An information fusion method for slope stability estimation

Abstract

The present invention discloses an information fusion method for slope stability estimation. The present invention monitors the changes in groundwater level and surface crack width based on liquid level meters and crack meters, combines them into characteristic variables, uses a clustering algorithm to obtain a reference value set of the characteristic variables, and constructs a corresponding slope stability grade reference confidence distribution. After obtaining the monitoring sample online, the reference confidence vector is fused according to the activation degree to obtain the fused stability confidence distribution, and finally the stability of the slope is determined according to the confidence maximization principle. The present invention constructs the influencing factors affecting the stability of the slope into a multidimensional characteristic variable, and the characteristic variable will match all reference value sets to obtain a reference confidence distribution. The reference confidence distribution is corrected and recursively fused to achieve slope stability estimation. This method improves the limitations of traditional assessment and alarm methods that rely on single monitoring data and can only match local reference values and cannot fully reflect the movement state of the slope.

Description

Technical Field

[0001] The invention belongs to the field of slope stability monitoring and alarming, and relates to an information fusion method for slope stability estimation. Background Art

[0002] Slope stability refers to the stability of slope rock and soil under certain conditions of height and angle. Unstable natural slopes and artificial slopes with excessively steep angles often slide or collapse under the weight of the rock and soil, water pressure, vibration, and other external forces. Large-scale slope rock and soil failure can cause traffic disruptions, building collapses, river blockages, and reservoir siltation, resulting in significant loss of life and property. The purpose of slope stability research is to predict the timing, scale, and severity of slope failure, enabling proactive preventive measures to mitigate geological hazards and ensure safe and economical artificial slope design. Therefore, slope stability monitoring and alarming are of great significance.

[0003] All slope instability involves the failure of slope rock and soil under shear stress. Therefore, factors affecting shear stress and the shear strength of rock and soil will affect slope stability. For example, the relationship between the occurrence of discontinuities such as faults, bedding planes, and unconformities on the slope and the slope inclination and dip angle; changes in slope size and shape; and fluctuations in groundwater levels within the slope rock and soil can all alter slope stability to some extent. Therefore, research on information fusion methods for slope stability estimation can be conducted by integrating changes in groundwater levels and surface crack widths to estimate slope stability, achieving early warning and preventing significant losses caused by landslides. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the present invention proposes an information fusion method for slope stability estimation.

[0005] The specific technical solutions of the present invention are as follows:

[0006] S1: Install a level meter and a crack meter on the slope to monitor the changes in the groundwater level x1 and the width of the surface crack x2 of the landslide body respectively;

[0007] S2: Combine x1 and x2 in S1 into a two-dimensional feature variable [x1, x2]. After obtaining the historical monitoring samples of [x1, x2], use the k-means clustering algorithm to obtain the reference value set of [x1, x2] and construct the corresponding reference confidence distribution of the slope stability grade.

[0008] S3: After the monitoring sample [x1(t), x2(t)] is obtained online, it is brought into the slope stability grade reference confidence distribution constructed in S2, the reference confidence vector is activated, and the reference confidence vector is fused according to the activation degree to obtain the fused stability confidence distribution;

[0009] S4: According to the principle of maximum reliability, the stability grade of the slope is determined from the reliability distribution obtained in S3.

[0010] Furthermore, step S1 monitors the changes in the groundwater level x1 and the surface crack width x2 of the landslide body through a liquid level meter and a crack meter, which are respectively recorded as X1 = {x1(k)|k = 1, 2, ...., K} and X2 = {x1(k)|k = 1, 2, ...., K}, where k is the sampling time and a total of K samples are collected.

[0011] Furthermore, the k-means clustering algorithm designed in step S2 obtains a reference value set of the two-dimensional characteristic variable [x1, x2] and constructs a corresponding reference confidence distribution of the slope stability grade, which is specifically implemented as follows:

[0012] S2-1: Combine x1 and x2 in S1 into a two-dimensional characteristic variable X = [x1, x2], and record the obtained historical monitoring sample set as S = {[x1(k), x2(k), y(k)]|k = 1, 2, ...., K}, where y is the slope stability grade corresponding to the characteristic variable [x1, x2], and the reference value set of y is Y = {y n |n=1,2,3}, where y1 represents "stable", y2 represents "moderately unstable", and y3 represents "very unstable". Extract [x1(k),x2(k)] from S to form the historical feature sample set S x ={[x1(k),x2(k)]|k=1,2,....,K}. For the historical feature sample set S x Perform K-means clustering on the samples in to obtain the cluster center set here is the number of cluster centers of the sample; let Then the historical feature sample set S x The reference value set is A={[A 1,1 ,A 2,1 ],[A 1,2 ,A 2,2 ],...,[A 1,J ,A 2,J ]},

