A refined construction method for shallow geological model of rock slope

Abstract

The application discloses a kind of rock mass slope shallow geological model fine construction method, comprising: step one, using ground penetrating radar to build rock mass internal model: step two, using unmanned aerial vehicle to build fine rock mass surface model: step three, model integration research: the rock mass surface three-dimensional model and internal three-dimensional model data of study area are coupled, and the digital three-dimensional model of slope for research is obtained.The present application uses unmanned aerial vehicle, ground penetrating radar to fine construction of shallow geological model in the range of 30m below the ground surface, identifies and processes rock mass structure surface by computer, establishes digital fine model of the study area rock mass, so that rock mass surface and internal structure surface information can be intuitively and efficiently obtained by computer, solving the shortcomings that traditional rock mass structure surface research can only be limited to partial area, research range is small, and rock mass internal research is not possible.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of rock mass slope structure analysis, in particular to a fine construction method of a rock mass slope shallow geological model. BACKGROUND

[0002] Large-scale engineering rock mass contains a large number of cross-scale joints, faults, weak interlayers, weathering fissures and other multi-level structural planes, which play a controlling role in the stability of the overall structure of the rock mass engineering. Under the influence of engineering disturbance, the structure framework and type of the rock mass change, leading to the deterioration and complication of the rock mass structure, which may further induce rock mass damage, instability and disaster. In engineering geological evaluation, structural planes are divided into five levels, and most of the structural planes that control the stability of the engineering rock mass are Ⅲ-Ⅳ level structural planes. Among the many factors affecting the stability of the rock mass slope, the rock mass structural plane is one of the most important factors, and it is also the most difficult to quantitatively describe. This is mainly due to the influence of many factors such as the geological conditions of the rock mass occurrence, the detection method and the accurate analysis of the detection results. Therefore, accurate research on the structural plane is an important aspect of evaluating the stability of the rock mass slope. At present, the contact or non-contact method is commonly used to study the distribution of the rock mass structural plane, and it is limited to the surface, and it is difficult to obtain the distribution and morphological characteristics of the internal structural plane of the rock mass. SUMMARY

[0003] The purpose of the present application is to provide a fine construction method of a rock mass slope shallow geological model, which uses an unmanned aerial vehicle and a ground penetrating radar to construct a fine shallow geological model within a range of 30m below the ground surface, and uses a computer to identify and process the rock mass structural plane, so as to solve the problems of inconvenience and low efficiency in obtaining the rock mass structural plane.

[0004] To achieve the above-mentioned purpose, the present application provides the following technical scheme:

[0005] A fine construction method of a rock mass slope shallow geological model, comprising the following steps:

[0006] Step 1: Constructing an internal model of the rock mass using a ground penetrating radar:

[0007] L1: According to the needs of the study area, use a ground penetrating radar with appropriate frequency to obtain internal data of the study area;

[0008] L2: Import the obtained relevant data into the ground penetrating radar software to perform noise reduction and optimization processing to obtain relevant radar images of the internal rock mass;

[0009] L3: Analyze the obtained internal radar images of the rock mass to determine the position and occurrence information of the internal fracture boundary surface;

[0010] Step 2: Constructing a fine rock mass surface model using an unmanned aerial vehicle:

[0011] W1, obtaining relevant information of the study area, determining the flight route, flight height and flight mode of the unmanned aerial vehicle;

[0012] W2, inputting the determined flight route and relevant flight parameters into the unmanned aerial vehicle, operating the unmanned aerial vehicle to perform the shooting task, and obtaining the ground image data;

[0013] W3, inputting the ground image data into the modeling software, the modeling software including Pix4D or Agisoft Metashape, obtaining the three-dimensional model of the ground surface of the study area after processing by the modeling software;

[0014] W4, processing the three-dimensional model by using the structural surface identification software to obtain the data of the occurrence, trace length and density of the structural surface, and then grouping the structural surface according to the occurrence information and analyzing to obtain the occurrence distribution function, equivalent disc diameter and bulk density characteristics of the dominant structural surface of each group;

[0015] Step three, model integration research:

[0016] Coupling the three-dimensional model data of the surface layer and the internal three-dimensional model of the rock mass in the study area to obtain the digital three-dimensional model of the slope for research.

