Rapid early warning method for regional geological disasters

Abstract

The invention discloses a quick early warning method for regional geological disasters, particularly relates to the field of early warning of geological disasters, and is used for solving the problems of early warning lag and inaccurate positioning caused by dependence on a single monitoring index in the prior art. According to the method, a multi-dimensional monitoring data stream with time-space continuity is constructed by acquiring a slope body deep inclination angle, earth sound event count and earth surface micro-displacement data of a geological early warning area; generating a regional deformation synchronization index based on phase-space reconstruction and correlation dimension analysis, and calculating a system entropy value through density clustering and displacement fluctuation analysis; performing real-time identification on the regional critical state by adopting a critical state judgment model of double-threshold grading; and finally, combining real-time rainfall intensity data to carry out spatial superposition to divide into multiple levels of warning subareas, and outputting corresponding early warning signals. According to the method, pre-disaster megasses can be captured from the aspect of internal structure instability, the accuracy and foresight of regional geological disaster pre-warning are improved, and regional public safety is guaranteed.

Description

Technical Field

[0001] This invention relates to the field of geological disaster early warning technology, and more specifically, to a method for rapid early warning of regional geological disasters. Background Technology

[0002] Currently, in complex mountainous and hilly areas, regional geological disasters such as landslides and collapses are characterized by their suddenness, destructiveness, and complex evolution mechanisms, posing a significant threat to regional public safety and infrastructure stability. Particularly in areas with frequent heavy rainfall, earthquakes, or significant geological tectonic activity, potential geological disaster areas are often located in sections of slopes with loose deep structures and weakened surface support capacity. Furthermore, the dense distribution of residential areas, highways, power transmission lines, and major engineering facilities such as water conservancy projects further increases the risk of disaster losses.

[0003] Existing regional geological disaster early warning technologies typically rely on single monitoring indicators or statistical patterns. However, in the actual disaster incubation process, internal structural damage often precedes significant surface deformation, leading to lag and high false alarm rates in monitoring methods that depend solely on a single indicator. Furthermore, existing early warning methods generally lack sufficient response to spatial heterogeneity, making it difficult to accurately locate critical instability zones and implement tiered measures, resulting in insufficient operability and precision in the early warning results. Especially in the context of landslides triggered by heavy rainfall, there are significant spatial differences between rainfall infiltration and overall slope structural weakening, requiring a linked analysis of rainfall dynamics and the state of potential instability areas to obtain more accurate early warning zoning results.

[0004] Therefore, there is an urgent need for a rapid regional geological disaster early warning method that can integrate multi-source monitoring data, comprehensively measure the evolution of the internal structure of the region, and combine the impact of rainfall to achieve rapid, accurate, and graded early warning, so as to improve disaster prevention and control capabilities and proactive risk management. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a method for rapid early warning of regional geological disasters to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A method for rapid early warning of regional geological disasters includes the following steps: S1. Collect data on the deep slope tilt angle, ground sound event count, and surface micro-displacement of the geological early warning area at a set frequency to construct a multi-dimensional monitoring data stream; S2. Reconstruct the phase space of the deep slope tilt angle sequence, calculate the deformation trajectory correlation dimension between adjacent monitoring units, and generate the regional deformation synchronization index. S3. Based on the spatiotemporal distribution of ground acoustic event counts, density clustering algorithm is used to identify acoustic emission clusters, and the system entropy is calculated by combining the surface micro-displacement vector field. S4. Input the deformation synchronization index and system entropy value into the critical state determination model based on dual threshold classification to identify the critical state of the geological early warning area; S5. Based on the geological warning area entering the critical state, extract real-time rainfall intensity data, divide the area into multiple warning zones through spatial overlay, and generate corresponding warning signals.

[0007] In a preferred embodiment, step S1, which involves collecting data on the deep slope inclination angle, ground sound event counts, and surface micro-displacement of the geological early warning area at a set frequency to construct a multi-dimensional monitoring data stream, specifically includes: In the geological early warning area, a sensor array centered on the main control node is deployed, with each sensor covering one monitoring unit; The geological early warning area is a geologically active area within the set monitoring range, and the monitoring unit is a pre-marked potential landslide body and a large unstable rock mass within the geological early warning area; The master control node sends synchronous acquisition commands to each monitoring unit at a set frequency, and the monitoring unit collects data on the deep slope tilt angle and surface micro-displacement. Continuously monitor the background vibration environment in the geological early warning area, and trigger the ground sound event count when the vibration signal amplitude exceeds the environmental noise baseline; The collected data is aggregated to the main control node, and the multi-source monitoring data is aligned according to the timestamp. The aligned data is integrated using a sliding time window mechanism to construct a multi-dimensional monitoring data stream with spatiotemporal continuity.

