Slope instability influence range prediction method for mountainous traffic corridor based on slope constraint

Abstract

The present application belongs to the technical field of machine learning, and specifically discloses a mountainous traffic corridor slope instability influence range prediction method based on slope constraint, constructs a slope coordinate system, introduces slope direction, slope, roughness, vegetation coverage and engineering disturbance intensity and other slope control factors, establishes a slope propagation impedance function, projects and modulates the numerical simulation or historical landslide influence range in the slope direction, and realizes the constraint expression of the landslide influence range in the real slope direction. On this basis, combined with the topographic and geomorphic factors, engineering disturbance factors and inducing factors along the traffic corridor, a training data set is constructed, and a slope instability influence range prediction model based on slope characteristics is trained to quickly predict the influence range parameters of slope instability. The present application can significantly improve the direction consistency and spatial accuracy of the prediction of the influence range of the slope instability of the mountainous traffic corridor, and provides technical support for traffic safety prevention and control and disaster warning.

Description

Technical Field

[0001] This invention belongs to the field of machine learning technology, specifically relating to a method for predicting the impact range of slope instability in mountainous transportation corridors based on slope constraints. Background Technology

[0002] In the construction of transportation engineering projects in mountainous areas, the terrain along the transportation corridors is highly undulating, with concentrated slope aspects and intense engineering disturbances, leading to frequent slope instability disasters. Slope instability can not only cause road blockages and vehicle damage, but also trigger secondary disasters such as landslides and debris flows, seriously impacting traffic safety and regional economic development. Accurately predicting the impact range of slope instability has become a key technical issue in the prevention and control of disasters in transportation corridors, and is of great significance for improving the safe operation capability of transportation engineering projects in mountainous areas and the level of geological disaster early warning.

[0003] Currently, methods for predicting the impact range of slope instability are mainly classified into four categories: statistical models, empirical formulas, numerical simulations, and machine learning-based surrogate models. Statistical models establish regression relationships between influencing factors and the impact range based on historical disaster data; empirical formulas provide empirical distances through simplified mechanical analysis; numerical simulations, such as discrete element or finite element methods, can simulate the entire process of landslide movement; and machine learning surrogate models achieve rapid predictions by learning from numerical simulation samples. In practical applications, researchers often use a planar coordinate system to express the accumulation range of a landslide, describing the area that may be affected by the landslide by setting buffer zones or calculating the maximum horizontal distance. Some methods introduce aspect correction, which involves simply rotating the impact range according to the slope direction to make the prediction results more closely match the actual terrain.

[0004] However, existing technologies still have significant limitations in mountainous transportation corridor scenarios. Traditional methods lack slope orientation constraints and often express landslide impact ranges in a planar coordinate system, leading to significant deviations between predicted results and actual landslide propagation directions, especially in canyon-type corridors with uniform slope aspect. Existing methods fail to effectively model the influence of engineering disturbances and inducing factors on propagation direction, neglecting the spatial heterogeneity of disturbance factors. Numerical simulations are typically based on a single slope or local area, making it difficult to form directionally consistent impact range samples at the transportation corridor scale. This results in insufficient training data for surrogate models or significant directional biases, making it difficult to directly use numerical simulation results for area prediction at the transportation corridor scale. Transportation corridors are distributed in a strip-like pattern, and landslide impact ranges exhibit clear directionality and channel characteristics. Existing methods often use planar buffer zones or empirical distances for description, failing to directly reflect the propagation characteristics of slope orientation. Summary of the Invention

[0005] The purpose of this invention is to address the problems of large topographic relief along the corridor, strong engineering disturbances, and the difficulty of accurately depicting the actual landslide propagation direction using traditional methods. This invention proposes a method for predicting the impact range of slope instability in mountainous transportation corridors based on slope constraints.

