An Automatic Extraction Method for Deformation Points of Railway Facilities Based on Multi-Attribute Clustering

Abstract

This invention discloses an automatic extraction method for deformation points of railway facilities based on multi-attribute clustering. The method includes: acquiring time-series SAR images along the railway line in the test area, performing time-series InSAR processing to obtain deformation point locations of the area and railway facilities; performing cluster analysis on the deformation points based on their multi-attribute information to obtain deformation point clusters of railway facilities; performing correlation analysis on the deformation points within each cluster, sorting the points according to their correlation level to obtain points with high correlation; performing correlation analysis using the mean of the top n highly correlated points and the top n+1 highly correlated points, and for each cluster satisfying the correlation requirement, selecting the point with the highest correlation as the representative of that cluster to obtain the individual deformation point and time-series deformation result of the corresponding railway facility. This invention offers high extraction accuracy, enabling automatic extraction of deformation points for railway facilities and improving the problem of multiple deformation points corresponding to a single facility.

Description

Technical Field

[0001] This invention relates to the field of SAR processing and interpretation technology for railway facility deformation, and more specifically to an automatic extraction method for railway facility deformation points based on multi-attribute clustering. Background Technology

[0002] Railway facility deformation monitoring is a crucial aspect of ensuring safe railway operation. Currently, railway facility deformation monitoring is mainly carried out through a combination of manual measurement and GNSS automatic monitoring. As the operating mileage of railways increases year by year, the workload and difficulty of manual measurement also increase annually. Furthermore, the deformation patterns obtained by manual measurement combined with GNSS automatic monitoring are mostly point-based, resulting in a long measurement cycle for facilities along the same line. There is room for improvement in terms of result consistency and obtaining facility deformation information over a wider range.

[0003] Railway facility deformation monitoring based on temporal SAR imagery consists of two stages: data processing and interpretation. Data processing uses InSAR technology to obtain deformation information within the region, while interpretation maps deformation points within the region to railway facilities to obtain temporal deformation information. Current work focuses primarily on the data processing stage, with limited research on interpretation. The fact that a single railway facility corresponds to multiple deformation points impacts the analysis of railway facility deformation.

[0004] Therefore, how to provide an automatic method for extracting deformation points of railway facilities with a wide range, high precision, strong consistency and interpretability is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] In view of this, the present invention provides an automatic extraction method for deformation points of railway facilities based on multi-attribute clustering, which has high extraction accuracy and can effectively improve the problem of a single facility corresponding to multiple deformation points.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] An automatic extraction method for deformation points of railway facilities based on multi-attribute clustering includes:

[0008] S1. Acquire SAR images along the railway line in the test area, perform InSAR processing, and obtain the deformation points of facilities within the area;

[0009] S2. Based on the multi-attribute information of the deformation points, cluster analysis is performed on the deformation points to obtain the deformation point clusters of railway facilities;

[0010] S3. Perform correlation analysis on the deformation points within the deformation point cluster of the facility. Sort the deformation points within the cluster according to their correlation level to obtain the points with higher correlation.

[0011] S4. Use the mean of the first n points with high correlation and the first n+1 points to perform correlation analysis. For point clusters that meet the correlation, select the point with the highest correlation in the point cluster as the representative of the point cluster to obtain the individual deformation point and deformation result of the corresponding railway facility.

[0012] Furthermore, in the above-mentioned method for automatic extraction of deformation points of railway facilities based on multi-attribute clustering, S1 includes:

[0013] S11. Collect SAR images of the railway in the test area that meet the preset spatial resolution according to the monitoring plan;

[0014] S12. The SAR data is processed using the temporal InSAR method. The processing flow includes image registration, differential interferometry, phase unwrapping, deformation calculation, geocoding, spatiotemporal filtering, etc., to obtain the deformation results of the study area and facilities.

[0015] S13. Geocode and project the obtained deformation results to give them geographic projection coordinate information; the deformation results include deformation rate and cumulative deformation information for each period; the final results are output in SHP file format.

[0016] Furthermore, in the above-mentioned method for automatic extraction of deformation points of railway facilities based on multi-attribute clustering, in S2, the multi-attribute information includes the elevation of the deformation point, the plane coordinates, and the plane coordinates are obtained by projection transformation of latitude and longitude and are normalized; the elevation is obtained by overlaying DEM and SAR images under the same coordinate system according to geocoding and is normalized.