[0008] In a preferred embodiment, step S2, which involves reconstructing the phase space of the deep slope tilt angle sequence, calculating the deformation trajectory correlation dimension between adjacent monitoring units, and generating a regional deformation synchronization index, specifically includes: The mutual information function is used to set the time delay parameter for phase space reconstruction, and the false nearest neighbor method is combined to calculate the optimal embedding dimension, so that the tilt angle time sequence of each monitoring unit is mapped to the deformation trajectory in the high-dimensional phase space. The dynamic time warping algorithm is used to measure the morphological matching degree between the deformation trajectories of adjacent monitoring units, and the correlation dimension value of each pair of adjacent units is calculated based on the trajectory matching results. The arithmetic mean of the correlation dimension values ​​of all adjacent monitoring unit pairs is taken, and this arithmetic mean is directly used as the regional deformation synchronization index.

[0009] In a preferred embodiment, step S3, which involves identifying acoustic emission clusters using a density clustering algorithm based on the spatiotemporal distribution of ground acoustic event counts, and calculating the system entropy value in conjunction with the surface micro-displacement vector field, specifically includes: A spatiotemporal distribution dataset is constructed based on the time of the triggered ground sound event count and the coordinates of the monitoring unit. A density-based clustering algorithm is used to process the spatiotemporal distribution dataset of ground acoustic event counts to identify acoustic emission clusters where the event density per unit spatiotemporal volume exceeds the average level of the population. Within the spatial range corresponding to each acoustic emission cluster area, surface micro-displacement data of all monitoring units in the area are extracted and converted into micro-displacement vectors. The dispersion of each vector direction and the fluctuation intensity of amplitude change are calculated. The local disorder of the cluster region is calculated based on the degree of dispersion and the intensity of the fluctuation. The system entropy is obtained by weighted averaging of the local disorder of all acoustic emission cluster regions.

[0010] In a preferred embodiment, step S4, inputting the deformation synchronization index and system entropy value into a critical state determination model based on a dual-threshold classification to identify the critical state of the geological early warning area specifically includes: A two-dimensional critical state determination plane is constructed using the deformation synchronization index and the system entropy value as coordinates; An elliptical safety boundary is defined within the critical state determination plane, with the historical stable state data cluster as its core, and a sector-shaped critical determination region concentric with the safety boundary is set outside the safety boundary. When real-time data points simultaneously exceed the safety boundary and fall into the critical judgment zone, the geological warning area is determined to have entered a critical state.

[0011] In a preferred embodiment, the range of the sector-shaped critical determination region is defined by both the radial boundary and the azimuth angle, specifically including: The radial boundary starts at the outer edge of the elliptical safety boundary, and its corresponding radial distance is set based on the historical extreme value fluctuation range of the regional deformation synchronization index and the system entropy value calculated by the monitoring unit based on the historical monitoring data of the monitoring unit. The azimuth angle is calculated by extracting the coordinates of the extreme points of the regional deformation synchronization index and regional instability recorded when historical geological disaster events occurred, and then calculating the angle between each extreme point and the line connecting the origin of the judgment plane coordinate system. The included angle is calculated by weighted average based on the scale and level of historical disaster events. The resulting weighted average value is set as the reference azimuth angle of the critical judgment zone of the sector, and the tolerance range is set as the sector unfolded surface of 45 degrees on each side of the reference azimuth angle axis.

[0012] In a preferred embodiment, step S5, which involves extracting real-time rainfall intensity data based on the geological warning area entering a critical state, dividing it into multi-level warning zones through spatial overlay, and generating corresponding warning signals, specifically includes: Acquire rainfall intensity data for the geological warning area, and simultaneously extract the spatial range of the area that has been determined to be in a critical state; By spatially overlaying the boundary of the critical state area with the real-time rainfall intensity contour lines, the actual early warning area within the rainfall impact range is determined, and the geological early warning area is divided into multi-level warning zones. Based on the warning zoning results, a corresponding vector warning zoning map is generated, and this warning map is integrated with the zoning warning level to form a warning signal.