[0013] S2-2: Use formula (1) to calculate the characteristic variables [x1(k), x2(k)] and the reference vector [A1,j ,A 2,j ]'s Euclidean distance d j (k), use formula (2) for all d j (k) is normalized to obtain the characteristic variables [x1(k), x2(k)] and the reference vector [A 1,j ,A 2,j ]'s matching degree a j (k)

[0014]

[0015]

[0016] Thus, the relationship between the characteristic variables [x1(k), x2(k)] and the stability level y(k) is equivalently converted into a reference value set A of [x1, x2] = {[A 1,j ,A 2,j ]|j=1,...J} and the reference value set Y={y n |n=1,2,3}. For the feature variable [x1(k),x2(k)], it will match each set of reference values in the reference value set A, and its similarity distribution with respect to the reference value set A is:

[0017] C([x1(k),x2(k)])={([A 1,j ,A 2,j ],α j (k))|j=1,…,J} (3)

[0018] S2-3: Based on the reference value set A and similarity distribution C obtained in steps S2-1 and S2-2, a reference confidence distribution of the slope stability grade is constructed as follows:

[0019] e j ={(y n ,β n,j )|j=1,...J;n=1,2,3} (4)

[0020] Where β n,j Indicates that when the characteristic variable takes the reference value [A 1,j ,A 2,j ], the stability level is y n The reliability of

[0021]

[0022]

[0023] where α n,j Indicates that [x1(k), x2(k), y(k)] matches [A1,j ,A 2,j ] and y(k)=y n All similarities α j (k) and.

[0024] Furthermore, in step S3, the classification result of slope stability is calculated using the confidence fusion rule, and the specific process is as follows:

[0025] S3-1: After obtaining the monitoring sample [x1(t), x2(t)] online, the matching degree α between [x1(t), x2(t)] and all reference value sets is calculated by equation (2) in step S2-2. j (t), and use the matching degree to modify the reference reliability distribution in formula (4) to obtain

[0026] e j ′={(y n ,β′ n,j )|j=1,...J;n=1,2,3} (7)

[0027] where β′ n,j =α j (t)β n,j .

[0028] S3-2: Use the belief fusion rule to distribute the J reference beliefs matching [x1(t), x2(t)] j ′ is fused to obtain the fusion confidence distribution e′ at the current time t (J) ={(y n ,β n,(J) )|n=1,2,3}, the specific steps are as follows:

[0029] (1) Fusion of e1′ and e2′ to obtain e′ (2) ={(y n ,β n,(2) )|n=1,2,3}, where

[0030] (2) Then e′ (2) Fusion with e3′ to obtain e′ (3) ={(y n ,β n,(3) )|n=1,2,3}, where

[0031] (3) Similarly, the fusion is repeated J-3 times to obtain the final fusion confidence distribution e′ (J) ={(y n ,β n,(J) )|n=1,2,3}, where

[0032] Furthermore, in step S4, the reliability maximization principle is determined, and the stability level of the slope is determined according to the reliability distribution obtained in step S3-2. The specific process is as follows:

[0033] For e′ (J) ={(y n ,β n,(J) )|n=1,2,3}, if β 1,(J) =max{β 1,(J) ,β 2,(J) ,β 3,(J)}, then the slope stability level corresponding to [x1(t),x2(t)] is “stable”; if β 2,(J) =max{β 1,(J) ,β 2,(J) ,β 3,(J)}, then the slope stability level corresponding to [x1(t),x2(t)] is “moderately unstable”; if β 3,(J) =max{β 1,(J) ,β 2,(J) ,β 3,(J)}, then the slope stability level corresponding to [x1(t),x2(t)] is “very unstable”; if β 1,(J) =β 2,(J) >β 3,(J) , then the slope stability level corresponding to [x1(t),x2(t)] is “moderately unstable”; if β 1,(J) =β 3,(J) >β 2,(J) , then the slope stability level corresponding to [x1(t),x2(t)] is “very unstable”; if β 2,(J) =β 3,(J) >β 1,(J) , then the slope stability level corresponding to [x1(t),x2(t)] is “very unstable”.

