A coal mine slope slip identification method and system based on radar monitoring

Abstract

The application discloses a coal mine slope slip identification method and system based on radar monitoring, and belongs to the field of coal mine slope slip identification. The method comprises the following steps: obtaining phase difference information in a denoising radar signal data set, calculating a millimeter-level displacement vector of a slope surface through interference measurement, and determining a micro-deformation information distribution map; fusing geological complex parameters for the micro-deformation information distribution map, classifying deformation feature types by using a support vector machine algorithm, and judging which deformation features belong to abnormal inclination changes; if the abnormal inclination changes exceed a preset threshold, extracting time sequence data from the micro-deformation information distribution map to obtain an inclination change trend sequence; comparing the inclination change trend sequence with historical slip hidden danger data, predicting a dynamic evaluation index by using a support vector machine algorithm, and determining an early signal strength level; integrating a stability analysis model according to the early signal strength level, and generating a slip hidden danger probability distribution map and a risk identification report.

Description

Technical Field

[0001] This invention belongs to the field of coal mine slope slippage identification technology, and particularly relates to a method and system for coal mine slope slippage identification based on radar monitoring. Background Technology

[0002] Slope slippage in coal mines is a significant hazard threatening safe production in open-pit mining, as its stability directly impacts personnel safety and equipment operating efficiency. Slope slippage can lead to substantial economic losses and even casualties; therefore, accurate monitoring and risk identification of slope stability have become core issues in the field of mining engineering.

[0003] In recent years, radar monitoring technology has been widely used in slope deformation monitoring due to its high precision and non-contact characteristics. Current slope monitoring methods largely rely on traditional measurement techniques, such as total stations or GPS positioning. These methods are often limited to point sampling when acquiring data, making it difficult to comprehensively capture the overall deformation characteristics of the slope surface. In the field of radar monitoring, acquiring micro-deformation information of the slope surface has become a key technical aspect. Micro-deformation information refers to millimeter-level deformation data of the slope surface. This data can reflect subtle changes in the slope angle and is an important basis for assessing the risk of landslide.

[0004] However, existing methods still face significant limitations in practical applications, failing to meet the needs of dynamic assessment of slippage risk under complex geological conditions. Especially in complex geological environments such as coal mine slopes, the slope is affected by mining disturbances, rainfall, or earthquakes, resulting in uneven deformation distribution. Traditional methods cannot effectively cover large areas of the slope, leading to a lack of comprehensive monitoring results. Furthermore, these methods are mostly discrete measurements, making it difficult to reflect the dynamic changes of the slope in real time and easily missing early signs of slippage. Under complex geological conditions, the slope material composition is diverse, such as a mixture of loose deposits and hard rock layers, resulting in significant differences in the reflection characteristics of radar signals, making it difficult to accurately extract micro-deformation information. Even more challenging is the complex and variable relationship between slope dip angle changes and slippage occurrence. How to extract the laws directly related to slippage from these changes and then quantify slope stability remains an unsolved problem. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention proposes a method and system for identifying coal mine slope slippage based on radar monitoring, thereby resolving the issues present in the prior art.

[0006] To achieve the above objectives, in a first aspect, the present invention provides a method for identifying coal mine slope slippage based on radar monitoring, comprising:

[0007] By acquiring surface reflection signals of coal mine slopes using synthetic aperture radar, and performing initial data aggregation based on the mixed reflection characteristics of loose soil and rock layers in geologically complex areas, the original radar signal dataset is obtained.

[0008] Wavelet transform filtering is used to process the signal interference of the original radar signal dataset to obtain a denoised radar signal dataset.

[0009] The phase difference information in the denoised radar signal dataset is obtained, and the millimeter-level displacement vector on the slope surface is calculated by interferometry to determine the distribution map of micro-deformation information.

[0010] To address the integration of micro-deformation information distribution maps with complex geological parameters, a support vector machine algorithm is employed to classify deformation feature types and determine which deformation features belong to dip angle anomalies.

[0011] If the change in tilt angle exceeds the preset threshold abnormally, time series data is extracted from the micro-deformation information distribution map to obtain the tilt angle change trend sequence;

[0012] By comparing the trend sequence of tilt angle changes with historical slip hazard data, the support vector machine algorithm is used to predict dynamic evaluation indicators and determine the early signal strength level.

[0013] Based on the early signal strength level, an integrated stability analysis model is used to generate a probability distribution map of slippage hazards and a risk identification report.

[0014] Preferably, the original radar signal dataset contains original echo datasets containing information about geologically complex areas.

[0015] Preferably, the process of obtaining the denoised radar signal dataset includes:

[0016] Wavelet transform is used to decompose the original radar signal dataset, separating the low-frequency component and the high-frequency noise component to obtain the denoised radar signal dataset.

