High and steep slope safety monitoring system and method

Abstract

The invention discloses a high and steep slope safety monitoring system and method, and relates to the field of high and steep slope safety monitoring, and the method comprises the steps: collecting sensing monitoring data of a target high and steep slope, an initial three-dimensional point cloud model, and the triaxial acceleration and angular velocity of a laser radar in a scanning process through deploying a multi-mode monitoring network; the point cloud configuration module generates a dynamic reference by using a laser radar and an IMU; the space-time alignment module realizes data fusion and feature extraction; the signal activation module triggers graded early warning based on stability evaluation; according to the application, data collected by different types of sensors are effectively integrated, key information reflecting the stability of the high and steep slope is extracted, the equipment layout is optimized, interference to monitoring equipment is reduced, and the accuracy and reliability of monitoring data are improved. And all-directional high-precision monitoring on the stability of the high and steep slope is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of high and steep slope safety monitoring, in particular to a high and steep slope safety monitoring system and method. BACKGROUND

[0002] Due to the complexity of the geological structure of high and steep slopes and the diversity of environmental factors, the stability of high and steep slopes is affected by various internal and external factors, including geological structure, rainfall, earthquake, weathering, etc., which may lead to slope instability and trigger geological disasters such as landslides and collapses. The conventional monitoring method is periodic detection, which has problems such as poor data accuracy and detection blind spots.

[0003] With the rapid development of technologies such as the Internet of Things and big data, high and steep slope safety monitoring technology is developing towards intelligence, automation and integration. However, the environment of high and steep slopes is complex and changeable, such as rainfall, earthquake and other natural factors, which may interfere with or damage the monitoring equipment, affecting the accuracy and reliability of the monitoring data, and the reliability of high and steep slope safety monitoring is limited.

[0004] In summary, the existing technology has the technical problems of complex and changeable environment of high and steep slopes, interference with monitoring equipment, affecting real-time evaluation of slope stability, and limited reliability of high and steep slope safety monitoring. SUMMARY

[0005] The purpose of the present application is to provide a high and steep slope safety monitoring system and method to realize comprehensive and high-precision monitoring of the stability of high and steep slopes.

[0006] To achieve the above-mentioned purpose, the present application provides the following solutions:

[0007] In a first aspect, the present application provides a high and steep slope safety monitoring system, comprising a data acquisition module, a point cloud configuration module, a space-time alignment module, a signal activation module and a safety reminding module;

[0008] The data acquisition module comprises a distributed strain sensor, a microseismic monitoring array, a groundwater level sensor, a laser radar and an inertial measurement unit; the distributed strain sensor is used to acquire strain information of a target high and steep slope; the microseismic monitoring array is used to acquire elastic wave signals generated by internal rupture events of the target high and steep slope; the groundwater level sensor is used to acquire groundwater level change signals of the target high and steep slope; the laser radar is used to scan an initial three-dimensional point cloud model of the target high and steep slope; the inertial measurement unit is used to measure three-axis acceleration and angular velocity of the laser radar during scanning;

[0009] The point cloud configuration module is configured to determine a dynamic reference point cloud model of the target high and steep slope according to the initial three-dimensional point cloud model and three-axis acceleration and angular velocity of the laser radar in a scanning process;

[0010] The space-time alignment module is configured to perform space-time alignment of the sensing monitoring data and the dynamic reference point cloud model, extract coupling features of internal stress field and displacement field of the target high and steep slope, and determine a slope stability evaluation index; the sensing monitoring data includes strain information, elastic wave signals and groundwater level change signals of the target high and steep slope; and the slope stability evaluation index includes a safety factor, a deformation rate, a stress ratio and a pore water pressure ratio;

[0011] The signal activation module is configured to identify a potential sliding surface according to the slope stability evaluation index, and activate a hierarchical early warning signal; the hierarchical early warning signal is determined according to a risk probability trigger logic; and the risk probability trigger logic is a risk probability interval of the early warning signal determined according to the slope stability evaluation index and meteorological factors;

[0012] The safety reminding module is configured to establish a response link through an edge computing node based on the hierarchical early warning signal to perform slope safety reminding.

[0013] In a second aspect, the application provides a high and steep slope safety monitoring method, which is applied to the high and steep slope safety monitoring system and includes the following steps:

[0014] Obtaining sensing monitoring data of a target high and steep slope, an initial three-dimensional point cloud model, and three-axis acceleration and angular velocity of a laser radar in a scanning process; the sensing monitoring data includes strain information, elastic wave signals and groundwater level change signals of the target high and steep slope;

[0015] Determining a dynamic reference point cloud model of the target high and steep slope according to the initial three-dimensional point cloud model and the three-axis acceleration and angular velocity of the laser radar in the scanning process;

[0016] Performing space-time alignment of the sensing monitoring data and the dynamic reference point cloud model, extracting coupling features of internal stress field and displacement field of the target high and steep slope, and determining a slope stability evaluation index; the slope stability evaluation index includes a safety factor, a deformation rate, a stress ratio and a pore water pressure ratio;

[0017] Identifying a potential sliding surface according to the slope stability evaluation index, and activating a hierarchical early warning signal; the hierarchical early warning signal is determined according to a risk probability trigger logic; and the risk probability trigger logic is a risk probability interval of the early warning signal determined according to the slope stability evaluation index and meteorological factors;

[0018] Based on the hierarchical early warning signal, a response link is established by an edge computing node to perform slope safety prompting.

[0019] According to the specific embodiments provided in the application, the application has the following technical effects:

[0020] The application provides a high and steep slope safety monitoring system and method. Sensing monitoring data, an initial three-dimensional point cloud model, and three-axis acceleration and angular velocity of a laser radar in a scanning process of a target high and steep slope are collected by deploying a multi-modal monitoring network. A point cloud configuration module generates a dynamic reference by using the laser radar and an IMU. A space-time alignment module realizes data fusion and feature extraction. A signal activation module triggers hierarchical early warning based on stability evaluation. A safety prompting module relies on an edge computing node for rapid response. The technical problems of complex and changeable environment of the high and steep slope, interference to the monitoring equipment, influence on real-time evaluation of slope stability, and limited reliability of high and steep slope safety monitoring are solved. The data collected by different types of sensors are effectively integrated, key information reflecting the stability of the high and steep slope is extracted, the equipment layout is optimized, the interference to the monitoring equipment is reduced, the accuracy and reliability of the monitoring data are improved, and omnibearing high-precision monitoring of the stability of the high and steep slope is realized. BRIEF DESCRIPTION OF DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the drawings needed in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0022] Figure 1 A structural schematic diagram of a high and steep slope safety monitoring system according to an embodiment of the application is provided.

[0023] Figure 2 A flowchart of a high and steep slope safety monitoring method according to an embodiment of the application is provided. DETAILED DESCRIPTION

[0024] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only some embodiments of the application, not all embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0025] The above purposes, features and advantages of the application will be more obvious and easy to understand. The application will be further described in detail below with reference to the drawings and specific embodiments.

[0026] In one exemplary embodiment, as shown in Figure 1 A high and steep slope safety monitoring system is provided, comprising a data acquisition module, a point cloud configuration module, a space-time alignment module, a signal activation module and a safety reminder module.

[0027] The data acquisition module comprises a distributed strain sensor, a microseismic monitoring array, a groundwater level sensor, a laser radar and an inertial measurement unit; the distributed strain sensor is used to acquire strain information of a target high and steep slope; the microseismic monitoring array is used to acquire elastic wave signals generated by internal rupture events of the target high and steep slope; the groundwater level sensor is used to acquire underground water level change signals of the target high and steep slope; the laser radar is used to scan an initial three-dimensional point cloud model of the target high and steep slope; and the inertial measurement unit is used to measure three-axis acceleration and angular velocity of the laser radar during scanning.

[0028] The point cloud configuration module is used to determine a dynamic reference point cloud model of the target high and steep slope according to the initial three-dimensional point cloud model and the three-axis acceleration and angular velocity of the laser radar during scanning.

[0029] The space-time alignment module is used to perform space-time alignment of sensing monitoring data and the dynamic reference point cloud model, extract coupling characteristics of internal stress field and displacement field of the target high and steep slope, and determine a slope stability evaluation index; the sensing monitoring data includes strain information, elastic wave signals and underground water level change signals of the target high and steep slope; and the slope stability evaluation index includes a safety factor, a deformation rate, a stress ratio and a pore water pressure ratio. The calculation core of the slope stability evaluation index is to determine whether the slope is stable by quantifying the relative relationship between the "anti-sliding force" and the "sliding force" of the slope, or analyzing the deformation characteristics. The most core index is the safety factor, and in addition, there are dynamic evaluation indexes based on deformation monitoring.

[0030] The signal activation module is used to identify a potential sliding surface according to the slope stability evaluation index, and activate a hierarchical early warning signal; the hierarchical early warning signal is determined according to a risk probability trigger logic; and the risk probability trigger logic is a risk probability interval of the early warning signal determined according to the slope stability evaluation index and meteorological factors.

[0031] The safety reminder module is used to establish a response link through an edge computing node based on the hierarchical early warning signal to perform slope safety reminding.

