Mine monitoring data preprocessing method and device, storage medium and computer equipment

Abstract

The application discloses a kind of mine monitoring data preprocessing method, device, storage medium and computer equipment, it is related to mine monitoring technical field, mainly in can improve the processing accuracy of mine monitoring data. Including: determining current data denoising window length, based on current data denoising window length to current to-be-processed mine monitoring data is carried out smooth processing, after the smooth processing current to-be-processed mine monitoring data is as current smooth mine monitoring data;Based on the variance of the smooth mine monitoring data in current data denoising window length, dynamically determine the current measurement noise matrix of current smooth mine monitoring data, and determine the process noise matrix corresponding to the to-be-processed mine monitoring data;Based on process noise matrix, determine the predicted initial mine monitoring data corresponding to current smooth mine monitoring data, based on current measurement noise matrix, to predicted initial mine monitoring data is corrected.The application is mainly used in mine data processing scene.

Description

Technical Field

[0001] This invention relates to the field of mine monitoring technology, and in particular to a method, apparatus, storage medium and computer equipment for mine monitoring data preprocessing. Background Technology

[0002] Currently, digital mines are rapidly evolving towards intelligent and unmanned operations, relying on technologies such as the Internet of Things (IoT), 5G, and artificial intelligence to build a closed-loop data system across the entire process. Through intelligent equipment, such as rail-guided unmanned transportation, ventilation systems, automated drilling rigs, and the application of digital twins and cloud platforms, mines are gradually achieving safe and efficient mining, optimized resource allocation, and green and low-carbon operations. Digital mines rely on monitoring data to support decision-making, covering geological displacement, equipment status, and environmental parameters. By analyzing monitoring data, disaster early warning, energy efficiency optimization, and automated control are achieved, ultimately forming an intelligent closed loop of "perception-analysis-decision-making." Therefore, to ensure safe production in mines and improve monitoring accuracy and decision-making reliability, preprocessing of mine monitoring data is necessary.

[0003] Currently, the average value of mine monitoring data is typically used to replace the maximum and minimum values ​​to preprocess the data. However, this mean replacement method does not consider the actual noise information in the data, resulting in low accuracy in the preprocessing of mine monitoring data. Summary of the Invention

[0004] This invention provides a method, apparatus, storage medium, and computer equipment for preprocessing mine monitoring data, which mainly improves the accuracy of mine monitoring data preprocessing.

[0005] According to a first aspect of the present invention, a method for preprocessing mine monitoring data is provided, comprising:

[0006] Acquire mine monitoring data to be processed;

[0007] Determine the current data denoising window length, and perform smoothing processing on the current mine monitoring data to be processed in the mine monitoring data to be processed based on the current data denoising window length. Use the smoothed current mine monitoring data to be processed as the current smoothed mine monitoring data.

[0008] Based on the variance of the smoothed mine monitoring data within the current data denoising window length, the current measurement noise matrix of the current smoothed mine monitoring data is dynamically determined, and the process noise matrix corresponding to the mine monitoring data to be processed is determined.

[0009] Based on the process noise matrix, the predicted initial mine monitoring data corresponding to the current smoothed mine monitoring data is determined, and based on the current measurement noise matrix, the predicted initial mine monitoring data is corrected, and the corrected predicted initial mine monitoring data is used as the preprocessed current mine monitoring data to be processed.

[0010] Optionally, the step of correcting the predicted initial mine monitoring data based on the current measurement noise matrix includes:

[0011] Determine the previous error covariance matrix of the previous smoothed mine monitoring data corresponding to the current smoothed mine monitoring data, and determine the current error covariance matrix of the predicted initial mine monitoring data based on the process noise matrix and the previous error covariance matrix;

[0012] Determine the measurement matrix of the current smoothed mine monitoring data, and based on the measurement matrix, the current measurement noise matrix, and the current error covariance matrix, determine the current correction coefficient for predicting the initial mine monitoring data;

[0013] Based on the current smoothed mine monitoring data and the current correction coefficient, the predicted initial mine monitoring data are corrected.

[0014] Optionally, the step of correcting the predicted initial mine monitoring data based on the current smoothed mine monitoring data and the current correction coefficient includes:

[0015] Determine the product between the measurement matrix and the predicted initial mine monitoring data, and use the difference between the current smoothed mine monitoring data and the product as the increment value of the predicted initial mine monitoring data;

[0016] The current correction factor is used to correct the increase value, and the corrected increase value is added to the predicted initial mine monitoring data to obtain the corrected predicted initial mine monitoring data.