[0013] The technical effects and advantages of the regional geological disaster rapid early warning method of the present invention are as follows: By integrating multi-source monitoring data such as deep slope tilt angle, ground acoustic event counts, and surface micro-displacements, a three-dimensional capture of the regional geological hazard incubation process is achieved, exhibiting higher sensitivity and reliability compared to traditional methods relying on single indicators. The regional deformation synchronization index, obtained through phase space reconstruction and deformation trajectory correlation analysis, can reflect the trend of coordinated degradation of internal structures before significant surface deformation occurs, significantly improving the ability to identify precursors of critical instability. The system entropy value calculated based on ground acoustic event clusters and surface displacement fluctuation intensity can quantify the degree of disorder in local energy release, providing real-time assessment of the mechanical instability process of geological hazards. The introduction of a dual-threshold graded critical state determination model enables accurate identification of spatially unevenly distributed potential instability areas, avoiding error propagation caused by overall regional averaging. Furthermore, real-time rainfall intensity is incorporated into the determination system, forming multi-level warning zones through spatial overlay, making the early warning results not only forward-looking in time but also spatially targeted and operable.

[0014] This invention improves the overall response speed and refined management capabilities of early warning systems, and can provide reliable technical support for regional security and disaster prevention and mitigation in major projects, thus having significant engineering application value. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of a rapid early warning method for regional geological disasters according to the present invention. Detailed Implementation

[0016] 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 of ordinary skill in the art without creative effort are within the scope of protection of the present invention. Example 1

[0017] Figure 1 This invention provides a rapid early warning method for regional geological disasters, which includes the following steps: S1. Collect data on the deep slope tilt angle, ground sound event count, and surface micro-displacement of the geological early warning area at a set frequency to construct a multi-dimensional monitoring data stream; S2. Reconstruct the phase space of the deep slope tilt angle sequence, calculate the deformation trajectory correlation dimension between adjacent monitoring units, and generate the regional deformation synchronization index. S3. Based on the spatiotemporal distribution of ground acoustic event counts, density clustering algorithm is used to identify acoustic emission clusters, and the system entropy is calculated by combining the surface micro-displacement vector field. S4. Input the deformation synchronization index and system entropy value into the critical state determination model based on dual threshold classification to identify the critical state of the geological early warning area; S5. Based on the geological warning area entering the critical state, extract real-time rainfall intensity data, divide the area into multiple warning zones through spatial overlay, and generate corresponding warning signals.

[0018] In S1, the slope inclination angle, ground sound event count, and surface micro-displacement data of the geological early warning area are collected at a set frequency to construct a multi-dimensional monitoring data stream.

[0019] A sensor array centered on a main control node is deployed within the geological early warning area. This area refers to a region with significant potential for crustal activity or surface deformation, defined by historical geological disaster records, fault structure distribution, stratigraphic structure, and slope conditions. To ensure continuous monitoring coverage and clear control of key areas, the geological early warning area is divided into several monitoring units. Each unit is a potential landslide boundary zone and area containing large unstable rock masses, pre-marked through preliminary geological reconnaissance, remote sensing image interpretation, and structural analysis. Examples include tension crack convergence zones at the toe of slopes, dip-side development sections of rock strata, and areas with interconnected rock joints. For each monitoring unit, an integrated monitoring device combining deep tilt sensors and surface micro-displacement sensors is installed in a stable area of ​​the surface rock mass. The installation depth reaches at least to the lower edge of the potential landslide zone or the root of the unstable rock mass. For example, the tilt sensors are buried 6 to 15 meters deep, configured according to the depth of the weak layers determined by on-site investigation. The surface displacement monitoring device is fixed to a solid base point on the surface of the soil and rock mass. Sensors from multiple monitoring units are connected to the main control node in a star-shaped structure.

