Study on the Weak Interlayer Identification Method Based on Borehole Photo-Acoustic Combined Measurement: Application to a Landslide Case Study

Abstract

A weak interlayer is the key factor in controlling slope stability. It is of great significance to effectively identify the weak interlayer in the study of spatial and temporal distribution law and the internal structure characteristics of a landslide. Considering the limitations of traditional optical imaging and wave speed test methods, this paper presents a weak interlayer identification method based on borehole photo-acoustic combination measurement. By using the combination of optical imaging and acoustic wave scanning, the multi-source data collection of borehole rock wall and borehole surrounding rock is realized, which effectively captures the comprehensive response characteristics of the weak interlayer. This paper first constructs a multi-source data acquisition technology based on the borehole photo-acoustic combination measurement to realize the visualization of the image information and acoustic data of the target area on the borehole rock wall. Subsequently, the optical image features and the acoustic response characteristics of the weak interlayer are clarified based on the optical image and the acoustic scanning data. The hole wall texture characteristic response function, hole wall integrity characteristic response function, hole wall acoustic characteristic response function and hole wall contour characteristic response function are constructed. Finally, the landslide weak interlayer identification method considering the texture characteristics, complete characteristics, acoustic response characteristics and contour characteristics of the borehole rock wall is proposed, which effectively distinguishes the types of rock mass structural surface and realizes the automatic identification of the weak interlayer. Combined with the case analysis, the correctness and reliability of the present method are verified. The results show that the method can identify the weak interlayer and provide scientific basis for landslide management, which can provide a feasible and effective new way to identify the landslide weak interlayer in practical engineering, with a good application prospect and promotion value.

1. Introduction

Landslides are a common natural geological disaster, and their harm is great. With the increasing frequency of human engineering activities, the landslide problem becomes more and more prominent, and it has been the concern of scholars in various countries [1,2]. The purpose of landslide research is to understand the development mechanism and movement failure mode of landslides, and then determine prevention and control measures. It mainly includes three aspects: landslide expression, landslide stability analysis and landslide engineering treatment. Among them, landslide expression is the basis of landslide research, and the purpose is to describe the spatial distribution characteristics of landslide mass, the position and shape of the sliding surface and the material composition and geological structure characteristics of landslide mass. It lays a foundation for the establishment of a landslide mechanics mathematical model, monitoring model and landslide stability evaluation. The existence of a weak interlayer in the landslide mass provides a potential sliding surface for the landslide mass, which is not conducive to the overall stability of the landslide mass [3,4,5]. When there are soft and hard rock deposits in the slope, under the influence of external factors such as rainfall and wind, the softer rock layer is more likely to be weathered and broken and to shed, forming soft structural planes or a weak interlayer, sometimes forming hard rock and soft rock interbedding, which is not conducive to the stability of the clastic rock slope. The stability of the clastic rock slope is worse than that of the conventional one because of the weak interlayer. After the construction of houses and roads are cut, the safety and stability of the slope is more threatened. On the one hand, the weak interlayer has strong water sensitivity, which is easily saturated and softened once it absorbs water, and the strength is greatly reduced, resulting in the sliding failure of the slope along the interlayer. On the other hand, the infiltration of rainwater increases the self weight of the slope, and creep deformation of the slope occurs, providing power for the slope to start sliding. Therefore, the weak interlayer is the key factor to control slope stability. In the study of the spatial-temporal distribution law and internal structural characteristics of the landslide, it is of great significance to effectively identify the weak interlayer [6,7,8].

The weak structural surface in the rock mass is characterized by weak fractures, thinness, easy disturbance and erosion, and the strength is significantly different from the rocks on both sides. It is difficult to directly identify the weak structural plane in the rock mass. At present, the main identification method is through outcrop observation or excavation sampling, which has a large workload and high cost. It is mainly applicable to the identification of weak structures in shallow strata. The rock structure at the corresponding depth is mainly obtained through modeling and geophysical exploration [9,10,11,12]. The conventional borehole method has a low success rate in taking samples of weak structural plane and poor positioning and qualitative effects. The survey method of large-diameter borehole or excavation of exploratory tunnels and shafts is adopted, which has a long construction period, high cost and is unsafe. In landslide exploration, conventional boreholes are a very mature exploration method, which are simple, convenient and practical. However, this is an indirect method. When encountering complex geological conditions with a low coring rate, it is difficult to obtain geological information of structural planes such as faults, weak planes, weak joints and weak zones in the borehole, which easily leads to misjudgment of deep geological conditions [13,14,15]. At present, with the rapid development of various sensors, optical imaging technology, acoustic testing technology, laser scanning technology, computer technology, etc. have been applied in geological borehole surveys. At the same time, geological radars, seismic waves, electromagnetic waves and other geophysical techniques have also made rapid development in geological borehole research, greatly facilitating the exploration and application of engineering geology [16,17]. In addition, ultrasonic testing technology has been applied to the integrity, stability and internal defects of rock and soil structure. Many research achievements have been made in fluid media, fine structure of rock and soil and acoustic emission characteristics, but there are still no good methods for imaging detection and visual analysis of soft structures in rock and soil.

Through the use of borehole digital imaging technology, the geological information in the borehole is more intuitive and clearer [18,19,20]. The digital imaging results can truly reflect the damage of the rock mass in the hole under the influence of structure, karst and sliding, and make up for the defect that the borehole method is not suitable for exposing the long-standing sliding surface in the rock mass and the thickness of the sliding zone is very thin. Although the borehole digital imaging technology makes up for the shortcomings of the conventional geological borehole and provides more intuitive and clear geological information in the borehole. Since the borehole imaging technology only reflects the image characteristics of the borehole rock wall surface and does not consider the differences and characteristics of the internal structure of the borehole rock mass, it is often difficult to identify the type of structural plane and provide more rich data sources for the identification and analysis of the structure surface [21,22,23]. Therefore, in view of the technical difficulties in judging the structure surface by image recognition in the geological structure, this paper proposes a method for identifying the weak interlayer based on the borehole photo-acoustic combination measurement, taking the weak intercalation of landslide as an example. Through the combination of optical imaging technology and borehole acoustic technology, the organic integration of borehole image and acoustic scanning data is realized. The borehole photo-acoustic combined measurement technology can complement and confirm the advantages of both optical image and acoustic scanning technology, which can compensate for the shortage of single data sources in the process of change information extraction, and combine the complementary characteristics of different data analysis methods to ensure the correctness and reliability of weak interlayer identification, which can provide a scientific basis for geologic hazard control.

2. Photo-Acoustic Combined Measurement Technology in Boreholes

The borehole photo-acoustic combined measurement technology is a comprehensive fine detection technology of rock mass structure in the borehole that combines optical image technology and acoustic circumferential scanning technology [24]. The optical image information of the borehole wall can be acquired by the optical measurement component. The digital image processing of the image information can finely reflect the optical texture characteristics of the borehole rock mass structure. The acoustic measurement component can capture the full pulse information reflected by the acoustic wave pulse after encountering the rock wall. The digital signal processing of the full pulse echo information can finely reflect the acoustic response characteristics of each scanning point on the borehole wall. On the basis of combining the acoustic time signal processing, it can effectively reflect the geometric characteristics of the borehole profile. The instruments and equipment corresponding to the borehole photo-acoustic combined measurement technology are mainly composed of five parts: in-hole probe, transmission cable, depth encoder, integrated control box and computer. The schematic diagram of the test principle of applying the borehole photo-acoustic combined measurement technology to the geological borehole is shown in Figure 1.