[0006] The technical solution of this invention is as follows: A method for predicting the impact range of slope instability in mountainous transportation corridors based on slope constraints includes: S1: Obtain topographic data, engineering disturbance data, and inducing factor data for each instability assessment unit along the mountain transportation corridor; construct the slope coordinate system and slope propagation impedance function for each instability assessment unit based on the topographic data and the engineering disturbance data; the inducing factor data includes at least one of rainfall, traffic vibration, seismic motion, and pore water pressure changes. S2: Extract the slope coordinate system features, propagation impedance features, and inducing factor features of each instability evaluation unit as input features of the training samples. Use the slope coordinate system and the slope propagation impedance function to project the historical landslide influence range or numerical simulation landslide influence range of each instability evaluation unit onto the slope constraint. Obtain the influence range parameters under the slope coordinate system as the output labels of the training samples to form a training sample set of slope constraints. S3: Use the training sample set to train an artificial intelligence agent model to obtain a prediction model for the impact range of slope instability; S4: Input the slope coordinate system characteristics, propagation impedance characteristics, and inducing factor characteristics of each instability evaluation unit along the traffic corridor to be predicted into the slope instability influence range prediction model to obtain the influence range parameters and influence probability of each instability evaluation unit in the slope coordinate system; the influence range parameters include the influence distance, influence width, and influence direction in the slope coordinate system. S5: Reconstruct the polygon of the influence range in the slope direction based on the influence range parameters, and determine the predicted influence range of slope instability.

[0007] Preferably, in S1, constructing the slope coordinate system includes: Based on the topographic data, the slope and aspect of each grid point in the instability evaluation unit are determined, and the aspect vector corresponding to each grid point is determined based on the slope and aspect. The corresponding slope normal vector is determined for each grid point through local curvature analysis. The slope vector is determined based on the aspect vector and the slope normal vector. The slope coordinate system is formed based on the aspect vector, the slope vector, and the slope normal vector of each grid point.

[0008] Preferably, in S1, constructing the slope propagation impedance function includes: Slope, roughness, and vegetation cover are extracted from the topographic data, and engineering disturbance intensity is extracted from the engineering disturbance data. The slope propagation impedance function is constructed based on the slope, roughness, vegetation cover, and engineering disturbance intensity. The slope propagation impedance function is expressed as follows: ; in: For slope, For roughness, For vegetation coverage, For engineering disturbance intensity, For slope impedance, For roughness impedance, For vegetation impedance, and These are the weighting coefficients for slope impedance, roughness impedance, vegetation impedance, and engineering disturbance intensity, respectively.

[0009] Preferably, in S2, the slope constraint projection includes: Transform the point set of the historical landslide influence range or the point set of the numerical simulation landslide influence range to the slope coordinate system; Obtain the projected distance of each point along the slope aspect vector in the slope coordinate system; The projection distance is modulated according to the slope propagation impedance function to obtain the slope constraint projection distance; The coordinates of each point in the slope coordinate system are obtained based on the slope constraint projection distance, forming a set of points representing the landslide influence range in the slope coordinate system.

[0010] Preferably, the influence range parameters include: influence distance, influence width, and influence direction.

[0011] Preferably, the slope coordinate system features include a slope vector, a slope aspect vector, and a slope normal vector; The propagation impedance characteristics include slope propagation impedance value, slope impedance, roughness impedance, vegetation impedance, and engineering disturbance intensity; The inducing factors include at least one of rainfall intensity, traffic vibration amplitude, peak ground acceleration, and pore water pressure variation.

[0012] Preferably, before training the AI ​​agent model using the training sample set, the method further includes: Calculate the slope consistency index for each training sample in the training sample set, and remove training samples with a slope consistency index lower than a preset threshold to obtain the preprocessed training sample set.

[0013] Preferably, the slope consistency index is calculated according to the following formula: ; In the formula, This represents the average slope propagation impedance actually used during the slope constraint projection process. The slope propagation impedance value is obtained by inversion based on historical landslide samples.

[0014] Preferably, the method further includes: A risk level map of slope instability for transportation corridors is constructed based on the predicted impact range of slope instability, and the potential road blockage risk is determined based on the interaction between the predicted impact range of slope instability and the road centerline.

[0015] Preferably, the road blockage risk includes blockage probability, blockage location, and blockage level, wherein the blockage level is determined based on the blockage probability and a preset level threshold. If the predicted impact range of slope instability covers the road, the blocking probability is the same as the impact probability; if the predicted impact range of slope instability does not cover the road, the blocking probability is 0.

