Fine identification method, device and system of coal mining subsidence area and storage medium

Abstract

The application relates to a coal mining subsidence area fine identification method, device and system and a storage medium, and relates to the technical field of mine detection. The method comprises the following steps: according to image data of a target mining area within a preset monitoring time range, extracting subsidence curve features of each pixel in the image data of the target mining area; the subsidence curve features represent the change relationship between the surface subsidence amount and time; according to the subsidence curve features of each pixel in the image data of the target mining area, determining a discrimination result of each pixel; the discrimination result comprises a coal mining subsidence area or a non-coal mining subsidence area; and aggregating each pixel with the discrimination result being the coal mining subsidence area to generate a spatial distribution map of the coal mining subsidence area in the target mining area.

Description

Technical Field

[0001] This application relates to the field of mine detection technology, and in particular to a method, device, system and storage medium for the refined identification of coal mining subsidence areas. Background Technology

[0002] Coal mining triggers the redistribution of internal stress within rock strata, leading to strata movement and the formation of surface deformation basins. This results in a series of problems, such as soil erosion, damage to surface structures, ecological damage, and environmental pollution. Therefore, accurate identification of surface deformation basins in large-scale mining goaf areas is crucial for timely adjustments to underground mining plans, the implementation of effective subsidence control measures, and the protection of surface structures. Research has revealed that large-scale wide-area deformation monitoring technology, primarily using TS-InSAR / D-InSAR techniques, provides information support and data assurance for identifying coal mining-induced moving deformation basins at the mining scale. However, most current methods fail to fully consider the temporal evolution characteristics and dynamic response mechanisms of surface subsidence in coal mining subsidence areas. Furthermore, they only use InSAR results as prior knowledge, failing to effectively integrate higher-precision time-series measured data, resulting in weak model applicability and transferability. Additionally, some models rely on point cloud and UAV monitoring results, which are limited by the difficulty of rapidly covering large areas with airborne measurements, thus restricting identification capabilities. Consequently, spatial domain-based identification methods struggle to fully utilize the information contained in long-term deformation results. Summary of the Invention

[0003] This invention provides a method, apparatus, system, and storage medium for refined identification of coal mining subsidence areas, aiming to at least solve the problem of low identification accuracy of coal mining subsidence areas due to the failure to fully explore and utilize long-term deformation characteristics and dynamic response mechanisms. The technical solution of this invention is as follows: According to a first aspect of the present invention, a method for refined identification of coal mining subsidence areas is provided. The method includes: extracting subsidence curve features of each pixel in the image data of the target mining area within a preset monitoring time range; the subsidence curve features characterize the relationship between surface subsidence and time; determining the discrimination result of each pixel based on the subsidence curve features of each pixel in the image data of the target mining area; the discrimination result includes coal mining subsidence areas or non-coal mining subsidence areas; and aggregating the pixels whose discrimination result is coal mining subsidence areas to generate a spatial distribution map of coal mining subsidence areas within the target mining area.

[0004] As one implementation method, before extracting the subsidence curve features of the target mining area, the method includes: generating simulated time-series subsidence data based on the measured surface subsidence data within the historical monitoring time range of multiple mining working faces, using a subsidence time function model; extracting time-series data of areas without significant deformation in non-coal mining areas within the historical monitoring time range to obtain non-subsidence area time-series data; performing time and space normalization processing on the simulated time-series subsidence data and non-subsidence area time-series data, extracting historical subsidence values ​​and historical maximum surface subsidence amounts at multiple preset monitoring time points within the historical monitoring time range, and constructing a time-series subsidence feature dataset.

[0005] In this implementation, given the imbalance of sample data between coal mining subsidence areas and non-subsidence areas in real mining areas, a subsidence time function model is adopted. Based on the parameters of multiple measured working faces, a large number and variety of simulated subsidence time series curves are generated, increasing the richness of positive samples, making the training set achieve a more reasonable ratio of positive and negative samples, improving the model's learning sufficiency for subsidence areas, and reducing the risk of overfitting.

[0006] As one implementation method, the first penalty parameter of the time-series samples of the subsidence area and the second penalty parameter of the time-series samples of the non-subsidence area are determined in the time-series subsidence feature dataset. Based on the time-series subsidence feature dataset, the preset classification model is trained according to the first penalty parameter and the second penalty parameter to obtain the coal mining subsidence area identification model. The coal mining subsidence area identification model is used to identify whether each surface unit in the target mining area belongs to the coal mining subsidence area.

[0007] In this implementation, different penalty parameters are set for samples from subsidence areas and samples from non-subsidence areas. By assigning higher penalty parameters to samples from subsidence areas, the class imbalance problem is specifically solved, and an asymmetric configuration of misclassification costs is achieved. This significantly improves the model's sensitivity to subsidence areas while maintaining the simplicity of the overall classification and its generalization ability across mining areas.

