A highway slope space-air-ground integrated monitoring system

Abstract

The present application relates to the field of space-time-ground data processing technology, in particular to a highway slope space-time-ground integrated monitoring system. According to the ground side data distribution of different sensors and the relative distance between each unit center position and the sensor, the cluster space deformation coherence coefficient is obtained; based on the deviation characteristics of the cluster space deformation coherence coefficient before and after the simulation fusion of adjacent ground side cluster nodes, the ground side composite monitoring cluster node is generated; according to the number of ground side cluster nodes in the ground side composite monitoring cluster node and the required cluster number, the candidate sky side area node is screened out; the time overlap ratio between the abnormal duration window of the ground side composite monitoring cluster node and the observation time window of the candidate sky side area node is obtained, the sky side and ground side data source with the lowest space-time mismatch degree is extracted, and the highway slope is monitored in a space-time-ground integrated manner. The present application accurately constructs the graph topology edge weight mechanism of fusion mismatch degree, and improves the effectiveness of integrated monitoring.

Description

Technical Field

[0001] This invention relates to the field of air-space-ground data processing technology, specifically to an integrated air-space-ground monitoring system for highway slopes. Background Technology

[0002] As a typical scenario of interaction between artificial structures and natural rock and soil, the stability of highway slopes is directly related to traffic operation safety. In existing technologies, integrated air-space-ground monitoring technology based on the fusion of satellite InSAR, UAV aerial surveys, and ground sensors is used for slope safety hazard investigation and pre-disaster collaborative early warning. However, the integrated monitoring system exposes the problem of heterogeneous data mismatch in open engineering environments. The air-space-ground monitoring data are naturally mismatched in terms of spatial representation area and temporal sampling frequency, and are easily affected by false alarm noise interference caused by isolated hardware failures. The system has difficulty in effectively linking macroscopic regional anomalies to microscopic local sensors, and often leads to collaborative misjudgment due to multi-source data mismatch, causing the monitoring system's collaborative early warning and accurate diagnosis capabilities to fail. Summary of the Invention

[0003] To address the technical problem of mismatched multi-source data (air, space, and ground) leading to misjudgments and thus the failure of collaborative early warning and accurate diagnosis capabilities in monitoring systems, this invention aims to provide an integrated air, space, and ground monitoring system for highway slopes. The specific technical solution adopted is as follows: This invention proposes an integrated air-space-ground monitoring system for highway slopes, the system comprising: The data acquisition module is used to acquire multi-source monitoring data of highway slopes from air, space, and ground. Based on the airside data, a three-dimensional baseline map of the slope is constructed, and the airside remote sensing data and groundside sensor data are mapped to the baseline map for spatial alignment. The deformation coherence analysis module is used to obtain multiple three-dimensional solid engineering meshes of highway slopes. The sensor set in each mesh is regarded as a ground-side cluster node, and the corresponding sky-side remote sensing grid is regarded as a sky-side regional node. Multiple cells of any three-dimensional solid engineering mesh are obtained. Based on the distribution of ground-side data from different sensors and the relative distance between the center of each cell and the sensor, the cluster spatial deformation coherence coefficient is obtained. The topology edge weight calculation module is used to perform adaptive grid merging based on the deviation characteristics of the cluster spatial deformation coherence coefficient before and after the simulation fusion of adjacent ground-side cluster nodes, generating ground-side composite monitoring cluster nodes; obtaining the required number of remotely observable clusters based on the area ratio relationship between the sky-side remote sensing grid and the 3D solid engineering grid; selecting candidate sky-side regional nodes based on the number of ground-side cluster nodes in the ground-side composite monitoring cluster nodes and the required number of clusters; obtaining the anomaly duration window of the ground-side composite monitoring cluster node and the time overlap ratio between the observation time windows of the candidate sky-side regional nodes as the graph topology edge weights between nodes. The integrated monitoring module is used to extract the sky-side and ground-side data sources with the lowest spatiotemporal mismatch based on the maximum edge weight matching relationship, and to carry out integrated air-space-ground monitoring of highway slopes.

[0004] Furthermore, the method for obtaining the cluster spatial deformation coherence coefficient includes: Based on the distribution of ground-side data from different sensors, the reliability factor of each sensor is obtained; Based on the confidence factor of each sensor and the relative distance between the center position of each unit and the sensor, the adaptive interpolation weight between the center position of each unit and each sensor is obtained. Based on the adaptive interpolation weights between the center position of each unit and different sensors, as well as the ground-side data of the sensors, the spatial deformation characteristic values ​​of each unit are obtained, and the cluster spatial deformation coherence coefficient is obtained.

[0005] Furthermore, the method for obtaining the reliability factor includes: The ground-side data of the sensors is normalized to obtain the difference between the normalized ground-side data of each sensor between each time series and the next time series, which is used as the residual value of each sensor in each time series. The mean of the product of the residuals between different sensors across all time series is obtained, and the absolute value of the mean of the product is used as the correlation of the monitoring data between different sensors. The mean value of the correlation between each sensor and the monitoring data of other sensors is obtained and normalized, and used as the reliability factor for each sensor.

