Geologic feature-based regional landslide hidden danger automatic identification system and method

Abstract

The invention discloses a geological feature-based regional landslide hidden danger automatic identification system and method, and relates to the technical field of intelligent sensors. The geological feature-based regional landslide hidden danger automatic identification method comprises the steps of S1, acquiring and preprocessing multi-source monitoring data and geological constraint data, and constructing a landslide monitoring database; s2, based on the weighted sensitivity and residual accumulation of the multi-source monitoring data, analyzing the credibility of the physical property parameters of the slip band; s3, spatial coupling analysis is carried out through multi-source deformation and geometric parameter data, and sliding band form recognition and connectivity adjustment operation are dynamically optimized; s4, dynamic instability analysis is carried out through deformation rate and sliding surface geometric fusion data, and the slip band instability trend is quantified; and S5, performing comprehensive judgment on the stable state of the monitoring unit by fusing the three-dimensional sliding surface probability body and the spatio-temporal evolution index. The problems of inaccurate slip zone parameter inversion, low slip surface judgment precision and early warning lag in regional landslide hidden danger identification are solved.

Description

Technical Field

[0001] This invention relates to the field of intelligent sensor technology, specifically to an automatic identification system and method for regional landslide hazards based on geological characteristics. Background Technology

[0002] Existing landslide monitoring and hazard identification technologies rely on a combination of multi-source sensor collaborative observation and geological constraint modeling. They utilize shallow drilling results, geological survey lines, and profile analysis to determine geometric parameters such as the burial depth of the slip surface, strike deviation angle, and dip deviation angle. Stability assessment and risk classification are then performed through comprehensive time-series analysis of deformation rate, pore water pressure, and rainfall response. These existing technologies have formed a comprehensive landslide hazard identification system integrating multi-source monitoring, parameter inversion, geometric modeling, and dynamic assessment.

[0003] For example, the invention patent with publication number CN112946240B discloses a landslide geological disaster gene identification and prediction system. This system includes a slope big data unit, a feature data analysis unit, and a landslide identification and prediction unit. Through slope big data, the system performs real-time identification and prediction of slope safety and stability. Without requiring quantitative calculation and analysis, it identifies, evaluates, and predicts landslides based on real-time and historical data from the sky, ground, and human environment in the slope area. Furthermore, by identifying and filtering slope disease genes, it can repair disaster-causing genes and prevent landslide disasters. Simultaneously, this invention overcomes the limitations of traditional landslide prediction methods when dealing with big data issues. By connecting with meteorological satellites and using the BeiDou system for positioning, it can achieve prediction and forecasting of landslide geological disasters. It is simple to operate, handles massive amounts of data, and boasts a high degree of automation, intelligence, and identification accuracy.

[0004] For example, the invention patent with publication number CN109470775A discloses a device and method for identifying and monitoring infrasound signals in soil landslides in the field. The device includes: several above-ground infrasound monitoring stations arranged in an array and one underground infrasound monitoring station; the above-ground and underground monitoring stations are respectively connected to a data acquisition center; the above-ground infrasound monitoring station includes: a monitoring pile (1), a base (3), a primary acoustic sensor (4), and a waterproof box (5). The monitoring pile (1) is installed and fixed on the base (3) on the ground surface, and the primary acoustic sensor (4) is installed in the waterproof box (5) on the monitoring pile (1). Based on the above device, this invention also provides a method for identifying and monitoring infrasound signals in soil landslides in the field, which can monitor landslides. This invention adopts a combined surface and underground monitoring method to eliminate interference from infrasound sources in the environment and improve the detection capability of infrasound signals in landslides.

[0005] However, existing technologies still have shortcomings in the reliability assessment and dynamic fusion of multi-source monitoring data. On the one hand, physical quantities collected by different sensors exhibit time drift, spatial unevenness, and response lag, resulting in insufficient stability and reliability of the sliding surface parameter inversion results. On the other hand, the dynamic evolution process of the sliding surface geometry is not coupled with the monitoring data in real time, making it difficult to achieve adaptive identification and connectivity correction of the sliding surface morphology.

[0006] Therefore, in order to address the above problems, there is an urgent need for an automatic identification system and method for regional landslide hazards based on geological characteristics. Summary of the Invention

[0007] Technical problems to be solved To address the shortcomings of existing technologies, this invention provides an automatic identification system and method for regional landslide hazards based on geological features, which solves the problems of inaccurate inversion of slip zone parameters, low accuracy of slip surface determination, and delayed early warning in the identification of regional landslide hazards.

[0008] Technical solution To achieve the above objectives, this invention provides the following technical solution: an automatic identification system and method for regional landslide hazards based on geological characteristics, comprising: S1, acquiring multi-source monitoring data and geological constraint data and preprocessing them to construct a landslide monitoring database; S2, analyzing the reliability of slip zone physical parameters based on the weighted sensitivity and residual accumulation of multi-source monitoring data, and dynamically evaluating and optimizing the slip zone physical parameter inversion results based on the reliability analysis results; S3, performing spatial coupling analysis through multi-source deformation and geometric parameter data, and dynamically optimizing slip zone morphology identification and connectivity adjustment operations based on the coupling analysis results; S4, performing dynamic instability analysis through deformation rate and slip surface geometry fusion data, quantifying the slip zone instability trend, and performing risk response and stability verification operations based on the instability analysis results; S5, comprehensively determining the stability state of monitoring units by integrating three-dimensional slip surface probability volume and spatiotemporal evolution indicators, and implementing hierarchical monitoring and early warning based on the determination results.

[0009] Furthermore, the specific steps for acquiring and preprocessing multi-source monitoring data and geological constraint data to construct a landslide monitoring database are as follows: Collecting multi-source monitoring data and geological constraint data required for landslide hazard identification. Multi-source monitoring data includes rainfall, three-day cumulative rainfall, slope angle, soil saturation, pore water pressure, TDR reflectance index, infiltration rate, volumetric water content, soil temperature, and landslide displacement. Geological constraint data includes shallow drilling results, landslide surface depth, landslide strike deviation angle, and landslide dip angle deviation angle. The time elapsed from the completion of one synchronous acquisition of multi-source monitoring data by the monitoring equipment to the completion of the next similar data acquisition is also considered. The time interval, used to characterize the smallest synchronous sampling unit of landslide monitoring data in the time dimension, is denoted as the sampling period; rainfall and three-day cumulative rainfall data are acquired, and the precipitation process is continuously recorded using a rain gauge sensor. Rainfall intensity, cumulative rainfall, and the three-day cumulative value of landslide data are obtained through time series statistical analysis; slope angle data are acquired, and the slope angle distribution is calculated based on topographic mapping and digital elevation models; infiltration rate and volumetric water content data are acquired, and soil moisture content changes are simultaneously monitored using a combination of infiltration rate probes and TDR probes. Infiltration rate and volumetric water content are obtained through time series analysis, and the monitoring sequence is compared with rainfall and... Comparative analysis of pore pressure changes yielded parameter inversion residuals; pore water pressure data was acquired by synchronously collecting pore pressure responses at different burial depths using a vibrating wire pore pressure gauge, and the characteristics of pore water pressure changes and parameter inversion residuals were obtained through time-series fluctuation analysis coupled with infiltration rate data; TDR reflectance index and soil saturation data were acquired, and soil saturation was derived and the reflectance signal change sequence was recorded based on TDR waveform characteristic analysis and volumetric water content calculation results; soil temperature data was acquired by collecting temperature changes at multiple depth locations using a thermistor probe, and the soil temperature distribution was obtained through trend analysis; and landslide displacement data was acquired by using tilting... The inclinometer and GNSS displacement monitoring device collect displacement time series, and obtain the monitored displacement through filtering and time alignment processing. The displacement series is then exported and analyzed to obtain the deformation rate. Strain data is acquired, and strain gauges are used to collect micro-strain changes within the sliding body. Strain gradient analysis is used to obtain the stress concentration distribution within the sliding zone. Geometric information of the sliding surface is obtained. Combined with shallow drilling results and geological survey data, spatial interpolation analysis is used to obtain the sliding surface burial depth. directional statistical analysis of geological survey and shallow drilling data is used to obtain the sliding surface strike deviation angle. Fitting analysis of geological profile survey lines and dip angle measurement data is used to obtain the sliding surface dip angle deviation angle. Parameters Specifically, this includes rainfall, three-day cumulative rainfall, slope angle, soil saturation, pore water pressure, and TDR reflectance index. Based on the monitoring data collection rules, the formats, timestamp accuracy, and spatial coding rules for rainfall, infiltration rate, volumetric water content, pore water pressure, soil temperature, landslide displacement, strain, and deformation rate are standardized to ensure consistency of monitored data in both time and space. Linear mapping normalization is performed on numerical parameters such as rainfall, infiltration rate, volumetric water content, pore pressure, displacement, and deformation rate. Angular and spatial parameters are mapped to corresponding levels based on geological profiles, ensuring all monitoring data maintain a uniform magnitude. After standardization and normalization, multi-source monitoring data and geological constraint data are stored to construct a landslide monitoring database.