[0017] Preferably, the process of determining the micro-deformation information distribution map includes:

[0018] The raw phase data from the radar signal dataset is obtained, and the denoised phase data is obtained by signal denoising processing.

[0019] The phase difference data is calculated from the denoised phase data using interferometry to obtain the slope surface displacement vector;

[0020] Based on the slope surface displacement vector and combined with the terrain surface features, displacement field data with millimeter-level accuracy is generated.

[0021] If the displacement vector in the displacement field data exceeds a preset threshold, it is marked as a micro-deformation region, and micro-deformation distribution data is obtained.

[0022] A spatial interpolation algorithm is used to generate a high-resolution micro-deformation information distribution map from micro-deformation distribution data.

[0023] Preferably, the process of determining which deformation features belong to abnormal tilt angle changes includes:

[0024] We acquire micro-deformation information distribution map data and geological parameter data, and use standardized processing to generate a fused dataset in a unified format.

[0025] The fused dataset is classified using the support vector machine algorithm to obtain the deformed feature types.

[0026] Features related to tilt angle changes are extracted from deformation feature types, and a preset threshold is used to determine abnormal tilt angle changes.

[0027] Preferably, the sequence of tilt angle change trends is obtained, including:

[0028] If the tilt angle change abnormally exceeds the preset threshold, the micro-deformation information distribution map is marked with spatiotemporal correlation to determine the correspondence between the elements in the map and the time dimension.

[0029] Based on the location of abnormal regions, spatial units related to abnormal tilt angle changes are selected from the distribution map;

[0030] Extract the tilt angle monitoring values ​​of the selected spatial units at each time point in chronological order;

[0031] The extracted discrete monitoring values ​​are serialized and organized to form a continuous trend sequence of tilt angle changes.

[0032] Preferably, the process of determining the early signal strength level includes:

[0033] Obtain the dip angle change trend sequence and historical slip hazard data, and perform time series analysis on both using sequence data processing technology to obtain standardized time series data;

[0034] Feature vectors are extracted from standardized time series data, and feature vector extraction techniques are used to generate a feature set containing dip angle change trends and slip hazard characteristics.

[0035] The feature set is trained using the support vector machine algorithm, and a prediction model is established through data correlation analysis to obtain the prediction results of dynamic evaluation indicators.

[0036] Based on the prediction results of dynamic evaluation indicators, the early signal strength level is calculated using signal strength evaluation technology to obtain the quantified value of signal strength.

[0037] Preferably, a probability distribution map of slippage hazards and a risk identification report are generated, including:

[0038] Early signal strength data is acquired by collecting environmental parameters through a sensor network to obtain the raw dataset;

[0039] A pre-defined stability analysis model is used to extract features from the original dataset to determine the signal feature set;

[0040] If the fluctuation value in the signal feature set exceeds the preset threshold, the support vector machine algorithm is used for classification to determine potential slip risks.

[0041] Based on the classification results, the probability distribution of slippage hazards is calculated using the probability density function to obtain probability distribution data;

[0042] By using probability distribution data, Bayesian networks are employed to analyze the spatial distribution characteristics of potential hazards and determine the risk area delineation.

[0043] Obtain the risk zone delineation results, combine them with a geographic information system for visualization processing, and generate a probability distribution map of landslide hazards;

[0044] For high-risk areas in the distribution map, cluster analysis is used to prioritize them and obtain the final risk identification results.

[0045] Secondly, the present invention also provides a coal mine slope slip identification system based on radar monitoring, comprising:

[0046] The signal acquisition module is used to acquire surface reflection signals of coal mine slopes using synthetic aperture radar algorithms. It performs initial data aggregation based on the mixed reflection characteristics of loose soil and rock layers in geologically complex areas to obtain the original radar signal dataset.

[0047] The signal processing module is used to process the signal interference of the original radar signal dataset using wavelet transform filtering to obtain a denoised radar signal dataset.

[0048] The phase analysis module is used to acquire phase difference information from the denoised radar signal dataset, calculate the millimeter-level displacement vector on the slope surface through interferometry, and determine the distribution map of micro-deformation information.

[0049] The feature classification module is used to classify deformation feature types by integrating complex geological parameters with micro-deformation information distribution maps and using the support vector machine algorithm to determine which deformation features belong to dip angle anomalies.

[0050] The sequence extraction module is used to extract time series data from the micro-deformation information distribution map if the tilt angle change abnormally exceeds a preset threshold, so as to obtain the tilt angle change trend sequence.

[0051] The prediction and assessment module is used to predict dynamic assessment indicators and determine the early signal strength level by comparing the dip angle change trend sequence with historical slip hazard data and using the support vector machine algorithm.