[0006] Furthermore, the method for obtaining the adaptive interpolation weights includes: The confidence factor of each sensor is negatively correlated and used as the first distance coefficient; The product of the first distance coefficient and the relative distance between the center position of each unit and each sensor is obtained as the equivalent spatial distance between the center position of each unit and each sensor. A negative correlation normalization mapping is performed on the center position of each unit and the equivalent spatial distance between different sensors, which serves as the adaptive interpolation weight between the center position of each unit and each sensor.

[0007] Furthermore, the method for obtaining the spatial deformation feature values ​​includes: The spatial deformation characteristic value of each unit is determined by the adaptive interpolation weights between the center position of each unit and different sensors, as well as the multiplication and summation of the normalized ground-side data of the sensors.

[0008] Furthermore, the method for obtaining the cluster spatial deformation coherence coefficient includes: For any three-dimensional solid engineering mesh, obtain the mean value of the spatial deformation characteristic value of all elements as the average deformation level; obtain the fluctuation characteristics of the spatial deformation characteristic value of all elements as the deformation fluctuation degree. Based on the proportional relationship between deformation volatility and average deformation level, the spatial variation coefficient is obtained, and the deformation volatility is positively correlated with the spatial variation coefficient. A negative correlation mapping is performed on the spatial variation coefficient, and the product between the negative correlation mapping result and the average deformation level is obtained as the cluster spatial deformation coherence coefficient.

[0009] Furthermore, the method for acquiring the ground-side composite monitoring cluster nodes includes: Obtain the absolute difference between the cluster spatial deformation coherence coefficients of each ground-side cluster node and its adjacent ground-side cluster nodes before and after the simulated fusion. Calculate the ratio of the absolute difference to the cluster spatial deformation coherence coefficients of each ground-side cluster node before the simulated fusion, and use this as the relative deviation change rate. If the relative deviation change rate corresponding to each ground-side cluster node is less than or equal to the preset change threshold, the grid corresponding to each ground-side cluster node and the adjacent ground-side cluster node are merged, and the sensor set in the merged grid is taken as the ground-side composite monitoring cluster node; if there is a ground-side cluster node whose relative deviation change rate is greater than the preset change threshold, merging is not performed, and the ground-side cluster node is taken as the ground-side composite monitoring cluster node.

[0010] Furthermore, the method for obtaining the required number of clusters includes: The ratio of the equivalent pixel area of ​​the surface of the sky-side remote sensing grid to the actual projected area of ​​the three-dimensional solid engineering grid is obtained as the first area ratio. The product of the first area ratio and the preset satellite phase perturbation sensitivity coefficient is obtained as the required number of clusters for remote sensing observation.

[0011] Furthermore, the method for obtaining the candidate celestial region nodes includes: If the number of ground-side cluster nodes in the ground-side composite monitoring cluster node is greater than or equal to the required number of cluster nodes, all sky-side regional nodes corresponding to the grid where the ground-side composite monitoring cluster node is located will be merged as candidate sky-side regional nodes.

[0012] Furthermore, the method for obtaining the time overlap ratio includes: The time when the ground-side data of the earliest sensor in the ground-side composite monitoring cluster node exceeds the preset warning threshold is taken as the start timestamp, and the time when the ground-side data of the last sensor exceeds the preset warning threshold is taken as the end timestamp, thus forming an abnormal duration window; the time when the satellite main radar image corresponding to the candidate sky-side regional node is captured is taken as the start time and the time when the secondary image is captured is taken as the end time, thus forming an observation time window; Obtain the overlap duration between the anomaly duration window and the observation time window, and calculate the ratio of the overlap duration to the total duration within the anomaly duration window as the time overlap ratio.

[0013] The present invention has the following beneficial effects: This invention obtains multiple three-dimensional solid engineering meshes of highway slopes, treats the sensor set within each mesh as a ground-side cluster node, and the corresponding sky-side remote sensing grid as a sky-side regional node. Through solid mesh partitioning and graph topology node mapping, microscopic discrete sensor point sources are integrated into a macroscopic surface observation cluster, eliminating the spatial scale mismatch barrier of multi-source data from air, space, and ground, and effectively isolating false alarm noise caused by single-point hardware failures. It obtains multiple cells of any three-dimensional solid engineering mesh, and based on the distribution of ground-side data from different sensors and the relative distance between the center of each cell and the sensors, obtains the cluster spatial deformation coherence coefficient, accurately quantifying the physical reality of regional coordinated deformation. By simulating the deviation characteristics of the coherence coefficient of spatial deformation of adjacent ground-side cluster nodes before and after fusion, a ground-side composite monitoring cluster node is generated, accurately restoring the spatial scale of the real landslide body. Based on the number of ground-side cluster nodes within the ground-side composite monitoring cluster node and the required number of clusters, candidate sky-side regional nodes are selected, effectively overcoming the signal dilution effect of macroscopic large pixels on microscopic local anomalies. The time overlap ratio between the anomaly duration window of the ground-side composite monitoring cluster node and the observation time window of the candidate sky-side regional nodes is obtained and used as the graph topological edge weights between nodes. Based on the maximum edge weight matching relationship, the sky-side and ground-side data sources with the lowest spatiotemporal mismatch are extracted for integrated air-space-ground monitoring of highway slopes. This invention improves the effectiveness of integrated monitoring by accurately constructing a graph topological edge weight mechanism for fusion mismatch. Attached Figure Description

