A method and system for monitoring surveying and mapping geographic information quality based on CORS big data

Abstract

The present application discloses a method and system for quality supervision of surveying and mapping geographic information based on CORS big data, which relates to the field of surveying and mapping geography, including: obtaining multiple track point data of CORS big data, obtaining the scope of supervised surveying and mapping geographic information projects, surveying and mapping unit information, and surveying and mapping project metadata; clustering multiple track point data using the DBSCAN clustering algorithm; using the boundary of each cluster obtained as the scope of supervised surveying and mapping geographic information projects to generate a quality supervision scope map; extracting and quantifying the attribute information of the surveying and mapping project metadata and the quality supervision scope map to form a standardized feature vector; constructing a training set to train a classification model based on a support vector machine; inputting new surveying and mapping projects to be supervised and the supervision scope into the classification model, and conducting quality supervision of surveying and mapping geographic information according to the predicted matching results. In view of the low efficiency of quality supervision of surveying and mapping geographic information in the prior art, the present application improves the supervision efficiency.

Description

Technical Field

[0001] The present application relates to the field of surveying and mapping geography, and in particular to a method and system for monitoring the quality of surveying and mapping geographic information based on CORS big data. Background Art

[0002] With the rapid development of modern surveying and mapping geographic information technology, various surveying and mapping projects are increasing, and the amount of collected geographic information data is exploding. At the same time, the quality requirements for surveying and mapping results are also increasing, and it is urgent to establish an efficient and strict surveying and mapping geographic information quality supervision system. However, the current surveying and mapping geographic information quality supervision still faces some problems and challenges that need to be solved.

[0003] Traditional quality supervision of surveying and mapping geographic information mainly relies on manual spot checks and audits, which have defects such as low efficiency and strong subjectivity. Manual supervision requires a lot of manpower and material resources, and it is difficult to fully cover the increasing number of surveying and mapping projects. At the same time, manual audits are easily affected by subjective factors, and the evaluation criteria of different supervisors may differ, making it difficult to ensure the consistency and fairness of supervision results. In addition, the traditional supervision model lacks effective data support and scientific analysis methods, making it difficult to discover hidden quality problems and risks.

[0004] With the popularization and application of satellite positioning technology, the Continuously Operating Satellite Positioning Service (CORS) has become an important means of collecting surveying and mapping geographic information. The CORS system provides users with high-precision and highly reliable navigation and positioning services by deploying multiple continuously operating GNSS base stations in the survey area, and is widely used in various surveying and mapping projects. During the surveying and mapping process, the CORS system will record a large amount of user receiver trajectory point data, which includes the temporal and spatial distribution information of the surveying and mapping operation. However, how to effectively use this massive amount of CORS trajectory data, explore the surveying and mapping operation rules contained therein, and serve the quality supervision of surveying and mapping geographic information, still lacks systematic research and application. Summary of the invention

[0005] In view of the problem of low efficiency in the quality supervision of surveying and mapping geographic information in the prior art, this application provides a method and system for the quality supervision of surveying and mapping geographic information based on CORS big data. The DBSCAN clustering algorithm is used in combination with spatiotemporal constraints to cluster the CORS trajectory point data to obtain a spatiotemporal two-dimensional cluster sequence representing the survey scope of the surveying and mapping project; then, a standardized quality supervision scope map is generated through coordinate transformation. On this basis, the multidimensional attribute characteristics of the surveying and mapping project metadata and the quality supervision scope map are extracted to construct a standardized feature vector of matching correlation. Then, representative surveying and mapping projects and supervision scopes are selected as training samples, and matching annotations are performed based on the standardized feature vectors to obtain a training data set. Finally, the training set is used in combination with geographic spatial correlation constraints to train an improved support vector machine classification model to improve the supervision efficiency.

[0006] The purpose of this application is achieved through the following technical solutions.

[0007] One aspect of the present application provides a method for quality supervision of surveying and mapping geographic information based on CORS big data, including: obtaining multiple trajectory point data of CORS big data, obtaining the scope of supervised surveying and mapping geographic information projects, surveying and mapping unit information, and surveying and mapping project metadata, the surveying and mapping project metadata including surveying and mapping time, surveying and mapping method, and surveying and mapping accuracy; clustering the multiple trajectory point data using the DBSCAN clustering algorithm, dividing the trajectory points into multiple clusters according to the spatial distribution and density characteristics of the trajectory points, each cluster representing the survey scope of a surveying and mapping geographic information project; when clustering, dividing the surveying and mapping time into multiple preset equal-length time periods, and clustering the trajectory points in the same time period into one category; taking the boundary of each obtained cluster as the scope of the supervised surveying and mapping geographic information project, performing coordinate transformation, and generating a quality supervision scope map; and The feature extraction and quantitative representation of the attribute information of the project metadata and the quality supervision scope map are carried out to form a standardized feature vector; wherein, the attribute information of the quality supervision scope map includes the spatial distribution characteristics of the clusters, the density characteristics and the division of the surveying and mapping time periods; based on the obtained clusters, multiple surveying and mapping projects and supervision scopes are selected as training samples, and the obtained standardized feature vectors are used to annotate them, and the annotated matching is carried out with relevant or irrelevant category labels; the feature vectors of the training samples and the corresponding category labels are combined into a training set; the training set is used to train a classification model based on a support vector machine; the input of the classification model is the standardized feature vectors of the surveying and mapping projects and supervision scopes, and the output is the predicted matching results; the new surveying and mapping projects and supervision scopes to be supervised are input into the classification model to obtain the predicted matching results, and the quality supervision of surveying and mapping geographic information is carried out according to the predicted matching results.

[0008] Among them, CORS (Continuously Operating Reference Stations) is a continuously operating GPS reference station network system. CORS big data refers to a large amount of GPS observation data collected by the CORS system, including the location and speed of GPS receivers. These data have the characteristics of large data volume and high temporal and spatial resolution. Surveying and mapping project metadata is descriptive information about surveying and mapping projects, usually including attribute information such as the name of the surveying and mapping project, surveying and mapping unit, surveying and mapping time, surveying and mapping location, surveying and mapping method, and surveying and mapping accuracy. Metadata helps to classify, retrieve, and manage surveying and mapping projects. Supervised surveying and mapping geographic information projects refer to surveying and mapping projects that have been included in the scope of supervision and have clear information such as surveying and mapping units and surveying and mapping time. By analyzing the supervised projects, we can understand the progress of surveying and mapping work and provide reference for subsequent surveying and mapping geographic information quality supervision. Surveying and mapping projects refer to the general term for field measurement, internal processing, and other work carried out to obtain geographic information. A complete surveying and mapping project usually includes multiple links, such as data collection, control measurement, local measurement, data processing, etc. Track points refer to a series of time and location points collected by CORS equipment during movement. The spatial distribution of trajectory points reflects the movement trajectory and surveying scope of the CORS device, while the density reflects the time and accuracy characteristics of the surveying. The surveying scope of the surveying and mapping project can be obtained through cluster analysis of trajectory points. The quality supervision scope map is a map generated by converting the cluster boundaries obtained after clustering the trajectory points into vector surface data. This map intuitively expresses the geographical scope of the surveying and mapping project and can be used for spatial positioning and scope delineation of quality supervision. The predicted matching result refers to the matching metric value obtained after feature extraction and model prediction between the new surveying and mapping project to be supervised and the existing supervision scope map. The matching result can be a binary classification (match / mismatch) or a correlation score, which is used to determine whether the new surveying and mapping project falls within the existing supervision scope. It is an important criterion for the quality supervision of surveying and mapping geographic information.

[0009] Furthermore, the DBSCAN clustering algorithm is used to cluster multiple trajectory point data, including: dividing the trajectory point data according to the surveying and mapping time attribute, and according to the preset time period length, using a sliding window method, sliding on the surveying and mapping time axis with a certain time step to draw multiple overlapping trajectory point subsets of equal-length time periods; wherein the time period length and the time step are set according to the time granularity and supervision frequency of the surveying and mapping project; selecting a subset of trajectory points located in each overlapping time period, extracting the two-dimensional spatial coordinates and timestamp of each trajectory point, and forming a spatiotemporal three-dimensional point set; calculating the relative time position of each trajectory point in the corresponding time period through the timestamp of the trajectory point; using the DBSCAN clustering algorithm on the spatiotemporal three-dimensional point set, by setting the spatiotemporal neighborhood radius Eps and the density threshold MinPts, according to the density distribution of the trajectory points in the spatiotemporal neighborhood, clustering the spatiotemporal three-dimensional point set into multiple density-connected clusters; wherein the spatiotemporal neighborhood radius Eps includes The spatial neighborhood radius and the temporal neighborhood radius are set according to the spatial span, temporal granularity and regulatory scope of the surveying and mapping project respectively; the density threshold MinPts is set according to the density of the spatiotemporal trajectory points; the spatiotemporal coordinate range of all trajectory points in each cluster is extracted as the spatiotemporal survey range of a surveying and mapping geographic information project in the current overlapping time period; the complete spatiotemporal two-dimensional survey range of the surveying and mapping project is obtained by performing a union operation on the spatiotemporal survey ranges in different overlapping time periods; the above steps are repeated until all subsets of trajectory points in all overlapping time periods are clustered, and according to the clustering results of each overlapping time period, a complete spatiotemporal two-dimensional survey range sequence of the surveying and mapping geographic information project is generated through a union operation; the above steps are repeated until all trajectory point data are clustered, and a spatiotemporal two-dimensional survey range sequence of multiple surveying and mapping projects is obtained; the spatiotemporal distribution and association matching of surveying and mapping projects are optimized by analyzing the spatiotemporal topological relationship between the survey ranges of different surveying and mapping projects.