[0052] The report generation module is used to integrate the stability analysis model based on the early signal strength level to generate a slip hazard probability distribution map and risk identification report.

[0053] Thirdly, the present invention also discloses a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in the first aspect.

[0054] Compared with the prior art, the present invention has the following advantages and technical effects:

[0055] This invention discloses a method for identifying coal mine slope slippage based on radar monitoring, solving the operational problem of accurately identifying micro-deformation features and dynamically assessing risks in geologically complex areas. The invention acquires synthetic aperture radar signals and aggregates the original dataset. Wavelet transform filtering is used to separate low-frequency micro-deformation signals from high-frequency noise, generating a denoised dataset. Subsequently, phase difference is extracted through interferometry, and millimeter-level displacement vectors are calculated to construct a micro-deformation distribution map. Support vector machines are used to classify deformation features and identify abnormal tilt angle changes. If the abnormality exceeds a threshold, time-series data is extracted to generate a trend sequence, which is compared with historical slippage hazards to predict dynamic assessment indicators, determine early signal strength, and finally integrate a stability model to generate a slippage hazard probability distribution map.

[0056] This invention achieves precise monitoring and dynamic risk prediction of slope micro-deformation through multi-source data fusion and intelligent analysis, significantly improving the efficiency and accuracy of coal mine slope safety management. Attached Figure Description

[0057] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0058] Figure 1 This is a flowchart of the coal mine slope slip identification method according to an embodiment of the present invention;

[0059] Figure 2 This is a schematic diagram of a coal mine slope slip recognition system according to an embodiment of the present invention. Detailed Implementation

[0060] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0061] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0062] Example 1

[0063] like Figure 1 As shown, this embodiment provides a method for identifying coal mine slope slippage based on radar monitoring, including:

[0064] S101. Collect surface reflection signals of coal mine slopes using synthetic aperture radar, and perform initial data aggregation based on the mixed reflection characteristics of loose soil and rock layers in geologically complex areas to obtain the original radar signal dataset.

[0065] Specifically, in coal mine slope monitoring, the synthetic aperture radar (SAR) algorithm is first used to scan the slope surface with a 10 GHz center frequency and 1 meter resolution parameters by deploying a high-resolution X-band radar sensor array to collect reflected signal data. The algorithm adopts a range-Doppler processing flow, including pulse compression to improve the signal-to-noise ratio to more than 20 dB and azimuth-to-Fourier transform to achieve imaging, generating a raw echo dataset containing information on geologically complex areas, i.e., the raw radar signal dataset.

[0066] S102. Wavelet transform filtering is used to process the signal interference of the original radar signal dataset to obtain a denoised radar signal dataset.

[0067] Wavelet transform is used to decompose the original radar signal dataset, separating the low-frequency component and the high-frequency noise component to obtain the denoised radar signal dataset.

[0068] Specifically, based on the processing of the original radar signal dataset, wavelet transform filtering technology is used to process signal interference in coal mine slope monitoring, separating low-frequency components related to micro-deformation and high-frequency noise components to generate a denoised radar signal dataset. First, for the acquired original radar signal dataset, assuming a data size of 800GB, the Discrete Wavelet Transform (DWT) algorithm is used, selecting the Daubechies wavelet (db4) as the basis function due to its good orthogonality and compact support characteristics, making it suitable for non-stationary signal analysis. The decomposition level is set to 5 levels. By performing multi-scale decomposition of the signal, the frequency range is divided into low-frequency approximate components (0-500Hz, containing micro-deformation information) and high-frequency detail components (above 500Hz, mainly noise). During the decomposition process, the energy spectrum of each level is calculated, revealing that the energy proportion of the low-frequency components is approximately 85%, while the high-frequency noise component accounts for 15%. Subsequently, a soft thresholding denoising method is used, setting the threshold to σ√(2lnN), where σ is the noise standard deviation, estimated at 0.02, and N is the signal length, approximately 10^7. By thresholding the coefficients of high-frequency components, retaining low-frequency components and removing noise coefficients, and using inverse wavelet transform (IDWT) to reconstruct the signal, a denoised dataset of approximately 750GB was generated. To ensure the denoising effect, the root mean square error (RMSE) of the signal before and after denoising was calculated, and the result was 0.03mm, indicating that micro-deformation information was preserved.

[0069] S103. Obtain the phase difference information in the denoised radar signal dataset, calculate the millimeter-level displacement vector on the slope surface through interferometry, and determine the distribution map of micro-deformation information.

