Geographic information data processing system based on cloud computing

Abstract

The invention relates to the technical field of geographic information data processing, in particular to a geographic information data processing system based on cloud computing, which comprises a data processing module, a division and fusion module, a difference detection module, a correlation analysis module, a feature prediction module and a data selection module. The data processing module preprocesses the received geographic information data; the division and fusion module performs iterative region division and fusion; the difference detection module detects a difference area; the association analysis module obtains an association relationship between the difference region and other regions; the feature prediction module obtains predicted geographic features of the difference region; and the data selection module selects final regional information data. According to the method, the data conflict region is identified through difference detection, the internal complex association of the geographic information data is mined, the geographic features of the difference region are predicted, the most suitable geographic information data is selected as the final data based on similarity calculation, and the intelligence and automation of the data processing process are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of geographic information data processing, and in particular to a geographic information data processing system based on cloud computing. Background Art

[0002] Geographic information data, also often referred to as spatial data, refers to data that describes various geographical elements on the earth's surface and their interrelationships. It is a type of data that contains location information and can be used to describe natural or human phenomena on the earth.

[0003] When integrating multi-source heterogeneous data, existing geographic information data processing technologies usually face problems such as uneven data quality, significant spatial heterogeneity, and insufficient mining of associations. Specifically, traditional methods often lack refined processing of geographic information data characteristics in the data preprocessing stage, resulting in low accuracy of subsequent analysis results; in the process of data fusion, simple and rough regional division and fusion strategies are difficult to ensure the integrity and accuracy of data, and are prone to introduce noise and errors; in view of the differences between different data sources in the same area, existing technologies often use simple superposition or averaging methods, ignoring the quality and reliability of data sources, and lack effective difference detection and processing mechanisms, and cannot accurately identify and locate data conflict areas; in addition, traditional methods do not adequately mine the complex associations inherent in geographic information data, resulting in the inability to fully release data value and limited prediction accuracy; more importantly, in the face of massive geographic information data, traditional methods have bottlenecks in computing efficiency and scalability, and are difficult to meet the needs of large-scale geographic information data processing. Therefore, how to effectively integrate multi-source heterogeneous geographic information data, improve data quality and fusion accuracy, fully mine the inherent associations of data, and improve data processing efficiency and scalability are key issues that need to be urgently solved in existing technologies. Summary of the invention

[0004] The present invention provides a geographic information data processing system based on cloud computing, which is used to solve the defects existing in the prior art.

[0005] The present invention provides a geographic information data processing system based on cloud computing, comprising: The data processing module is used to receive geographic information data from different data sources and pre-process each geographic information data to obtain pre-processed data.

[0006] The division and fusion module is used to iteratively divide and fuse the preprocessed data to generate preliminary complete geographic information data.

[0007] The difference detection module is used to extract the geographical features of each area, and detect the difference values ​​of geographical information data from different data sources in the same area based on the geographical features, set the difference threshold, and regard the area with a difference value higher than the difference threshold as the difference area.

[0008] The association analysis module is used to obtain the association relationship between the difference area and other areas based on geographical features. The association relationship includes spatial association, attribute association and time association.

[0009] The feature prediction module is used to build a prediction model based on a recurrent neural network, input the previous moment's geographical features of the difference area, the area with the maximum spatial correlation of the difference area, the area with the maximum attribute correlation of the difference area, and the area with the maximum temporal correlation of the difference area, and output the predicted geographical features of the difference area.

[0010] The data selection module is used to calculate the similarity between the geographical features of the difference area and the predicted geographical features, and use the information data in the area where the geographical information data with the highest similarity between the difference area and the predicted geographical features is the final regional information data.

[0011] According to a cloud computing-based geographic information data processing system provided by the present invention, the data processing module includes a multi-source data access unit, a data cleaning unit and a data conversion unit. The multi-source data access unit is used to receive geographic information data from different data sources. The data cleaning unit is used to perform cleaning processing on the received geographic information data to obtain cleaned data, and the cleaning processing includes missing value processing and abnormal value processing. The data conversion unit is used to convert the cleaned data into a unified data type as preprocessed data.

[0012] According to a geographic information data processing system based on cloud computing provided by the present invention, the process of iteratively dividing and fusing preprocessed data includes: Initial region division: Use the grid division method to perform initial division on the preprocessed data to obtain N initial regions; Regional data fusion: For each initial region, the Kalman filter algorithm is used to fuse data from different data sources to obtain a fusion result; Fusion effect evaluation: evaluate the fusion quality of the fusion results, which includes data integrity and data consistency; Repeat the initial region division, regional data fusion and fusion effect evaluation, and set the iteration stop criteria, which include the fusion quality reaching a preset threshold.

[0013] According to a geographic information data processing system based on cloud computing provided by the present invention, the difference detection module includes a geometric difference detection unit and an attribute difference detection unit. The geometric difference detection unit is used to detect the difference in geometric shapes of the same area in different data sources. The attribute difference detection unit is used to detect the difference in attribute values ​​of the same area in different data sources.