[0016] The beneficial effects of this invention are: 1. The method for predicting the impact range of slope instability in mountain transportation corridors based on slope constraints of the present invention constructs a slope coordinate system composed of slope aspect vector, slope gradient vector and slope normal vector, and transforms the expression of the impact range in the traditional plane coordinate system into the expression of slope direction, so that the landslide propagation direction matches the actual terrain slope aspect height, effectively solving the technical problem of large prediction deviation of landslide propagation direction in mountain transportation corridors.

[0017] 2. The method for predicting the impact range of slope instability in mountain transportation corridors based on slope constraints proposed in this invention utilizes a slope propagation impedance function that comprehensively considers multiple factors such as slope, roughness, vegetation cover, and engineering disturbance intensity. This function can quantitatively characterize the energy dissipation and resistance changes during landslide propagation along the slope. By projecting historical or simulated landslide results onto the slope coordinate system and performing impedance modulation through slope constraint, the accuracy of the predicted impact range is improved. Introducing the engineering disturbance intensity into the slope propagation impedance function, and simultaneously inputting the characteristics of inducing factors as independent features into the surrogate model, enables the prediction model to reflect the dynamic impact of human engineering activities and environmental changes on the landslide propagation path. Attached Figure Description

[0018] Figure 1 The diagram shows a flowchart of a method for predicting the impact range of slope instability in mountain transportation corridors based on slope constraints. Figure 2 The diagram shown illustrates the construction of the slope coordinate system. Figure 3The diagram shows the construction of the slope propagation impedance function; Figure 4 The diagram shown is a flowchart of slope constraint projection. Detailed Implementation

[0019] Exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be understood that the embodiments shown and described in the drawings are merely exemplary and are intended to illustrate the principles and spirit of the invention, and are not intended to limit the scope of the invention.

[0020] This invention provides a method for predicting the impact range of slope instability in mountainous transportation corridors based on slope constraints. The core of this method lies in constructing a slope coordinate system and introducing slope control factors such as aspect, gradient, roughness, vegetation cover, and engineering disturbance intensity. A slope propagation impedance function is established, and the impact range of numerically simulated or historical landslides is projected onto the slope direction and modulated with impedance to achieve a constrained expression of the landslide impact range in the actual slope direction. Based on this, a training dataset is constructed by combining topographic factors, engineering disturbance factors, and inducing factors along the transportation corridor. This training dataset is then used to train a slope instability impact range prediction model for rapidly predicting the impact range parameters of slope instability.

[0021] Please see Figure 1 , Figure 1 The diagram shows a flowchart of a method for predicting the impact range of slope instability in mountainous transportation corridors based on slope constraints; Figure 1 As shown in this embodiment, the method for predicting the impact range of slope instability in mountain transportation corridors based on slope constraints includes the following steps: S1: Obtain topographic data, engineering disturbance data, and inducing factor data for each instability assessment unit along the mountain transportation corridor. Based on the topographic data and engineering disturbance data, construct the slope coordinate system and slope propagation impedance function for each instability assessment unit.

[0022] In this embodiment, the instability evaluation unit is a strip-shaped slope unit constructed along the traffic corridor, which is jointly determined by the first watershed on both sides of the road and the slope unit. It is a slope unit with hydrological significance and slope consistency.

[0023] Understandably, the width of the instability assessment unit is a hydrologically driven adaptive width, automatically determined by terrain constraints: based on the road centerline and bounded by the first watershed on both sides of the road. The farther the watershed is from the road, the wider the instability assessment unit; the steeper the terrain, the closer the watershed is, and the narrower the instability assessment unit.

[0024] For example, topographic data includes topographic factors such as digital elevation model (DEM), slope, aspect, and curvature; engineering disturbance data includes engineering disturbance factors such as road cut excavation, embankment filling, drainage system, and retaining wall structure; and induced factor data includes induced factors such as rainfall, traffic vibration, seismic motion, and pore water pressure changes.

[0025] In this embodiment, a slope coordinate system is constructed for each grid point of each instability evaluation unit to express the slope direction.

[0026] Optionally, the slope and aspect are calculated based on the DEM, and the slope normal vector is determined through local curvature analysis, constructing a slope coordinate system composed of the aspect vector s, the slope vector d, and the slope normal vector n. For example, the local curvature is calculated using the second derivative of the DEM, or directly extracted using ArcGIS's Curvature tool.