[0017] Further, in L1, the study area is determined to be a shallow slope of the surface layer 0m-30m according to the frequency of the ground penetrating radar, a 40MHz-200MHz frequency antenna is used for detection, and multiple parallel lines are set at fixed distances according to the model of the ground penetrating radar antenna on the top of the study area.

[0018] Further, in L2, the ground penetrating radar supporting software is used to process the measured data obtained from L1 through band pass filtering, removing direct wave, background removal, linear gain and smoothing gain to obtain radar images that can be used for interpretation.

[0019] Further, in L3, the spatial position distribution and occurrence of the interface are calculated according to the structural surface plane information appearing in each line by integrating multiple line images.

[0020] Further, in W1, the relevant information of the study area includes: the relative height of the rock mass slope, the overall slope angle of the rock mass slope, the position of the coordinate reference point or control point, and the 1:1000 or 1:500 scale digital map of the study area.

[0021] Further, in W2, the unmanned aerial vehicle shooting adopts the method of oblique photography to study the study area, sets the camera angle perpendicular to the overall slope angle of the slope, sets 5 flight strips, the ground image resolution-GSD≤3cm / pixel, the heading overlap rate≥70%, the lateral overlap rate≥60%, and the flight height is selected according to the overall height of the slope, the shape and the GSD data.

[0022] Further, in W4, the model built in step W3 is first segmented according to geological zoning, and then the segmented model is imported into a structural plane identification software.

[0023] Further, the specific method in step three is divided into:

[0024] (1) When the ground penetrating radar image is clear and the relevant structural plane can be accurately judged, the model obtained in step one and step two is integrated and reconstructed by a three-dimensional modeling software;

[0025] (2) If the ground penetrating radar cannot obtain a clear and easily identifiable image, a discrete fracture network-DFN simulation form is used to reconstruct the internal fracture, and the fine model is obtained through the existing data verification.

[0026] Further, the construction characteristics of the discrete fracture network-DFN model are as follows:

[0027] (1) The DFN fracture is modeled using a 3D disc;

[0028] (2) The created DFN model takes each group of dominant structural planes obtained in W4 as the creation target;

[0029] (3) The fracture position is discrete in space, i.e. a Poisson distribution model;

[0030] (4) Each group of fracture sets has a specified density;

[0031] (5) The occurrence distribution of the fracture set is a 3D symmetric simulation of normal distribution, i.e. Fisher distribution;

[0032] (6) The DFN is created by combining a DFN template, a density term and a random seed.

[0033] Compared with the prior art, the beneficial effects of the present application are:

[0034] The rock mass slope shallow geological model fine construction method of the present application reconstructs the rock mass digital model by combining unmanned aerial vehicle high-precision oblique photography and ground penetrating radar and other geophysical prospecting methods, solves the shortcomings that the traditional rock mass structural plane research can only be limited to part of the area, the research range is small, and the internal rock mass cannot be researched; secondly, the present application digitally models the rock mass, which can be studied multiple times in subsequent work, avoiding repeated field work by artificial, saving manpower and material resources; in addition, the model constructed by the present application has higher accuracy than the geological model constructed by the traditional method, and is more intuitive, and the real-time advantage can provide relevant basis for slope stability evaluation, treatment and other projects faster. BRIEF DESCRIPTION OF DRAWINGS

[0035] Figure 1 It is the technical roadmap of the rock mass slope shallow geological model fine construction method of the present application embodiment.

