Integrated slope monitoring system and method

Abstract

The present invention discloses an integrated slope monitoring system and method, which relates to the technical field of highway slope monitoring, and includes the following steps: dividing the highway slope geological area to be monitored into multiple sub-areas; in each sub-area, scanning with an initial preset point cloud density by a laser radar installed on an unmanned aerial vehicle to obtain the initial surface morphological information of the entire highway slope; for each sub-area, using the point cloud data generated by the laser radar to obtain the surface morphological information under the sub-area. By dividing the sub-areas and combining the pre-trained machine learning model to predict abnormal surface changes, the present invention can timely identify early signs and issue early warnings at the early stage of risks, thereby improving the prevention and emergency response capabilities of geological disasters. At the same time, the point cloud density is dynamically adjusted to improve the monitoring accuracy for high-risk areas, avoid waste of resources, greatly improve monitoring efficiency and accuracy, and optimize risk management.

Description

Technical Field

[0001] The invention relates to the technical field of highway slope monitoring, and in particular to an integrated slope monitoring system and method. Background Art

[0002] Integrated slope monitoring is a system that combines multiple monitoring technologies and early warning methods to conduct comprehensive and continuous monitoring and analysis of slope stability. The system is based on a point-line-network monitoring and early warning technology system. It uses modern sensors and Internet of Things technology to collect data in real time, and uses drones to patrol slopes on fixed cruise routes to verify the accuracy of early warning information. In addition, the system uses commercial maps such as Baidu and AutoNavi to collect vehicle driving characteristics, further verify early warning information and issue alerts. The integrated monitoring system can run simultaneously when the network is unobstructed and interrupted, ensuring reliability under extreme conditions. Through the online and offline parallel early warning mode, combined with multi-dimensional data verification and sound and light warning equipment, the system can quickly issue early warning information and guide the public to avoid risks. This integrated approach not only improves the accuracy and timeliness of early warnings, but also provides multi-dimensional data support for emergency rescue decisions.

[0003] Point: Based on modern sensing technology and the Internet of Things, the smallest engineering unit that realizes monitoring and early warning through the network.

[0004] Line: Use drones to patrol slopes along fixed cruise routes and verify early warning information.

[0005] Network: Use commercial maps such as Baidu and Amap to collect vehicle driving characteristics, verify monitoring and early warning information, and issue alarms.

[0006] In recent years, the rapid development of UAV technology has provided a new direction for monitoring geological disasters on highway slopes. In existing technologies, UAVs are usually equipped with LiDAR for monitoring. The principle of LiDAR is to send laser pulses to the ground or target objects through a laser transmitter, and the receiver captures the reflected light signal and calculates the distance based on the time difference of the light signal returning. This technology can generate high-precision three-dimensional point cloud data and record the surface morphology changes of the slope in real time. UAVs combined with LiDAR technology can achieve all-weather, all-round and efficient slope monitoring, especially in complex terrain or extreme weather conditions.

[0007] When using LiDAR to monitor highway slope geology, existing technologies usually use a preset point cloud density for monitoring. This initial setting can usually meet the detection needs under normal conditions. However, when there are small abnormal morphological changes in the local area of ​​the slope geology, such as surface subsidence, crack expansion, soil loosening or slippage, continuing to use a fixed point cloud density for monitoring may have serious consequences. The initial point cloud density is mainly designed for normal monitoring. The lower resolution may not be able to capture these small but crucial details when facing complex geological changes. In this case, the abnormal area may not be discovered in time, or it may be misjudged as risk-free in the monitoring data, causing the early warning system to miss the best response time. Delayed warnings may directly lead to an escalation of disaster risks. For example, cracks or landslide precursors are ignored. Once a landslide actually occurs, the monitoring system will not be able to provide sufficient response time for rescue and emergency operations, greatly increasing the severity of the disaster.

[0008] Point cloud density represents the number of laser pulses emitted by the lidar per second and is usually measured in millions of points per second.

[0009] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0010] The purpose of the present invention is to provide an integrated slope monitoring system and method, which can timely identify early abnormal signs by dividing the highway slope geological area into multiple sub-areas, and combine pre-trained machine learning models to analyze and predict abnormal surface morphological changes in each sub-area, ensure that early warnings are issued in the early stages of risks, and significantly improve the prevention and emergency response capabilities of geological disasters. In addition, the present invention improves monitoring accuracy and avoids waste of resources in high-risk areas by identifying and marking abnormal areas and dynamically adjusting the point cloud density. This precise monitoring mechanism greatly improves monitoring efficiency and accuracy, optimizes risk management under limited resources, and solves the problems in the above-mentioned background technology.

[0011] In order to achieve the above object, the present invention provides the following technical solution: an integrated slope monitoring method, comprising the following steps:

[0012] Divide the highway slope geological area to be monitored into multiple sub-areas;

[0013] In each sub-area, the laser radar installed on the drone is used to scan with an initial preset point cloud density to obtain the initial surface morphology information of the entire highway slope;

[0014] For each sub-area, the point cloud data generated by the laser radar is used to obtain the surface morphological information of the sub-area;

[0015] Extract features of the surface morphology information in the sub-area, and based on the extracted features, use the pre-trained machine learning model to analyze and predict abnormal surface morphology changes in the sub-area under the monitoring window, and identify early signs of abnormal surface morphology changes;

[0016] Based on the prediction results of the machine learning model, all sub-areas under the monitoring window are divided into abnormal morphological change areas and normal morphological change areas. The normal morphological change areas are areas with stable geological changes, and the abnormal morphological change areas are areas with the risk of abnormal morphological changes.