[0020] The master control node sets a fixed acquisition frequency based on the requirement for continuous temporal monitoring during the geological disaster formation process, sending a synchronization acquisition command to all monitoring units every minute to ensure that the deep tilt angle and surface micro-displacement data acquired by each monitoring unit have a consistent time reference. Upon receiving the synchronization acquisition command, each monitoring unit initiates its internal measurement process, continuously measuring the tilt angle change within the installation depth range using tilt sensors. The tilt angle measurement accuracy is no less than 0.01 degrees, fully characterizing the slow change process of the deep slope structure's attitude under gravity. Surface micro-displacement measurement employs high-precision displacement gauges or GNSS micro-displacement modules, measuring the minute relative movement of monitoring points through fixed reference points. The displacement change detection range is no less than ±10 mm, with a minimum resolution of 0.1 mm. To avoid interference from accidental factors in single acquisitions, each monitoring unit continuously acquires several sets of data and performs averaging and filtering processing.

[0021] During data acquisition, the background vibration environment of the geological early warning area is continuously monitored. Ground acoustic sensors deployed at the monitoring unit locations record high-frequency acoustic emission signals generated by micro-fractures in the earth's crust in real time. This background vibration signal is composed of a mixture of factors such as natural wind vibration, small-scale rockfalls, and disturbance from flora and fauna. To identify abnormal signals related to internal structural damage, at least 72 consecutive hours of background vibration data are collected, and the background noise baseline range is calculated. When the amplitude of the real-time detected signal continuously exceeds the upper limit of the background noise baseline for a duration of not less than 0.5 seconds, a ground acoustic event count is triggered. Each trigger automatically records the event time and the corresponding monitoring unit coordinates, achieving spatiotemporal localization of ground acoustic activity.

[0022] Upon receiving the deep tilt angle, surface micro-displacement, and ground sound event count data transmitted from each monitoring unit, the master control node immediately performs data aggregation and unified management. It achieves consistency across multiple data sources in the time dimension by precisely aligning timestamps, enabling the establishment of correspondences among the monitoring data. Subsequently, to construct a continuous data structure, the synchronized and aligned multi-type monitoring data are integrated into a sliding time window. The length of the sliding time window is set based on the geological disaster development characteristics, for example, a 30-minute period as a single window cycle, updated in 5-minute increments. This allows the window to smoothly slide along a continuous time axis while ensuring sufficient data volume and keeping lag within a reasonable range.

[0023] In S2, the phase space of the deep slope tilt angle sequence is reconstructed, the deformation trajectory correlation dimension between adjacent monitoring units is calculated, and the regional deformation synchronization index is generated.

[0024] After continuous acquisition of deep tilt angles, phase space reconstruction processing was performed on the tilt angle sequence corresponding to each monitoring unit to extract the evolution law of the internal structure of the slope from the time series data. First, the mutual information function was called to analyze the time dependence of the tilt angle sequence. The reconstruction time delay parameter was determined based on the location where the mutual information value first reached a local minimum. The optimal delay range was between 2 and 6 sampling periods to ensure sufficient dynamic differences between adjacent mapped points. After setting the time delay, the embedding dimension was evaluated for each time series using the spurious nearest neighbor method. When the proportion of spurious nearest neighbors dropped below 5%, the corresponding dimension was taken as the optimal embedding dimension. Based on the determined time delay and embedding dimension, the tilt angle time series was mapped to a high-dimensional phase space, forming a deformation trajectory with a continuous distribution of point sets. The connection between trajectory points reflects the continuous deformation trend under the influence of changes in the internal forces of the slope. During the reconstruction process, a unified processing rule was adopted, and no smoothing compression that might weaken morphological features was performed on the data to avoid incomparability of trajectories due to parameter differences.

[0025] After phase space reconstruction, trajectory correlation analysis is performed on each pair of adjacent monitoring units within the geological early warning area to quantify the coupling relationship between deformation evolution. First, a dynamic time warping algorithm is used to search for matching paths between the two trajectory sequences. Dynamic time warping constructs a cumulative cost matrix and finds the globally optimal matching path, ensuring alignment of attitude changes even with local scaling on the time axis. The matching process uses the Euclidean distance between the reconstructed trajectory points as the cost metric, setting the maximum allowable time offset ratio to 20% of the preset window length. In the correlation dimension calculation step, the General Process (GP) algorithm is used as the core method. By progressively expanding the neighborhood radius, the algorithm statistically analyzes the changing patterns of the number of trajectory points near high-dimensional state points, estimating the correlation dimension characteristics of the two trajectories on the joint attractor. Specifically, multiple progressively increasing neighborhood scales are set within the phase space, and the number of trajectory point pairs satisfying the neighborhood constraints within the joint point set of the two trajectories is counted. The slope of the double logarithmic trend fitting curve between the neighborhood extreme radius and the cumulative count is used as the correlation dimension corresponding to the pair of monitoring units. The closer this dimension value is to the low-dimensional structure, the more consistent the evolutionary trends of the two trajectories are.