[0010] Furthermore, the coordinates of the scope of the supervised surveying and mapping geographic information projects are transformed to generate a quality supervision scope map, including: extracting the geographic coordinate parameters of the scope of the supervised surveying and mapping geographic information projects, the geographic coordinate parameters include the coordinate system type, projection type, central meridian and origin latitude; according to the coordinate conversion formula or the seven-parameter space coordinate conversion formula, the geographic coordinate parameters are converted into coordinate values ​​under a unified target coordinate system; according to the converted coordinate values, vector polygons are constructed to generate a standardized supervision scope for each surveying and mapping project, and the standardized supervision scope has the same size; the standardized supervision scopes of all projects are vector-merged to generate a quality supervision scope map, which represents the spatial scope of quality supervision of surveying and mapping geographic information.

[0011] Furthermore, the coordinate transformation formula is as follows: ; ; ; Wherein, X, Y, Z: coordinate values ​​in the spatial rectangular coordinate system, respectively representing the coordinate values ​​of the ground point in the geocentric spatial rectangular coordinate system centered at the origin of the coordinate system; It represents the latitude of the earth, which is the latitude angle of a ground point, that is, the angle between the normal line of the ground point and the equatorial plane. It represents geodetic longitude, which means the longitude angle of a ground point; H represents geodetic height, which means the height from the ground point to the reference ellipsoid surface, measured along the normal direction of the ellipsoid, in meters. Indicates the difference in geodetic height. = represents the radius of curvature of the equator; the equator is the circle on the ellipsoid that is orthogonal to the meridian. On the ellipsoid, the radius of curvature of the equator changes with the latitude. This is because the ellipsoid is not a strict sphere, and the radius of curvature near the equator is larger than the radius of curvature near the poles. e represents the first eccentricity of the ellipsoid, which indicates the flattening of the reference ellipsoid; Indicates the correction amount in the direction of the curvature radius of the azimuth circle.

[0012] Furthermore, the seven-parameter space coordinate transformation formula is as follows: ;in, Represents the spatial rectangular coordinate value in the source coordinate system, indicating the three-dimensional coordinate of the ground point in the source coordinate system; Represents the spatial rectangular coordinate value in the target coordinate system, indicating the three-dimensional coordinate of the ground point in the target coordinate system; It represents the translation of the coordinate origin, which indicates the position offset of the origin of the source coordinate system relative to the origin of the target coordinate system; m represents the scale factor, which indicates the scale difference between the source coordinate system and the target coordinate system, and reflects the proportional relationship between the two coordinate systems; Indicates the rotation angle of the three coordinate axes, indicating the rotation angle of the source coordinate system relative to the target coordinate system, rotating around the X, Y, and Z axes respectively; It indicates the coordinate correction in the source coordinate system, which is used to correct the source coordinate deviation caused by surveying and mapping errors, changes in observation conditions, etc. It indicates the coordinate correction in the target coordinate system, which is used to correct the target coordinate deviation caused by surveying and mapping benchmarks, crustal movement and other factors;

[0013] Furthermore, feature extraction and quantitative representation are performed on the attribute information of the surveying and mapping project metadata and the quality supervision scope map to form a standardized feature vector, including: extracting the spatial vector data of each surveying and mapping project scope in the quality supervision scope map, using vector overlay analysis to calculate the overlapping area between each surveying and mapping project scope and the corresponding supervision scope, and obtaining the area ratio feature reflecting the spatial overlap; respectively calculating the edit distance between the surveying and mapping project name and the project name of the supervision record in the attribute information of the quality supervision scope map , and the edit distance between the surveying unit name and the surveying unit name in the regulatory record ,according to and Calculate text similarity features ; Extract the surveying and mapping accuracy in the metadata of the surveying and mapping project, and convert it into the corresponding numerical level; extract the surveying and mapping accuracy level of the supervision requirements in the attribute information of the quality supervision scope map, calculate the difference between the accuracy level of the surveying and mapping project and the accuracy level of the supervision requirements, and obtain the accuracy compliance feature that reflects whether the surveying and mapping project meets the supervision accuracy requirements; extract the surveying and mapping start and end time in the metadata of the surveying and mapping project, and convert it into a timestamp format; extract the start and end time of the supervision time period in the attribute information of the quality supervision scope map, and convert it into a timestamp format, calculate the timestamp difference between the center points of the two time periods, and obtain the time feature that reflects the degree of proximity between the implementation time of the surveying and mapping project and the supervision time period; perform minimum-maximum normalization processing on the area ratio feature, accuracy compliance feature and time feature; perform Z-score normalization processing on the text similarity feature; obtain a standardized feature vector.

[0014] Among them, vector overlay analysis is a common method in GIS spatial analysis. It refers to the process of superimposing two or more vector layers according to certain rules to obtain a new vector layer. In overlay analysis, the spatial relationship (such as intersection, inclusion, etc.) and attribute relationship (such as equality, greater than, etc.) between different layers can be used to extract new information or establish new spatial relationships. In this scheme, vector overlay analysis is mainly used to calculate the spatial overlap between the scope of surveying and mapping projects and the scope of quality supervision. Specifically, the vector surface data representing the scope of surveying and mapping projects is superimposed with the vector surface data representing the scope of quality supervision, and the overlapping area of ​​the two is calculated, and the ratio of the overlapping area to the area of ​​the supervision scope is used as a quantitative indicator of spatial overlap. The higher the spatial overlap, the better the consistency between the actual operating scope of the surveying and mapping project and the scope of supervision requirements. Conversely, it means that the surveying and mapping project may have a problem that the operating scope does not meet the supervision requirements. Therefore, spatial overlap can be used as an important reference indicator for the quality supervision of surveying and mapping geographic information.

[0015] Among them, surveying and mapping accuracy refers to the accuracy of surveying and mapping results, reflecting the degree of closeness between surveying and mapping results and the actual geographical status. Surveying and mapping accuracy is usually expressed in a certain level. Different surveying and mapping scales and different surveying and mapping methods will have different accuracy level requirements. For example, in 1:500 topographic map surveying, the mean error in plane position is less than or equal to 0.15m, and the mean error in elevation is less than or equal to 0.10m to meet the corresponding accuracy requirements. In 1:10000 topographic map surveying, the mean error in plane position is required to be less than or equal to 2.5m, and the mean error in elevation is required to be less than or equal to 1.5m. In this scheme, the surveying and mapping accuracy level is mainly used to compare whether the actual accuracy of the surveying and mapping project meets the requirements of quality supervision. By converting the surveying and mapping accuracy recorded in the metadata of the surveying and mapping project into the corresponding numerical level and comparing it with the regulatory accuracy requirement level recorded in the quality supervision scope map attributes, a quantitative indicator reflecting the degree of compliance with the accuracy of the surveying and mapping project can be obtained. If the actual accuracy level of the surveying and mapping project is higher than or equal to the accuracy level required by supervision, it can be considered that the project meets the quality requirements in terms of accuracy. Otherwise, it means that there is a problem that the accuracy of the surveying and mapping results does not meet the standards, which requires further verification and processing.

[0016] Furthermore, the text similarity features are calculated : ;in, and They are the weight coefficients of the surveying and mapping project name and the surveying and mapping unit name respectively.

[0017] Further, based on the obtained clusters, multiple surveying and mapping projects and regulatory areas are selected as training samples, and the obtained standardized feature vectors are used to annotate them, and the annotations match relevant or irrelevant category labels, including: for each cluster obtained by clustering, the number of surveying and mapping projects n and the number of regulatory areas m in the cluster are counted, and the number of selected training samples is calculated according to the preset sample selection ratio r. ; K-means++ algorithm is used to select k surveying and mapping projects from each cluster as training samples; for each selected surveying and mapping project, one is selected from the regulatory scope with the largest area ratio feature in the corresponding cluster as the matching regulatory scope to form a training sample; the surveying and mapping project metadata and regulatory scope attribute information of each training sample are extracted, and the corresponding standardized feature vector is calculated; the matching label of each training sample is determined according to the preset regulatory conditions, and the standardized feature vector and the matching label are combined into training data, the positive sample label is 1, and the negative sample label is 0; the preset regulatory conditions include area ratio features, text similarity features, accuracy compliance features and time proximity features that meet the set threshold requirements respectively; repeat until all clusters have completed the selection and annotation of training samples, and the training data of all training samples are aggregated to obtain a complete training data set; the training data set is randomly divided into a training set and a validation set, which are used to train and evaluate the classification model for matching surveying and mapping projects with regulatory scopes, and realize intelligent supervision of the quality of surveying and mapping geographic information.