[0017] All sub-areas that are judged to be areas of abnormal morphological changes are marked so that they can be focused on in subsequent monitoring;

[0018] For the sub-areas that are judged to have normal morphological changes, the initially set point cloud density is continued to be used for monitoring;

[0019] For sub-areas marked as having the risk of abnormal morphological changes, high point cloud density is reallocated to them based on the geological prediction results of the sub-areas;

[0020] Taking advantage of the reallocated high point cloud density, all marked anomaly areas are further scanned.

[0021] Preferably, when the geological area of ​​the highway slope to be monitored is divided into multiple sub-areas, a zoning method based on geological characteristics and terrain data is adopted, and high-precision three-dimensional terrain data of the entire area is obtained by lidar. Then, the area is classified in combination with geological survey data, and GIS or spatial analysis software is used to divide the highway slope into multiple sub-areas with relatively consistent geological conditions.

[0022] Preferably, feature extraction is performed on the surface morphological information under the sub-area, wherein the extracted features include surface crack density, width and degree of surface area subsidence; in the monitoring window, after analyzing the surface crack density, width and degree of surface area subsidence, surface crack quantification values ​​and settlement deformation quantification values ​​are generated respectively; the risk of crack extension is quantified by the surface crack quantification values, and the subsidence risk of the surface area is quantified by the settlement deformation quantification values; the surface crack quantification values ​​and settlement deformation quantification values ​​are input into a pre-trained machine learning model, and a surface morphological change coefficient is generated based on the machine learning model; the surface abnormal morphological changes are analyzed and predicted by the surface morphological change coefficient, and early signs of abnormal surface morphological changes are identified.

[0023] Preferably, under the monitoring window, the surface morphology change coefficient generated after analyzing and predicting the abnormal surface morphology change in the sub-area based on the machine learning model is compared with a preset reference threshold of the surface morphology change coefficient to divide the sub-area. The specific division process is as follows:

[0024] If the surface morphology change coefficient is greater than or equal to the reference threshold of the surface morphology change coefficient, the sub-region is divided into an abnormal morphology change region, and the normal morphology change region is a region with stable geological changes;

[0025] If the surface morphology change coefficient is less than the surface morphology change coefficient reference threshold, the sub-region is divided into a normal morphology change region, and the abnormal morphology change region is a region with the risk of abnormal morphology change.

[0026] Preferably, in the monitoring window, after analyzing the density and width of the surface cracks, the specific steps of generating the quantitative value of the surface cracks are as follows:

[0027] The density and width information of surface cracks are extracted from the point cloud data obtained by the lidar. The crack density indicates the number of cracks per unit area, and the crack width indicates the lateral opening width of the crack. The calculation expression of the crack density is: , where is the crack density, is the number of cracks, A is the area of ​​the sub-region, and the crack width is directly calculated from the point cloud data. express;

[0028] The crack expansion rate is calculated to reflect the expansion trend of the crack within the monitoring window period. The crack expansion rate is the product of the crack density and the width change rate, which indicates the speed of crack expansion. The calculation expression is: , where is the crack growth rate, and are the start and end time of the monitoring window, and They represent the crack density and crack width at time t respectively;

[0029] Risk weight coefficients are introduced for different fracture extension conditions. The impact of fractures is weighted according to the changes in fracture density and width. The risk weight coefficient is based on fracture density, width and geological sensitivity in the sub-region. The calculation expression is: , where is the risk weight factor;

[0030] The final surface crack quantification value is calculated by combining the crack extension rate and risk weight coefficient to quantify the risk of crack extension. The calculation expression is: , where Quantify the value of surface cracks.

[0031] Preferably, in the monitoring window, after analyzing the subsidence degree of the surface area, the specific steps of generating the quantitative value of the subsidence deformation are as follows:

[0032] The sub-areas are monitored regularly to obtain the surface subsidence data at different time points, which are recorded as ,in Indicates the coordinates in the sub-region at time t The degree of subsidence;

[0033] Calculate the change in the speed of surface subsidence. The calculation expression is: , where represents the sinking speed at time t;

[0034] In order to quantify the sinking risk more accurately, the change rate of the sinking speed is further calculated. The calculation expression is: , where represents the sinking acceleration at time t;

[0035] Based on the sinking velocity and sinking acceleration, the settlement deformation quantification value is defined and calculated. The calculation expression is:

[0036] ,

[0037] In the formula, Represents the quantitative value of settlement deformation, It means to quadrature the nonlinear changes in the monitoring window to capture the overall sinking trend in the monitoring window.

[0038] Preferably, all sub-regions determined to be regions with abnormal morphological changes are marked, and the marking method includes assigning a special identifier to the region or setting it as a high-priority monitoring region in the drone system and the monitoring platform.