[0014] Figure 1 This is a structural block diagram of an integrated air-ground-space monitoring system for highway slopes provided in one embodiment of the present invention; Figure 2 This is a flowchart illustrating a method for obtaining the spatial deformation coherence coefficient of a cluster, as provided in an embodiment of the present invention. Detailed Implementation

[0015] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0016] The following description, in conjunction with the accompanying drawings, details a specific scheme for an integrated air-space-ground monitoring system for highway slopes provided by this invention.

[0017] Please see Figure 1The diagram illustrates a structural block diagram of an integrated air-space-ground monitoring system for highway slopes according to an embodiment of the present invention. The system specifically includes: a data acquisition module 101, a deformation coherence analysis module 102, a topology edge weight calculation module 103, and an integrated monitoring module 104. The data acquisition module 101 is used to acquire multi-source monitoring data of highway slopes from air, space, and ground, construct a three-dimensional baseline map of the slope based on the airside data, and map the airside remote sensing data and groundside sensing data to the baseline map for spatial alignment.

[0018] In embodiments of the present invention, multi-source monitoring data of highway slopes from air, space, and ground are acquired, including InSAR satellite sky-side data, UAV air-side data, and ground sensor ground-side data. Sky-side data is obtained by acquiring time-series SAR images from radar remote sensing satellites, interpreting the satellite-to-ground distance using time-series InSAR technology, and obtaining the radar interferometric coherence coefficient of the sky-side remote sensing grid. When the radar interferometric coherence coefficient is less than a preset coherence threshold, the sky-side remote sensing grid data is deemed invalid and removed. Based on relevant historical experience, the preset coherence threshold is set to 0.35. Air-side data is obtained using UAVs deployed along the highway. The fully automated hangar performs periodic adaptive data collection. During inspections, the UAV follows a pre-planned route and uses its onboard oblique photography camera and LiDAR to perform high-frequency scanning. Based on the highway red line and the top and toe lines of the slope excavation, a three-dimensional spatial mask clipping frame is established. Only geometric slices of key damaged areas such as the slope grid beams are extracted and recorded to construct a three-dimensional baseline map of the slope. Ground-side data is collected in real time by IoT sensors deployed on the highway graded walkways and anti-slide piles, and median filtering and adaptive variance sliding window are introduced for online cleaning. The specific methods are well known to those skilled in the art and will not be described in detail here.

[0019] It should be noted that, in terms of spatial scale adaptation, a unified three-dimensional rectangular coordinate system for slopes is established using the DEM model; to address the perspective shrinkage and overlapping geometric distortion caused by the side-view imaging of the Tian-side satellite InSAR, the system uses the satellite orbit vector skewness matrix to perform inverse geocoding reprojection on the two-dimensional radar projection grid, mapping it to the three-dimensional slope coordinates of the highway, so that the different spatial scales corresponding to the monitoring data are mapped to a consistent spatial geographic reference.

[0020] The deformation coherence analysis module 102 is used to obtain multiple three-dimensional solid engineering grids of highway slopes, taking the sensor set in each grid as a ground-side cluster node and the corresponding sky-side remote sensing grid as a sky-side regional node; obtaining multiple cells of any three-dimensional solid engineering grid, and obtaining the cluster spatial deformation coherence coefficient based on the distribution of ground-side data from different sensors and the relative distance between the center position of each cell and the sensor.

[0021] A single pixel grid of InSAR from a space-side satellite covers a very large area, while ground-based sensors are just isolated, tiny points. Points and surfaces cannot be directly compared. Clustering all sensors within a large grid into a ground-side cluster node essentially upgrades and integrates discrete point data into a continuous surface data, thereby forcibly aligning it with the macroscopic remote sensing grid in space at a geometric scale. Multiple three-dimensional engineering grids of a highway slope are obtained, and the sensor set within each grid is taken as a ground-side cluster node, while the corresponding space-side remote sensing grid is taken as a space-side regional node.

[0022] It should be noted that the longitudinal engineering chainage interval of the highway is used as the one-dimensional control axis, the top and bottom lines of the graded stepped walkway are used as the two-dimensional spatial boundary lines, and the edges of the rows of anti-slide piles and anchor cable grid beams are used as the three-dimensional stiffness boundaries. Based on the above physical boundaries, the entire highway slope is spatially cut and discretized into several closed three-dimensional solid engineering grids without overlap. All sensors contained in the grid are used as a ground-side cluster node of the algorithm topology layer. According to the spatial resolution of radar interferometric satellite pixels, multiple two-dimensional remote sensing grids covering the solid engineering grids of the entire highway slope are mapped to the sky-side region nodes in the graph topology structure.