[0070] The raw phase data from the radar signal dataset is acquired, and denoised phase data is obtained through signal denoising. Phase difference data is calculated from the denoised phase data using interferometry to obtain the slope surface displacement vector. Based on the slope surface displacement vector and combined with topographic surface features, millimeter-level precision displacement field data is generated. If the displacement vector in the displacement field data exceeds a preset threshold, it is marked as a micro-deformation region, yielding micro-deformation distribution data. A spatial interpolation algorithm is used to generate a high-resolution micro-deformation information distribution map from the micro-deformation distribution data.

[0071] For example, in coal mine slope monitoring scenarios, acquiring phase data from the raw radar signal dataset is a crucial first step. Raw phase data typically contains electromagnetic wave information reflected from the slope surface, and is affected by terrain undulations, vegetation cover, and atmospheric disturbances, resulting in significant noise. Phase data can be extracted from the raw radar signal using a phase unwrapping algorithm.

[0072] For example, using a synthetic aperture radar (SAR) system operating in the C-band at a wavelength of 5.6 cm, the acquired phase data has a resolution of 0.1 radians and a data volume of approximately 500 GB. This data directly reflects the minute deformations of the slope surface, but further processing is required to remove noise.

[0073] Specifically, signal denoising can employ adaptive filtering methods, such as Kalman filtering, to smooth random disturbances in the phase data. In one implementation, assuming the noise is Gaussian white noise with a standard deviation of 0.05 radians, the estimated phase values ​​are iteratively updated to generate a denoised phase dataset of approximately 450 GB. This method effectively preserves phase information related to slope deformation while filtering out high-frequency noise, providing a reliable foundation for subsequent calculations.

[0074] In one embodiment, the calculation of phase difference data based on denoised phase data can be achieved using interferometry. Interferometry utilizes the difference between two phase images to generate interference fringes, reflecting the relative displacement of the slope surface at different time points.

[0075] For example, by using two SAR images taken 24 hours apart, the phase difference can be calculated with a resolution of 0.01 radians. The phase difference data directly corresponds to the displacement changes on the slope surface. This method can capture deformation information at the millimeter level and is suitable for high-precision monitoring.

[0076] For example, when converting phase difference data into a slope surface displacement vector, geometric calculations can be performed using radar incident angle and wavelength information. Assuming an incident angle of 30 degrees, a wavelength of 5.6 cm, and a phase difference of 0.5 radians, the displacement vector can be calculated to be approximately 2 mm. Combined with terrain surface features, such as elevation data provided by a digital elevation model (DEM) with a resolution of 1 m, displacement field data with millimeter-level accuracy can be generated. This displacement field data can intuitively reflect the spatial deformation distribution of the slope surface.

[0077] Specifically, if the displacement vector in the displacement field data exceeds a preset threshold, such as 3 mm, it is marked as a micro-deformation region.

[0078] In one possible implementation, spatial clustering algorithms, such as K-means clustering, are used to group displacement vectors by region, generating micro-deformation distribution data. The data shows that the displacement vectors in the top region of the slope are concentrated in the 2-4 mm range, consistent with micro-deformation characteristics. This distribution data provides crucial information for slope stability assessment.

[0079] In one embodiment, a Kriging spatial interpolation algorithm is used to generate a high-resolution micro-deformation information distribution map from micro-deformation distribution data. The Kriging algorithm utilizes spatial autocorrelation to interpolate and generate a distribution map with a resolution of 0.5m.

[0080] For example, the interpolation reveals a 10m² micro-deformation zone in the middle of the slope, with a displacement of approximately 3.5mm. This distribution map visually demonstrates the spatial characteristics of slope deformation, facilitating engineers' identification of potential risk areas.

[0081] For example, the micro-deformation information distribution map generated through the above processing can be used to guide slope reinforcement measures, such as arranging monitoring points or reinforcement structures for marked micro-deformation areas. This method improves the accuracy and efficiency of slope monitoring and provides data support for safe production.

[0082] S104. Based on the micro-deformation information distribution map fused with complex geological parameters, the support vector machine algorithm is used to classify the deformation feature types and determine which deformation features belong to dip angle change anomalies.

[0083] Micro-deformation information distribution data and geological parameter data were acquired and standardized to generate a fused dataset in a unified format. The fused dataset was then classified using a support vector machine algorithm to determine deformation feature types. Dip change-related features were extracted from these deformation feature types, and a preset threshold was used to identify anomalous dip change features.

[0084] For example, in coal mine slope monitoring, acquiring micro-deformation information distribution maps and geological parameter data are crucial steps. Micro-deformation information distribution maps typically contain displacement vector data of the slope surface, with a resolution of 0.5m and a data volume of approximately 200GB, reflecting the spatial deformation characteristics of the slope. Geological parameter data includes rock mass strength, fracture distribution, and groundwater level, and the data formats are diverse; for example, rock mass strength is expressed in MPa, and fracture distribution is expressed as the number of fractures per square meter. Standardization is the process of converting these heterogeneous data into a unified format.

