Complex deformation body same-scale time sequence deformation monitoring method

Abstract

The invention discloses a method for monitoring same-scale time sequence deformation information of a complex deformation body under a multi-source earth observation technology fusion background. The same-scale deformation information is accurately extracted from the multi-source observation data based on the multi-scale and multi-dimensional deformation characteristics of a complex deformation body, and the method has important application value for building safety operation, smart city construction, geological disaster monitoring and early warning and the like. According to the method, a high-resolution earth observation system and a single Beidou navigation positioning system in China are taken as main surveying and mapping technical means, and an intelligent extraction method for same-scale time sequence deformation of massive, complex-scale and multi-dimensional monitoring data is established according to an Msplit classification estimation theory, a time sequence model and a median function.

Description

Technical Field

[0001] The present invention relates to a method for monitoring deformation of complex deformable bodies using surveying and mapping technology to provide geographic information services. Specifically, the present invention relates to a method for monitoring deformation of complex deformable bodies using multi-source observation technologies such as a high-resolution earth observation system (hereinafter referred to as a high-resolution system), a single Beidou navigation and positioning system (hereinafter referred to as a single Beidou system), and three-dimensional laser scanning to obtain the same-scale time series deformation of complex deformable bodies, thereby accurately extracting the same-scale deformation information of the multi-dimensional deformation of the complex deformable body. Background Art

[0002] Deformation monitoring of complex objects, such as structures, landslides, and dams, is essential to ensure their stable operation and protect the lives and property of the people. With the development of my country's economy and society, high-rise buildings, reservoirs, dams, and other structures have proliferated. Monitoring geological hazards such as landslides and debris flows, which threaten human life and property, has become a crucial scientific and technological task. However, current monitoring methods, which rely on manually deployed benchmarks and a single observation technique to obtain deformation information on objects, suffer from high workload and limited information, failing to meet the actual needs of my country's social development. Complex objects exhibit highly variable deformation characteristics, and accurately acquiring uniformly distributed deformation information over a fixed period is crucial for revealing deformation patterns, predicting deformation trends, and preventing disasters. With the advancement of my country's surveying and mapping technology, particularly the mature application of high-resolution and single-source BeiDou satellite systems, autonomous monitoring of complex objects is essential for ensuring social security and stability and promoting the development of the geographic information industry. Summary of the Invention

[0003] The purpose of the invention is to address the technical bottlenecks faced by the extraction of deformation information of complex deformable bodies in the same-scale time series, and the application difficulties existing in the autonomous monitoring of deformable bodies by my country's surveying and mapping technology. The present invention provides a same-scale time series deformation monitoring method for complex deformable bodies, and provides a programmatic solution to break through the application bottleneck of autonomous monitoring and prediction of deformable bodies in the development of geographic information services by surveying and mapping technology, and solve the following technical problems involved: 1. Autonomous identification of monitoring points and layout of deformation monitoring networks; 2. Baseline reduction of deformation data and accuracy of three-dimensional deformation fields; 3. Autonomous identification and classification of same-scale deformation data based on time series models.

[0004] The principle of the invention is to autonomously screen monitoring points with significant land features based on the free network adjustment model and the S conversion model, and to build a deformation monitoring network by combining the method of manually arranging monitoring points. Using Gaussian plane coordinates and elevation as the three-dimensional benchmark data to describe deformation, the deformation monitoring data is normalized for the conversion and benchmark transformation of air-based monitoring data and ground-based observation data respectively. In order to accurately calibrate the deformation field described by the time series model, the theory of irregular triangulation is applied to generate a three-dimensional deformation field based on the measured data, and the deformation field after registration is calibrated in combination with the F detection model. Apply M split(q＝2) Estimation and median function theory are used to extract the deformation information of the same scale in different dimensions, and the iterative method is used to realize the cyclic discrimination of the deformation information of the same scale.

[0005] Technical Solution: A method for monitoring deformation of complex deformable objects at the same scale in a time series. The technical solution is: registration-detection-comparison-classification. The steps of this solution are as follows: Given the lack of a unified spatial benchmark for space-based and ground-based observation data involved in deformation monitoring, a reference benchmark is established through autonomous identification of monitoring points, supplemented by manual deployment, effectively overcoming the many shortcomings of current methods. Based on the verification results of existing practices, the use of a single Beidou system for deformation monitoring can effectively avoid the spatial benchmark conversion used by different navigation and positioning systems. The spatial data unification of airborne three-dimensional data and ground-based distance and angle data is achieved through conformal projection and elevation anomaly fitting. The F-test method based on probability statistics is used to detect the significance of the time series model, and the optimal model is accurately identified through the cyclic identification method. Based on the irregular triangulation and Kriging interpolation theory, a three-dimensional deformation field based on the measured data is generated. The three-dimensional deformation field fitted by the time series model is then compared by statistically analyzing the deformation errors of the monitoring points. Based on the normalized benchmark, model detection, and data comparison, the massive deformation data is scale-classified. According to the time series model, the M-type with two splitting numbers is used. split(q＝2) Estimation and median function theory are used to construct a same-scale deformation extraction method with iterative classification.