[0017] Optionally, after correcting the predicted initial mine monitoring data based on the current measurement noise matrix, the method further includes:

[0018] Based on the collection time of the mine monitoring data, a preset number of preprocessed historical mine monitoring data to be processed are determined corresponding to the preprocessed current mine monitoring data to be processed. The preprocessed current mine monitoring data to be processed and each preprocessed historical mine monitoring data to be processed are collectively referred to as preprocessed mine monitoring data.

[0019] The slope of any two adjacent mine monitoring data in the preprocessed mine monitoring data is determined, and the mean slope of each slope is determined. Based on the mean slope, the preprocessing effect of the current mine monitoring data to be processed is verified. The method for verifying the preprocessing effect of the current mine monitoring data to be processed based on the mean slope includes:

[0020] If the average slope is less than a preset slope threshold, the verification result of the preprocessed current mine monitoring data is deemed qualified; otherwise, the verification result of the preprocessed current mine monitoring data is deemed unqualified.

[0021] Optionally, the smoothing process of the current mine monitoring data to be processed in the mine monitoring data to be processed based on the current data denoising window length includes:

[0022] Based on the current data denoising window length, multiple mine monitoring data within the current data denoising window are determined, wherein the multiple mine monitoring data include the current mine monitoring data to be processed and the smoothed historical mine monitoring data preceding the current mine monitoring data to be processed;

[0023] Determine the mean value of the current mine monitoring data after smoothing the historical mine monitoring data and the current mine monitoring data to be processed, and replace the current mine monitoring data to be processed with the mean value of the current mine monitoring data, and use the replaced current mine monitoring data to be processed as the current smoothed mine monitoring data.

[0024] Optionally, before smoothing the current mine monitoring data in the mine monitoring data to be processed based on the current data denoising window length, the method further includes:

[0025] Take any one of the mine monitoring data to be processed as a target mine monitoring data, determine the surrounding mine monitoring data within a preset range corresponding to the target mine monitoring data, and determine the average distance between the target mine monitoring data and the surrounding mine monitoring data.

[0026] Based on the average distance, the field radius is dynamically determined, and the target data density of the mine monitoring data to be processed within the field radius corresponding to the target mine monitoring data to be processed is determined.

[0027] Determine the density distribution information of the data density corresponding to each mine monitoring data to be processed, and dynamically determine the preset density threshold based on the density distribution information;

[0028] If the target data density is less than the preset density threshold, then the target mine monitoring data to be processed is determined to be abnormal mine monitoring data, and the abnormal mine monitoring data is cleaned.

[0029] Optionally, determining the mean value of the smoothed historical mine monitoring data and the current mine monitoring data to be processed includes:

[0030] Historical mine monitoring data is determined, and based on the correlation of the historical mine monitoring data under different time delays, the autocorrelation information of the mine monitoring data to be processed is determined. Based on the distribution of the historical mine monitoring data on different frequency components, the power spectral density of the mine monitoring data to be processed is determined.

[0031] Based on the autocorrelation information and the power spectral density, the number of data points is determined. Based on the number of data points, reference mine monitoring data for mean calculation is selected from the smoothed historical mine monitoring data and the current mine monitoring data to be processed, and the mean of the reference mine monitoring data is used as the mean of the current mine monitoring data.

[0032] According to a second aspect of the present invention, a mine monitoring data preprocessing apparatus is provided, comprising:

[0033] The acquisition unit is used to acquire mine monitoring data to be processed.

[0034] A smoothing unit is used to determine the current data denoising window length, perform smoothing processing on the current mine monitoring data to be processed in the mine monitoring data to be processed based on the current data denoising window length, and use the smoothed current mine monitoring data to be processed as the current smoothed mine monitoring data.

[0035] The determining unit is used to dynamically determine the current measurement noise matrix of the current smoothed mine monitoring data based on the variance of the smoothed mine monitoring data within the current data denoising window length, and to determine the process noise matrix corresponding to the mine monitoring data to be processed.

[0036] The correction unit is used to determine the predicted initial mine monitoring data corresponding to the current smoothed mine monitoring data based on the process noise matrix, and to correct the predicted initial mine monitoring data based on the current measurement noise matrix, and to use the corrected predicted initial mine monitoring data as the preprocessed current mine monitoring data to be processed.

[0037] According to a third aspect of the present invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the above-described mine monitoring data preprocessing method.