[0010] Furthermore, based on the weighted sensitivity and residual accumulation of multi-source monitoring data, the specific steps for analyzing the reliability of the sliding band physical parameters are as follows: Obtaining parameters... The parameter inversion residual, parameter Specifically, this includes rainfall, three-day cumulative rainfall, slope angle, soil saturation, pore water pressure, and TDR reflection index; the weighted residual is obtained by multiplying the parameter inversion residual of parameter j by the sensitivity weighting factor, and the weighted residuals of each parameter are summed to obtain the weighted residual accumulation value. The confidence value of the slip zone parameter is obtained by subtracting the weighted residual accumulation value from the sum of the weighted residual accumulation values.

[0011] Furthermore, the specific steps for realizing the dynamic evaluation and optimization correction of the inversion results of the physical property parameters of the sliding zone according to the credible analysis results are as follows: By comparing the credible values of the sliding zone parameters with the trust thresholds T1 and T2 in real time, the dynamic control and optimization of the spatial inversion quality of the sliding zone parameters can be achieved. When the credible value of the sliding zone parameter < T1, the number of combined units of rain gauges, vibrating wire piezometers, soil moisture probes, infiltration rate probes and TDR probes is increased to improve the sampling density of rainfall, pore water pressure, volumetric water content and infiltration rate monitoring data: Infiltration rate and volumetric water content probes are arranged at different elevation positions along the vertical direction of the slope, and shallow and deep piezometers are configured at the corresponding positions, and a vertical monitoring chain is formed in cooperation with inclinometers; During rainfall, the infiltration rate, pore pressure and deformation rate are recorded synchronously with the sampling period, and the AI spatial interpolation algorithm is linked to update the monitoring data in real time and recalculate the credible value of the sliding zone parameter. If the credible value of the sliding zone parameter is still less than the trust threshold, verification tests are carried out on key areas, infiltration experiments and shear tests are carried out, the value range of the monitoring data is revised and the physical property continuous field is updated; Generate an on-site review warning report and file it in the landslide slip surface monitoring database; When T1 ≤ credible value of the sliding zone parameter < T2, maintain the existing monitoring points and sampling frequencies,剔除异常数据并重建时间序列,在参数计算中参数的敏感度权重因子,重新计算黏聚力、内摩擦角、渗透率与饱和度代表值,进行短期渗透试验与剪切试验以验证修正结果;当滑带参数可信值≥T2时,发布滑带物性连续场与不确定性分布结果,对孔隙水压力、入渗率与体积含水率监测数据每周执行一次漂移校验,在每次降雨过程结束后重新计算滑带参数可信值,当重新计算的滑带参数可信值连续n次低于降雨前的平均滑带参数可信值时发布预警报告。

[0012] Furthermore, the specific steps for performing spatial coupling analysis through multi-source deformation and geometric parameter data are as follows: Obtain the credible value of the sliding zone parameter, the monitored displacement, the deviation angle of the sliding surface trend, the deviation angle of the sliding surface inclination and the buried depth of the sliding surface; After summing the deviation angle of the sliding surface trend and the deviation angle of the inclination, divide by the buried depth of the sliding surface, then multiply by the angle correction factor, and take the negative number as the exponential power for exponential calculation to obtain the spatial geometric error attenuation value; Calculate the product result of the credible value of the sliding zone parameter, the monitored displacement and the spatial geometric error attenuation value to obtain the sliding surface geometric fusion value.

[0013] It should be noted that there is an unclear part in the translation of which says "剔除异常数据并重建时间序列,在参数计算中参数的敏感度权重因子,重新计算黏聚力、内摩擦角、渗透率与饱和度代表值,进行短期渗透试验与剪切试验以验证修正结果;当滑带参数可信值≥T2时,发布滑带物性连续场与不确定性分布结果,对孔隙水压力、入渗率与体积含水率监测数据每周执行一次漂移校验,在每次降雨过程结束后重新计算滑带参数可信值,当重新计算的滑带参数可信值连续n次低于降雨前的平均滑带参数可信值时发布预警报告。", please check and correct it according to the original Chinese text for a more accurate translation.Furthermore, based on the coupling analysis results, the specific steps for dynamically optimizing the slip surface morphology recognition and connectivity adjustment are as follows: Real-time comparison of the slip surface geometric fusion value and the fusion threshold. When the slip surface geometric fusion value is less than the fusion threshold, the 3D search strategy is readjusted: the depth, strike, and dip angle search steps are increased within the slope unit range, and the curvature extremum lines and ditch edge lines are re-extracted to correct the slip surface projection; shallow drilling points and inclinometers are deployed in areas with concentrated strike and dip angle differences to supplement deformation data and recalculate the slip surface geometric fusion value. If the slip surface geometric fusion value is still less than the fusion threshold after m iterations, a spatial criterion anomaly report is generated. It is recommended to conduct on-site verification and archive the data to the landslide monitoring database. When the sliding surface geometric fusion value is greater than or equal to the fusion threshold, all voxels that meet the conditions are connected according to their spatial adjacency. Adjacent voxels whose strike difference and dip difference are both less than the geometric difference control value are used as connecting nodes, and connected blocks are formed in sequence. The nodes in the connected blocks are sorted from largest to smallest according to the sliding surface geometric fusion value. Starting with the largest sliding surface geometric fusion value, the main sliding zone skeleton is generated along the path with the strongest directional consistency. Voxels whose dip difference and strike difference are both less than the limit value are selected on both sides of the skeleton for supplementary splicing. The output is a three-dimensional sliding surface probability volume.

[0014] Furthermore, the specific steps for quantifying the slip zone instability trend through dynamic instability analysis using fused data of deformation rate and slip surface geometry are as follows: obtain the fused values ​​of deformation rate and slip surface geometry; multiply the fused values ​​of slip surface geometry by the geometric amplification factor and use the result as an exponent to calculate the geometric influence term; multiply the deformation rate by the geometric influence term to obtain the risk driving base quantity; add one to the obtained risk driving base quantity and take the natural logarithm to obtain the time-varying instability driving value.

[0015] Furthermore, the specific steps for performing risk response and stability verification operations based on the instability analysis results are as follows: By comparing the time-varying instability driving value with the stability threshold in real time, when the time-varying instability driving value is less than the stability threshold, the monitoring deployment and sampling frequency are maintained, and the monitoring data of inclinometers, pore pressure gauges, and rain gauges are checked for consistency, the data records are updated and smoothed to form a continuous time-series curve to track the deformation trend; when the time-varying instability driving value is greater than or equal to the stability threshold, short-term trend fitting is performed on the monitoring data to extract the pore pressure growth rate and deformation acceleration; drainage pipes and strain gauges are deployed in the time-varying instability driving value growth area, and a small-scale load reduction test is carried out to verify the monitoring response and form a time-varying instability driving response rate curve; when the time-varying instability driving value calculated n times consecutively is less than the stability threshold, the normal sampling interval is restored, and the time-series driving correction data is recorded and archived to the landslide monitoring database.

[0016] Furthermore, the specific steps for comprehensively judging the stability state of monitoring units by integrating the three-dimensional slip surface probability volume and spatiotemporal evolution indicators, and implementing hierarchical monitoring and early warning based on the judgment results, are as follows: Combining the spatial distribution characteristic map of slip zone parameters, the three-dimensional slip surface probability volume, time-varying instability driving values, slip surface geometric fusion values, and slip zone parameter reliability values, the stability state of monitoring points is comprehensively judged and risk level classification is implemented; when the time-varying instability driving value is less than the stability threshold, the area is designated as a general monitoring area, and data on deformation rate, pore water pressure, and rainfall are regularly summarized to update the risk assessment file; when the time-varying instability driving value is greater than or equal to the stability threshold, and the slip surface geometric fusion value is greater than the fusion threshold, the area is marked as a key early warning area, a risk early warning report is pushed out, and on-site investigation, drainage, and reinforcement engineering measures are coordinated; based on the reliability distribution pattern shown in the spatial distribution characteristic map of slip zone parameters, zoned and intensified monitoring and increased sampling frequency are implemented in areas with low reliability values, generating risk time-series change reports and archiving them to the landslide monitoring database.

