Landslide slippage mode identification method and device based on multistage data matrix

Abstract

The invention relates to the technical field of landslide sliding mode recognition, in particular to a landslide sliding mode recognition method and device based on a multistage data matrix, and the method comprises the steps: obtaining landslide safety monitoring apparent displacement and deep displacement monitoring data, and carrying out the time sequence interpolation of the monitoring data; constructing an apparent displacement-deep displacement first-level data matrix according to the principle that the apparent and deep measuring points are the same in position; constructing a secondary data matrix according to the actual monitoring frequency; determining a data analysis sub-cycle, and constructing a three-level data matrix; calculating a correlation matrix of different deep positions and apparent deformation based on the three-level data matrix; and based on the correlation matrix, qualitatively judging the correlation between the apparent displacement and the deep displacement, and giving a depth range of the overall slip of the landslide and a slip key time node. According to the method, the slip depth range can be identified, the slip key time node can be captured, the spatial-temporal change trend of the landslide can be scientifically mastered, and a reference is provided for laying of safety treatment measures of the landslide.

Description

Technical Field

[0001] The present application relates to the technical field of landslide sliding pattern recognition, and in particular to a landslide sliding pattern recognition method and device based on a multi-level data matrix. Background Art

[0002] The sliding process of reservoir landslides is a complex dynamic process. Such landslides are widely distributed and have the characteristics of large number and volume. At present, the scale of large-scale hydropower projects in southwest my country is constantly expanding. The landslide time-dependent deformation phenomenon is obvious in the areas where the high-dam and large-reservoir hydropower projects under construction and completed are located. It is of great significance to determine the depth range of landslides and capture the key time nodes of sliding for the development of landslide safety prevention work.

[0003] Landslide deformation monitoring generally includes two aspects: apparent deformation and deep deformation monitoring. Commonly used means of apparent deformation monitoring include GNSS monitoring and robot monitoring, and commonly used deep deformation monitoring includes tension wire meters and inclinometers. Such monitoring data can instantly reflect the deformation characteristics of the landslide body in different spaces and at different times. Spearman correlation analysis is a rank-related non-parametric measurement method. It performs correlation analysis by replacing the "level" of the observed data with the "level", reflecting the strength of the monotonic relationship between variables. Spearman correlation analysis does not require prior knowledge to accurately obtain the sampling probability distribution of X and Y. If there are no repeated values ​​in the data and when the two variables are completely monotonically correlated, the Spearman correlation coefficient is +1 or -1. It is a mature and reliable correlation analysis method.

[0004] At present, there are relatively few studies on the potential connection information mining between apparent deformation and deep deformation monitoring data, and it is necessary to solve the problems of non-fixed monitoring frequency and non-uniform monitoring time of different types of data. At the same time, in the existing research on the potential connection between apparent deformation data and deep deformation data, the spatial variability and timeliness characteristics of data series are often ignored, and there are few studies on the identification of important locations and key time nodes of spatial variation of the connection between data. Summary of the invention

[0005] To achieve the above objectives, this application provides the following technical solutions:

[0006] According to a first aspect of the present invention, the present invention claims a method for identifying landslide sliding patterns based on a multi-level data matrix, comprising:

[0007] Perform landslide safety monitoring, collect in-situ apparent displacement time series and in-situ deep displacement time series, and construct an apparent displacement time series matrix and a deep displacement time series matrix based on the in-situ apparent displacement time series and in-situ deep displacement time series;

[0008] Based on the apparent displacement time series matrix and the deep displacement time series matrix, a primary data matrix is ​​obtained through measuring point position identification;

[0009] Obtaining a time window length factor, combining the first-level data matrix, to obtain a second-level data matrix;

[0010] Obtaining a data analysis sub-cycle, combining the secondary data matrix, and obtaining a tertiary data matrix;

[0011] Performing correlation analysis on the three-level data matrix to obtain an apparent depth correlation matrix;

[0012] According to the apparent depth correlation matrix, the correlation between the apparent displacement and the deep displacement is qualitatively judged, and the depth range of the landslide and the key time node of the sliding are given.

[0013] Furthermore, performing landslide safety monitoring, collecting in-situ apparent displacement time series and in-situ deep displacement time series, constructing an apparent displacement time series matrix and a deep displacement time series matrix based on the in-situ apparent displacement time series and in-situ deep displacement time series, further comprising:

[0014] When performing landslide safety monitoring, select the appropriate interpolation method based on the monitoring frequency, monitoring cycle, and data deformation trend of the on-site monitoring data;

[0015] The interpolation method at least comprises:

[0016] Linear interpolation, polynomial interpolation, Fourier series;

[0017] The interpolated data should meet the characteristics of surface data and deep data grouping and time series intervals.