[0034] The beneficial effect of the present invention lies in that the information fusion method of the present invention is mainly reflected in that it constructs the influencing factors affecting slope stability into multidimensional characteristic variables, and the characteristic variables will match all reference value sets to obtain a reference confidence distribution, and the reference confidence distribution will be corrected and recursively fused to achieve slope stability estimation. This method improves the limitations of traditional assessment and alarm methods that rely on single monitoring data, can only match local reference values, and cannot fully reflect the slope movement state. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. By referring to the drawings, the features and advantages of the present invention will be more clearly understood. The drawings are schematic and should not be understood as limiting the present invention in any way. Those skilled in the art can derive other drawings based on these drawings without inventive effort. Among them:

[0036] Figure 1 This is a flowchart of the information fusion method for slope stability estimation of the present invention;

[0037] Figure 2 is a graph showing changes in groundwater level and crack width in an embodiment of the method of the present invention;

[0038] Figure 3 1 is a graph showing the slope stability estimation results in an embodiment of the method of the present invention. DETAILED DESCRIPTION

[0039] In order to more clearly understand the above-mentioned objects, features and advantages of the present invention, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0040] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the scope of protection of the present invention is not limited to the specific embodiments disclosed below.

[0041] like Figure 1 As shown, the present invention provides a slope stability estimation method that integrates the changes in multiple slope influencing factors, particularly an information-fused slope stability estimation method for factors such as groundwater level and crack width. Based on the changes in groundwater level and surface crack width obtained through monitoring by a liquid level meter and a crack meter, the present invention combines these changes into characteristic variables. A clustering algorithm is used to obtain a reference value set for the characteristic variables, and a corresponding reference confidence distribution for the slope stability grade is constructed. After obtaining monitoring samples online, the reference confidence vectors are fused according to the activation level to obtain a fused stability confidence distribution. Ultimately, the slope stability grade is determined according to the confidence maximization principle.

[0042] In order to facilitate understanding of the above technical solutions of the present invention, the above technical solutions of the present invention are described in detail below through specific embodiments.

[0043] Example:

[0044] An information fusion method for slope stability estimation, the flow chart of which is as follows Figure 1 As shown, it includes the following steps:

[0045] S1: Install a level meter and a crack meter on the slope to monitor the changes in the groundwater level x1 and the width of the surface crack x2 of the landslide body respectively;

[0046] S2: Combine x1 and x2 in S1 into a two-dimensional feature variable [x1, x2]. After obtaining the historical monitoring samples of [x1, x2], use the k-means clustering algorithm to obtain the reference value set of [x1, x2] and construct the corresponding reference confidence distribution of the slope stability grade.

[0047] S3: After the monitoring sample [x1(t), x2(t)] is obtained online, it is brought into the slope stability grade reference confidence distribution constructed in S2, the activated reference confidence vector, and the reference confidence vector is fused according to the activation degree to obtain the fused stability confidence distribution;

[0048] S4: According to the principle of maximum reliability, the stability grade of the slope is determined from the reliability distribution obtained in S3.

[0049] Furthermore, step S1 monitors the changes in the groundwater level x1 and the surface crack width x2 of the landslide body through a liquid level meter and a crack meter, which are respectively recorded as X1 = {x1(k)|k = 1, 2, ...., K} and X2 = {x1(k)|k = 1, 2, ...., K}, where k is the sampling time and a total of K samples are collected.

[0050] By installing liquid level meters and crack meters on the slope, the changes in the groundwater level and surface crack width of the landslide body were monitored in real time. A total of K = 1355 samples were collected. The changes in groundwater level and crack width are as follows: Figure 2 shown.

[0051] Furthermore, the k-means clustering algorithm designed in step S2 obtains a reference value set of the two-dimensional characteristic variable [x1, x2] and constructs a corresponding reference confidence distribution of the slope stability grade, which is specifically implemented as follows:

[0052] S2-1: Combine x1 and x2 in S1 into a two-dimensional characteristic variable X = [x1, x2], and record the obtained historical monitoring sample set as S = {[x1(k), x2(k), y(k)]|k = 1, 2, ...., K}, where y is the slope stability grade corresponding to the characteristic variable [x1, x2], and the reference value set of y is Y = {y n |n=1,2,3}, where y1 represents "stable", y2 represents "moderately unstable", and y3 represents "very unstable". Extract [x1(k),x2(k)] from S to form the historical feature sample set S x={[x1(k),x2(k)]|k=1,2,....,K}. For the historical feature sample set S x Perform K-means clustering on the samples in to obtain the cluster center set here is the number of cluster centers of the sample; let Then the historical feature sample set S x The reference value set is A={[A 1,1 ,A 2,1 ],[A 1,2 ,A 2,2 ],...,[A 1,J ,A 2,J ]},

[0053] Specifically, the k-means clustering algorithm described in step S2 obtains a reference value set of the two-dimensional feature variable [x1, x2], the number of cluster centers is 3, S x The reference value set of is A = {[0.718, 18.163], [1.114, 18.318], [2.344, 19.423], [3.184, 19.631], [3.386, 19.716]}, and the reference value set of y is Y = {y1, y2, y3}.