[0039] Preferably, for the sub-region marked as having the risk of abnormal morphological change, based on the geological prediction result of the sub-region, the high point cloud density is reallocated thereto, and the specific steps are as follows:

[0040] For the sub-regions that have been marked as having the risk of abnormal morphological changes, the initial point cloud density used in routine monitoring of the sub-region is extracted and recorded as ;

[0041] Based on the surface morphology monitoring data of the sub-region and the results predicted by the machine learning model, the surface morphology change coefficient is extracted and recalibrated as ;

[0042] The calculated surface morphology change coefficient is compared with the preset reference threshold of the surface morphology change coefficient to calculate the surface abnormal morphology change deviation. The calculation expression is:

[0043] ,

[0044] In the formula, Indicates the deviation of abnormal surface morphology change, Indicates the reference threshold of the surface morphology change coefficient, Indicates the change amplification factor, which is used to amplify significant abnormal surface changes and enhance the identification of risks;

[0045] In order to accurately capture the abnormal morphological changes in the sub-area, the point cloud density increment is calculated according to the abnormal morphological change deviation of the surface, and the point cloud density is dynamically adjusted through the point cloud density increment. The calculation expression of the point cloud density increment is:

[0046] ,

[0047] In the formula, represents the point cloud density increment, k Represents a constant coefficient, which controls the increment between the point cloud density and the surface morphology change coefficient. represents the nonlinear enhancement factor, which adjusts the sensitivity of the surface anomaly morphology change to the point cloud density increment. Represents the density adjustment coefficient, which is used to process the smooth transition of point cloud density adjustment;

[0048] Based on the calculated point cloud density increment, a new point cloud density is assigned to the sub-area to ensure a more refined monitoring effect. The calculation expression of the new point cloud density is: , where is the new point cloud density, which indicates the monitoring density after redistribution.

[0049] An integrated slope monitoring system includes a region division module, an initial scanning module, a surface morphology acquisition module, a feature extraction and prediction module, a sub-region division module, an abnormal region marking module, a normal region monitoring module, a point cloud density adjustment module and a depth scanning module;

[0050] The regional division module divides the highway slope geological area to be monitored into multiple sub-areas;

[0051] The initial scanning module uses the laser radar installed on the drone to scan in each sub-area with an initial preset point cloud density to obtain the initial surface morphology information of the entire highway slope;

[0052] The surface morphology acquisition module uses the point cloud data generated by the laser radar to obtain the surface morphology information of each sub-area;

[0053] The feature extraction and prediction module extracts features from the surface morphology information in the sub-area. Based on the extracted features, the pre-trained machine learning model is used to analyze and predict the abnormal surface morphology changes in the sub-area under the monitoring window to identify early signs of abnormal surface morphology changes.

[0054] The sub-region division module divides all sub-regions under the monitoring window into abnormal morphological change regions and normal morphological change regions based on the prediction results of the machine learning model. The normal morphological change regions are regions with stable geological changes, and the abnormal morphological change regions are regions with the risk of abnormal morphological changes.

[0055] The abnormal area marking module marks all sub-areas that are judged to be areas of abnormal morphological changes so that they can be focused on in subsequent monitoring;

[0056] The normal area monitoring module continues to monitor the sub-areas that are judged to have normal morphological changes using the initially set point cloud density;

[0057] Point cloud density adjustment module, for sub-areas marked with the risk of abnormal morphological changes, reallocates high point cloud density to them based on the geological prediction results of the sub-areas;

[0058] The deep scanning module uses the reallocated high point cloud density to further scan all marked abnormal areas.

[0059] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0060] The present invention divides the highway slope geological area into multiple sub-areas, and combines a pre-trained machine learning model to analyze and predict abnormal surface morphological changes in each sub-area. It can identify early signs of abnormal surface morphological changes, allowing the system to issue an early warning at the early stage of the risk, avoiding missing the best warning time due to failure to identify minor changes, thereby significantly improving the prevention and emergency response capabilities of geological disasters.

[0061] The present invention identifies and marks areas with abnormal morphological changes, uses the initial point cloud density for areas with normal morphological changes, and reallocates a higher point cloud density to sub-areas with abnormal risks. This dynamic adjustment mechanism not only effectively avoids waste of resources, but also ensures more detailed monitoring in high-risk areas and captures more subtle geological changes. This precise monitoring method greatly improves the efficiency and accuracy of slope monitoring, and helps to achieve optimal risk management under limited resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0063] Figure 1 The present invention is a method flow chart of an integrated slope monitoring method.

[0064] Figure 2 It is a module schematic diagram of an integrated slope monitoring system of the present invention. DETAILED DESCRIPTION

[0065] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present disclosure will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art.

[0066] The present invention provides Figure 1 An integrated slope monitoring method is shown, comprising the following steps:

[0067] Divide the highway slope geological area to be monitored into multiple sub-areas;

[0068] When monitoring highway slopes, the entire geological area is divided into multiple sub-areas, and the geological conditions in each sub-area are relatively consistent. This division strategy can greatly improve the accuracy of monitoring. Through such refinement, targeted analysis can be carried out on the geological characteristics of different areas, thereby improving the efficiency and effectiveness of overall monitoring. The benefit of ensuring that the geological conditions of each sub-area are relatively consistent is that monitoring errors under different geological conditions can be avoided, allowing the system to capture local abnormal changes more accurately. In addition, this division method also facilitates independent analysis of local risks, and monitoring plans can be adjusted more flexibly in areas with large changes in geological risks or more potential problems. The division of sub-areas not only helps to quickly identify early signs of surface anomalies, but also provides accurate data support for subsequent emergency responses, ensuring that customized response measures are taken for different types of geological risks in large-scale slope environments, and achieving more efficient and accurate disaster prevention and management.

[0069] When dividing the geological area of ​​the highway slope to be monitored into multiple sub-areas, a zoning method based on geological characteristics and terrain data can be used. First, high-precision three-dimensional terrain data of the entire area is obtained through LiDAR, and then the area is classified in combination with geological survey data, such as rock layer distribution, soil type, slope, historical geological disaster records, etc. Using GIS (Geographic Information System) or spatial analysis software, the highway slope is divided into multiple sub-areas with relatively consistent geological conditions based on these data. This not only allows for a more accurate analysis of the risks of each area, but also provides targeted treatment solutions for subsequent monitoring and early warning.