[0014] According to a geographic information data processing system based on cloud computing provided by the present invention, the difference detection module also includes a spatial relationship feature extraction unit and a statistical feature extraction unit. The spatial relationship feature extraction unit is used to extract the spatial relationship features between each area and its adjacent areas. The statistical feature extraction unit is used to extract the statistical features of the internal attribute values ​​of each area.

[0015] According to a cloud computing-based geographic information data processing system provided by the present invention, spatial association includes proximity and connectivity. Proximity means calculating the physical distance between the difference area and other areas through a spatial weight matrix. Connectivity means determining the road connection between the difference area and other areas through a graph theory algorithm.

[0016] According to a geographic information data processing system based on cloud computing provided by the present invention, the attribute association degree includes attribute similarity and correlation. Attribute similarity indicates the similarity between the difference area and other areas calculated by the Jaccard similarity coefficient. Correlation indicates the attribute statistical dependency relationship between the difference area and other areas calculated by the chi-square test and mutual information.

[0017] According to a geographic information data processing system based on cloud computing provided by the present invention, the temporal correlation includes time series similarity and causal relationship. Time series similarity indicates the similarity of time series data between the difference area and other areas calculated by dynamic time warping and mutual correlation coefficient. Causal relationship indicates the mutual influence relationship between the difference area and other areas analyzed by Granger causality test.

[0018] According to a geographic information data processing system based on cloud computing provided by the present invention, the process of building a prediction model based on a recurrent neural network includes: Collect the geographical features of the difference areas, the areas with the maximum spatial correlation of the difference areas, the areas with the maximum attribute correlation of the difference areas, and the areas with the maximum temporal correlation of the difference areas at historical moments, and mark the geographical features of the difference areas at the next moment corresponding to the geographical features at the historical moment.

[0019] A basic model of a recurrent neural network is constructed, and the geographical features of the difference area, the area with the maximum spatial correlation of the difference area, the area with the maximum attribute correlation of the difference area, and the area with the maximum temporal correlation of the difference area at historical moments are taken as input, and the geographical features of the difference area at the next moment corresponding to the geographical features of the difference area at the historical moment are taken as output. The basic model of the recurrent neural network is trained; the model parameters that meet the test accuracy are retained to obtain the prediction model.

[0020] According to a geographic information data processing system based on cloud computing provided by the present invention, the process of calculating the similarity between the geographic features of the difference area and the predicted geographic features includes: The geographical features of the difference areas and the predicted geographical features are decomposed into numerical feature sets and categorical feature sets respectively.

[0021] The numerical similarity between the numerical feature set in the geographical features of the difference area and the numerical feature set in the predicted geographical features is calculated by the Euclidean distance algorithm.

[0022] The classification similarity between the classification feature set in the geographical features of the difference area and the classification feature set in the predicted geographical features is calculated by the Jaccard similarity coefficient.

[0023] The weighted sum method was used to calculate the similarity between the geographical features of the difference area and the predicted geographical features.

[0024] The present invention provides a geographic information data processing system based on cloud computing. Through the synergy of various modules, the system realizes the automatic and intelligent fusion of multi-source heterogeneous geographic information data, and effectively improves the data quality and processing efficiency. It is compatible with multiple data formats, breaks the data island effect, and realizes the unified access of multi-source heterogeneous data. It adopts an iterative regional division and fusion strategy, and can dynamically adjust the regional boundaries according to the actual distribution of data, so as to realize more refined regional division. The data conflict area can be accurately identified through the difference detection module, and the complex association inherent in the geographic information data can be mined through the association analysis module, so as to provide a decision basis for data selection and fusion. By constructing a prediction model based on a recurrent neural network, the geographical features of the difference area can be effectively predicted, and the most suitable geographic information data can be selected as the final data based on similarity calculation, so as to realize the intelligence and automation of the data processing process. Based on the cloud computing platform, it has good scalability and high concurrent processing capability, and can meet the needs of large-scale geographic information data processing. It has a higher degree of automation, higher processing accuracy, stronger intelligent decision-making ability and better scalability. Therefore, the system can be widely used in urban planning, resource management, environmental protection, disaster management, transportation and other fields, providing users with better geographic information services and providing strong support for scientific research and decision-making in related fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0026] Figure 1 It is a structural diagram of a geographic information data processing system based on cloud computing provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] Combine the following Figure 1 The present invention describes a geographic information data processing system based on cloud computing.

[0029] Figure 1 It is a structural diagram of a geographic information data processing system based on cloud computing provided by an embodiment of the present invention.

[0030] like Figure 1 As shown, an embodiment of the present invention provides a geographic information data processing system based on cloud computing, including a data processing module, a division and fusion module, a difference detection module, an association analysis module, a feature prediction module and a data selection module.