[0054] S2-2: Use formula (1) to calculate the characteristic variables [x1(k), x2(k)] and the reference vector [A 1,j ,A 2,j ]'s Euclidean distance d j (k), use formula (2) for all d j (k) is normalized to obtain the characteristic variables [x1(k), x2(k)] and the reference vector [A 1,j ,A 2,j ]'s matching degree a j (k)

[0055]

[0056]

[0057] Thus, the relationship between the characteristic variables [x1(k), x2(k)] and the stability level y(k) is equivalently converted into a reference value set A of [x1, x2] = {[A 1,j ,A 2,j ]|j=1,...J} and the reference value set Y={y n|n=1,2,3}. For the feature variable [x1(k),x2(k)], it will match each set of reference values in the reference value set A, and its similarity distribution with respect to the reference value set A is:

[0058] C([x1(k),x2(k)])={([A 1,j ,A 2,j ],α j (k))|j=1,...,J} (3)

[0059] S2-3: Based on the reference value set A and similarity distribution C obtained in steps S2-1 and S2-2, a reference confidence distribution of the slope stability grade is constructed as follows:

[0060] e j ={(y n ,β n,j )|j=1,...J;n=1,2,3} (4)

[0061] Where β n,j Indicates that when the characteristic variable takes the reference value [A 1,j ,A 2,j ], the stability level is y n The reliability of

[0062]

[0063]

[0064] where α n,j Indicates that [x1(k), x2(k), y(k)] matches [A 1,j ,A 2,j ] and y(k)=y n All similarities α j (k) and.

[0065] Specifically, in step S2, the reference confidence distribution of the slope stability grade is constructed according to the reference value set A and the similarity distribution C as shown in Table 1.

[0066] Table 1 Reference reliability distribution table of characteristic variables [x1(k), x2(k)]

[0067]

[0068] Furthermore, the specific process of step S3 is as follows:

[0069] S3-1: In step S3, when the classification result of slope stability is calculated using the confidence fusion rule, after the monitoring sample [x1(t), x2(t)] is obtained online, the matching degree α between [x1(t), x2(t)] and all reference value sets is calculated by formula (2) in step S2-2. j (t), and use the matching degree to modify the reference reliability distribution in formula (4) to obtain

[0070] e j ′={(y n ,β′ n,j )|j=1,...J;n=1,2,3} (7)

[0071] where β′ n,j =α j (t)β n,j .

[0072] Specifically, after obtaining the monitoring sample [x1(t), x2(t)] = [1.407, 18.511] online, the matching degrees α1(t) to α5(t) of [x1(t), x2(t)] with all reference value sets are calculated by formula (2), which are α1(t) = 0.155, α2(t) = 0.753, α3(t) = 0.054, α4(t) = 0.021, and α5(t) = 0.017 respectively; the reference reliability distribution in Table 1 is corrected by using the matching degree, and the corrected reference reliability e can be obtained by formula (7) ′1={(y1,0.134),(y2,0.020),(y3,0.001)}, e′2={(y1,0.654),(y2,0.095),(y3,0.004)}, e′3={(y1,0.001),(y2, 0.051),(y3,0.002)}, e′4={(y1,0.001),(y2,0.004),(y3,0.017)}, e′5={(y1,0.001),(y2,0.003),(y3,0.014)}.