The detection process of weak intercalation in geological borehole is as follows: first, lower the probe in the hole to the upper edge of the AB area to be detected in the geological borehole, and complete the optical image and acoustic scanning signal acquisition of the A area. Then, lower the probe in the hole at a constant speed. After the optical and acoustic signal acquisition in this area is completed, continue to lower the probe in the hole, and repeat this process until the probe in the hole has acquired the optical and acoustic signals at different depths in the AB area. Finally, through post analysis and processing, the optical image signal and acoustic detection signal of rock wall at each depth section are synthesized to realize the fine detection of the geological borehole.

In the aspect of data processing, the borehole photo-acoustic combined measurement technology mainly uses digital image expansion technology to realize optical image processing. In the captured video image, the image near the center reflects the rock wall image at the upper end of the drilling rock wall, and the image far away from the center reflects the rock wall image at the lower end of the drilling rock wall. The hole wall image closer to the center in the video image is more blurred, and the hole wall image farther away from the center in the video image is clearer, which is mainly caused by the principle of cone mirror imaging. For this reason, in image processing, the image far from the center of the video image is mainly used as the analysis data source. First, cut the video image along the geographical north direction. In the process of cutting, the center point selects the center point position of the video image to build a polar coordinate system. The tangential unfolding process can be realized by the following formula:

(1)$\{\begin{array}{l}{X}_{\theta }(\rho )=X_o+\rho \ast \cos (\theta )\\ Y_{\theta }(\rho )=Y_o+\rho \ast \sin (\theta )\end{array}$

where $X_{\theta }(\rho )$ and $Y_{\theta }(\rho )$, respectively, represent the image coordinates of the borehole wall when the angle is $\theta$ and the radius is $\rho$. (X₀, Y₀) is the center coordinate of the borehole $\theta \in [0,\hspace{0.17em}2\pi ]$. When $\rho$ changes from small to large, the pixel points on each circumference are sampled iteratively on $\theta \in [0,\hspace{0.17em}2\pi ]$. Then, the radial unwrapping technology of fish eye image is used to expand the image by interpolation. The image of borehole optical image acquisition is shown in Figure 2a, the schematic diagram of center coordinate search is shown in Figure 2b, and the optical expansion of the processed hole wall is shown in Figure 2c.

The core technology of acoustic wave scanning data processing is the circumferential visualization of scanning data. The circumferential visualization of discrete data points is realized by establishing the scanning matrix of horizontal section of the borehole. The innermost circle of the cross-section scanning matrix ring chart represents the acoustic reflection signal of the borehole surface, and the outermost circle represents the scanning maximum value of the reflection signal, that is, the maximum test range of the borehole. The length of a single scan is twice the maximum test range, and the length of a single scan is the product of the number of sampling points, the sampling interval time and the acoustic propagation speed of the rock. The other inner circles are arranged from the inside to the outside according to the time, and the size is the amplitude of the acoustic echo signal. Specifically, if a complete scanning wave has n sampling points, the sector area $S|{}_{h,i,1}$ formed by the innermost circle and the inner second circle displays the information represented by the first sampling point of the scanning signal, and the sector area $S|{}_{h,i,1}$ formed by the outermost second circle and the outer circle displays the information represented by the nth sampling point of the scanning signal. If the number of scanning lines of each section is m, then the surrounding rock structure information of the j-th sampling point on the i-th scanning line at the depth h is represented by sector area $S|{}_{h,i,1}$, where 1 ≤ i ≤ m, 1 ≤ j ≤ n, the span angle of sector area $S|{}_{h,i,1}$ is 2π/m, and the span width of sector area $S|{}_{h,i,1}$ is Δt. Δt is the sampling interval time of the signal. The schematic diagram of borehole circumferential scanning data is shown in Figure 3a, the schematic diagram of circular scanning matrix is shown in Figure 3b, and the visualization of processed scanning data is shown in Figure 3c.

3. Weak Interlayer Identification Method

3.1. Optical Image Characteristics of Weak Interlayer

In the borehole image, if the light is uniform, the different colors in the image mainly come from the differences in the composition and structure of the rock mass. The greater the difference in composition and structure of rock mass, the greater the difference in the corresponding image gray value and gradient value. The optical image can intuitively show the difference of rock color and filling connection. The darker the rock texture, the darker the corresponding image. If the rock mass structure is not filled, the corresponding image is also darker. The image matrix information can effectively reflect these differences. There is a difference between weak interlayer and non weak interlayer in filling composition, that is, the difference in texture characteristics of hole wall is greater. Therefore, the response function TF of texture characteristics of hole wall is constructed to describe the difference in composition of rock mass and the degree of filling cement. In addition, the complete rock mass structure has a certain thickness, which is shown as the continuity of gray value on the image, and the weak interlayer has a certain discontinuity. To this end, the hole wall integrity characteristic response function is constructed to describe the difference in the continuity of the gray value of the rock mass image.

Before the hole wall texture feature response function TF and the hole wall integrity feature response function IF are calculated, the optical unfolded image to be analyzed needs to be gridded first, and the specific size of the grid should be adjusted according to the actual situation. If the optical expansion image in Figure 4a is analyzed, the image with pixel point M × N is divided into p × q basic units, and the pixel point corresponding to each basic unit is M/p × N/q, as shown in Figure 4b.

A two-dimensional function $f_h(x,y)$ is defined to represent the borehole image at the borehole depth h. The upper left corner of the borehole image is a dot of the corresponding coordinate system, the upper edge of the image is the x-axis direction, that is, the image expansion direction, and the left side of the image is the depth direction, that is, the depth direction of the borehole. $\nabla f_h$ represents the gradient value at the midpoint (m, n) of the coordinate system, and its corresponding expression is:

(2)$\nabla f_h=grad(f_h)=\left[\begin{array}{l}{G}_{h,x}(m,n)\\ G_{h,y}(m,n)\end{array}\right]=\left[\begin{array}{l}{\displaystyle \sum _{k=-1}^1[f_h(m+k,n+1)-f_h(m+k,n-1)]}\\ {\displaystyle \sum _{k=-1}^1[f_h(m+1,n+k)-f_h(m-1,n+k)]}\end{array}\right]$

where $G_{h,x}(m,n)$ is the gradient value of the pixel point (m, n) on the borehole image at the borehole depth h in the x direction, and $G_{h,y}(m,n)$ is the gradient value of the pixel point (m, n) on the borehole image at the borehole depth h in the y direction, and the amplitude value is:

(3)$\left|\nabla f_h(x,y)\right|=\sqrt{G_{h,x}{(m,n)}^2+G_{h,y}{(m,n)}^2}$

The borehole unfolding image represents the rock mass structure information of 360 degrees of the borehole. In order to further refine the distribution of weak intercalations of the rock mass in different directions, the image can be decomposed into multiple images along different directions. It is assumed that the pixel point of the borehole image at the borehole depth h is M × N. Divided into q groups in the x-axis direction, i.e., q column scanning lines, then the number of pixel points on the borehole image corresponding to each column scanning line is M × (N/q). Since the hole wall texture feature response function TF needs to describe the composition of rock mass per unit depth at the borehole depth h and the degree of filling cement, and the gradient distribution on each scanning line needs to be considered, the hole wall texture feature response function $TF(h,i)$ takes the average value of RGB gradient of the image on the i-th scanning line per unit length at h, and its expression is:

(4)$TF(h,i)=1-\frac{\left(\nabla f_{h,i}|{}_r+\nabla f_{h,i}|{}_g+\nabla f_{h,i}|{}_b\right)}{3}$

where $\nabla f_{h,i}|{}_r$ represents the red image gradient corresponding to the i-th column scanning line at the borehole depth h. $\nabla f_{h,i}|{}_g$ represents the green image gradient corresponding to the i-th column scanning line at the borehole depth h. $\nabla f_{h,i}|{}_b$ represents the blue image gradient corresponding to the i-th column scanning line at the borehole depth h. The smaller the value of the hole wall texture feature response function, the more obvious the corresponding texture mutation, and the darker the corresponding color in the texture feature image. The larger the value of the hole wall texture feature response function, the less obvious the corresponding texture mutation, and the lighter the corresponding color in the texture feature image.

The borehole expansion image can show the integrity characteristics of the borehole wall, but due to the interference of the environment and the equipment itself, the formed image has some noise. It is necessary to enhance the image of the target area to highlight the abnormal area, so as to realize the integrity characteristics of the system hole wall. Because optical imaging is affected by different wavelengths, the component images corresponding to different wavelengths also have different characteristics. Therefore, in the calculation of the hole wall integrity characteristic response function IF, the local and global characteristics of the image are comprehensively considered from the image characteristics of different wavelengths. By establishing an image search window, the difference between the maximum value and the minimum value of the long wavelength information intensity in the image search window area of the whole borehole is searched and scanned, that is, the expression of the local feature L(h, i) of the image is:

(5)$L(h,i)=\underset{h,i\in \Omega }{\max }(S_r)-\underset{h,i\in \Omega }{\min }(S_r)$

where Ω is the search window, and the search window scans the optical image of the whole structure in a single pixel step. The long wavelength information in the optical image is used to characterize the local characteristics of the image, and the global contrast of the short wavelength information is extracted by using the feature of the weakest attenuation of the short wavelength optical information to form a significant global difference feature, that is, the expression of the image global feature O(h, i) is:

(6)$O(h,i)={\displaystyle \sum _{m=1}^M{\displaystyle \sum _{n=1}^N\Vert S_b(h,i)-S_b(m,n)\Vert }}$

where M is the number of rows of pixels in the borehole expanded image, and N is the number of columns of pixels in the borehole expanded image. In the aspect of hole wall integrity feature calculation of borehole unfolding image, a variety of image features are fused. The hole wall integrity feature response function $IF(h,i)$ considering the local feature L and the global feature O of the image is constructed. The hole wall integrity characteristic response function $IF(h,i)$ represents the scale of the maximum complete rock mass in the unit length corresponding to the i-th scanning line at the borehole depth h, and its expression is:

(7)$IF(h,i)=L(h,i)\times O(h,i)$

Among them, h and i need to be normalized according to the values of p and q, that is, h∈[1, p], i∈[1, q]. If h is greater than p, h is averaged and rounded. If i is greater than q, i is averaged and rounded. The smaller the value of the hole wall integrity characteristic response function, the worse the corresponding rock wall integrity, and the darker the corresponding color in the integrity characteristic image. The larger the value of the hole wall integrity characteristic response function, the better the integrity of the corresponding rock wall, and the lighter the corresponding color in the integrity characteristic image.

3.2. Acoustic Scanning Characteristics of Weak Interlayer

Through the acoustic scanning signal, the acoustic characteristics of the scanned rock wall can be perceived and the corresponding geometric contour characteristics of the scanned rock wall can be reflected. Therefore, the acoustic scanning characteristics of the weak interlayer are analyzed based on the two effective information of the acoustic scanning signal. Before the acoustic scanning characteristic analysis of the weak interlayer, the circumferential acoustic scanning information needs to be converted into matrix information, and the corresponding schematic diagram is shown in Figure 5. Figure 5a is the acoustic wave scanning data of a certain horizontal section, and the grid shown in Figure 5b is formed by superimposing the data of adjacent p horizontal scanning sections, where q represents the number of scanning lines of the horizontal section. In order to be consistent with the optical feature analysis, the number of scanning lines is normalized and averaged, and then the acoustic signal is mapped to the corresponding rectangular cells. Each rectangular cell retains the complete acoustic scanning data, providing basic data for the subsequent acoustic scanning feature analysis of weak interlayer.

In terms of the acoustic characteristics of a rock wall, there are certain differences between the weak interlayer and non-weak interlayer in acoustic response characteristics. The key to distinguish them is to find the acoustic characteristic function of rock. The power spectrum of an ultrasonic wave has the characteristics of highlighting the main frequency and reflecting the power distribution of signal energy. Therefore, the acoustic characteristic response function AF of the hole wall is established by using the scanning data of ultrasonic wave and the power spectrum characteristics of ultrasonic wave. Ultrasonic power spectrum analysis refers to the analysis of ultrasonic signals in the frequency domain. The power spectrum is obtained by Fourier transform of ultrasonic time-domain signals. Power spectrum is the abbreviation of power spectral density function, which is defined as the signal power in the unit frequency band. It shows the change of signal power with frequency, that is, the distribution of signal power in frequency domain.

The power spectrum represents the variation of signal power with frequency. For the time domain signal f(t) with sample length T, if FFT is used for analysis, the power spectrum is expressed as:

(8)$G(f)=\frac{2}{T}{\left|F(f)\right|}^2$

where $F(f)$ is the Fourier transform of $f(t)$. According to the physical meaning of spectrum analysis, the power spectrum reflects the energy distribution in the frequency domain. If the frequency spectrum is high, the sound wave attenuation of this section of rock on the surface is small, and the received signal is strong. On the contrary, if the frequency spectrum is low, the attenuation is large, and the received signal is weak. In order to extract the characteristic information of the received echo signal more accurately, an acoustic characteristic function $E(f)$ is proposed, which is defined as:

(9)$E(f)=\frac{{\displaystyle {\mathsf{\int }}_{f_1}^{f}G(f)df}}{{\displaystyle {\mathsf{\int }}_{f_1}^{f_2}G(f)df}}$

where f₁ is the minimum frequency value of the analysis frequency interval, f₂ is the maximum frequency value of the analysis frequency interval, f is the change frequency of the characteristic function, and $G(f)$ is the power spectrum function. In order to make the acoustic characteristic function independent of the absolute amplitude of the ultrasonic echo signal and only extract the relative change of the echo signal frequency, it is necessary to normalize the analyzed echo signal. The acoustic response characteristics corresponding to different rocks are different, and the acoustic characteristic response function AF of the hole wall is reflected by the area enclosed between the acoustic characteristic function $E(f)$ and the frequency axis. $AF(h,i)$ represents the acoustic response characteristics of the hole wall in the unit length corresponding to the i-th scanning line at the borehole depth h, and the corresponding expression is:

(10)$AF(h,i)={\displaystyle {\mathsf{\int }}_{f_{\min }}^{f_{\max }}E(f)|{}_{h,i}df}$

where f_max represents the maximum frequency corresponding to the bandwidth of the acoustic signal, and f_min represents the minimum frequency corresponding to the bandwidth of the acoustic signal. $E(f)|{}_{h,i}$ represents the acoustic characteristic function corresponding to the i-th column scanning line at the borehole depth h.