[0018] Further, the training set is used to train a classification model based on a support vector machine, including: selecting a support vector machine (SVM) algorithm as a machine learning binary classification algorithm, and constructing a support vector machine classification model based on training samples in the training set; wherein the standardized feature vector of the training sample is used as an input feature of the SVM, and the matching label of the training sample is used as a target output of the SVM; the standardized feature vector and matching label of the training sample are input into the SVM algorithm, and the optimal classification hyperplane and support vector of the SVM classification model are obtained by solving the following convex quadratic programming problem: ; Constraints: ;in, is the normal vector of the classification hyperplane, is the bias term of the classification hyperplane, is the slack variable of the i-th training sample, is the penalty coefficient, which is used to balance the classification error and model complexity; is the feature mapping function, which takes the d-dimensional input feature vector Mapped to m-dimensional feature space; represents the matching label of the i-th training sample, N is the total number of training samples;

[0019] By solving the above optimization problem, we can obtain the parameters of the optimal classification hyperplane and , forming the SVM classification decision function: ;in, The d-dimensional normalized feature vector representing the surveying and mapping items to be classified; s: is a geospatial correlation measurement function that measures the d-dimensional mapping project feature vector With the p-dimensional regulatory range feature vector The degree of proximity and overlap in geographic space can be obtained by using appropriate spatial similarity calculation methods based on the spatial information such as the geographic coordinates and boundary shapes of the two. The larger the value of , the stronger the geographic spatial correlation between the two; is the weight coefficient of geographic spatial correlation, which controls the influence of geographic spatial correlation on classification decision and can be optimized and selected through methods such as cross-validation.

[0020] For new mapping project feature vector , by calculating the improved classification decision function The sign of The predicted matching label of , then the prediction surveying and mapping project Matching regulatory scope ;like , then the prediction surveying and mapping project Mismatch in regulatory scope . Indicates surveying and mapping projects and regulatory scope The geospatial correlation measure measures the proximity and overlap between the two in geographic space. It can be obtained by using appropriate spatial similarity calculation methods, such as Overlap Index and Hausdorff distance, based on the spatial information such as the geographic coordinates and boundary shapes of the surveying and mapping projects and the regulatory scope. The larger the value, the stronger the geospatial correlation between the surveying and mapping project and the regulatory scope. is the weight coefficient of geographic spatial correlation, which controls the influence of geographic spatial correlation on classification decision. The value of can be optimized and selected through methods such as cross-validation.

[0021] Another aspect of the present application also provides a surveying and mapping geographic information quality supervision system based on CORS big data, which is used to execute a surveying and mapping geographic information quality supervision method based on CORS big data of the present application.

[0022] Compared with the prior art, the advantages of this application are:

[0023] Cluster analysis is performed using trajectory point data in CORS big data to automatically identify the survey scope of surveying and mapping geographic information projects, avoiding the tediousness and subjectivity of manual calibration of survey scopes and improving the efficiency and accuracy of regulatory scope determination. By dividing the surveying and mapping time and considering the time attributes of trajectory points, the temporal correlation and semantic integrity of clustering results are enhanced. Through coordinate conversion processing, the geographic spatial information of different surveying and mapping projects and regulatory records is unified into a standard regulatory scope map, eliminating the spatial position deviation caused by different coordinate systems, laying the foundation for subsequent spatial feature extraction and matching analysis.

[0024] Multi-dimensional feature information is extracted from the metadata of surveying and mapping projects and the attributes of the quality supervision scope map, including spatial overlap, text similarity, accuracy compliance, time proximity, etc., which comprehensively describes the correlation characteristics between surveying and mapping projects and supervision scope in terms of space, semantics, quality, time, etc., and provides rich feature representation for intelligent matching analysis. The introduction of text similarity features based on edit distance overcomes the differences in the expression of text attributes such as project names and unit names, and improves the robustness of semantic matching. The weighted fusion similarity calculation method can flexibly adjust the importance of different text attributes.

[0025] Through normalization, features of different dimensions and scales are mapped to a unified numerical range, eliminating the influence of feature numerical size, making the contribution of each dimension feature to matching analysis more balanced, and improving the effectiveness of feature representation. A training sample selection strategy based on clustering results is designed, and matching samples are constructed by selecting surveying and mapping projects and regulatory areas with the largest overlap within the cluster, which not only covers the diversity of surveying and mapping projects, but also obtains high-quality matching samples, improving the representativeness and effectiveness of training samples.

[0026] The SVM algorithm is used to train the classification model for matching surveying and mapping projects with regulatory scopes. The nonlinear decision boundary between surveying and mapping project features and matching labels is learned through the optimal classification hyperplane, which has good classification generalization ability. The introduction of geospatial correlation features further improves the spatial discrimination of matching predictions. By using the improved SVM classification decision function and comprehensively considering the multi-dimensional feature information and geospatial correlation of surveying and mapping projects, rapid matching predictions between surveying and mapping projects and any regulatory scope can be achieved, supporting real-time intelligent supervision of the quality of surveying and mapping geographic information, and greatly improving the timeliness and accuracy of supervision. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The present application will be further described in the form of exemplary embodiments, which will be described in detail by the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same number represents the same structure, wherein:

[0028] Figure 1It is an exemplary flow chart of a method for quality supervision of surveying and mapping geographic information based on CORS big data according to some embodiments of the present application;

[0029] Figure 2 is an exemplary flow chart of obtaining a time-space two-dimensional survey range sequence of multiple surveying and mapping projects according to some embodiments of the present application;

[0030] Figure 3 is an exemplary flow chart for generating a quality control range map according to some embodiments of the present application;

[0031] Figure 4 is an exemplary flow chart of generating a standardized feature vector according to some embodiments of the present application. DETAILED DESCRIPTION

[0032] The method and system provided in the embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0033] Figure 1 This is an exemplary flow chart of a method for quality supervision of surveying and mapping geographic information based on CORS big data as shown in some embodiments of the present application, including: obtaining multiple trajectory point data of CORS big data, obtaining the scope of supervised surveying and mapping geographic information projects, surveying and mapping unit information, and surveying and mapping project metadata; the surveying and mapping project metadata includes surveying and mapping time, surveying and mapping method and surveying and mapping accuracy; clustering multiple trajectory point data using the DBSCAN clustering algorithm, dividing the trajectory points into multiple clusters according to the spatial distribution and density characteristics of the trajectory points, each cluster representing the survey scope of a surveying and mapping geographic information project; when clustering, dividing the surveying and mapping time into multiple preset equal-length time periods, and clustering the trajectory points in the same time period into one category; taking the boundary of each obtained cluster as the scope of the supervised surveying and mapping geographic information project, performing coordinate transformation, and generating a quality supervision scope map ; Feature extraction and quantitative representation of the attribute information of the surveying and mapping project metadata and the quality supervision scope map to form a standardized feature vector; wherein the attribute information of the quality supervision scope map includes the spatial distribution characteristics of the clusters, the density characteristics and the division of the surveying and mapping time periods; based on the obtained clusters, multiple surveying and mapping projects and supervision scopes are selected as training samples, and the obtained standardized feature vectors are used to annotate them, and the annotated matching is performed with relevant or irrelevant category labels; the feature vectors of the training samples and the corresponding category labels form a training set; the training set is used to train a classification model based on a support vector machine; the input of the classification model is the standardized feature vectors of the surveying and mapping projects and supervision scopes, and the output is the predicted matching results; the new surveying and mapping projects and supervision scopes to be supervised are input into the classification model to obtain the predicted matching results, and the quality supervision of surveying and mapping geographic information is performed according to the predicted matching results.

[0034] S1, obtain multiple trajectory point data of CORS big data, and sign a data sharing agreement with the CORS operation and maintenance unit to regularly obtain CORS station observation data for all or specific areas and specific time periods from the CORS data center. The data format is RINEX (Receiver Independent Exchange Format). The RINEX data is solved and processed to obtain a high-precision three-dimensional coordinate time series for each CORS station. The coordinate format is XYZ coordinates under ECEF (Earth-Centered, Earth-Fixed). According to the identification ID (such as station name, station number) and observation time of each CORS station, the coordinate time series data is converted into trajectory point data. Each trajectory point is recorded in the format of (ID, X, Y, Z, T), where T is UTC (Coordinated Universal Time). The trajectory point data is cleaned and filtered to eliminate abnormal data such as missing coordinates, time confusion, and position offset to obtain a trajectory point data set with reliable quality. , where N is the total number of trajectory points.

[0035] Obtain the scope of the supervised surveying and mapping geographic information projects, retrieve the list of surveying and mapping projects that have been accepted through the business system or database of the surveying and mapping geographic information management department, and extract the unique number PID, project name PNAME, survey area range RANGE and other attribute information of each surveying and mapping project. The survey area range RANGE is recorded in GeoJSON format and represented as a closed polygon area. Each vertex of the polygon is defined by longitude and latitude coordinates (LONG, LAT). For the survey area range in GeoJSON format, use a coordinate conversion tool (such as PROJ) to convert the longitude and latitude coordinates into ECEF-XYZ coordinates to keep them consistent with the coordinates of the trajectory points. Extract the converted XYZ coordinate sequence, build a spatial index (such as R-tree) for the survey area range, and accelerate subsequent spatial relationship calculations.