[0023] Due to cost and terrain limitations, sensors installed on slopes are often sparse discrete points, and there are many blind spots that cannot be detected within the grid. Therefore, the large grid is divided into small units to obtain multiple units of any three-dimensional solid engineering grid. It should be noted that, in the embodiments of the present invention, the entire grid surface is cut into multiple non-overlapping units according to a preset high resolution, such as a discrete accuracy of 0.1m × 0.1m.

[0024] Real slope landslides are physically interconnected, and the deformation within the entire grid should be consistent and coordinated. If a sensor malfunctions or is subject to local interference, the data distribution within the grid will be extremely uneven. Based on the ground-side data distribution of different sensors and the relative distance between the center of each unit and the sensor, the cluster spatial deformation coherence coefficient is obtained.

[0025] Preferably, in one embodiment of the present invention, the method for obtaining the cluster spatial deformation coherence coefficient is described in [reference needed]. Figure 2 It illustrates a flowchart of a method for obtaining the coherence coefficient of cluster spatial deformation, including: Step S201: Obtain the reliability factor for each sensor based on the distribution of ground-side data from different sensors.

[0026] Preferably, in one embodiment of the present invention, the method for obtaining the reliability factor includes: The first step is to normalize the ground-side data of the sensors and obtain the difference between the normalized ground-side data of each sensor between each time series and the next time series, which is used as the residual value of each sensor in each time series. It should be noted that, in the embodiments of the present invention, maximum and minimum value normalization is adopted, that is, the maximum and minimum values ​​of all time-series ground-side data of all sensors within the specified acquisition period and sampling interval are selected and normalized to the range of 0-1. The data sampling period and sampling interval can be adjusted according to the actual scenario requirements; the specific means are technical means well known to those skilled in the art, and will not be described in detail here.

[0027] The second step is to obtain the mean of the product of the residual values ​​between different sensors across all time series, and use the absolute value of the mean of the product as the correlation of the monitoring data between different sensors. The formula for the correlation of monitoring data is expressed as: ;in, Indicates the first The sensor and the first Correlation of monitoring data between individual sensors; Indicates the number of timing sequences; Indicates the first The sensor at the first The residual values ​​of the timing sequence; Indicates the first The sensor at the first The residual value of the time series, ; This indicates taking the absolute value.

[0028] It should be noted that the larger the absolute value of the mean of the residual products, the higher the synergy between different sensors in the dynamic change trend, and the greater the similarity of the monitoring data. During the accumulation process, if two sets of data show consistent changes in the same direction or consistent changes in opposite directions, the same sign will be accumulated, and the result obtained after taking the absolute value is highly correlated, which can reflect strong mechanical synergy. If the sensor changes randomly due to random noise interference, the positive and negative signs will alternate and cancel each other out during accumulation, and the result will approach zero after taking the absolute value, with extremely low correlation.

[0029] The third step is to obtain the mean value of the correlation between each sensor and other sensors' monitoring data, and then normalize it as the reliability factor for each sensor.

[0030] The correlation of monitoring data reflects the overall correlation between the data monitored by the sensors. The greater the correlation, the greater the reliability factor. It should be noted that in the embodiments of the present invention, maximum and minimum value normalization is used to analyze the mean value of the correlation between each sensor and other sensors, and select the maximum and minimum values ​​to normalize to the range of 0-1. The specific means are well known to those skilled in the art and will not be described in detail here.

[0031] Step S202: Based on the confidence factor of each sensor and the relative distance between the center position of each unit and the sensor, obtain the adaptive interpolation weight between the center position of each unit and each sensor.

[0032] Preferably, in one embodiment of the present invention, the method for obtaining adaptive interpolation weights includes: The first step is to perform a negative correlation mapping on the confidence factor of each sensor, which serves as the first distance coefficient; It should be noted that, in the embodiments of the present invention, the exponential function with the natural constant as its base is... A negative correlation mapping is performed. The higher the confidence factor, the higher the correlation between the sensor monitoring data and the overall sensor cluster, and the smaller the distance between the sensors. The specific methods are well known to those skilled in the art and will not be elaborated here.

[0033] The second step is to obtain the product of the first distance coefficient and the relative distance between the center position of each unit and each sensor, which is used as the equivalent spatial distance between the center position of each unit and each sensor. The formula for equivalent spatial distance is expressed as: ;in, Indicates the first The center position of the unit is related to the first The equivalent spatial distance between the sensors; Indicates the first The center position of the unit is related to the first The relative distance between the sensors; Indicates the first The reliability factor of each sensor; Indicates the first distance coefficient; This represents an exponential function with the natural constant as its base.

[0034] It should be noted that in some embodiments of the present invention, the relative distance is obtained by existing distance calculation methods such as Euclidean distance or Manhattan distance; the specific means are well known to those skilled in the art and will not be described in detail here.