[0038] According to a fourth aspect of the present invention, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the above-described mine monitoring data preprocessing method.

[0039] According to the present invention, a method, apparatus, storage medium, and computer equipment for preprocessing mine monitoring data, compared with the current method of simply replacing the maximum and minimum values ​​of mine monitoring data with the mean value, the present invention pre-processes the mine monitoring data based on the length of the data denoising window. This smooths short-term fluctuations in the mine monitoring data and removes high-frequency noise and random interference. Based on the variance of the smoothed mine monitoring data within the current data denoising window, a measurement noise matrix is ​​dynamically determined, and the mine monitoring data is preprocessed based on this dynamically determined measurement noise matrix. This method of dynamically determining the measurement noise matrix can adapt to the complex and ever-changing mine environment in real time, thereby more effectively eliminating noise in the data and improving the accuracy of data preprocessing. By combining smoothing and dynamic noise processing to preprocess the mine monitoring data, multi-level smoothing from coarse to fine can be achieved, thereby improving the accuracy of data preprocessing and enhancing data quality. Attached Figure Description

[0040] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0041] Figure 1 A flowchart of a mine monitoring data preprocessing method provided by an embodiment of the present invention is shown;

[0042] Figure 2 This invention provides a flowchart of another mine monitoring data preprocessing method according to an embodiment of the present invention.

[0043] Figure 3 This diagram illustrates the structure of a mine monitoring data preprocessing device according to an embodiment of the present invention.

[0044] Figure 4 This invention provides a schematic diagram of another mine monitoring data preprocessing device according to an embodiment of the invention.

[0045] Figure 5 A schematic diagram of the physical structure of a computer device provided in an embodiment of the present invention is shown. Detailed Implementation

[0046] The present invention will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the present application can be combined with each other.

[0047] Currently, the method of using the mean value of mine monitoring data to replace the maximum and minimum values ​​for preprocessing mine monitoring data does not consider the actual noise information of the data, resulting in low accuracy of mine monitoring data preprocessing.

[0048] To address the aforementioned problems, embodiments of the present invention provide a method for preprocessing mine monitoring data, such as... Figure 1 As shown, the method includes:

[0049] 101. Obtain the mine monitoring data to be processed.

[0050] The mine monitoring data to be processed includes, but is not limited to, data on geological displacement, equipment status parameters, and environmental parameters. This mine monitoring data can be acquired in real time using sensors or other devices, or it can be obtained from a database. In this embodiment of the invention, the mine monitoring data to be processed is time-series data of mine monitoring data.

[0051] 102. Determine the current data denoising window length, and perform smoothing on the current mine monitoring data to be processed based on the current data denoising window length. Use the smoothed current mine monitoring data as the current smoothed mine monitoring data.

[0052] In this embodiment of the invention, the current data denoising window length can be determined based on the smoothing requirements, data characteristics, and data type of the mine monitoring data. Data characteristics include, for example, the determined current data denoising window length is 15 data points. Simultaneously, the current data denoising window length can be dynamically adjusted based on the current mine monitoring data sequence to be processed. For example, the noise level of the current mine monitoring data sequence to be processed can be calculated, and the noise level can be reflected by the standard deviation. Then, the current data denoising window length is dynamically adjusted based on this noise level. If the noise level is higher, the current data denoising window length should be appropriately increased to retain more details in the mine monitoring data. Then, the current mine monitoring data to be processed is taken as the last mine monitoring data within the current data denoising window length. The mean value of all mine monitoring data within the current data denoising window length is calculated, and this mean value replaces the current mine monitoring data to be processed, thus obtaining the smoothed current mine monitoring data to be processed. This embodiment of the invention, by pre-smoothing the mine monitoring data based on the data denoising window length, can smooth short-term fluctuations in the mine monitoring data and remove high-frequency noise and random interference from the data.

[0053] 103. Based on the variance of the smoothed mine monitoring data within the current data denoising window length, dynamically determine the current measurement noise matrix of the current smoothed mine monitoring data, and determine the process noise matrix corresponding to the mine monitoring data to be processed.

