based on each of the time similarity and spatial similarity, and to determine the edge connection relationship and edge weight between the two adjacent nodes based on each of the comprehensive correlation degrees. The building unit is used to construct an initial spatiotemporal relation graph based on the coupling relationship, edge connection relationship and edge weight of multiple nodes.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images