[0027] Please see Figure 2 , Figure 2 The diagram shows the construction of the slope coordinate system. Figure 2 As shown, constructing the slope coordinate system specifically includes: determining the slope and aspect of each grid point of the instability evaluation unit based on topographic data, determining the aspect vector corresponding to each grid point based on the slope and aspect, determining the corresponding slope normal vector for each grid point through local curvature analysis, determining the slope vector based on the aspect vector and slope normal vector, and forming the corresponding slope coordinate system based on the aspect vector, slope vector, and slope normal vector of each grid point.

[0028] It should be noted that the slope coordinate system is used to transform the landslide propagation direction from the traditional planar expression to the slope direction expression, so that the subsequent impact range prediction can truly reflect the slope geometry.

[0029] Please see Figure 3 , Figure 3 The diagram shows the construction of the slope propagation impedance function. Figure 3 As shown, the construction of the slope propagation impedance function specifically includes: extracting slope, roughness, and vegetation cover from topographic data; extracting engineering disturbance intensity from engineering disturbance data; and constructing the slope propagation impedance function based on slope, roughness, vegetation cover, and engineering disturbance intensity.

[0030] In this embodiment, the slope propagation impedance function is expressed as: ; in: For slope, For roughness, For vegetation coverage, For engineering disturbance intensity, For slope impedance, For roughness impedance, For vegetation impedance, and These are the weighting coefficients for slope impedance, roughness impedance, vegetation impedance, and engineering disturbance intensity, respectively.

[0031] For example, slope Roughness can be calculated using the ArcGIS Slope tool. Vegetation cover can be obtained based on DEM roughness or local elevation standard deviation. It could be the normalized vegetation index or the intensity of engineering disturbance. It can be calculated by superimposing normalized indicators such as cut depth, retaining wall height, and drainage ditch density, and belongs to dimensionless normalized indicators.

[0032] For example, cut depth normalization: , This represents the actual depth of the roadbed. The maximum cut depth within the instability assessment unit; retaining wall height normalized: , This is the actual height of the retaining wall. The highest retaining wall height within the instability assessment unit; normalized drainage ditch density: , The total length of the drainage ditch within the instability assessment unit. The total length of the instability evaluation unit. Engineering disturbance intensity. : , , and The weights for each normalized indicator can be set to 0.45, 0.35, and 0.20.

[0033] In this implementation, slope impedance It can be calculated using an exponential function or a piecewise linear function; roughness impedance It can be calculated based on the surface roughness index; vegetation impedance It can be calculated based on NDVI or vegetation density.

[0034] For example, , This is the slope resistance coefficient; ; , This represents the vegetation impedance coefficient.

[0035] In this embodiment, and Historical landslide samples can be used to obtain data through data-driven inversion and calibration. For example, the inversion method can be least squares or an interpretable machine learning method.

[0036] The four types of factors given by the slope propagation impedance function in this embodiment are the minimum set of factors that take into account availability, stability and representativeness in the context of a transportation corridor. In other optional embodiments, other factors can also be introduced as needed, such as factors related to geological parameters such as lithology, structural planes, and weathering degree, which are not limited here.

[0037] It should be noted that the slope propagation impedance values ​​of all instability assessment units in the mountain transportation corridor constitute a continuous slope propagation impedance field in the slope coordinate system. The slope propagation impedance field can reflect the differences in the propagation distance and energy dissipation capacity of the sliding body along the slope under different slope conditions, providing a macroscopically equivalent physical constraint for subsequent slope direction projection, without the need to explicitly solve the slope rock and soil instability dynamics process.

[0038] S2: Extract the slope coordinate system features, propagation impedance features, and inducing factor features of each instability evaluation unit as input features of the training samples. Use the slope coordinate system and slope propagation impedance function to project the historical landslide influence range or numerical simulation landslide influence range of each instability evaluation unit onto the slope constraint. Obtain the influence range parameters under the slope coordinate system as the output labels of the training samples, forming a training sample set of slope constraints.

[0039] In this embodiment, based on the topographic data, engineering disturbance data, and inducing factor data of each instability evaluation unit, the corresponding slope coordinate system features, propagation impedance features, and inducing factor features are extracted.