[0006] The specific steps of the method are as follows:

[0007] Step 1. Utilize high-resolution monitoring data from two or more phases to autonomously identify monitoring points, combined with manual deployment of monitoring points to construct a deformation monitoring network. Step 2. Coordinate conversion and deviation correction are performed between the airborne 3D monitoring data and the ground-based monitoring data. Using the minimum deformation point as the reference point, the deformation values of the monitoring points are converted to relative deformation values between the points.

[0008] Step 3. Construct a three-dimensional deformation field based on the calculated relative deformation between monitoring points, perform significance detection of the time series model and compare the accuracy of the three-dimensional deformation field predicted by the model.

[0009] Step 4. Apply M according to the time series model after detection split(q＝2) Estimation and median function theory are used to perform iterative cycle extraction of deformation information of the same scale.

[0010] In Step 1, a free network adjustment model based on a centroidal datum is first applied. Based on the deformation values of the monitoring points, the S-transformation model and t-test (not limited to t-test) are combined to autonomously identify the monitoring points obtained by the high-resolution system without a fixed datum. If the number of autonomously identified monitoring points meets the requirement that the number of points is greater than twice the number of observations required to solve the time series model, and the ground equipment at the same station can see three or more monitoring points, then they are used to form the deformation control network. If these requirements are not met, the number of monitoring points should be supplemented by manually deploying them.

[0011] Furthermore, the manually laid out monitoring points must meet the operational requirements for the layout of Class B control points in the Global Positioning System Measurement Specification (GB / T18314-2009).

[0012] In Step 2, the conversion of airborne 3D monitoring data into plane coordinates and elevations proceeds as follows: 3D spatial coordinates - geodetic coordinates - plane coordinates and elevations. When converting geodetic coordinates into plane coordinates, the longitude of the deformation object is used as the central meridian of the projection zone to minimize projection deformation. The geodetic height is corrected for vertical deviation, and the corresponding normal direction is corrected to the vertical direction corresponding to the elevation. Elevation anomalies are corrected using the Chinese geoid model. Ground-based distance observations are corrected for earth curvature and, combined with angle observations, are converted into plane coordinates and elevations. Elevation values are converted into elevations using the same elevation anomaly correction parameters as for the 3D spatial coordinate conversion.

[0013] Furthermore, when using the Gauss-Krüger projection to convert geodetic coordinates into plane coordinates, the longitude value of the meridian where the deformation body is located should be accurate to one hundredth of a second. Based on at least two periods of monitoring data, the F test is used to dynamically screen the benchmark points with the minimum deformation value as a parameter, and the relative deformation value of the monitoring point is calculated based on this.

[0014] In Step 3, a three-dimensional deformation field of the relative deformation of the monitoring points is generated based on the irregular triangulation network and kriging interpolation theory. A Gauss-Markov model is applied, with the deformation variables in plane coordinates and elevation as the model's dependent variables. Based on the principle of proceeding from low-order terms to high-order terms, the time series model is tested for significant cycles using the F-test. The time series model that passes the significance test is solved using the least squares algorithm, and the three-dimensional deformation field is estimated. The mean square error of the deformation at each monitoring point is calculated, and the three-dimensional deformation field is then calibrated.

[0015] Furthermore, the deformation field estimated by the model and the deformation field generated by the monitoring data should be registered, and the reference points for registration are the reference points dynamically selected in Step 2.

[0016] In Step 4, the relative deformation values of the three dimensions are used as the dependent variables and the deformation period is used as the variable to extract the information of the same scale deformation. The least squares estimation of the time series model is used as the initial value. According to M split(q＝2) The residual vector of the dependent variable is calculated by estimating the two sets of estimated values. The midpoint function is applied according to a certain step size to define the domain value parameters for same-scale discrimination. The t-test (not limited to the t-test) is applied to discriminate the same-scale deformation values, and such observations are eliminated from the observation sample, and the same-scale discrimination is repeated.

[0017] Advantages of the invention

[0018] 1. Applying airborne Earth observation systems, establish an autonomous identification method for monitoring points in the deformation control network, effectively reduce the technical difficulty of the current control network deployment, and overcome the many limitations of deploying monitoring points for complex deformation bodies.