[0008] As one implementation method, the subsidence curve features of the target mining area are extracted based on image data of the target mining area within a preset monitoring time range. This includes: acquiring image data of the target mining area within a preset monitoring time range; using small baseline set temporal interferometry to extract the temporal subsidence information of each pixel from the image data; the temporal subsidence information includes the surface subsidence amount and the maximum surface subsidence amount at each preset monitoring time point within the preset monitoring time range; normalizing the temporal subsidence information based on the total cumulative number of days within the preset monitoring time range to obtain the target temporal subsidence information; and performing temporal and spatial normalization on the target temporal subsidence information to extract the subsidence curve features of each pixel in the target mining area.

[0009] In this implementation, SBAS-InSAR technology was used to achieve high-precision deformation monitoring of coal mining subsidence areas over a large area and over a long time period, at the pixel level. The dynamic subsidence process of each pixel was fully preserved, thereby extracting the subsidence curve characteristics of each pixel. Since the curve morphology of coal mining subsidence areas differs significantly from that of non-coal mining subsidence areas, the coal mining subsidence areas in the mining area can be effectively identified based on the differences in curve morphology and subsidence curve characteristics, thus improving the identification accuracy.

[0010] As one implementation method, the target time-series settlement information is subjected to temporal and spatial normalization processing to extract the settlement curve features of each pixel in the target mining area. This includes: obtaining the time-normalized target time series based on the ratio of the cumulative settlement days at each preset monitoring time point corresponding to the target pixel to the total settlement days; obtaining the spatially normalized target settlement series based on the ratio of the settlement value at each preset monitoring time point to the maximum surface settlement corresponding to the target time series curve; and constructing the settlement curve features of the target pixel based on the target time series and the target settlement series.

[0011] In this implementation, time normalization and spatial normalization are combined to obtain a settlement characteristic curve with relative time as the abscissa and relative settlement as the ordinate. This curve completely eliminates the absolute scale differences in time and amplitude, retaining only the dynamic shape of the settlement process, thus achieving unified comparison and high-precision identification of different pixels.

[0012] As one implementation method, the discrimination result of each pixel in the image of the target mining area is determined based on historical time-series subsidence information. This includes: inputting the historical time-series subsidence information of each pixel in the image data of the target mining area into the coal mining subsidence area identification model to obtain the discrimination result of each pixel; the discrimination result indicates whether the surface corresponding to each pixel is a surface deformation unit caused by coal mining.

[0013] As one implementation method, based on the measured time-series subsidence data of the surface within the historical monitoring time range of multiple mining working faces, a subsidence time function model is adopted to generate simulated time-series subsidence data, including: the subsidence time function model includes growth curve function, Weber function, power function and logic function.

[0014] According to a second aspect of the present invention, a refined identification device for coal mining subsidence areas is provided, the device comprising: The feature extraction unit is configured to extract the subsidence curve features of the target mining area based on the image data of the target mining area within a preset monitoring time range; the subsidence curve features characterize the relationship between the amount of surface subsidence and time.

[0015] The discrimination unit is configured to determine the discrimination result of each pixel based on the subsidence curve characteristics of each pixel in the image data of the target mining area; the discrimination result includes coal mining subsidence area or non-coal mining subsidence area.

[0016] The result generation unit is configured to aggregate each pixel that is identified as a coal mining subsidence area to generate a spatial distribution map of the coal mining subsidence area within the target mining area.

[0017] According to a third aspect of the present invention, a refined identification system for coal mining subsidence areas is provided, the system being configured to perform a refined identification method for coal mining subsidence areas as described in the first aspect and any possible implementation thereof.

[0018] According to a fourth aspect of the present invention, a refined identification device for coal mining subsidence areas is provided, the device being configured to perform a refined identification method for coal mining subsidence areas as described in the first aspect and any possible implementation thereof.

[0019] According to a fifth aspect of the present invention, a computer-readable storage medium is provided, on which instructions are stored, such that when the instructions in the computer-readable storage medium are executed by a processor of a coal mining subsidence area fine identification device, the coal mining subsidence area fine identification device is able to perform a coal mining subsidence area fine identification method as described in the first aspect and any possible implementation thereof.

[0020] According to a sixth aspect of the present invention, a computer program product is provided, the computer program product including computer instructions, which, when the computer instructions are run on a coal mining subsidence area fine identification device, cause the coal mining subsidence area fine identification device to execute the coal mining subsidence area fine identification method of the first aspect and any possible implementation thereof.

[0021] The technical solution provided by the embodiments of the present invention brings at least the following beneficial effects: The present invention utilizes long-term deformation information obtained from InSAR image data, fully considering the temporal evolution characteristics and dynamic response mechanism of surface subsidence in coal mining subsidence areas. By extracting the subsidence curve features of each pixel within the target mining area, the dynamic information of the entire process from initial subsidence to gradual stabilization of the surface is completely preserved. Since there are significant differences in the temporal curves between coal mining subsidence areas and non-coal mining subsidence areas, using subsidence curve features as the discrimination criterion can effectively reduce the probability of missing subsidence areas, thereby significantly improving the identification accuracy of subsidence areas. On this basis, the present invention performs pixel-by-pixel discrimination based on the independent subsidence curve features of each pixel, and aggregates the discrimination results to generate a spatial distribution map of coal mining subsidence areas, thereby automatically identifying coal mining subsidence areas within the target mining area and achieving effective backtracking of surface subsidence areas at a large-scale mining area.