[0040] In this embodiment, the slope coordinate system features include the slope vector, aspect vector, and slope normal vector. Propagation impedance features include slope propagation impedance value, slope impedance, roughness impedance, vegetation impedance, and engineering disturbance intensity. Inducing factor features include at least one of rainfall intensity, traffic vibration amplitude, peak ground acceleration, and pore water pressure variation.

[0041] It is understandable that, for an instability evaluation unit, the slope coordinate system characteristics and propagation impedance characteristics of all the grid points it contains can be aggregated to obtain the slope coordinate system characteristics and propagation impedance characteristics corresponding to that instability evaluation unit.

[0042] Please see Figure 4 , Figure 4 The diagram shown is a flowchart of slope constraint projection. Figure 4 As shown, the slope constraint projection specifically includes the following steps: Step 1: Transform the point set of the historical landslide impact area or the point set of the numerical simulation landslide impact area to the slope coordinate system; Step 2: Obtain the projected distance of each point along the slope direction vector in the slope coordinate system.

[0043] Step 3: Modulate the projection distance according to the slope propagation impedance function to obtain the slope constraint projection distance.

[0044] In this embodiment, the modulation process can be represented as: ; In the formula, The projected distance is constrained by the slope. This is the projected distance, i.e., the geometric projected distance without considering the slope resistance effect. This represents the impedance field propagating on the slope.

[0045] It is understood that in other embodiments, the modulation process may also employ a more complex path integral form.

[0046] Step 4: Obtain the coordinates of each point in the slope coordinate system based on the slope constraint projection distance, and form a set of points representing the landslide influence range in the slope coordinate system.

[0047] Let the location of the landslide-affected area in the geographic coordinate system be... Relative to the slope reference point The vector is: In the slope coordinate system, this point is along the slope direction vector. The unmodulated projection distance is: The coordinates of each point in the slope coordinate system can be determined by the slope aspect component and the normal component, forming a set of points within the influence range under slope constraints.

[0048] In this embodiment, the influence range parameters in the slope coordinate system are extracted from the set of landslide influence range points in the slope coordinate system and used as the output labels for the training samples. The influence range parameters include: influence distance, influence width, and influence direction.

[0049] Understandably, slope constraint projections can be performed on the point sets of the landslide impact range under different inducing factor characteristics to form a slope constraint sample set with multiple scenarios and slope aspects.

[0050] S3: Train an artificial intelligence agent model using the training sample set to obtain a prediction model for the impact range of slope instability.

[0051] In this embodiment, before training the artificial intelligence agent model using the training sample set, the method further includes: calculating the slope consistency index of each training sample in the training sample set, removing training samples with a slope consistency index lower than a preset threshold, and obtaining a preprocessed training sample set.

[0052] In this embodiment, to improve sample quality, the Slope Consistency Index (SCI) is introduced to eliminate low-confidence samples with excessive deviations in slope direction.

[0053] Alternatively, the slope consistency index can be determined based on aspect deviation, slope deviation, or impedance consistency.

[0054] In this embodiment, impedance consistency is selected to calculate the slope consistency index. Specifically, the slope consistency index is calculated according to the following formula: ; In the formula, This represents the average slope propagation impedance actually used during the slope constraint projection process. The slope propagation impedance value is obtained by inversion based on historical landslide samples.

[0055] Optionally, the AI ​​agent model is a machine learning model, a deep learning model, or a combination thereof built based on slope features. For example, it could be a model architecture such as XGBoost, RandomForest, LightGBM, MLP, or ONNX deep learning models.

[0056] S4: Input the slope coordinate system characteristics, propagation impedance characteristics and inducing factor characteristics of each instability evaluation unit along the traffic corridor to be predicted into the slope instability influence range prediction model to obtain the influence range parameters and influence probability of each instability evaluation unit in the slope coordinate system.

[0057] In this embodiment, the acquisition of slope coordinate system characteristics, propagation impedance characteristics, and inducing factor characteristics of each instability evaluation unit along the traffic corridor to be predicted is similar to the feature acquisition process when constructing the training sample set, and will not be described in detail here.