[0017] Furthermore, in the above-mentioned method for automatically extracting deformation points of railway facilities based on multi-attribute clustering, in S2, the clustering analysis is a multi-attribute clustering analysis algorithm, and the calculation formula used is:

[0018]

[0019] in Let x be the normalized abscissa of the deformation point to be clustered. The normalized ordinate of the point in the plane. This is the normalized elevation of the point; The normalized abscissa of the cluster center point. Its normalized planar ordinate, Its normalized elevation; is the Euclidean distance between the point to be clustered and the center point of the cluster.

[0020] Furthermore, in the above-mentioned method for automatically extracting deformation points of railway facilities based on multi-attribute clustering, in S2, the clustering analysis requires specifying the distance threshold between the cluster deformation center point and other deformation points, as well as the number of the fewest points within the cluster. The distance is Euclidean distance, and the number of the fewest points within the cluster is obtained based on the statistical law of the number of deformation points on railway facilities.

[0021] Furthermore, in the above-mentioned method for automatic extraction of railway facility deformation points based on multi-attribute clustering, in S3, the correlation analysis of deformation points within a cluster involves taking the mean of the time-series deformation data of all deformation points within the same cluster to obtain the average deformation sequence of the cluster. Then, the correlation coefficient between the time-series deformation data of each deformation point within the cluster and the average deformation sequence is calculated to obtain the correlation set of each point within the cluster with the mean sequence. The sorting of deformation points according to their correlation level involves sorting the correlation sets of deformation points within the cluster from high to low to obtain a set of points with higher correlation for subsequent screening of facility deformation points.

[0022] Furthermore, in the above-mentioned method for automatically extracting deformation points of railway facilities based on multi-attribute clustering, in S4, the correlation analysis of the mean values ​​of the first n points and the first n+1 points refers to taking the mean value sequence of the time-series deformation data of the first n points and the mean value sequence of the time-series deformation data of the first n+1 points of the set of points with high correlation, and calculating the correlation coefficient between the two mean value sequences; the cluster of points that satisfies the correlation refers to the cluster of points where the correlation coefficient between the two mean value sequences is greater than the set correlation threshold.

[0023] Furthermore, in the above-mentioned method for automatically extracting deformation points of railway facilities based on multi-attribute clustering, S4 also includes: n is not greater than the minimum number of samples within the cluster.

[0024] As can be seen from the above technical solution, compared with the prior art, the present invention discloses an automatic extraction method for deformation points of railway facilities based on multi-attribute clustering, which has the following beneficial effects:

[0025] This invention extracts deformation points from remote sensing images along railway lines, enabling the automatic acquisition of deformation information within the region and improving the automation and efficiency of deformation information acquisition within the region.

[0026] This invention performs multi-attribute clustering on deformation points within railway facilities. By adding attribute information, the clustering of point clusters can be made more accurate, while effectively reducing the clustering of unrelated deformation points, thus accurately obtaining the deformation point locations corresponding to railway facilities.

[0027] This invention utilizes an automated method to obtain a cluster of deformation points for railway facilities. By filtering, it identifies a unique deformation point corresponding to each railway facility, thus improving the phenomenon of a single facility having multiple deformation points. Simultaneously, it obtains deformation information of railway facilities, enhancing the interpretability of railway facility deformation. Attached Figure Description

[0028] Figure 1 A flowchart of the automatic extraction method for deformation points of railway facilities provided by the present invention;

[0029] Figure 2 and Figure 3 These are all schematic diagrams showing the results of deformation points of railway facilities provided by this invention. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] like Figure 1 As shown in the figure, this invention discloses an automatic extraction method for deformation points of railway facilities based on multi-attribute clustering, including:

[0032] S1. Acquire SAR images along the railway line in the test area, perform InSAR processing, and obtain the deformation points of the area and facilities;

[0033] S2. Based on the plane coordinates and elevation of the deformation points, perform cluster analysis on the deformation points to obtain the deformation point clusters of railway facilities;

[0034] S3. Perform correlation analysis on the deformation points within the deformation point cluster of the facility. Sort the deformation points within the cluster according to their correlation level to obtain the points with higher correlation.

[0035] S4. Correlation analysis is performed using the mean of the top n points and the top n+1 points with high correlation. For point clusters that satisfy the correlation criteria, the point with the highest correlation within the cluster is selected as the representative of that cluster, yielding the individual deformation point and deformation result of the corresponding railway facility. For example... Figure 2 and Figure 3 The image shown is a schematic diagram illustrating the results of automatic extraction of deformation points for railway facilities.