[0070] In each sub-area, the laser radar installed on the UAV is used to scan with an initial preset point cloud density (initial point cloud density) to obtain the initial surface morphological information of the entire highway slope;

[0071] The initial point cloud density is set based on monitoring requirements under normal conditions, which usually provides sufficient resolution to identify terrain changes over a large area. UAV technology can efficiently cover large areas of slopes, and combined with high-precision scanning of LiDAR, ensures the comprehensiveness and accuracy of the data.

[0072] For each sub-area, the point cloud data generated by the laser radar is used to obtain the surface morphological information of the sub-area;

[0073] Using laser radar (LiDAR) technology, each sub-area of ​​the highway slope is scanned to generate three-dimensional point cloud data of the area. These point cloud data calculate the precise position of each point through the time difference between the emission and reception of laser pulses, thereby constructing a detailed terrain model of the sub-area. In this way, surface morphological information of the surface can be obtained, such as ground height changes, slopes, surface undulations, etc. The high precision and long-range sensing capabilities of LiDAR enable it to capture subtle geological changes in the slope, especially in complex or dangerous terrain, providing basic data support for subsequent geological monitoring and risk analysis.

[0074] Extract features of the surface morphology information in the sub-area, and based on the extracted features, use the pre-trained machine learning model to analyze and predict abnormal surface morphology changes in the sub-area under the monitoring window, and identify early signs of abnormal surface morphology changes;

[0075] The surface morphological information of the sub-area is subjected to feature extraction, wherein the extracted features include surface crack density, width and surface area subsidence degree. Under the monitoring window, after analyzing the surface crack density, width and surface area subsidence degree, surface crack quantification value and settlement deformation quantification value are generated respectively. The risk of crack extension is quantified by the surface crack quantification value, and the subsidence risk of the surface area is quantified by the settlement deformation quantification value. The surface crack quantification value and settlement deformation quantification value are input into a pre-trained machine learning model, and a surface morphological change coefficient is generated based on the machine learning model. The surface abnormal morphological changes are analyzed and predicted by the surface morphological change coefficient, and the early signs of abnormal surface morphological changes are identified.

[0076] The degree of subsidence refers to the vertical displacement of the ground surface at that specific location relative to its initial state, and usually indicates the downward settlement of the ground.

[0077] When the density and width of surface cracks in a sub-area continue to increase, it usually indicates that there is a risk of abnormal morphological changes on the surface of the area. The increase in crack density means an increase in the number of cracks per unit area, reflecting the instability of the surface structure, which may be the result of changes in pressure inside the soil or rock layer. This increase in cracks may be an early sign of gradual loosening or rupture of the surface. The increase in crack width indicates the expansion of cracks, indicating that not only the number of cracks has increased, but also the depth or opening of the cracks has expanded. This often indicates that there has been a significant displacement of the surface or underground layer, indicating that the geological structure may be gradually becoming unstable. The simultaneous occurrence of both means that the pressure or stress on the geological structure is increasing, which may lead to geological disasters such as landslides and collapses. Therefore, when the density and width of cracks increase at the same time, it is necessary to be highly vigilant about the potential risk of abnormal geological changes in the area.

[0078] In the monitoring window, after analyzing the density and width of surface cracks, the specific steps to generate the quantitative value of surface cracks are as follows:

[0079] The density and width information of surface cracks are extracted from the point cloud data obtained by the lidar. The crack density indicates the number of cracks per unit area, and the crack width indicates the lateral opening width of the crack. The calculation expression of the crack density is: , where is the crack density, is the number of cracks, A is the area of ​​the sub-region, and the crack width is directly calculated from the point cloud data. express;

[0080] The actual width of each crack on the surface can be measured using the point cloud data generated by LiDAR. LiDAR generates three-dimensional point cloud data by emitting laser pulses and receiving reflected signals, from which the geometric shape of the surface can be accurately calculated. By analyzing the height and position differences on both sides of the crack in these point cloud data, the horizontal opening width of the crack can be obtained.

[0081] Calculate the crack expansion rate to reflect the expansion trend of the crack within the monitoring window period (such as 10 days, 20 days, one month, etc.). The crack expansion rate is the product of the crack density and the width change rate, which indicates the speed of crack expansion. The calculation expression is: , where is the crack growth rate, and are the start and end time of the monitoring window, and They represent the crack density and crack width at time t respectively;

[0082] Risk weight coefficients are introduced for different fracture extension conditions. The impact of fractures is weighted according to the changes in fracture density and width. The risk weight coefficient is based on fracture density, width and geological sensitivity in the sub-region. The calculation expression is: , where is the risk weight factor;

[0083] This formula is a Sigmoid function that is used to control the impact of changes in crack density and width on the overall risk, making the index change smoother. However, when the crack density and width increase, the index will rise rapidly.

[0084] The final surface crack quantification value is calculated by combining the crack extension rate and risk weight coefficient to quantify the risk of crack extension. The higher the surface crack quantification value, the greater the risk of crack extension in the area. The calculation expression is: , where Quantify the value of surface cracks.

