A Smart Slope Monitoring System and Method

Abstract

This invention relates to an intelligent slope monitoring system and method. The intelligent slope monitoring system includes: a camera installed on the electrical connection infrastructure of the slope monitoring area to continuously capture original images of the slope; and a processor comprising a feature extraction module, a risk identification module, and a real-time early warning module. The feature extraction module extracts feature parameters of defects from the original images; the risk identification module analyzes the feature parameters to obtain risk indicators; and the real-time early warning module receives and monitors the feature parameters and risk indicators, and sends monitoring data and early warning signals for the corresponding abnormal area when the feature parameters and / or risk indicators exceed a preset abrupt change threshold. The intelligent slope monitoring system of this invention can continuously and stably monitor slope conditions.

Description

Technical Field

[0001] This invention relates to the field of slope monitoring, and in particular to an intelligent slope monitoring system and method. Background Technology

[0002] As a crucial transportation infrastructure, highways are often constructed and operated in complex environments, particularly along their slopes. Mountainous highway slopes, in particular, are prone to landslides, collapses, and rockfalls due to various factors including geological structure, rainfall, weathering, and human engineering activities. These geological hazards pose a serious threat to highway safety and structural stability. Therefore, effective and timely monitoring and early warning systems for highway slopes in mountainous areas are crucial for ensuring traffic safety and minimizing disaster losses.

[0003] Currently, there are various technologies and methods for slope monitoring. Traditional slope monitoring techniques mainly rely on regular manual inspections and the deployment of contact sensors, such as displacement gauges, inclinometers, stress gauges, and pore water pressure gauges. However, manual inspections suffer from problems such as strong subjectivity, low efficiency, high risk, and difficulty in achieving continuous monitoring around the clock, especially for steep slopes or areas with severe weather conditions, where the difficulty and risk of inspections are even greater. Furthermore, the deployment of contact sensors is costly, requiring drilling into the slope structure, which may disturb the slope's inherent stability.

[0004] Therefore, existing slope monitoring technologies widely employ non-contact monitoring methods, utilizing remote sensing and unmanned aerial vehicle (UAV) technologies. For example, satellite remote sensing (such as InSAR technology), terrestrial laser scanning (TLS), or UAVs equipped with cameras and lidar are used to map and monitor slope deformation, thereby acquiring three-dimensional morphology and change information of the slope surface. However, satellite remote sensing technology has limited spatial and temporal resolution, making it difficult to capture early signs of local deformation, and is easily affected by cloud and fog obstruction. Terrestrial laser scanning equipment is expensive, complex to operate, and typically can only perform periodic scans, making continuous real-time monitoring difficult. UAV operations are significantly affected by adverse weather conditions, making flight operations difficult in extreme weather. Furthermore, UAVs have limited flight time and require professional operators. For applications requiring long-term, continuous, and all-weather monitoring, their monitoring costs are high, and the continuity of monitoring data cannot be guaranteed.

[0005] In summary, existing slope monitoring technologies still have many shortcomings in terms of all-weather adaptability and monitoring continuity, making it difficult to fully meet the needs of slope health management throughout its entire life cycle. Summary of the Invention

[0006] Based on this, the purpose of the present invention is to provide an intelligent slope monitoring system and method that combines stability and monitoring continuity.

[0007] This invention provides an intelligent slope monitoring system, comprising:

[0008] The imaging equipment is installed on the electrical connection infrastructure in the slope monitoring area to continuously capture raw images of the slope.

[0009] The processor includes a feature extraction module, a risk identification module, and a real-time early warning module, wherein...

[0010] The feature extraction module is used to extract feature parameters of defects in the original image;

[0011] The risk identification module is used to analyze and obtain risk indicators based on the characteristic parameters of the defects.

[0012] The real-time early warning module is used to receive and monitor characteristic parameters and risk indicators, and when the characteristic parameters and / or risk indicators exceed the preset mutation threshold, it sends monitoring data and early warning signals for the corresponding abnormal area.

[0013] This invention places the imaging equipment on the electrical connection infrastructure of the slope monitoring area to avoid causing additional impact on the slope, while mitigating the impact of weather on the acquisition of the original images, thereby improving the stability and continuity of slope monitoring.

[0014] Furthermore, the processor also includes a periodic evaluation module for periodically predicting the short-term future trend of the slope based on feature parameters and the original image, and obtaining a periodic evaluation report.