[0031] The data processing module is used to receive geographic information data from different data sources, and pre-process each geographic information data to obtain pre-processed data.

[0032] The data processing module includes a multi-source data access unit, a data cleaning unit and a data conversion unit. The multi-source data access unit is used to receive geographic information data from different data sources. The data cleaning unit is used to perform cleaning processing on the received geographic information data to obtain cleaned data, and the cleaning processing includes missing value processing and abnormal value processing. The data conversion unit is used to convert the cleaned data into a unified data type as preprocessing data.

[0033] In this embodiment, the multi-source data access unit needs to identify which channels the data comes from, such as commercial geographic information system databases, remote sensing image data, multi-source geographic data, sensor network data, etc. For different data sources, corresponding connection channels need to be established, using different technologies and protocols, including API interfaces, database connections, and file reading. The required geographic information data is extracted from the data source, including spatial range data, attribute field selection, and time range filtering. The data cleaning unit receives the geographic information data transmitted by the multi-source data access unit, and processes errors, incompleteness, and inconsistencies in the data to improve data quality. For different missing situations, if the proportion of missing values ​​is very small and has little effect on the analysis, the records containing missing values ​​can be directly deleted. Or the missing values ​​can be filled by filling methods such as mean and median filling. Use statistical methods to detect whether there is data in the data that is obviously deviated from the normal range as an outlier. Statistical methods can include standard deviation method and box plot method. If the outlier is confirmed to be erroneous data, it is directly deleted. If the correct value of the outlier can be determined, it is corrected. Use filtering methods to remove unnecessary fluctuations in the data and highlight the main trends and characteristics of the data. Check and remove data records that are completely duplicated or duplicated in key attributes to ensure data uniqueness. Unify data formats through standardization to resolve inconsistencies in data sets caused by inconsistent units, naming conflicts, and other issues. The data conversion unit converts cleaned data into a unified data type and format for subsequent regional division and fusion, including data type conversion, coordinate system conversion, data format conversion, and data standardization.

[0034] By receiving geographic information data from different data sources and preprocessing them, the quality of subsequent data fusion and analysis is significantly improved. It is compatible with multiple data formats and realizes unified access to multi-source heterogeneous data. In view of the characteristics of geographic information data, the preprocessing process can effectively clean the data, thereby improving the accuracy and reliability of the data. Through data conversion and standardization, data of different scales and units are unified under the same standard, laying the foundation for subsequent regional division and fusion.

[0035] The division and fusion module is used to iteratively divide and fuse the preprocessed data to generate preliminary complete geographic information data.

[0036] The process of iterative region division and fusion of preprocessed data includes: Initial region division: Use the grid division method to perform initial division on the preprocessed data to obtain N initial regions; Regional data fusion: For each initial region, the Kalman filter algorithm is used to fuse data from different data sources to obtain a fusion result; Fusion effect evaluation: evaluate the fusion quality of the fusion results, which includes data integrity and data consistency; Repeat the initial region division, regional data fusion and fusion effect evaluation, and set the iteration stop criteria, which include the fusion quality reaching a preset threshold.

[0037] In this embodiment, the goal of iterative region division and fusion is to obtain a complete and highly consistent geographic information data set by continuously adjusting the region division method and fusing data from different data sources. The entire study area is divided into regular grids, such as square or rectangular grids. The grid size is selected according to the density and spatial distribution characteristics of the data. N initial regions are obtained, each of which contains geographic information data from different data sources. For each initial region, the Kalman filter algorithm is used to fuse the data from different data sources to generate a unified data set. The process includes: Define the state variables that need to be estimated, including the height, position, and attribute values ​​of the features.

[0038] An observation model is established to describe the relationship between observations from different data sources and state variables.

[0039] Establish a process model to describe the change pattern of state variables over time.

[0040] Initialize the state variables and the error covariance matrix.

[0041] The Kalman filter algorithm is used for iterative updating to continuously correct the estimated value of the state variable.

[0042] Based on the process model, predict the state variables at the next moment.

[0043] The predicted values ​​are compared with the actual observed values, the Kalman gain is calculated, and the state variables and the error covariance matrix are updated.

[0044] For each initial region, a fused dataset is obtained, which contains the optimal estimated state of the region.

[0045] Evaluate the quality of the fusion results and determine whether the expected fusion effect has been achieved. Evaluate the missingness of each attribute in the fused data set, including the proportion of missing values ​​and the geographical scope of the data coverage. Evaluate whether there are contradictions or conflicts between different attributes and different data sources in the fused data set, including attribute consistency: whether the type, name, elevation and other attributes of the same feature are consistent. Spatial consistency: whether the geometric shape and location of the same feature are consistent. Perform statistical analysis on the fused data and calculate indicators such as missing rate and consistency.