[0036] The process of automatically extracting deformation points of railway facilities in this embodiment of the invention is as follows:

[0037] Using InSAR technology, deformation information in the railway line area is automatically extracted. Cluster analysis is performed based on the plane coordinates and elevation of the deformation information to obtain deformation point clusters of railway facilities. Correlation analysis is then performed on the deformation point clusters of railway facilities to obtain deformation points with high correlation. Correlation analysis is also performed by analyzing the mean sequence of time-series deformation data of the first n points and the first n+1 points. For point clusters that meet the requirements, the point with the highest correlation is selected as the representative of the point cluster, thus obtaining the deformation points and deformation information corresponding to the railway facilities.

[0038] The steps described above will be explained in further detail below.

[0039] S1: Acquire SAR images along the railway line in the test area, perform InSAR processing, and obtain the deformation points of facilities within the area.

[0040] S11. Railway SAR remote sensing images of the test area that meet the preset spatial resolution according to the monitoring plan; the acquisition time is determined according to the satellite shooting cycle and the location of the railway line, and the acquisition cycle of the two phases shall not exceed one month. The spatial resolution should be better than 3m to ensure that the structural features of railway bridges, roadbeds and other facilities can be clearly distinguished.

[0041] S12. The SAR data is processed using the temporal InSAR method. The processing flow includes image registration, differential interferometry, phase unwrapping, deformation calculation, geocoding, spatiotemporal filtering, etc., to obtain the deformation results of the study area and facilities.

[0042] In one embodiment, S2, the multi-attribute information of the deformation point includes:

[0043] S21. Project and transform the latitude and longitude of the deformation points to obtain their coordinates in the WGS84 coordinate system; unify the DEM and SAR images to the same coordinate system and overlay them to obtain the elevation information of each deformation point.

[0044] S22. Normalize the plane coordinates and elevations of the deformation points to obtain plane coordinates and elevations under the same dimensions. The normalization formula is:

[0045]

[0046] in A value representing a specific attribute of a deformation point (such as the horizontal coordinate, vertical coordinate, or elevation). This is the mean of the plane coordinates or elevation of the deformation point. This represents the standard deviation of the plane coordinates or elevation of the deformation point.

[0047] S23. Perform cluster analysis based on the normalized plane coordinates and elevations of the deformation points. The clustering method is DBSCAN, and the formula is:

[0048]

[0049] Where num represents the minimum number of points in each cluster (i.e., the min_samples parameter in DBSCAN), which can be determined experimentally based on the typical size of the target railway facility and the spatial density of deformation points. For example, in a preferred embodiment, this value is set to 3; exp (i.e., the eps parameter in DBSCAN) represents the distance threshold for judging the neighborhood during clustering, A represents the normalized plane coordinates and elevation combination of the deformation points, and d is the clustering result of the point cluster. The distance formula is as follows:

[0050]

[0051] Where x is the normalized abscissa of the deformation point to be clustered, y is the normalized ordinate of the point, h is the normalized elevation of the point; xi is the normalized abscissa of the cluster center point, yi is its normalized ordinate, hi is its normalized elevation; p is the Euclidean distance between the deformation point to be clustered and the cluster center point.

[0052] In one embodiment, S3 includes the following:

[0053] S31. Take the average value of the temporal deformation information of points within a certain point cluster, for a total of 25 periods;

[0054]

[0055] Where m represents the number of deformation points within the cluster, and z represents the deformation information of each deformation point in each period, such as... This represents the deformation information of the third point in a certain cluster during phase 2.

[0056] S32. Correlation analysis is performed between each deformation point within the cluster and the mean. The Pearson method is used for correlation analysis, and the formula is:

[0057]

[0058] in This provides temporal deformation information for each deformation point within the point cluster. The mean of temporal deformation information of points within the cluster.

[0059] The correlation coefficients between the deformation information and the mean of each point within the cluster are sorted according to their relative values.

[0060] In one embodiment, in S4, the selection of a single deformation result for the railway facility includes:

[0061] S41. Obtain the time series information of the top 2 and top 3 deformation points with high correlation coefficients with the mean within each point cluster, for a total of 25 periods, and obtain the mean of the top 2 and top 3 deformation information.

[0062]

[0063] Where n represents the number of deformation points, and z represents the deformation information of each deformation point in each period, such as This represents the deformation information of the third point in a certain cluster during phase 2.