In terms of the geometric profile characteristics of the rock wall, the borehole profile has a certain impact on the optical imaging and acoustic wave scanning results. When the borehole profile is in the characterization circle, the illumination is uniform, and the color characteristics of the collected optical image can effectively reflect the changes of the rock mass structure, which is helpful to the identification of the weak interlayer. However, when the borehole profile is a non-standard circle, such as the phenomenon of falling blocks, the illumination intensity of the rock surface area is small, which affects the effective identification of the weak interlayer. Similarly, when the borehole profile is in the characterization circle, the acoustic energy attenuation of the acoustic wave in different directions is relatively consistent, and the weak interlayer and the non-weak interlayer can be effectively distinguished according to the difference characteristics of the energy. However, when the borehole profile is a non-standard circle, such as in the block falling phenomenon the energy change of the acoustic wave will be different from that of the normal position due to the longer or closer propagation distance of the acoustic wave. Therefore it is easy to affect the effective identification of weak interlayer. The acoustic scanning data can effectively obtain the contour information of the borehole. Therefore, the borehole contour feature response function CF is established to reflect the borehole contour feature of the measurement area.

In the conversion process from Figure 4a to Figure 4b, the waveform needs to be matrixed to realize the process of converting the waveform into an image matrix. If the acoustic scanning data at the same level is represented by $U|{}_{h,a}$, the time corresponding to the first occurrence of strong reflection signal is represented by $t|{}_{h,a}$, representing the time taken for the acoustic signal to and from. Time $t|{}_{h,a}$ can be determined by threshold method and energy method, that is, by the maximum peak value or energy maximum value of the first echo signal. Then the expression of the distance $R|{}_{h,a}$ between the drilling center position and the rock wall scanning point corresponding to the scanning line is:

(11)$R|{}_{h,a}=\frac{c\ast t|{}_{h,a}+d}{2}$

In the above formula, c is the sound velocity corresponding to the acoustic wave propagation medium in the borehole. If the acoustic wave medium in the borehole is in a non mud environment, it is generally 1480 m/s. d is the diameter of the probe in the hole. If the diameter of the geological borehole is D, the expression of the corresponding hole wall profile characteristic response function $CF(h,a)$ at the location where the geographic direction is a at the depth h is:

(12)$CF(h,a)=\left|R|{}_{h,a}-\frac{D}{2}\right|$

Among them, h and a need to be normalized according to the values of p and q, that is, h∈[1, p], a∈[1, q]. If h is greater than p, h is averaged and rounded. If a is greater than q, a is averaged and rounded.

3.3. Identification of a Weak Interlayer in Geological Borehole

There are differences in texture characteristics, integrity characteristics and acoustic response characteristics of the weak interlayer in the geological borehole. The contour characteristics of the geological borehole have a certain impact on the interpretation of the results of optical imaging and acoustic scanning in the borehole. Therefore, based on the optical image and acoustic scanning data, considering the texture characteristics, integrity characteristics, acoustic response characteristics and contour characteristics of the borehole wall, this paper proposes a recognition method suitable for the weak interlayer. The position information of the weak interlayer in the borehole rock mass structure is highlighted by organically fusing the hole wall texture feature response function TF, the hole wall integrity feature response function IF, the hole wall acoustic feature response function AF, and the hole wall contour feature response function CF. The corresponding weak interlayer identification block diagram is shown in Figure 6.

In this paper, before the identification of the weak interlayer in the geological borehole, based on the original hole wall image fi, the hole wall texture feature response function TF, the hole wall integrity feature response function IF, the hole wall acoustic feature response function AF, and the hole wall contour feature response function CF are fused to reconstruct the corresponding rock wall image. Construct the borehole wall reconstruction image FI considering various hole wall features, and the expression is:

(13)$FI(i,j)={\lambda }_1\cdot TF(i,j)+{\lambda }_2\cdot IF(i,j)+{\lambda }_3\cdot AF(i,j)+{\lambda }_4\cdot CF(i,j)$

Among them, ${\lambda }_1$, ${\lambda }_2$, ${\lambda }_3$ and ${\lambda }_4$ are the weighted values of different response functions, usually taken as the average value, and can also be adjusted appropriately according to the actual situation. At the same time, the four parameters also need to meet the basic requirements of ${\lambda }_1+{\lambda }_2+{\lambda }_3+{\lambda }_4=1$. When the hole wall texture feature and the hole wall integrity feature are obvious, that is, the larger the value of the corresponding hole wall texture feature response function TF and the hole wall integrity feature response function if, the smaller the possibility of corresponding weak interlayer. When the hole wall texture feature and the hole wall integrity feature are not obvious, that is, the corresponding values of the hole wall texture feature response function TF and the hole wall integrity feature response function IF are small, the greater the possibility of a corresponding weak interlayer. When the acoustic characteristics of the hole wall and the contour characteristics of the hole wall are obvious, that is, the larger the value of the corresponding acoustic characteristic response function AF of the hole wall and the contour characteristic response function CF of the hole wall, the greater the possibility of corresponding weak interlayer. When the acoustic characteristics of the hole wall and the contour characteristics of the hole wall are not obvious, that is, the values of the corresponding acoustic characteristic response function AF of the hole wall and the contour characteristic response function CF of the hole wall are small, the possibility of corresponding weak interlayer is smaller. Through the grayscale processing of 256 values for the four-hole wall feature values, the smaller the FI value of the borehole wall reconstruction image formed, the greater the possibility of corresponding weak interlayer, that is, the darker the borehole wall reconstruction image FI, the higher the probability of corresponding weak interlayer. Combined with the subsequent image processing, the weak interlayer can be recognized.

The identification of weak interlayer in geological borehole is a process of binarization of borehole wall reconstruction image FI considering various borehole wall characteristics. The process of binarization is to segment the image by threshold processing. According to different threshold selection methods, the main algorithms of image segmentation are the histogram threshold method, iteration method and Ostu method [25,26]. Wavelet transform has the ability to detect local mutation, and is a good tool for edge detection. Its characteristic is to combine multi-scale information to detect. The weak interlayer area in the reconstructed image is the place with the largest change rate, and the first derivative of the smoothing function can be used as the wavelet function [27,28,29]. Set up θ(x₁, x₂) is a two-dimensional smoothing function, and its first derivative along x₁ and x₂ directions is taken as two basic wavelets. For any two-dimensional function $f(x_1,x_2)\in L^2(R^2)$, its wavelet transform has two components:

(14)$\mathrm{Along} {\mathrm{x}}_1 \mathrm{direction}: W{T}^{(1)}f(a,x_1,x_2)=f(x_1,x_2)\ast \ast {\psi }_a^{(1)}(x_1,x_2)$