[0035] The third step is to perform a negative correlation normalization mapping on the equivalent spatial distance between the center position of each unit and the sensor, which serves as the adaptive interpolation weight between the center position of each unit and each sensor.

[0036] It should be noted that, in the embodiments of the present invention, the reciprocal of the square of the equivalent spatial distance is used for negative correlation mapping. In order to avoid the equivalent spatial distance being 0, a very small positive number with consistent dimensions is added to the denominator. The value of the number is specifically set according to the range of the denominator, which will not be elaborated here. The sum of the squares of the negative correlation mapping results of the equivalent spatial distance between each unit center position and all sensors is obtained. The ratio of the negative correlation mapping result of the equivalent spatial distance between each unit center position and each sensor to the sum of the squares is calculated, that is, normalized, and used as the adaptive interpolation weight between each unit center position and each sensor.

[0037] Step S203: Based on the adaptive interpolation weights between the center position of each unit and different sensors, and the ground-side data of the sensors, obtain the spatial deformation characteristic values ​​of each unit, and obtain the cluster spatial deformation coherence coefficient.

[0038] Preferably, in one embodiment of the present invention, the method for obtaining spatial deformation feature values ​​includes: The spatial deformation characteristic value of each unit is determined by the adaptive interpolation weights between the center position of each unit and different sensors, as well as the multiplication and summation of the normalized ground-side data of the sensors.

[0039] The formula for the eigenvalue of spatial deformation is expressed as: ;in, Indicates the first Spatial deformation characteristic values ​​of each unit; Indicates the first The center position of the unit is related to the first Adaptive interpolation weights between sensors; Indicates the first Normalized ground-side data from each sensor; Indicates the number of sensors.

[0040] Based on this, the larger the adaptive interpolation weight, the larger the ground-side data, the larger the spatial deformation characteristic value, and the greater the deformation damage intensity.

[0041] Preferably, real slope landslides are physically interconnected, and the deformation within the entire grid surface should be consistent and coordinated; if a sensor malfunctions or is subject to local interference, the data distribution within the grid will be extremely uneven, resulting in a smaller cluster spatial deformation coherence coefficient; therefore, the consistency of deformation is quantified by analyzing the cluster spatial deformation coherence coefficient; in one embodiment of the present invention, the method for obtaining the cluster spatial deformation coherence coefficient includes: The first step is to obtain the mean value of the spatial deformation characteristic value of all elements for any three-dimensional solid engineering mesh, as the average deformation level; and to obtain the fluctuation characteristics of the spatial deformation characteristic value of all elements, as the deformation fluctuation degree.

[0042] It should be noted that, in one embodiment of the present invention, the fluctuation characteristics are characterized by calculating the standard deviation. The larger the standard deviation, the larger the fluctuation characteristics, and the greater the degree of non-uniformity of deformation. The smaller the standard deviation, the smaller the fluctuation characteristics, and the smaller the degree of non-uniformity of deformation. In other embodiments of the present invention, the fluctuation characteristics can also be characterized by calculating the variance. The specific means are well known to those skilled in the art and will not be described in detail here.

[0043] The second step is to obtain the spatial variation coefficient based on the proportional relationship between the deformation fluctuation degree and the average deformation level. The deformation fluctuation degree and the spatial variation coefficient are positively correlated. It should be noted that, considering that when the slope is in a stable and deformation-free state, the average value of the spatial deformation characteristics is 0, when calculating the ratio, a very small positive number with consistent dimensions is added to the denominator. The value of this number is set according to the specific range of the denominator, which will not be elaborated here.

[0044] The third step is to perform a negative correlation mapping on the spatial variation coefficient, and obtain the product between the negative correlation mapping result and the average deformation level, which is used as the cluster spatial deformation coherence coefficient.

[0045] It should be noted that, in the embodiments of the present invention, negative correlation mapping is performed by an exponential function with the natural constant as the base. The larger the spatial variation coefficient, the smaller the deformation uniformity and the smaller the cluster spatial deformation coherence coefficient. The specific means are well known to those skilled in the art and will not be described in detail here.

[0046] The formula for the cluster spatial deformation coherence coefficient is expressed as: ;in, Represents the spatial deformation coherence coefficient of the cluster; Indicates the average level of deformation; Indicates the degree of deformation fluctuation; Represents a very small positive number; Indicates the coefficient of spatial variation; Represents the natural constant.

[0047] The topology edge weight calculation module 103 is used to perform adaptive grid merging based on the deviation characteristics of the cluster spatial deformation coherence coefficient before and after the simulation fusion of adjacent ground-side cluster nodes, and generate ground-side composite monitoring cluster nodes; obtain the required number of remotely observable clusters according to the area ratio relationship between the sky-side remote sensing grid and the three-dimensional solid engineering grid; select candidate sky-side regional nodes according to the number of ground-side cluster nodes in the ground-side composite monitoring cluster nodes and the required number of clusters; obtain the time overlap ratio between the anomaly duration window of the ground-side composite monitoring cluster node and the observation time window of the candidate sky-side regional node as the graph topology edge weight between the nodes.