[0058] In this embodiment, the influence range parameters of each instability evaluation unit in the slope coordinate system include the influence distance, influence width, and influence direction in the slope coordinate system.

[0059] S5: Reconstruct the polygon of the influence range in the slope direction based on the influence range parameters to determine the predicted influence range of slope instability.

[0060] In this embodiment, the method for predicting the impact range of slope instability in mountainous transportation corridors based on slope constraints further includes: constructing a risk level map of slope instability in the transportation corridor based on the predicted impact range of slope instability, and determining potential road blockage risks based on the interaction between the predicted impact range of slope instability and the road centerline.

[0061] Specifically, the road blockage risk includes the blockage probability, the blockage location, and the blockage level. In this embodiment, the blockage level is determined based on the blockage probability and a preset level threshold; the blockage probability is determined based on the interaction between the predicted slope instability impact range and the road centerline. Specifically, the interaction between the two is obtained through spatial overlay analysis. If the predicted slope instability impact range covers the road, the blockage probability is consistent with the impact probability; if the predicted slope instability impact range does not cover the road, the blockage probability is 0.

[0062] For example, three threshold levels are set: 0.75, 0.5, and 0.25. A blocking probability ≥ 0.75 indicates a high blocking probability, with the slope instability affecting the main traffic area of ​​the road, and is classified as high risk; 0.5 ≤ blocking probability < 0.7 indicates a medium blocking probability, with the slope instability affecting the edge of the road, and is classified as medium risk; 0.25 ≤ blocking probability < 0.5 indicates a low blocking probability, with the slope instability affecting the road not reaching it, and is classified as low risk; when the blocking probability < 0.25, it indicates a blocking probability close to 0, with the slope instability affecting the road not reaching it, and is classified as no risk.

[0063] This invention presents a method for predicting the impact range of slope instability in mountainous transportation corridors based on slope constraints. It establishes a slope coordinate system and introduces slope control factors such as aspect, gradient, roughness, vegetation cover, and engineering disturbance intensity. A slope propagation impedance function is established, and the impact range of numerically simulated or historical landslides is projected onto the slope direction and modulated with impedance to achieve a constrained expression of the landslide impact range in the actual slope direction. Based on this, a training dataset is constructed by combining topographic factors, engineering disturbance factors, and inducing factors along the transportation corridor. This dataset is then used to train a slope instability impact range prediction model based on slope features, which is used to quickly predict the impact distance, width, and direction of slope instability. By constructing a slope coordinate system composed of aspect vector, gradient vector, and slope normal vector, the impact range expression in the traditional planar coordinate system is transformed into a slope direction expression, ensuring that the landslide propagation direction closely matches the actual topographic slope aspect. This effectively solves the technical problem of large prediction deviations in the landslide propagation direction in mountainous transportation corridors. It can significantly improve the directional consistency and spatial accuracy of predicting the impact range of slope instability in mountain transportation corridors, providing technical support for traffic safety prevention and disaster early warning.

[0064] Furthermore, the effectiveness of the slope constraint-based method for predicting the impact range of slope instability in mountain transportation corridors based on the present invention is illustrated through simulation experiments.

[0065] 1. Basic data: By collecting historical disaster data, 12 typical slopes along the K section that have experienced landslides were identified. The data included: digital elevation model (DEM, 5m resolution), historical landslide impact range (manual interpretation + UAV orthophoto), engineering disturbance data (cut depth, retaining wall height, drainage ditch layout), and rainfall and traffic vibration records.

[0066] 2. Model Validation Scheme: The 12 historical landslides were randomly divided into two groups, with 8 landslides used for model training and 4 used for independent validation. To verify the effectiveness of the slope constraint model, three methods were compared as shown in Table 1.

[0067] Table 1

[0068] To comprehensively evaluate the predictive effectiveness, the following indicators were used: (1) Direction Deviation Angle (DDA), used to measure the deviation between the predicted propagation direction and the actual direction. , This is the slope aspect vector of the actual landslide propagation direction in the slope coordinate system, determined by the principal axis direction of the actual landslide influence area. The slope aspect vector in the slope coordinate system represents the predicted landslide propagation direction, determined by the influence direction predicted by the model. The DDA value is the angle between two vectors (usually an acute angle, measured in degrees), used to measure the deviation between the predicted direction and the true direction. The smaller the DDA value, the more accurate the direction prediction.