(15)$\mathrm{Along} {\mathrm{x}}_2 \mathrm{direction}: W{T}^{(2)}f(a,x_1,x_2)=f(x_1,x_2)\ast \ast {\psi }_a^{(2)}(x_1,x_2)$

where ** represents two-dimensional convolution, and the vector form of its specific meaning is:

(16)$\left[\begin{array}{l}W{T}^{(1)}f(a,x_1,x_2)\\ W{T}^{(2)}f(a,x_1,x_2)\end{array}\right]=a\left[\begin{array}{l}\frac{\partial }{\partial x_1}[f(x_1,x_2)\ast \ast {\theta }_a(x_1,x_2)]\\ \frac{\partial }{\partial x_2}[f(x_1,x_2)\ast \ast {\theta }_a(x_1,x_2)]\end{array}\right]$

where $f^s(x_1,x_2)$ is the image obtained by smoothing $f(x_1,x_2)$ by ${\theta }_a(x_1,x_2)$. The above formula indicates that WT⁽¹⁾ and WT⁽²⁾ respectively reflect the gradient of this image along the (x₁, x₂) direction. Generally, a is taken as $2^j(j\in z)$, and the corresponding vector formula is:

(17)$\left[\begin{array}{l}W{T}^{(1)}f(2^j,x_1,x_2)\\ W{T}^{(2)}f(2^j,x_1,x_2)\end{array}\right]=WTf(2^j,x_1,x_2)$

The relation (18) is the binary wavelet transform of $f(x_1,x_2)$. The modulus of $f(x_1,x_2)$ is:

(18)$Mod\left[WTf(2^j,x_1,x_2)\right]={\left[{\left|W{T}^{(1)}f(2^j,x_1,x_2)\right|}^2+{\left|W{T}^{(2)}f(2^j,x_1,x_2)\right|}^2\right]}^{1/2}$

Its argument angle (included angle with x₁ direction) is:

(19)$Arg\left[WTf(2^j,x_1,x_2)\right]=t{g}^{-1}\left[\frac{W{T}^{(2)}f(2^j,x_1,x_2)}{W{T}^{(1)}f(2^j,x_1,x_2)}\right]$

Through the above steps, the borehole wall reconstruction image FI considering various hole wall characteristics is binarized, and the distinction between weak interlayer and non-weak interlayer in the geological borehole mass can be realized. The black area is used to represent the weak interlayer in the geological borehole of the rock mass, and the white area is used to represent the non weak interlayer area in the geological borehole of the rock mass, so as to provide a reference for the judgment of the sliding surface.

4. Case Analysis

4.1. Data Acquisition

The instruments and equipment used in this case are the borehole photo-acoustic combined measurement system independently developed by Institute of Rock and Soil Mechanics, Chinese Academy of Sciences. The instrument and equipment are mainly composed of five parts: probe ①, transmission cable ②, depth encoder ③, integrated control box ④ and computer ⑤. The physical image of the overall instrument and equipment is shown in Figure 7a. The core components of the probe in the hole are the optical measurement component and the acoustic scanning measurement component. The optical measurement component is mainly composed of an optical camera and an LED diaphragm component. The physical picture of the component is shown in Figure 7b. The acoustic scanning component is mainly composed of acoustic transducer, electronic compass, scanning motor and other components. The physical picture of the component is shown in Figure 7c. The communication mode is network signal transmission, and armored optical fiber cable is used as the signal transmission medium.

The data source used in this case analysis is a rock landslide in Jinping County, Guizhou Province, China. The landform is broken and the terrain is high in the northwest and low in the southeast. The west and southwest are middle and low mountain areas, the middle and northeast are low mountain canyon areas, and the southeast is a low mountain hilly basin dam area. In the investigation and study of determining the weak layer, if only the traditional digital borehole camera system is used to collect the data of the borehole wall, the distribution position of the structural plane can only be roughly judged, and the type of the structural plane cannot be accurately identified. Therefore, in order to further identify the position distribution of the structural plane and make a correct judgment on whether it is a weak interlayer, the corresponding instruments and equipment of this technology are used to collect the data of optical image and acoustic scanning. The optical image and acoustic wave scanning data of some hole walls are shown in Figure 8. Figure 8a is a screenshot of some rock wall video images at the borehole depth of 28.5~29.5 m. Figure 8b is a pseudo color image of some borehole acoustic scanning data at the borehole depth of 28.5–29.5 m.

4.2. Data Processing

Digital image processing is performed on the video image acquired by this technology; after the coordinates of the center of the circle in the image are determined, the origin of the computer coordinate system is transformed to the center of the circle image. The video images of different depths are tangentially expanded, and the images are radially expanded to form the borehole wall expansion image as shown in Figure 9. The horizontal direction of the image in Figure 8 represents the depth of the borehole, and the depth of the borehole gradually increases from left to right. The left side of the image represents the rock wall image at a depth of 28.5 m, and the right side of the image represents the rock wall image at a depth of 29.5 m. From top to bottom, Figure 8 represents that the borehole is unfolded from north to anticlockwise, that is, the borehole image represents the rock wall azimuth image of “North West South East North” borehole from top to bottom.

During the analysis of the hole wall characteristics of the image, the image was divided into 360 × 1000 basic units, that is, the number of scanned lines in the borehole rock wall was 360, and the angle between the two adjacent scanning lines was 1°. In the borehole depth direction, the distance between the two adjacent foundation units was 1 mm. In this paper, the image of hole wall texture feature response function TF and the hole wall integrity feature response function IF were calculated, respectively. The results of the image after the matrix are shown in Figure 10a. The results of the image corresponding hole wall integrity feature matrix are shown in Figure 10b. The acoustic scan data acquired using the present technique was transformed into matrix information and f₁ = 0.2 MHz, f₂ = 2 MHz are selected. The acoustic characteristic function was calculated separately for the transmit signal and the received echo signal respectively. The acoustic features of the hole wall at 28.5~29.5 m are shown in Figure 11a. The profiles of the hole wall at 28.5~29.5 m are shown in Figure 11b.

After completing the calculation of hole wall texture feature response function TF, IF, hole wall acoustic feature response function AF and CF, the matrix to form hole wall texture feature, hole wall integrity feature, hole wall acoustic feature and hole wall profile feature image, the four images were fused according to the weights of 0.25, 0.25, 0.25, 0.25 and 0.25 to reconstruct a new borehole rock wall expansion image, as shown in Figure 12a. Subsequently, the borehole rock wall reconstruction image FI, considering a variety of hole wall features, was binarized to form a binary image, as shown in Figure 12b. From the binary image, we can distinguish between the weak interlayer and realize the identification of the weak interlayer. The black area is used to indicate the weak interlayer in the geological borehole of the landslide, and the white area is used in the landslide geological borehole, so as to provide a reference for judging the sliding surface of the landslide. The technical method in this paper shows that there is a weak interlayer at the borehole depth of about 29.1 m, which needs special attention. The thickness of the sandwich is about 5 cm.