[0054] In this embodiment of the invention, the current data denoising window includes the currently smoothed mine monitoring data and the previously smoothed historical mine monitoring data arranged before the current smoothed mine monitoring data. The variance of all smoothed mine monitoring data within the current data denoising window is determined; the variance reflects the degree of data dispersion, and a larger variance indicates greater measurement noise. Then, a variance adjustment coefficient is determined, and multiplying the variance by the variance yields the current measurement noise matrix R of the current smoothed mine monitoring data. Simultaneously, based on the characteristics of the mine environment and the characteristics of the sensors and other measuring devices collecting the mine monitoring data to be processed, a fixed process noise matrix Q is determined. This process noise matrix Q can adapt to changes in sensor accuracy. This embodiment of the invention dynamically determines the measurement noise matrix based on the variance of the smoothed mine monitoring data within the current data denoising window. This method of dynamically determining the measurement noise matrix can adapt to the complex and ever-changing mine environment in real time, thereby more effectively eliminating noise in the data and improving the accuracy of data preprocessing. This embodiment of the invention achieves accurate denoising of mine monitoring data by determining suitable measurement noise matrices for different mine monitoring data.

[0055] 104. Based on the process noise matrix, determine the predicted initial mine monitoring data corresponding to the current smoothed mine monitoring data, and correct the predicted initial mine monitoring data based on the current measurement noise matrix. Use the corrected predicted initial mine monitoring data as the preprocessed current mine monitoring data to be processed.

[0056] In this embodiment of the invention, based on the previous smoothed mine monitoring data corresponding to the current smoothed mine monitoring data (i.e., the state estimate of the smoothed mine monitoring data at the previous moment, i.e., the predicted monitoring data and the process noise matrix), the state value at the current moment is predicted, i.e., the predicted initial mine monitoring data. Then, based on the current measurement noise matrix and the current smoothed mine monitoring data, the predicted initial mine monitoring data is corrected to obtain the preprocessed current mine monitoring data to be processed. This embodiment of the invention preprocesses mine monitoring data by combining smoothing processing and dynamic noise processing, achieving multi-level smoothing from coarse to fine, thereby improving the accuracy of data preprocessing and enhancing data quality.

[0057] According to the present invention, a preprocessing method for mine monitoring data, compared with the current method of simply replacing the maximum and minimum values ​​of mine monitoring data with the mean value, this invention pre-processes the mine monitoring data based on the length of the data denoising window. This smooths short-term fluctuations in the mine monitoring data and removes high-frequency noise and random interference. Based on the variance of the smoothed mine monitoring data within the current data denoising window, a measurement noise matrix is ​​dynamically determined, and the mine monitoring data is preprocessed based on this dynamically determined measurement noise matrix. The method of dynamically determining the measurement noise matrix can adapt to the complex and ever-changing mine environment in real time, thereby more effectively eliminating noise in the data and improving the accuracy of data preprocessing. By combining smoothing and dynamic noise processing to preprocess the mine monitoring data, multi-level smoothing from coarse to fine can be achieved, thereby improving the accuracy of data preprocessing and enhancing data quality.

[0058] Furthermore, to better illustrate the above-described process of preprocessing mine monitoring data, as a refinement and extension of the above embodiments, this invention provides another method for preprocessing mine monitoring data, such as... Figure 2 As shown, the method includes:

[0059] 201. Obtain the mine monitoring data to be processed.

[0060] Specifically, the mine monitoring data of the target mine is measured in real time using sensors and other measuring devices to obtain the mine monitoring data to be processed.

[0061] 202. Determine the current data denoising window length. Based on the current data denoising window length, smooth the current mine monitoring data to be processed in the mine monitoring data to be processed. Use the smoothed current mine monitoring data to be processed as the current smoothed mine monitoring data.

[0062] In this embodiment of the invention, to avoid the resources and time consumed in analyzing abnormal data in mine monitoring data, it is first necessary to identify and process abnormal data in the mine monitoring data to be processed. Based on this, the method includes: taking any one of the mine monitoring data to be processed as a target mine monitoring data; determining the surrounding mine monitoring data within a preset range corresponding to the target mine monitoring data; determining the average distance between the target mine monitoring data and the surrounding mine monitoring data; dynamically determining the domain radius based on the average distance; determining the target data density of the mine monitoring data within the domain radius corresponding to the target mine monitoring data; determining the density distribution information of the data density corresponding to each mine monitoring data; dynamically determining a preset density threshold based on the density distribution information; determining whether the target data density is less than the preset density threshold; if so, determining the target mine monitoring data as abnormal mine monitoring data; and cleaning the abnormal mine monitoring data.

[0063] The preset range is set according to actual needs. For example, it can be a preset time range set according to the collection frequency and collection scenario of mine monitoring data.