[0015] This invention also relates to an intelligent slope monitoring method, comprising:

[0016] The original images of the slope are obtained by continuously capturing images of the slope using imaging equipment installed on the electrical connection foundation in the slope monitoring area.

[0017] Extract feature parameters of defects from the original image;

[0018] Risk indicators are obtained based on the characteristic parameters of the defects;

[0019] It receives and monitors characteristic parameters and risk indicators. When the characteristic parameters and / or risk indicators exceed the preset mutation threshold, it sends monitoring data and early warning signals for the corresponding abnormal area.

[0020] Furthermore, the extraction of feature parameters of defects in the original image includes:

[0021] A deep learning-based convolutional neural network model is used to analyze the original image to identify and locate one or more surface defects. For each identified surface defect, a set of corresponding feature parameters is extracted, including crack feature parameters, displacement feature parameters, and deformation feature parameters. The crack feature parameters include the crack length. and average width The displacement characteristic parameter includes the displacement magnitude d of the characteristic point; the deformation characteristic parameter includes the area of ​​the overall or local deformation of the slope. .

[0022] Furthermore, the risk indicators include a comprehensive risk indicator for specific types of defects and an overall risk indicator. The specific methods for obtaining the risk indicators based on the characteristic parameters of the defects include:

[0023] Receive the characteristic parameters of each defect; according to the preset image space division rules, divide the image of the slope monitoring area into an N×M grid array, with each grid module... Representing a sub-region of the slope, where i is the row index and j is the column index; based on known slope engineering geological assessment methods, risk correlation criteria, and expert evaluation, each grid module in the grid array is assigned a specific value. Geospatial importance weighting coefficients assigning this module a degree of influence on the overall stability of the slope. Combine the characteristic parameters of each defect with the corresponding geospatial importance weighting coefficient. The corresponding comprehensive risk index for specific types of defects is calculated; the overall risk index is further calculated based on the weight of each comprehensive risk index for specific types of defects.

[0024] Furthermore, regarding the detected cracks The total number is bar, length is The average width is ,crack The "basic severity" Represented as:

[0025] ;

[0026] in and These are the weighting coefficients for length and average width, respectively;

[0027] When cracks Completely located in a single mesh module When inside, the crack The risk indicators are:

[0028] ;

[0029] When cracks Spanning multiple mesh modules At this time, assuming a crack Divided into grid array Segments, each segment is [length missing] ,in It's a crack. The sequence number spanning the grid modules, Located in the mesh module Its geospatial importance weighting coefficient is Then the crack Effective weights for:

[0030] ;

[0031] The crack The risk indicators are:

[0032] ;

[0033] After unfolding, we get:

[0034] ;

[0035] Therefore, the comprehensive risk index of cracks is calculated. for:

[0036] ;

[0037] in It is the crack number. .

[0038] Furthermore, for the detected displacement feature points The total number is One, displacement magnitude is Assuming this displacement point Located in the mesh module Within this range, the comprehensive risk index of displacement is calculated. get:

[0039] ;

[0040] in It is a displacement point The geospatial importance weighting coefficient of the grid module in which it is located. It is the index of the displacement characteristic point. .

[0041] Furthermore, regarding the detected deformation areas The total number is A block, with an area of ;

[0042] When the deformation area Completely located in a single mesh module When the deformation occurs, the risk index for that area is:

[0043] ;

[0044] When the deformation area When spanning multiple mesh modules, the deformation region is assumed to be... Divided into grid array Each block has an area of ​​[area missing]. ,in It is a deformable area The sequence number spanning the grid modules, Located in the mesh module Its geospatial importance weighting coefficient is Then the deformation region The risk indicators are:

[0045] ;

[0046] Therefore, the comprehensive deformation risk index is calculated. for:

[0047] ;

[0048] And the overall risk index was calculated. ,in These are the weighting coefficients for the set comprehensive risk indicators for cracks, displacement, and deformation, respectively.

[0049] Furthermore, when any one of the single feature parameters or single risk indicators exceeds the preset mutation threshold, monitoring data and early warning signals for the corresponding abnormal area are sent.

[0050] Furthermore, it also includes:

[0051] Periodic assessment reports are generated by predicting the short-term future trend of slopes based on characteristic parameters and original images, including:

[0052] After each review cycle, the images collected in the current cycle are spatially aligned with those in the previous cycle. The feature parameters of the current cycle and the previous cycle are compared, and the changes and rates of change of the feature parameters are calculated. For feature parameters of multiple consecutive cycles, time series analysis is applied to fit the feature evolution trend and make preliminary predictions on short-term future trends. Based on the preset slope stability evaluation model, the safety level of the slope is assessed, and the stability of the corresponding slope is determined according to the different safety levels. A periodic assessment report containing slope status information, feature quantification statistics, change trend analysis, safety level assessment results, and corresponding maintenance recommendations is generated.