[0036] Figure 2 The working schematic diagram of the rock mass slope shallow geological model refinement construction method of the embodiment of the present application;

[0037] Figure 3 The schematic diagram of the ground penetrating radar scanning rock mass interior research method adopted by the method of the present application;

[0038] Figure 4 The ground penetrating radar scanning survey line diagram of the embodiment of the present application;

[0039] Figure 5 The research area survey line 1 interpretation diagram of the embodiment of the present application;

[0040] Figure 6 The research area survey line 2 interpretation diagram of the embodiment of the present application;

[0041] Figure 7 The research area survey line 3 interpretation diagram of the embodiment of the present application;

[0042] Figure 8 The research area survey line 4 interpretation diagram of the embodiment of the present application;

[0043] Figure 9 The research area survey line 2 analysis diagram of the embodiment of the present application;

[0044] Figure 10 The research area survey line 3 analysis diagram of the embodiment of the present application;

[0045] Figure 11 The research area survey line 4 analysis diagram of the embodiment of the present application;

[0046] Figure 12 The research area 3D image analysis diagram of the embodiment of the present application;

[0047] Figure 13 The schematic diagram of the unmanned aerial vehicle rock mass surface modeling research method adopted by the method of the present application;

[0048] Figure 14 The rock mass surface layer unmanned aerial vehicle three-dimensional modeling diagram of the embodiment of the present application;

[0049] Figure 15 The rock mass surface layer slope surface structure plane digitization identification diagram of the embodiment of the present application;

[0050] Figure 16 The pole equal density cloud diagram of the occurrence of the embodiment of the present application;

[0051] Figure 17 The joint rose diagram of the embodiment of the present application;

[0052] Figure 18 The research area slope model diagram of the embodiment of the present application;

[0053] Figure 19 A simulation diagram of a regional slope interface for the embodiment of the present application is shown in the figure;

[0054] Figure 20 A DFN simulation diagram of two groups of structural surfaces in a study area for the embodiment of the present application is shown in the figure;

[0055] Figure 21 A stereographic projection diagram of DFN simulated fissures for the embodiment of the present application is shown in the figure;

[0056] Figure 22 A simplified DFN model diagram for the embodiment of the present application is shown in the figure;

[0057] Figure 23 A deterministic-non-deterministic structural surface coupling model diagram for the embodiment of the present application is shown in the figure;

[0058] Figure 24 A deterministic-non-deterministic structural surface coupling tangent diagram for the embodiment of the present application is shown in the figure;

[0059] Figure 25 A flowchart of the method for fine construction of a rock mass slope shallow geological model according to the embodiment of the present application is shown in the figure.

[0060] In the figure: 1, rock mass slope; 2, unmanned aerial vehicle; 3, ground penetrating radar. DETAILED DESCRIPTION

[0061] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts fall within the scope of the present application.

[0062] Referring to Figures 1-25 The method for fine construction of a rock mass slope shallow geological model provided in the embodiments of the present application includes the following steps:

[0063] Step 1: Constructing a rock mass internal model by using a ground penetrating radar

[0064] L1, according to the needs of the study area, using a ground penetrating radar with a suitable frequency to obtain internal data of the study area; specifically, according to the principle that the higher the frequency of the ground penetrating radar, the shallower the detection depth, and the higher the detection accuracy, in this embodiment, the main study is the shallow slope of the surface 0m~-30m, a 40MHz~200MHz frequency antenna is used for detection, and a plurality of parallel lines are set at fixed intervals on the top of the study area according to the model of the ground penetrating radar antenna;

[0065] The embodiment selects a 100m long region of a mine excavation slope as a study area, the ground penetrating radar model is RIS ground penetrating radar of Italy Botek, 20MHz-80MHz antenna is selected, 4 lines are set at every 3m on the top of the slope, the speed of the person and the equipment is controlled at about 1m / s during the measurement process; each record length is 20ns, 512 time sampling points; 9 point segmentation gain is adopted, linear gain from shallow to deep; the measurement detection mode is adopted, the horizontal position is controlled by the system recorded channel information, as shown in Figure 3 、 Figure 4 .