[0085] Specifically, the numerical range of microdeformation data and geological parameter data can be mapped to the interval between 0 and 1.

[0086] For example, displacement vector values ​​from 0 to 5 mm are mapped to 0 to 1, and rock mass strengths from 10 to 50 MPa are also mapped to 0 to 1. This method facilitates subsequent algorithm processing and improves data compatibility.

[0087] In one embodiment, the generation of the fusion dataset requires integrating standardized microdeformation data and geological parameter data.

[0088] For example, the micro-deformation distribution map data is stored in raster form, with each raster point containing a displacement vector value; the geological parameter data is recorded in point cloud form, with each point corresponding to rock mass strength or fracture density. Using spatial interpolation methods, such as inverse distance weighting, the geological parameter data and micro-deformation data are aligned to generate a unified format fused dataset, approximately 250GB in size. Each data point in the fused dataset contains multi-dimensional features such as displacement vector, rock mass strength, and fracture density, making it suitable for machine learning algorithm analysis.

[0089] For example, when using the support vector machine algorithm to classify the features of a fused dataset, the data can be divided into three categories: stable, slightly deformable, and potential landslides.

[0090] Specifically, support vector machines classify data points in a multidimensional feature space by constructing a hyperplane.

[0091] For example, areas with a displacement vector greater than 3 mm and a crack density higher than 5 cracks / m² may be classified as potential landslide areas. During the classification process, a kernel function, such as a radial basis function, needs to be set to handle nonlinear features. The classification results generate a deformation feature type dataset containing category labels for each region, with a data size of approximately 10 GB. This classification method can effectively distinguish different slope conditions.

[0092] In one embodiment, when extracting dip change-related features from deformation feature types, dip change can be calculated based on displacement vectors and terrain data.

[0093] For example, using digital elevation model data with a resolution of 1m, the elevation difference between adjacent grid points is calculated to obtain the local tilt change value, in degrees. Suppose the tilt change in a certain area increases from 2 degrees to 5 degrees, indicating significant deformation. When using a preset threshold to identify anomalous features, the tilt change threshold can be set to 3 degrees. If the tilt change in a certain area exceeds 3 degrees, it is marked as an anomalous feature area.

[0094] For example, a 10-square-meter area in the middle of a slope with a dip angle change of 4 degrees is marked as an anomaly. This method can accurately identify potential risk areas and provide a reliable basis for slope stability assessment.

[0095] S105. If the tilt angle change abnormally exceeds the preset threshold, time series data is extracted from the micro-deformation information distribution map to obtain the tilt angle change trend sequence.

[0096] If the tilt angle change abnormally exceeds the preset threshold, the micro-deformation information distribution map is marked with spatiotemporal correlation to determine the correspondence between elements in the map and the time dimension; based on the location of the abnormal area, spatial units related to the tilt angle change abnormality are selected from the distribution map; the tilt angle monitoring values ​​of the selected spatial units at each time node are extracted in chronological order; the extracted discrete monitoring values ​​are serialized and organized to form a continuous tilt angle change trend sequence.

[0097] Specifically, when the coal mine slope monitoring system detects an abnormal change in dip angle exceeding a preset threshold in a certain area, the system's spatiotemporal correlation program is activated. This program matches and correlates each monitoring point on the micro-deformation information distribution map with its corresponding monitoring time record, assigning a time identifier to each monitoring point on the micro-deformation information distribution map. The specific slope area with the abnormal dip angle is determined using a slope anomaly localization algorithm. All dip angle monitoring units and surrounding associated monitoring points within this area are selected from the micro-deformation information distribution map as target spatial units. In chronological order, the measured dip angle data of these target spatial units for each monitoring period before and after the anomaly occurs are extracted from the monitoring database. The extracted discrete dip angle data are then processed, and invalid interference data is removed. The data is then arranged along the time axis to form a trend sequence of continuous change in dip angle over time in the slope anomaly area. This sequence is stored in a specialized analysis library and triggers the subsequent slip risk assessment process.

[0098] S106. By comparing the trend sequence of tilt angle changes with historical slip hazard data, the support vector machine algorithm is used to predict dynamic evaluation indicators and determine the early signal strength level.

[0099] The process involves acquiring dip angle change trend sequences and historical slip hazard data, performing time series analysis on both using sequence data processing techniques to obtain standardized time series data. Feature vectors are extracted from the standardized time series data, and feature vector extraction techniques are used to generate a feature set containing dip angle change trends and slip hazard characteristics. A support vector machine algorithm is employed to train the feature set, and a prediction model is established through data correlation analysis to obtain dynamic assessment index prediction results. Based on the dynamic assessment index prediction results, signal strength assessment techniques are used to calculate early signal strength levels, obtaining quantified signal strength values.