[0046] Set thresholds for data integrity and data consistency. When the fusion quality reaches the threshold, stop the iteration. For data integrity, check whether there are enough data points in each region, set a minimum data point threshold, and the region below this threshold is marked as incomplete data. For data consistency, calculate the standard deviation between different data sources in the same region. The smaller the standard deviation, the more consistent the data. Set a maximum standard deviation threshold, and the region exceeding this value needs to be re-divided. According to the quality assessment results, for areas with low data integrity or poor data consistency, use a smaller grid for division to capture data features more finely. According to the quality assessment results of each iteration, set a feedback mechanism to decide whether to continue iteration or adjust the grid size. If the fusion effect does not meet the iteration stop standard, repeat the initial region division, regional data fusion and fusion effect evaluation. In each iteration, adjust the grid size or Kalman filter parameters to optimize the fusion effect.

[0047] Through iterative regional division and fusion, the problem of difficult determination of regional division granularity in traditional methods is solved, and the integrity and accuracy of data fusion are improved. Traditional static regional division methods are often difficult to adapt to the complex and changeable characteristics of geographic information data, which easily leads to inaccurate regional boundaries or excessive heterogeneity of data within the region. The iterative division and fusion strategy adopted can dynamically adjust the regional boundaries according to the actual distribution of the data, thereby achieving more refined regional division. During the division process, the module will comprehensively consider the spatial distribution and attribute characteristics of the data to ensure the homogeneity of the data within the region. Effectively handle the stitching problem of regional boundaries to ensure the continuity and integrity of the data. Through iterative division and fusion, preliminary complete geographic information data can be generated, which provides a reliable data foundation for subsequent difference detection and association analysis, and also greatly improves the effect of data visualization and application analysis.

[0048] The difference detection module is used to extract the geographical features of each area, and detect the difference values ​​of geographical information data from different data sources in the same area based on the geographical features, set the difference threshold, and regard the area with a difference value higher than the difference threshold as the difference area.

[0049] The difference detection module includes a geometric difference detection unit and an attribute difference detection unit. The geometric difference detection unit is used to detect the difference in geometric shapes of the same area in different data sources. The attribute difference detection unit is used to detect the difference in attribute values ​​of the same area in different data sources. The difference detection module also includes a spatial relationship feature extraction unit and a statistical feature extraction unit. The spatial relationship feature extraction unit is used to extract the spatial relationship features between each area and its adjacent areas. The statistical feature extraction unit is used to extract the statistical features of the internal attribute values ​​of each area.

[0050] In this embodiment, the process of the geometric difference detection unit detecting the difference in geometric shape of the same area in different data sources includes: aligning the geometric data of different data sources to ensure that they are in the same coordinate system. If the coordinate systems of the data sources are different, coordinate conversion is required. The Douglas-Peucker algorithm is used to reduce the number of geometric vertices and simplify the geometric data. A geometric difference measurement method is selected, including: area difference, calculating the area difference of the same area in different data sources. Perimeter difference, calculating the perimeter difference of the same area in different data sources. Shape index, calculating the shape index difference of the same area in different data sources. The shape index includes: compactness index, describing the compactness of the regional shape. Fractal dimension, describing the complexity of the regional shape. The calculated geometric difference value is compared with a preset geometric difference threshold. If the difference value is higher than the threshold, it is considered that there is a geometric difference in the area and it is marked as a difference area.

[0051] The process of the attribute difference detection unit detecting the difference in attribute values ​​of the same region in different data sources includes: matching attribute names and understanding semantics, determining the attributes to be compared in different data sources, converting attribute values ​​of different data sources into a unified data type, and selecting attribute difference measurement methods, including numerical attributes, category attributes, and text attributes.

[0052] Numerical attributes include: absolute difference, which calculates the absolute difference of the same attribute in different data sources. Relative difference, which calculates the relative difference of the same attribute in different data sources. Standardized difference, which standardizes the attribute value and then calculates the difference. Category attributes include: consistency ratio, which calculates the ratio of the same attribute in different data sources with consistent values. Kappa coefficient, which measures the classification consistency of different data sources on category attributes. Text attributes include: edit distance, which calculates the edit distance between two text strings to reflect their similarity. Semantic similarity, which uses natural language processing technology to calculate the semantic similarity between two text strings. Compare the calculated attribute difference value with the preset attribute difference threshold. If the difference value is higher than the threshold, it is considered that there is an attribute difference in the area and it is marked as a difference area.

[0053] Perform data analysis on the historical geographic information data of the target area, and use the difference value distribution of each area in the historical data to calculate the mean and standard deviation of the difference value. Set the difference threshold to the mean plus K times the standard deviation, with K ranging from 1.5 to 2, to filter out normal fluctuations. During the processing, dynamically adjust the difference threshold according to the changes in real-time data. Including, if the difference value in a certain period of time generally increases, then appropriately increase the threshold.