[0066] L2, the obtained relevant data is imported into the ground penetrating radar software for noise reduction, optimization processing to obtain the internal relevant radar image of the rock mass; wherein, the measured data obtained by L1 is processed by band pass filtering, removing direct wave, background removal, linear gain and smoothing gain through the matching software of the ground penetrating radar, to obtain the radar image which can be used for interpretation, as shown in Figures 5-8 , the interpretation image of line 1-line 4, the horizontal axis of the image is the length of the line (from left to right), and the vertical axis of the image is the detection depth.

[0067] L3, the obtained internal relevant radar image of the rock mass is analyzed to determine the position and occurrence information of the internal fracture interface; wherein, the spatial position distribution and occurrence are calculated according to the structural plane information appearing in each line by integrating the images of multiple lines;

[0068] As can be seen from Figures 5-8 , the line 1 is almost distributed with medium and low frequency signals on the whole measuring surface, which means that the internal slope of the study area does not contain water; and the red and blue lines are interlaced, the red line means that the reflection coefficient is positive and the reflection wave phase is positive, the blue line means that the reflection coefficient is negative and the reflection wave phase is negative, the interlaced red and blue lines mean that there are fracture surfaces, and the same phase axis is discontinuous and the waveform is relatively disordered, and the amplitude is large, so it can be judged that the whole measuring surface of line 1 belongs to the joint and fracture dense zone; it can be seen from line 2 that the image below the depth of 12m is mainly medium and high frequency signal, the signal frequency changes little, the same phase axis is continuous and the waveform is uniform, and the amplitude is low, so it can be judged that the lower part of 12m is a complete rock mass, and there is a layered surface at the depth of 12m; similarly, it can be judged that there are similar layered surfaces at the depth of 10m of line 3 and the depth of 6.5m of line 4, the layered lines of the broken region and the complete rock mass are connected through 3D analysis of the four lines, and the occurrence of the layered line is SW253∠30° analysis image and 3D image, as shown in Figures 9-12 .

[0069] Step two, a fine rock mass surface model is built by using an unmanned aerial vehicle

[0070] W1, obtaining relevant information of the study area, determining the flight route, flight height and flight mode of the unmanned aerial vehicle, specifically, obtaining the relative height of the rock mass slope of the study area, the overall slope angle of the rock mass slope, the position of the coordinate reference point or control point and the 1:1000 or 1:500 scale digital map of the study area and the like;

[0071] In this embodiment, DJI unmanned aerial vehicle M3E is adopted, the lens is 4 / 3 CMOS, the effective pixel is 20 million, the RTK module is adopted for accurate positioning, according to the relevant information of a mine, the relative height of the slope of the study area is 15 m, the slope direction of the slope is SW253°, the inclination angle is 40°, and the platform width is 20 m.

[0072] W2, inputting the determined flight route and related flight parameters into the unmanned aerial vehicle, operating the unmanned aerial vehicle to perform a shooting task, and obtaining ground image data; specifically, the study area is studied by using the method of oblique photography, the angle of view of the camera should be perpendicular to the overall slope angle of the slope as much as possible, 5 flight strips are set, the ground image resolution, hereinafter referred to as GSD, is kept at ≤3cm / pixel, the heading overlap rate is ≥70%, the lateral overlap rate is ≥60%, and the flight height should be selected comprehensively considering the slope height, shape and GSD data;

[0073] In this embodiment, the shooting is performed by using the method of oblique photography, 5 flight routes are set, the lateral overlap rate is 65%, the heading overlap rate is 75%, the gimbal pitch angle is set to-55°, the flight height is 80 m, the GSD is 2.5cm / pixel, the unmanned aerial vehicle is operated to perform a shooting task, and ground image data is obtained, as shown in Figure 13 .