[0085] It can be seen from the quantitative value of surface cracks that, under the monitoring window, the larger the performance value of the quantitative value of surface cracks generated after analyzing the density and width of surface cracks, it usually indicates that the risk of surface morphological changes in the sub-area is greater. By analyzing the density and width of cracks during the monitoring window, the quantitative value of surface cracks comprehensively reflects the expansion speed and severity of cracks. When the quantitative value of surface cracks is high, it means that the number and width of cracks are increasing rapidly, which is usually a signal of unstable geological structure and may indicate that the risk of potential geological disasters such as landslides and collapses is increasing. On the contrary, a low quantitative value of surface cracks indicates that the expansion speed of cracks is slow, the surface morphology is relatively stable, and the risk of geological disasters is small. Therefore, the larger the quantitative value of surface cracks, the greater the risk of abnormal surface morphological changes.

[0086] If the subsidence of a sub-area is large, it usually indicates that the geological structure under the sub-area may be at risk of abnormal morphological changes. Surface subsidence may be caused by a variety of factors, such as groundwater loss, soil compaction, loosening of geological structures, or the formation of underground cavities. These phenomena may lead to insufficient surface support, causing slope slip, crack expansion, and even large-scale landslides. Therefore, when the subsidence of the surface area is significantly higher than normal, it is often an early sign of slope instability, indicating that the geological morphology of the area is undergoing potentially dangerous changes. Timely detection and analysis of these subsidence characteristics can help warn of geological disasters and provide a basis for subsequent emergency response to avoid greater disasters. Therefore, when the surface subsidence is large, it does indicate that there is a risk of abnormal morphological changes in the area.

[0087] In the monitoring window, after analyzing the subsidence degree of the surface area, the specific steps to generate the quantitative value of settlement deformation are as follows:

[0088] The sub-areas are monitored regularly to obtain the surface subsidence data at different time points, which are recorded as ,in Indicates the coordinates in the sub-region at time t The degree of subsidence at the location (the degree of subsidence refers to the vertical displacement of the ground surface at that specific location relative to the initial state);

[0089] As time goes by, the degree of surface subsidence at each coordinate point will change differently. The subsidence data is obtained through lidar scanning and is based on time and space distribution.

[0090] Calculate the change in the speed of surface subsidence. The calculation expression is: , where represents the sinking speed at time t;

[0091] This step can capture the dynamic changes of surface subsidence by performing differential processing on data at consecutive time points.

[0092] In order to quantify the sinking risk more accurately, the change rate of the sinking speed is further calculated. The calculation expression is: , where represents the sinking acceleration at time t;

[0093] This calculation formula represents the acceleration of sinking, measuring the change in sinking speed through the second derivative. If the acceleration increases significantly, it indicates that the risk of sinking in the area is high.

[0094] Based on the sinking velocity and sinking acceleration, the settlement deformation quantification value is defined and calculated. The calculation expression is:

[0095] ,

[0096] In the formula, Represents the quantitative value of settlement deformation, Indicates the integration of nonlinear changes within the monitoring window to capture the overall sinking trend within the monitoring window ( and The start and end time of the monitoring window respectively);

[0097] The integral interval of this calculation formula starts from the start time of the monitoring window To end time , which represents the integral within the range of acceleration change, is used to capture the sinking dynamics of the area within the monitoring time window. This formula combines the effects of sinking speed and acceleration, weighs the nonlinear changes of speed and acceleration, and can identify risks more sensitively.

[0098] The settlement deformation quantification value is an important indicator used to quantify the risk of surface subsidence. The larger the settlement deformation quantification value generated by analyzing the subsidence degree of the surface area, the higher the potential risk of subsidence and geological instability. This is because the larger settlement deformation quantification value reflects a more significant surface subsidence speed and acceleration, indicating possible geological disasters such as landslides or collapses. On the contrary, the smaller settlement deformation quantification value indicates that the subsidence degree and speed of the surface are relatively stable, and the risk of abnormal morphological changes is small. Therefore, the settlement deformation quantification value can be effectively used to assess and warn of the risk of abnormal morphological changes on the surface.

[0099] Machine learning models, including regression analysis, support vector machines (SVM) or neural networks, are used to predict slope stability and assess the risk of geological hazards. This process improves the accuracy of the analysis of complex geological changes and supplements the deficiencies of traditional monitoring methods.

[0100] A pre-trained machine learning model means that before actual application, the model has been trained and optimized through a large amount of historical data so that it can accurately predict changes in surface morphology. In this process, the model will learn the relationship between input parameters (such as surface crack quantification values, settlement deformation quantification values) and abnormal surface changes. Through training, the model establishes an ability to judge and predict based on new input data. In the present invention, the pre-trained model can generate a surface morphology change coefficient based on the real-time input surface crack quantification values ​​and settlement deformation quantification values, and predict whether the surface will change abnormally based on this coefficient, thereby identifying early risks.

[0101] The machine learning model is not specifically limited here, and can quantify the surface cracks and settlement deformation quantification value Comprehensive analysis to generate surface morphology change coefficient In order to realize the technical solution of the present invention, the present invention provides a specific implementation method;

[0102] Coefficient of surface morphology change The resulting calculation formula is:

[0103] ,

[0104] In the formula, , Quantitative value of surface cracks and settlement deformation quantification value The preset scaling factor of , Both are greater than 0.

[0105] It can be seen from the surface morphology change coefficient that, under the monitoring window, the larger the performance value of the surface crack quantification value generated after analyzing the density and width of the surface cracks, the larger the performance value of the settlement deformation quantification value generated after analyzing the subsidence degree of the surface area. In other words, the larger the performance value of the surface morphology change coefficient generated after analyzing and predicting the abnormal surface morphology changes in the sub-area under the monitoring window, the greater the risk of abnormal surface morphology changes in the sub-area, and vice versa.