[0017] Furthermore, a second aspect of the present invention provides an automatic identification system for regional landslide hazards based on geological features, applied to an automatic identification method for regional landslide hazards based on geological features, comprising: a landslide data acquisition and preprocessing module for acquiring and preprocessing multi-source monitoring data and geological constraint data to construct a landslide monitoring database; a physical parameter inversion and slip zone feature estimation module for analyzing the reliability of slip zone physical parameters based on the weighted sensitivity and residual accumulation of multi-source monitoring data, and realizing dynamic evaluation and optimization correction of slip zone physical parameter inversion results based on the reliability analysis results; and a slip surface geometry estimation module. The connectivity and discontinuity optimization module is used to perform spatial coupling analysis using multi-source deformation and geometric parameter data, and dynamically optimize the slip surface morphology identification and connectivity adjustment operations based on the coupling analysis results. The dynamic evolution and uncertainty update module is used to perform dynamic instability analysis using deformation rate and slip surface geometry fusion data, quantify slip surface instability trends, and perform risk response and stability verification operations based on the instability analysis results. The risk level assessment and early warning module is used to integrate three-dimensional slip surface probability volume and spatiotemporal evolution indicators to comprehensively determine the stability state of the monitoring unit, and implement hierarchical monitoring and early warning based on the determination results.

[0018] The present invention has the following beneficial effects: (1) This invention, by comprehensively considering the influence of each monitoring parameter in the sliding zone feature inversion and its deviation performance, and based on the sensitivity weight and parameter residual information, carries out a credible analysis, thereby realizing the quantitative expression and dynamic update of the credibility of the sliding zone physical parameters, effectively solving the problem that the reliability of parameter inversion results is difficult to evaluate in the prior art.

[0019] (2) This invention adaptively adjusts the sampling density, monitoring deployment and physical property parameter range by comparing the reliable value of the sliding belt parameter with the trust threshold, thereby realizing the dynamic optimization of the sliding belt parameter inversion result and the linkage of on-site verification, effectively solving the problems of insufficient monitoring accuracy and data correction lag in the prior art.

[0020] (3) This invention, by comprehensively analyzing the reliable values ​​of sliding strip parameters, monitored displacement, sliding surface deviation angle, sliding surface inclination deviation angle, and sliding surface burial depth, extracts spatial geometric feature relationships, thereby achieving continuous identification of the spatial morphology of the sliding strip and dynamic correction of the sliding surface geometry, effectively solving the problems of blurred sliding surface boundaries and insufficient spatial connectivity in the prior art.

[0021] (4) This invention combines the deformation rate with the geometric fusion value of the sliding surface to reflect the coupling change trend of the deformation of the sliding body and the geometric constraints, thereby realizing the effect of time-series characterization and trend quantification of the slip zone instability process, effectively solving the problems of lagging identification of landslide instability evolution law and slow risk response in the prior art.

[0022] Of course, any product implementing this invention does not necessarily need to achieve all the advantages described above simultaneously. Spatial distribution characteristic diagram of sliding strip parameters. Attached Figure Description

[0023] Figure 1 This is a flowchart of the automatic identification method for regional landslide hazards based on geological features according to the present invention; Figure 2 This is a structural diagram of the regional landslide hazard automatic identification system based on geological features of the present invention; Figure 3 This is a spatial distribution feature map of the slip zone parameters of the regional landslide hazard automatic identification system and method based on geological features of the present invention; Figure 4 This is a flowchart of the optimization process for the geometric fusion and spatial connectivity discrimination of the regional landslide hazard automatic identification system based on geological features, as described in this invention. Detailed Implementation

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

[0025] Please see Figures 1-4This invention provides a technical solution: an automatic identification system and method for regional landslide hazards based on geological features, comprising: S1, acquiring multi-source monitoring data and geological constraint data and preprocessing them to construct a landslide monitoring database; S2, analyzing the reliability of slip zone physical parameters based on the weighted sensitivity and residual accumulation of multi-source monitoring data, and dynamically evaluating and optimizing the slip zone physical parameter inversion results based on the reliability analysis results; S3, performing spatial coupling analysis through multi-source deformation and geometric parameter data, and dynamically optimizing slip zone morphology identification and connectivity adjustment operations based on the coupling analysis results; S4, performing dynamic instability analysis through deformation rate and slip surface geometry fusion data, quantifying the slip zone instability trend, and performing risk response and stability verification operations based on the instability analysis results; S5, comprehensively determining the stability state of monitoring units by integrating three-dimensional slip surface probability volume and spatiotemporal evolution indicators, and implementing hierarchical monitoring and early warning based on the determination results.

[0026] Specifically, the steps for acquiring and preprocessing multi-source monitoring data and geological constraint data to construct a landslide monitoring database are as follows: First, collect the multi-source monitoring data and geological constraint data required for landslide hazard identification. The multi-source monitoring data includes rainfall, three-day cumulative rainfall, slope angle, soil saturation, pore water pressure, TDR reflectance index, infiltration rate, volumetric water content, soil temperature, and landslide displacement. This multi-source monitoring data is acquired simultaneously by monitoring equipment deployed at different elevations and depths to ensure the representativeness and continuity of the data in both vertical and horizontal spaces. The geological constraint data includes shallow drilling results, landslide surface depth, landslide strike deviation angle, and landslide dip deviation angle. The stratigraphic structure and landslide surface spatial morphology are jointly verified through field sampling, geological mapping, and borehole exposure. The time interval between the completion of a single synchronous acquisition of multi-source monitoring data by the monitoring equipment and the completion of the next similar acquisition is used to characterize the smallest synchronous sampling unit of landslide monitoring data in the time dimension, and is denoted as the sampling period. This sampling period definition ensures that all sensor-collected data maintain strict alignment in timestamps and spatial encoding, providing a time reference for subsequent multi-source data fusion. When acquiring rainfall and three-day cumulative rainfall data, rain gauge sensors are used to continuously record the precipitation process, and time series statistical analysis is combined to obtain rainfall intensity, cumulative rainfall, and the three-day cumulative value of landslide, which is used to reveal the hysteresis effect of rainfall on changes in infiltration rate and pore pressure. Slope angle data is acquired, and the slope angle distribution is calculated based on topographic mapping and digital elevation models. The results are compared and verified with the shallow drilling exposed profile to ensure that the slope angle parameters reflect the true topographic undulations of the landslide body. Infiltration rate and volumetric water content data were acquired. A combination of an infiltration rate probe and a TDR probe was used to synchronously monitor soil water content changes. The monitoring process was time-aligned according to the sampling cycle. Infiltration rate and volumetric water content were obtained through time series analysis. The monitoring sequence was compared with rainfall and pore pressure changes to obtain parameter inversion residuals, which were used for subsequent calculation of slip zone parameter reliability values. Pore water pressure data were acquired. A vibrating wire pore pressure gauge was used to synchronously collect pore pressure responses at different burial depths. Time series fluctuation analysis was coupled with infiltration rate data to obtain pore water pressure change characteristics and parameter inversion residuals, thus characterizing the dynamic impact of rainfall infiltration on groundwater pressure. TDR reflectance index and soil saturation data were acquired. Based on TDR waveform feature analysis and volumetric water content calculation results, soil saturation was derived, and the reflectance signal change sequence was recorded to identify abrupt changes in soil water content and changes in saturation layer depth. Soil temperature data was acquired by collecting temperature changes at multiple depth locations using thermistor probes. Trend analysis revealed the soil temperature distribution, and the thermal-water coupling characteristics were analyzed in conjunction with changes in infiltration rate and saturation. Landslide displacement data was also acquired by collecting displacement time series data using inclinometers and GNSS displacement monitoring devices. Filtering and time alignment were performed to obtain the monitored displacement values, and the displacement series was exported and analyzed to obtain the deformation rate, reflecting the landslide's movement trend and stability changes.Strain data was acquired by using strain gauges to collect micro-strain changes within the slip zone. Strain gradient analysis revealed the stress concentration distribution within the slip zone, providing physical constraints for slip zone morphology identification. Geometric information of the slip surface was obtained by combining shallow drilling data with geological survey data. Spatial interpolation analysis determined the slip surface depth, and statistical analysis of the directions of geological surveys and shallow drilling data yielded the slip surface strike deviation angle. Fitting analysis of geological profile survey lines and dip angle measurement data provided the slip surface dip angle deviation angle, offering fundamental geometric parameters for constructing a fusion value of the slip surface geometry. Parameter j specifically included rainfall, three-day cumulative rainfall, slope angle, soil saturation, pore water pressure, and TDR reflectance index. All parameters were collected synchronously according to the sampling cycle and time-stamped. Based on monitoring data acquisition rules, the formats, timestamp accuracy, and spatial coding rules for rainfall, infiltration rate, volumetric water content, pore water pressure, soil temperature, slip zone displacement, strain, and deformation rate were standardized to ensure consistency of monitored quantities in both time and space. Numerical parameters such as rainfall, infiltration rate, volumetric water content, pore pressure, displacement, and deformation rate are normalized by linear mapping. Angular and spatial parameters are mapped to corresponding levels based on geological profiles to ensure that all monitoring data maintain a uniform magnitude. After standardization and normalization, multi-source monitoring data and geological constraint data are stored to construct a landslide monitoring database, providing high-precision data support for subsequent reliable value analysis of slip zone parameters, geometric fusion calculation of slip surface, and determination of instability risk.