[0022] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0023] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure, and are not intended to unduly limit this disclosure.

[0024] Figure 1 This is a schematic diagram illustrating a refined identification system for coal mining subsidence areas according to an exemplary embodiment; Figure 2 This is a flowchart illustrating a refined identification method for coal mining subsidence areas according to an exemplary embodiment; Figure 3 This is a schematic diagram of a time-series characteristic settlement curve according to an exemplary embodiment; Figure 4 This is a schematic diagram of a time-series characteristic settlement curve of a coal mining subsidence area according to an exemplary embodiment; Figure 5 This is a schematic diagram illustrating the curved shape of a non-coal mining subsidence area according to an exemplary embodiment; Figure 6 This is a schematic diagram illustrating the identification results of coal mining subsidence areas based on TS-InSAR and subsidence curve features according to an exemplary embodiment; Figure 7 This is a block diagram illustrating a refined identification device for coal mining subsidence areas according to an exemplary embodiment; Figure 8 This is a schematic diagram illustrating a refined identification device for coal mining subsidence areas according to an exemplary embodiment. Detailed Implementation

[0025] To enable those skilled in the art to better understand the technical solutions of this disclosure, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings.

[0026] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0027] Before providing a detailed introduction to the refined identification method for coal mining subsidence areas provided in this application embodiment, let's briefly introduce the application scenarios and implementation environment involved in this application embodiment.

[0028] Currently, coal resources still occupy a dominant position in China's energy supply. However, coal mining triggers the redistribution of internal stress within rock strata, leading to strata movement and the formation of surface deformation basins. This causes a series of problems, such as soil erosion, damage to surface structures, ecological damage, and environmental pollution. Therefore, accurate identification of surface deformation basins in large-scale mining areas is crucial for timely adjustments to underground mining plans, the implementation of effective subsidence control measures, and the protection of surface structures.

[0029] To study the response mechanism of surface movement and deformation to coal mining, numerous surface movement observation stations have been established both domestically and internationally to monitor surface movement and deformation caused by coal mining, primarily including subsidence and horizontal movement observations. For example, the "Black Duck" surface movement observation station, established in 1955 in the Linxi mining area of ​​eastern China, obtained valuable early measured data on surface movement and deformation, significantly promoting the development of the discipline of mining subsidence. Based on high-frequency, long-term field observations and data analysis by field personnel, studies have found that the temporal subsidence curve induced by coal mining exhibits an "S"-shaped variation characteristic. This "S"-shaped curve can be divided into three stages based on a subsidence rate of 1.67 mm / d: the initial stage, the acceleration stage, and the decay stage. Based on these findings, scholars both domestically and internationally have successively proposed various subsidence time function models, such as Knothe, Weibull, Logistic, and MMF (growth curve function), as well as improved models such as Power Knothe, double Weibull, and piecewise time functions. These studies have laid an important foundation for a deeper understanding of the dynamic response mechanism of surface movement and deformation to coal mining. High-frequency, long-term observations incur high labor costs and make it difficult to acquire large-scale deformation information in a short period. Synthetic Aperture Radar Interferometry (SAR) offers large-scale monitoring over short periods with millimeter-level accuracy, providing an effective technical means for monitoring surface deformation in large mining areas. In recent years, based on satellite data from Sentinel-1A / B, Alos1 / 2, Radset-2, TerraSAR, and Lutan-1A / B, InSAR technology has been widely applied in mining area deformation monitoring. A research approach integrating space, air, and ground systems has gradually formed, using time-series monitoring technologies (SBAS, PS, DS, etc.) as the core monitoring method and prior models of the mining area (probability integral method, time function model) and field observation data as constraints.

[0030] Research has shown that TS-InSAR or D-InSAR technologies are the primary monitoring methods for wide-area deformation monitoring at large scales, providing information support and data assurance for identifying moving deformation basins caused by coal mining at the mining area scale. An automatic identification method for coal mine subsidence areas is proposed based on the Dynamic Time Warping (DTW) algorithm. This method calculates the distance between the time-series subsidence data of pixels in the InSAR monitoring dataset of the subsidence area to be identified and a threshold, identifying the subsidence area pixel by pixel, thus achieving large-scale identification of coal mine subsidence areas. A method for constructing moving deformation basins in coal mining subsidence areas based on airborne lidar monitoring data is also presented. This method constructs a deep neural network model, extracts characteristic stable regions of the subsidence area, and uses interpolation algorithms to fit the model, obtaining the complete subsidence basin. Based on Sentinel-1 SAR data, a U-Net network is constructed to identify the subsidence boundaries of coal mining subsidence areas using interferogram features. Boundary extraction is performed on low-coherence areas or areas with chaotic interferometric fringes in the interferogram. Finally, a coal mining subsidence area identification method integrating DETR and D-InSAR technologies is also presented.