[0074] W3, inputting the ground image data into modeling software, the modeling software including Pix4D or Agisoft Metashape, after processing by the modeling software, obtaining the three-dimensional model of the study area, as shown in Figure 14 .

[0075] W4, processing the three-dimensional model by using structure surface identification software, obtaining the data of the occurrence, diameter and density of the structure surface, and then grouping the structure surface according to the occurrence information and analyzing to obtain the occurrence distribution function, equivalent disc diameter and bulk density characteristics of the dominant structure surface of each group; specifically, the model of step W3 is segmented according to the geological zoning, and then the segmented model is imported into the structure surface identification software;

[0076] In this embodiment, the structure surface information of the study area is obtained, which is 36 groups, the occurrence of the structure surface is counted in the stereographic projection diagram, and the polar point isodensity cloud diagram and the joint rose diagram of the slope surface are obtained, as shown in Figures 15-17As shown, by structural plane identification software analysis, the occurrence distribution characteristics are obtained, the structural plane trace length simulation curve is obtained, the structural plane diameter is researched by using the disc assumption and the fitting parameters are obtained, and the structural plane bulk density is calculated by using the structural plane line density and the structural plane surface density.

[0077] Step three, model integration research:

[0078] The rock mass surface three-dimensional model and internal three-dimensional model data of the research area are coupled to obtain a digital three-dimensional model of the slope for research; when the ground penetrating radar image is clear and the relevant structural plane can be accurately judged, the model obtained in step one and step two is integrated and reconstructed by three-dimensional modeling software; if the ground penetrating radar cannot obtain a clear and easily identifiable image due to frequency problems, the internal fracture is reconstructed in the form of discrete fracture network (Discrete Fracture Network, hereinafter referred to as DFN) simulation, and a fine model is obtained through existing data verification;

[0079] The scale and geological structure of the research area in this embodiment are simple, and a three-dimensional model with a SW253° inclination, a 40° inclination angle, a 20m platform width, and a 15m relative height is drawn by using the built-in command stream of three-dimensional modeling, as shown in Figure 18 .

[0080] In the embodiment of the application, since the frequency of the ground penetrating radar is single, the internal structural plane information obtained is insufficient to meet the requirement of fine modeling, therefore, the model is obtained in the form of digital simulation structural plane in step three.

[0081] According to the analysis result of the ground penetrating radar detection, it can be concluded that there is an obvious rock layer boundary surface at a depth of 8-12m in the internal slope, the occurrence of which is SW253∠30°, the upper rock mass of the boundary surface is relatively broken, the structural plane is relatively dense, the signal frequency of the lower rock mass changes little, the phase axis is continuous and the waveform is uniform, and the amplitude is low, which can be inferred as a relatively complete rock mass, the boundary surface is drawn on the model to obtain the model as shown in Figure 19 .

[0082] In the embodiment of the application, the DFN construction is based on the 3D fracture bulk density inferred from the observed 1D fracture density, wherein the main features of the DFN model created in the research area are as follows:

[0083] (1) the DFN fracture is modeled using a 3D disc;

[0084] (2) the DFN model created this time takes each group of dominant structural planes obtained in W4 as the creation target;

[0085] (3) the fracture position is discrete in space (i.e. Poisson distribution model);

[0086] (4) the DFN model created this time takes each group of dominant structural planes obtained in W4 as the creation target;(4) Each set of cracks needs to have a specified density;

[0087] (5) The occurrence distribution of the fracture set is a 3D symmetric simulation of the normal distribution (i.e., Fisher distribution);

[0088] (6) DFN implementations are created through a combination of DFN templates, density terms, and random seeds.