[0027] This implementation plan achieves systematic collection, calibration, and fusion of multi-source monitoring data and geological constraint data within the landslide monitoring area, establishing a basic database with temporal consistency and spatial comparability. Various monitoring data, such as rainfall, three-day cumulative rainfall, slope angle, soil saturation, pore water pressure, TDR reflectance index, infiltration rate, volumetric water content, soil temperature, and landslide displacement, are acquired synchronously within the sampling period. Through time series analysis and spatial interpolation correction, strict alignment of data in timestamps, sampling frequencies, and spatial coding is ensured. After standardization and normalization, different types of monitoring parameters are converted to a unified magnitude system, eliminating biases caused by differences in monitoring scale, units, and burial depth. By co-storing monitoring data with geological constraint parameters, a landslide monitoring database that combines geological structural attributes with temporal dynamic characteristics is formed. This database not only provides high-quality data support for the subsequent calculation of reliable values ​​of slip zone parameters, geometric fusion analysis of slip surface, and assessment of time-varying instability driving values, but also significantly improves the completeness, accuracy, and timeliness of data acquisition in the process of landslide hazard identification. In turn, it realizes the unified management and efficient integration of basic landslide monitoring data, laying a reliable technical foundation for dynamic risk assessment.

[0028] Specifically, the steps for analyzing the reliability of the sliding band physical property parameters based on the weighted sensitivity and residual accumulation of multi-source monitoring data are as follows: Obtain the parameters. The parameter inversion residual, parameter Specifically, the parameters include rainfall, three-day cumulative rainfall, slope angle, soil saturation, pore water pressure, and TDR reflectance index. The residuals of each parameter are calculated by comparing measured data with geologically constrained inversion results, used to characterize the deviation of monitoring data from reflecting the slip zone's physical properties. Based on historical records of each monitoring parameter and slip zone response in the landslide monitoring database, correlation coefficients, partial correlation coefficients, and mutual information values ​​are calculated to measure the influence of each monitoring parameter on slip zone stability. Within the sliding time window, the ratio of the change in each monitoring parameter to the deformation change rate is statistically analyzed to obtain the single-parameter sensitivity value. The sensitivity values ​​and correlation indices are weighted, fused, and normalized to ensure the sum of all weight factors is one, thus obtaining the sensitivity weight factor for each parameter. The product of the parameter inversion residual for parameter j and the sensitivity weight factor is calculated to obtain the weighted residual for each parameter in the comprehensive reliability analysis. The sensitivity weight factor is determined based on samples from previous landslide events and slip zone response characteristics, used to reflect the intensity of the influence of each monitoring parameter on changes in slip zone physical properties. By summing the weighted residuals of all parameters, a weighted residual sum is obtained, which quantifies the overall error level under the coupling effect of different monitoring parameters. Subtracting the weighted residual sum from the result yields the reliability value of the sliding band parameters, reflecting the reliability of the sliding band physical parameter inversion results under current monitoring conditions. This process, through a comprehensive analysis of multi-source data weighting, error normalization, and residual superposition, effectively achieves a dynamic quantitative assessment of the reliability of monitoring parameters, providing a basis for subsequent sliding band parameter optimization, correction, and risk classification.

[0029] The specific formula for calculating the reliable value of the sliding band parameter is as follows:

[0030] In the formula, This represents the confidence value of the slip zone parameters, which is used to characterize the overall confidence level of the inversion results of key mechanical and hydrological parameters within the slip zone area; Indicates the first The sensitivity weighting factor of each parameter is used to reflect the relative influence of the parameter on the results of the slip zone stability analysis. Indicates the first The parameter inversion residuals of each parameter are used to characterize the volatility and uncertainty of the estimated values ​​of the parameters obtained in the hierarchical Bayes inversion process. This indicates the total number of parameters for which parameter data was obtained and participated in the calculation.

[0031] Get parameters The parameter inversion residuals and sensitivity weighting factors were calculated. For monitoring point 1, the rainfall-weighted residual was 0.013, the three-day cumulative rainfall-weighted residual was 0.012, the slope angle-weighted residual was 0.012, the soil saturation-weighted residual was 0.011, the pore water pressure-weighted residual was 0.012, and the TDR reflectance index-weighted residual was 0.010. The calculated confidence value for the slip zone parameters was 0.93. For monitoring point 2, the rainfall-weighted residual was 0.052, the three-day cumulative rainfall-weighted residual was 0.045, and the slope angle-weighted residual was... The soil saturation weighted residual is 0.032, the pore water pressure weighted residual is 0.035, the TDR reflectance weighted residual is 0.048, and the TDR reflectance weighted residual is 0.037, resulting in a confidence value of 0.66 for the slip zone parameters. At monitoring point 3, the rainfall weighted residual is 0.090, the three-day cumulative rainfall weighted residual is 0.080, the slope angle weighted residual is 0.085, the soil saturation weighted residual is 0.073, the pore water pressure weighted residual is 0.095, and the TDR reflectance weighted residual is 0.099. The slip zone parameters were calculated as follows: The confidence value is 0.48; for monitoring point 4, the weighted residuals of rainfall, cumulative three-day rainfall, slope angle, soil saturation, pore water pressure, and TDR reflectance are 0.035, 0.045, 0.040, 0.031, 0.050, and 0.049, respectively, resulting in a confidence value of 0.75 for the slip zone parameters; for monitoring point 5, the weighted residuals of rainfall, cumulative three-day rainfall, slope angle, and soil saturation are 0.029, 0.021, 0.019, and 0.019, respectively, respectively. The weighted residuals for slope angle, pore water pressure, and TDR reflectance were 0.031, 0.050, and 0.030, respectively, yielding a confidence value of 0.82 for the slip zone parameter. At monitoring point 6, the weighted residuals for rainfall, three-day cumulative rainfall, slope angle, soil saturation, pore water pressure, and TDR reflectance were 0.100 and 0.153, respectively, yielding a confidence value of 0.39 for the slip zone parameter.

[0032] Table 1. Statistical table of weighted residuals and reliable values ​​of sliding parameters at monitoring points

[0033] like Figure 3 The figure shows the spatial distribution characteristics of the slip zone parameters provided in this application embodiment. According to Table 1 and the monitoring points in the figure, there are significant differences in the weighted residual distribution at different monitoring points, reflecting the spatial inhomogeneity of the mechanical and hydrological parameter inversion results within the region. For example, the residual values ​​at monitoring point 1 are all low, and its slip zone parameter confidence value K... pThe value reached 0.93, indicating that the parameter inversion results in this region are highly stable and have high reliability; the weighted residuals of monitoring points 4 and 5 are both less than 0.05, corresponding to... The residuals for monitoring points 3 and 6 were 0.75 and 0.82 respectively, falling within the medium-to-high confidence range. Conversely, the weighted residuals for monitoring points 3 and 6 were significantly higher than those for other points, especially monitoring point 6, where the maximum residual reached 0.153. The residuals for all points were generally large, leading to... A value of only 0.39 indicates a low confidence level, suggesting significant fluctuations in the inversion of mechanical or hydrological parameters at this location. This reflects potential monitoring anomalies, complex geological structures, or sparse sampling data. Overall, the confidence value of the slip zone parameters is... Spatially, the data exhibits a gradient distribution from high to low. High-confidence points are mostly distributed in areas with dense monitoring data and relatively homogeneous geological structures, while low-confidence points are often located in sections with significant parameter residuals and large data fluctuations. This provides data support and spatial criteria for improving the accuracy of subsequent slip zone identification, verifying key areas, and deploying early warning systems.

[0034] In this implementation plan, a quantitative evaluation of the reliability of landslide physical parameters is achieved through weighted sensitivity and residual accumulation analysis of multi-source monitoring data. By weighting the inversion residuals of different monitoring parameters, the influence intensity of each parameter on landslide stability is comprehensively considered, enabling the evaluation results to reflect the coupling relationship and overall reliability between multi-source monitoring information. This process effectively reduces the interference of single parameter anomalies on the overall judgment, ensuring the stability and consistency of landslide physical property inversion results. This provides a scientific basis for subsequent dynamic assessment, parameter correction, and risk classification, achieving the effect of accurate quantification and dynamic updating of landslide monitoring data reliability.