4.3. Interpretation of Result

The image results of the above section effectively highlight the optical image of rock wall texture and rock integrity characteristics. At the same time, combined with the rock response and hole wall profile characteristics in the acoustic scanning, the comprehensive image fusion of various factors is realized and thus provides a more abundant data source for the judgment of the weak interlayer. Before reconstructing the borehole rock wall expansion image, as shown in Figure 13a, and the post-fused image is shown in Figure 13b. As can be seen from Figure 13, the rock wall structural surface in Figure 13a is a more structured surface texture, and the less structured surface texture is shown in Figure 13b. This is mainly due to the multiple types of structural surfaces on the hole wall, and not all structural surfaces are a weak interlayer. Through the technical processing in this paper, the real software structure can be highlighted and the effective screening of information can be realized. Therefore, the structural surface of the hole wall is less informative and more likely to be a weak feature.

The image of the borehole rock wall, not processed by the present technique, is shown in Figure 14a. Structural feature regions were identified by fusing synthetic images of the borehole rock wall with multi-source data, and the results are shown in Figure 14b. Through the comparison, it can be seen that to identify the image considering multi-source factors, only the region with color difference can only be identified through the color depth information. The results are polyanalytical. At the same time, the target features are not focused, and the specific types of structural features cannot be distinguished. Through this technology, the structural characteristic region containing multiple information can be identified, which can reflect the essential difference characteristics of the rock, and effectively identify the weak and non-weak areas of the rock.

Combined with the energy difference characteristics of the image pixels, the contour 3D map of the rock wall image was established, as shown in Figure 15. Figure 15a shows the 3D drawing before the borehole wall fusion, and Figure 15a is the 3D figure after the borehole wall fusion. As can be seen from Figure 15, the 3D contour map surface of the hole wall was not smooth, and the vertical depth jumps in more areas before the fusion technology treatment presented here. After the fusion technique, the 3D contour map surface of the hole wall was smooth, and the longitudinal depth jump area was small. This is mainly due to the fusion treatment; the target areas to be analyzed are more concentrated. Through three-dimensional high treatment, the rock characteristics of rock walls can be presented more intuitively, which can facilitate the division and identification of weak areas.

To verify the correctness of this method, the reconstructed image formed by the present method is compared with the traditional method, as shown in Figure 16. The blue curve in the figure represents the values obtained by the optical borehole TV, the green curve represents the values obtained by the method of this paper, and the red curve represents the values obtained with the wave speed test technique. To compare the different values, the three data were normalized separately. The image data were averaged over the gray scale values in the same depth direction and then converted into values between 0 and 1, and the larger the value, the larger the gray scale, and the closer to the non-weak layer rock. The sound speed value was normalized to the sound speed value of the rock at different depths and converted into values between 0 and 1. The larger the value, the greater the gray scale, and the closer it is to the non-weak layer rock. The smaller the value in Figure 16, the greater the probability of weak layer rock. By comparing the paper method with the traditional optical borehole image method and the rock sound speed method, it can be seen that the results of the method and the traditional two methods in the weak interlayer area is more consistent, and can effectively reflect the geographical depth of the weak interlayer and provide more accurate data for the location of the weak interlayer in the borehole.

The image before and after the fusion is curled along the central axis of the borehole to form the image shown in Figure 17. Figure 16a is an image acquired by the optical imaging mode, presenting a virtual core image of the N-S orientation. Figure 17b is an image acquired by an optical imaging mode, presenting a virtual core image of the S-N orientation. Figure 17c shows the image reconstructed using the present method, presenting a virtual core image of the N-S orientation. Figure 17d shows the image reconstructed using the present method, presenting a virtual core image of the S-N orientation. In order to more intuitively present the changes of the stratum with the depth, the 3D stereo diagram of the stratum was reconstructed by the virtual core image, and the stratum map at the depth 28.5–29.5 m is shown in Figure 18. Figure 18a shows the 3D images of the strata before using the method of this paper. Four structural features can be intuitively seen in Figure 18a. Figure 18b shows the 3D images of the strata. One structural feature can be intuitively seen in Figure 18b. Through the comparison, we can see that this method can effectively distinguish the types of structure features, and more effectively identify the weak structure surface from the structure types. This is mainly due to the synthesis of a variety of rock characteristic parameters, and to realize the fusion of multi-source data the results are more specific features.

In terms of the verification of the accuracy of the weak structure identification, the present method is compared with the results of the optical imaging acquisition and the rock sound speed test. In terms of the weak structure surface identification of optical imaging, the probability distribution of the weak structure at different depths is shown in Figure 19a. In terms of the structure and weak structure surface identification of this method, the probability distribution of the weak structure at different depths is shown in Figure 19b. In terms of the structure weak structure surface identification of the sound speed test, the probability distribution of the weak structure at different depths is shown in Figure 19c. The abscissa in Figure 19 shows the likelihood that the formation is a weak layer, and the greater the value. The ordinate in Figure 19 indicates the depth of the formation. As can be seen from Figure 19, conventional optical imaging techniques acquire more areas of structural characteristic features, with five areas of concern at hole 28.5–29.5 m when the possibility threshold is set to 0.6. When the possibility threshold is set to 0.8, there are 4 areas of concern at hole 28.5–29.5 m. Traditional acoustic testing techniques acquire smaller areas of structural features, with 1 area of concern at hole 28.5–29.5 m when the possibility threshold is set to 0.6. When the possibility threshold is set to 0.8, there is 1 area of concern at hole 28.5–29.5 m. The number of structural feature regions obtained using the technical method is centered, with 3 areas of concern at hole 28.5–29.5 m when the likelihood threshold is set to 0.6. When the possibility threshold is set to 0.8, there is 1 area of concern at hole 28.5–29.5 m.

From the comparison of the three sets of data in Figure 19, it can be seen that the number of weak regions identified by this method is somewhere between the number of structural regions obtained by the optical image method and the acoustic test method. This method identifies fewer weak regions than those identified by conventional optical image methods. That is, the present weak structural surface identification method is more accurate than the traditional optical borehole imaging technique, which can effectively distinguish the types of structural surface and extract the weak structural surface. This is mainly due to the fact that the traditional optical image method mainly senses the color difference of the borehole rock wall, while the structural area in the very optical image may be the texture characteristics of the rock itself, and it is not weak. This technical method combines the acoustic scanning data to consider the rock response and hole wall profile characteristics in the acoustic scanning by superimposing the rock acoustic information inside the rock wall. This method identifies more weak regions than conventional sound speed measurement methods. That is, the identification method of the weak structure surface is more abundant than the traditional acoustic wave test technology, which can effectively characterize the potential sub-weak structure region. This is mainly due to the traditional acoustic test method which mainly measures the difference of the sound speed of the rock to search for the area with the lowest sound speed of the rock, while for the local crack area, the identification ability is weak. In the fine survey, it is not easy to reflect more possible weak areas. The technical approach of this paper combines optical image data to consider rock wall texture and rock integrity features. By stacking the rock wall texture difference features, the potential weak structure regions can be identified and presented.