[0048] The cluster spatial deformation coherence coefficient reflects the comprehensive evaluation index of the overall and true deformation intensity within the grid. The landslide instability of highway slopes is essentially a co-dynamic rheological process of continuous medium in geotechnical mechanics, and has spatial continuity. Based on the deviation characteristics of the cluster spatial deformation coherence coefficient before and after the simulation fusion of adjacent ground-side cluster nodes, grid adaptive merging is performed to generate ground-side composite monitoring cluster nodes.

[0049] Preferably, in one embodiment of the present invention, the method for obtaining ground-side composite monitoring cluster nodes includes: Obtain the absolute difference between the cluster spatial deformation coherence coefficients of each ground-side cluster node and its adjacent ground-side cluster nodes before and after the simulated fusion. Calculate the ratio of the absolute difference to the cluster spatial deformation coherence coefficients of each ground-side cluster node before the simulated fusion, and use this as the relative deviation change rate. It should be noted that each ground-side cluster node is simulated and merged with its adjacent ground-side cluster nodes to become a new ground-side cluster node. The simulated and merged cluster spatial deformation coherence coefficient is obtained based on the method for obtaining the cluster spatial deformation coherence coefficient.

[0050] If the relative deviation change rate corresponding to each ground-side cluster node is less than or equal to the preset change threshold, the grid corresponding to each ground-side cluster node and the adjacent ground-side cluster node are merged, and the sensor set in the merged grid is taken as the ground-side composite monitoring cluster node; if there is a ground-side cluster node whose relative deviation change rate is greater than the preset change threshold, merging is not performed, and the ground-side cluster node is taken as the ground-side composite monitoring cluster node.

[0051] It should be noted that the absolute difference refers to the absolute value of the difference. In one embodiment of the present invention, in order to give the system tolerance to normal sensor measurement errors and minor environmental fluctuations, a strict spatial mechanical consistency boundary is constructed to ensure that only truly adjacent meshes under the same continuous sliding mechanism will be topology merged. Based on relevant historical experience, a preset change threshold is set to... In other embodiments of the present invention, the size of the preset change threshold can be set according to specific circumstances, and will not be limited or described in detail here.

[0052] By calculating the area ratio of the large sky grid to the small ground grid, we can inversely calculate how many small ground grids need to deform simultaneously for their superimposed energy to be sufficient to break through the lower limit of satellite signal recognition; based on the area ratio of the sky-side remote sensing grid and the three-dimensional solid engineering grid, we can obtain the required number of clusters that can be observed by remote sensing.

[0053] Preferably, in one embodiment of the present invention, the method for obtaining the required number of clusters includes: The ratio of the equivalent pixel area of ​​the surface of the sky-side remote sensing grid to the actual projected area of ​​the three-dimensional solid engineering grid is obtained as the first area ratio. The product of the first area ratio and the preset satellite phase perturbation sensitivity coefficient is obtained as the required number of clusters for remote sensing observation.

[0054] It should be noted that the first area ratio reflects the ratio of the equivalent pixel area data of the land surface to the actual area. The larger the first area ratio, the more clusters are required. The required number of clusters effectively overcomes the signal dilution effect of macroscopic large pixels on microscopic local anomalies and eliminates blind fusion caused by scale misalignment.

[0055] It should be noted that, in the embodiments of the present invention, the preset satellite phase perturbation sensitivity coefficient is used to control the relative proportion and to define the strictness of the number of clusters. In order to ensure that the abnormal deformation signal dominates the echo energy of the satellite single pixel grid, it is set to 0.5 based on relevant historical experience. In other embodiments of the present invention, the magnitude of the preset satellite phase perturbation sensitivity coefficient can be set according to the specific situation, and is not limited or described here.

[0056] The actual engineering grid is divided based on the actual slope pile foundation and walkway. The area corresponding to the actual grid is much smaller than the area corresponding to the pixel grid in the satellite imaging field of view. Therefore, in order for the anomalies detected by the sensors in the actual engineering grid to be captured by the satellite, a sufficient number of small-area actual engineering grids are needed to form a large-area actual engineering grid so that the deformation energy reflected by the grid will not be diluted or submerged in the satellite echo calculation. Based on the number of ground-side cluster nodes in the ground-side composite monitoring cluster nodes and the required number of clusters, candidate satellite-side regional nodes are selected.

[0057] Preferably, in one embodiment of the present invention, the method for obtaining candidate celestial region nodes includes: If the number of ground-side cluster nodes in the ground-side composite monitoring cluster node is greater than or equal to the required number of cluster nodes, all sky-side regional nodes corresponding to the grid where the ground-side composite monitoring cluster node is located will be merged as candidate sky-side regional nodes.

[0058] The more likely the anomaly duration window of the ground-side composite monitoring cluster node is to reflect the actual landslide occurrence period, the more likely the observation time window is to reflect the time range of cumulative deformation analyzed by the satellite; the time overlap ratio reflects the degree of temporal agreement between the deformation captured by the satellite and the deformation recorded by the ground sensor; the time overlap ratio between the anomaly duration window of the ground-side composite monitoring cluster node and the observation time window of the candidate sky-side regional node is obtained as the graph topology edge weight between the nodes.