[0053] Compared with existing monitoring methods, the intelligent slope monitoring method of this invention not only achieves stability and continuity detection, but also further optimizes the extraction of feature parameters. By combining the extracted feature parameters with weights and classification criteria, different risk indicators are obtained. It can also obtain the periodic changes of the slope to predict the safety trend of the slope. By monitoring risk indicators and predicting safety trends, it achieves highly accurate intelligent slope monitoring.

[0054] To better understand and implement this invention, the following detailed description is provided in conjunction with the accompanying drawings. Attached Figure Description

[0055] Figure 1 This is a schematic diagram of the slope intelligent monitoring system of the present invention.

[0056] Figure 2 This is a flowchart of the intelligent slope monitoring system of the present invention.

[0057] Figure 3 This is a schematic diagram of the camera installed on a street lamp in the intelligent slope monitoring system of the present invention.

[0058] Figure 4 This is a schematic diagram of the slope monitoring area of ​​the intelligent slope monitoring system of the present invention. Detailed Implementation

[0059] This invention carefully analyzes existing slope monitoring technologies and finds that their poor stability and monitoring continuity are due to excessive influence from weather or weather conditions affecting the slope itself, making continuous monitoring impossible. To address this, this invention attempts to install cameras on highway streetlights, sharing the same circuit with the streetlights to avoid additional impact on the slope, reduce the influence of weather on the acquisition of raw images, and thus improve the stability and continuity of slope monitoring. Furthermore, the image processing workflow for the raw images acquired by the cameras is adjusted to optimize the recognition and analysis of the raw images, ensuring the accuracy of the recognition results.

[0060] Based on this, combined Figure 1 and Figure 2The present invention provides an intelligent slope monitoring system, including a camera 1 and a processor (not shown).

[0061] Camera 1 is mounted on a street lamp along the highway and continuously captures images of the slope to obtain raw images. Camera 1 is a high-definition camera with night vision and image enhancement functions; please refer to [link / reference]. Figure 3 The street light 7 includes a base 71, a pole 72, and a light-emitting element 73. The camera 1 is mounted on the pole 72 of the street light 7 and connected to the existing power system inside the street light 7 for power supply, thereby enabling real-time imaging of the slope of the highway 8; please refer to Figure 4 The width and condition of the highway change with location. On the sloping section 81, there is still some space along the outer edge of the highway where the streetlights are installed; this space constitutes the slope. On the non-sloping section 82, there is almost no extra space after the streetlights are installed, meaning it does not have a slope. The camera 1 does not need to be installed on all the streetlights 7; it only needs to be installed on the sloping section 81 to detect the slope. Therefore, the streetlights with slope locations constitute the slope monitoring area 9, and the camera 1 is installed on the streetlights within the slope monitoring area 9.

[0062] Furthermore, camera 1 is not limited to being mounted on street lamp 7. For slopes with other electrical connection infrastructure, camera 1 can also be mounted on other electrical connection infrastructure, as long as it can provide power and support for camera 1. Similarly, the device for acquiring the raw image of the slope can also be other types of shooting equipment, as long as it can be mounted on electrical connection infrastructure such as street lamps and meets the shooting requirements.

[0063] The processor includes a preprocessing module 2, a feature extraction module 3, a risk identification module 4, a real-time early warning module 5, and a periodic evaluation module 6.

[0064] Preferably, before feature extraction, the preprocessing module 2 performs denoising, enhancement, and distortion correction on the original image to obtain a slope image. Specifically, median filtering is used to denoise the original image, effectively filtering out random noise to obtain a denoised image; histogram equalization is used to enhance the denoised image, adjusting the gray-level distribution to improve its contrast and clarity, making the slope's contours and details more prominent, resulting in a clear image; based on a pre-calibrated camera intrinsic model, geometric transformation is performed on the clear image to correct image distortion caused by lens distortion, thus obtaining the slope image. This ensures the slope image's size and shape are accurate and provides high-quality image data for subsequent processing.