[0019] 2. Establish a unified spatial benchmark reduction method for deformation monitoring using information-based observation technology, realize the construction of a coordinate benchmark in three dimensions: plane coordinates and elevation, and establish a technical solution to maintain benchmark consistency.

[0020] 3. By extracting the same-scale deformation information of complex deformable bodies, the deformation characteristics of the monitored object are revealed, and the technical difficulties faced in deformation processing, interpretation, and prediction of mixed deformation monitoring information are solved. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A landslide in southwest my country for deformation monitoring; Figure 2 4 is a flowchart of the monitoring method of the present invention. DETAILED DESCRIPTION

[0022] The following, with reference to the accompanying drawings, illustrates the technical implementation scheme of the present invention using deformation monitoring of a landslide in southwest my country as an example. This should not be construed as limiting the technical implementation scheme of the present invention. The time span from the start of deformation monitoring of the landslide to its collapse was fifteen months. The observation methods involved in deformation monitoring included:

[0023] (1) A total of 42 high-resolution observation images were acquired, and nine deformation monitoring points were autonomously identified. Since the autonomously identified monitoring points cannot meet the requirement of Step 1 of the technical solution of the present invention that the ground equipment at the same station can see more than three monitoring points, two additional monitoring points were manually deployed to form two groups of visible monitoring points.

[0024] (2) A single BeiDou system was used for control measurement, and a total of 28 observations were conducted on the artificially deployed monitoring points.

[0025] (3) Twenty-eight observations were conducted at eleven monitoring points using a prism-free total station.

[0026] (4) Twenty-eight observations of the landslide were carried out using a three-dimensional laser scanner, and three to five points were selected from eleven monitoring points to align the scan data.

[0027] The deformation data after monitoring data reduction and benchmark normalization are listed in Table 1.

[0028] Table 1 Settlement data of monitoring points during some monitoring periods (unit: mm)

[0029]

[0030] like Figure 2 The figure shows the flow chart of the monitoring method of the present invention. Based on this chart, the deformation information of the landslide body at the same scale is proposed. The cumulative settlement of some monitoring points is listed in Table 2.

[0031] Table 2 Same scale sedimentation extracted by the method of the present invention (unit: mm)

[0032] 1-5mm 5-10mm 10-15 15-20 Greater than 20 -0.9 -5.6 -12.9 -16.4 -21.3 -2 -7 -12.6 -16.3 -21.8 -1.2 -8.1 -14.5 -18.9 -33.2 0.9 -7.6 10.8 -16.6 -21.1 -0.8 -8.9 -11.1 -17.9 -30.1 -1.1 -6.4 -10.4 -15.4 -40.9 -1.4 -7.2 -11.8 -15.5 -21.3 -1.4 -9.1 -11.3 -19.4 -31.2 -2.1 -6.2 -13.1 -17.5 -20.4

Claims

1. A method for monitoring deformation of complex deformable bodies in a time series of the same scale, characterized by: According to the structural characteristics of the deformable body, a monitoring point layout scheme with autonomous identification as the main method and manual layout as the auxiliary method is adopted. The high-resolution earth observation system (hereinafter referred to as the high-resolution system) and the single Beidou navigation and positioning system (hereinafter referred to as the single Beidou system) are used as the main monitoring technology methods, supplemented by ground-based three-dimensional laser scanning and total station technology methods to obtain high-frequency and multi-dimensional deformation information. The time series model is applied, with the monitoring period as the variable and the deformation value of the monitoring point as the dependent variable. According to M split The estimation theory classifies and estimates the deformation in three dimensions respectively, and extracts the time series deformation of the same scale of the complex deformation body.