[0106] Based on the prediction results of the machine learning model, all sub-areas under the monitoring window are divided into abnormal morphological change areas and normal morphological change areas. The normal morphological change areas are areas with stable geological changes, and the abnormal morphological change areas are areas with the risk of abnormal morphological changes.

[0107] In the monitoring window, the surface morphology change coefficient generated by analyzing and predicting the abnormal surface morphology changes in the sub-area based on the machine learning model is compared with the pre-set reference threshold of the surface morphology change coefficient to divide the sub-area. The specific division process is as follows:

[0108] If the surface morphology change coefficient is greater than or equal to the reference threshold of the surface morphology change coefficient, the sub-region is divided into an abnormal morphology change region, and the normal morphology change region is a region with stable geological changes;

[0109] If the surface morphology change coefficient is less than the surface morphology change coefficient reference threshold, the sub-region is divided into a normal morphology change region, and the abnormal morphology change region is a region with the risk of abnormal morphology change.

[0110] In this way, geological risk areas can be accurately located, the waste of global monitoring resources can be avoided, and more monitoring resources can be concentrated in high-risk areas.

[0111] All sub-areas that are judged to be areas of abnormal morphological changes are marked so that they can be focused on in subsequent monitoring;

[0112] Marking can include assigning special identifiers to these areas, or setting them as high-priority monitoring areas in drone systems and monitoring platforms. The marking step ensures that abnormal areas can be quickly identified and handled during subsequent monitoring to avoid the expansion of risks.

[0113] For the sub-areas that are judged to have normal morphological changes, the initially set point cloud density is continued to be used for monitoring;

[0114] This is to save computing resources and processing time, ensuring that resources are concentrated on those abnormal areas that require higher resolution. At the same time, continuous monitoring of normal areas can ensure that there are no new changes in their geological status, maintaining the comprehensiveness and continuity of overall monitoring.

[0115] For sub-areas marked as having the risk of abnormal morphological changes, high point cloud density is reallocated to them based on the geological prediction results of the sub-areas;

[0116] For the sub-areas marked with the risk of abnormal morphological changes, high point cloud density is reallocated to them based on the geological prediction results of the sub-areas. The specific steps are as follows:

[0117] For the sub-regions that have been marked as having the risk of abnormal morphological changes, the initial point cloud density used in routine monitoring of the sub-region is extracted and recorded as ;

[0118] The initial point cloud density is usually set based on conventional geological conditions and historical monitoring experience to ensure sufficient monitoring accuracy in most areas.

[0119] Based on the surface morphology monitoring data of the sub-region and the results predicted by the machine learning model, the surface morphology change coefficient is extracted and recalibrated as ;

[0120] The calculated surface morphology change coefficient is compared with the preset reference threshold of the surface morphology change coefficient to calculate the surface abnormal morphology change deviation. The calculation expression is:

[0121] ,

[0122] In the formula, Indicates the deviation of abnormal surface morphology change, Indicates the reference threshold of the surface morphology change coefficient, Indicates the change amplification factor, which is used to amplify significant abnormal surface changes and enhance the identification of risks;

[0123] In order to accurately capture the abnormal morphological changes in the sub-area, the point cloud density increment is calculated according to the abnormal morphological change deviation of the surface, and the point cloud density is dynamically adjusted through the point cloud density increment. The calculation expression of the point cloud density increment is:

[0124] ,

[0125] In the formula, represents the point cloud density increment, k Represents a constant coefficient, which controls the increment between the point cloud density and the surface morphology change coefficient. represents the nonlinear enhancement factor, which adjusts the sensitivity of the surface anomaly morphology change to the point cloud density increment. Represents the density adjustment coefficient, which is used to process the smooth transition of point cloud density adjustment;

[0126] Based on the calculated point cloud density increment, a new point cloud density is assigned to the sub-area to ensure a more refined monitoring effect. The calculation expression of the new point cloud density is: , where is the new point cloud density, which indicates the monitoring density after redistribution.

[0127] Generally, abnormal areas require higher point cloud density to obtain more detailed terrain data in order to capture subtle changes that may affect slope stability. By increasing the point cloud density, geological changes in these areas can be observed more accurately, especially disaster precursors such as crack expansion or surface slip.

[0128] All marked anomaly areas are further scanned using the reassigned high point cloud density;

[0129] This process can capture more detailed surface morphological information and identify previously undetected minor geological changes, such as loosening of surface soil and enlargement of cracks. By improving monitoring accuracy, it can ensure a comprehensive grasp of abnormal morphological changes in slopes and provide accurate data support for subsequent early warning and response measures.

[0130] The present invention divides the highway slope geological area into multiple sub-areas and combines a pre-trained machine learning model to analyze and predict abnormal surface morphological changes. It can identify early signs of abnormal surface morphological changes, allowing the system to issue an early warning at the early stage of the risk, avoiding missing the best warning time due to failure to identify minor changes, thereby significantly improving the prevention and emergency response capabilities of geological disasters.

[0131] The present invention identifies and marks areas with abnormal morphological changes, uses the initial point cloud density for areas with normal morphological changes, and reallocates a higher point cloud density to sub-areas with abnormal risks. This dynamic adjustment mechanism not only effectively avoids waste of resources, but also ensures more detailed monitoring in high-risk areas and captures more subtle geological changes. This precise monitoring method greatly improves the efficiency and accuracy of slope monitoring, and helps to achieve optimal risk management under limited resources.