[0054] The spatial relationship feature extraction unit extracts the spatial relationship features between each region and its adjacent regions, including: determining adjacent regions based on topological relationships, including adjacency, inclusion, and intersection. Extracting the spatial relationship features between each region and its adjacent regions. The features include a spatial weight matrix and a spatial autocorrelation coefficient. The spatial weight matrix describes the strength of the spatial relationship between regions. It includes: an adjacency matrix, where the weight is 1 if two regions are adjacent, otherwise 0. A distance weight matrix, where the weight is inversely proportional to the distance between regions. The spatial autocorrelation coefficient measures the degree of spatial aggregation of regional attribute values.

[0055] The statistical feature extraction unit extracts the statistical features of the internal attribute values ​​of each region, including: selecting the attributes whose statistical features need to be extracted, and calculating the statistical features of the internal attribute values ​​of each region. Including: the mean value, median, standard deviation, variance, minimum value, maximum value and quantile of the attribute value, such as 25% quantile and 75% quantile. And create a histogram of the attribute value, extract the peak, skewness and kurtosis features of the histogram, and reflect the distribution of the attribute value.

[0056] By detecting the difference values ​​of different data sources in the same area and extracting the geographical features of each area, the data conflict area is effectively identified and located, providing a decision-making basis for subsequent data selection and fusion. Due to differences in collection time, collection method, data accuracy, etc., geographic information data from different data sources often have data conflicts, such as the location, attributes or shape of the same geographic feature in different data sources are inconsistent. Traditional simple superposition or averaging methods cannot effectively resolve these data conflicts and easily lead to erroneous analysis results. By setting a difference threshold, this module can accurately identify areas with difference values ​​higher than the threshold and focus on analyzing these areas as difference areas. In addition, this module can also extract the geographical features of each area, such as terrain, landform, vegetation, water system, transportation, etc., to provide basic data for subsequent association analysis. By comprehensively analyzing the geographical features and difference values ​​of the difference area, this module can effectively determine the cause of the data conflict and provide a decision-making basis for subsequent data selection and fusion. For example, if the difference area is located in a high-incidence area of ​​geological disasters, it may be necessary to give priority to the latest data or a data source with higher reliability. This difference detection mechanism can effectively improve the accuracy of data fusion and provide users with more reliable geographic information data.

[0057] The association analysis module is used to obtain the association relationship between the difference area and other areas based on geographical features. The association relationship includes spatial association, attribute association and time association.

[0058] Spatial association includes proximity and connectivity. Proximity means calculating the physical distance between the difference area and other areas through the spatial weight matrix. Connectivity means judging the road connection between the difference area and other areas through the graph theory algorithm.

[0059] In this embodiment, spatial association analysis is used to determine the degree of spatial association between the difference area and other areas, including calculating proximity, calculating connectivity and spatial association synthesis. Among them, the proximity is calculated by constructing a spatial weight matrix to describe the proximity relationship between areas. Including: Adjacency weight matrix: If two areas are adjacent, the weight is 1, otherwise it is 0. Distance weight matrix: The weight is inversely proportional to the distance between the areas. K nearest neighbor weight matrix: For each area, select the K nearest areas as its neighbors, with a weight of 1 and the rest are 0. Based on the spatial weight matrix, the physical distance between the difference area and other areas is calculated using the Euclidean distance or Manhattan distance measurement method. Use the inverse of the distance as the proximity to convert the physical distance into proximity.

[0060] The process of calculating connectivity includes: constructing the road data of the study area into a graph, where nodes represent road intersections and edges represent roads. Using graph theory algorithms to determine the road connections between the difference area and other areas, including: shortest path algorithm, calculating the shortest path length between the difference area and other areas. Connected component analysis, determining whether the difference area and other areas belong to the same connected component.

[0061] The weighted average method is used to combine proximity and connectivity to obtain the final spatial association.

[0062] Attribute association includes attribute similarity and correlation. Attribute similarity means calculating the similarity between the difference area and other areas through the Jaccard similarity coefficient. Correlation means calculating the attribute statistical dependency between the difference area and other areas through the chi-square test and mutual information.

[0063] In this embodiment, attribute association analysis is used to determine the degree of association between the difference area and other areas in terms of attributes, including calculating attribute similarity, calculating correlation and combining attribute association. Calculating attribute similarity includes representing the land use type of the area as a set through the Jaccard similarity coefficient, and calculating the similarity of the land use types of the two areas. The Jaccard similarity coefficient is defined as the size of the intersection of the two sets divided by the size of the union. It can also include cosine similarity, representing the attributes of the area as a vector, and using cosine similarity to calculate the similarity of the attribute vectors of the two areas.

[0064] Calculation of correlation includes chi-square test to test whether there is a significant correlation between two categorical variables. Mutual information measures the degree of mutual dependence between two variables. The greater the mutual information, the higher the degree of correlation between the two variables. Analyze the correlation between categorical and numerical attributes. At the same time, use the information gain rate to correct the preference of mutual information for high cardinality attributes.