[0069] (2) Intersection over Union (IoU) , This represents the area of ​​the polygon representing the predicted landslide impact area. The area of ​​the polygon representing the actual landslide impact area.

[0070] (3) Slope Consistency Index (SCI), used to measure impedance consistency. , This represents the average slope propagation impedance actually used during the slope constraint projection process. The slope propagation impedance value is obtained by inversion based on historical landslide samples.

[0071] 3. Experimental Results Table 2 shows the comparison results of Direction Deviation Angle (DDA).

[0072] Table 2

[0073] Based on the results in Table 2, the average directional deviations were further calculated as follows: M1: 31.6°, M2: 21.6°, M3: 9.0° (a decrease of 57.1%).

[0074] Table 3 shows the comparison results of the overlap of influence range (IoU).

[0075] Table 3

[0076] Based on the results in Table 3, the average IoU was further calculated as follows: M1: 0.41, M2: 0.52, M3: 0.71 (an increase of 36.5%).

[0077] Table 4 shows the comparison results of the Slope Consistency Index (SCI).

[0078] Table 4

[0079] Based on the results in Table 4, the average SCI was further calculated as follows: M1: 0.63, M2: 0.74, M3: 0.89 (an improvement of 20.3%).

[0080] Experimental results analysis: The directional deviation angle decreased from 31.6° to 9.0°, indicating that the slope coordinate system can effectively reflect the true propagation direction, demonstrating that the slope coordinate system significantly improves directional consistency. The SCI improved by 20.3%, indicating that the impedance function can reflect the influence of slope, roughness, vegetation, and engineering disturbances on the propagation path, demonstrating that the slope propagation impedance function can effectively characterize the ease or difficulty of propagation. The IoU improved by 36.5%, indicating that the slope constraint model can more accurately reconstruct the landslide influence range, and the method of this invention can significantly improve the spatial accuracy of the influence range.

[0081] Due to the concentrated slope aspect and strong terrain dissection of transportation corridors, the slope constraint mechanism proposed in this invention can effectively solve the problem of large directional deviation in traditional methods. The experimental results above show that the slope constraint mechanism can significantly reduce the directional deviation of the affected area, verifying that the method of this invention is significantly superior to traditional methods in terms of directional consistency, spatial accuracy, and slope consistency. Therefore, the method of this invention is effective and has engineering application value in predicting the affected area of ​​slope instability in mountainous transportation corridors.

[0082] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations are intended to cover non-exclusive inclusion, such that an article or device comprising a list of elements includes not only those elements but also other elements not expressly listed. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the article or device comprising said element. Terms such as "connected" or "linked" are not limited to physical or mechanical connections but can include electrical connections, whether direct or indirect. The orientations or positional relationships indicated by terms such as "upper," "lower," "left," and "right" are based on the orientations or positional relationships shown in the accompanying drawings and are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as limiting the invention.

[0083] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features or characteristics described may be combined in any suitable manner in one or more embodiments or examples. In addition, those skilled in the art can combine and integrate the different embodiments or examples described in this specification.

[0084] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the spirit of the invention, and these modifications and combinations are still within the scope of protection of this invention.