[0065] The feature extraction module 3 is used to extract feature parameters of defects in the original image or slope image. Specifically, a deep learning-based convolutional neural network model is used to analyze the original image or slope image to identify and locate one or more surface defects. For each identified surface defect, corresponding feature parameters are extracted, including crack feature parameters, displacement feature parameters, and deformation feature parameters. The crack feature parameters include the crack length. and average width The displacement feature parameters include the displacement magnitude d of the feature point, which is specifically obtained by comparing the original image or slope image with historical images; the deformation feature parameters include the area of ​​the overall or local deformation of the slope. .

[0066] The risk identification module 4 is used to analyze and obtain risk indicators based on the feature parameters of defects. These risk indicators include the comprehensive risk of specific types of defects and the overall risk indicator. Specifically, the risk identification module receives the feature parameters of each defect output by the feature extraction module; and according to preset original image or slope image spatial division rules, divides the captured image of the slope monitoring area into an N×M grid array, with each grid module... Representing a sub-region of the slope, where i is the row index and j is the column index; based on known slope engineering geological assessment methods, risk correlation criteria, and expert evaluation, the grid modules in the grid array are defined. Assign geospatial importance weighting coefficients to quantify the module's impact on overall slope stability. Combine the characteristic parameters of each defect with the corresponding geospatial importance weighting coefficient. The corresponding comprehensive risk index for specific types of defects is calculated; and the overall risk index is calculated based on the weight of each comprehensive risk index for specific types of defects. In this invention, the comprehensive risk index for specific types of defects includes a comprehensive risk index for cracks, a comprehensive risk index for displacement, and a comprehensive risk index for deformation.

[0067] For the detected cracks The total number is bar, length is The average width is ,crack The "basic severity" Represented as:

[0068]

[0069] in and These are the weighting coefficients for length and average width, respectively;

[0070] When cracks Completely located in a single mesh module When inside, the crack The risk indicators are:

[0071]

[0072] When cracks Spanning multiple mesh modules At this time, assuming a crack Divided into grid array Segments, each segment is [length missing] ,in It's a crack. The sequence number spanning the grid modules, Located in the mesh module Its weight is Then the crack Effective weights for:

[0073]

[0074] The crack The risk indicators are:

[0075]

[0076] After unfolding, we get:

[0077]

[0078] Therefore, the comprehensive risk index of cracks is calculated. for:

[0079]

[0080] in It is the crack number. .

[0081] For the detected displacement feature points The total number is One, displacement magnitude is Assuming the displacement point Located in the mesh module Within this range, the comprehensive risk index of displacement is calculated. get:

[0082]

[0083] in It is a displacement point The geospatial importance weighting coefficient of the grid module in which it is located. It is the index of the displacement characteristic point. .

[0084] For the detected deformation area The total number is A block, with an area of ;

[0085] When the deformation area Completely located in a single mesh module When the deformation occurs, the risk index for that area is:

[0086]

[0087] When the deformation area When spanning multiple mesh modules, the deformation region is assumed to be... Divided into grid array Each block has an area of ​​[area missing]. ,in It is a deformable area The sequence number spanning the grid modules, Located in the mesh module Its weight is Then the deformation region The risk indicators are:

[0088]

[0089] Therefore, the comprehensive deformation risk index is calculated. for:

[0090]

[0091] The overall risk index is thus calculated. ,in These are the weighting coefficients for each comprehensive risk indicator, which can be adjusted according to actual needs.

[0092] The real-time early warning module 5 is used to receive and monitor the characteristic parameters and risk indicators of defects. These risk indicators include the comprehensive risk indicator for a specific type of defect and the overall risk indicator calculated above. When a single characteristic parameter, the comprehensive risk indicator for a single specific type of defect, or the overall risk indicator exceeds a preset mutation threshold, monitoring data and an early warning signal for the corresponding abnormal area are sent. Specifically, when the characteristic parameters of the defect, the comprehensive risk indicator for a specific type of defect, and the overall risk indicator meet any of the following conditions, the real-time early warning module 5 sends an early warning signal and simultaneously sends monitoring data for the corresponding abnormal area to facilitate observation of the data in the abnormal area:

[0093] When the mutation of a single feature parameter exceeds a preset mutation threshold, at this time... > ,in The mutation threshold preset for a certain feature parameter; specifically, it can be set. , ,

[0094] When the comprehensive risk index of a specific type of defect exceeds the preset mutation threshold, at this time... ,in The mutation threshold preset for a comprehensive risk index of a specific type of defect can be set. , ， .

[0095] When the overall risk index exceeds the preset mutation threshold, at this time ,in The mutation threshold preset for the overall risk indicator.