[0032] In actual application, under the constraint of the layout density of slope geological parameters, a multi-modal sensor network (data acquisition module) is deployed, which includes a distributed strain sensor, a microseismic monitoring array and a groundwater level sensor.

[0033] Distributed strain sensors refer to sensors that continuously measure the internal strain distribution of materials or structures, commonly used to monitor the deformation and stress state of structures, and convert strain information into electrical or optical signals for transmission and processing through optical fiber or strain gauge sensors; Microseismic monitoring array is a monitoring system composed of multiple microseismic sensors, which is used to detect the elastic wave signals generated by the micro-fracture events in the rock-soil mass, and thus analyze the stability of the rock-soil mass; Groundwater level sensor is used to monitor the change of groundwater level in real time, and the fluctuation of groundwater level will affect the strength and stability of rock-soil mass, which is one of the important parameters for slope stability monitoring; Multi-modal sensor network refers to the integration of multiple different types of sensors to achieve comprehensive monitoring of various physical quantities and state information of the slope.

[0034] Under the constraint of the layout density of the slope geological parameters, reasonable deployment of multi-modal sensor network is the basis for realizing the safety monitoring of high and steep slope. Specifically, according to the geological parameters of the slope (such as lithology, structural plane, slope, etc.) and the layout density requirements, the installation position and quantity of distributed strain sensors, microseismic monitoring array and groundwater level sensors are determined. Generally, distributed strain sensors can be arranged along the potential slip surface and key structural parts of the slope, such as in the middle, lower part and deformation-prone areas of the slope, to monitor the strain distribution and deformation trend inside the slope; Microseismic monitoring array can be arranged around and at the bottom of the slope to capture information such as the occurrence position, size and frequency of microseismic events inside the slope; Groundwater level sensors are installed in sensitive areas of groundwater level change in the slope, such as aquifers and rock-soil mass permeability good parts, to monitor the rise and fall of groundwater level in real time.

[0035] By reasonably deploying multi-modal sensor network, comprehensive and accurate information of stress field, displacement field and hydrological conditions inside the slope can be obtained, which is the basis for accurately evaluating the stability of the slope. At the same time, microseismic monitoring array can monitor the occurrence of microseismic events inside the slope in real time, providing an important basis for analyzing the fracture mechanism and stability of the rock-soil mass of the slope; Groundwater level sensors can reflect the change of groundwater level in real time, and timely understand the influence of groundwater on the stability of the slope, so as to take corresponding measures to cope with it. In summary, reasonable deployment of multi-modal sensor network is the basis for improving the accuracy and reliability of monitoring data and realizing comprehensive and accurate monitoring of the stability of high and steep slope.

[0036] In an embodiment, the data acquisition module further comprises: a vibration optical fiber sensor array and an inclination sensor cluster; the layout direction of the vibration optical fiber sensor array is orthogonal to the potential slip direction; the inclination sensor cluster is embedded in the surface anchoring point of the target high and steep slope in a honeycomb topology structure.

[0037] The vibration fiber sensor array is used for detecting vibration and deformation inside the target high and steep slope.

[0038] The inclination sensor cluster is used for measuring the inclination change signal of the surface layer anchorage point.

[0039] In practical applications, according to the fissure development degree of the slope rock mass and the groundwater level distribution, a high-risk monitoring sub-region (target high and steep slope) is determined; a vibration fiber sensor array is configured in the high-risk monitoring sub-region, and the layout direction of the vibration fiber sensor array is orthogonal to the potential sliding direction; an inclination sensor cluster is synchronously deployed, and the inclination sensor cluster is embedded in the surface layer anchorage point of the slope in a honeycomb topology structure; based on the rainfall infiltration coefficient of the slope, the sampling frequency and data uploading interval of the groundwater level sensor are adjusted.

[0040] Specifically, the fissure development degree of the rock mass refers to the distribution density, width, length and other characteristics of the fissures (such as joints, cracks, etc.) in the rock mass, which reflects the integrity and permeability of the rock mass. The more, the wider and the longer the fissures are, the poorer the stability of the rock mass is in general; the groundwater level distribution refers to the position and change of the groundwater in the slope rock-soil mass, and the height and change of the groundwater level have an important influence on the slope stability, and high water level will reduce the shear strength of the rock-soil mass; the vibration fiber sensor array is a monitoring system based on optical fiber vibration sensing technology, which can detect the micro vibration and deformation inside the rock-soil mass, and the layout direction thereof is orthogonal to the potential sliding direction in order to more sensitively capture the early signs of landslide; the inclination sensor cluster is a cluster composed of multiple inclination sensors, which is used for measuring the inclination change of the surface layer anchorage point of the slope, and the honeycomb topology structure is a network topology structure similar to a honeycomb, which has high connectivity and stability, and can ensure the data transmission and cooperative work between the sensor clusters; the rainfall infiltration coefficient of the slope is a parameter describing the difficulty of rainfall water penetrating into the slope rock-soil mass, and the slope rainfall infiltration coefficient affects the change of the groundwater level and the saturation of the rock-soil mass, and further affects the stability of the slope.

[0041] Through geological exploration and geophysical prospecting means such as drilling, geological radar, etc., the development of rock mass fissures is obtained, including parameters such as fissure density, spacing, and strike. At the same time, by analyzing the distribution and change trend of the underground water level in the slope, if the underground water level in the location of the fissure development zone rises significantly during the rainy season, the region is determined as a high-risk monitoring sub-region, and a vibration optical fiber sensor array is configured in the region, which is arranged orthogonally to the potential sliding direction, and can effectively capture the slight vibration and deformation of the rock-soil mass in the potential sliding direction. Preferably, when the rock-soil mass slips, it will produce obvious vibration and strain change in the sliding direction, and the orthogonally arranged sensor array can more sensitively detect these changes. For example, when the potential sliding direction is northeast-southwest, the vibration optical fiber sensor array is arranged along the northwest-southeast direction. When the rock-soil mass slips, the sensor can detect the vibration signal perpendicular to the sliding direction, thereby warning the landslide risk in advance.

[0042] A cluster of inclination sensors is synchronously deployed to embed the slope surface anchor points in a cellular topology structure, which can monitor the inclination changes of the slope surface in real time. Specifically, the cellular topology structure of the cluster of inclination sensors ensures high connectivity between the sensors and stability of data transmission. Preferably, the cluster of inclination sensors is arranged in a cellular manner at the anchor points on the slope surface, and each sensor node forms a stable communication connection with adjacent nodes. When the slope surface changes in inclination, the inclination sensor can measure the change in inclination angle in real time and transmit data to the monitoring center through the cellular topology network.

[0043] Based on the rainfall infiltration coefficient of the slope, the sampling frequency and data upload interval of the underground water level sensor are adjusted, which can improve the timeliness and pertinence of the monitoring data. Specifically, the rainfall infiltration coefficient reflects the speed of rainfall water entering the slope rock-soil mass. By analyzing this coefficient, the influence time and degree of rainfall on the underground water level can be predicted. In the slope area with a high rainfall infiltration coefficient, rainfall water can quickly penetrate into the rock-soil mass, causing the underground water level to rise rapidly. Exemplarily, before the rainy season or the arrival of rainfall, the sampling frequency of the underground water level sensor is adjusted from once an hour to once every 10 minutes to obtain the underground water level change data more timely, thereby providing key data support for analyzing the influence of rainfall on the stability of the slope.

[0044] By determining the high-risk monitoring sub-region according to the degree of development of the slope rock mass fissures and the distribution of the underground water level, and configuring a vibration optical fiber sensor array and a cluster of inclination sensors in the region, and adjusting the sampling frequency and data upload interval of the underground water level sensor based on the rainfall infiltration coefficient, precise and efficient monitoring of the high and steep slope can be achieved, thereby providing strong data support for the safety management and early warning decision of the slope.

[0045] In an embodiment, the point cloud configuration module comprises:

[0046] a filtering unit configured to filter the three-axis acceleration and angular velocity of the laser radar during scanning to obtain attitude change information and position change information of the laser radar.

[0047] a point cloud distortion compensation unit configured to perform geometric transformation on the initial three-dimensional point cloud model according to the attitude change information and the position change information to obtain a dynamic reference point cloud model of the target high and steep slope; the geometric transformation includes rotation, translation and scaling.

[0048] a cumulative displacement error calculation unit configured to calculate the cumulative displacement error of the corrected point cloud data and the last reference point cloud data.

[0049] a model updating unit configured to, when the cumulative displacement error exceeds a preset error threshold, cause the laser radar to re-scan the target high and steep slope to update the dynamic reference point cloud model of the target high and steep slope.

[0050] In actual application, the point cloud configuration module is configured to scan the deformation features of the slope surface using the laser radar and correct the spatial drift error synchronously using the IMU unit to configure the dynamic reference point cloud.