[0100] For example, when acquiring dip angle change trend sequences and historical landslide hazard data, raw data can be collected from geological monitoring points via a sensor network. Dip angle change trend sequences typically come from inclinometers installed on slopes, recording dip angle changes every 12 hours, ranging from 0.05 to 0.25 degrees per meter. Historical landslide hazard data is extracted from geological archives, including the number of landslide events and displacement amounts recorded over the past 5 years; for example, a region might have recorded 3 landslide events, each with a displacement ranging from 10 to 50 centimeters. These data are aligned using timestamps to form a unified time-series dataset.

[0101] In one possible implementation, standardized time series data can be processed using the z-score standardization method. The mean is subtracted from both the tilt change and slip displacement values, and then divided by the standard deviation to obtain a standardized series with a mean of 0 and a standard deviation of 1.

[0102] For example, after standardization, the 30-day tilt change sequence of a monitoring point is converted from 0.05 to 0.25 degrees per meter to -2 to 2. The slip displacement sequence is also standardized, with the original values ​​of 10 to 50 centimeters converted to -1.5 to 1.5. This standardization facilitates subsequent feature extraction and model training by eliminating dimensional differences.

[0103] For example, principal component analysis (PCA) can be used to extract key features from standardized time series. For dip angle change series, extracted features include the series mean, maximum rate of change, and periodic fluctuation frequency. For instance, the maximum rate of change for a series might be 0.2 degrees per meter per day, and the periodic fluctuation frequency might be 0.02 Hz. For slip hazard data, extracted features include the frequency of slip events and the average displacement. For instance, the frequency of slip events in a certain area might be 0.6 times per year, and the average displacement might be 30 centimeters. These features are combined into a feature set containing a 10-dimensional vector, covering multiple aspects of dip angle and slip.

[0104] In one possible implementation, the Support Vector Machine (SVM) algorithm is used to train the feature set and build a predictive model. The training data contains 500 sets of feature vectors, each corresponding to the tilt and slip features of a monitoring point. The model maps the features to a high-dimensional space using a kernel function to optimize the classification boundary.

[0105] For example, the model can distinguish between high-risk and low-risk areas and output dynamic assessment indicators, such as risk probability values ​​between 0 and 1. A predicted risk probability of 0.8 for a certain monitoring point indicates a high-risk tendency. This model can capture the correlation between dip angle changes and potential slippage hazards.

[0106] For example, signal strength assessment techniques can calculate early signal strength levels using a weighted average method. A quantified value is then calculated by combining the risk probability output by the prediction model with the rate of change characteristics within the feature set.

[0107] For example, if the risk probability of a monitoring point is 0.8, the dip angle change rate is 0.18 degrees per meter per day, and the weighted signal strength value is 0.75, it indicates a relatively strong early warning signal. This quantitative value facilitates intuitive judgment of potential geological risks.

[0108] In one possible implementation, the generated early signal strength values ​​can be combined with a geological parameter database to verify the reliability of the prediction results.

[0109] For example, the soil shear strength at a high-risk monitoring point is 22 MPa, which is below the safety threshold of 25 MPa. This, along with the signal strength value of 0.75, supports the high-risk assessment. These results can be stored as a geographic information system (GIS) layer for automated monitoring and early warning, significantly improving the efficiency of geological disaster prevention and control.

[0110] S107. Based on the early signal strength level, integrate the stability analysis model to generate a slip hazard probability distribution map and risk identification report.

[0111] Early signal strength data is acquired by collecting environmental parameters through a sensor network to obtain the raw dataset. A pre-defined stability analysis model is used to extract features from the raw dataset to determine the signal feature set. If the fluctuation values ​​in the signal feature set exceed a preset threshold, a support vector machine algorithm is used for classification to identify potential slippage hazards. Based on the classification results, a probability density function is used to calculate the probability distribution of slippage hazards, obtaining probability distribution data. Using this probability distribution data, a Bayesian network is used to analyze the spatial distribution characteristics of the hazards and determine the risk area division. The risk area division results are obtained and visualized using a geographic information system to generate a slippage hazard probability distribution map. For high-risk areas in the distribution map, cluster analysis is used to prioritize them, obtaining the final risk identification result.