[0065] The attribute similarity and correlation are combined using the weighted average method to obtain the final attribute association.

[0066] Temporal correlation includes time series similarity and causality. Time series similarity refers to the similarity of time series data between the difference area and other areas calculated by dynamic time warping and mutual correlation coefficient. Causality refers to the analysis of the mutual influence relationship between the difference area and other areas through Granger causality test.

[0067] In this embodiment, the temporal correlation analysis is used to determine the temporal correlation between the difference region and other regions, including calculating the time series similarity, analyzing the causal relationship and the temporal correlation synthesis. The time series similarity represents the similarity of the time series data between the difference region and other regions calculated by dynamic time warping and the mutual correlation coefficient. Dynamic time warping can measure the time series similarity, allowing the time series to bend on the time axis to find the best matching path, and is suitable for time series of different lengths. The mutual correlation coefficient can measure the correlation between two time series at different time lags, and can be used to discover the lag relationship between time series.

[0068] Causality means analyzing the mutual influence between the difference area and other areas through Granger causality test. Use Granger causality test to determine whether a time series can be used to predict another time series. If a time series X has Granger causality to another time series Y, it means that the past value of X can be used to predict the current value of Y.

[0069] The weighted average method is used to combine the time series similarity and causal relationship to obtain the final time correlation.

[0070] By obtaining the spatial, attribute and temporal correlation between the difference area and other areas, the inherent complex correlation of geographic information data is effectively mined, providing an important basis for model prediction and data selection. Geographic information data often has complex spatial dependence, attribute correlation and temporal evolution. For example, the topography of adjacent areas is often similar, different types of geographical elements may have mutual influence, and geographical elements in the same area may change at different times. Traditional analysis methods often ignore these inherent correlations, resulting in the accuracy and reliability of the analysis results being limited. This module can accurately obtain these correlations by comprehensively analyzing the spatial distance, attribute similarity and temporal correlation between the difference area and other areas. For example, through spatial correlation analysis, the correlation between the difference area and the topography, vegetation coverage and other aspects of the surrounding areas can be found; through attribute correlation analysis, the correlation between the difference area and the surrounding areas in terms of land use type, population density and other aspects can be found; through temporal correlation analysis, the evolution law of the difference area in different periods can be found. These correlations can provide important input features for model prediction and improve the accuracy and reliability of prediction.

[0071] The feature prediction module is used to build a prediction model based on a recurrent neural network, input the previous moment's geographical features of the difference area, the area with the maximum spatial correlation of the difference area, the area with the maximum attribute correlation of the difference area, and the area with the maximum temporal correlation of the difference area, and output the predicted geographical features of the difference area.

[0072] The process of building a prediction model based on a recurrent neural network includes: Collect historical geographical feature data of different areas. Geographic features include: land use type, vegetation coverage, population density, economic indicators, meteorological data and POI data.

[0073] Collect historical geographic feature data of the area with the largest spatial correlation, the area with the largest attribute correlation, and the area with the largest temporal correlation with the difference area, and divide the collected data into training sets and test sets.

[0074] Set the time span and time granularity. The time span should be long enough to capture the long-term trend of geographic features. The time granularity should be selected based on data availability and prediction needs.

[0075] The collected data are processed for missing values, outliers, data standardization and time series processing.

[0076] The historical geographic features of the difference area, the area with the largest spatial correlation in the difference area, the area with the largest attribute correlation in the difference area, and the area with the largest temporal correlation in the difference area are combined into the input features of the model. Assuming the time granularity is monthly, the geographic feature data of the past 12 months can be used as the input of the model. The geographic features of the next moment corresponding to the geographic features of the difference area at the historical moment are used as the output labels of the model. For example, if the input is the geographic feature data of the past 12 months, the output is the geographic feature data of the 13th month.

[0077] Design the structure of the basic model of the recurrent neural network, including the input layer, which receives the input features. It contains multiple recurrent layers. The number of recurrent layers and the number of recurrent units in each layer are hyperparameters that need to be adjusted. The output layer outputs the prediction results.

[0078] The tanh activation function is used for the recurrent layer, and the linear activation function is used for the output layer.

[0079] The recurrent neural network basic model is trained on the training set, the validation set is used to evaluate the performance of the recurrent neural network basic model during the training process, and the test set is used to test the generalization ability of the trained recurrent neural network basic model.

[0080] The model parameters that meet the test accuracy are retained to obtain the prediction model.

[0081] The trained prediction model is deployed on the cloud computing platform, and the previous moment geographic features of the difference area, the area with the maximum spatial correlation of the difference area, the area with the maximum attribute correlation of the difference area, and the area with the maximum temporal correlation of the difference area are input, and the predicted geographic features of the difference area are output.