[0132] The present invention provides Figure 2 An integrated slope monitoring system shown includes a region division module, an initial scanning module, a surface morphology acquisition module, a feature extraction and prediction module, a sub-region division module, an abnormal region marking module, a normal region monitoring module, a point cloud density adjustment module and a depth scanning module;

[0133] The regional division module divides the highway slope geological area to be monitored into multiple sub-areas;

[0134] The initial scanning module uses the laser radar installed on the drone to scan in each sub-area with an initial preset point cloud density to obtain the initial surface morphology information of the entire highway slope;

[0135] The surface morphology acquisition module uses the point cloud data generated by the laser radar to obtain the surface morphology information of each sub-area;

[0136] The feature extraction and prediction module extracts features from the surface morphology information in the sub-area. Based on the extracted features, the pre-trained machine learning model is used to analyze and predict the abnormal surface morphology changes in the sub-area under the monitoring window to identify early signs of abnormal surface morphology changes.

[0137] The sub-region division module divides all sub-regions under the monitoring window into abnormal morphological change regions and normal morphological change regions based on the prediction results of the machine learning model. The normal morphological change regions are regions with stable geological changes, and the abnormal morphological change regions are regions with the risk of abnormal morphological changes.

[0138] The abnormal area marking module marks all sub-areas that are judged to be areas of abnormal morphological changes so that they can be focused on in subsequent monitoring;

[0139] The normal area monitoring module continues to monitor the sub-areas that are judged to have normal morphological changes using the initially set point cloud density;

[0140] Point cloud density adjustment module, for sub-areas marked with the risk of abnormal morphological changes, reallocates high point cloud density to them based on the geological prediction results of the sub-areas;

[0141] The deep scanning module uses the reallocated high point cloud density to further scan all marked abnormal areas;

[0142] An integrated slope monitoring method provided in an embodiment of the present invention is implemented by the above-mentioned integrated slope monitoring system. The specific method and process of an integrated slope monitoring system are detailed in the embodiment of the above-mentioned integrated slope monitoring method, which will not be repeated here.

[0143] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0144] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0145] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. An integrated slope monitoring method, characterized in that: The following steps are involved: Divide the highway slope geological area to be monitored into multiple sub-areas; In each sub-area, the laser radar installed on the drone is used to scan with an initial preset point cloud density to obtain the initial surface morphology information of the entire highway slope; For each sub-area, the point cloud data generated by the laser radar is used to obtain the surface morphological information of the sub-area; Extract features of the surface morphology information in the sub-area, and based on the extracted features, use the pre-trained machine learning model to analyze and predict abnormal surface morphology changes in the sub-area under the monitoring window, and identify early signs of abnormal surface morphology changes; Based on the prediction results of the machine learning model, all sub-areas under the monitoring window are divided into abnormal morphological change areas and normal morphological change areas. The normal morphological change areas are areas with stable geological changes, and the abnormal morphological change areas are areas with the risk of abnormal morphological changes. All sub-areas that are judged to be areas of abnormal morphological changes are marked so that they can be focused on in subsequent monitoring; For the sub-areas that are judged to have normal morphological changes, the initially set point cloud density is continued to be used for monitoring; For sub-areas marked as having the risk of abnormal morphological changes, high point cloud density is reallocated to them based on the geological prediction results of the sub-areas; Taking advantage of the reallocated high point cloud density, all marked anomaly areas are further scanned.

2. The integrated slope monitoring method according to claim 1, characterized in that: When dividing the geological area of ​​the highway slope to be monitored into multiple sub-areas, a zoning method based on geological characteristics and terrain data is adopted. High-precision three-dimensional terrain data of the entire area is obtained through lidar. Then, combined with geological survey data, the area is classified and GIS or spatial analysis software is used to divide the highway slope into multiple sub-areas with relatively consistent geological conditions.

3. The integrated slope monitoring method according to claim 1, characterized in that: The surface morphological information of the sub-area is subjected to feature extraction, wherein the extracted features include surface crack density, width and surface area subsidence degree. Under the monitoring window, after analyzing the surface crack density, width and surface area subsidence degree, surface crack quantification value and settlement deformation quantification value are generated respectively. The risk of crack extension is quantified by the surface crack quantification value, and the subsidence risk of the surface area is quantified by the settlement deformation quantification value. The surface crack quantification value and settlement deformation quantification value are input into a pre-trained machine learning model, and a surface morphological change coefficient is generated based on the machine learning model. The surface abnormal morphological changes are analyzed and predicted by the surface morphological change coefficient, and the early signs of abnormal surface morphological changes are identified.

4. The integrated slope monitoring method according to claim 3, characterized in that: In the monitoring window, the surface morphology change coefficient generated by analyzing and predicting the abnormal surface morphology changes in the sub-area based on the machine learning model is compared with the pre-set reference threshold of the surface morphology change coefficient to divide the sub-area. The specific division process is as follows: If the surface morphology change coefficient is greater than or equal to the reference threshold of the surface morphology change coefficient, the sub-region is divided into an abnormal morphology change region, and the normal morphology change region is a region with stable geological changes; If the surface morphology change coefficient is less than the surface morphology change coefficient reference threshold, the sub-region is divided into a normal morphology change region, and the abnormal morphology change region is a region with the risk of abnormal morphology change.