[0018] Furthermore, based on the apparent displacement time series matrix and the deep displacement time series matrix, a primary data matrix is ​​obtained through measuring point position identification, which also includes:

[0019] The primary data matrix is ​​a submatrix of the interpolated apparent displacement time series matrix, and a new matrix is ​​formed by combining corresponding submatrices in the interpolated deep displacement time series matrix based on the principle of the same measuring point positions.

[0020] Further, obtaining a time window length factor, combining the primary data matrix to obtain a secondary data matrix, also includes:

[0021] When constructing the secondary data matrix, a suitable time window length factor t is selected according to the actual monitoring frequency, and the factor is set to the shortest duration or average duration of the time interval between two monitorings at all measuring points.

[0022] Further, obtaining a data analysis sub-cycle, combining the secondary data matrix to obtain a tertiary data matrix, also includes:

[0023] The third-level data matrix is ​​a submatrix of the second-level data matrix. The dimension of the third-level data matrix is ​​based on the data analysis sub-period l. The period setting is adjusted according to specific problems in actual engineering. The period l is an integer multiple of the time window length factor t, and at the same time satisfies l / t>=12.

[0024] Further, correlation analysis is performed on the three-level data matrix to obtain an apparent depth correlation matrix, which also includes:

[0025] The apparent depth correlation matrix is ​​a matrix constructed by calculating the Spearman correlation coefficients between the deformation data at different deep positions in the three-level data matrix and the apparent deformation in the three-level data matrix.

[0026] Furthermore, according to the apparent depth correlation matrix, the correlation between the apparent displacement and the deep displacement is qualitatively determined, and the depth range of the landslide and the key time node of the landslide are given, which also includes:

[0027] The depth range of the landslide and the key time node of the sliding are first calibrated by the correlation coefficient result to determine the sliding state, and then calculated by combining the time window length factor t and various parameters.

[0028] According to a second aspect of the present invention, the present invention claims a landslide sliding pattern recognition device based on a multi-level data matrix, comprising:

[0029] one or more processors;

[0030] A memory having one or more programs stored thereon, when the one or more programs are executed by the one or more processors, the one or more processors implement the landslide slip pattern recognition method based on a multi-level data matrix.

[0031] The present application relates to the technical field of landslide sliding pattern recognition, and in particular to a landslide sliding pattern recognition method and device based on a multi-level data matrix, obtaining the apparent displacement and deep displacement monitoring data of landslide safety monitoring, and performing time series interpolation on the monitoring data; constructing a primary data matrix of apparent displacement-deep displacement according to the principle that the positions of the apparent and deep measuring points are the same; constructing a secondary data matrix according to the actual monitoring frequency; determining the data analysis sub-cycle, and constructing a tertiary data matrix; based on the tertiary data matrix, calculating the correlation matrix between different deep positions and the apparent deformation; based on the correlation matrix, qualitatively judging the correlation between the apparent displacement and the deep displacement, and giving the depth range of the overall sliding of the landslide and the key time node of the sliding. The present invention can identify the sliding depth range and capture the key time node of the sliding, can scientifically grasp the temporal and spatial variation trend of the landslide, and provide a reference for the layout of safety management measures for the landslide. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 A workflow diagram of a landslide sliding pattern recognition method based on a multi-level data matrix as claimed in an embodiment of the present application;

[0033] Figure 2 A second working flow chart of a landslide sliding pattern recognition method based on a multi-level data matrix as claimed in an embodiment of the present application;

[0034] Figure 3 A monitoring point arrangement diagram of a landslide sliding pattern recognition method based on a multi-level data matrix as claimed in an embodiment of the present application. DETAILED DESCRIPTION

[0035] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0036] The terms "first", "second" and "third" in this application are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Thus, the features defined as "first", "second" and "third" can explicitly or implicitly include at least one of the features. In the description of this application, the meaning of "multiple" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined. All directional indications in the embodiments of the present application (such as up, down, left, right, front, back...) are only used to explain the relative position relationship, movement, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication also changes accordingly. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally also includes steps or units that are not listed, or optionally also includes other steps or units inherent to these processes, methods, products or devices.