[0082] In this embodiment, by constructing a prediction model based on a recurrent neural network and inputting the geographical features of the difference area and its associated area at the previous moment, the geographical features of the difference area can be effectively predicted, providing an important reference basis for the data selection module. Recurrent neural networks are particularly good at processing time series data and can capture the laws of geographical features changing over time. In geographic information data processing, this module makes full use of this advantage of RNN and can effectively predict the geographical features of the difference area, such as land use type, vegetation coverage, population density, etc. By inputting the geographical features of the difference area and its spatial, attribute and time correlation area at the previous moment, the model can comprehensively consider the influence of multiple factors, thereby improving the accuracy and reliability of the prediction. The prediction results can provide an important reference basis for the data selection module, for example, the geographic information data with the highest similarity to the predicted geographical features can be selected as the final data of the area. This model prediction mechanism can effectively improve the rationality of data selection, reduce the need for manual intervention, and thus improve the automation and efficiency of data processing. At the same time, the model can also be adaptively adjusted according to actual conditions, for example, the accuracy of the prediction can be improved by adjusting the network structure, optimizing training parameters, etc.

[0083] The data selection module is used to calculate the similarity between the geographical features of the difference area and the predicted geographical features, and use the information data in the area where the geographical information data with the highest similarity between the difference area and the predicted geographical features is the final regional information data.

[0084] The process of calculating the similarity between the geographical features of the difference area and the predicted geographical features includes: The geographical features of the difference areas and the predicted geographical features are decomposed into numerical feature sets and categorical feature sets respectively.

[0085] The numerical similarity between the numerical feature set in the geographical features of the difference area and the numerical feature set in the predicted geographical features is calculated by the Euclidean distance algorithm.

[0086] The classification similarity between the classification feature set in the geographical features of the difference area and the classification feature set in the predicted geographical features is calculated by the Jaccard similarity coefficient.

[0087] The weighted sum method was used to calculate the similarity between the geographical features of the difference area and the predicted geographical features.

[0088] The geographic information data with the highest similarity to the predicted geographic features in the difference area is taken as the final data, so as to obtain the complete and most accurate geographic information data of the target area.

[0089] In this embodiment, by calculating the similarity between the geographical features of the difference area and the predicted geographical features, and selecting the geographical information data with the highest similarity as the final regional information data, the data processing process is intelligent and automated, and the data quality and analysis accuracy are improved. Traditional geographical information data fusion often relies on manual judgment and selection, is easily affected by subjective factors, and has low efficiency. However, by introducing similarity calculation, this module can objectively evaluate the quality and reliability of different data sources, and select the data closest to the predicted results as the final data. This data selection strategy can effectively reduce the demand for manual intervention and improve the automation of data processing. In addition, this module can also select different similarity calculation methods according to different application scenarios, such as Euclidean distance, cosine similarity or correlation coefficient, etc., to adapt to different data characteristics and analysis requirements. By selecting the most suitable geographical information data as the final data, this module can effectively improve the accuracy of data fusion and provide a more reliable data basis for subsequent analysis and application. This intelligent data selection mechanism can effectively improve the efficiency and quality of data processing and provide users with better geographical information services.

[0090] In summary, this embodiment provides a geographic information data processing system based on cloud computing. Through the synergy of various modules, the system realizes the automatic and intelligent fusion of multi-source heterogeneous geographic information data, and effectively improves the data quality and processing efficiency. It is compatible with multiple data formats, breaks the data island effect, and realizes the unified access of multi-source heterogeneous data. It adopts an iterative regional division and fusion strategy, and can dynamically adjust the regional boundaries according to the actual distribution of the data, so as to achieve more refined regional division. The difference detection module can accurately identify the data conflict area, and the association analysis module can mine the inherent complex association of the geographic information data, providing a decision basis for data selection and fusion. By constructing a prediction model based on a recurrent neural network, the geographical features of the difference area can be effectively predicted, and the most suitable geographic information data can be selected as the final data based on the similarity calculation, realizing the intelligence and automation of the data processing process. Based on the cloud computing platform, it has good scalability and high concurrent processing capabilities, and can meet the needs of large-scale geographic information data processing. It has a higher degree of automation, higher processing accuracy, stronger intelligent decision-making ability and better scalability. Therefore, the system can be widely used in urban planning, resource management, environmental protection, disaster management, transportation and other fields, providing users with better geographic information services and providing strong support for scientific research and decision-making in related fields.

[0091] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0092] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A geographic information data processing system based on cloud computing, characterized in that: include: A data processing module, used for receiving geographic information data from different data sources, and preprocessing each of the geographic information data to obtain preprocessed data; A division and fusion module, used for iteratively dividing and fusion of the pre-processed data to generate preliminary complete geographic information data; A difference detection module is used to extract the geographical features of each area, and detect the difference values ​​of geographical information data from different data sources in the same area according to the geographical features, set a difference threshold, and regard the area with a difference value higher than the difference threshold as a difference area; An association analysis module, used for obtaining the association relationship between the difference area and other areas according to the geographical features, wherein the association relationship includes spatial association, attribute association and time association; A feature prediction module is used to construct a prediction model based on a recurrent neural network, input the geographical features of the difference area, the area with the maximum spatial correlation of the difference area, the area with the maximum attribute correlation of the difference area, and the area with the maximum temporal correlation of the difference area at the previous moment, and output the predicted geographical features of the difference area; The data selection module is used to calculate the similarity between the geographical features of the difference area and the predicted geographical features, and use the information data in the area where the geographical information data with the highest similarity between the difference area and the predicted geographical features is the final area information data.