[0036] Get the metadata of the surveying and mapping project, and query the metadata record of each project in the surveying and mapping business management system through the number PID of the surveying and mapping project. Extract the key metadata attributes of the surveying and mapping project, including: Surveying and mapping time (Stime, Etime): the start and end time of the actual operation of the project, in the format of "YYYY-MM-DD". Surveying and mapping method METHOD: describes the main surveying and mapping technical methods used in the project, such as "RTK", "CORS", "MMS", etc. Surveying and mapping accuracy PRECISION: describes the plane and elevation accuracy indicators of the project results, such as "plane ±5cm, elevation ±10cm". Extract each metadata attribute into a structured record (PID, Stime, Etime, METHOD, PRECISION), which constitutes a complete project description information with the spatial scope of the surveying and mapping project. Get the surveying and mapping unit information of the surveying and mapping project, and associate the surveying and mapping unit registration management system through the number PID of the surveying and mapping project to obtain the undertaking unit information of each surveying and mapping project. Extract the unique identifier UID (such as the unified social credit code) and unit name UNAME of the surveying and mapping unit. Associate the surveying and mapping unit information with the surveying and mapping project information to obtain complete project-unit metadata (PID, UID, UNAME, Stime, Etime, METHOD, PRECISION).

[0037] Figure 2 It is an exemplary flow chart of obtaining a time-space two-dimensional survey range sequence of multiple surveying and mapping projects according to some embodiments of the present application, using the DBSCAN clustering algorithm to cluster multiple trajectory point data, and dividing the trajectory points into multiple clusters according to the spatial distribution and density characteristics of the trajectory points, each cluster represents the survey range of a surveying and mapping geographic information project; when clustering, the surveying and mapping time is divided into multiple preset equal-length time periods, and the trajectory points in the same time period are clustered into one category; including: dividing the trajectory point data according to the surveying and mapping time attribute, and the trajectory point data set is recorded as ,in is the UTC timestamp of each track point. According to the start and end time (Stime, Etime) of the surveying and mapping project, a subset of track points falling within the interval [Stime, Etime] is extracted from D. . Set the time period length to L and the time step to S, where the values ​​of L and S need to be determined according to the time granularity and supervision frequency of the surveying and mapping project. Usually, L is in the order of hours to days, and S is in the order of hours to hours. For example, for urban road surveying projects with daily supervision frequency, L=24 hours and S=2 hours can be taken; for pipeline census projects with monthly supervision frequency, L=7 days and S=1 day can be taken.

[0038] The sliding window method is used to divide the overlapping time periods. The start and end time of the surveying and mapping project are set as [Stime, Etime], and the starting time of the sliding window is defined as , the initial value is Stime. Time: Define the time period of the current sliding window as ,in . Let the current time period be , where j is the number of the time period, the initial value is 1, and it increases by 1 each time. Move S backward, that is . Continue the next cycle until After the loop is completed, a series of overlapping time periods on the time axis of the surveying and mapping project are obtained. , where M is the total number of time periods. For example, if Stime = "2024-06-01 00:00:00", Etime = "2024-06-05 00:00:00", L = 1 day, S = 0.5 day, the following 7 time periods can be obtained:

[0039] W_1=["2024-06-0100:00:00""2024-06-0200:00:00"],

[0040] W_2=["2024-06-0112:00:00""2024-06-0212:00:00"],

[0041] W_3=["2024-06-0200:00:00" "2024-06-0300:00:00"],

[0042] W_4=["2024-06-0212:00:00" "2024-06-0312:00:00"],

[0043] W_5=["2024-06-0300:00:00" "2024-06-0400:00:00"],

[0044] W_6=["2024-06-0312:00:00" "2024-06-0412:00:00"],

[0045] W_7=["2024-06-0400:00:00" "2024-06-0500:00:00"].

[0046] Extract the coordinates and timestamps of the trajectory points in each time period and traverse each time period , from the trajectory point subset Extracting satisfaction from Falling The trajectory point inside ,in is the track point number in the current time period, is the total number of trajectory points in the current time period. For each trajectory point P_jk, extract its two-dimensional space coordinates and UTC timestamp , forming a three-dimensional point in space and time Finally, we get The corresponding three-dimensional point set in space and time , a total of M point sets.

[0047] Calculate the relative time position of each trajectory point within the time period. For each trajectory point P'_jk in : ,in, The relative value in the interval indicates the position of the trajectory point in the time period, 0 corresponds to the start of the time period, and 1 corresponds to the end of the time period. The relative time position of the trajectory point is added to the space-time 3D point set to obtain the final space-time 4D point. . Each time period The corresponding four-dimensional point set in space-time is denoted as , a total of M point sets.

[0048] The DBSCAN clustering algorithm is used on the spatiotemporal three-dimensional point set. For each overlapping time period The three-dimensional point set in space and time , DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm is used for clustering. The DBSCAN algorithm requires setting two key parameters: spatiotemporal neighborhood radius Eps and density threshold MinPts. The spatiotemporal neighborhood radius Eps includes the spatial neighborhood radius and the time neighborhood radius , respectively controlling the spatial distance and temporal distance thresholds between points in the clustering process. The value of is determined according to the spatial span of the surveying and mapping project and the scale of the regulatory scope, usually in the order of tens to hundreds of meters, such as . The value of is determined according to the time granularity of the surveying and mapping project, usually in the order of minutes to hours, such as The density threshold MinPts controls the minimum number of cluster points in the clustering process. It is set according to the density of the spatiotemporal trajectory points and is usually an integer between 3 and 10, such as MinPts=5. Divide into multiple density-connected clusters ,in is the total number of clusters in the current time period. There may also be a small number of unclustered discrete noise points in the clustering results, which are recorded as the noise point set .

[0049] Extract the spatiotemporal coordinate range of each cluster and traverse each time period Cluster within , extract the two-dimensional spatial coordinates and timestamps of all trajectory points in the cluster, and get the cluster The minimum bounding rectangle of : ;in, and Cluster The minimum and maximum coordinates on the X and Y axes, and Cluster The start and end timestamps of the . Represents a surveying and mapping geographic information project in the time period The space-time measurement range within is denoted as .

[0050] Combine the testing scopes in different time periods and Within the testing range Perform a union operation to obtain the complete two-dimensional time and space survey range R of the surveying and mapping project: ;in, ; Through the union operation, the scattered survey areas in different time periods are merged into a complete spatiotemporal trajectory surface of the surveying and mapping project. Repeat the clustering and merging process until all overlapping time periods are A subset of trajectory points within DBSCAN clustering is completed. For each time period , a series of testing ranges are obtained based on the clustering results Through the union operation, the survey range sequences of all time periods are merged to obtain the complete spatiotemporal two-dimensional survey range sequence of the surveying and mapping project. }, where Q is the total number of surveying and mapping projects.

[0051] Preferably, the temporal topological relationship of the surveying and mapping projects is analyzed, and the surveying range of any two surveying and mapping projects is and , extract its time range and . Calculate the overlap between two time ranges : ;in, It is a value in the interval [0, 1], indicating the degree of overlap between two surveying and mapping projects in the time dimension, 0 means no overlap at all, and 1 means complete overlap. Set the time overlap threshold , indicating that if the overlap of the surveying time of two surveying and mapping projects exceeds 80%, it is considered that there is a significant temporal topological relationship. Traverse all surveying and mapping projects and calculate the temporal overlap between each project. and with the threshold By comparison, we identify pairs of items (i, j) with significant temporal topological relationships.

[0052] Preferably, the spatial topological relationship of the surveying and mapping projects is analyzed, and the surveying range of any two surveying and mapping projects is and , extract its spatial extent and . Calculates the intersection area of ​​two spatial ranges : ;in, Indicates the intersection area of ​​two surveying and mapping projects in the spatial dimension, in square meters. Calculates the intersection ratio of two spatial ranges : ;in, and Surveying and mapping projects and The measurement area, It is a value in the interval [0, 1], indicating the degree of intersection of two surveying and mapping projects in the spatial dimension. 0 means no intersection at all, and 1 means that one project is completely covered by the other project. Set the spatial intersection ratio threshold , , indicating that if the intersection area of ​​the survey scope of two surveying and mapping projects exceeds 20% of the area of ​​the smaller project, it is considered that there is a significant spatial intersection topological relationship. Traverse all surveying and mapping projects and calculate the spatial intersection ratio between the two projects and with the threshold By comparison, the project pairs (i, j) with significant spatial intersection topological relationships can be identified. Similarly, the minimum distance between the boundaries of the survey scope of two surveying and mapping projects can also be calculated. , set the spatial adjacency distance threshold , identify pairs of items (i, j) with significant spatial adjacency topological relationships.