[0059] Preferably, in one embodiment of the present invention, the method for obtaining the time overlap ratio includes: The time when the ground-side data of the earliest sensor in the ground-side composite monitoring cluster node exceeds the preset warning threshold is taken as the start timestamp, and the time when the ground-side data of the last sensor exceeds the preset warning threshold is taken as the end timestamp, thus forming an abnormal duration window. It should be noted that, in the embodiments of the present invention, the preset warning threshold refers to the critical judgment value of the road slope soil and rock mass moving from safe and stable to unstable and failed in terms of mechanical deformation or environmental state, which is set a priori based on the survey and design specifications of road slopes and the physical and mechanical parameters of soil and rock.

[0060] The observation time window is formed by taking the satellite main radar image capture time corresponding to the candidate skyside regional node as the starting point and the secondary image capture time as the ending point. It should be noted that in InSAR interferometry, satellite radar images inevitably come with attribute parameter files. The capture time of the main and secondary images is an inherent public attribute of the data, which can be directly parsed and read.

[0061] Obtain the overlap duration between the anomaly duration window and the observation time window, and calculate the ratio of the overlap duration to the total duration within the anomaly duration window as the time overlap ratio.

[0062] Therefore, the higher the temporal overlap ratio, the more the deformation captured by the satellite and the deformation recorded by the ground sensor are in time, thus avoiding the influence of false signals that are spatially aligned but temporally misaligned.

[0063] The integrated monitoring module 104 is used to extract the sky-side and ground-side data sources with the lowest spatiotemporal mismatch based on the maximum edge weight matching relationship, and to carry out integrated air-space-ground monitoring of highway slopes.

[0064] The larger the edge weight, the higher the spatiotemporal fit and the lowest mismatch. Therefore, extracting the sky and ground data sources with the lowest spatiotemporal mismatch through the maximum edge weight matching relationship is more helpful for integrated air-space-ground monitoring of highway slopes.

[0065] It should be noted that the graph topology edge weights constructed from the obtained time overlap ratios are sorted by numerical value, and the data sources corresponding to the highest edge weight values ​​are extracted from the sky and ground sides. Since it has the optimal edge weight data association, it represents the highest time overlap ratio and obtains the data source with the lowest spatiotemporal mismatch, enabling integrated air-space-ground monitoring of highway slopes.

[0066] In summary, this invention obtains multiple elements of any three-dimensional solid engineering mesh. Based on the distribution of ground-side data from different sensors and the relative distance between the center of each element and the sensor, it obtains the cluster spatial deformation coherence coefficient. Based on the deviation characteristics of the cluster spatial deformation coherence coefficient before and after simulation fusion of adjacent ground-side cluster nodes, it generates ground-side composite monitoring cluster nodes. Based on the number of ground-side cluster nodes within the ground-side composite monitoring cluster nodes and the required number of clusters, it selects candidate sky-side regional nodes. It obtains the time overlap ratio between the anomaly duration window of the ground-side composite monitoring cluster node and the observation time window of the candidate sky-side regional nodes, using this as the graph topology edge weight between nodes. Based on the maximum edge weight matching relationship, it extracts the sky-side and ground-side data sources with the lowest spatiotemporal mismatch, enabling integrated air-space-ground monitoring of highway slopes. This invention improves the effectiveness of integrated monitoring by accurately constructing a graph topology edge weight mechanism for fusion mismatch.