[0026] Finally, the correlation dimensions corresponding to each pair of spatially adjacent monitoring units are statistically analyzed. During the statistical process, an arithmetic mean is calculated for all correlation dimensions. This involves summing the dimension values ​​obtained for each adjacent pair and dividing the sum by the total number of adjacent pairs; the average value is the regional deformation synchronization index. This index directly reflects the degree of consistency in deformation behavior at the dynamic level across the entire region, eliminating the need for additional thresholds or weights and thus avoiding biases caused by manually set parameters. To ensure stable calculation results, invalid dimension records resulting from data interruptions are excluded before averaging. Furthermore, the number of valid adjacent pairs within each time window is required to be no less than 90% of the total number of adjacent pairs. If this percentage is lower, the data quality is considered unsatisfactory, triggering an automatic re-acquisition mechanism. The resulting deformation synchronization index is stored as a continuously changing curve in the master node and calculated once per update cycle, ensuring that the index changes accurately describe the evolutionary trend of regional deformation synergy over time.

[0027] In S3, based on the spatiotemporal distribution of ground acoustic event counts, a density clustering algorithm is used to identify acoustic emission clusters, and the system entropy is calculated by combining the surface micro-displacement vector field.

[0028] All triggered ground acoustic event counts were spatiotemporally aggregated to construct a spatiotemporally distributed dataset containing the event occurrence time and the coordinates of the monitoring unit where the event occurred. Each record includes data in a uniform format, namely the event trigger time and the corresponding three-dimensional coordinates, ensuring accurate indexing of events in both time and spatial dimensions. Subsequently, a density-based clustering algorithm was used to analyze and process this spatiotemporally distributed dataset. Clustering uses distance constraints and a minimum number of events to determine cluster units. For example, the spatial radius was set to a threshold within the range of 5 to 15 meters, and the number of events within a local time period was required to be no less than 3, exceeding the random distribution level of background noise. This resulted in the aggregation of dense acoustic emission clusters within a unit spatiotemporal volume. The clustering algorithm execution order was as follows: merging event point sets based on spatial neighborhood relationships, removing sparse points that did not meet the density requirements as noise records, and finally generating multiple acoustic emission anomaly clusters with clear geometric boundaries and time periods. Each acoustic emission cluster is composed of its spatial outer envelope and the event time period, with its boundary directly taken from the envelope of the point set aggregated by the clustering algorithm.

[0029] After identifying acoustic emission clusters, time-series data of surface micro-displacements corresponding to all monitoring units within the spatial area covered by each cluster are extracted. The observed coordinate changes are uniformly converted into vector form, where a vector is defined as the instantaneous offset direction and length of a monitoring point in a fixed reference coordinate system. Within the same time window, the set of displacement vectors for all monitoring units in the region is obtained, and the direction of each vector is normalized to polar angle. Subsequently, the dispersion of all vector directions in the angular distribution is calculated. The dispersion is directly and quantitatively represented by standard methods such as the variance of the direction distribution or the reciprocal of the length of the mean vector on the unit circle. Higher dispersion indicates stronger inconsistency in the direction of local structural deformation. Simultaneously, the amplitude changes of all displacement vectors within the same cluster at continuous time sampling points are extracted, and their fluctuation intensity is statistically analyzed. The fluctuation intensity is calculated based on the standard deviation of the vector length changes between adjacent time points. Both the dispersion and fluctuation intensity are calculated using unified rules. The two are used as equivalent dimensional indicators for local disorder calculation. After normalization, they are directly added together. The sum of the dispersion value and the fluctuation intensity value is used as the local disorder result for the cluster area, which fully reflects the randomness and instability of surface displacement changes in the area.