[0051] Specifically, the laser radar (LiDAR) is a remote sensing technology that measures the distance and surface features of target objects by emitting laser beams and receiving reflected light, has the characteristics of high precision and high resolution, and can quickly obtain three-dimensional deformation information of the slope surface; the unmanned aerial vehicle is an aircraft that can fly autonomously or under remote control, has the characteristics of strong maneuverability and high flexibility, and can be used to carry the laser radar to scan the slope surface; the three-dimensional point cloud model refers to a three-dimensional model constructed by a large amount of point cloud data obtained by scanning using the laser radar, and can intuitively display the geometric shape and features of the slope surface; the IMU (Inertial Measurement Unit) refers to a sensor that measures the three-axis acceleration, angular velocity and magnetic field intensity of an object in real time, can be used to correct the spatial drift error of the laser radar during scanning, and improve the accuracy of the scanning data; point cloud distortion compensation refers to correcting the distortion of the laser radar scanning data caused by factors such as device movement and vibration to improve the accuracy and reliability of the point cloud data; dynamic reference point cloud refers to reference point cloud data updated at certain time intervals or conditions, used to compare and analyze the changes in the deformation of the slope surface. The reference point cloud updating mechanism refers to updating the reference point cloud data according to certain rules and conditions to ensure the timeliness and accuracy of the reference point cloud.

[0052] The laser radar is used for scanning the surface of the slope to obtain a large amount of point cloud data, and the point cloud data contains three-dimensional coordinate information and deformation characteristics of the surface of the slope. However, during the scanning process, due to the movement, vibration of the laser radar equipment and the influence of the surrounding environment, spatial drift errors may occur in the scanning data, thereby affecting the accuracy of the deformation monitoring. Therefore, the IMU is used to measure the acceleration and angular velocity of the laser radar during the scanning process in real time, and the spatial drift error of the scanning data of the laser radar is corrected by using a filtering algorithm and other data processing methods, so as to improve the accuracy and reliability of the point cloud data. Specifically, when the unmanned aerial vehicle is used to carry the laser radar to scan the slope, the flight attitude and speed change of the unmanned aerial vehicle may cause errors in the scanning data, and the IMU can capture these changes in real time and feed them back to the data processing system, so as to accurately correct the errors in the scanning data.

[0053] After the unmanned aerial vehicle is used to carry the laser radar, the surface of the slope can be scanned in all directions and at multiple angles according to the preset flight path and scanning task. Each scanning covers a certain area, and finally a large amount of point cloud data is obtained. The point cloud data contains three-dimensional coordinate information (X coordinate, Y coordinate, Z coordinate) and reflected light intensity information of each scanning point on the surface of the slope. By processing and splicing these data, an initial three-dimensional point cloud model is generated. The generated initial three-dimensional point cloud model can intuitively show the geometric shape of the surface of the slope, including cracks, protrusions, depressions and other geological features.

[0054] The key nodes are usually selected at positions such as the top, middle and bottom of the slope and the potential sliding surface. The motion state of these positions has an important influence on the stability of the slope. The IMU measures the acceleration and angular velocity of the key nodes in three axial directions in real time. The data acquisition frequency is generally set to 100 Hz (i.e. 100 groups of data are collected per second) to ensure the continuity and accuracy of the data. Preferably, the IMU deployed at the top of the slope can measure the acceleration of the top of the slope in the horizontal direction (X axis and Y axis) and the angular velocity around the vertical axis (Z axis). These data reflect the vibration and swing of the top of the slope. Through a wireless communication module (such as a 4G / 5G module), the IMU transmits the collected data to the data processing system of the monitoring center in real time.

[0055] After obtaining high-precision point cloud data, a dynamic reference point cloud is configured, i.e. the point cloud data obtained at different times is compared and analyzed to establish a dynamically updated reference point cloud model.

[0056] Due to the inevitable attitude changes (such as roll, pitch, yaw) and acceleration changes (such as acceleration, deceleration, turning) of the UAV during flight, these changes will cause the scanning direction and position of the laser radar to deviate, thereby causing distortion of the point cloud data. The process of distortion compensation is as follows: through Kalman filtering algorithm, the acceleration and angular velocity data collected by the IMU are filtered to remove noise interference and obtain accurate attitude and position change information; according to these change information, geometric transformation (such as rotation, translation, scaling) is performed on the initial three-dimensional point cloud model to correct the distorted point cloud data to the real geometric position and ensure the accuracy of the point cloud data.

[0057] After each completion of point cloud distortion compensation, the corrected point cloud data is compared with the previous reference point cloud data to calculate the cumulative displacement error, which can be obtained by calculating the average Euclidean distance between the same named points (i.e. points at the same geographical position) in the point cloud data; the preset error threshold is set according to the stability requirements and monitoring accuracy requirements of the slope, usually between 5-10 cm; when the monitoring system detects that the cumulative displacement error reaches the preset error threshold, a full-area rescan task is triggered immediately; the UAV carries the laser radar again to perform comprehensive scanning on the entire slope area, generating a new reference point cloud model, and updating the attitude and position reference data of the IMU; the new reference point cloud model obtained by rescan accurately reflects the surface morphology change after the slope landslide, ensuring the long-term accuracy and reliability of the slope surface deformation monitoring.

[0058] By using the UAV loaded with laser radar to perform multi-angle scanning of the slope surface, deploying IMU at key nodes for data collection, performing point cloud distortion compensation, and establishing a reference point cloud updating mechanism, high-precision and dynamic monitoring of the surface deformation of high and steep slopes can be achieved, effectively addressing the deformation challenges of slopes in complex natural environments, and providing reliable data basis for slope stability analysis and early warning.

[0059] Further, when the slope surface deforms, the real-time scanned point cloud data is compared with the dynamic reference point cloud to accurately detect the deformation characteristics and change trend of the slope surface. Specifically, by comparing and analyzing the point cloud data before and after the rainy season, it is found that local sliding and crack expansion phenomenon occurs on the slope surface, and the instability risk of the slope is timely warned. In addition, when the cumulative displacement error exceeds the preset error threshold, full-area rescan is triggered to ensure the accuracy and reliability of the reference point cloud, effectively addressing the complexity and uncertainty in the slope deformation process, and providing high-precision surface deformation monitoring data for subsequent slope stability evaluation and early warning.

[0060] By comparing the point cloud data of the slope before and after the rainy season, it is found that the slope surface has local slip. Further, the three-dimensional coordinate deviation of the corresponding area (10m*10m grid unit) in the real-time point cloud and the dynamic reference point cloud is calculated. If the average displacement is greater than or equal to 5cm (adjusted according to the slope stability level, 3cm for high-risk areas), and the displacement direction is consistent, generally, the displacement direction is along the slope surface downward or along the inclined direction of the potential sliding surface, it is determined that there is local slip. Differently, the slip area needs to show continuous spatial distribution, that is, the points with displacement exceeding the threshold form a connected domain and the connected domain area meets the minimum connected domain area threshold. If the dispersed isolated points exceed the threshold, it is determined that it is measurement error or local rock mass loosening, but not slip. In short, by distinguishing the displacement distribution pattern, the displacement of each point in the slip area is uniform, that is, the slip is a surface whole displacement.

[0061] By comparing the point cloud data of the slope before and after the rainy season, it is found that the slope surface has local crack expansion phenomenon. Further, the crack length extracted from the real-time point cloud is compared with the reference point cloud. If the maximum length increment is greater than the crack length threshold value, it is determined that it is longitudinal expansion. The transverse distance of the points on both sides of the crack is analyzed by point cloud slicing. If the maximum width increment is greater than the crack width threshold value, it is determined that it is transverse expansion. Combined with the echo intensity data of the laser radar, if the reflectivity of the point cloud in the crack area suddenly decreases, it indicates that the crack deepening leads to the increase of laser penetration. The depth estimation value is determined by fitting the crack profile of the multi-view point cloud. If the depth estimation value is greater than the crack depth threshold value, it is determined that it is depth expansion. In short, by distinguishing the displacement distribution pattern, there is only significant displacement along the two sides of the crack line in the crack area, and the middle is a broken zone without displacement, that is, the crack is a linear discrete cracking.

[0062] In an embodiment, the space-time alignment module comprises:

[0063] An energy extraction unit is configured to perform wavelet packet decomposition on the elastic wave signal to extract energy distribution features of different frequency bands.

[0064] An interpolation unit is configured to convert the strain information into an equivalent plastic strain field, and map the equivalent plastic strain field to a three-dimensional finite element grid model of the target high and steep slope by using a finite element interpolation method.

[0065] The alignment unit is used to align the groundwater level change signal with the surface displacement monitoring sequence after mapping the equivalent plastic strain field to the three-dimensional finite element mesh model of the target high-steep slope, combining the energy distribution characteristics of different frequency bands, quantifying the coupling strength between the sensing monitoring data, and generating a fusion weight matrix. The surface displacement monitoring sequence is obtained by a specific monitoring technology, which precisely measures the surface position of the target area (such as a slope, building, roadbed, etc.) at different time points. These discrete "time-position" data are arranged in chronological order to form a continuous surface displacement data set (i.e., monitoring sequence) that changes over time. The purpose of the fusion weight matrix is to reasonably weight the multi-source data (or multi-model, multi-feature) to highlight high-value information and suppress redundant / noisy information, ultimately improving the accuracy, reliability, and robustness of the fusion result.