[0089] The analysis of the model is obtained by digital simulation of the structural surface. The relevant parameters of the structural surface are shown in Table 1, where lmax is the maximum value of the disk diameter, lmin is the minimum value of the disk diameter, and lm is the expected value of the disk diameter. According to the parameters in the table, the DFN model created is as follows Figure 20 As shown:

[0090] Table 1DFN crack parameters

[0091]

[0092]

[0093] The DFN fissure generation area shown in the figure is slightly larger than the slope model of the study area. The black DFN fissures represent the dominant group 1, and the light gray DFN fissures represent the dominant group 2. The overall DFN fissures created are randomly generated based on random number seeds under the statistical distribution parameters in Table 1. A total of 1236 DFN fissures were created. By drawing their extreme points of occurrence and density, the DFN fissures are randomly generated based on random number seeds under the statistical distribution parameters in Table 1. Figure 21 , which is the same density as the pole measured in step 2 Figure 16 In comparison, under the premise of ignoring the structural surfaces outside the dominant group, it can be seen that the distribution of the simulated fractures is very similar to the measured ones.

[0094] In this embodiment, the DFN model using the number of cracks and the drill line density as the generation termination conditions may generate too many cracks, which will affect the subsequent coupling of deterministic and non-deterministic structural surfaces and further affect the mechanical modeling. Therefore, it is necessary to simplify the generated DFN joints, including deleting and merging them, as follows:

[0095] (1) Some cracks that are negligible in terms of application should be removed;

[0096] (2) Adjust the DFN model based on the connectivity of the cracks, where connectivity refers to the distance from the specified structure to the crack, calculated as the path through the center of the intersection;

[0097] (3) Cracks that are combined along a common plane can be merged.

[0098] The DFN model constructed by taking the body density as the fracture generation termination condition generally does not need to consider the adjustment and simplification of the DFN model, but since the model of the application includes the layered surface detected by the ground penetrating radar, the fractures develop on the upper part of the layered surface, and the lower part is the complete rock mass, the application takes the layered surface as the research object to simplify the DFN model:

[0099] Firstly, a layered surface (2D model) needs to be created as an initial structure, secondly, the connection distance between the fractures and the initial structure is calculated, then the fractures with too small connection distance are deleted, and finally, within a certain connection distance, the fractures with close distance and little change in dip angle are merged to reduce the body density of the structural surface near the layered surface, which is consistent with the actual situation; the adjusted DFN model is shown in Figure 22 , in which the medium gray color is the layered surface structure, compared with the original DFN model, a total of 22 DFN fractures are optimized.

[0100] In summary, the embodiment models the deterministic structural surface and the random structural surface of the research area, in which the deterministic structural surface is the layered surface obtained by radar detection, and the random structural surface is the DFN model constructed according to the occurrence, size and density parameters obtained by statistical analysis of the surface structural surface detected by the unmanned aerial vehicle, the coupling of the two can be realized by cutting the structural surface by the slope model, and the slope three-dimensional model obtained after coupling is shown in Figure 23 , the corresponding model tangent line is shown in Figure 24 , and finally a total of 193 blocks are cut out.

[0101] The embodiment of the application utilizes the unmanned aerial vehicle and the ground penetrating radar to finely construct the shallow geological model within the range of 30m below the ground surface, identifies and processes the rock mass structure surface by the computer, establishes the digital fine model of the rock mass in the research area, so that the information of the surface and internal structure surface of the rock mass can be directly and efficiently obtained by the computer, provides certain technical support for the research of the rock mass structure surface, and solves the problems of inconvenience and low efficiency in obtaining the rock mass structure surface.

[0102] The above merely describes the preferred specific embodiments of the application, but the protection scope of the application is not limited thereto, any person skilled in the art can make equivalent replacement or change according to the technical solution and the inventive concept of the application within the technical range disclosed by the application, which should be covered within the protection scope of the application.