2. A geographic information data processing system based on cloud computing according to claim 1, characterized in that: The data processing module includes a multi-source data access unit, a data cleaning unit and a data conversion unit; the multi-source data access unit is used to receive geographic information data from different data sources; the data cleaning unit is used to perform cleaning processing on the received geographic information data to obtain cleaned data, and the cleaning processing includes missing value processing and outlier processing; the data conversion unit is used to convert the cleaned data into a unified data type as preprocessing data.

3. The geographic information data processing system based on cloud computing according to claim 1, characterized in that: The process of iteratively dividing and fusing the preprocessed data includes: Initial region division: using a grid division method to perform initial division on the preprocessed data to obtain N initial regions; Regional data fusion: for each of the initial regions, a Kalman filter algorithm is used to fuse data from different data sources to obtain a fusion result; Fusion effect evaluation: evaluating the fusion quality of the fusion result, wherein the fusion quality includes data integrity and data consistency; The initial region division, regional data fusion and fusion effect evaluation are repeated, and an iteration stop criterion is set, wherein the iteration stop criterion includes that the fusion quality reaches a preset threshold.

4. The geographic information data processing system based on cloud computing according to claim 1, characterized in that: The difference detection module includes a geometric difference detection unit and an attribute difference detection unit; the geometric difference detection unit is used to detect the difference in geometric shapes of the same area in different data sources; the attribute difference detection unit is used to detect the difference in attribute values ​​of the same area in different data sources.

5. The geographic information data processing system based on cloud computing according to claim 1, characterized in that: The difference detection module also includes a spatial relationship feature extraction unit and a statistical feature extraction unit; The spatial relationship feature extraction unit is used to extract the spatial relationship features between each region and its adjacent regions; the statistical feature extraction unit is used to extract the statistical features of the internal attribute values ​​of each region.

6. A geographic information data processing system based on cloud computing according to claim 4, characterized in that: The spatial association includes proximity and connectivity; the proximity represents the physical distance between the difference area and other areas calculated by a spatial weight matrix; the connectivity represents the road connection between the difference area and other areas determined by a graph theory algorithm.

7. The geographic information data processing system based on cloud computing according to claim 1, characterized in that: The attribute association degree includes attribute similarity and correlation; the attribute similarity represents the similarity between the difference area and other areas calculated by the Jaccard similarity coefficient; the correlation represents the attribute statistical dependency relationship between the difference area and other areas calculated by the chi-square test and mutual information.

8. The geographic information data processing system based on cloud computing according to claim 1, characterized in that: The temporal correlation includes time series similarity and causal relationship; the time series similarity represents the similarity of the time series data between the difference area and other areas calculated by dynamic time warping and mutual correlation coefficient; the causal relationship represents the mutual influence relationship between the difference area and other areas analyzed by Granger causality test.

9. The geographic information data processing system based on cloud computing according to claim 1, characterized in that: The process of building a prediction model based on a recurrent neural network includes: Collect the geographical features of the difference area, the area with the largest spatial correlation in the difference area, the area with the largest attribute correlation in the difference area, and the area with the largest temporal correlation in the difference area at the historical moment, and mark the geographical features of the difference area at the next moment corresponding to the geographical features at the historical moment; A recurrent neural network basic model is constructed, and the geographical features of the difference area, the area with the maximum spatial correlation of the difference area, the area with the maximum attribute correlation of the difference area, and the area with the maximum time correlation of the difference area at historical moments are taken as input, and the geographical features of the difference area at the next moment corresponding to the geographical features at the historical moments are taken as output, and the recurrent neural network basic model is trained; the model parameters that meet the test accuracy are retained to obtain a prediction model.

10. The geographic information data processing system based on cloud computing according to claim 1, characterized in that: The process of calculating the similarity between the geographical features of the difference area and the predicted geographical features includes: Decomposing the geographical features of the difference area and the predicted geographical features into a numerical feature set and a categorical feature set respectively; Calculating the numerical similarity between the numerical feature set in the geographical features of the difference area and the numerical feature set in the predicted geographical features by using the Euclidean distance algorithm; Calculate the classification similarity between the classification feature set in the geographical features of the difference area and the classification feature set in the predicted geographical features by using the Jaccard similarity coefficient; The weighted sum method is used to calculate the similarity between the geographical features of the difference area and the predicted geographical features.

Images