[0066] Coupling strength is a quantitative indicator that measures the "strength" of the interaction between systems, and coupling characteristics are qualitative / quantitative manifestations that describe the "way and attributes" of the interaction between systems. The relationship between the two is that coupling strength determines the prominence of coupling characteristics, and coupling characteristics reflect the action form of coupling strength.

[0067] The purpose of quantifying the coupling strength between sensing monitoring data is to uncover the correlation rules of elements within the system from seemingly independent monitoring data, providing data support for system state evaluation, abnormal diagnosis, optimization control, and prediction decision-making.

[0068] In practical applications, the space-time alignment module is used to perform space-time alignment on the sensing monitoring data collected by the multi-modal sensor network and the dynamic reference point cloud, extract the coupling characteristics of the internal stress field and displacement field of the slope, and configure the slope stability evaluation index.

[0069] Specifically, space-time alignment refers to the unified space-time coordinate transformation and matching of multi-source data obtained at different times and different spatial positions, so that these data can be fused and analyzed in the same space-time framework. The internal stress field of the slope refers to the stress distribution state of each point in the slope rock-soil mass, including normal stress, shear stress, etc., which reflects the evolution of the internal mechanical state of the slope. The displacement field refers to the position change of each point in the slope rock-soil mass, including horizontal displacement, vertical displacement, etc., which is an important representation of slope deformation. The coupling characteristics refer to the relationship and characteristics of the interaction and influence between the stress field and the displacement field. By analyzing the coupling characteristics, the mechanical behavior and stability of the slope can be more accurately understood. The slope stability evaluation index is a parameter system for quantitatively evaluating the stability of the slope, usually including safety factor, deformation rate, stress ratio, etc., which is used to judge the stability state and risk level of the slope.

[0070] The data collected by the multi-modal sensors such as distributed strain sensors, microseismic monitoring arrays and groundwater level sensors are unified in time and space. Specifically, in the time dimension, the data collected by different sensors at different frequencies are time-synchronized, and further, the data collected by the distributed strain sensors every second, the data collected by the microseismic monitoring arrays every minute and the data collected by the groundwater level sensors every hour are unified into the same time sequence for subsequent data analysis and comparison. In the spatial dimension, the data collected by the sensors are matched with the three-dimensional geographic coordinate system of the slope, so that these data can be fused and visually displayed in the same spatial coordinate framework. Further, through the geographic information system (GIS) platform, the strain data, microseismic event location data and groundwater level data collected by the sensors are spatially superimposed with the topographic and geomorphic data of the slope to form a complete spatial distribution map of the slope monitoring data.

[0071] The coupling characteristics of the stress field and displacement field inside the slope are extracted. Specifically, through the analysis of the strain sensor data, the stress distribution and changes inside the slope are obtained. Through the processing of the microseismic monitoring array data, the occurrence position, size and frequency of the microseismic events are determined, and then the rupture and sliding trend of the slope rock-soil mass are inferred. In combination with the data of the groundwater level sensor, the influence of the change of the groundwater level on the stress field and displacement field of the slope can be analyzed. Through the fusion analysis of these data, the coupling relationship between the stress field and displacement field can be extracted. Specifically, the stress concentration area is often accompanied by large displacement changes, and the rise of the groundwater level may lead to the reduction of the shear strength at the bottom of the slope, thereby exacerbating the deformation and instability risk of the slope.

[0072] Through the spatio-temporal alignment and coupling feature extraction of the multi-modal sensor data, it is found that the stress concentration area in the middle of the slope highly coincides with the abnormal change area of the displacement field, and during the rise of the groundwater level in the rainy season, the strain and displacement rate of this area increase significantly, indicating that this area is a potential sliding weak part of the slope.

[0073] The slope stability evaluation index is configured. Specifically, according to the geological conditions of the slope, the monitoring data and the engineering experience, appropriate stability evaluation indexes such as safety factor, deformation rate, stress ratio and pore water pressure ratio are selected. The safety factor is a commonly used index to measure the stability of the slope, and the larger the value is, the more stable the slope is. The deformation rate reflects the speed of the slope deformation, and when the deformation rate exceeds a certain threshold, the slope may have a risk of instability. The stress ratio is the ratio of the actual stress in the slope rock-soil mass to the ultimate resistance, which is used to evaluate whether the stress state of the slope is close to the limit equilibrium. The pore water pressure ratio reflects the influence of the groundwater on the stability of the slope.

[0074] The collected sensing monitoring data of the multi-modal sensor network and the dynamic reference point cloud are spatio-temporally aligned, the coupling characteristics of the internal stress field and displacement field of the slope are extracted, and the slope stability evaluation index is configured. According to the analysis of the slope stability evaluation index, the stability of the slope can be quantitatively evaluated, which provides a scientific basis for the safety management and early warning decision of the slope, and further realizes the all-around precision monitoring of the stability of the high and steep slope, and provides strong technical support for the safe operation and disaster prevention of the slope.

[0075] The collected sensing monitoring data of the multi-modal sensor network and the dynamic reference point cloud are spatio-temporally aligned, and specifically include:

[0076] The microseismic event signal is subjected to wavelet packet decomposition to extract energy distribution characteristics of different frequency bands; the strain sensor data is converted into an equivalent plastic strain field and mapped to a three-dimensional grid model by finite element interpolation; after mapping to the three-dimensional grid model, the underground water level change curve and the surface displacement monitoring sequence are aligned, and the coupling strength between the sensing monitoring data is quantified and a fusion weight matrix is generated in combination with the energy distribution characteristics of different frequency bands.

[0077] The microseismic event signal refers to the elastic wave signal generated by the micro-fracture event in the rock-soil mass, which contains information such as the position, size and direction of the rock-soil mass fracture; wavelet packet decomposition can decompose the signal into sub-bands of different frequency bands to extract the characteristics of the signal in different frequency ranges; the energy distribution characteristics refer to the energy proportion and distribution of the signal in different frequency bands, which reflect the frequency characteristics and energy concentration degree of the signal; the equivalent plastic strain field refers to the strain data collected by the strain sensor, which is converted into an equivalent strain field that can reflect the plastic deformation of the material, and is used to analyze the plastic deformation of the rock-soil mass; finite element interpolation mapping refers to the interpolation mapping of data from one spatial position to another spatial position by finite element grid, realizing the spatial fusion of data; the three-dimensional grid model refers to the division of the geometry of the slope into a three-dimensional finite element grid model, which is used for numerical simulation and data fusion; the underground water level change curve refers to the curve of the change of the underground water level with time, which reflects the dynamic change of the underground water level; the surface displacement monitoring sequence refers to the sequence formed by arranging the surface displacement monitoring data of the slope in time sequence, which reflects the spatio-temporal variation characteristics of the surface displacement of the slope; the coupling strength refers to the degree of mutual correlation and mutual influence between different monitoring data; the fusion weight matrix refers to the weight coefficient matrix determined according to the coupling strength between the data, which is used to determine the weight of each data source in data fusion.

[0078] Microseismic event signals are typically acquired using sensors in a microseismic monitoring array, with the sampling frequency generally set to 1kHz-10kHz. The acquired microseismic monitoring signals undergo preprocessing, including noise removal and filtering. Preferably, a Butterworth filter is used for bandpass filtering to remove noise components. The preprocessed microseismic monitoring signals are then subjected to wavelet packet decomposition. Appropriate wavelet basis functions and decomposition levels (usually 3-4 levels) are selected to decompose the signal into sub-bands of different frequency bands. For example, the resulting frequency bands could be 0Hz-10Hz, 10Hz-50Hz, 50Hz-100Hz, 100Hz-500Hz, 5... The energy values ​​of sub-band signals in the 00Hz-1000Hz range are calculated, typically using the integral square method. Energy distribution characteristics of different frequency bands are extracted, i.e., the proportion of energy in each band to the total energy is calculated. If the energy proportion in the 50Hz-100Hz band reaches 35%, and the energy proportion in the 100Hz-500Hz band is 40%, these high-frequency bands indicate that the microseismic event may be related to rapid fracturing and frictional sliding within the soil and rock mass. By analyzing the energy distribution characteristics of different microseismic events, different types of microseismic events can be distinguished, such as rock fracturing, rock mass sliding, and groundwater flow, providing an important basis for slope stability analysis.

[0079] Strain sensors are typically distributed at multiple locations within a slope, and the collected data reflects the strain conditions in a local area. To convert these discrete strain data into a continuous equivalent strain field, a three-dimensional finite element mesh model of the slope is established. Based on the geological conditions and geometry of the slope, a three-dimensional mesh is generated using finite element software (such as ABAQUS, ANSYS, etc.), typically using eight-node hexahedral elements. Then, the data collected by the strain sensors is converted into equivalent strain values. The equivalent strain can be calculated using the von Mises strain criterion or other suitable criteria. For example, for three principal strains ε1, ε2, and ε3 collected by a three-dimensional strain sensor, the equivalent strain ε... eq You can follow the formula The calculation is performed; the equivalent strain value is interpolated from the sensor position to the entire three-dimensional mesh model using finite element interpolation. During the interpolation process, the strain value is distributed to the adjacent mesh nodes according to the position of the sensor in the mesh, using methods such as linear interpolation or spline interpolation, and the equivalent strain field of the entire model is formed by assembling the elements.