5. The integrated slope monitoring method according to claim 3, characterized in that: In the monitoring window, after analyzing the density and width of surface cracks, the specific steps to generate the quantitative value of surface cracks are as follows: The density and width information of surface cracks are extracted from the point cloud data obtained by the lidar. The crack density indicates the number of cracks per unit area, and the crack width indicates the lateral opening width of the crack. The calculation expression of the crack density is: , where is the crack density, is the number of cracks, A is the area of ​​the sub-region, and the crack width is directly calculated from the point cloud data. express; The crack expansion rate is calculated to reflect the expansion trend of the crack within the monitoring window period. The crack expansion rate is the product of the crack density and the width change rate, which indicates the speed of crack expansion. The calculation expression is: , where is the crack growth rate, and are the start and end time of the monitoring window, and They represent the crack density and crack width at time t respectively; Risk weight coefficients are introduced for different fracture extension conditions. The impact of fractures is weighted according to the changes in fracture density and width. The risk weight coefficient is based on fracture density, width and geological sensitivity in the sub-region. The calculation expression is: , where is the risk weight factor; The final surface crack quantification value is calculated by combining the crack extension rate and risk weight coefficient to quantify the risk of crack extension. The calculation expression is: , where Quantify the value of surface cracks.

6. The integrated slope monitoring method according to claim 5, characterized in that: In the monitoring window, after analyzing the subsidence degree of the surface area, the specific steps to generate the quantitative value of settlement deformation are as follows: Regularly monitor the sub-area to obtain the surface subsidence degree data at different time points, recorded as ,in Indicates the coordinates in the sub-region at time t The degree of subsidence; Calculate the change in the speed of surface subsidence. The calculation expression is: , where represents the sinking speed at time t; In order to quantify the sinking risk more accurately, the change rate of the sinking speed is further calculated. The calculation expression is: , where represents the sinking acceleration at time t; Based on the sinking velocity and sinking acceleration, the settlement deformation quantification value is defined and calculated. The calculation expression is: ， In the formula, Represents the quantitative value of settlement deformation, It means to quadrature the nonlinear changes in the monitoring window to capture the overall sinking trend in the monitoring window.

7. The integrated slope monitoring method according to claim 1, characterized in that: All sub-areas determined to be areas of abnormal morphological change are marked, and the marking method includes assigning identifiers to the areas or setting them as high-priority monitoring areas in the UAV system and monitoring platform.

8. The integrated slope monitoring method according to claim 4, characterized in that: For the sub-areas marked with the risk of abnormal morphological changes, high point cloud density is reallocated to them based on the geological prediction results of the sub-areas. The specific steps are as follows: For the sub-regions that have been marked as having the risk of abnormal morphological changes, the initial point cloud density used in routine monitoring of the sub-region is extracted and recorded as ; Based on the surface morphology monitoring data of the sub-region and the results predicted by the machine learning model, the surface morphology change coefficient is extracted and recalibrated as ; The calculated surface morphology change coefficient is compared with the preset reference threshold of the surface morphology change coefficient to calculate the surface abnormal morphology change deviation. The calculation expression is: ， In the formula, Indicates the deviation of abnormal surface morphology change, Indicates the reference threshold of the surface morphology change coefficient, Indicates the change amplification factor, which is used to amplify significant abnormal surface changes and enhance the identification of risks; In order to accurately capture the abnormal morphological changes in the sub-area, the point cloud density increment is calculated according to the abnormal morphological change deviation of the surface, and the point cloud density is dynamically adjusted through the point cloud density increment. The calculation expression of the point cloud density increment is: ， In the formula, represents the point cloud density increment, k Represents a constant coefficient, which controls the increment between the point cloud density and the surface morphology change coefficient. represents the nonlinear enhancement factor, which adjusts the sensitivity of the surface anomaly morphology change to the point cloud density increment. Represents the density adjustment coefficient, which is used to process the smooth transition of point cloud density adjustment; Based on the calculated point cloud density increment, a new point cloud density is assigned to the sub-area to ensure a more refined monitoring effect. The calculation expression of the new point cloud density is: , where is the new point cloud density, which indicates the monitoring density after redistribution.

9. An integrated slope monitoring system, used to implement the integrated slope monitoring method according to any one of claims 1 to 8, characterized in that: It includes area division module, initial scanning module, surface morphology acquisition module, feature extraction and prediction module, sub-area division module, abnormal area marking module, normal area monitoring module, point cloud density adjustment module and depth scanning module; The regional division module divides the highway slope geological area to be monitored into multiple sub-areas; The initial scanning module uses the laser radar installed on the drone to scan in each sub-area with an initial preset point cloud density to obtain the initial surface morphology information of the entire highway slope; The surface morphology acquisition module uses the point cloud data generated by the laser radar to obtain the surface morphology information of each sub-area; The feature extraction and prediction module extracts features from the surface morphology information in the sub-area. Based on the extracted features, the pre-trained machine learning model is used to analyze and predict the abnormal surface morphology changes in the sub-area under the monitoring window to identify early signs of abnormal surface morphology changes. The sub-region division module divides all sub-regions under the monitoring window into abnormal morphological change regions and normal morphological change regions based on the prediction results of the machine learning model. The normal morphological change regions are regions with stable geological changes, and the abnormal morphological change regions are regions with the risk of abnormal morphological changes. The abnormal area marking module marks all sub-areas that are judged to be areas of abnormal morphological changes so that they can be focused on in subsequent monitoring; The normal area monitoring module continues to monitor the sub-areas that are judged to have normal morphological changes using the initially set point cloud density; Point cloud density adjustment module, for sub-areas marked with the risk of abnormal morphological changes, reallocates high point cloud density to them based on the geological prediction results of the sub-areas; The deep scanning module uses the reallocated high point cloud density to further scan all marked abnormal areas.

Images