[0053] Optimize the spatiotemporal distribution of surveying and mapping projects, integrate the results of topological relationship analysis such as time overlap, spatial intersection, and spatial adjacency, and construct a spatiotemporal topological relationship network for surveying and mapping projects. In the spatiotemporal topological relationship network, each node represents a surveying and mapping project, and the edge between nodes represents the spatiotemporal topological relationship between the two projects. The weight of the edge can be used to represent the significance of the topological relationship. Perform community detection on the spatiotemporal topological relationship network to identify surveying and mapping project communities with dense spatiotemporal distribution and close topological relationships. For surveying and mapping projects in the same community, avoid excessive overlap of their survey scope in the time dimension by optimizing scheduling, such as implementing projects in the community in batches and staggered peaks. For surveying and mapping projects between different communities, arrange the spatial edge connection order of the survey reasonably according to their spatial adjacency relationship to achieve seamless connection and continuous coverage of surveying and mapping projects in the spatial dimension.

[0054] Realize the correlation matching of surveying and mapping project data, and further analyze the metadata attributes of different surveying and mapping projects based on the spatiotemporal topological relationship network, such as surveying and mapping scale, surveying and mapping accuracy, surveying and mapping content, etc. Set the attribute threshold of correlation matching, such as scale difference threshold , indicating that the scale level difference between the two surveying and mapping projects does not exceed 2 levels before they can be associated and matched. For the surveying and mapping project pairs (i, j) in the spatiotemporal neighborhood, their metadata attributes are judged one by one to see whether they meet the associated matching threshold, and the project pairs that can be associated and matched are identified. For the surveying and mapping projects that can be associated and matched, the land features within their surveying scope are analyzed, the identification information of the common land features is extracted, and the land feature mapping relationship between different projects is established. The land feature mapping relationship is used to realize the data fusion and integration of different surveying and mapping projects, such as integrating the geometry, attributes, and topological information of the same-name land features of different projects to generate a consistent land feature library. Based on the integrated land feature library, the quality inspection, change analysis, information mining and other applications of surveying and mapping results can be further carried out to realize the comprehensive management and utilization of surveying and mapping geographic information in the spatiotemporal dimension.

[0055] Figure 3 The exemplary flow chart of generating a quality supervision scope map according to some embodiments of the present application is as follows. The boundary of each cluster is used as the scope of the supervised surveying and mapping geographic information project, and coordinate conversion is performed to generate a quality supervision scope map; including: extracting the geographic coordinate parameters of the scope of the supervised surveying and mapping geographic information project, for each supervised surveying and mapping geographic information project, extracting the boundary coordinate sequence of its survey scope , where N is the number of boundary coordinate points. Extract geographic coordinate parameters from the metadata of the surveying and mapping project, including: Coordinate system type: such as WGS84 geocentric coordinate system, CGCS2000 national geodetic coordinate system, etc. Projection type: such as UTM projection, Gauss-Krüger projection, etc. Central meridian: such as 120°E, 123°E, etc. Origin latitude: such as 0° (equator), 40°N, etc. Geographic coordinate parameters are usually stored in EPSG encoding or OGC WKT format, such as "EPSG:4326" for WGS84 geographic coordinate system, "EPSG:32650" for WGS84 UTM 50N projection coordinate system. Convert geographic coordinates to a unified target coordinate system, select the target coordinate system of the quality supervision scope map, such as the Gauss-Krüger projection coordinate system under CGCS2000, the central meridian is 117°E, and the origin latitude is 0°. For each supervised surveying and mapping geographic information project, the boundary coordinate sequence , coordinate transformation is performed according to its geographic coordinate parameters and target coordinate system parameters.

[0056] In a specific embodiment of the present application, a target coordinate system is selected, and the quality supervision range map adopts the CGCS2000 national geodetic coordinate system as the spatial reference frame. Under the CGCS2000 framework, the Gauss-Krüger projection is selected as the plane coordinate system. According to the spatial position of the supervision range, the central meridian of the Gauss-Krüger projection is determined to be 117°E. The origin latitude of the Gauss-Krüger projection is the equator, that is, 0°. Extract the geographic coordinates of the supervised projects, and for each supervised surveying and mapping geographic information project, extract the geographic coordinate sequence of the boundary of its survey range. ,in: The geodetic latitude of the boundary coordinate point in degrees. The geodetic longitude of the boundary coordinate point in degrees. is the geodetic height of the boundary coordinate point in meters. N is the number of boundary coordinate points. Geographic coordinates are usually based on WGS84, CGCS2000 or other reference ellipsoids, and the ellipsoid parameters need to be obtained from the metadata of the surveying and mapping project.

[0057] Calculate auxiliary parameters for each boundary coordinate point According to the CGCS2000 ellipsoid parameters, the following auxiliary parameters are calculated: the major semi-axis of the ellipsoid a = 6378137m, the minor semi-axis of the ellipsoid b = 6356752.3141m. The first eccentricity of the ellipsoid . Radius of curvature of the yoke . , in radians. Latitude , in radians. , in meters, can be calculated based on the elevation anomaly model, or approximated to 0. The normal radius of curvature of the meridian, .

[0058] Apply the coordinate transformation formula to each boundary coordinate point , substitute into the given coordinate transformation formula to calculate its spatial rectangular coordinates : ; ; ;in, is the earth height anomaly, which can be calculated according to the elevation anomaly model, or approximated to 0. The angle unit in the formula is radian, , and The unit is meter. Iterate the conversion until all boundary coordinate points of the supervised surveying and mapping geographic information projects have completed the conversion of spatial rectangular coordinates. Finally, the spatial rectangular coordinate sequence of each project is obtained. .

[0059] When the ellipsoid parameters of the original coordinate system and the target coordinate system are different (such as WGS84 ellipsoid and CGCS2000 ellipsoid), it is necessary to perform seven-parameter space coordinate conversion first. Convert to geodetic coordinates in the target coordinate system Obtain the parameters of the original coordinate system and the target coordinate system, and determine the ellipsoid parameters of the original coordinate system (such as WGS84) and the target coordinate system (such as CGCS2000), including: the major semi-axis of the ellipsoid of the original coordinate system , the minor semiaxis , flattening Etc. Target coordinate system ellipsoid major semi-axis , the minor semiaxis , flattening Etc. Get seven parameters between the original coordinate system and the target coordinate system, including: translation parameters , in meters. Scale parameter m, dimensionless. Rotation parameter , in radians. The seven parameters are generally published by authoritative surveying and mapping departments, such as the seven parameters between CGCS2000 and WGS84 published by the China National Surveying and Mapping Product Quality Inspection and Testing Center.

[0060] Calculate the spatial rectangular coordinates in the original coordinate system, and extract the geodetic coordinates of the boundary points of the surveying range for each supervised surveying and mapping geographic information project , where N is the number of boundary points.

[0061] According to the ellipsoid parameters of the original coordinate system, the geodetic coordinates Convert to rectangular coordinates : Calculate the eccentricity of the ellipsoid Calculate the radius of curvature of the yoke . Calculate the spatial rectangular coordinates: ; ; ; Apply the seven-parameter space coordinate transformation formula, for each boundary point's space rectangular coordinate , substitute into the seven-parameter model, and obtain the spatial rectangular coordinates in the target coordinate system :

[0062] Construct the rotation matrix R: .

[0063] Compute the coordinate transformation: ;in, is the coordinate residual, which can usually be ignored. Convert the spatial rectangular coordinates to geodetic coordinates, and convert the spatial rectangular coordinates to geodetic coordinates according to the ellipsoid parameters of the target coordinate system. Convert to geodetic coordinates : Calculate the first eccentricity of the ellipsoid . Calculate the second eccentricity of the ellipsoid .

[0064] Calculate the naturalized latitude The initial value of iteration: ; Iterative calculation of naturalized latitude , until the given accuracy requirement is met: ; Calculate geodetic latitude . Calculate geodetic longitude Calculate the geodetic height .

[0065] Construct a standardized regulatory scope based on the geodetic coordinate sequence in the converted target coordinate system , construct the vector polygon scope of each supervised surveying and mapping geographic information project. Select a uniform vector polygon size, such as 100m×100m or 500m×500m, and generate a standardized rectangular supervision scope based on the actual survey scope size of each project. The rectangular boundary coordinates of the standardized supervision scope are ,in: , represents the minimum latitude of the regulatory scope, rounded down to an integer multiple of Δ. , represents the maximum latitude of the regulatory scope, rounded up to an integer multiple of Δ. and Similarly, ensure that the standardized rectangle covers the outer rectangle of the original survey scope. Δ is the latitude or longitude span of the standardized rectangle, such as 0.001° or 0.005°. For projects that exceed the size of the standardized rectangle, a regulatory scope composed of multiple standardized rectangles can be constructed based on their horizontal and vertical spans.

[0066] Generate a quality supervision scope map, and perform vector merging of the standardized supervision scopes of all supervised surveying and mapping geographic information projects in the target coordinate system to obtain the total quality supervision scope. The quality supervision scope map uses a coordinate reference system consistent with the target coordinate system, such as CGCS2000. In the quality supervision scope map, each standardized rectangle represents a supervised surveying and mapping project, and adjacent or overlapping rectangles represent projects that are close or intersecting in space. Use GIS software or tools to visualize the quality supervision scope map, generate a vector layer of the supervision scope, and overlay it with other geographic base maps to intuitively display the spatial distribution of surveying and mapping geographic information quality supervision.