[0080] Aligning groundwater level change curve and surface displacement monitoring sequence: Groundwater level change curve and surface displacement monitoring sequence usually have different time sampling frequencies, which need to be aligned in time series; By interpolating groundwater level data, it has the same time sampling frequency as surface displacement data; The aligned groundwater level change curve and surface displacement monitoring sequence are fused with the energy distribution characteristics of microseismic monitoring signals for analysis, specifically, during the rainy season, the rising groundwater level leads to the reduction of the strength of the slope rock-soil body, and the surface displacement rate accelerates, while the high-frequency energy proportion in the microseismic event signal increases; By establishing a correlation model between groundwater level, surface displacement and microseismic energy distribution, the coupling strength between them is quantified; The coupling strength can be measured by calculating the Pearson correlation coefficient, mutual information and other indicators; According to the Pearson correlation coefficient between groundwater level change and surface displacement rate, and the mutual information value between surface displacement rate and high-frequency energy proportion of microseismic events, a fusion weight matrix is generated, the elements of the fusion weight matrix represent the weight coefficients of each monitoring data in the slope stability evaluation, and the weight coefficients can be determined by principal component analysis, analytic hierarchy process and other methods.

[0081] By wavelet packet decomposition of microseismic event signals, converting strain sensor data into equivalent plastic strain field and mapping to three-dimensional grid model, and quantifying the coupling strength between monitoring data and generating fusion weight matrix, the deep fusion and stability comprehensive evaluation of multi-source monitoring data of high and steep slope are realized, which provides comprehensive and accurate data support and technical support for the safety monitoring and early warning of slope.

[0082] In practical application, the signal activation module is used to identify potential slip surface with the slope stability evaluation index, activate hierarchical warning signal, the hierarchical warning signal is associated with risk probability trigger logic, and the risk probability trigger logic is coupled with meteorological data.

[0083] Specifically, the slope stability evaluation index is a parameter system determined according to the geological conditions, stress field, displacement field and other information of the slope, which is used to quantitatively evaluate the stability state of the slope; The potential slip surface refers to the sliding failure surface that may occur in the rock-soil mass of the slope, and its position and form are the key to slope stability analysis; The hierarchical warning signal is the warning information of different levels according to the risk degree of slope stability, which is used to prompt to take appropriate measures in time; The risk probability trigger logic is to determine the risk probability threshold of the warning signal according to the slope stability evaluation index and related factors, when the monitoring data reaches or exceeds the threshold, the corresponding warning signal is triggered; Meteorological data coupling refers to the combination of meteorological factors (such as rainfall, rainfall intensity, wind speed, etc.) with slope stability evaluation and warning, because meteorological conditions have a significant impact on slope stability, such as rainfall may cause groundwater level rise, rock-soil body strength reduction, etc.

[0084] According to the slope stability evaluation index system, the stability of the slope is quantitatively evaluated. Specifically, by calculating the safety factor, deformation rate, stress ratio and other indicators of the slope, comparison is made with the pre-set threshold value; when these indicators exceed the normal range, it indicates that the slope may have a risk of instability. Then, by using numerical simulation and data analysis methods, the potential sliding surface of the slope is identified, the stress distribution and deformation of the slope are analyzed by finite element analysis, and the position and shape of the potential sliding surface are determined in combination with the geological conditions and monitoring data. Preferably, by analyzing the slope stability evaluation index, it is found that the potential sliding surface exists in the stress concentration area and the displacement abnormal change area in the middle of the slope, the safety factor is lower than the design requirement, and the deformation rate and stress ratio exceed the early warning threshold.

[0085] According to the risk degree of the slope stability, the early warning signals are divided into different levels, such as low-risk early warning, medium-risk early warning and high-risk early warning; when the slope stability evaluation index reaches or exceeds the threshold value of the corresponding level, the corresponding early warning signal is triggered, and the graded early warning signal is associated with the risk probability trigger logic. According to historical data and statistical analysis, the risk probability interval corresponding to different early warning signals is determined. Specifically, the risk probability corresponding to the low-risk early warning is 10%-30%, the risk probability corresponding to the medium-risk early warning is 30%-60%, and the risk probability corresponding to the high-risk early warning is more than 60%. In addition, the risk probability trigger logic is coupled with meteorological data; real-time meteorological data such as rainfall, rainfall intensity and wind speed are obtained, and the influence of meteorological conditions on the stability of the slope is analyzed. Preferably, during the rainy season, when the rainfall intensity exceeds a certain threshold value and the duration is relatively long, the risk probability of slope instability is increased, and the level of the early warning signal is adjusted accordingly. In summary, by identifying the potential sliding surface through the slope stability evaluation index, activating the graded early warning signal, and associating it with the risk probability trigger logic and coupling it with meteorological data, real-time evaluation and early warning of the safety risk of high and steep slopes can be realized, providing strong support for the safety management and disaster prevention and control of slopes.

[0086] Further, since the graded early warning signal is associated with the risk probability trigger logic, the signal activation module is also used to:

[0087] According to the rock mass strength parameters and joint surface occurrence data, the pore water pressure distribution under different rainfall conditions is predicted, and the sliding surface search direction is iteratively corrected to determine the minimum safety factor; connect the historical landslide event library to build a landslide event knowledge graph; when the minimum safety factor is lower than the safety factor threshold, determine the risk probability interval according to the matching similar historical landslide scene in the landslide event knowledge graph, and the risk probability interval is used to trigger the graded early warning signal.

[0088] Specifically, the rock mass strength parameters refer to parameters that describe the rock mass's resistance to deformation and failure, including compressive strength, shear strength, elastic modulus, etc., reflecting the mechanical properties of the rock mass; the joint surface occurrence data refer to the spatial geometric characteristics of the joint surface (cracks or weak planes in the rock mass), including the strike, dip, and inclination of the joint surface, as well as the roughness and spacing of the joint surface, etc., which often serve as potential sliding surfaces for rock slope landslides; the backpropagation neural network refers to a neural network that adjusts network weights through error backpropagation to achieve nonlinear mapping of input-output relationships, commonly used for prediction and classification tasks; the pore water pressure distribution refers to the spatial distribution of the pressure generated by groundwater in the pores of rock-soil bodies, and changes in pore water pressure will affect the effective stress of rock-soil bodies, thereby affecting the stability of slopes; the sliding surface search direction refers to the direction and path of searching for potential sliding surfaces in slope stability analysis, and by continuously adjusting the search direction, the location and shape of the sliding surface can be more accurately determined; the minimum safety factor refers to the minimum safety factor value obtained through calculation in slope stability analysis, reflecting the stability of the slope under the most unfavorable conditions, and the smaller the safety factor, the higher the risk of slope instability; the historical landslide event library refers to a database that stores relevant data of past landslide events (such as landslide time, location, scale, and triggering factors); the landslide event knowledge graph is a knowledge graph constructed by mining and analyzing historical landslide event data, showing the correlation between landslide events, characteristic patterns, and evolution laws; the risk probability interval refers to the probability range of the risk of slope landslide determined based on slope stability analysis and historical landslide event knowledge graph, used to trigger graded warning signals.

[0089] Collect the rock mass strength parameters (such as compressive strength, shear strength, etc.) and joint surface occurrence data (such as strike, dip, inclination, etc.) of the slope, construct a backpropagation neural network model, the input layer neurons include rock mass strength parameters and joint surface occurrence data, and the output layer neurons are pore water pressure distribution characteristic parameters (such as pore water pressure value, distribution range, etc.); select appropriate training samples (such as historical monitoring data, numerical simulation data, etc.) to train the backpropagation neural network model, for example, use the pore water pressure monitoring data of the slope under different rainfall amounts (such as 50mm, 100mm, 150mm) and rainfall durations (such as 1 hour, 6 hours, 12 hours) as training samples, and after multiple iterations of training, the prediction error of the backpropagation neural network model is controlled within 5%; after training, use the backpropagation neural network model to predict the pore water pressure distribution under different rainfall conditions, providing key data support for subsequent slope stability analysis.

[0090] Meanwhile, the minimum safety factor is determined by iteratively correcting the search direction of the slip surface, and a mechanical model of the slope is established by using a numerical simulation method (such as a finite element method, a finite difference method, etc.), the initial search direction of the slip surface can be determined based on the occurrence data of the joint surface and the structural characteristics of the rock mass, for example, the initial slip surface is assumed to be a circular arc-shaped slip surface along the inclination direction of the joint surface, in the calculation process, the shape and position of the slip surface are adjusted according to the simulation results and the predicted value of the pore water pressure distribution, the search direction of the slip surface is iteratively corrected, for example, it is found that the safety factor corresponding to the initial slip surface is relatively high, but by considering the influence of the pore water pressure distribution, it is found that there is a more dangerous slip surface, after adjusting the search direction of the slip surface, a smaller safety factor is obtained by re-calculation, which reflects the stability of the slope under the most unfavorable conditions.