Claims

1. A method for finely constructing a shallow geological model of a rock mass slope, characterized by: The following steps are involved: Step 1: Use ground penetrating radar to build a rock mass internal model: L1. According to the needs of the study area, use ground penetrating radar of appropriate frequency to obtain data inside the study area; L2. Import the acquired relevant data into the ground penetrating radar software for noise reduction and optimization processing to obtain relevant radar images inside the rock mass; L3. Analyze the obtained radar images of the rock mass interior to determine the location and occurrence of internal fracture interfaces; Step 2: Use drones to build a detailed rock surface model: W1. Obtain relevant information about the study area and determine the flight route, altitude, and mode of the UAV; W2. Input the determined route and related flight parameters into the UAV, operate the UAV to perform the shooting mission, and obtain surface image data; W3. Input the surface image data into modeling software, such as Pix4D or Agisoft Metashape. After processing by the modeling software, a three-dimensional surface model of the study area is obtained. W4. Use structural surface recognition software to process the three-dimensional model to obtain data on the occurrence, trace length, and density of the structural surfaces. Then, the structural surfaces are grouped according to the occurrence information and analyzed to obtain the occurrence distribution function, equivalent disk diameter, and volume density characteristics of each group of dominant structural surfaces; Step 3: Model integration research: The surface 3D model and internal 3D model data of the rock mass in the study area are coupled to obtain a digital 3D model of the slope for study; The specific methods in step three are as follows: (1) When the GPR image is clear and the relevant structural surfaces can be accurately identified, the models obtained in steps 1 and 2 are integrated and remodeled using 3D modeling software; (2) If the ground penetrating radar cannot obtain a relatively clear and easily distinguishable image, the internal fracture reconstruction is carried out in the form of discrete fracture network (DFN) simulation, and the refined model is obtained by verification with existing data; The construction characteristics of the discrete fracture network (DFN) model are as follows: (1) Model each group of DFN cracks according to the disk model; (2) The DFN model created takes the groups of dominant structural surfaces obtained in W4 as the creation target; (3) The crack location is discrete in space, i.e., the Poisson distribution model; (4) Each set of cracks has a specified density; (5) The occurrence distribution of the fracture set is a 3D symmetric simulation of the normal distribution, namely the Fisher distribution; (6) DFN is created by combining DFN template, density term and random seed.

2. The method for constructing a refined shallow geological model of a rock mass slope according to claim 1, wherein: In L1, the study area is determined to be a shallow slope with a surface depth of 0 to -30 m according to the frequency of the ground penetrating radar. A 40 MHz to 200 MHz frequency antenna is used for detection, and multiple parallel survey lines are set at fixed distances at the top of the slope in the study area according to the model of the ground penetrating radar antenna.

3. The method for constructing a refined shallow geological model of a rock mass slope according to claim 2, wherein: Ground penetrating radar supporting software is used in L2. The measured data obtained by L1 are then processed through bandpass filtering, direct wave removal, background removal, linear gain and smooth gain to obtain a radar image that can be used for interpretation.

4. The method for constructing a refined shallow geological model of a rock mass slope according to claim 1, wherein: In L3, by integrating images of multiple survey lines, the spatial position distribution and the occurrence of the interface are inferred based on the plane information of the structural surface appearing in each survey line.

5. The method for constructing a refined shallow geological model of a rock mass slope according to claim 1, wherein: The relevant information of the study area obtained in W1 includes: the relative height of the rock slope in the study area, the overall slope angle of the rock slope, the location of the coordinate reference point or control point, and the digital map of the study area at a scale of 1:1000 or 1:

500.

6. The method for constructing a refined shallow geological model of a rock mass slope according to claim 1, wherein: In W2, the UAV photography used the oblique photography method to study the study area. The camera’s viewing angle was set perpendicular to the overall slope angle, and five flight strips were set. The oblique ground image resolution - GSD ≤ 3 cm / pixel, the heading overlap rate ≥ 70%, the lateral overlap rate ≥ 60%, and the flight altitude was selected based on the slope height, morphology, and GSD data.

7. The method for constructing a refined shallow geological model of a rock mass slope according to claim 1, wherein: In W4, the modeling in step W3 is first segmented according to geological partitions, and then the segmented model is imported into the structural surface recognition software.

Images