[0030] After obtaining the local disorder degrees corresponding to all acoustic emission clusters, to ensure consistency and continuity in the overall regional assessment, a weighted average is performed based on the number of monitoring units in each cluster to generate the final system entropy value. Specifically, the weighting method is as follows: the local disorder degree of each cluster is multiplied by the proportion of the number of monitoring units in that cluster to the total number of monitoring units. The sum of all weighted values ​​is the system entropy value. The weighting ratio is fixed between 0 and 1, and the sum is 1; the ratio value is determined by the number of monitoring units.

[0031] In step S4, the deformation synchronization index and system entropy value are input into a critical state determination model based on dual threshold grading to identify the critical state of the geological early warning area.

[0032] Based on the continuous updating of the deformation synchronization index and system entropy, a two-dimensional plane for determining critical states is constructed, using the deformation synchronization index as the abscissa and the system entropy as the ordinate. This plane serves as a unified mapping space for distinguishing regional geological states. To ensure that the determination criteria match the actual stability of the geological structure, a safe judgment area is first constructed in this two-dimensional plane based on historical stable state data clusters. The historical stable state data clusters are sets of monitoring records from long-term monitoring data that have not experienced any geological disasters or significant deformation changes. The central region of these data clusters is extracted as the stable center point through statistical clustering. Then, the lengths of the major and minor axes of an ellipse are calculated based on the standard deviations of this set in the two coordinate directions, forming an elliptical safe boundary with the stable center point as its geometric center. This elliptical boundary completely encompasses all historical observations under normal stable conditions, and the outer edge of the boundary indicates that the regional structure has deviated from the critical range of the stable state. To avoid errors from manual correction of the boundary curve, the elliptical boundary parameters are automatically generated entirely from the statistical distribution of actual historical data. The lengths of both the major and minor axes are expanded outwards by a safety margin coefficient increased by 20% based on the dispersion of the stable state data, ensuring that the safe boundary has sufficient capacity to accommodate normal fluctuations and preventing misjudgments. Once the elliptical boundary is constructed, all real-time data points within that region are considered to be in a stable or substable geological state.

[0033] A concentric sector-shaped critical decision zone is established outside the elliptical safety boundary. The sector boundary consists of two parts: a radial boundary, representing the degree of deviation from the stable region; and an azimuth boundary, representing the reliable statistical trend of the deviation direction. The radial boundary starts at any point on the outer edge of the elliptical safety boundary and extends outward in any direction within the two-dimensional plane. The extension distance is determined based on the extreme range of the joint fluctuation of the deformation synchronization index and system entropy value in historical monitoring data. The minimum radial extension distance is selected as the range corresponding to the simultaneous peak values ​​of the two indicators in historical records, ensuring that the sector only covers the observed significant instability trend range. After determining the radial boundary, to extract directional information with real disaster-causing significance from the data evolution trend, critical record points of actual geological disaster events in historical data are extracted. The coordinates of these record points in the two-dimensional decision plane are collected, connected to the origin, and the geometric angle between each line and the horizontal axis is calculated. Subsequently, each angle is assigned a corresponding weight according to the severity of the disaster event, and a weighted average is performed. The resulting angle is used as the reference azimuth of the sector-shaped decision zone. To account for natural biases caused by monitoring noise and inconsistencies in index evolution, the permissible azimuth deviation range is set as a sector-shaped region of 45 degrees on each side of the reference azimuth axis, ensuring sufficient statistical robustness. This constructed sector-shaped critical decision region will be used to identify state points exhibiting significant instability trends.

[0034] After the elliptical safety boundary and the sector-shaped critical decision region are constructed, dynamic critical decision-making is performed on the real-time monitoring data. When a new monitoring cycle arrives, the deformation synchronization index and system entropy value corresponding to that cycle are first calculated and updated. The real-time data points are mapped onto the two-dimensional critical state decision plane, and the spatial relationship between the point and the elliptical safety boundary and the sector-shaped critical decision region is detected. If the real-time point still falls within the elliptical safety boundary, the current state is considered a normal stable state, and no critical response is triggered. When the real-time point crosses the elliptical safety boundary but does not enter the sector-shaped critical decision region, it indicates that although the regional state deviates from the normal range, it has not yet met the requirements for the direction of instability, and is marked as a transitional state. When the real-time data point simultaneously meets both conditions, i.e., it both crosses the elliptical safety boundary and falls within the sector-shaped critical decision region, the region is directly identified as having entered a critical state.