[0091] The historical landslide event library stores a large amount of landslide event data, including the time, location, scale, triggering factors (such as rainfall, earthquake intensity, etc.) and precursor characteristics of the landslide, etc.; for example, in the historical landslide event library of a certain area, the rainfall, rainfall duration, landslide volume, landslide location and other geological information of each landslide event are recorded in detail; by using data mining and machine learning algorithms, the data in the historical landslide event library is analyzed to construct a landslide event knowledge graph, the construction process of the knowledge graph includes entity recognition (such as landslide events, rainfall, slope types, etc.), relationship extraction (such as triggering relationships, spatio-temporal relationships, etc.) and knowledge fusion (integrating data from different sources into a unified knowledge framework), the precursor characteristics of joint development high and steep rock slope usually include surface crack propagation and groundwater level rise, when the minimum safety factor of the slope is lower than the set safety factor threshold, at this time, the risk probability of the current slope is determined according to the similar historical landslide scene matched in the landslide event knowledge graph, by calculating the risk probability under the similar historical landslide scene, and combining the monitoring data and prediction results of the current slope; a graded early warning signal is triggered according to the risk probability interval, which can make full use of existing experience data to improve the accuracy and reliability of slope risk assessment.

[0092] In summary, by predicting the pore water pressure distribution according to the rock mass strength parameters and the occurrence data of the joint surface, iteratively correcting the search direction of the slip surface to determine the minimum safety factor, and combining the historical landslide event library to construct a landslide event knowledge graph, the precise assessment and graded early warning of the risk probability of high and steep slopes are realized, which provides strong support for the safety management and disaster prevention of slopes.

[0093] Further, the signal activation module is further configured to:

[0094] The characteristic parameter distribution interval of the normal working condition and the early warning working condition in the historical monitoring data is acquired; a classification hyperplane is built based on the characteristic parameter distribution interval of the normal working condition and the early warning working condition, and the classification hyperplane is used for dynamically dividing the boundary condition of the safe domain and the dangerous domain; the real-time rainfall intensity and duration are introduced as threshold correction factors, and a differentiated early warning triggering condition is set, when the characteristic parameters of the continuous N monitoring periods cross the classification hyperplane, the risk escalation mechanism in the risk probability triggering logic is triggered, and N is greater than or equal to 3.

[0095] Specifically, the historical monitoring data refers to the data accumulated by the slope in the past monitoring process, including various monitoring indexes under normal working condition and early warning working condition; the normal working condition characteristic parameter distribution interval refers to the fluctuation range of each monitoring index when the slope is in a stable state; the early warning working condition characteristic parameter distribution interval refers to the fluctuation range of each monitoring index when the slope shows signs of instability; the support vector machine is used for classification and regression analysis, and different categories of data are separated by building a classification hyperplane; the classification hyperplane refers to a plane in high-dimensional space used to divide different categories of data, which is used to distinguish the safe domain and the dangerous domain of the slope; the threshold correction factor refers to the factor for adjusting the early warning threshold, and the real-time rainfall intensity and duration are key factors affecting the stability of the slope; the differentiated early warning triggering condition refers to the early warning condition adjusted according to different factors, so that the early warning is more accurate; the risk escalation mechanism refers to increasing the early warning level when the monitoring data meets more stringent conditions, and prompting more serious risks.

[0096] The monitoring data of the slope in the normal state and the impending instability state is collected and sorted, which covers the strain, displacement, pore water pressure and other aspects of the slope; through statistical analysis of a large amount of historical data, the fluctuation range of each monitoring index under normal working condition and early warning working condition is determined, generally, the internal strain of the slope under normal working condition fluctuates between 20-50 micro-strain, and the strain under early warning working condition exceeds 80 micro-strain; based on the above distribution interval, a classification hyperplane is built by support vector machine, which is used for dynamically dividing the boundary condition of the safe domain and the dangerous domain; the characteristic parameters of the normal working condition and the early warning working condition are input into the SVM model as a training set, and the model optimizes the classification hyperplane so that the data points of the normal working condition and the early warning working condition are effectively separated by the hyperplane.

[0097] When the monitored characteristic parameters are located on one side of the classification hyperplane, it is determined that the slope is in the safe domain; if located on the other side, it is determined that it is in the dangerous domain; a set of real-time monitored characteristic parameters is substituted into the classification based on the SVM, which can effectively process high-dimensional data, and the classification hyperplane performs well in processing nonlinear problems, further, the real-time rainfall intensity and duration are introduced as threshold correction factors, and a differentiated early warning trigger condition is set, rainfall is a key factor affecting the stability of the slope, real-time rainfall intensity and duration can significantly change the stress and hydrological conditions of the slope, and the two factors are used as threshold correction factors to dynamically adjust the early warning trigger condition, for example, under light rainfall conditions, the number of continuous periods N in which the characteristic parameters cross the classification hyperplane is set to 5; under high-intensity rainfall conditions, due to the increased risk of slope instability, the value of N is reduced to 3, making the early warning more sensitive. This differentiated early warning trigger condition can adjust the early warning strategy in a timely manner according to changes in actual environmental factors, improving the timeliness and accuracy of early warning.

[0098] When the characteristic parameters of the continuous N monitoring periods cross the classification hyperplane, the risk escalation mechanism in the risk probability trigger logic is triggered, aiming to timely upgrade the warning level when the monitoring data continuously shows that the slope state is deteriorating; in this way of dynamically adjusting the early warning condition and the risk escalation mechanism, the stability of the slope under different environments can be effectively dealt with, and the intelligent level of slope safety monitoring is improved.

[0099] Further, the signal activation module is further used for:

[0100] The surface crack propagation graph of the target high and steep slope is obtained, and the crack length, width and propagation rate characteristics are extracted to generate a crack development trend prediction index; when the crack propagation rate under the crack development trend prediction index exceeds a preset rate threshold, a local monitoring intensification mode is triggered at the corresponding edge computing node, and a reinforcement reminder signal is sent. The surface crack propagation graph is obtained by the image acquisition device.

[0101] Specifically, the image acquisition device refers to a device for acquiring slope surface images, such as a camera; the surface crack propagation graph refers to an image or chart showing the expansion of the slope surface cracks over time; the crack length, width and propagation rate characteristics refer to the geometric size changes of the cracks and their rate of change over time; the crack development trend prediction index refers to a quantitative index for predicting the future development trend of the cracks according to the crack characteristic parameters; the edge computing node refers to a computing unit close to the data source for data processing and analysis, which can quickly respond to local events; the local monitoring intensification mode refers to a mode of intensifying the monitoring frequency and accuracy of a specific area; the reinforcement reminder signal refers to a signal prompting the need for reinforcement of the slope.

[0102] The image acquisition device is laid out, and appropriate equipment and installation positions are selected according to the size, shape and geological characteristics of the slope. An unmanned aerial vehicle take-off and landing platform is deployed at key positions at the top and bottom of the slope. The image acquisition device should have high resolution and good optical zoom capability to clearly capture the fine cracks on the surface of the slope. The protection level of the image acquisition device should reach IP66 or above to adapt to outdoor harsh environments. Image data of the surface of the slope is obtained through the image acquisition device, and the image is preprocessed using image processing software, including grayscale, noise removal, edge enhancement and other operations. Gaussian filtering is used to remove random noise in the image, and a Canny operator is used to detect crack edges to generate a binary crack image. The length, width and expansion rate characteristics of the crack are extracted through image analysis algorithms.

[0103] A linear regression model or a more complex machine learning model (such as a long short-term memory network) is used to predict the future development trend of the crack. For example, a linear regression model is trained using one month of crack length and width data, and the crack development trend prediction index can be defined as a comprehensive index of crack length and width expansion rates. The length expansion rate and the width expansion rate are weighted and fused to obtain the crack expansion rate. When the crack expansion rate exceeds a preset threshold, it is determined that the crack is in a rapid expansion state, and monitoring needs to be strengthened and reinforcement measures need to be considered.

[0104] The edge computing node determines whether to trigger a local monitoring intensification mode according to a preset rate threshold after receiving the crack development trend prediction index. The edge computing node immediately triggers the local monitoring intensification mode, which includes increasing the image acquisition frequency. At the same time, the edge computing node generates a reinforcement reminder signal and sends the reminder signal to the mobile terminal and the monitoring center through the wireless communication network. The reminder signal should include the location, current state, development trend of the crack and the recommended reinforcement measures (such as anchor reinforcement, crack grouting, etc.), so that appropriate measures can be taken quickly to prevent disasters caused by slope instability. In summary, by laying out image acquisition devices to monitor crack expansion, generating crack development trend prediction indexes, triggering local monitoring intensification modes and sending reinforcement reminder signals, dangerous changes in surface cracks of the slope can be discovered in a timely manner, reinforcement measures can be taken in advance, and the initiative and effectiveness of slope safety monitoring can be effectively improved.