Claims

1. A method for predicting the impact range of slope instability in mountainous transportation corridors based on slope constraints, characterized in that, include: S1: Obtain topographic data, engineering disturbance data, and inducing factor data of each instability evaluation unit along the mountain transportation corridor, and construct the slope coordinate system and slope propagation impedance function of each instability evaluation unit based on the topographic data and the engineering disturbance data. The triggering factor data includes at least one of rainfall, traffic vibration, seismic motion, and pore water pressure changes; Constructing the slope coordinate system includes: Based on the topographic data, the slope and aspect of each grid point in the instability evaluation unit are determined, and the aspect vector corresponding to each grid point is determined based on the slope and aspect. For each grid point, the corresponding slope normal vector is determined through local curvature analysis. The slope vector is determined based on the aspect vector and the slope normal vector. The slope coordinate system corresponding to each grid point is formed by the aspect vector, the slope vector, and the slope normal vector. Constructing the slope propagation impedance function includes: Slope, roughness, and vegetation cover are extracted from the topographic data, and engineering disturbance intensity is extracted from the engineering disturbance data. The slope propagation impedance function is constructed based on the slope, roughness, vegetation cover, and engineering disturbance intensity. The slope propagation impedance function is expressed as follows: ； in: For slope, For roughness, For vegetation coverage, For engineering disturbance intensity, For slope impedance, For roughness impedance, For vegetation impedance, and These are the weighting coefficients for slope resistance, roughness resistance, vegetation resistance, and engineering disturbance intensity, respectively. S2: Extract the slope coordinate system features, propagation impedance features, and inducing factor features of each instability evaluation unit as input features of the training samples. Use the slope coordinate system and the slope propagation impedance function to project the historical landslide influence range or numerical simulation landslide influence range of each instability evaluation unit onto the slope constraint. Obtain the influence range parameters under the slope coordinate system as the output labels of the training samples to form a training sample set of slope constraints. S3: Use the training sample set to train an artificial intelligence agent model to obtain a prediction model for the impact range of slope instability; S4: Input the slope coordinate system characteristics, propagation impedance characteristics, and inducing factor characteristics of each instability evaluation unit along the traffic corridor to be predicted into the slope instability influence range prediction model to obtain the influence range parameters and influence probability of each instability evaluation unit in the slope coordinate system; the influence range parameters include the influence distance, influence width, and influence direction in the slope coordinate system. S5: Reconstruct the polygon of the influence range in the slope direction based on the influence range parameters, and determine the predicted influence range of slope instability.

2. The method for predicting the impact range of slope instability in mountain transportation corridors based on slope constraints according to claim 1, characterized in that, In S2, the slope constraint projection includes: Transform the point set of the historical landslide influence range or the point set of the numerical simulation landslide influence range to the slope coordinate system; Obtain the projected distance of each point along the slope aspect vector in the slope coordinate system; The projection distance is modulated according to the slope propagation impedance function to obtain the slope constraint projection distance; The coordinates of each point in the slope coordinate system are obtained based on the slope constraint projection distance, forming a set of points representing the landslide influence range in the slope coordinate system.

3. The method for predicting the impact range of slope instability in mountain transportation corridors based on slope constraints according to claim 1, characterized in that, The parameters of the influence range include: influence distance, influence width, and influence direction.

4. The method for predicting the impact range of slope instability in mountain transportation corridors based on slope constraints according to claim 1, characterized in that, The features of the slope coordinate system include the slope vector, the aspect vector, and the slope normal vector; The propagation impedance characteristics include slope propagation impedance value, slope impedance, roughness impedance, vegetation impedance, and engineering disturbance intensity; The inducing factors include at least one of rainfall intensity, traffic vibration amplitude, peak ground acceleration, and pore water pressure variation.

5. The method for predicting the impact range of slope instability in mountain transportation corridors based on slope constraints according to claim 1, characterized in that, Before training the AI ​​agent model using the training sample set, the following steps are also included: Calculate the slope consistency index for each training sample in the training sample set, and remove training samples with a slope consistency index lower than a preset threshold to obtain the preprocessed training sample set.

6. The method for predicting the impact range of slope instability in mountain transportation corridors based on slope constraints according to claim 5, characterized in that, The slope consistency index is calculated according to the following formula: ； In the formula, This represents the average slope propagation impedance actually used during the slope constraint projection process. The slope propagation impedance value is obtained by inversion based on historical landslide samples.

7. The method for predicting the impact range of slope instability in mountain transportation corridors based on slope constraints according to claim 1, characterized in that, The method further includes: Based on the predicted impact range of slope instability, a risk level map of slope instability in the transportation corridor is constructed, and the potential road blockage risk is determined based on the interaction between the predicted impact range of slope instability and the road centerline.

8. The method for predicting the impact range of slope instability in mountain transportation corridors based on slope constraints according to claim 7, characterized in that, The road blockage risk includes the probability of blockage, the location of blockage, and the level of blockage, wherein the level of blockage is determined based on the probability of blockage and a preset level threshold. If the predicted impact range of slope instability covers the road, the blocking probability is the same as the impact probability; if the predicted impact range of slope instability does not cover the road, the blocking probability is 0.

Images