[0073] S3-2: Use the belief fusion rule to distribute the J reference beliefs matching [x1(t), x2(t)] j ′ is fused to obtain the fusion confidence distribution e′ at the current time t (J) ={(y n ,β n,(J) )|n=1,2,3}, the specific steps are as follows:

[0074] (1) Fusion of e1′ and e2′ to obtain e′ (2) ={(y n ,β n,(2) )|n=1,2,3}, where

[0075] (2) Then e′ (2) Fusion with e3′ to obtain e′ (3) ={(y n ,β n,(3) )|n=1,2,3}, where

[0076] (3) Similarly, the fusion is repeated J-3 times to obtain the final fusion confidence distribution e′ (J) ={(y n ,β n,(J) )|n=1,2,3}, where

[0077] Specifically, the five reference reliability distributions e that match [x1(t), x2(t)] are transformed into j ′ to fuse, first fuse e1′ and e2′ to get e′ (2) ={(y1,0.874),(y2,0.071),(y3,0.002)}; then e′ (2) Fusion with e3′ to obtain e′ (3) ={(y1,0.823),(y2,0.125),(y3,0.003)}; then e′ (3) Fusion with e4′ to obtain e′ (4) ={(y1,0.811),(y2,0.129),(y3,0.012)}; Finally, e′ (4) Combined with e5′, the final fusion confidence distribution e′ is obtained (5) ={(y1,0.799),(y2,0.133),(y3,0.020)}.

[0078] Furthermore, in step S4, the reliability maximization principle is determined, and the stability level of the slope is determined according to the reliability distribution obtained in step S3-2. The specific process is as follows:

[0079] For e′ (J) ={(y n ,β n,(J) )|n=1,2,3}, if β 1,(J) =max{β 1,(J) ,β 2,(J) ,β 3,(J)}, then the slope stability level corresponding to [x1(t),x2(t)] is “stable”; if β 2,(J) =max{β 1,(J) ,β 2,(J) ,β 3,(J)}, then the slope stability level corresponding to [x1(t),x2(t)] is “moderately unstable”; if β 3,(J) =max{β 1,(J) ,β 2,(J) ,β 3,(J)}, then the slope stability level corresponding to [x1(t),x2(t)] is “very unstable”; if β 1,(J) =β 2,(J) >β 3,(J) , then the slope stability level corresponding to [x1(t),x2(t)] is “moderately unstable”; if β 1,(J) =β 3,(J) >β 2,(J) , then the slope stability level corresponding to [x1(t),x2(t)] is “very unstable”; if β 2,(J) =β 3,(J) >β 1,(J) , then the slope stability level corresponding to [x1(t),x2(t)] is “very unstable”.

[0080] Specifically, after obtaining the monitoring sample [x1(t), x2(t)] = [1.407, 18.511] online, the five reference reliability distributions e that match [x1(t), x2(t)] are fused using the reliability fusion rule. j ′ is fused to obtain the fusion confidence distribution e′ at the current time t (5) ={(y1,0.799),(y2,0.133),(y3,0.020)}. Since max{0.799,0.133,0.020}=0.799, the slope stability level corresponding to [x1(t),x2(t)] is determined to be “stable”.

[0081] By obtaining the changes of groundwater level and crack width of disaster factors online, the slope stability estimation result is obtained through the credibility fusion rule. Since the change pattern of online samples is consistent with that of historical samples, the historical samples are used as online samples and brought into the method of the present invention. The results obtained are as follows Figure 3 shown.