[0037] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0038] According to a first embodiment of the present invention, the present invention claims a landslide sliding pattern recognition method based on a multi-level data matrix, referring to Figure 1 ,include:

[0039] Perform landslide safety monitoring, collect in-situ apparent displacement time series and in-situ deep displacement time series, and construct an apparent displacement time series matrix and a deep displacement time series matrix based on the in-situ apparent displacement time series and in-situ deep displacement time series;

[0040] Based on the apparent displacement time series matrix and the deep displacement time series matrix, a primary data matrix is ​​obtained through measuring point position identification;

[0041] Obtaining a time window length factor, combining the first-level data matrix, to obtain a second-level data matrix;

[0042] Obtaining a data analysis sub-cycle, combining the secondary data matrix, and obtaining a tertiary data matrix;

[0043] Performing correlation analysis on the three-level data matrix to obtain an apparent depth correlation matrix;

[0044] According to the apparent depth correlation matrix, the correlation between the apparent displacement and the deep displacement is qualitatively judged, and the depth range of the landslide and the key time node of the sliding are given.

[0045] Furthermore, performing landslide safety monitoring, collecting in-situ apparent displacement time series and in-situ deep displacement time series, constructing an apparent displacement time series matrix and a deep displacement time series matrix based on the in-situ apparent displacement time series and in-situ deep displacement time series, further comprising:

[0046] When performing landslide safety monitoring, select the appropriate interpolation method based on the monitoring frequency, monitoring cycle, and data deformation trend of the on-site monitoring data;

[0047] The interpolation method at least comprises:

[0048] Linear interpolation, polynomial interpolation, Fourier series;

[0049] The interpolated data should meet the characteristics of surface data and deep data grouping and time series intervals.

[0050] In this embodiment, the in-situ apparent displacement time series a i is a column vector with dimension k, which is the in-situ monitoring data of the apparent monitoring point numbered i, where i = 1, 2, ..., n, that is, there are n apparent monitoring data, and the apparent observation point numbered i has k observation results; the in-situ deep displacement time series b j The dimension is p*q j The matrix is ​​the in-situ monitoring data of the deep measuring point numbered j, where j = 1, 2, ..., m, that is, there are m deep deformation monitoring data, the deep measuring point numbered j has p observation results, and there are q sub-measuring points at different depths at this measuring point.

[0051] The apparent displacement time series matrix A is composed of the interpolated apparent displacement time series a i 'Constitute, A={a1',a2',…a i '…,a n '}, is a matrix with dimension t'*n, t' is the number of time points in the interpolated time series; the deep displacement time series matrix B is composed of the deep displacement time series b after interpolation j 'Constitute, B = {b1', b2', ... b j '…,b m '}, is the dimension After interpolation, the apparent data and deep data time series should have the same number of rows and the same time intervals for all data, that is, the apparent displacement time series matrix A and the deep displacement time series matrix B should have the same number of rows.

[0052] When constructing the apparent displacement time series matrix and the deep displacement time series matrix, it is necessary to add the measurement point location mark to the matrix;

[0053] The measurement point location identifier is added by adding a new row above matrix A and matrix B to indicate the measurement point location of each column of data. i A total of n measurement point location identifiers c_b of matrix B j total Surface observation points and deep measurement points located at the same position should be marked with the same measurement point location identifier.

[0054] Furthermore, based on the apparent displacement time series matrix and the deep displacement time series matrix, a primary data matrix is ​​obtained through measuring point position identification, which also includes:

[0055] The primary data matrix is ​​a submatrix of the interpolated apparent displacement time series matrix, and a new matrix is ​​formed by combining corresponding submatrices in the interpolated deep displacement time series matrix based on the principle of the same measuring point positions.

[0056] In this embodiment, a primary data matrix of apparent displacement-deep displacement is constructed, and the apparent and deep displacement data at the same position in matrices A and B are selected by judging the position identifiers of the measuring points in matrices A and B to construct a primary data matrix F k ,k=1,2,.... That is, c_a i = =c_b j When F k ={a i ',b j '}.

[0057] Further, obtaining a time window length factor, combining the primary data matrix to obtain a secondary data matrix, also includes:

[0058] When constructing the secondary data matrix, a suitable time window length factor t is selected according to the actual monitoring frequency, and the factor is set to the shortest duration or average duration of the time interval between two monitorings at all measuring points.

[0059] In this embodiment, a secondary data matrix S of apparent displacement-deep displacement is constructed. k ,k=1,2,...,n, window length factor t, S k It is a matrix with f+1 rows and 2 columns, which is used to calculate the value of f. The factor can be set as the shortest or average time interval between two monitoring times of all measuring points.