[0031] However, there are still four limitations: First, most methods fail to fully consider the temporal evolution characteristics and dynamic response mechanism of surface subsidence in coal mining subsidence areas; second, they only use InSAR results as prior knowledge and fail to effectively integrate more accurate time-series measured data, resulting in generally weak applicability and transferability of the models; third, some models rely on monitoring results such as point clouds and UAVs, and their recognition ability is limited because airborne measurements cannot quickly cover large areas; fourth, spatial domain-based recognition methods cannot fully utilize the information contained in long-term deformation results.

[0032] To address the aforementioned issues, this application proposes a refined identification method for coal mining subsidence areas. By integrating measured deformation data from both subsidence and non-subsidence areas with InSAR observations, a standardized time-series subsidence curve feature dataset is constructed. This dataset is then trained using a weighted support vector machine model, ultimately achieving accurate identification of coal mining subsidence areas. This provides reliable information support and technical means for mine hazard identification, subsidence control, and disaster early warning, and also offers a scientific basis for land reclamation and environmental restoration.

[0033] Secondly, the implementation architecture involved in this application will be briefly introduced below.

[0034] Figure 1 This is a schematic diagram of a refined identification system for coal mining subsidence areas provided in this application. Figure 1 As shown, the refined identification system for coal mining subsidence areas includes a feature extraction module 11, a judgment module 12, and a result generation module 13.

[0035] The feature extraction module 11, the judgment module 12, and the result generation module 13 are connected by communication.

[0036] The feature extraction module 11 is configured to extract the subsidence curve features of the target mining area based on the image data of the target mining area within a preset monitoring time range; the subsidence curve features characterize the relationship between the amount of surface subsidence and time.

[0037] The data processing module 12 is configured to determine the discrimination result of each pixel based on the subsidence curve characteristics of each pixel in the image data of the target mining area; the discrimination result includes coal mining subsidence area or non-coal mining subsidence area.

[0038] The result visualization module 13 is configured to aggregate each pixel that is identified as a coal mining subsidence area to generate a spatial distribution map of the coal mining subsidence area within the target mining area.

[0039] For ease of understanding, the following section provides a detailed description of the refined identification method for coal mining subsidence areas provided in this application, in conjunction with the accompanying drawings.

[0040] Figure 2 This is a flowchart illustrating a refined identification method for coal mining subsidence areas according to an exemplary embodiment, such as... Figure 2 As shown, the refined identification method for coal mining subsidence areas includes the following steps.

[0041] S21. Based on the image data of the target mining area within the preset monitoring time range, extract the settlement curve features of each pixel in the image data of the target mining area.

[0042] Among them, the subsidence curve characteristics characterize the relationship between the amount of surface subsidence and time.

[0043] The image data is InSAR data.

[0044] In one implementation, image data of the target mining area within a preset monitoring time range is acquired. Using small baseline set temporal interferometry, surface temporal subsidence information for each pixel is extracted from the image data. Based on the total cumulative number of days within the preset monitoring time range, the surface temporal subsidence information is normalized to obtain the target temporal subsidence information. The target temporal subsidence information is then subjected to temporal and spatial normalization to extract the subsidence curve features of each pixel in the target mining area.

[0045] Among them, the temporal subsidence information includes the surface subsidence amount and the maximum surface subsidence amount at each preset monitoring time point within the preset monitoring time range.

[0046] Small baseline set temporal interferometry, also known as SBAS-InSAR technology.

[0047] SBAS-InSAR (Small Baseline Set Temporal Interferometry) can process multiple SAR images. By combining small baselines from multiple master images, it effectively suppresses interference such as atmospheric delay and decorrelation noise, acquiring continuous subsidence sequences for each pixel within the monitoring time range with millimeter-level accuracy. It provides spatially continuous (covering the entire mining area) and temporally dense (each image corresponds to a specific time point) surface deformation data, laying a solid data foundation for subsequent pixel-level feature extraction.

[0048] First, because InSAR data may have different total durations of subsidence sequences for each pixel due to differences in satellite revisit cycles and data start points, it is difficult to unify them by directly comparing absolute times. Furthermore, the duration of coal mining subsidence from the start to stabilization is affected by mining depth, thickness, and geological conditions, resulting in differences in mining progress. Therefore, to eliminate the influence of inconsistent observation durations between different pixels or different mining areas, the surface temporal subsidence information is time-normalized, aligning all subsidence processes to a unified relative time axis [0,1].

[0049] Specifically, the target time series after time normalization is obtained by using the ratio of the cumulative number of settlement days to the total number of settlement days at each preset monitoring time point corresponding to the target pixel. The time normalization process is specifically represented by the following formula (1).

[0050] (1).

[0051] in, The cumulative number of days for the nth observation; This represents the maximum cumulative number of days, i.e., the total cumulative number of days.

[0052] Secondly, the maximum subsidence in coal mining subsidence areas can range from a few centimeters to tens of meters, while in non-subsidence areas it may only be millimeter-level noise, a huge difference in magnitude. If the absolute value is used directly, the classifier will be biased towards the magnitude and ignore the curve shape. By comparing the subsidence value with the maximum subsidence of that pixel, the relative subsidence ratio between [0,1] can be obtained. Regardless of the total subsidence, the "completion progress" can be compared over time, highlighting the dynamic process.