[0035] Specifically, the specific steps for realizing the dynamic evaluation and optimization correction of the inversion results of the physical property parameters of the sliding zone according to the credible analysis results are as follows: By comparing the credible values of the sliding zone parameters with the trust thresholds T1 and T2 in real time, the dynamic control and optimization adjustment of the spatial inversion quality of the sliding zone parameters are realized. When the credible value of the sliding zone parameter < T1, it indicates that there are deficiencies in the spatial consistency of the monitoring data and the accuracy of physical property inversion, and it is necessary to densify the observation layout and data collection in the monitoring area. At this time, by adding a rain gauge sensor, a vibrating wire piezometer, a soil moisture probe, and a combined unit of an infiltration rate probe and a TDR probe, the sampling density of rainfall, pore water pressure, volumetric water content, and infiltration rate is increased; infiltration rate and volumetric water content probes are arranged at different elevation positions in the vertical direction of the slope surface, and shallow and deep piezometers are configured at corresponding depths to form a vertical monitoring chain coordinated with the inclinometer. During rainfall, the infiltration rate, pore pressure, and deformation rate are synchronously recorded based on the sampling period as the time reference, and the AI spatial interpolation algorithm is used to dynamically interpolate and trend-correct the real-time data, update the monitoring database in real time, and recalculate the credible value of the sliding zone parameter; if the credible value is still lower than the trust threshold after recalculation, further verification tests are implemented on key areas, the value range of the monitoring parameters is revised through infiltration experiments and shear tests, and the continuous field of the physical properties of the sliding zone is updated, and a field review warning report is generated and archived in the landslide monitoring database. When T1 ≤ credible value of the sliding zone parameter < T2, it means that the monitoring quality is within an acceptable range, the existing monitoring points and sampling frequencies are maintained, abnormal data are剔除 and the time series is reconstructed, the sensitivity weight factor in the parameter calculation is readjusted, and the representative values of cohesion, internal friction angle, permeability, and saturation are recalculated in combination with the results of short-term infiltration tests and shear tests to ensure the rationality and stability of the physical property parameters. When the credible value of the sliding zone parameter ≥ T2, it indicates that the inversion result already has a high degree of credibility, the continuous field of the physical properties of the sliding zone and the uncertainty distribution result are released, the drift verification of the monitoring data of pore water pressure, infiltration rate, and volumetric water content is performed weekly, and the credible value of the sliding zone parameter is recalculated after each rainfall process; if the calculation results are continuously lower than the average credible value before rainfall for n times, the early warning mechanism is automatically triggered and a risk prompt report is generated, so as to realize the continuous dynamic perception and hierarchical early warning of the evolution process of landslide hazards. The parameter n is used to limit the number of judgments of the continuous decrease of the credible value of the sliding zone parameter, and its value is 3, which is determined according to the time resolution of historical monitoring data and the rainfall response period.

[0036] It should be noted that the Chinese word "剔除" in the translation is a literal translation. In a more natural English expression, it might be better to use a phrase like "exclude" or "remove". Also, the overall translation aims to be as accurate as possible while following the rules, but there could still be room for further refinement in terms of language fluency and naturalness in some parts.This implementation scheme achieves dynamic evaluation and quality optimization of the slip zone property parameter inversion results, enabling the monitoring system to automatically adjust monitoring density and sampling frequency based on changes in the reliability values ​​of slip zone parameters. By comparing the reliability values ​​of slip zone parameters with the trust threshold in real time, the system can adaptively identify the reliability level of monitoring data. When the reliability is low, it automatically increases the deployment of sensing units and strengthens data acquisition; when the reliability is moderate, it performs parameter recalculation and sequence repair; and when the reliability is high, it publishes the continuous field of physical properties and uncertainty distribution results and conducts drift verification. Simultaneously, by setting a determination mechanism for the number of consecutive calculations (n), it can promptly identify the continuous downward trend of the slip zone reliability value and trigger an early warning response. This process effectively improves the automation and timeliness of landslide monitoring, enhances the stability and reliability of the physical property parameter inversion results, and achieves dynamic quality control and intelligent early warning in the landslide hazard identification process.

[0037] Specifically, the steps for spatial coupling analysis using multi-source deformation and geometric parameter data are as follows: Acquire reliable values ​​of slip zone parameters, monitor displacement, slip surface strike deviation angle, slip surface dip deviation angle, and slip surface depth to construct the basis for joint analysis of the slip surface's spatial geometric characteristics and physical parameters. Reliable values ​​of slip zone parameters reflect the reliability of inversion of key physical and hydrological parameters within the slip zone area; monitored displacement reflects the cumulative deformation amplitude of the slip body during the monitoring period; slip surface strike deviation angle and dip deviation angle characterize the spatial difference between the slip zone's geometry and the theoretical slip surface model; and slip surface depth represents the geological constraint characteristics of the slip zone in the vertical profile. Summing the slip surface strike deviation angle and dip deviation angle and dividing by the slip surface depth yields the normalized ratio of angle difference to depth. This ratio is then multiplied by an angle correction factor to eliminate the influence of angle variations between different geological units. A negative exponent is used for exponential calculation to obtain the spatial geometric error attenuation value, which describes the nonlinear attenuation characteristics of slip surface geometric differences with depth. Subsequently, the reliable values ​​of the sliding strip parameters, the monitored displacement, and the spatial geometric error attenuation value are multiplied together to form the sliding surface geometric fusion value, which comprehensively reflects the coupling relationship between the reliability of the sliding strip properties, the intensity of the deformation response, and the geometric consistency. This can provide key criteria for subsequent optimization of sliding strip connectivity and identification of the main skeleton of the sliding surface.

[0038] The specific formula for calculating the surface geometry blending value is as follows:

[0039] In the formula, This represents the surface geometry fusion value, which is used to characterize the comprehensive evaluation result of the reliability of the sliding band parameters and the degree of matching between the surface geometry and the surface geometry. This indicates the reliability of the sliding zone parameters, reflecting the degree of reliability of key physical parameters such as cohesion, internal friction angle, permeability, and saturation. It represents the monitored displacement, used to describe the cumulative deformation amplitude of the sliding body within the statistical period; This indicates the angle of deviation of the slip surface, in degrees. This indicates the angle of deviation of the sliding surface, in degrees. Indicates the depth of the sliding surface, used to characterize the spatial scale of the sliding strip's position; This represents the angle correction factor, used to adjust the rate at which the difference in direction and tilt affects the decay rate of the fusion value.

[0040] This implementation plan establishes a quantitative coupling relationship between the physical properties of the slip zone and its spatial geometry, enabling a comprehensive assessment of the geometric matching degree of the slip surface. By integrating the reliable values ​​of slip zone parameters, monitored displacement, and spatial geometric error attenuation values, it can simultaneously reflect the stability of physical property inversion within the slip zone region, the response intensity of slip body deformation, and the consistency of the slip surface geometry, thus forming a comprehensive index that can dynamically characterize the internal coupling state of the slip body. This process not only unifies the spatial expression of slip surface geometric features and monitoring data but also provides a quantitative basis for subsequent slip zone connectivity judgment, slip surface main framework identification, and spatial probability volume generation, thereby improving the accuracy and reliability of the fusion analysis of geometric and physical properties in landslide hazard identification.

[0041] Specifically, the steps for dynamically optimizing the sliding surface morphology recognition and connectivity adjustment based on the coupling analysis results are as follows: Real-time comparison of the sliding surface geometric fusion value and fusion threshold is used to determine the rationality and spatial continuity of the matching degree between the sliding surface geometry and physical property parameters, such as... Figure 4This is a flowchart of the optimization process for the geometric fusion and spatial connectivity discrimination of the regional landslide hazard automatic identification system based on geological features of this invention. When the geometric fusion value of the slip surface is less than the fusion threshold, it indicates that there is uncertainty or local anomaly in the geometric structure of the slip surface. At this time, it is necessary to readjust the three-dimensional search strategy: increase the search step size of depth, direction and dip angle within the slope unit range, and set the lower limit of the step size to 0.5m and 0.5° respectively to ensure that the search range is sufficient and the results are smooth; at the same time, the curvature extreme value line and the gully edge line are re-extracted to correct the slip surface projection shape; for areas with concentrated direction difference and dip angle difference, shallow drilling points and inclinometers are set up to supplement deformation data to improve the accuracy of spatial geometric constraints, and the geometric fusion value of the slip surface is recalculated to verify the correction effect. If the geometric fusion value of the slip surface remains below the fusion threshold after m consecutive calculations, a spatial criterion anomaly report is generated, indicating potential structural inconsistencies or data deviations in the area, and recommending on-site verification and supplementary surveys. The value of m is limited to 1 to 5 calculations to ensure the correction process reflects local geological differences while avoiding excessive iteration and computational redundancy. All results are archived in the landslide monitoring database. When the geometric fusion value of the slip surface is greater than or equal to the fusion threshold, it indicates that the slip surface geometry and physical property distribution have reached a stable matching level. At this point, all voxels meeting the conditions are assessed for connectivity based on their spatial adjacency. Adjacent voxels with strike difference and dip difference both less than the geometric difference control value are used as connection nodes, gradually expanding to form spatial connected blocks. The control values ​​for strike difference and dip difference are determined based on the spatial fluctuation characteristics of the slip surface geometric parameters. A statistical analysis of the standard deviations of strike difference and dip difference for all voxels within the monitoring area is performed, and the mean is weighted by 0.5 times the standard deviation as the standard deviation. The constraints are as follows: the strike difference is controlled between 1° and 3°, and the dip difference is controlled between 0.5° and 2°. This ensures both the geometric consistency of the slip surface and avoids excessively restricting the expansion range of connectivity determination, thus generating connected components that possess both geometric continuity and spatial representativeness. Nodes within the connected components are sorted from largest to smallest slip surface geometric fusion value. Starting from the point with the largest slip surface geometric fusion value, the main slip zone skeleton is generated along the path with the strongest directional consistency. Voxels with dip and strike differences less than the constraints are selected on both sides of the skeleton for supplementary splicing to improve the slip zone boundary structure. The final output is a three-dimensional slip surface probability volume, used to characterize the continuity and reliability of the slip surface in spatial distribution, providing a data foundation for subsequent refined identification and stability analysis of landslide hazard areas.