[0112] Specifically, based on early signal strength levels, sensor data such as vibration frequency (0.5 Hz) and displacement (1.2 mm) were collected. First, a Gaussian mixture model algorithm was used to cluster the signal, dividing the data into stable and unstable groups. The stable group had a mean of 0.3 Hz and a variance of 0.1, while the unstable group had a mean of 1.0 Hz and a variance of 0.5. The posterior probability was calculated using the expectation-maximization algorithm iterated 10 times, yielding an initial stability index of 0.75, indicating that the overall system is in a medium-risk state. Subsequently, the stability analysis model was integrated, and a Monte Carlo simulation method was used to generate 10,000 random paths. Input parameters included a soil friction coefficient of 0.8 and a slope angle of 15 degrees. The simulation results, integrating signal strength and early signal strength levels, were used as the foundational data for evaluating the stability analysis model. The stability factor was shown to have an average value of 0.82 and a standard deviation of 0.15. Combined with a Bayesian network to update the prior probability of 0.6, the posterior risk probability was derived to be 0.35. Based on this, a risk identification report was generated. The training dataset for the support vector machine classifier contained 500 historical samples, achieving an accuracy of 92.5%. The report quantified the risk level as medium, estimated the potential loss at 5 million yuan, and recommended increasing the monitoring frequency to once per hour. Finally, the probability distribution map of the slippage hazard was determined. The simulation output was smoothed using the kernel density estimation method with a bandwidth of 0.05, and the probability density function graph was plotted. The peak appeared at a probability of 0.4. The cumulative distribution function showed that the area of ​​the hazard with a probability of more than 0.5 accounted for 25%. A high-risk area was defined by a threshold of 0.3, which assisted in decision-making and deployment of reinforcement measures such as anchor bolt spacing of 1 meter. The entire process formed a closed-loop thinking chain from data collection to visualization and prediction, ensuring early warning and prevention of slippage hazards.

[0113] Example 2

[0114] like Figure 2 As shown, based on the same inventive concept, this embodiment also provides a radar-based coal mine slope slip identification system, the system comprising:

[0115] The signal acquisition module is used to acquire surface reflection signals of coal mine slopes using synthetic aperture radar algorithms. It performs initial data aggregation based on the mixed reflection characteristics of loose soil and rock layers in geologically complex areas to obtain the original radar signal dataset.

[0116] The signal processing module is used to process the signal interference of the original radar signal dataset using wavelet transform filtering to obtain a denoised radar signal dataset.

[0117] The phase analysis module is used to acquire phase difference information from the denoised radar signal dataset, calculate the millimeter-level displacement vector on the slope surface through interferometry, and determine the distribution map of micro-deformation information.

[0118] The feature classification module is used to classify deformation feature types by integrating complex geological parameters with micro-deformation information distribution maps and using the support vector machine algorithm to determine which deformation features belong to dip angle anomalies.

[0119] The sequence extraction module is used to extract time series data from the micro-deformation information distribution map if the tilt angle change abnormally exceeds a preset threshold, so as to obtain the tilt angle change trend sequence.

[0120] The prediction and assessment module is used to predict dynamic assessment indicators and determine the early signal strength level by comparing the dip angle change trend sequence with historical slip hazard data and using the support vector machine algorithm.

[0121] The report generation module is used to integrate the stability analysis model based on the early signal strength level to generate a slip hazard probability distribution map and risk identification report.

[0122] Example 3

[0123] This embodiment also discloses a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the method described in Embodiment 1.

[0124] The above are merely preferred embodiments 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 technical scope 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.

Claims

1. A method for identifying coal mine slope slippage based on radar monitoring, characterized in that, Includes the following steps: By acquiring surface reflection signals of coal mine slopes using synthetic aperture radar, and performing initial data aggregation based on the mixed reflection characteristics of loose soil and rock layers in geologically complex areas, the original radar signal dataset is obtained. Wavelet transform filtering is used to process the signal interference of the original radar signal dataset to obtain a denoised radar signal dataset. The phase difference information in the denoised radar signal dataset is obtained, and the millimeter-level displacement vector on the slope surface is calculated by interferometry to determine the distribution map of micro-deformation information. To address the integration of micro-deformation information distribution maps with complex geological parameters, a support vector machine algorithm is employed to classify deformation feature types and determine which deformation features belong to dip angle anomalies. If the change in tilt angle exceeds the preset threshold abnormally, time series data is extracted from the micro-deformation information distribution map to obtain the tilt angle change trend sequence; By comparing the trend sequence of tilt angle changes with historical slip hazard data, the support vector machine algorithm is used to predict dynamic evaluation indicators and determine the early signal strength level. Based on the early signal strength level, an integrated stability analysis model is used to generate a probability distribution map of slippage hazards and a risk identification report.

2. The method according to claim 1, characterized in that, The original radar signal dataset contains original echo datasets containing information about geologically complex areas.

3. The method according to claim 1, characterized in that, The process of obtaining a denoised radar signal dataset includes: Wavelet transform is used to decompose the original radar signal dataset, separating the low-frequency component and the high-frequency noise component to obtain the denoised radar signal dataset.