[0096] Specifically, when an alert is triggered, the real-time alert module 5 will automatically package information such as the location of the abnormal area, image or video clips, risk type, and characteristic parameters and / or risk indicators of the abnormal area and send them to a display screen or other device with display function in the form of an alert data packet, so that the specific changes in the abnormal area where the risk has occurred can be observed intuitively.

[0097] In addition to real-time monitoring, to achieve better intelligent monitoring of slopes, the prediction of future slope trends is also considered. To this end, the periodic assessment module 6 is used to periodically predict the short-term future trend of the slope based on characteristic parameters and the original image or slope image, and obtain a periodic assessment report.

[0098] Specifically, in each review cycle After completion, the images acquired in the current period are spatially aligned with those from the previous period to ensure the accuracy of feature information comparison; the current period is then compared. Compared with the previous cycle Feature parameters Calculate the change = With rate of change For characteristic parameters across multiple consecutive periods, time series analysis is applied to fit the evolution trend of the characteristics, and preliminary predictions of short-term future trends are made; based on a pre-defined slope stability evaluation model... Assess the safety level of the slope. ,in This is a vector representing the changes in slope characteristic parameters. This is the rate-of-change vector of slope characteristic parameters; based on the safety level Different judgments correspond to the stability level of the slope, which includes stable, basically stable, understability and unstable; and the periodic assessment module 6 also generates a periodic assessment report containing slope status information, characteristic quantitative statistics, change trend analysis, safety level assessment results and corresponding maintenance suggestions.

[0099] The intelligent slope monitoring system of this invention acquires slope-related information by installing cameras on streetlights in road sections with slopes. It also extracts specific feature parameters from this information and calculates different risk indicators based on different weights and classification standards. By comprehensively judging the anomalies of risk indicators and feature parameters, and predicting the safety trend of slopes through periodic changes, it achieves high-precision intelligent monitoring.

[0100] The embodiments described above merely illustrate the preferred implementation of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and the present invention also intends to include these modifications and variations.

Claims

1. A slope intelligent monitoring system, comprising: The imaging equipment is installed on the electrical connection infrastructure in the slope monitoring area to continuously capture raw images of the slope. The processor includes a feature extraction module, a risk identification module, and a real-time early warning module, wherein... The feature extraction module is used to extract feature parameters of defects in the original image; The risk identification module is used to analyze and obtain risk indicators based on the characteristic parameters of the defects. The real-time early warning module is used to receive and monitor characteristic parameters and risk indicators, and when the characteristic parameters and / or risk indicators exceed the preset mutation threshold, it sends monitoring data and early warning signals for the corresponding abnormal area. The characteristic parameters include crack characteristic parameters, which include crack length and average width; the risk indicators include a comprehensive risk indicator for a specific type of defect, and the risk indicators obtained from the analysis of the defect characteristic parameters specifically include: Receive the characteristic parameters of each defect; according to the preset image space division rules, divide the image of the slope monitoring area into an N×M grid array, with each grid module... Representing a sub-region of the slope, where i is the row index and j is the column index; based on known slope engineering geological assessment methods, risk correlation criteria, and expert evaluation, each grid module in the grid array is assigned a specific value. Geospatial importance weighting coefficients assigning this module a degree of influence on the overall stability of the slope. Combine the characteristic parameters of each defect with the corresponding geospatial importance weighting coefficient. The corresponding comprehensive risk index for specific types of defects is calculated. For the detected cracks The total number is bar, length is The average width is ,crack "Basic severity" Represented as: ； in and These are the weighting coefficients for length and average width, respectively; When cracks Completely located in a single mesh module When inside, the crack The risk indicators are: ； When cracks Spanning multiple mesh modules At this time, assuming a crack Divided into grid array Segments, each segment is [length missing] ,in It's a crack. The sequence number spanning the grid modules, Located in the mesh module Its geospatial importance weighting coefficient is Then the crack Effective weights for: ； The crack The risk indicators are: ； After unfolding, we get: ； Therefore, the comprehensive risk index of cracks is calculated. for: ； in It is the crack number. .

2. The intelligent slope monitoring system according to claim 1, characterized in that: The processor also includes a periodic evaluation module for periodically predicting the short-term future trend of the slope based on feature parameters and the original image, and obtaining a periodic evaluation report.