[0042] In this implementation scheme, by continuously comparing the fusion value of the slip surface geometry with the fusion threshold, discontinuous areas or local anomalies in the slip zone geometry can be identified in a timely manner, triggering a 3D search adjustment and on-site data supplementation process to ensure accurate correction of the slip surface projection and spatial morphology. Multi-round iterative calculations and the constraint setting of the step size lower limit achieve a balance between spatial accuracy and computational stability in the slip surface search, improving the consistency and reliability of slip zone geometry identification. When the slip surface geometry fusion value reaches the fusion threshold, spatial connectivity analysis and slip zone skeleton reconstruction are performed, extracting the directional consistency structure of the main slip zone and its neighboring voxels, and outputting a high-confidence 3D slip surface probability volume. This achieves dynamic reconstruction and reliable representation of slip surface geometry information from discrete points to a continuous spatial volume, effectively improving the accuracy and stability of slip zone spatial determination in landslide hazard identification.

[0043] Specifically, the steps for quantifying the slip zone instability trend through dynamic instability analysis using fused data of deformation rate and slip surface geometry are as follows: First, obtain the fused values ​​of deformation rate and slip surface geometry to comprehensively reflect the coordinated change pattern of the slip zone's deformation characteristics and spatial geometric characteristics over time. Second, obtain the geometric amplification factor by performing joint regression and time-decay weighted analysis on the fused values ​​of historical deformation rate and slip surface geometry. Third, multiply the fused values ​​of slip surface geometry by the geometric amplification factor and use the result as an exponent to calculate the geometric influence term, reflecting the nonlinear amplification effect of slip surface geometric morphology changes on the instability process. Finally, multiply the deformation rate by the geometric influence term to obtain the risk-driving fundamental quantity, which measures the instability-driving effect of the slip zone caused by the combined effects of geometric disturbance and deformation acceleration during its temporal evolution. Finally, by adding one to the risk-driven basis and taking the natural logarithm, the time-varying instability driving value is obtained. This value can serve as a quantitative indicator of the overall instability trend of the sliding strip in the time dimension. It can realize the dynamic coupling analysis of the deformation acceleration characteristics and geometric response effect of the sliding strip, and provide a highly timely and accurate driving parameter basis for subsequent risk assessment and stability verification.

[0044] The specific formula for calculating the variable instability driving value is as follows:

[0045] In the formula, This represents the time-varying instability driving value, used to characterize the overall instability trend of the sliding strip in the time dimension; The deformation rate is used to reflect the motion intensity and acceleration trend of the sliding body at the monitoring time. This represents the surface geometry fusion value, used to characterize the degree of matching between the surface geometry and physical property parameters; This represents the geometric amplification factor, used to control the nonlinear amplification effect of the sliding surface geometric fusion value on time-varying instability drive.

[0046] In this implementation scheme, the deformation rate of the slip zone and the geometric fusion value of the slip surface are dynamically coupled and analyzed to extract their coordinated change patterns in the temporal and spatial dimensions, quantifying the accelerated deformation trend of the slip zone under geometric disturbance. By introducing a geometric amplification factor, the amplification effect of the slip surface morphology on the instability evolution is expressed nonlinearly, thereby forming a time-varying instability driving value that reflects the comprehensive instability trend of the slip zone. This achieves continuous characterization and quantitative assessment of the dynamic evolution process of the landslide, providing key criteria for subsequent risk classification and early warning response.

[0047] Specifically, the steps for performing risk response and stability verification based on the instability analysis results are as follows: The stability of the slip zone over time is dynamically determined by comparing the time-varying instability driving value with the stability threshold in real time. When the time-varying instability driving value is less than the stability threshold, the existing monitoring deployment and sampling frequency are maintained. Consistency checks are performed on the multi-source monitoring data from inclinometers, pore manometers, and rain gauges. Drift and noise data are removed. Moving average and smoothing algorithms are used to optimize the monitoring curves over time, forming a continuous and traceable deformation trend curve, thus enabling long-term tracking of the slip zone's stable phase. When the time-varying instability driving value is greater than or equal to the stability threshold, a short-term dynamic response mechanism is activated. Trend fitting analysis is performed on the latest monitoring data to extract key characteristic quantities such as the pore water pressure growth rate and deformation acceleration, and to identify potential instability acceleration stages. In areas where the time-varying instability driving value continues to rise, auxiliary facilities such as drainage pipes and strain gauges are deployed. The sensitivity of the monitoring response is verified through local load reduction tests, forming an instability driving response rate curve describing the stress-strain coupling relationship of the slip zone. If the time-varying instability driving value calculated n times consecutively is less than the stability threshold, it indicates that the slip zone has recovered to a relatively stable state. At this time, the regular sampling interval is restored, and the corrected time-series driving data and verification results are archived in the landslide monitoring database for subsequent trend evolution and risk tracking analysis.

[0048] In this implementation scheme, automatic identification and phased response control of the slip zone's stability state are achieved through dynamic comparison of time-varying instability driving values ​​and stability thresholds. During the stable phase, routine monitoring and data smoothing are maintained to ensure the continuity and traceability of the deformation trend curve. During the instability phase, short-term trend analysis and local response tests are automatically triggered to accurately capture the rise in pore pressure and the accelerated deformation process. During the slip zone recovery phase, the sampling strategy is adjusted in a timely manner and monitoring data is updated, achieving closed-loop dynamic management of the entire process from stability to instability to re-stabilization. This process effectively improves the sensitivity and response accuracy of landslide time-varying monitoring, thereby enabling intelligent identification of the slip zone's stability state and continuous tracking of its evolution.

[0049] Specifically, the steps for comprehensively determining the stability of monitoring units by integrating three-dimensional slip surface probability volume and spatiotemporal evolution indicators, and implementing hierarchical monitoring and early warning based on the determination results, are as follows: Combining the spatial distribution characteristic map of slip zone parameters, three-dimensional slip surface probability volume, time-varying instability driving value, slip surface geometric fusion value, and slip zone parameter reliability value, a multi-dimensional comprehensive determination of the stability of monitoring points is performed, and dynamic risk level classification is implemented. During the calculation process, the spatial heterogeneity and temporal evolution characteristics of the internal parameters of the slip zone are comprehensively considered, and quantitative identification of the stability of monitoring units is achieved through multi-source data fusion. When the time-varying instability driving value is less than the stability threshold, it indicates that the area is in a stable or slightly deformed state, and it is classified as a general monitoring area. Key monitoring data such as deformation rate, pore water pressure, and rainfall are regularly summarized to update the risk assessment file, achieving dynamic maintenance of basic monitoring. When the time-varying instability driving value is greater than or equal to the stability threshold, and the sliding surface geometric fusion value is greater than the fusion threshold, the system automatically determines that the area has a significant instability trend, marks it as a key early warning area, pushes a risk warning report in real time, and coordinates with on-site investigation, drainage, and reinforcement engineering measures to form a closed-loop prevention and control process of "data identification - response intervention - safety verification". At the same time, based on the spatial variation law of the confidence value in the spatial distribution characteristic map of the slip zone parameters, the system implements intensified monitoring and increased sampling frequency operations for areas with low confidence values, generates time-series change reports reflecting the risk evolution trend, ensures the targeting and efficiency of the monitoring network, and ultimately achieves graded control of landslide risk and closed-loop archiving of dynamic early warning information to the landslide monitoring database.