[0035] In step S5, real-time rainfall intensity data is extracted based on the geological warning area entering the critical state, multi-level warning zones are divided by spatial overlay, and corresponding warning signals are generated.

[0036] After identifying the critical state of the geological early warning area, real-time rainfall intensity data covering the entire area is acquired. Rainfall intensity includes the current cumulative rainfall and the real-time rainfall rate. Spatial interpolation is used to obtain the rainfall intensity isosurface corresponding to the monitoring unit grid. Simultaneously, the coordinates of monitoring data points that have entered the critical state are extracted from the two-dimensional critical judgment plane. Using the origin as a stable state reference point, the distance between the monitoring data point and the origin is calculated using a geometric distance method. This distance is used as a critical deviation index to quantify the degree of disruption to the regional stability. Subsequently, rainfall intensity and critical deviation are standardized to ensure they are on the same numerical scale. Rainfall triggering factors and internal structural instability factors are jointly characterized, and their product is used as the comprehensive early warning index value. The comprehensive early warning index can simultaneously reflect the true risk state resulting from the superposition of the external triggering effect of rainfall factors and the degree of internal deterioration due to structural deformation within the region. Based on the distribution patterns of comprehensive early warning indicators, the region is divided into different alert levels. Each monitoring cycle dynamically adjusts the zoning according to real-time calculation results, so that the boundaries of critical areas can shrink or expand as structural deterioration intensifies or rainfall eases, ultimately forming a multi-level dynamic early warning zoning that matches the actual risk trend.

[0037] Vectorized early warning zoning maps are generated based on the comprehensive indicator results for each region. Specifically, this involves: using the comprehensive early warning indicator as a layer attribute, assigning differentiated color labels to regions within different indicator value ranges to achieve visual identification of zoning levels; and recording the spatial boundaries, indicator values, and update time of each region as structured attributes in the vector data, ensuring the traceability and statistical compatibility of the early warning areas. Subsequently, the early warning zoning maps and comprehensive early warning indicator levels are uniformly encoded and packaged into a real-time early warning signal using a combination of text, numerical values, and graphics, enabling analysis through regional monitoring systems, information platforms, or emergency command and dispatch terminals. The early warning signal clearly records the basis for the region entering a state of alert, namely the comprehensive indicator calculation results formed by combining current rainfall conditions and critical deviation, allowing responders to quickly locate the source of risk.

[0038] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0039] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0040] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0041] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0042] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0043] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0044] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0045] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0046] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0047] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A regional geological disaster rapid early warning method, characterized in that, The method comprises the following steps: S1, collecting slope deep inclination angle, ground sound event count and surface micro-displacement data of the geological warning area at a set frequency, and constructing a multi-dimensional monitoring data stream; S2, phase space reconstruction is performed on the slope deep inclination angle sequence, the correlation dimension of the deformation trajectory between adjacent monitoring units is calculated, and a regional deformation synchronization index is generated; S3, according to the spatio-temporal distribution of the ground sound event count, a density clustering algorithm is used to identify a sound emission cluster area, and a system entropy value is calculated in combination with a surface micro-displacement vector field; S4, the deformation synchronization index and the system entropy value are input into a critical state judgment model based on double-threshold grading to identify the critical state of the geological warning area; S5, based on the real-time rainfall intensity data of the geological warning area entering the critical state, a multi-level warning partition is divided through spatial superposition, and a corresponding warning signal is generated.

2. The regional geological disaster early warning method according to claim 1, characterized in that, In S1, collecting slope deep inclination angle, ground sound event count and surface micro-displacement data of the geological warning area at a set frequency, and constructing a multi-dimensional monitoring data stream specifically comprises: A sensor array with a master node as the core is arranged in the geological warning area, and each sensor covers a monitoring unit; The geological warning area is a geological structure active area within a set monitoring range, and the monitoring unit is a potential landslide body and a large dangerous rock mass pre-labeled in the geological warning area; The master node sends a synchronous acquisition instruction to each monitoring unit at a set frequency, and the monitoring unit collects slope deep inclination angle and surface micro-displacement data; The background vibration environment of the geological warning area is continuously monitored, and the ground sound event count is triggered when the vibration signal amplitude exceeds the environmental noise baseline; The collected data is converged to the master node, and multi-source monitoring data is aligned according to the time stamp. The aligned data is integrated using a sliding time window mechanism to construct a multi-dimensional monitoring data stream with spatio-temporal continuity.