[0053] Specifically, the spatially normalized target settlement sequence is obtained by comparing the settlement value at each preset monitoring time point with the maximum surface settlement corresponding to the target time series curve.

[0054] Furthermore, based on the target time series and the target settlement series, the settlement curve features of the target pixels are constructed.

[0055] like Figure 3 The time-series characteristic settlement curve accurately reflects the S-shaped curve of the coal mining subsidence area, which shows a rapid initial settlement followed by a slower, gradually stabilizing process, and the dynamic process of its settlement rate.

[0056] Thus, by combining time normalization and spatial normalization, a settlement characteristic curve is obtained with relative time as the horizontal axis and relative settlement as the vertical axis. This curve completely eliminates the absolute scale differences in time and amplitude, retaining only the dynamic shape of the settlement process.

[0057] The curve morphology of coal mining subsidence areas and non-coal mining subsidence areas is significantly different, such as... Figure 4 The subsidence curves of the coal mining subsidence area shown indicate that the subsidence at all measuring points continuously increases over time, with the isosurfaces rising layer by layer, and the subsidence accumulating over time. Figure 5 The curve morphology of the non-coal mining subsidence area shown indicates that no fixed, continuous subsidence center is formed at each measuring point; the high and low peak values ​​are randomly interspersed, and there is no spatial gradient of gradually decreasing subsidence from the center outwards. Based on this, the coal mining subsidence area in the mining area can be effectively identified according to the difference in curve morphology between coal mining subsidence areas and non-coal mining subsidence areas, thereby improving the identification accuracy.

[0058] Optionally, the number N of preset monitoring time points is determined by experimentally balancing the model identification accuracy and data collection difficulty, with N=10 being preferred; additionally, the maximum settlement is treated as an independent feature, bringing the total number of features to N+1=11.

[0059] S22. Based on the settling curve characteristics of each pixel in the image data of the target mining area, determine the discrimination result of each pixel.

[0060] The determination results include coal mining subsidence areas or non-coal mining subsidence areas.

[0061] By integrating measured deformation data from coal mining subsidence areas and non-subsidence areas with InSAR observation results, a standardized time-series subsidence curve feature dataset is constructed. This dataset is then used to train a weighted support vector machine model, enabling accurate identification of coal mining subsidence areas. Specifically, this involves the following three steps.

[0062] First, construct a time-series settlement feature dataset.

[0063] In one implementation, simulated time-series subsidence data is generated based on historical monitoring data of surface subsidence from multiple mining faces within a given timeframe. A subsidence time function model is used to generate simulated time-series subsidence data. Time-series data of non-subsidence areas without significant deformation are extracted from non-coal mining areas within the historical monitoring timeframe to obtain non-subsidence area time-series data. Temporal and spatial normalization processing is performed on the simulated time-series subsidence data and the non-subsidence area time-series data. Historical subsidence values ​​and historical maximum surface subsidence amounts at multiple preset monitoring time points within the historical monitoring timeframe are extracted to construct a time-series subsidence feature dataset.

[0064] In this implementation, since the extent of coal mining subsidence areas is relatively limited in real mining areas, while non-subsidence areas are dominant, directly using measured data would result in a significant disparity in the number of positive and negative samples, and the trained weighted SVM might still be biased towards the majority class. Therefore, a subsidence time function model is adopted, generating a large number and variety of simulated subsidence time-series curves based on parameters from multiple measured working faces. This increases the richness of positive samples, achieves a more reasonable ratio of positive to negative samples in the training set, improves the model's sufficiency in learning about subsidence areas, and reduces the risk of overfitting.

[0065] Optionally, the settling time function model includes the MMF function (growth curve function), the Weibull function, the PowerKnothe function (power function), or the Logistic function.

[0066] Temporal data of areas without significant deformation are extracted from the temporal InSAR monitoring results of non-coal mining areas as non-subsidence area datasets, i.e., non-subsidence area temporal data.

[0067] Secondly, a weighted support vector machine model is trained to construct a coal mining subsidence area identification model.

[0068] A first penalty parameter and a second penalty parameter are determined for time-series samples from subsidence areas and time-series samples from non-subsidence areas in the time-series subsidence feature dataset. Based on the time-series subsidence feature dataset, a pre-defined classification model is trained according to the first and second penalty parameters to obtain a coal mining subsidence area identification model.

[0069] Among them, the coal mining subsidence area identification model is used to identify whether each surface unit in the target mining area belongs to a coal mining subsidence area.

[0070] Weighted Support Vector Machine (SVM) is used as the classification model.

[0071] In one implementation, the constructed time-series subsidence feature dataset is used as training samples, and different penalty parameters are assigned to samples from coal mining subsidence areas and samples from non-coal mining subsidence areas to handle the class imbalance problem, thereby training a coal mining subsidence area identification model.

[0072] The objective function of the weighted support vector machine is represented by the following formula (2).

[0073] (2).