At present, in the process of landslide surveys, the formation information is usually disclosed according to the well depth, trough exploration and borehole, but in the case of low core extraction rate, discontinuity or incomplete investigation information, the process can easily lead to the inaccurate identification of the weak interlayer. In the case of unclear rock mass structure type, it is easy to misjudge the landslide type. Although the traditional borehole optical imaging technology can make up for the shortage of conventional geological boreholes and provide more intuitive borehole geological information, because the borehole imaging technology only reflects the image characteristics of the borehole wall surface, it does not consider the internal structure of borehole rock characteristics, so it finds it difficult to identify the type of structural surface and cannot provide more rich data sources for the identification and analysis of landslide sliding surface. The traditional wave speed test method can only roughly judge the sound speed change of rock at different strata depths, which reflects the comprehensive results of rock at the same depth, and does not have directional characteristics. As a result, the traditional wave speed test method can only roughly judge the position of the weak interlayer, which can not fully reflect the distribution and contour characteristics of the structural characteristics, and has a limited ability to identify a thin, weak interlayer. Therefore, to meet the limitations of traditional optical imaging and wave speed testing methods, this paper proposes the combination of optical imaging and sound wave testing to comprehensively reflect and analyze the weak interlayer. This paper adopts the borehole optical and acoustic combination measurement technology, combines optical imaging and acoustic scanning modes, and effectively synchronously obtains the optical image information of the borehole rock wall and the acoustic full pulse reflection information of 360 degrees around the borehole circle, capturing the more comprehensive optical image information and acoustic scanning data. This technology also has some limitations. Because this technology uses acoustic testing technology, which needs to be carried out in the presence of water, if there is no water in the test borehole, it may be difficult to directly use this technology. When the distribution area of groundwater is not clear, the depth change of the groundwater level in the borehole can be determined through acoustic testing technology, and the surrounding rock data in the borehole area can be effectively collected when there is water. When there is no water in the test borehole, the optical image of the borehole wall can be collected normally, and the borehole can be filled with water through an external water supplement to achieve the data acquisition of borehole acoustic wave. When carrying out optical imaging in the presence of water, it is necessary to ensure the uniformity of water, that is, the water needs to have a certain light transmittance. The clearer the water is, the higher the optical image quality is. If it is muddy water or there are a lot of particles in the water, it is necessary to conduct borehole washing to ensure the water is clear.

5. Conclusions

In this paper, based on the combined measurement technology of borehole image and sound, a method of identifying weak interlayer is proposed, which combines the borehole optical image information and the rock wall acoustic response data. The two technologies adopted in this paper can complement and confirm each other’s advantages (of optical image and acoustic scanning technology), compensate for the shortage of a single data source in the process of changing information extraction, and combine the complementary characteristics of different data analysis methods to improve the accuracy and objectivity of rock mass structure data interpretation and interpretation, which can provide a scientific basis for landslide control. Combined with the case analysis, the main conclusions are as follows: (1) by adopting the borehole survey mode, combining optical imaging and acoustic scanning, the optical image information of the borehole wall and the 360-degree acoustic full pulse reflection information of the borehole circumference can be effectively captured, and more comprehensive multi-source data can be provided for the identification of weak interlayer. (2) The identification result of the weak structural plane of this method is more accurate than that of the traditional optical borehole imaging technology. It can solve the problem that the traditional optical imaging method does not consider the difference characteristics of the internal structure of the borehole rock mass, thereby effectively distinguishing the types of structural planes and dividing the areas where the weak structural plane is located. (3) The identification results of the weak structural plane in this method are more abundant than those of the traditional wave velocity testing technology, which can improve the shortcoming that the traditional wave velocity testing method can only roughly judge the sound velocity changes of rocks at different formation depths, solve the comprehensive reflection characteristics without directionality, and effectively identify and characterize the potential weak structural areas. (4) The photo-acoustic combined measurement technology brings applied innovation to the exploration of the weak interlayer of the landslide. On the one hand, it enriches and develops the existing weak layer detection technology of the landslide geological body; on the other hand, it provides a feasible way for the classification of the structural plane types in the borehole, which can improve the accurate identification ability of the weak interlayer in the landslide geological body.

Author Contributions

Conceptualization, J.W.; Formal analysis, J.W.; Funding acquisition, J.W.; Investigation, J.H.; Resources, H.M., F.W., Y.W. and Z.H.; Writing—review & editing, J.W. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

All data, models, and code generated or used during the study appear in the submitted article.

Acknowledgments

All the images and data are from our actual tests and permitted by the owners. We are compliant with ethical standards, and all authors declare that this paper has no conflict of interest. Finally, we are grateful for the many helpful and constructive comments from many anonymous reviewers.

Conflicts of Interest

The authors declare no conflict of interest.

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Figures

Figure 1. Schematic diagram of working principle.

Figure 2. Schematic diagram of borehole image expansion. (a) Acquired image, (b) center coordinate, (c) expansion diagram.

Figure 3. Schematic diagram of acoustic wave scanning data processing. (a) Scanning data, (b) circular matrix, (c) circumferential visualization.

Figure 4. Optical image and grid division. (a) Optical unfolded image, (b) grid division of optical unfolded image.

Figure 5. Acoustic wave scanning data and its matrix processing. (a) Horizontal section acoustic wave scanning data, (b) schematic diagram of matrix processing.

Figure 6. Identification block diagram of weak interlayer.

Figure 7. Physical diagram of instrument and equipment. (a) Complete system, (b) optical imaging measurement module, (c) acoustic scanning measurement module.

Figure 8. Partial data collected. (a) Video image screenshot of rock wall, (b) rock wall acoustic wave scanning data.

Figure 9. Rock wall expansion image.

Figure 10. Rock wall optical image characteristics. (a) Hole wall texture feature image, (b) hole wall integrity feature image.

Figure 11. Acoustic scanning characteristics of rock walls. (a) Hole wall acoustic feature image, (b) hole wall contour feature image.

Figure 12. Reconstructed image and binary image. (a) Fusion image, (b) binarized image.

Figure 13. Image comparison before and after well wall fusion. (a) Optical image (b) processed image.

Figure 14. Comparison of recognition before and after hole wall image fusion. (a) Optical image recognition, (b) fusion hole wall image recognition.

Figure 15. 3D images before and after borehole wall fusion. (a) 3D images before fusion, (b) 3D graphics after fusion.

Figure 16. Numerical comparison of different methods.

Figure 17. Before and after borehole wall fusion. (a) Pre-fusion image (N-S azimuth), (b) pre-fusion image (S-N azimuth), (c) post-fusion image (N-S azimuth), (d) post-fusion image (S-N azimuth).

Figure 18. 3D formation images. (a) Formation image before fusion, (b) formation image after fusion.

Figure 19. Probability value distribution of the identified results by the different methods. (a) Image recognition method, (b) this article’s recognition method, (c) sound speed recognition method.

References

1. Lateh, H.; Ahmad, J.; Mohamad, N.A. The prevalence of post-traumatic stress disorder among landslide victims. J. Psychol. Clin. Psychiatry; 2018; 9, pp. 160-167.

2. Gurung, L.; McGowran, P. Local perspectives on landslide prevention and management in Kalimpong district, West Bengal, India. Disaster Prev. Manag.; 2022; 31, pp. 193-201. [DOI: https://dx.doi.org/10.1108/DPM-02-2021-0056]

3. Xin, Q.I.; Huang, B.L.; Liu, G.N.; Wang, S.C. Landslide Susceptibility Assessment in the Three Gorges Area, China, Zigui Synclinal Basin, Using Gis Technology and Frequency Ratio Model. J. Geomech.; 2022; 23, pp. 97-104.