[0067] Figure 4 It is an exemplary flow chart for generating standardized feature vectors according to some embodiments of the present application, and features are extracted and quantified for the attribute information of the surveying and mapping project metadata and the quality supervision scope map to form a standardized feature vector; wherein the attribute information of the quality supervision scope map includes the spatial distribution characteristics and density characteristics of the clusters and the division of the surveying and mapping time periods; it includes: extracting the spatial vector data of the surveying and mapping project scope, and extracting the spatial vector data corresponding to each surveying and mapping project scope from the quality supervision scope map, such as vector polygon elements. The spatial vector data should contain the geometric information and attribute information of the surveying and mapping project scope, such as project number, project name, surveying and mapping unit, surveying and mapping accuracy, surveying and mapping time, etc. Perform topological checking and error correction on the extracted spatial vector data to ensure the validity and integrity of the geometric objects.

[0068] Calculate the spatial overlap characteristics and use the vector overlay analysis method to overlay the vector polygons of each surveying and mapping project scope with the vector polygons of the corresponding supervision scope in the quality supervision scope map. Calculate the overlapping area A_overlap of the two vector polygons and the total area of ​​the surveying and mapping project scope . Calculate area ratio feature : , The value range of is [0, 1], which indicates the spatial overlap between the surveying and mapping project scope and the regulatory scope.

[0069] Calculate the text similarity features and extract the project name and surveying and mapping unit name from the surveying and mapping project metadata and quality supervision scope map attribute information. Calculate the edit distance between the surveying and mapping project name and the supervision record project name Common edit distance algorithms include Levenshtein distance, Jaro-Winkler distance, etc. Calculate the edit distance between the surveying and mapping unit name and the regulatory record surveying and mapping unit name , using the same edit distance algorithm.

[0070] according to and Calculate text similarity features : ,in, and They are the weight coefficients of the surveying and mapping project name and the surveying and mapping unit name, respectively, which are used to balance the contribution of the two parts of the edit distance to the comprehensive similarity. The value range of is [0, 1], which indicates the similarity between the surveying and mapping project and the regulatory record in terms of text attributes. Calculate the accuracy to meet the characteristics, extract the surveying and mapping accuracy information from the surveying and mapping project metadata, such as "1:500" and "1:1000" and other scale forms, and convert them into corresponding numerical levels, such as "500" and "1000". Extract the surveying and mapping accuracy level required by the regulatory requirements from the attribute information of the quality regulatory scope map, such as "500" and "1000". Calculate the accuracy level of the surveying and mapping project Accuracy level according to regulatory requirements The difference: , The value of may be positive, zero, or negative, indicating that the accuracy of the surveying and mapping project is higher, equal to, or lower than the regulatory requirements, respectively.

[0071] According to the accuracy level difference The calculation accuracy meets the characteristics :when hour, , indicating that the surveying and mapping project meets the regulatory accuracy requirements. hour, , indicating that the surveying and mapping project does not meet the regulatory accuracy requirements. Calculate the time feature and extract the surveying and mapping start and end time from the surveying and mapping project metadata and , and convert it into a timestamp format, such as Unix timestamp. Extract the start and end time of the supervision period from the attribute information of the quality supervision range map and , and convert it to timestamp format. Calculate the timestamp of the center point of the surveying and mapping project time period : , calculate the timestamp of the center point of the supervision time period : ; Calculate the timestamp difference between the time center point of the surveying and mapping project and the time center point of the supervision : ; Δt represents the degree of deviation between the implementation time of the surveying and mapping project and the supervision period, measured in seconds.

[0072] Normalized feature vector, area ratio feature , accuracy meets characteristics and time characteristics Perform minimum-maximum normalization: ; (No normalization required); ; For text similarity features Perform Z-score normalization: ;in, and They are The mean and standard deviation of . Get the standardized feature vector , which indicates the degree of match between the surveying and mapping project and the quality supervision scope in the four dimensions of space, text, accuracy and time.

[0073] S5, based on the obtained clusters, select multiple surveying and mapping projects and regulatory areas as training samples, use the obtained standardized feature vectors to annotate them, and annotate and match relevant or irrelevant category labels; combine the feature vectors of the training samples and the corresponding category labels into a training set; including: determining the number of training samples for each cluster, and for each cluster obtained by clustering, counting the number of surveying and mapping projects n and the number of regulatory areas m in the cluster. According to the preset sample selection ratio r, calculate the number of selected training samples k: k = min (n, m) * r; where r is the sample selection ratio, the value range is (0, 1], which is used to control the number of training samples. When n and m are very different, select the smaller one as the benchmark to ensure the balance of positive and negative samples. The value of the sample selection ratio r needs to be adjusted according to the actual situation. The following are some reference principles: When the amount of data is large, you can choose a smaller r value, such as 0.1 or 0.2, to reduce the number of training samples and improve the calculation efficiency. When the amount of data is small, you can choose a larger r value, such as 0.5 or 0.8, to increase the number of training samples and improve the generalization ability of the model. When the ratio of positive and negative samples is unbalanced, you can choose different r values ​​to sample positive and negative samples separately, such as taking r=0.8 for positive samples and r=0.2 for negative samples to balance the number of positive and negative samples. By reasonably setting the sample selection ratio r, you can strike a balance between the quantity and quality of training samples, ensuring the representativeness and diversity of samples while controlling the calculation cost and memory usage.

[0074] Select surveying and mapping projects as training samples, and use the K-means++ algorithm to select k surveying and mapping projects from each cluster as training samples. The steps of the K-means++ algorithm are as follows: Randomly select a surveying and mapping project as the first cluster center. For each remaining surveying and mapping project, calculate its shortest distance to the selected cluster center. According to probability Select the next cluster center, where The sum of the squares of the shortest distances between all remaining surveying and mapping projects and the selected clustering center. Repeat until k surveying and mapping projects are selected as training samples. The advantage of the K-means++ algorithm is that the clustering center is selected by probability, making the clustering center more dispersed and avoiding the problem of over-concentration of clustering centers. Selecting surveying and mapping projects that are far away from the selected clustering center as new clustering centers makes the selected training samples more representative and diverse. Compared with the traditional K-means algorithm, the K-means++ algorithm can better initialize the clustering center and improve the effect and stability of clustering. When implementing the K-means++ algorithm, the following details should be noted: When randomly selecting the first clustering center, uniform random sampling or weighted random sampling can be used to ensure the fairness of the selection. When calculating the distance between the surveying and mapping project and the clustering center, the Euclidean distance, Manhattan distance, cosine similarity and other metrics can be used, and the appropriate distance metric can be selected according to the characteristics of the feature vector. In order to improve the convergence speed of the algorithm, the maximum number of iterations or convergence threshold can be set, and the algorithm can be terminated early when the conditions are met. The training samples selected by the K-means++ algorithm can better cover the feature space of the surveying and mapping projects in each cluster, improving the quality and representativeness of the training data.

[0075] Matching supervision ranges constitute training samples. For each selected surveying and mapping project, the supervision range with the largest area ratio feature is selected from the corresponding cluster as the matching object to form a training sample. Area ratio feature It represents the spatial overlap between the surveying and mapping project scope and the supervision scope. The calculation formula is: ;in, The overlapping area between the surveying and mapping project scope and the supervision scope. The total area of ​​the surveying project scope. The value range of is [0, 1]. The closer to 1, the higher the spatial overlap, and the closer to 0, the lower the spatial overlap. For each mapping project, calculate its area ratio characteristics with all regulatory areas in the cluster, and select The largest regulatory scope is used as the matching object.

[0076] The matching process can be achieved by the following steps: For each mapping project, initialize the maximum area scale feature If it is 0, the matching regulatory scope is empty. Traverse all regulatory scopes in the cluster, and for each regulatory scope: calculate the overlapping area between the surveying and mapping project scope and the regulatory scope . Calculate area ratio feature .if , then update for , taking the current regulatory scope as the matching object. After traversing all regulatory scopes, the regulatory scope with the largest area ratio feature matching the surveying and mapping project is the matching object. The regulatory scope and surveying and mapping project selected through the matching process constitute a training sample, which represents the matching relationship between the surveying and mapping project and the regulatory scope under the condition of the highest spatial overlap.

[0077] Extract the feature vector of the training sample; for each training sample, extract the metadata of the corresponding surveying and mapping project and the attribute information of the matching supervision scope to construct a standardized feature vector. The standardized feature vector includes the following four features: area ratio feature : ;in, is the original area ratio feature, and For all training samples, The minimum and maximum values ​​of .

[0078] Through minimum-maximum normalization, Map to the interval [0, 1] to eliminate the dimension effect. Text similarity feature : ;in, is the original text similarity feature, and For all training samples, The mean and standard deviation of Transformed to a standard normal distribution with a mean of 0 and a variance of 1 to highlight relative differences. Accuracy meets features : ;in, The original accuracy meets the feature, with a value of 0 or 1, indicating whether the surveying and mapping project meets the surveying and mapping accuracy required by the regulator. The accuracy meets the feature without normalization, and the original value can be used directly.