4. The method according to claim 1, characterized in that, The process of determining the distribution map of micro-deformation information includes: The raw phase data from the radar signal dataset is obtained, and the denoised phase data is obtained by signal denoising processing. The phase difference data is calculated from the denoised phase data using interferometry to obtain the slope surface displacement vector; Based on the slope surface displacement vector and combined with the terrain surface features, displacement field data with millimeter-level accuracy is generated. If the displacement vector in the displacement field data exceeds a preset threshold, it is marked as a micro-deformation region, and micro-deformation distribution data is obtained. A spatial interpolation algorithm is used to generate a high-resolution micro-deformation information distribution map from micro-deformation distribution data.

5. The method according to claim 1, characterized in that, The process of determining which deformation features belong to anomalies in dip angle includes: We acquire micro-deformation information distribution map data and geological parameter data, and use standardized processing to generate a fused dataset in a unified format. The fused dataset is classified using the support vector machine algorithm to obtain the deformed feature types. Features related to tilt angle changes are extracted from deformation feature types, and a preset threshold is used to determine abnormal tilt angle changes.

6. The method according to claim 1, characterized in that, The sequence of trends in tilt angle change was obtained, including: If the tilt angle change abnormally exceeds the preset threshold, the micro-deformation information distribution map is marked with spatiotemporal correlation to determine the correspondence between the elements in the map and the time dimension. Based on the location of abnormal regions, spatial units related to abnormal tilt angle changes are selected from the distribution map; Extract the tilt angle monitoring values ​​of the selected spatial units at each time point in chronological order; The extracted discrete monitoring values ​​are serialized and organized to form a continuous trend sequence of tilt angle changes.

7. The method according to claim 1, characterized in that, The process of determining early signal strength levels includes: Obtain the dip angle change trend sequence and historical slip hazard data, and perform time series analysis on both using sequence data processing technology to obtain standardized time series data; Feature vectors are extracted from standardized time series data, and feature vector extraction techniques are used to generate a feature set containing dip angle change trends and slip hazard characteristics. The feature set is trained using the support vector machine algorithm, and a prediction model is established through data correlation analysis to obtain the prediction results of dynamic evaluation indicators. Based on the prediction results of dynamic evaluation indicators, the early signal strength level is calculated using signal strength evaluation technology to obtain the quantified value of signal strength.

8. The method according to claim 1, characterized in that, Generate a probability distribution map of slippage hazards and a risk identification report, including: Early signal strength data is acquired by collecting environmental parameters through a sensor network to obtain the raw dataset; A pre-defined stability analysis model is used to extract features from the original dataset to determine the signal feature set; If the fluctuation value in the signal feature set exceeds the preset threshold, the support vector machine algorithm is used for classification to determine potential slip risks. Based on the classification results, the probability distribution of slippage hazards is calculated using the probability density function to obtain probability distribution data; By using probability distribution data, Bayesian networks are employed to analyze the spatial distribution characteristics of potential hazards and determine the risk area delineation. Obtain the risk zone delineation results, combine them with a geographic information system for visualization processing, and generate a probability distribution map of landslide hazards; For high-risk areas in the distribution map, cluster analysis is used to prioritize them and obtain the final risk identification results.

9. A coal mine slope slip identification system based on radar monitoring, characterized in that, The system for implementing the method according to any one of claims 1-8 comprises: The signal acquisition module is used to acquire surface reflection signals of coal mine slopes using synthetic aperture radar algorithms. It performs initial data aggregation based on the mixed reflection characteristics of loose soil and rock layers in geologically complex areas to obtain the original radar signal dataset. The signal processing module is used to process the signal interference of the original radar signal dataset using wavelet transform filtering to obtain a denoised radar signal dataset. The phase analysis module is used to acquire phase difference information from the denoised radar signal dataset, calculate the millimeter-level displacement vector on the slope surface through interferometry, and determine the distribution map of micro-deformation information. The feature classification module is used to classify deformation feature types by integrating complex geological parameters with micro-deformation information distribution maps and using the support vector machine algorithm to determine which deformation features belong to dip angle anomalies. The sequence extraction module is used to extract time series data from the micro-deformation information distribution map if the tilt angle change abnormally exceeds a preset threshold, so as to obtain the tilt angle change trend sequence. The prediction and assessment module is used to predict dynamic assessment indicators and determine the early signal strength level by comparing the dip angle change trend sequence with historical slip hazard data and using the support vector machine algorithm. The report generation module is used to integrate the stability analysis model based on the early signal strength level to generate a slip hazard probability distribution map and risk identification report.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the method according to any one of claims 1-8.

Images