[0060] Further, obtaining a data analysis sub-cycle, combining the secondary data matrix to obtain a tertiary data matrix, also includes:

[0061] The third-level data matrix is ​​a submatrix of the second-level data matrix. The dimension of the third-level data matrix is ​​based on the data analysis sub-period l. The period setting is adjusted according to specific problems in actual engineering. The period l is an integer multiple of the time window length factor t, and at the same time satisfies l / t>=12.

[0062] In this embodiment, a three-level data matrix T of apparent displacement-deep displacement is constructed. ku , k=1,2,...,n;u=1,2,...,fl / t+1,select the secondary data matrix S according to the data analysis sub-period l k The period l must be an integer multiple of the time window length factor t, and l / t>=12.

[0063] Further, correlation analysis is performed on the three-level data matrix to obtain an apparent depth correlation matrix, which also includes:

[0064] The apparent depth correlation matrix is ​​a matrix constructed by calculating the Spearman correlation coefficients between the deformation data at different deep positions in the three-level data matrix and the apparent deformation in the three-level data matrix.

[0065] In this embodiment, the correlation matrix is ​​a three-level data matrix T ku The apparent factor vectors x in are calculated with the deep factor vectors y pairwise by Spearman correlation analysis.

[0066] Assuming the Spearman coefficient, the specific calculation formula is shown in formula (1) to formula (2);

[0067]

[0068] Among them, r kuq' is the Spearman coefficient between the apparent factor vector x and the deep factor vector y, R xi and R yi is the rank of the apparent factor vector x and the deep factor vector y, and For R xi and R yi The mean value of Note: Level difference Then convert formula (1) into:

[0069]

[0070] Where n is the number of pairs of variables; d is the difference in rank, q' is the layout number, q'=1,2,…,qj .

[0071] Furthermore, according to the apparent depth correlation matrix, the correlation between the apparent displacement and the deep displacement is qualitatively determined, and the depth range of the landslide and the key time node of the landslide are given, which also includes:

[0072] The depth range of the landslide and the key time node of the sliding are first calibrated by the correlation coefficient result to determine the sliding state, and then calculated by combining the time window length factor t and various parameters.

[0073] In this embodiment, the slip state is calibrated according to the following:

[0074] When 0.6≤r kuq' When ≤1.0, it indicates strong correlation or extremely strong correlation;

[0075] 0.0≤r kuq' When <0.6, it indicates weak or moderate correlation;

[0076] r kuq' When ==NA, it means there is no slip here.

[0077] R ku =[r ku1 ,r ku2 ,...,r kuq '];

[0078] Where q'=1,2,...,q j ;k=1,2,...,n;u=1,2,...,fl / t+1.

[0079] The depth range of the overall sliding of the landslide and the key time node of the sliding are calculated based on the time window length factor t and various parameters. The depth range of the overall sliding is depth = q'*Δd, and the key time node is time = (u-1)t+1, Δd is the self-test point layout interval of the deep test point, q' is the layout number, and u is the third-level data matrix number. k For each correlation coefficient in , when the correlation coefficient changes greatly compared with the adjacent coefficient or produces a state change, the corresponding depth and time node are the key points for calibrating the depth range of the overall sliding of the landslide and the key time node for sliding.

[0080] According to a second embodiment of the present invention, the present invention claims a landslide sliding pattern recognition device based on a multi-level data matrix, comprising:

[0081] one or more processors;

[0082] A memory having one or more programs stored thereon, when the one or more programs are executed by the one or more processors, the one or more processors implement the landslide slip pattern recognition method based on a multi-level data matrix.

[0083] like Figure 2 As shown, the three-level data matrix sliding pattern recognition method based on landslide safety monitoring data of the embodiment includes:

[0084] Step 1: First, obtain the apparent displacement and deep displacement monitoring data of landslide safety monitoring, perform time series interpolation on the monitoring data, and add the measuring point location mark to the interpolated matrix;

[0085] like Figure 3 In this embodiment, a large-volume landslide between two cascade hydropower stations in a river basin in my country is selected as a calculation and analysis case. Two apparent displacement monitoring points (BG1-1, BG3-1) and two deep displacement monitoring points (SB1-1, SB3-1) are arranged at the front edge of the landslide body. The in-situ apparent displacement time series of BG1-1 is a column vector with a dimension of 222, denoted as a1, that is, there are 222 observation results at this measuring point; the in-situ apparent displacement time series of BG3-1 is a column vector with a dimension of 200, denoted as a2, that is, there are 200 observation results at this measuring point. The SB1-1 in-situ deep displacement time series is a matrix with a dimension of 111*79, denoted as b1, that is, there are 111 observation results at this measuring point and 79 sub-measuring points in total; the SB3-1 in-situ deep displacement time series is a matrix with a dimension of 130*75, denoted as b2, that is, there are 130 observation results at this measuring point and 75 sub-measuring points in total.