[0079] Time characteristics : ; Among them, Δt is the original time feature, which indicates the time deviation between the implementation time of the surveying and mapping project and the supervision period. Δt is mapped to the interval [0, 1] through minimum-maximum normalization to eliminate the dimension effect. By extracting and normalizing the multidimensional features of the surveying and mapping project and the supervision scope, the standardized feature vector is obtained , which represents the feature representation of training samples in four dimensions: space, text, accuracy and time.

[0080] Determine the matching labels of the training samples, determine the matching labels of each training sample according to the preset supervision conditions, and form the training data with the standardized feature vector and the matching labels. The preset supervision conditions include: area ratio feature Meet threshold requirements, such as . Text similarity features Meet threshold requirements, such as . Accuracy meets features Satisfy the requirements, that is . Temporal proximity features Meet threshold requirements, such as days. When a training sample meets all regulatory conditions at the same time, it is marked as a positive sample with a matching label of 1; otherwise it is marked as a negative sample with a matching label of 0. Summarize the training data set and repeat until all clusters have completed the training sample selection and labeling. Summarize the training data of all training samples to obtain a complete training data set. The training data set includes the standardized feature vector and the corresponding matching label of each training sample. Divide the training set and validation set, and randomly divide the training data set into a training set and a validation set. The ratio can be set according to actual conditions, such as 7:3 or 8:2. The training set is used to train classification models that match surveying and mapping projects with regulatory scopes, such as logistic regression, support vector machines, decision trees, etc. The validation set is used to evaluate the performance of the trained classification model, such as accuracy, precision, recall, F1 score, etc.

[0081] S6, using the training set, trains a classification model based on a support vector machine; the input of the classification model is the standardized feature vector of the surveying and mapping project and the regulatory scope, and the output is the predicted matching result; including: selecting the support vector machine SVM algorithm as the machine learning binary classification algorithm, and building a support vector machine classification model based on the training samples in the training set. The standardized feature vector of the training sample is used as the input feature of the SVM, and the matching label of the training sample is used as the target output of the SVM. The standardized feature vector is , where d is the feature dimension, representing the multidimensional attribute information of the surveying and mapping project. The matching label is ,in Indicates that the surveying and mapping project matches the regulatory scope, Indicates that the surveying and mapping project does not match the regulatory scope. The standardized feature vector and matching label of the training sample are input into the SVM algorithm, and the optimal classification hyperplane and support vector of the SVM classification model are obtained by solving the following convex quadratic programming problem: ,in, is the normal vector of the classification hyperplane, is the bias term of the classification hyperplane, is the slack variable of the i-th training sample, Is the penalty coefficient, which is used to balance the classification error and model complexity, with the following constraints: ,in, is the feature mapping function, which takes the d-dimensional input feature vector Mapped to m-dimensional feature space; represents the matching label of the i-th training sample, is the total number of training samples.

[0082] By solving the above optimization problem, we can obtain the parameters of the optimal classification hyperplane and , forming the SVM classification decision function: ,in The d-dimensional normalized feature vector representing the surveying and mapping items to be classified; is a geographic spatial correlation measurement function that measures the proximity and overlap between the d-dimensional surveying and mapping project feature vector x and the p-dimensional regulatory scope feature vector r in geographic space. It can be obtained by using an appropriate spatial similarity calculation method based on the spatial information of the two, such as their geographic coordinates and boundary shapes. The larger the value of , the stronger the geographic spatial correlation between the two; is the weight coefficient of geographic spatial correlation, which controls the influence of geographic spatial correlation on classification decision and can be optimized and selected through methods such as cross-validation.

[0083] For the new surveying and mapping project feature vector x, by calculating the improved classification decision function The symbol of , and obtain its predicted matching label with a given regulatory scope: , then the predicted surveying and mapping project x matches the regulatory scope r; if , then the prediction of surveying and mapping project x does not match the regulatory scope r. By selecting the SVM algorithm as a binary classifier, introducing the geospatial correlation measurement function, and solving the convex quadratic programming problem, a classification model for matching surveying and mapping projects and regulatory scopes considering geospatial information can be constructed. This model not only utilizes the multidimensional attribute characteristics of surveying and mapping projects, but also integrates the correlation between surveying and mapping projects and regulatory scopes in geospatial space, thus improving the accuracy and interpretability of matching classification.

[0084] S7, input the new surveying and mapping project to be regulated and the scope of supervision into the classification model to obtain the prediction matching result. For the new surveying and mapping project to be regulated, extract its standardized feature vector , where d is the characteristic dimension, representing the multidimensional attribute information of the surveying and mapping project. For a given regulatory scope, extract its standardized feature vector r, where p is the characteristic dimension of the regulatory scope, representing the multidimensional attribute information of the regulatory scope. And the regulatory range feature vector r is input into the trained SVM classification decision function to calculate the predicted matching label. represents the predicted matching label, Indicates that the forecast surveying and mapping project matches the regulatory scope, Indicates that the predictive mapping project does not match the regulatory scope. and is the optimal classification hyperplane parameter obtained through training. is a feature mapping function that maps the feature vector of the surveying and mapping project to a high-dimensional feature space It is a geospatial correlation measurement function that measures the proximity and overlap between the surveying and mapping project and the regulatory scope in geospatial space. is the weight coefficient of geographic spatial correlation, which controls the influence of geographic spatial correlation on classification decision. According to the prediction matching results, the quality supervision of surveying and mapping geographic information is carried out. , indicating that the surveying and mapping project matches the supervision scope, indicating that the surveying and mapping project is consistent with the supervision scope in terms of attribute characteristics and geographic space, and can be included in the supervision scope for quality supervision. Conduct a quality assessment on the matched surveying and mapping projects to check whether the quality indicators such as the completeness, accuracy, and timeliness of their surveying and mapping results meet the supervision requirements. If the quality assessment results are qualified, the surveying and mapping project is considered to have passed the supervision, and its surveying and mapping results can be included in the geographic information database, and the supervision records will be updated. If the quality assessment results are unqualified, it is necessary to further analyze the causes of the quality problems, and require the surveying and mapping units to make rectifications and resubmit the surveying and mapping results until the quality requirements are met.

[0085] If the prediction matches the label , indicating that the surveying and mapping project does not match the regulatory scope, indicating that the surveying and mapping project does not match the regulatory scope in terms of attribute characteristics or geographic space, and may not need to be included in the current regulatory scope. Further analysis is conducted on the unmatched surveying and mapping projects to determine whether they belong to other regulatory scopes or are new regulatory objects. If they belong to other regulatory scopes, they will be transferred to the corresponding regulatory department for processing. If it is a new regulatory object, it is necessary to evaluate its importance and necessity for supervision, decide whether to include it in the regulatory scope, and update the regulatory plan and resource allocation accordingly. Through the analysis and disposal of the prediction matching results, accurate supervision of the quality of surveying and mapping geographic information can be achieved, and the efficiency and effectiveness of supervision can be improved. For surveying and mapping projects that match the regulatory scope, quality assessment and supervision are focused to ensure that the quality of surveying and mapping results meets the requirements. For surveying and mapping projects that do not match the regulatory scope, unnecessary waste of regulatory resources is avoided, while potential new regulatory objects are also discovered, and the regulatory scope and regulatory strategy are dynamically optimized.