[0064] S42. Correlation analysis is performed on the deformation information of deformation points within a cluster. Pearson correlation analysis is used. For clusters with a correlation greater than 0.9, the point with the highest correlation coefficient between the time-series deformation information and the mean is selected as the representative of the cluster, and the individual deformation points and deformation information corresponding to the railway facilities are obtained.

[0065] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0066] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An automatic extraction method for deformation points of railway facilities based on multi-attribute clustering, characterized in that, include: S1. Acquire SAR images along the railway line in the test area, perform InSAR processing, and obtain the deformation points of the area and facilities; S2. Based on the multi-attribute information of the deformation points, cluster analysis is performed on the deformation points to obtain the deformation point clusters of railway facilities; S3. Perform correlation analysis on the deformation points within the deformation point cluster of the facility. Sort the deformation points within the cluster according to their correlation level to obtain the points with higher correlation. S4. Use the mean of the first n points with high correlation and the first n+1 points to perform correlation analysis. For point clusters that meet the correlation, select the point with the highest correlation in the point cluster as the representative of the point cluster to obtain the individual deformation point and deformation result of the corresponding railway facility.

2. The method for automatically extracting deformation points of railway facilities based on multi-attribute clustering according to claim 1, characterized in that, S1 includes: S11. Collect SAR images of the railway in the test area that meet the preset spatial resolution according to the monitoring plan; S12. The SAR data is processed using the temporal InSAR method. The processing flow includes image registration, differential interferometry, phase unwrapping, deformation calculation, geocoding, and spatiotemporal filtering to obtain the deformation results of the study area and facilities. S13. Geocode and project the obtained deformation results to give them geographic projection coordinate information; the deformation results include deformation rate and cumulative deformation information for each period; the final results are output in SHP file format.

3. The method for automatically extracting deformation points of railway facilities based on multi-attribute clustering according to claim 1, characterized in that, In S2, the multi-attribute information includes the elevation of the deformation point, the plane coordinates, and the plane coordinates are obtained by projection transformation of latitude and longitude and are normalized; the elevation is obtained by overlaying the digital elevation model (DEM) and SAR image under the same coordinate system according to the geocoding and is normalized. The cluster analysis requires specifying the neighborhood radius (eps) and the minimum number of samples within the cluster (min_samples) threshold.

4. The method for automatically extracting deformation points of railway facilities based on multi-attribute clustering according to claim 1, characterized in that, The clustering analysis is a multi-attribute clustering analysis algorithm, and the calculation formula used is: ； in Let x be the normalized abscissa of the deformation point to be clustered. The normalized ordinate of the point in the plane. This is the normalized elevation of the point; The normalized abscissa of the cluster center point. Let its normalized planar ordinate be... Its normalized elevation; is the Euclidean distance between the point to be clustered and the center point of the cluster.

5. The method for automatically extracting deformation points of railway facilities based on multi-attribute clustering according to claim 1, characterized in that, The clustering analysis requires specifying the neighborhood radius (eps) and the minimum number of samples within a cluster (min_samples) threshold. The distance is Euclidean distance, and the minimum number of samples within a cluster is determined based on the spatial distribution statistical law of deformation points on railway facilities.

6. The method for automatically extracting deformation points of railway facilities based on multi-attribute clustering according to claim 1, characterized in that, The correlation analysis of deformation points within the aforementioned facility deformation point cluster involves taking the mean of the time-series deformation data of all deformation points within the same cluster to obtain the average deformation sequence of the cluster. Then, the Pearson correlation coefficient method is used to calculate the correlation coefficient between the time-series deformation data of each deformation point within the cluster and the average deformation sequence, thus obtaining the correlation set between each point within the cluster and the mean sequence. The deformation points are sorted according to their correlation, which means sorting the set of deformation points within the point cluster from high to low correlation, and obtaining the set of points with high correlation for subsequent screening of facility deformation points.

7. The method for automatically extracting deformation points of railway facilities based on multi-attribute clustering according to claim 1, characterized in that, The correlation analysis between the mean values ​​of the first n points and the first n+1 points refers to taking the mean value sequence of the time-series deformation data of the first n points of the set of points with high correlation obtained in claim 6, and calculating the correlation coefficient between the mean value sequence of the time-series deformation data of the first n+1 points; the cluster of points that satisfies the correlation refers to the cluster of points where the correlation coefficient between the two mean values ​​is greater than the set correlation threshold.

8. The method for automatically extracting deformation points of railway facilities based on multi-attribute clustering according to claim 5, characterized in that, The number n is not greater than the minimum number of samples within the cluster.

Images