[0050] This implementation plan integrates multi-source information, including spatial distribution feature maps of slip zone parameters, three-dimensional slip surface probability volumes, time-varying instability driving values, slip surface geometric fusion values, and reliable slip zone parameter values, to achieve comprehensive assessment and risk classification early warning of the stability status of landslide monitoring units. It can automatically identify stable and potentially unstable areas based on changes in monitoring indicators, perform periodic risk file updates for stable areas, and implement focused monitoring and on-site intervention for abnormal areas. Simultaneously, based on the spatial distribution patterns of reliable slip zone parameter values, it automatically intensifies monitoring and increases sampling frequency in areas with low reliability, thus forming a complete closed loop of data collection, risk identification, early warning linkage, and information archiving. This step effectively improves the spatial accuracy and dynamic response capability of landslide hazard identification, thereby achieving hierarchical control and intelligent early warning of regional landslide risks.

[0051] Specifically, this embodiment provides an automatic identification system for regional landslide hazards based on geological features. Applying the aforementioned automatic identification method for regional landslide hazards based on geological features, it includes a landslide data acquisition and preprocessing module, a physical parameter inversion and slip zone feature estimation module, a slip surface geometry inference and connectivity optimization module, a dynamic evolution and uncertainty update module, and a risk level assessment and early warning module. The slip surface geometry inference and connectivity optimization module and the risk level assessment and early warning module work together. The landslide data acquisition and preprocessing module acquires multi-source monitoring data and geological constraint data, performs cleaning, calibration, and format unification processing, and constructs a structured landslide monitoring database to achieve standardization and consistency of data in the spatiotemporal dimensions. The physical parameter inversion and slip zone feature estimation module quantitatively assesses the reliability of parameters such as rainfall, slope angle, pore water pressure, and TDR reflectance index based on the weighted sensitivity and residual accumulation of multi-source monitoring data. It also performs real-time evaluation and optimization correction of the physical property inversion results based on the dynamic changes in the reliability values ​​of slip zone parameters to ensure the reliability of the slip zone physical property field distribution. The slip surface geometry inference and connectivity optimization module fuses reliable values ​​of slip zone parameters, monitored displacement, and slip surface geometric feature data. Through spatial coupling analysis, it calculates the fused slip surface geometry value, adjusts the 3D search strategy and step size, optimizes slip zone connectivity identification, and outputs a high-precision 3D slip surface probability volume. The dynamic evolution and uncertainty update module calculates time-varying instability driving values ​​based on deformation rate and slip surface geometry fused data, dynamically characterizes the slip zone stability change trend, and performs experimental verification and data drift correction, enabling continuous tracking and uncertainty updates of the landslide evolution process. The risk level assessment and early warning module integrates the spatial distribution feature map of slip zone parameters, the 3D slip surface probability volume, and spatiotemporal evolution indicators to perform multi-level judgments on the stability state of monitoring units. It divides the area into general monitoring areas and key early warning areas, links on-site investigation and reinforcement measures, generates risk assessment and early warning reports, and achieves hierarchical management and intelligent early warning of regional landslide hazards.

[0052] This implementation plan achieves intelligent and refined management of the entire landslide hazard identification process by constructing five core modules: landslide data acquisition and preprocessing, physical parameter inversion and slip zone characteristic estimation, slip surface geometry inference and connectivity optimization, dynamic evolution and uncertainty update, and risk level assessment and early warning. Based on multi-source monitoring data, it enables data standardization and structured processing, dynamically assesses the reliability of slip zone physical parameters, accurately identifies slip surface geometry and connectivity characteristics, tracks the instability evolution trend of the slip zone in real time, and combines risk indicators to achieve hierarchical early warning responses. Through the collaborative operation of multiple modules, this step effectively improves the reliability and inversion accuracy of landslide monitoring data, enhances the spatial analysis and temporal response capabilities of landslide hazard identification, and thus achieves the effects of dynamic identification, accurate assessment, and proactive early warning of regional landslide risks.

[0053] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0054] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. An automatic identification method for regional landslide hazards based on geological characteristics, characterized in that, Includes the following steps: S1. Acquire multi-source monitoring data and geological constraint data and preprocess them to construct a landslide monitoring database; S2, based on the weighted sensitivity and residual accumulation of multi-source monitoring data, analyzes the reliability of the sliding belt physical parameters, and realizes dynamic evaluation and optimization correction of the sliding belt physical parameter inversion results based on the reliability analysis results; S3 performs spatial coupling analysis using multi-source deformation and geometric parameter data, and dynamically optimizes the sliding strip morphology recognition and connectivity adjustment operations based on the coupling analysis results. S4 performs dynamic instability analysis by fusing deformation rate and surface geometry data, quantifies the instability trend of the sliding zone, and performs risk response and stability verification operations based on the instability analysis results. S5 integrates three-dimensional slip surface probability volume and spatiotemporal evolution indicators to comprehensively determine the stable state of the monitoring unit, and implements hierarchical monitoring and early warning based on the determination results.

2. The automatic identification method for regional landslide hazards based on geological characteristics according to claim 1, characterized in that: The specific steps for acquiring multi-source monitoring data and geological constraint data, performing preprocessing, and constructing a landslide monitoring database are as follows: The process involves collecting multi-source monitoring data and geological constraint data required for landslide hazard identification. Multi-source monitoring data includes rainfall, three-day cumulative rainfall, slope angle, soil saturation, pore water pressure, TDR reflectance index, infiltration rate, volumetric water content, soil temperature, and landslide displacement. Geological constraint data includes shallow drilling results, landslide surface depth, landslide surface strike deviation angle, and landslide surface dip deviation angle. The time interval between the completion of one synchronous acquisition of multi-source monitoring data by the monitoring equipment and the completion of the next similar data acquisition is used to characterize the smallest synchronous sampling unit of landslide monitoring data in the time dimension, and is denoted as the sampling period. Acquire rainfall and three-day cumulative rainfall data, continuously record the precipitation process using a rain gauge sensor, and obtain rainfall intensity, cumulative rainfall and sliding three-day cumulative value through time series statistical analysis; Acquire slope angle data and calculate the slope angle distribution based on topographic mapping and digital elevation model; Data on infiltration rate and volumetric water content were acquired. The soil water content changes were monitored synchronously using a combination unit of infiltration rate probe and TDR probe. The infiltration rate and volumetric water content were obtained through time series analysis. The monitoring sequence was compared and analyzed with rainfall and pore pressure changes to obtain the parameter inversion residual. Pore ​​water pressure data were acquired, and pore pressure response was synchronously collected at different burial depths using a vibrating wire pore pressure gauge. The pore water pressure variation characteristics and parameter inversion residuals were obtained through time-series fluctuation analysis coupled with infiltration rate data. TDR reflectance index and soil saturation data were obtained. Based on TDR waveform feature analysis and volume water content calculation results, soil saturation was derived and the reflectance signal change sequence was recorded. Soil temperature data is acquired by using thermistor probes deployed at multiple depths to collect temperature changes and obtaining soil temperature distribution through trend analysis. The sliding body displacement data is obtained by collecting displacement time series using inclinometers and GNSS displacement monitoring devices. The monitored displacement is obtained through filtering and time alignment processing, and the displacement series is exported and analyzed to obtain the deformation rate. Obtain strain data, collect the micro-strain changes inside the landslide body using strain gauges, and obtain the stress concentration distribution within the slip zone through strain gradient analysis; Obtain the geometric information of the slip surface. Combine the results of shallow borehole disclosures and geological survey data, obtain the buried depth of the slip surface through spatial interpolation analysis, obtain the strike deviation angle of the slip surface through directional statistical analysis of geological survey and shallow borehole disclosure data, and obtain the dip deviation angle of the slip surface through fitting analysis of geological profile survey lines and dip measurement data; parameter Specifically, this includes rainfall, three-day cumulative rainfall, slope angle, soil saturation, pore water pressure, and TDR reflectance index. Based on the monitoring data collection rules, the formats, timestamp accuracy, and spatial coding rules for rainfall, infiltration rate, volumetric water content, pore water pressure, soil temperature, landslide displacement, strain, and deformation rate are standardized to ensure consistency of monitored data in both time and space. Linear mapping normalization is performed on numerical parameters such as rainfall, infiltration rate, volumetric water content, pore pressure, displacement, and deformation rate. Angular and spatial parameters are mapped to corresponding levels based on geological profiles, ensuring all monitoring data maintain a uniform magnitude. After standardization and normalization, multi-source monitoring data and geological constraint data are stored to construct a landslide monitoring database.

3. The automatic identification method for regional landslide hazards based on geological characteristics according to claim 1, characterized in that: The specific steps for analyzing the reliability of the physical property parameters of the slip zone based on the weighted sensitivity and residual accumulation of multi-source monitoring data are as follows: Get parameters The parameter inversion residual, parameter Specifically, this includes rainfall, three-day cumulative rainfall, slope angle, soil saturation, pore water pressure, and TDR reflectance index; Calculate the product of the parameter inversion residual amount and the sensitivity weight factor of parameter j respectively to obtain the weighted residual amount. Sum and accumulate the weighted residual amounts of each parameter to obtain the weighted residual accumulation value. Calculate one minus the weighted residual accumulation value to obtain the reliability value of the slip zone parameters.