Claims

1. An information fusion method for slope stability estimation, characterized in that: The following steps are included: S1: Install a level meter and a crack meter on the slope to monitor the changes in the groundwater level x1 and the width of the surface crack x2 of the landslide body respectively; S2: Combine x1 and x2 in S1 into a two-dimensional feature variable [x1, x2]. After obtaining the historical monitoring samples of [x1, x2], use the k-means clustering algorithm to obtain the reference value set of [x1, x2] and construct the corresponding reference confidence distribution of the slope stability grade. S3: After the monitoring sample [x1(t), x2(t)] is obtained online, it is brought into the slope stability grade reference confidence distribution constructed in S2, the reference confidence vector is activated, and the reference confidence vector is fused according to the activation degree to obtain the fused stability confidence distribution; S4: According to the principle of maximum reliability, the stability of the slope is determined from the stability reliability distribution obtained in S3; The k-means clustering algorithm designed in step S2 obtains a reference value set of the two-dimensional characteristic variables [x1, x2] and constructs the corresponding reference confidence distribution of the slope stability grade, as follows: S2-1: Combine x1 and x2 in S1 into a two-dimensional characteristic variable X = [x1, x2], and record the obtained historical monitoring sample set as S = {[x1(k), x2(k), y(k)]|k = 1, 2, ...., K}, where y is the slope stability grade corresponding to the characteristic variable [x1, x2], and the reference value set of y is Y = {y n |n=1,2,3}, where y1 represents "stable", y2 represents "moderately unstable", and y3 represents "very unstable"; Extract [x1(k), x2(k)] from S to form the historical feature sample set S x ={[x1(k),x2(k)]|k=1,2,....,K}; for the historical feature sample set S x Perform K-means clustering on the samples in to obtain the cluster center set here is the number of cluster centers of the sample; let Then the historical feature sample set S x The reference value set is A={[A 1,1 ,A 2,1 ],[A 1,2 ,A 2,2 ],...,[A 1,J ,A 2,J ]}, S2-2: Use formula (1) to calculate the characteristic variables [x1(k), x2(k)] and the reference vector [A 1,j ,A 2,j ]'s Euclidean distance d j (k), use formula (2) for all d j (k) is normalized to obtain the characteristic variables [x1(k), x2(k)] and the reference vector [A 1,j ,A 2,j ]'s matching degree a j (k) Thus, the relationship between the characteristic variables [x1(k), x2(k)] and the stability level y(k) is equivalently converted into a reference value set A of [x1, x2] = {[A 1,j ,A 2,j ]|j=1,...J} and the reference value set Y={y n |n=1,2,3}; for the feature variable [x1(k),x2(k)], it will match each set of reference values in the reference value set A, and its similarity distribution with respect to the reference value set A is: C([x1(k),x2(k)])={([A 1,j ,FLUENT 2,j ],α j (k))|j=1,…,J} (3) S2-3: Based on the reference value set A and similarity distribution C obtained in steps S2-1 and S2-2, a reference confidence distribution of the slope stability grade is constructed as follows: e j {(y n ,β n,j )|j=1,...J=n=1,2,3} (4) Where β n,j Indicates that when the characteristic variable takes the reference value [A 1,j ,A 2,j ], the stability level is y n The reliability of where α n,j Indicates that [x1(k), x2(k), y(k)] matches [A 1,j ,A 2,j ] and y(k)=y n All similarities α j (k) and.

2. The information fusion method for slope stability estimation according to claim 1, characterized in that: In step S1, the changes of the groundwater level x1 and the surface crack width x2 of the landslide body are monitored by a liquid level meter and a crack meter, which are respectively recorded as X1 = {x1(k)|k = 1, 2, ...., K} and X2 = {x1(k)|k = 1, 2, ...., K}, where k is the sampling time and a total of K samples are collected.

3. The information fusion method for slope stability estimation according to claim 1, characterized in that: The specific process of step S3 is: S3-1: After obtaining the monitoring sample [x1(t), x2(t)] online, the matching degree α between [x1(t), x2(t)] and all reference value sets is calculated by equation (2) in step S2-2. j (t), and use the matching degree to modify the reference reliability distribution in formula (4) to obtain e j ′={(y n ,β′ n,j )|j=1,...J=n=1,2,3} (7) among them n,j =a j (t)b n,j ; S3-2: Use the belief fusion rule to distribute the J reference beliefs matching [x1(t), x2(t)] j ′ is fused to obtain the fusion confidence distribution e′ at the current time t (J) ={(y n ,β n,(J) )|n=1,2,3}, the specific steps are as follows: (1) Fusion of e1′ and e2′ to obtain e′ (2) ={(y n ,β n,(2) )|n=1,2,3}, where (2) Then e′ (2) Fusion with e′3 to obtain e′ (3) ={(y n ,β n,(3) )|n=1,2,3}, where (3) Similarly, the fusion is repeated J-3 times to obtain the final fusion confidence distribution e′ (J) ={(y n ,β n,(J) )|n=1,2,3}, where 4. The information fusion method for slope stability estimation according to claim 3 is characterized in that: The specific process of step S4 is: For e′ (J) ={(y n ,β n,(J) )|n=1,2,3}, if β 1,(J) =max{β 1,(J) ,β 2,(J) ,β 3,(J) }, then the slope stability level corresponding to [x1(t),x2(t)] is "stable"; if β 2,(J) =max{β 1,(J) ,β 2,(J) ,β 3,(J) }, then the slope stability level corresponding to [x1(t),x2(t)] is "moderately unstable"; if β 3,(J) =max{β 1,(J) ,β 2,(J) ,β 3,(J) }, then the slope stability level corresponding to [x1(t),x2(t)] is "very unstable"; if β 1,(J) =β 2,(J) >β 3,(J) , then the slope stability level corresponding to [x1(t),x2(t)] is "moderately unstable"; if β 1,(J) =β 3,(J) >β 2,(J) , then the slope stability level corresponding to [x1(t),x2(t)] is "very unstable"; if β 2,(J) =β 3,(J) >β 1,(J) , then the slope stability level corresponding to [x1(t),x2(t)] is "very unstable".

Images