Claims

1. A method for monitoring the quality of surveying and mapping geographic information based on CORS big data, characterized in that: include: Obtain multiple track point data from CORS big data, obtain the scope of supervised surveying and mapping geographic information projects, surveying and mapping unit information, and surveying and mapping project metadata; The metadata of surveying and mapping projects includes surveying and mapping time, surveying and mapping method and surveying and mapping accuracy; The DBSCAN clustering algorithm is used to cluster multiple trajectory point data. According to the spatial distribution and density characteristics of the trajectory points, the trajectory points are divided into multiple clusters, each cluster representing the survey scope of a surveying and mapping geographic information project; when clustering, the surveying and mapping time is divided into multiple preset equal-length time periods, and the trajectory points in the same time period are clustered into one category; The boundary of each cluster is taken as the scope of the supervised surveying and mapping geographic information project, and coordinate transformation is performed to generate a quality supervision scope map; Feature extraction and quantitative representation are performed on the attribute information of the surveying and mapping project metadata and the quality supervision scope map to form a standardized feature vector; the attribute information of the quality supervision scope map includes the spatial distribution characteristics and density characteristics of the clusters and the division of the surveying and mapping time period; Based on the obtained clusters, multiple surveying and mapping projects and regulatory areas are selected as training samples, and the obtained standardized feature vectors are used to annotate them, and the annotated matching related or irrelevant category labels are used; the feature vectors of the training samples and the corresponding category labels are combined into a training set; Using the training set, a classification model based on support vector machine is trained; the input of the classification model is the standardized feature vector of the surveying and mapping project and the regulatory scope, and the output is the predicted matching result; Input new surveying and mapping projects to be supervised and the scope of supervision into the classification model to obtain the prediction and matching results, and conduct surveying and mapping geographic information quality supervision based on the prediction and matching results; S6, using the training set, trains a classification model based on a support vector machine, including: The support vector machine (SVM) algorithm is selected as the machine learning binary classification algorithm, and a support vector machine classification model is constructed based on the training samples in the training set; the standardized feature vector of the training sample is used as the input feature of the SVM, and the matching label of the training sample is used as the target output of the SVM; The standardized feature vectors and matching labels of the training samples are input into the SVM algorithm, and the optimal classification hyperplane and support vector of the SVM classification model are obtained by solving the following convex quadratic programming problem: Constraints: y i (W T φ(x i )+b)≥1-ξ i ,x i ≥0,i=1,2,......,N Where W∈R m is the normal vector of the classification hyperplane, b∈R is the bias term of the classification hyperplane, ξ i ∈R is the slack variable of the i-th training sample, C∈R + is the penalty coefficient, which is used to balance the classification error and model complexity; φ: R d →R m is the feature mapping function, which takes the d-dimensional input feature vector x i Mapped to m-dimensional feature space; y i ∈{-1,1} represents the matching label of the i-th training sample, N∈N + is the total number of training samples; By solving the above optimization problem, we can obtain the parameter W of the optimal classification hyperplane. * ∈R m and b * ∈R, forming the SVM classification decision function: f(x)=sign(W T φ(x)+b*+λ×s(x,r)) Where x∈R d The d-dimensional normalized feature vector representing the surveying and mapping items to be classified; s:R d ×R p →[0,1] is the geographic spatial correlation measurement function. The larger the value of s(x,r), the stronger the geographic spatial correlation between the two. + is the weight coefficient of geographic spatial correlation, which controls the influence of geographic spatial correlation on classification decision; For a new mapping project feature vector x∈R d , by calculating the sign of the improved classification decision function f(x), we can get its relationship with the given regulatory range r∈R p Prediction matching label: if f(x)>0, the predicted mapping project x matches the regulatory scope r; if f(x)<0, the predicted mapping project x does not match the regulatory scope r.

2. The method for monitoring the quality of surveying and mapping geographic information based on CORS big data according to claim 1 is characterized in that: S2, using DBSCAN clustering algorithm to cluster multiple trajectory point data, including: The trajectory point data is divided according to the surveying and mapping time attribute. According to the preset time period length, a sliding window method is used to slide and draw multiple overlapping trajectory point subsets of equal length time periods on the surveying and mapping time axis with a certain time step; wherein the time period length and time step are set according to the time granularity and supervision frequency of the surveying and mapping project; Select a subset of trajectory points within each overlapping time period, extract the two-dimensional spatial coordinates and timestamp of each trajectory point, and form a three-dimensional point set in time and space; calculate the relative time position of each trajectory point in the corresponding time period through the timestamp of the trajectory point; The DBSCAN clustering algorithm is used on the spatiotemporal 3D point set. By setting the spatiotemporal neighborhood radius Eps and the density threshold MinPts, the spatiotemporal 3D point set is clustered into multiple density-connected clusters according to the density distribution of trajectory points in the spatiotemporal neighborhood. Extract the spatiotemporal coordinate range of all trajectory points in each cluster as the spatiotemporal survey range of a surveying and mapping geographic information project in the current overlapping time period; obtain the complete spatiotemporal two-dimensional survey range of the surveying and mapping project by performing a union operation on the spatiotemporal survey ranges in different overlapping time periods; Repeat the above steps until all the trajectory point subsets in the overlapping time periods are clustered. According to the clustering results of each overlapping time period, a complete spatiotemporal two-dimensional survey range sequence of the surveying and mapping geographic information project is generated through a union operation; Repeat the above steps until all trajectory point data are clustered to obtain the spatiotemporal two-dimensional survey range sequence of multiple surveying and mapping projects; optimize the spatiotemporal distribution and correlation matching of surveying and mapping projects by analyzing the spatiotemporal topological relationship between the survey ranges of different surveying and mapping projects.

3. The method for monitoring the quality of surveying and mapping geographic information based on CORS big data according to claim 2 is characterized in that: S3, coordinate conversion of the scope of the supervised surveying and mapping geographic information project to generate a quality supervision scope map, including: Extract the geographic coordinate parameters of the scope of the supervised surveying and mapping geographic information project, which include the coordinate system type, projection type, central meridian and origin latitude; According to the coordinate conversion formula or the seven-parameter space coordinate conversion formula, the geographic coordinate parameters are converted into coordinate values ​​in a unified target coordinate system; Construct vector polygons based on the converted coordinate values ​​to generate standardized regulatory ranges for each surveying and mapping project, and the standardized regulatory ranges have the same size; The standardized supervision scopes of all projects are vector-merged to generate a quality supervision scope map, which represents the spatial scope of quality supervision of surveying and mapping geographic information.

4. The method for monitoring the quality of surveying and mapping geographic information based on CORS big data according to claim 3 is characterized in that: The coordinate transformation formula is as follows: Among them, X, Y, and Z represent the coordinate values ​​of the ground point in the geocentric rectangular coordinate system centered at the origin of the coordinate system; represents the geodetic latitude; λ represents the geodetic longitude; H represents the geodetic height; ΔH represents the difference in geodetic height; ν represents the radius of curvature of the ellipsoid; e represents the first eccentricity of the ellipsoid; ΔN represents the correction in the direction of the radius of curvature of the ellipsoid.

5. The method for monitoring the quality of surveying and mapping geographic information based on CORS big data according to claim 4 is characterized in that: The seven-parameter space coordinate transformation formula is as follows: Among them, X1, Y1, Z1 represent the spatial rectangular coordinate values ​​in the source coordinate system; X2, Y2, Z2 represent the spatial rectangular coordinate values ​​in the target coordinate system; ΔX, ΔY, ΔZ represent the translation of the coordinate origin; m represents the scale factor; ε X ,ε Y ,ε Z They represent the rotation angles of the three coordinate axes respectively; δX1, δY1, δZ1 represent the coordinate corrections in the source coordinate system; δx, δy, δz represent the coordinate corrections in the target coordinate system.

6. The method for monitoring the quality of surveying and mapping geographic information based on CORS big data according to any one of claims 1 to 5, characterized in that: S4. Form a standardized feature vector, including: Extract the spatial vector data of each surveying and mapping project scope in the quality supervision scope map, use vector overlay analysis to calculate the overlapping area of ​​each surveying and mapping project scope and the corresponding supervision scope, and obtain the area ratio characteristics reflecting the spatial overlap; Calculate the edit distance d between the surveying and mapping project name and the project name recorded in the quality supervision scope map attribute information p , and the edit distance d between the surveying and mapping unit name and the surveying and mapping unit name in the regulatory record u , according to d p and d u Calculate text similarity features s t ; Extract the surveying and mapping accuracy in the metadata of the surveying and mapping project and convert it into the corresponding numerical level; extract the regulatory requirement surveying and mapping accuracy level in the attribute information of the quality supervision scope map, calculate the difference between the surveying and mapping project accuracy level and the regulatory requirement accuracy level, and obtain the accuracy compliance feature that reflects whether the surveying and mapping project meets the regulatory accuracy requirements; Extract the surveying and mapping start and end time from the surveying and mapping project metadata and convert it into a timestamp format; extract the supervision time period start and end time from the quality supervision scope map attribute information and convert it into a timestamp format, calculate the timestamp difference between the center points of the two time periods, and obtain the time characteristics that reflect the degree of proximity between the surveying and mapping project implementation time and the supervision period; The area ratio feature, precision compliance feature and time feature are normalized by minimum-maximum normalization; the text similarity feature is normalized by Z-score normalization; and a standardized feature vector is obtained.

7. The method for monitoring the quality of surveying and mapping geographic information based on CORS big data according to claim 6 is characterized in that: Calculate text similarity features s t : Among them, w p and w u They are the weight coefficients of the surveying and mapping project name and the surveying and mapping unit name respectively.

8. The method for monitoring the quality of surveying and mapping geographic information based on CORS big data according to claim 7 is characterized in that: S5, constitutes the training set, including: For each cluster obtained by clustering, the number of surveying and mapping projects n and the number of regulatory areas m in the cluster are counted, and the number of selected training samples k = min (n, m) × r is calculated according to the preset sample selection ratio r; The K-means++ algorithm is used to select k surveying and mapping projects from each cluster as training samples; For each selected mapping project, one regulatory range with the largest area ratio characteristic in the corresponding cluster is selected as the matching regulatory range to form a training sample; Extract the mapping project metadata and regulatory scope attribute information of each training sample and calculate the corresponding standardized feature vector; According to the preset supervision conditions, the matching label of each training sample is determined, and the standardized feature vector and the matching label are combined into training data, the positive sample label is 1, and the negative sample label is 0; the preset supervision conditions include that the area ratio feature, the text similarity feature, the accuracy conformity feature, and the time proximity feature meet the set threshold requirements respectively; Repeat until all clusters have completed the selection and labeling of training samples, and aggregate the training data of all training samples to obtain a complete training data set; The training data set is randomly divided into training set and validation set, which are used to train and evaluate the classification model that matches the surveying and mapping projects with the regulatory scope, so as to realize the intelligent supervision of the quality of surveying and mapping geographic information.

Images