[0105] Further, the signal activation module is further configured to:

[0106] The micro-fracture signal of the rock-soil mass is monitored, the potential landslide risk area is identified based on the frequency domain characteristics of the micro-fracture signal of the rock-soil mass, the time-frequency joint analysis is performed on the micro-fracture signal, the main frequency band energy mutation point and the signal attenuation coefficient are extracted, the three-dimensional fracture propagation vector is generated according to the main frequency band energy mutation point and the signal attenuation coefficient in combination with the focal location result of the microseismic monitoring array, the rock-soil mass damage cumulative index is determined according to the three-dimensional fracture propagation vector and the displacement gradient field of the dynamic reference point cloud, the rock-soil mass damage cumulative index is input into the finite element model of the slope stability evaluation index, the spatial geometric parameters of the potential sliding surface are dynamically corrected, the encrypted risk topology graph is synchronously uploaded to the edge computing node, and the risk topology graph is used to adjust the priority order of the reinforcement prompt signal. In this embodiment, the acoustic emission sensor is used to monitor the micro-fracture signal of the rock-soil mass.

[0107] Specifically, the acoustic emission sensor is a device capable of detecting the elastic wave signal generated by the micro-fracture event inside the rock-soil mass, and its working principle is based on the piezoelectric effect. When the rock-soil mass micro-fractures, the stress wave generated by the micro-fracture excites the piezoelectric crystal in the sensor, generates an electric signal, and is then collected and recorded. The frequency domain characteristics refer to the performance of the signal in the frequency domain, such as the energy distribution of different frequency components, which can reflect the frequency characteristics and energy concentration degree of the micro-fracture event. Time-frequency joint analysis refers to a method combining time and frequency analysis, which can provide information on the signal in time and frequency at the same time, such as wavelet transform and Hilbert-Huang transform, which can be used for time-frequency joint analysis, so as to more comprehensively understand the characteristics of the signal. The main frequency band energy mutation point refers to the position where the energy of the main frequency band of the signal changes significantly at a certain time point, which is usually related to the sudden change of the internal structure of the rock-soil mass, such as the expansion of micro-fracture or the formation of new cracks. The signal attenuation coefficient refers to the degree of energy attenuation of the signal in the propagation process, which is related to the physical properties of the medium (such as density, elastic modulus) and obstacles on the propagation path, and can be used to evaluate the integrity and fracture development degree of the rock-soil mass. The three-dimensional fracture propagation vector is a vector containing the direction, speed and intensity information of the fracture, which is used to describe the propagation process of the micro-fracture in the rock-soil mass. The rock-soil mass damage cumulative index is an index for quantifying the damage degree of the rock-soil mass, which considers factors such as the frequency, energy and fracture propagation range of the micro-fracture event, and reflects the damage accumulation of the rock-soil mass under long-term stress. The displacement gradient field of the dynamic reference point cloud refers to the displacement change rate distribution of each point in the dynamic reference point cloud, which reflects the rate and trend of the slope surface deformation, and can be used to evaluate the stability of the slope.

[0108] The acoustic emission sensor is arranged at a potential sliding surface and a high-risk area of a slope, such as inside the rock-soil mass in the middle and lower part of the slope, and an array of sensors is arranged at appropriate intervals. During monitoring, the sensors capture a plurality of micro-cracking signals. Through preliminary analysis of the signals, effective signals are screened for subsequent processing. Preferably, based on the frequency domain characteristics of the micro-cracking signals, the time domain signals are converted into frequency domain signals by using fast Fourier transform, the energy values of different frequency components are obtained, and the main frequency band with higher energy is identified. If the energy of the main frequency band suddenly increases in a certain area, it indicates that the micro-cracking activity inside the rock-soil mass in the area is intensified, and new cracks or existing cracks may expand, thereby identifying the potential landslide risk area.

[0109] The micro-cracking signals are subjected to time-frequency joint analysis, and a continuous wavelet transform method is used to decompose the signals into components at different times and frequencies, and to extract the main frequency band energy mutation point and signal attenuation coefficient. The position information of the micro-cracking signals captured by the acoustic emission sensor is combined with the source location results of the micro-seismic monitoring array. The source location of the micro-seismic event is determined by the time difference of the signals received by the multiple sensors of the micro-seismic monitoring array using a cross-location algorithm. The source locations are combined with the characteristic parameters of the acoustic emission signals (such as the main frequency band energy mutation point and the signal attenuation coefficient) to generate a three-dimensional cracking propagation vector, which more directly displays the propagation process of the micro-cracking.

[0110] According to the displacement gradient field of the three-dimensional cracking propagation vector and the dynamic reference point cloud, the rock-soil mass damage accumulation index is determined. The three-dimensional cracking propagation vector provides dynamic information of the internal cracking of the rock-soil mass, and the displacement gradient field of the dynamic reference point cloud reflects the deformation rate of the slope surface. By establishing a correlation model between the two, the damage degree of the rock-soil mass is comprehensively evaluated. As known, when the three-dimensional cracking propagation vector shows that the micro-cracking activity is frequent and the propagation range is expanding, and the displacement gradient field shows that the deformation rate of the slope surface is accelerating, the rock-soil mass damage accumulation index correspondingly increases, indicating that the rock-soil mass damage is intensified and the stability is reduced.

[0111] The rock-soil mass damage accumulation index is input into the finite element model of the slope stability evaluation index, the model parameters are updated in real time, the spatial geometric parameters of the potential sliding surface are dynamically corrected, such as the shape, position and inclination of the sliding surface, and the corrected model data and risk information are converted into an encrypted format and uploaded to the edge computing node through a secure communication protocol. The edge computing node decrypts and analyzes the received risk topology graph, adjusts the priority order of the reinforcement reminder signal according to the size and trend of the rock-soil mass damage accumulation index, and effectively reduces the risk of slope landslide to ensure the safety and stability of the slope.

[0112] In practical applications, the safety reminder module is used to establish a response link through the edge computing node based on the hierarchical early warning signal to remind the safety of the slope.

[0113] Specifically, the edge computing node refers to a computing node that performs data processing and analysis close to the data source or device end. Compared with the traditional cloud computing center, the edge computing node has the characteristics of low delay, high bandwidth and strong real-time performance; the response link refers to the information transmission path and communication connection between the monitoring system and the user or relevant departments, used for timely conveying the slope safety warning information; the slope safety reminder refers to sending the slope safety warning information to relevant personnel through various ways (sound and light alarm).

[0114] When the hierarchical warning signal is activated, the edge computing node immediately receives the signal; the edge computing node is deployed near the slope monitoring device, which can quickly process and analyze monitoring data and reduce data transmission delay. Specifically, in the slope monitoring system, the edge computing node can be installed in the field control box or data acquisition station near the monitoring device. Then, the edge computing node establishes the corresponding response link according to the level of the warning signal and the preset response strategy. The response link can include various communication methods, such as 4G / 5G wireless communication, satellite communication, Ethernet, etc., to ensure that the warning information can be timely and reliably conveyed to relevant personnel.

[0115] According to the level and emergency degree of the warning signal, different reminding methods are adopted. Specifically, for low-risk warning signals, reminder information can be sent through email or APP push; for medium-risk warning signals, sound and light alarm can be used to take corresponding preventive measures in time, such as evacuating dangerous areas, strengthening support, etc.; for high-risk warning signals, emergency alarm information is sent through multiple channels (such as sound and light alarm, broadcast, etc.). In summary, based on the hierarchical warning signal, the response link is established through the edge computing node to remind the slope safety, realizing the rapid transmission and effective response of the high and steep slope safety warning information, and ensuring the safety of the high and steep slope.

[0116] Due to the adoption of the sensor deployment module, the multi-modal sensor network is deployed under the constraint of the layout density of the slope geological parameters; the point cloud configuration module is used to scan the slope surface deformation characteristics by using the laser radar, and the dynamic reference point cloud is configured by synchronously correcting the spatial drift error by the IMU unit; the space-time alignment module is used to perform space-time alignment on the dynamic reference point cloud and the sensing monitoring data obtained by the multi-modal sensor network, extract the coupling characteristics of the internal stress field and displacement field of the slope, and configure the slope stability evaluation index; the signal activation module is used to identify the potential sliding surface according to the slope stability evaluation index, activate the hierarchical early warning signal, and associate the hierarchical early warning signal with the risk probability trigger logic, and the risk probability trigger logic is coupled with the meteorological data; the safety reminding module is used to establish a response link through the edge computing node based on the hierarchical early warning signal to remind the safety of the slope, and the application provides a high and steep slope safety monitoring system and method, effectively integrates the data collected by different types of sensors, extracts the key information reflecting the stability of the high and steep slope, optimizes the equipment layout, reduces the interference to the monitoring equipment, improves the accuracy and reliability of the monitoring data, and realizes the all-around precision monitoring of the stability of the high and steep slope

[0117] In one embodiment, as shown in Figure 2 The application provides a high and steep slope safety monitoring method, which is applied to the high and steep slope safety monitoring system, and comprises the following steps:

[0118] S1: Obtain the sensing monitoring data of the target high and steep slope, the initial three-dimensional point cloud model, and the three-axis acceleration and angular velocity of the laser radar in the scanning process; the sensing monitoring data includes strain information, elastic wave signals, and groundwater level change signals of the target high and steep slope.

[0119] S2: Determine the dynamic reference point cloud model of the target high and steep slope according to the initial three-dimensional point cloud model and the three-axis acceleration and angular velocity of the laser radar in the scanning process.

[0120] S3: Perform space-time alignment on the sensing monitoring data and the dynamic reference point cloud model, extract the coupling characteristics of the internal stress field and displacement field of the target high and steep slope, and determine the slope stability evaluation index; the slope stability evaluation index includes a safety factor, a deformation rate, a stress ratio, and a pore water pressure ratio.