[0067] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A highway slope integrated air-space-ground monitoring system, characterized in that, The system includes: The data acquisition module is used to acquire multi-source monitoring data of highway slopes from air, space, and ground. Based on the airside data, a three-dimensional baseline map of the slope is constructed, and the airside remote sensing data and groundside sensor data are mapped to the baseline map for spatial alignment. The deformation coherence analysis module is used to obtain multiple three-dimensional solid engineering meshes of highway slopes. The sensor set in each mesh is regarded as a ground-side cluster node, and the corresponding sky-side remote sensing grid is regarded as a sky-side regional node. Multiple cells of any three-dimensional solid engineering mesh are obtained. Based on the distribution of ground-side data from different sensors and the relative distance between the center of each cell and the sensor, the cluster spatial deformation coherence coefficient is obtained. The method for obtaining the cluster spatial deformation coherence coefficient includes: Based on the distribution of ground-side data from different sensors, the reliability factor of each sensor is obtained; Based on the confidence factor of each sensor and the relative distance between the center position of each unit and the sensor, the adaptive interpolation weight between the center position of each unit and each sensor is obtained. Based on the adaptive interpolation weights between the center position of each unit and different sensors, as well as the ground-side data of the sensors, the spatial deformation characteristic values ​​of each unit are obtained, and the cluster spatial deformation coherence coefficient is obtained. The method for obtaining the spatial deformation feature values ​​includes: The spatial deformation characteristic value of each unit is determined by the adaptive interpolation weights between the center position of each unit and different sensors, as well as the multiplication and summation of the normalized ground-side data of the sensors. For any three-dimensional solid engineering mesh, obtain the mean value of the spatial deformation characteristic value of all elements as the average deformation level; obtain the fluctuation characteristics of the spatial deformation characteristic value of all elements as the deformation fluctuation degree. Based on the proportional relationship between deformation volatility and average deformation level, the spatial variation coefficient is obtained, and the deformation volatility is positively correlated with the spatial variation coefficient. A negative correlation mapping is performed on the spatial variation coefficient, and the product between the negative correlation mapping result and the average deformation level is obtained as the cluster spatial deformation coherence coefficient; The topology edge weight calculation module is used to perform adaptive grid merging based on the deviation characteristics of the cluster spatial deformation coherence coefficient before and after the simulation fusion of adjacent ground-side cluster nodes, generating ground-side composite monitoring cluster nodes; obtaining the required number of remotely observable clusters based on the area ratio relationship between the sky-side remote sensing grid and the 3D solid engineering grid; selecting candidate sky-side regional nodes based on the number of ground-side cluster nodes in the ground-side composite monitoring cluster nodes and the required number of clusters; obtaining the anomaly duration window of the ground-side composite monitoring cluster node and the time overlap ratio between the observation time windows of the candidate sky-side regional nodes as the graph topology edge weights between nodes. The method for obtaining the nodes of the ground-side composite monitoring cluster includes: Obtain the absolute difference between the cluster spatial deformation coherence coefficients of each ground-side cluster node and its adjacent ground-side cluster nodes before and after the simulated fusion. Calculate the ratio of the absolute difference to the cluster spatial deformation coherence coefficients of each ground-side cluster node before the simulated fusion, and use this as the relative deviation change rate. If the relative deviation change rate corresponding to each ground-side cluster node is less than or equal to the preset change threshold, the grid corresponding to each ground-side cluster node and the adjacent ground-side cluster node is merged, and the sensor set in the merged grid is taken as the ground-side composite monitoring cluster node; if there is a ground-side cluster node whose relative deviation change rate is greater than the preset change threshold, merging is not performed, and the ground-side cluster node is taken as the ground-side composite monitoring cluster node. The method for obtaining the required number of clusters includes: The ratio of the equivalent pixel area of ​​the surface of the sky-side remote sensing grid to the actual projected area of ​​the three-dimensional solid engineering grid is obtained as the first area ratio. The product of the first area ratio and the preset satellite phase perturbation sensitivity coefficient is obtained as the required number of clusters for remote sensing observation; The integrated monitoring module is used to extract the sky-side and ground-side data sources with the lowest spatiotemporal mismatch based on the maximum edge weight matching relationship, and to carry out integrated air-space-ground monitoring of highway slopes.

2. The integrated air-space-ground monitoring system for highway slopes according to claim 1, characterized in that, The methods for obtaining the reliability factor include: The ground-side data of the sensors is normalized to obtain the difference between the normalized ground-side data of each sensor between each time series and the next time series, which is used as the residual value of each sensor in each time series. The mean of the product of the residuals between different sensors across all time series is obtained, and the absolute value of the mean of the product is used as the correlation of the monitoring data between different sensors. The mean value of the correlation between each sensor and the monitoring data of other sensors is obtained and normalized, and used as the reliability factor for each sensor.

3. The integrated air-space-ground monitoring system for highway slopes according to claim 2, characterized in that, The method for obtaining the adaptive interpolation weights includes: The confidence factor of each sensor is negatively correlated and used as the first distance coefficient; The product of the first distance coefficient and the relative distance between the center position of each unit and each sensor is obtained as the equivalent spatial distance between the center position of each unit and each sensor. A negative correlation normalization mapping is performed on the center position of each unit and the equivalent spatial distance between different sensors, which serves as the adaptive interpolation weight between the center position of each unit and each sensor.

4. The integrated air-space-ground monitoring system for highway slopes according to claim 1, characterized in that, The method for obtaining the candidate celestial region nodes includes: If the number of ground-side cluster nodes in the ground-side composite monitoring cluster node is greater than or equal to the required number of cluster nodes, all sky-side regional nodes corresponding to the grid where the ground-side composite monitoring cluster node is located will be merged as candidate sky-side regional nodes.

5. The integrated air-space-ground monitoring system for highway slopes according to claim 1, characterized in that, The method for obtaining the time overlap ratio includes: The time when the ground-side data of the earliest sensor in the ground-side composite monitoring cluster node exceeds the preset warning threshold is taken as the start timestamp, and the time when the ground-side data of the last sensor exceeds the preset warning threshold is taken as the end timestamp, thus forming an abnormal duration window. The observation time window is formed by taking the satellite main radar image capture time corresponding to the candidate skyside regional node as the starting point and the secondary image capture time as the ending point. Obtain the overlap duration between the anomaly duration window and the observation time window, and calculate the ratio of the overlap duration to the total duration within the anomaly duration window as the time overlap ratio.

Images