[0086] Since the monitoring time interval is relatively large in the initial stage of monitoring and the monitoring frequency is increased in the middle and late stages of monitoring, the time range covered by the data is very large. This embodiment uses a linear interpolation method to ensure the authenticity of the data information used for subsequent analysis. The apparent displacement time series matrix A formed by the interpolated data is a matrix with a dimension of 3046*2, and B is a matrix with a dimension of 3046*154.

[0087] There are 2 measuring point position identifications in matrix A, BG1-1 measuring point position identification is recorded as 1, BG3-1 measuring point position identification is recorded as 2, because the surface observation points and deep measuring points located at the same position should be marked with the same measuring point position. There are also 2 measuring point position identifications in matrix B, SB1-1 measuring point is recorded as 1, and SB3-1 measuring point is recorded as 2. The final dimension of matrix A is 3047*2, and the dimension of matrix B is 3047*154.

[0088] Step 2: According to the principle that the positions of the apparent and deep measuring points are the same, a primary data matrix of apparent displacement-deep displacement is constructed; a secondary data matrix is ​​constructed based on the actual monitoring frequency; a data analysis sub-cycle is determined to construct a tertiary data matrix;

[0089] By judging the position identification of the measuring points in matrix A and matrix B, the apparent and deep displacement data at the same position in matrix A and matrix B are selected to construct a primary data matrix, and the submatrix with all identifications as 1 is selected to construct a primary data matrix F1, F1 = {a1', b1'}, with a dimension of 3047*80, and the corresponding F2 dimension is 3047*76.

[0090] After completing the classification identification, the subsequent processing method of the primary data matrix constructed based on the identification is the same. The following will take F1 constructed by position identification 1 as an example to describe the specific implementation process.

[0091] According to the actual monitoring frequency of the two types of data and the previous interpolation method, in order to accurately reflect the correlation of apparent deep deformation and avoid the pseudo-correlation caused by the interpolation method, this embodiment sets the window influence factor t=28, then f=108, and the S1 dimension is 109*76.

[0092] In order to ensure the reliability of the Spearman coefficient, in this embodiment, the data analysis sub-period l=336, the T1 u dimension is 12*80, the maximum value of u is 97, and there are 97 T1 u matrices in total.

[0093] Step 3: Based on the three-level data matrix, calculate the correlation matrix between different deep positions and apparent deformation; based on the correlation matrix, qualitatively judge the correlation between apparent displacement and deep displacement, and give the depth range of the overall sliding of the landslide and the key time nodes of sliding.

[0094] The apparent factor vector x in T1 u was analyzed with the deep factor vector y by pairwise Spearman correlation to obtain r 1uq' , according to r 1uq' The correlation matrix R1u is obtained, R1u = [r 1u1 ,r 1u2 ,...,r 1uq ]; q=1,2,...,79; u=1,2,...,97.

[0095] The specific calculation formula of Spearman coefficient is shown in formula (1) to formula (2);

[0096]

[0097] in Note: Level difference but

[0098]

[0099] Where n is the number of pairs of variables; d is the difference in rank, q' is the layout number of the deep factor vector, q' = 1, 2, ..., q j .

[0100] The slip state calibration is based on the following:

[0101] When 0.6≤r kuq' When ≤1.0, it indicates strong correlation or extremely strong correlation;

[0102] 0.0≤r kuq' When <0.6, it indicates weak or moderate correlation;

[0103] r kuq' When ==NA, it means there is no slip here.

[0104] The depth range of the overall sliding of the landslide and the key time node of sliding are calculated according to the time window length factor t and various parameters, depth = q′*Δd, time = (u-1)t+1, and the self-test point layout interval of the deep test points in this embodiment is Δd = 1.

[0105] After arranging the calculation results of this embodiment, it can be finally obtained that before January 16, 2016, the apparent deformation and deep deformation showed a strong correlation, and the landslide showed the characteristics of overall sliding within the monitoring range. From January 16, 2016 to October 22, 2016, the sliding depth was about 32-33m; from October 22, 2016, the sliding depth was about 22m, and after November 17, 2018, it was about 21m. In several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point, the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0106] In addition, each functional unit in each embodiment of the present application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of software functional units. The above is only an implementation method of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the specification and drawings of this application, or directly or indirectly used in other related technical fields, is also included in the patent protection scope of the present application.