[0121] S4: Identify the potential sliding surface according to the slope stability evaluation index, and activate the hierarchical early warning signal; the hierarchical early warning signal is determined according to the risk probability trigger logic; the risk probability trigger logic is the risk probability interval of the early warning signal determined according to the slope stability evaluation index and meteorological factors.

[0122] S5: Based on the hierarchical early warning signal, a response link is established by an edge computing node to carry out a slope safety reminder.

[0123] The technical features of the above embodiments can be combined in any manner. To make the description concise, not all possible combinations of the technical features in the above embodiments are described, but it should be considered that any combination of the technical features is within the scope of the present disclosure as long as the combination does not result in contradictions.

[0124] The principles and implementation modes of the present application are described by using specific examples herein, and the above embodiments are only used to help understand the method and its core idea of the present application. Meanwhile, for those skilled in the art, the specific implementation modes and application ranges will be changed according to the idea of the present application. In summary, the content of the present description should not be understood as a limitation of the present application.

Claims

1. A safety monitoring system for steep slopes, characterized in that, include: Data acquisition module, point cloud configuration module, spatiotemporal alignment module, signal activation module, and safety alert module; The data acquisition module includes a distributed strain sensor, a microseismic monitoring array, a groundwater level sensor, a lidar, and an inertial measurement unit. The distributed strain sensor is used to acquire strain information of the target steep slope. The microseismic monitoring array is used to acquire elastic wave signals generated by internal rupture events of the target steep slope. The groundwater level sensor is used to acquire groundwater level change signals of the target steep slope. The lidar is used to scan the initial three-dimensional point cloud model of the target steep slope. The inertial measurement unit is used to measure the triaxial acceleration and angular velocity of the lidar during the scanning process. The point cloud configuration module is used to determine the dynamic reference point cloud model of the target steep slope based on the initial three-dimensional point cloud model and the three-axis acceleration and angular velocity of the lidar during the scanning process. The spatiotemporal alignment module is used to spatiotemporally align the sensor monitoring data with the dynamic benchmark point cloud model, and extract the coupling characteristics of the internal stress field and displacement field of the target steep slope to determine the slope stability evaluation index; the sensor monitoring data includes strain information, elastic wave signal and groundwater level change signal of the target steep slope; the slope stability evaluation index includes safety factor, deformation rate, stress ratio and pore water pressure ratio; The signal activation module is used to identify potential slip surfaces based on the slope stability evaluation index and activate graded early warning signals; the graded early warning signals are determined based on risk probability triggering logic; the risk probability triggering logic is based on the risk probability range of the early warning signals determined by the slope stability evaluation index and meteorological factors. The safety alert module is used to provide slope safety alerts by establishing a response link through edge computing nodes based on the graded early warning signal.

2. The high and steep slope safety monitoring system according to claim 1, characterized in that, The data acquisition module further includes: a vibration fiber optic sensor array and a tilt sensor cluster; the layout direction of the vibration fiber optic sensor array is orthogonal to the potential slip direction; the tilt sensor cluster is embedded in the surface anchoring point of the target steep slope in a honeycomb topology. The vibration fiber optic sensor array is used to detect vibration and deformation inside the target steep slope; The cluster of tilt sensors is used to measure the tilt change signal of the surface anchoring point.

3. The high and steep slope safety monitoring system according to claim 1, characterized in that, The point cloud configuration module includes: The filtering unit is used to filter the triaxial acceleration and angular velocity of the lidar during the scanning process to obtain the attitude change information and position change information of the lidar. A point cloud distortion compensation unit is used to perform geometric transformations on the initial three-dimensional point cloud model based on the attitude change information and the position change information to obtain a dynamic reference point cloud model of the target steep slope; the geometric transformations include rotation, translation and scaling; The cumulative displacement error calculation unit is used to calculate the cumulative displacement error between the corrected point cloud data and the previous reference point cloud data. The model update unit is used to rescan the target steep slope by the lidar and update the dynamic reference point cloud model of the target steep slope when the cumulative displacement error exceeds a preset error threshold.

4. The high and steep slope safety monitoring system according to claim 1, characterized in that, The spatiotemporal alignment module includes: An energy extraction unit is used to perform wavelet packet decomposition on the elastic wave signal to extract energy distribution characteristics of different frequency bands. An interpolation unit is used to convert the strain information into an equivalent plastic strain field, and the equivalent plastic strain field is mapped to the three-dimensional finite element mesh model of the target steep slope using the finite element interpolation method. The alignment unit is used to map the equivalent plastic strain field to the three-dimensional finite element mesh model of the target steep slope, align the groundwater level change signal and the surface displacement monitoring sequence, combine the energy distribution characteristics of different frequency bands, quantify the coupling strength between the sensing and monitoring data, and generate a fusion weight matrix.

5. The high and steep slope safety monitoring system according to claim 1, characterized in that, The data acquisition module further includes: an image acquisition device; The image acquisition device is used to acquire a surface crack propagation map of the target steep slope, extract crack length, width and propagation rate features, generate crack development trend prediction index, and when the crack propagation rate under the crack development trend prediction index exceeds a preset rate threshold, a local monitoring and reinforcement mode is triggered at the corresponding edge computing node and a reinforcement reminder signal is sent.

6. The high and steep slope safety monitoring system according to claim 5, characterized in that, The data acquisition module also includes: an acoustic emission sensor; The acoustic emission sensor is used to monitor the micro-fracture signals of the soil and rock mass of the target steep slope in order to adjust the priority of the reinforcement reminder signal.

7. A method for safety monitoring of steep slopes, characterized in that, The aforementioned high and steep slope safety monitoring method is applied to the high and steep slope safety monitoring system according to any one of claims 1-6, and the high and steep slope safety monitoring method includes: The system acquires sensor monitoring data of the target steep slope, an initial three-dimensional point cloud model, and triaxial acceleration and angular velocity of the lidar during the scanning process; the sensor monitoring data includes strain information, elastic wave signals, and groundwater level change signals of the target steep slope. Based on the initial three-dimensional point cloud model and the triaxial acceleration and angular velocity of the lidar during the scanning process, the dynamic reference point cloud model of the target steep slope is determined. The sensor monitoring data is spatiotemporally aligned with the dynamic benchmark point cloud model, and the coupling characteristics of the internal stress field and displacement field of the target steep slope are extracted to determine the slope stability evaluation index; the slope stability evaluation index includes safety factor, deformation rate, stress ratio and pore water pressure ratio; Potential slip surfaces are identified based on the slope stability evaluation indicators, and graded early warning signals are activated. The graded early warning signals are determined based on risk probability triggering logic. The risk probability triggering logic is based on the risk probability range of the early warning signals determined by the slope stability evaluation indicators and meteorological factors. Based on the aforementioned graded early warning signals, a response link is established through edge computing nodes to provide slope safety reminders.

8. The method for safety monitoring of steep slopes according to claim 7, characterized in that, Based on the initial 3D point cloud model and the triaxial acceleration and angular velocity of the lidar during the scanning process, a dynamic reference point cloud model of the target steep slope is determined, specifically including: The three-axis acceleration and angular velocity of the lidar during the scanning process are filtered to obtain the attitude change information and position change information of the lidar. Based on the attitude change information and the position change information, the initial three-dimensional point cloud model is geometrically transformed to obtain the dynamic reference point cloud model of the target steep slope; the geometric transformation includes rotation, translation and scaling; Calculate the cumulative displacement error between the corrected point cloud data and the previous reference point cloud data; the corrected point cloud data is obtained by geometric transformation of the initial three-dimensional point cloud model; When the cumulative displacement error exceeds a preset error threshold, the lidar rescans the target steep slope and updates the dynamic reference point cloud model of the target steep slope.

9. The method for safety monitoring of steep slopes according to claim 7, characterized in that, The sensor monitoring data is spatiotemporally aligned with the dynamic benchmark point cloud model, and the coupling characteristics of the internal stress field and displacement field of the target steep slope are extracted to determine the slope stability evaluation index, specifically including: The elastic wave signal is subjected to wavelet packet decomposition to extract the energy distribution characteristics of different frequency bands; The strain information is converted into an equivalent plastic strain field, and the equivalent plastic strain field is mapped to the three-dimensional finite element mesh model of the target steep slope using the finite element interpolation method. After mapping the equivalent plastic strain field to the three-dimensional finite element mesh model of the target steep slope, the groundwater level change signal and the surface displacement monitoring sequence are aligned. Combining the energy distribution characteristics of different frequency bands, the coupling strength between the sensing and monitoring data is quantified and a fusion weight matrix is ​​generated.

10. The method for safety monitoring of steep slopes according to claim 7, characterized in that, Also includes: Obtain a surface crack propagation map of the target steep slope; Based on the surface crack propagation map, crack length, crack width, and crack propagation rate are extracted to generate crack development trend prediction indicators. When the crack propagation rate exceeds a preset rate threshold, a local monitoring and reinforcement mode is triggered at the corresponding edge computing node, and a reinforcement reminder signal is sent.

Images