3. The regional geological disaster early warning method according to claim 1, characterized in that, In S2, phase space reconstruction is performed on the slope deep inclination angle sequence, the correlation dimension of the deformation trajectory between adjacent monitoring units is calculated, and a regional deformation synchronization index is generated specifically comprising: The time delay parameter of phase space reconstruction is set by using the mutual information function, the best embedding dimension is calculated by combining the false nearest neighbor method, and the inclination angle time sequence of each monitoring unit is mapped into the deformation trajectory in the high-dimensional phase space; The morphological matching degree between the deformation trajectories of adjacent monitoring units is measured by using the dynamic time warping algorithm, and the correlation dimension value of each pair of adjacent units is calculated based on the trajectory matching result; The correlation dimension values of all adjacent monitoring unit pairs are taken as the arithmetic mean value, and the arithmetic mean value is directly taken as the regional deformation synchronization index.

4. The regional geological disaster early warning method according to claim 1, characterized in that, In S3, according to the spatio-temporal distribution of the ground sound event count, a density clustering algorithm is used to identify a sound emission cluster area, and a system entropy value is calculated in combination with a surface micro-displacement vector field specifically comprising: A spatio-temporal distribution data set is constructed according to the time of triggering the ground sound event count and the monitoring unit coordinates; The spatio-temporal distribution data set of the ground sound event count is processed by using a density-based clustering algorithm to identify a sound emission cluster area with an event density exceeding the average level in a unit spatio-temporal volume; In the spatial range corresponding to each acoustic emission cluster, the ground micro-displacement data of all monitoring units in the region are extracted and converted into micro-displacement vectors, and the discrete degree and fluctuation intensity of the amplitude change of each vector direction are calculated; Based on the discrete degree and fluctuation intensity, the local disorder degree of the cluster is calculated, and the system entropy value is obtained by weighted average of the local disorder degrees of all acoustic emission clusters.

5. The regional geological disaster early warning method according to claim 1, characterized in that, In S4, the deformation synchronization index and the system entropy value are input into a critical state judgment model based on double-threshold grading, and the critical state of the geological warning area is identified, which specifically includes: A two-dimensional critical state judgment plane is constructed with the deformation synchronization index and the system entropy value as coordinates; An elliptical safety boundary with a historical stable state data cluster as the core is defined in the critical state judgment plane, and a concentric fan-shaped critical judgment area is set outside the safety boundary; When the real-time data point simultaneously breaks through the safety boundary and falls into the critical judgment area, it is determined that the geological warning area enters the critical state.

6. The regional geological disaster early warning method according to claim 5, characterized in that, The range of the fan-shaped critical judgment area is jointly defined by the radial boundary and the azimuth angle, which specifically includes: The radial boundary starts from the outer edge of the elliptical safety boundary, and its corresponding radial distance is set based on the historical extreme value fluctuation range of the regional deformation synchronization index and the system entropy value calculated from the corresponding historical monitoring data of the monitoring unit; The azimuth angle is calculated by extracting the extreme point coordinates of the regional deformation synchronization index and the regional instability degree recorded when the historical geological disaster event occurs, and the angle between the connecting line of each extreme point and the origin of the judgment plane coordinate system is calculated; The weighted average of the angles is calculated according to the scale grade of the historical disaster event, and the obtained weighted average value is set as the reference azimuth angle of the fan-shaped critical judgment area, and the tolerance range is set as a fan-shaped expansion surface of forty-five degrees on both sides of the reference azimuth angle axis.

7. The regional geological disaster early warning method according to claim 1, characterized in that, In S5, real-time rainfall intensity data is extracted based on the geological warning area that has entered the critical state, and multi-level warning subareas are divided through spatial superposition, and corresponding warning signals are generated, which specifically includes: Rainfall intensity data of the geological warning area is obtained, and the spatial range of the area determined to be in a critical state is extracted; The boundary of the critical state area is spatially superimposed with the real-time rainfall intensity contour line to determine the actual warning area within the rainfall influence range, and the geological warning area is divided into multi-level warning subareas; According to the warning subarea result, the corresponding vector warning subarea map is generated, and the vector warning subarea map and the warning level of the subarea are integrated into a warning signal.

Images