[0107] The specific implementation methods of the invention are described in detail above, but they are only examples, and the present application is not limited to the specific implementation methods described above. For those skilled in the art, any equivalent modification or substitution of the invention is also within the scope of the present application, and therefore, the equalization, modification, and improvement made without departing from the spirit and principle of the present application should be included in the scope of the present application.

Claims

1. A landslide pattern recognition method based on multi-level data matrix, characterized in that: include: Perform landslide safety monitoring, collect in-situ apparent displacement time series and in-situ deep displacement time series, and construct an apparent displacement time series matrix and a deep displacement time series matrix based on the in-situ apparent displacement time series and in-situ deep displacement time series; Based on the apparent displacement time series matrix and the deep displacement time series matrix, a primary data matrix is ​​obtained through measuring point position identification; Obtaining a time window length factor, combining the first-level data matrix, to obtain a second-level data matrix; Obtaining a data analysis sub-cycle, combining the secondary data matrix, and obtaining a tertiary data matrix; Performing correlation analysis on the three-level data matrix to obtain an apparent depth correlation matrix; According to the apparent depth correlation matrix, the correlation between the apparent displacement and the deep displacement is qualitatively judged, and the depth range of the landslide and the key time node of the sliding are given.

2. A landslide slip pattern recognition method based on a multi-level data matrix according to claim 1, characterized in that: Performing landslide safety monitoring, collecting in-situ apparent displacement time series and in-situ deep displacement time series, constructing an apparent displacement time series matrix and a deep displacement time series matrix based on the in-situ apparent displacement time series and in-situ deep displacement time series, and also including: When performing landslide safety monitoring, select the appropriate interpolation method based on the monitoring frequency, monitoring cycle, and data deformation trend of the on-site monitoring data; The interpolation method at least comprises: Linear interpolation, polynomial interpolation, Fourier series; The interpolated data should meet the characteristics of surface data and deep data grouping and time series intervals.

3. A landslide slip pattern recognition method based on a multi-level data matrix according to claim 1, characterized in that: Based on the apparent displacement time series matrix and the deep displacement time series matrix, a primary data matrix is ​​obtained through measuring point position identification, and further includes: The primary data matrix is ​​a submatrix of the interpolated apparent displacement time series matrix, and a new matrix is ​​formed by combining corresponding submatrices in the interpolated deep displacement time series matrix based on the principle of the same measuring point positions.

4. The landslide pattern recognition method based on a multi-level data matrix according to claim 1 is characterized in that: Obtaining the time window length factor, combining the primary data matrix to obtain the secondary data matrix, also includes: When constructing the secondary data matrix, a suitable time window length factor t is selected according to the actual monitoring frequency, and the factor is set to the shortest duration or average duration of the time interval between two monitorings at all measuring points.

5. The landslide pattern recognition method based on multi-level data matrix according to claim 1 is characterized in that: Acquiring a data analysis sub-cycle, combining the secondary data matrix to obtain a tertiary data matrix, further comprising: The third-level data matrix is ​​a submatrix of the second-level data matrix. The dimension of the third-level data matrix is ​​based on the data analysis sub-period l. The period setting is adjusted according to specific problems in actual engineering. The period l is an integer multiple of the time window length factor t, and at the same time satisfies l / t>=12.

6. The landslide pattern recognition method based on multi-level data matrix according to claim 1 is characterized in that: Performing correlation analysis on the three-level data matrix to obtain an apparent depth correlation matrix also includes: The apparent depth correlation matrix is ​​a matrix constructed by calculating the Spearman correlation coefficients between the deformation data at different deep positions in the three-level data matrix and the apparent deformation in the three-level data matrix.

7. The landslide pattern recognition method based on multi-level data matrix according to claim 1 is characterized in that: According to the apparent depth correlation matrix, the correlation between the apparent displacement and the deep displacement is qualitatively judged, and the depth range of the landslide and the key time node of the landslide are given, which also includes: The depth range of the landslide and the key time node of the sliding are first calibrated by the correlation coefficient result to determine the sliding state, and then calculated by combining the time window length factor t and various parameters.

8. A landslide pattern recognition device based on a multi-level data matrix, characterized in that: include: one or more processors; A memory having one or more programs stored thereon, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement a landslide slip pattern recognition method based on a multi-level data matrix according to any one of claims 1 to 7.

Images