3. A method for intelligent slope monitoring, comprising: The original images of the slope are obtained by continuously capturing images of the slope using imaging equipment installed on the electrical connection foundation in the slope monitoring area. Extract feature parameters of defects from the original image; Risk indicators are obtained based on the characteristic parameters of the defects; Receive and monitor characteristic parameters and risk indicators. When the characteristic parameters and / or risk indicators exceed the preset mutation threshold, send monitoring data and early warning signals for the corresponding abnormal area. The characteristic parameters include crack characteristic parameters, which include crack length and average width; the risk indicators include a comprehensive risk indicator for a specific type of defect, and the risk indicators obtained from the analysis of the defect characteristic parameters specifically include: Receive the characteristic parameters of each defect; according to the preset image space division rules, divide the image of the slope monitoring area into an N×M grid array, with each grid module... Representing a sub-region of the slope, where i is the row index and j is the column index; based on known slope engineering geological assessment methods, risk correlation criteria, and expert evaluation, each grid module in the grid array is assigned a specific value. Geospatial importance weighting coefficients assigning this module a degree of influence on the overall stability of the slope. Combine the characteristic parameters of each defect with the corresponding geospatial importance weighting coefficient. The corresponding comprehensive risk index for specific types of defects is calculated. For the detected cracks The total number is bar, length is The average width is ,crack "Basic severity" Represented as: ； in and These are the weighting coefficients for length and average width, respectively; When cracks Completely located in a single mesh module When inside, the crack The risk indicators are: ； When cracks Spanning multiple mesh modules At this time, assuming a crack Divided into grid array Segments, each segment is [length missing] ,in It's a crack. The sequence number spanning the grid modules, Located in the mesh module Its geospatial importance weighting coefficient is Then the crack Effective weights for: ； The crack The risk indicators are: ； After unfolding, we get: ； Therefore, the comprehensive risk index of cracks is calculated. for: ； in It is the crack number. .

4. The intelligent slope monitoring method according to claim 3, characterized in that: The feature parameters for extracting defects from the original image include: A deep learning-based convolutional neural network model is used to analyze the original image to identify and locate one or more surface defects. For each identified surface defect, a set of corresponding feature parameters are extracted, including crack feature parameters, displacement feature parameters, and deformation feature parameters. The displacement feature parameters include the displacement magnitude of the feature point, and the deformation feature parameters include the area of ​​the slope's overall or local deformation.

5. The intelligent slope monitoring method according to claim 4, characterized in that: The risk indicators also include an overall risk indicator, which is further calculated based on the weight of the comprehensive risk indicator for each specific type of defect.

6. The intelligent slope monitoring method according to claim 5, characterized in that: For the detected displacement feature points The total number is One, displacement magnitude is Assuming this displacement point Located in the mesh module Within this range, the comprehensive risk index of displacement is calculated. get: ； in It is a displacement point The geospatial importance weighting coefficient of the grid module in which it is located. It is the index of the displacement characteristic point. .

7. The intelligent slope monitoring method according to claim 6, characterized in that: For the detected deformation area The total number is A block, with an area of ; When the deformation area Completely located in a single mesh module When the deformation occurs, the risk index for that area is: ； When the deformation area When spanning multiple mesh modules, the deformation region is assumed to be... Divided into grid array Each block has an area of ​​[area missing]. ,in It is a deformable area The sequence number spanning the grid modules, Located in the mesh module Its geospatial importance weighting coefficient is Then the deformation region The risk indicators are: ； Therefore, the comprehensive deformation risk index is calculated. for: ； And the overall risk index was calculated. ,in These are the weighting coefficients for the set comprehensive risk indicators for cracks, displacement, and deformation, respectively.

8. The intelligent slope monitoring method according to claim 7, characterized in that: When any one of the single feature parameters or single risk indicators exceeds the preset mutation threshold, the monitoring data and early warning signal of the corresponding abnormal area are sent.

9. The intelligent slope monitoring method according to claim 8, characterized in that: Also includes: Periodic assessment reports are generated by predicting the short-term future trend of slopes based on characteristic parameters and original images, including: After each review cycle, the images collected in the current cycle are spatially aligned with those in the previous cycle. The feature parameters of the current cycle and the previous cycle are compared, and the changes and rates of change of the feature parameters are calculated. For feature parameters of multiple consecutive cycles, time series analysis is applied to fit the feature evolution trend and make preliminary predictions on short-term future trends. Based on the preset slope stability evaluation model, the safety level of the slope is assessed, and the stability of the corresponding slope is determined according to the different safety levels. A periodic assessment report containing slope status information, feature quantification statistics, change trend analysis, safety level assessment results, and corresponding maintenance recommendations is generated.

Images