[0074] in, To mitigate structural risk, maximize the classification margin; Risk based on experience; is the normal vector of the hyperplane, which determines the orientation of the decision boundary; b is the bias term; As a slack variable, it allows some samples to be misclassified or fall within the interval.

[0075] Its constraints are specifically represented by the following formula (3).

[0076] (3).

[0077] in, Let i be the true class label of the i-th sample; is the normal vector of the hyperplane; b is the bias term; These are slack variables; This is the feature mapping function.

[0078] Based on the above weighted support vector machine model, different penalty parameters are assigned to samples from coal mining subsidence areas and samples from non-coal mining subsidence areas, as specifically represented by the following formula (4).

[0079] (4).

[0080] Where C is the basic penalty parameter; Label the sample category; The first penalty parameter for samples from coal mining subsidence areas; The second penalty parameter for samples from non-coal mining subsidence areas; This represents the number of samples from the coal mining subsidence area. This represents the number of samples from non-coal mining subsidence areas.

[0081] SSA decomposition is performed on feature point settlement data to break down complex non-stationary time series into multiple sub-components with specific trends and periodicity, thereby effectively extracting different feature information contained in the data. Thus, after model training is completed, prior InSAR observation data for the target mining area is not required for identification of the mining area.

[0082] In this implementation, the number of non-subsidence area samples is typically much greater than the number of subsidence area samples in the time-series subsidence feature dataset. If a uniform penalty parameter is used, the model will favor non-subsidence areas, resulting in a low recognition rate for subsidence areas. By setting different penalty parameters for subsidence area samples and non-subsidence area samples respectively, during model training, misclassifying a subsidence area sample incurs a greater loss, thus forcing the classification boundary to shift towards non-subsidence areas, making the model pay more attention to the feature distribution of subsidence areas. In this way, by assigning a higher penalty parameter to subsidence area samples, the class imbalance problem is specifically addressed, achieving an asymmetric configuration of misclassification costs. This significantly improves the model's sensitivity to subsidence area recognition while maintaining the simplicity of the overall classification and its generalization ability across mining areas.

[0083] Third, based on the coal mining subsidence area identification model, the subsidence curve features of each pixel in the image data of the target mining area are classified and judged to determine the judgment result of each pixel.

[0084] The discrimination result indicates whether the surface corresponding to each pixel is a surface deformation unit caused by coal mining.

[0085] Specifically, the historical time-series subsidence information of each pixel in the image data of the target mining area is input into the coal mining subsidence area identification model to obtain the discrimination result of each pixel.

[0086] S23, aggregate the pixels that are identified as coal mining subsidence areas to generate a spatial distribution map of coal mining subsidence areas within the target mining area.

[0087] like Figure 6 As shown, by determining whether a surface deformation unit is caused by coal mining on a pixel-by-pixel basis, the discrimination results of all pixels are output, and a spatial distribution map of the subsidence and movement basin in the coal mining subsidence area is obtained.

[0088] In summary, this invention can quickly and automatically identify subsidence basins in coal mining subsidence areas without relying on prior monitoring data of the target mining area, effectively tracing surface subsidence basins at a large-scale mining area. Furthermore, boundary verification based on measured surface data of the working face shows that this invention significantly improves the extraction accuracy of 10mm subsidence boundaries compared to existing methods, increasing accuracy by approximately 80%, demonstrating its practical engineering value in accurately delineating the scope of subsidence influence.

[0089] To achieve the above functions, the refined identification device for coal mining subsidence areas includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art will readily recognize that, based on the algorithmic steps of the examples described in conjunction with the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware 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.

[0090] This disclosure also provides an embodiment such as Figure 7 The coal mining subsidence area fine identification device shown includes a feature extraction unit 301, a discrimination unit 302, and a result generation unit 303.

[0091] The feature extraction unit 301 is configured to extract the subsidence curve features of the target mining area based on the image data of the target mining area within a preset monitoring time range; the subsidence curve features characterize the relationship between the amount of surface subsidence and time.

[0092] The discrimination unit 302 is configured to determine the discrimination result of each pixel based on the subsidence curve characteristics of each pixel in the image data of the target mining area; the discrimination result includes coal mining subsidence area or non-coal mining subsidence area.

[0093] The result generation unit 303 is configured to aggregate each pixel that is identified as a coal mining subsidence area to generate a spatial distribution map of the coal mining subsidence area within the target mining area.

[0094] As one implementation method, the discrimination unit 302 is specifically configured to, before extracting the subsidence curve features of the target mining area, include the following steps: based on the measured time-series subsidence data of the surface within the historical monitoring time range of multiple mining working faces, a subsidence time function model is used to generate simulated time-series subsidence data; time-series data of areas without obvious deformation in non-coal mining areas within the historical monitoring time range are extracted to obtain non-subsidence area time-series data; time and space normalization processing is performed on the simulated time-series subsidence data and the non-subsidence area time-series data; historical subsidence values ​​and historical maximum surface subsidence amounts at multiple preset monitoring time points within the historical monitoring time range are extracted to construct a time-series subsidence feature dataset.