4. The automatic identification method for regional landslide hazards based on geological features according to claim 1, characterized in that: The specific steps for realizing the dynamic evaluation and optimization correction of the inversion results of the physical property parameters of the slip zone according to the reliability analysis results are as follows: By comparing the reliability value of the slip zone parameters with the trust thresholds T1 and T2 in real time, the dynamic control and optimization of the spatial inversion quality of the slip zone parameters can be achieved. When the reliability value of the slip zone parameters < T1, increase the number of combined units of rain gauges, vibrating wire piezometers, soil moisture probes, infiltration rate probes, and TDR probes to improve the sampling density of rainfall, pore water pressure, volumetric water content, and infiltration rate monitoring data: arrange infiltration rate and volumetric water content probes at different elevation positions along the vertical direction of the slope, and configure shallow and deep piezometers at the corresponding positions, and cooperate with inclinometers to form a vertical monitoring chain; synchronously record the infiltration rate, pore pressure, and deformation rate at the sampling period during rainfall,联动AI空间插值算法实时更新监测数据并重新计算滑带参数可信值,若滑带参数可信值仍小于信任阈值,则对重点区域实施核验试验,开展渗透实验与剪切测试,修订监测数据取值范围并更新物性连续场;生成现场复核预警报告并将归档至滑坡滑面监测数据库;当T1≤滑带参数可信值<T2时,维持现有监测点位与采样频次,剔除异常数据并重建时间序列,在参数计算中参数的敏感度权重因子,重新计算黏聚力、内摩擦角、渗透率与饱和度代表值,进行短期渗透试验与剪切试验以验证修正结果; When the reliability value of the slip zone parameters ≥ T2, release the continuous field and uncertainty distribution results of the physical properties of the slip zone, perform a drift check on the monitoring data of pore water pressure, infiltration rate, and volumetric water content once a week, recalculate the reliability value of the slip zone parameters after each rainfall process ends, and issue a warning report when the recalculated reliability value of the slip zone parameters is continuously lower than the average reliability value of the slip zone parameters before rainfall for n times.

5. The method for automatic identification of regional landslide hazards based on geological characteristics according to claim 1, characterized in that: The specific steps for performing spatial coupling analysis through multi-source deformation and geometric parameter data are as follows: Obtain the reliability value of the slip zone parameters, the monitored displacement amount, the strike deviation angle of the slip surface, the dip deviation angle of the slip surface, and the buried depth of the slip surface; The spatial geometric error attenuation value is obtained by summing the slip surface deviation angle and the dip angle deviation angle, dividing by the slip surface burial depth, multiplying by the angle correction factor, taking the negative number as the exponent, and performing exponential calculation. The geometric fusion value of the sliding surface is obtained by multiplying the reliable value of the sliding surface parameters, the monitored displacement, and the spatial geometric error attenuation value.

6. The automatic identification method for regional landslide hazards based on geological characteristics according to claim 1, characterized in that: The specific steps for dynamically optimizing the sliding band morphology recognition and connectivity adjustment operations based on the coupling analysis results are as follows: The sliding surface geometric fusion value is compared with the fusion threshold in real time. When the sliding surface geometric fusion value is less than the fusion threshold, the three-dimensional search strategy is readjusted: the depth, strike and dip angle search step size is increased within the slope unit, and the curvature extreme value line and gully edge line are re-extracted to correct the sliding surface projection; shallow drilling points and inclinometers are set up in the areas where the strike difference and dip angle difference are concentrated, deformation data is collected and the sliding surface geometric fusion value is recalculated. If the sliding surface geometric fusion value is still less than the fusion threshold after m times, a spatial criterion anomaly report is generated and on-site verification is recommended, and the data is archived to the landslide monitoring database. When the sliding surface geometry fusion value is greater than or equal to the fusion threshold, all voxels that meet the conditions are connected according to their spatial adjacency. Adjacent voxels whose direction difference and tilt difference are both less than the geometric difference control value are used as connecting nodes, and connected blocks are formed in sequence. The nodes in the connected blocks are sorted from largest to smallest according to the sliding surface geometry fusion value. Starting with the largest sliding surface geometry fusion value, the main sliding zone skeleton is generated along the path with the strongest directional consistency. Voxels whose tilt difference and direction difference are both less than the limit value are selected on both sides of the skeleton for supplementary splicing. The output is a three-dimensional sliding surface probability volume.

7. The automatic identification method for regional landslide hazards based on geological characteristics according to claim 1, characterized in that: The specific steps for performing dynamic instability analysis and quantifying the instability trend of the sliding band by fusing deformation rate and sliding surface geometry data are as follows: Obtain deformation rate and surface geometry fusion values; Multiply the geometric fusion value of the slip surface by the geometric amplification factor and use it as an exponent to calculate the geometric influence term. Then multiply the deformation rate by the geometric influence term to obtain the risk driving basis. Add one to the obtained risk driving basis and take the natural logarithm to obtain the time-varying instability driving value.

8. The automatic identification method for regional landslide hazards based on geological characteristics according to claim 1, characterized in that: The specific steps for performing risk response and stability verification operations based on the instability analysis results are as follows: By comparing the time-varying instability driving value with the stability threshold in real time, when the time-varying instability driving value is less than the stability threshold, the monitoring deployment and sampling frequency are maintained, the consistency of the monitoring data of inclinometers, pore pressure gauges and rain gauges is checked, the data records are updated and smoothed to form a continuous time series curve, thereby realizing the tracking of deformation trends. When the time-varying instability driving value is greater than or equal to the stability threshold, short-term trend fitting is performed on the monitoring data to extract the pore pressure growth rate and deformation acceleration. Drainage pipes and strain gauges were installed in the area where the time-varying instability driving value increased, and a small-scale load reduction test was carried out to verify the monitoring response and form a variable instability driving response rate curve. When the time-varying instability driving value calculated for n consecutive times is less than the stability threshold, the normal sampling interval was restored, and the time-series driving correction data was recorded and archived to the landslide monitoring database.

9. The method for automatic identification of regional landslide hazards based on geological characteristics according to claim 1, characterized in that: The specific steps for comprehensively determining the stable state of the monitoring unit by integrating the three-dimensional slip surface probability volume and spatiotemporal evolution index, and implementing hierarchical monitoring and early warning based on the determination results, are as follows: By combining the spatial distribution feature map of the sliding band parameters, the three-dimensional sliding surface probability volume, the time-varying instability driving value, the sliding surface geometric fusion value, and the reliable value of the sliding band parameters, the stability status of the monitoring point is comprehensively judged and the risk level is classified. When the time-varying instability driving value is less than the stability threshold, the area is designated as a general monitoring area, and data on deformation rate, pore water pressure and rainfall are regularly summarized to update the risk assessment file; When the time-varying instability driving value is greater than or equal to the stability threshold, and the sliding surface geometric fusion value is greater than the fusion threshold, the area is marked as a key early warning area, a risk early warning report is pushed out, and on-site investigation, drainage and reinforcement engineering measures are coordinated; based on the confidence distribution pattern shown in the spatial distribution characteristic map of the slip zone parameters, the low confidence value area is subject to zoned intensified monitoring and increased sampling frequency operation, a risk time series change report is generated and archived in the landslide monitoring database.

10. An automatic identification system for regional landslide hazards based on geological features, employing the automatic identification method for regional landslide hazards based on geological features as described in any one of claims 1-9, characterized in that... ,include: The landslide data acquisition and preprocessing module is used to acquire multi-source monitoring data and geological constraint data, perform preprocessing, and construct a landslide monitoring database. The physical parameter inversion and slip zone characteristic estimation module is used to analyze the reliability of slip zone physical parameters based on the weighted sensitivity and residual accumulation of multi-source monitoring data, and to realize the dynamic evaluation and optimization correction of slip zone physical parameter inversion results based on the reliability analysis results. The sliding surface geometry inference and connectivity optimization module is used to perform spatial coupling analysis through multi-source deformation and geometric parameter data, and dynamically optimize the sliding surface morphology recognition and connectivity adjustment operations based on the coupling analysis results. The dynamic evolution and uncertainty update module is used to perform dynamic instability analysis by fusing deformation rate and sliding surface geometry data, quantify the instability trend of the sliding zone, and perform risk response and stability verification operations based on the instability analysis results. The risk level assessment and early warning module is used to integrate three-dimensional slip surface probability volume and spatiotemporal evolution indicators to comprehensively determine the stable state of the monitoring unit, and to implement hierarchical monitoring and early warning based on the determination results.

Images