4. Kafle, L.; Xu, W.-J.; Zeng, S.-Y.; Nagel, T. A numerical investigation of slope stability influenced by the combined effects of reservoir water level fluctuations and precipitation: A case study of the Bianjiazhai landslide in China. Eng. Geol.; 2022; 297, pp. 508-521. [DOI: https://dx.doi.org/10.1016/j.enggeo.2021.106508]

5. Ma, C.; Hu, K.-H.; Song, G.-H.; Tian, M. Processes and Characteristics of Debris Flows Induced by Maoqiaozi Landslide in Wenchuan Earthquake Stricken Area. J. Earth Sci. Environ.; 2013; 35, pp. 98-103.

6. Ma, C.; Zhan, H.; Zhang, T.; Yao, W. Investigation on shear behavior of weak interlayers by ring shear tests. Eng. Geol.; 2019; 254, pp. 34-42. [DOI: https://dx.doi.org/10.1016/j.enggeo.2019.04.002]

7. Xiao, Y.; Wang, T.; Li, Y. Formation mechanism of complex landslide with weak interlayer reservoir bank. Coal Geol. Explor.; 2018; 46, pp. 133-137.

8. Guo, X.; Yao, Z.; Ma, W.; He, X. Mathematical model and application of engineering geological hazard risk assessment of mountain tunnel opening section. Tunn. Undergr. Eng. Disaster Prev.; 2019; 1, 10.

9. Jeon, S.; Kim, J.; Seo, Y.; Hong, C. Effect of a fault and weak plane on the stability of a tunnel in rock—A scaled model test and numerical analysis. Int. J. Rock Mech. Min. Sci.; 2004; 41, pp. 486-492. [DOI: https://dx.doi.org/10.1016/j.ijrmms.2003.12.021]

10. Innaurato, N.; Mancini, R.; Cardu, M. On the influence of rock mass quality on the quality of blasting work in tunnel driving. Tunn. Undergr. Space Technol.; 1998; 13, pp. 81-89. [DOI: https://dx.doi.org/10.1016/S0886-7798(98)00027-3]

11. Shane, M.W.; Pate, J.S. Survival strategies of the root tuberous geophyte Chamaescilla corymbosa in a Mediterranean-climate rock-outcrop environment. Aust. J. Bot.; 2015; 63, pp. 497-511. [DOI: https://dx.doi.org/10.1071/BT14220]

12. Jindra, M.; Velický, M.; Bouša, M.; Abbas, G.; Kalbáč, M.; Frank, O. Localized Spectroelectrochemical Identification of Basal Plane and Defect-Related Charge-Transfer Processes in Graphene. J. Phys. Chem. Lett.; 2022; 13, pp. 642-648. [DOI: https://dx.doi.org/10.1021/acs.jpclett.1c03466] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35020405]

13. Jiao, Y.-Y.; Wu, K.; Zou, J.; Zheng, F.; Zhang, X.; Wang, C.; Li, X.; Zhang, C. On the strong earthquakes induced by deep coal mining under thick strata-a case study. Geomech. Geophys. Geo-Energy Geo-Resour.; 2021; 7, 97. [DOI: https://dx.doi.org/10.1007/s40948-021-00301-1]

14. Tang, Z.; Wu, Z.; Zou, J. Appraisal of the number of asperity peaks, their radii and heights for three-dimensional rock fracture. Int. J. Rock Mech. Min. Sci.; 2022; 153, 105080. [DOI: https://dx.doi.org/10.1016/j.ijrmms.2022.105080]

15. Zou, J.; Jiao, Y.-Y.; Tan, F.; Lv, J.; Zhang, Q. Complex hydraulic-fracture-network propagation in a naturally fractured reservoir. Comput. Geotech.; 2021; 135, 104165. [DOI: https://dx.doi.org/10.1016/j.compgeo.2021.104165]

16. Tao, Y.; Liu, W.; Zhou, A. Identification and control strategy of landslide in engineering geological survey. Resour. Inf. Eng.; 2016; 31, 2.

17. Tian, Z. Content and treatment of landslide engineering geological survey. Resour. Inf. Eng.; 2017; 32, 2.

18. Prioul, R.; Jocker, J. Fracture characterization at multiple scales using borehole images, sonic logs, and walkaround vertical seismic profile. AAPG Bull.; 2009; 93, pp. 1503-1516. [DOI: https://dx.doi.org/10.1306/08250909019]

19. Lofi, J.; Pezard, P.; Loggia, D.; Garel, E.; Gautier, S.; Merry, C.; Bondabou, K. Geological discontinuities, main flow path and chemical alteration in a marly hill prone to slope instability: Assessment from petrophysical measurements and borehole image analysis. Hydrol. Process.; 2012; 26, pp. 2071-2084. [DOI: https://dx.doi.org/10.1002/hyp.7997]

20. Nováková, L.; Novák, P.; Brož, M.; Sosna, K.; Pitrák, M.; Kasíková j Rukavičková, L.; Maňák, L. The Results of Borehole Acoustic Imaging from a Granite in the Jihlava District, Czech Republic: Implications for Structural Geological Research. J. Geogr. Geol.; 2012; 4, 92.ISSN 1916-9779 [DOI: https://dx.doi.org/10.5539/jgg.v4n4p92]

21. Wang, J.; Wang, C.; Hu, S.; Han, Z. Study on the extraction method of structural surface parameters of hole wall borehole images. Geotech. Mech.; 2017; 38, pp. 3074-3080.

22. Wang, J.; Wang, C.; Hu, S. Detection method of borehole surrounding rock structure based on directional sound wave scanning. J. Sichuan Univ. (Eng. Sci. Ed. ); 2020; 52, pp. 118-125.

23. Wang, J.; Wang, C.; Tang, X.; Han, Z.; Wang, Y.; Hu, S. Research on karst cave detection method based on multifrequency borehole sonar. Geotech. Mech.; 2020; 40, 8.

24. Wang, J.; Xu, H.; Zou, J. Fine detection technology of rock mass structure based on borehole acousto-optic combined measurement. Measurement; 2022; 187, 110259. [DOI: https://dx.doi.org/10.1016/j.measurement.2021.110259]

25. Song, S.S.; Jia, Z.H.; Yang, J.; Nikola, K.B. Minimum spanning tree image segmentation algorithm combined with the Ostu thresholding method. Comput. Eng. Appl.; 2019; 55, 6.

26. Liu, C.; Cao, H.; Fan, Y.; Fang, W.; Shao, S. Adaptive DCP image optimization algorithm combined with the OSTU thresholding method. Fire Command Control; 2022; 47, 9.

27. Xiao, W.; Xiao, Z.; Guo, X.; Zhang, Z. Blasting crack development speed based on wavelet image processing. J. Rock Mech. Eng.; 2003; 22, 2057.

28. Wang, J. Application of wavelet transform image processing technology in financial stock analysis. J. Intell. Fuzzy Syst.; 2021; 40, pp. 2017-2027. [DOI: https://dx.doi.org/10.3233/JIFS-189204]

29. Dong, X.; Yue, X. Based on Wavelet Transform Image Reconstruction to Study Raindrops Spectrum Image Processing System. Energy Procedia; 2011; 11, pp. 939-946.