[0095] As one implementation method, the discrimination unit 302 is specifically configured to determine the first penalty parameter of the time-series samples of the subsidence area and the second penalty parameter of the time-series samples of the non-subsidence area in the time-series subsidence feature dataset; based on the time-series subsidence feature dataset, the preset classification model is trained according to the first penalty parameter and the second penalty parameter to obtain the coal mining subsidence area identification model; the coal mining subsidence area identification model is used to identify whether each surface unit in the target mining area belongs to the coal mining subsidence area.

[0096] As one implementation method, the feature extraction unit 301 is specifically configured to extract the subsidence curve features of the target mining area based on the image data of the target mining area within a preset monitoring time range, including: acquiring image data of the target mining area within a preset monitoring time range; using small baseline set temporal interferometry to extract the temporal subsidence information of each pixel from the image data; the temporal subsidence information includes the surface subsidence amount and the maximum surface subsidence amount at each preset monitoring time point within the preset monitoring time range; normalizing the temporal subsidence information according to the total cumulative number of days within the preset monitoring time range to obtain the target temporal subsidence information; and performing temporal and spatial normalization processing on the target temporal subsidence information to extract the subsidence curve features of each pixel in the target mining area.

[0097] As one implementation method, the feature extraction unit 301 is specifically configured to perform temporal and spatial normalization processing on the target time-series settlement information, and extract the settlement curve features of each pixel in the target mining area, including: obtaining the time-normalized target time series based on the ratio of the cumulative settlement days at each preset monitoring time point corresponding to the target pixel to the total settlement days; obtaining the spatially normalized target settlement sequence based on the ratio of the settlement value at each preset monitoring time point to the maximum surface settlement corresponding to the target time series curve; and constructing the settlement curve features of the target pixel based on the target time series and the target settlement sequence.

[0098] As one implementation method, the discrimination unit 302 is specifically configured to determine the discrimination result of each pixel in the image of the target mining area based on historical time-series subsidence information, including: inputting the historical time-series subsidence information of each pixel in the image data of the target mining area into the coal mining subsidence area identification model to obtain the discrimination result of each pixel; the discrimination result characterizes whether the surface corresponding to each pixel is a surface deformation unit caused by coal mining.

[0099] Regarding the apparatus in the above embodiments, the specific manner in which each unit module performs its operations has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0100] Figure 8 This is a schematic diagram of a refined identification device for coal mining subsidence areas provided in this application. Figure 8 The refined identification device 50 for coal mining subsidence areas includes: a first processor 501, a communication bus 502, a memory 503, a communication interface 504, an output device 505, an input device 506, and a second processor 507.

[0101] The refined identification device 50 for coal mining subsidence areas may include at least one first processor 501 and a memory 503 for storing processor-executable instructions. The first processor 501 is configured to execute the instructions in the memory 503 to implement the refined identification method for coal mining subsidence areas in the following embodiments.

[0102] In addition, the refined identification equipment 50 for coal mining subsidence areas may also include a communication bus 502, at least one communication interface 504, an input device 506, and an output device 505.

[0103] The first processor 501 may be a processor (central processing unit, CPU), a microprocessor unit, an ASIC, or one or more integrated circuits for controlling the execution of programs according to the present application.

[0104] The communication bus 502 may include a path for transmitting information between the aforementioned components.

[0105] Communication interface 504 uses any transceiver-like device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0106] Input device 506 is used to receive input signals and output device 505 is used to output signals.

[0107] Memory 503 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed discs, laser discs, optical discs, digital universal discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. Memory may exist independently and be connected to the processing unit via a bus. Memory may also be integrated with the processing unit.

[0108] The memory 503 stores instructions for executing the scheme of this application, and the execution is controlled by the first processor 501. The first processor 501 executes the instructions stored in the memory 503 to realize the functions of the method of this application.

[0109] In a specific implementation, as one example, the first processor 501 may include one or more CPUs, for example... Figure 8 CPU0 and CPU1 in the CPU.

[0110] In a specific implementation, as one example, the refined identification device 50 for coal mining subsidence areas may include multiple processors, such as... Figure 8 The first processor 501 and the second processor 507 are described. Each of these processors can be a single-core processor or a multi-core processor. A processor here can refer to one or more devices, circuits, and / or processing cores used to process data (such as computer program instructions).

[0111] The sophisticated identification equipment for coal mining subsidence areas, such as Figure 8 The diagram includes a first processor 501 and a memory 503 for storing executable instructions of the first processor 501. The first processor 501 is configured to execute the executable instructions to implement the refined identification method for coal mining subsidence areas as described in any of the possible embodiments above. Since the same technical effects can be achieved, further details are omitted here to avoid repetition.

[0112] This application also provides a computer-readable storage medium. When the instructions in the computer-readable storage medium are executed by the processor of a coal mining subsidence area fine identification device, the coal mining subsidence area fine identification device can perform the coal mining subsidence area fine identification method as described in any of the above possible embodiments, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0113] This application also provides a computer program product, including a computer program or instructions, which are executed by a processor as described in any of the possible implementations above for a refined identification method for coal mining subsidence areas. Furthermore, it achieves the same technical effects, and to avoid repetition, it will not be described again here.