2. The method for monitoring deformation of a complex deformable body at the same scale and in a time series according to claim 1 is characterized by: The specific implementation steps of this monitoring method are as follows: Step 1. Use autonomous identification combined with manual deployment to deploy monitoring points and build a deformation monitoring network. Based on the structural characteristics of the deformable body, two or more phases of observation data from the high-resolution system are used. Based on the free network adjustment model and the S-transformation model, monitoring points with significant ground features are selected. Manual deployment of monitoring points supplements the number and location of autonomously identified points. The deployment of points should meet the requirements of high-precision satellite positioning and comply with the operational requirements for the deployment of Class B control points in the "Global Positioning System Measurement Specification" (GB / T18314-2009). Step 2. Normalize the deformation monitoring data, and calculate and detect the relative deformation. Normalize all types of monitoring data, and the normalization benchmark is the horizontal direction and the vertical direction. The normalization order of the air-based deformation monitoring data is three-dimensional rectangular coordinates (XYZ) - geodetic coordinates (BLH) - plane coordinates (xy) and elevation (h1). The normalization order of the foundation deformation monitoring data is length (S) and angle (horizontal angle α, vertical angle β) - plane coordinates (xy) and height (h2). In order to ensure that the relative deformation between monitoring points is accurately obtained through deformation monitoring, the deformation value of each monitoring point is calculated based on the minimum deformation point. The deformation monitoring operation must comply with the second-class horizontal and vertical deformation operation requirements in the "Code for Building Deformation Measurement" (JGJ8-2016). Step 3. Construct a three-dimensional deformation field of the deformable body to perform saliency detection of the time series model and three-dimensional accuracy of deformation prediction. Based on massive amounts of three-dimensional (x, y, and h) deformation data acquired through airborne observations (such as the Gaofen system and the Beidou system) and ground-based observations (such as 3D laser scanning and total stations), a 3D deformation field is generated using triangulated irregular networks and kriging interpolation theory. The time series model is recast as a Gauss-Markov model, and the model is tested for significance using the F-test. The deformation field is then simulated using the time series model and compared with the measured deformation field. Step 4. M of the time series model split Classification estimation, defining the discrimination domain value of the same-scale deformation, and extracting the same-scale deformation information in three dimensions. Apply M with two split numbers (q, q = 2) split(q＝2) Estimation: Initial classification is performed on the deformation values of the three dimensions (x, y, and h1). Based on the classification results, the median function is applied to define the scale parameter, identify deformation values of the same scale, and perform iterative cycle discrimination.

3. The method for monitoring deformation of a complex deformable object in a same-scale time series according to claim 2 is characterized by: In Step 1, the autonomous identification of monitoring points based on the high-resolution system adopts the method without fixed benchmarks, and the free network adjustment model uses the centroid benchmark condition to construct the constraint equation. The monitoring points are screened according to the deformation value of the monitoring points, combined with the S transformation model and t detection. The conditions for the layout of monitoring points are as follows: (1) The total number of monitoring points should be greater than twice the number of observations required to solve the time series model. If the number of autonomously identified monitoring points is insufficient, monitoring points should be manually deployed to supplement them. (2) The ground equipment at the same station can observe more than three monitoring points through visual communication. If the number of autonomously identified monitoring points is insufficient, monitoring points should be manually deployed to supplement them.

4. The method for monitoring deformation of a complex deformable object in a same-scale time series according to claim 2 is characterized by: In Step 2, the airborne three-dimensional coordinates are converted to the geodetic height (H) of the geodetic coordinates. The geodetic height must be corrected for vertical deviation, and the normal direction corresponding to the geodetic height must be corrected to the vertical. At the same time, the Chinese geoid model is applied to correct elevation anomalies and convert the geodetic height to elevation (h1). The longitude and latitude (BL) of the geodetic coordinates are converted to plane coordinates (xy) using the Gauss-Krüger projection. The central meridian of the projection zone must be the meridian of the deformation body (accurate to one hundredth of a second). Ground-based observations are converted to plane coordinates and elevations. Distance observations must be corrected for earth curvature. The height values (h2) are converted to elevations (h1) using the same geoid model used for the spatial three-dimensional coordinate conversion. The benchmark points for calculating the relative deformation of the monitoring points must be dynamically screened using the F-test with the minimum deformation value based on the monitoring data from every two periods.

5. The method for monitoring deformation of a complex deformable object in a same-scale time series according to claim 2 is characterized by: In Step 3, using the fixed monitoring period as the variable and the deformation variables in the three dimensions (x, y, and h1) as the dependent variables, a significance test is performed using the Gauss-Markov model using the F test: First, the significance test of the quadratic polynomial model is performed. If the test passes, the test is stopped. If the test fails, the cubic polynomial model test is performed, and so on. The time series model is solved using the Gauss-Markov model and the least squares algorithm. The simulated 3D deformation field and the measured deformation field must first be aligned using the reference points selected in Step 2. Then, the mean square error of the simulated 3D deformation field is calculated based on the relative deformation values of each monitoring point.

6. The method for monitoring deformation of a complex deformable object in a same-scale time series according to claim 2 is characterized by: In Step 4, the M of the time series model split The estimation must take the least squares solution of the model in Step 3 as the initial value. The technical steps for screening deformations of the same scale are: (1) According to M split The two sets of estimated values are used to calculate the residual vector of the dependent variable; (2) according to a certain step size, the median function is applied to define the domain value parameters of the same scale discrimination; (3) Apply the t-test, and the deformation values corresponding to the residuals smaller than the threshold are judged as deformations of the same scale, and such observations are eliminated from the observation sample. The deformation values corresponding to the residuals larger than the threshold are iteratively judged according to the above steps.

Images