[0114] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the following claims.

[0115] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A method for refined identification of coal mining subsidence areas, characterized in that, The method includes: Based on the image data of the target mining area within a preset monitoring time range, the subsidence curve features of each pixel in the image data of the target mining area are extracted; the subsidence curve features characterize the relationship between the amount of surface subsidence and time. Based on the subsidence curve characteristics of each pixel in the image data of the target mining area, the discrimination result of each pixel is determined; the discrimination result includes coal mining subsidence area or non-coal mining subsidence area; The pixels of the coal mining subsidence area are aggregated according to the discrimination result to generate a spatial distribution map of the coal mining subsidence area in the target mining area.

2. The method for refined identification of coal mining subsidence areas according to claim 1, characterized in that, Before extracting the subsidence curve features of the target mining area, the method includes: Based on the measured time-series subsidence data of the surface within the historical monitoring time range of multiple mining working faces, a subsidence time function model is adopted to generate simulated time-series subsidence data; Extract time series data of areas without obvious deformation in non-coal mining areas within the historical monitoring time range to obtain time series data of non-subsidence areas; The simulated time-series settlement data and the time-series data of the non-subsidence area are subjected to time and space normalization processing. The historical settlement values ​​and the historical maximum surface subsidence at multiple preset monitoring time points within the historical monitoring time range are extracted to construct a time-series settlement feature dataset.

3. The method for refined identification of coal mining subsidence areas according to claim 2, characterized in that, The method further includes: Determine the first penalty parameter for time-series samples in the subsidence area and the second penalty parameter for time-series samples in the non-subsidence area in the time-series subsidence feature dataset, respectively. Based on the time-series subsidence feature dataset, a preset classification model is trained according to the first penalty parameter and the second penalty parameter to obtain a coal mining subsidence area identification model; the coal mining subsidence area identification model is used to identify whether each surface unit in the target mining area belongs to a coal mining subsidence area.

4. The method for refined identification of coal mining subsidence areas according to claim 1, characterized in that, The step of extracting the subsidence curve features of the target mining area based on image data of the target mining area within a preset monitoring time range includes: The image data of the target mining area within the preset monitoring time range is acquired; the temporal subsidence information of each pixel is extracted from the image data using small baseline set temporal interferometry; the temporal subsidence information includes the surface subsidence amount and the maximum surface subsidence amount at each preset monitoring time point within the preset monitoring time range; Based on the total cumulative number of days within the preset monitoring time range, the temporal subsidence information of the land surface is normalized to obtain the target temporal subsidence information; The target temporal settlement information is subjected to temporal and spatial normalization processing to extract the settlement curve features of each pixel in the target mining area.

5. The method for refined identification of coal mining subsidence areas according to claim 4, characterized in that, The step of performing temporal and spatial normalization processing on the target temporal settlement information to extract the settlement curve features of each pixel in the target mining area includes: The target time series after time normalization is obtained by the ratio of the cumulative number of settlement days at each preset monitoring time point corresponding to the target pixel to the total number of settlement days. Based on the ratio of the settlement value at each preset monitoring time point to the maximum surface settlement corresponding to the target time series curve, the spatially normalized target settlement sequence is obtained. Based on the target time series and the target settlement series, the settlement curve features of the target pixel are constructed.

6. The method for refined identification of coal mining subsidence areas according to any one of claims 2 to 4, characterized in that, The step of determining the discrimination result of each pixel in the image of the target mining area based on the historical time-series subsidence information includes: The historical temporal subsidence information of each pixel in the image data of the target mining area is input into the coal mining subsidence area identification model to obtain the discrimination result of each pixel; the discrimination result indicates whether the surface corresponding to each pixel is a surface deformation unit caused by coal mining.

7. The method for refined identification of coal mining subsidence areas according to claim 2, characterized in that, Based on historical monitoring data of surface subsidence from multiple mining faces over a given time period, simulated time-series subsidence data were generated using a subsidence time function model, including: The settling time function model includes growth curve function, Weber function, power function and logistic function.

8. A refined identification device for coal mining subsidence areas, characterized in that, The device includes: The feature extraction unit is configured to extract the subsidence curve features of the target mining area based on image data of the target mining area within a preset monitoring time range; the subsidence curve features characterize the relationship between surface subsidence and time. The discrimination unit is configured to determine the discrimination result of each pixel based on the subsidence curve characteristics of each pixel in the image data of the target mining area; the discrimination result includes coal mining subsidence area or non-coal mining subsidence area; The result generation unit is configured to aggregate each pixel that is identified as a coal mining subsidence area to generate a spatial distribution map of the coal mining subsidence area within the target mining area.

9. A refined identification system for coal mining subsidence areas, characterized in that, It is configured to perform the refined identification method for coal mining subsidence areas as described in any one of claims 1-7.

10. A computer-readable storage medium storing instructions thereon, characterized in that, When the instructions in the computer-readable storage medium are executed by the processor of the electronic device, the electronic device is able to perform the refined identification method for coal mining subsidence areas as described in any one of claims 1-7.

Images