[0064] Specifically, for example, if mine monitoring data is collected once per second, five data points before and after the target mine monitoring data can be taken as surrounding mine monitoring data. The absolute distance between the target mine monitoring data and each data point in the surrounding mine monitoring data is calculated, and the average of the distance sequences is obtained. The average distance is multiplied by a dynamic adjustment coefficient to obtain the radius of the affected area, where the dynamic adjustment coefficient can be set based on the noise level of the mine. A circular region centered on the target mine monitoring data and with the radius of the affected area is determined. The data density of the mine monitoring data within this circular region is determined. Following the above method, the data density of each mine monitoring data can be determined, along with its mean and standard deviation. A preset density threshold is then determined based on the mean and standard deviation. If the data density of the mine monitoring data is less than the preset density threshold, it is considered abnormal and removed from the mine monitoring data. If the data density is greater than or equal to the preset density threshold, it is considered normal. Since mine monitoring data is affected by factors such as geological conditions and mining progress, the overall distribution may change dynamically over time. In the process of abnormal data detection, the embodiments of the present invention can adapt to the distribution changes of mine monitoring data by dynamically determining the area radius and preset density threshold, thereby improving the accuracy of abnormal data identification.

[0065] Furthermore, it is necessary to smooth the mine monitoring data after anomaly processing. Based on this, step 202 specifically includes: determining multiple mine monitoring data within the current data denoising window based on the current data denoising window length, wherein the multiple mine monitoring data includes the current mine monitoring data to be processed and the smoothed historical mine monitoring data preceding the current mine monitoring data to be processed; determining the mean value of the current mine monitoring data of the smoothed historical mine monitoring data and the current mine monitoring data to be processed, and replacing the current mine monitoring data to be processed with the mean value of the current mine monitoring data, and using the replaced current mine monitoring data to be processed as the current smoothed mine monitoring data.

[0066] Specifically, for example, if the mine monitoring data to be processed is an ambient temperature sequence, and the temperature sensor continuously outputs the following temperature data to be processed: 25.3, 25.7, 24.9, 26.1, 25.5, 25.8, 26.3, 25.6, 25.4, 26.0. If 25.3 is taken as the current mine monitoring data to be processed, and if the current data denoising window length is 5 data points, then based on the current data denoising window length, the smoothed historical temperature data for the four consecutive historical moments before the acquisition time corresponding to 25.3 are determined to be 25.2, 26.1, 26.2, and 25.9. The average value of the smoothed historical temperature data for the four consecutive historical moments and the current mine monitoring data to be processed is determined to be 25.74, then 25.74 is the current smoothed mine monitoring data.

[0067] In another embodiment of the present invention, if the current data denoising window length is less than a preset threshold, the mean of all data within the current data denoising window length can be calculated during the smoothing process. If the current data denoising window length is greater than or equal to the preset threshold, the computational load is large during the mean calculation process. To reduce the computational load, a certain number of data points need to be selected within the current data denoising window length for mean calculation. Based on this, the method includes: determining historical mine monitoring data; determining the autocorrelation information of the mine monitoring data to be processed based on the correlation of the historical mine monitoring data under different time delays; determining the power spectral density of the mine monitoring data to be processed based on the distribution of the historical mine monitoring data on different frequency components; determining the number of data points based on the autocorrelation information and the power spectral density; selecting reference mine monitoring data for mean calculation from the smoothed historical mine monitoring data and the current mine monitoring data to be processed based on the number of data points; and using the mean of the reference mine monitoring data as the mean of the current mine monitoring data.

[0068] Specifically, historical mine monitoring data is acquired, and the similarity of the same type of data at different times within this historical monitoring data is determined as relevant information. :

[0069]

[0070] in, Here, t represents historical mine monitoring data at time t, and T represents the total time range. For different time delays, The average value of historical mine monitoring data within time period T. Historical mine monitoring data Time delay The historical mine monitoring data is used as a basis. Further, in determining the power spectral density, the historical mine monitoring data is first divided into k segments, each of length L. A Hanning window is applied to each segment, and a Discrete Fourier Transform (DFT) is calculated. Then, the power spectrum of each segment is calculated based on the DFT-transformed data. Finally, the power spectra of each segment are averaged to obtain the power spectral density. Next, the autocorrelation feature vector of the autocorrelation information and the density feature vector of the power spectral density are determined. Feature-level aggregation is performed on the autocorrelation and density feature vectors to obtain a feature aggregation vector; element-level aggregation is performed on the autocorrelation and density feature vectors to obtain an element aggregation vector; basic-level aggregation is performed on the autocorrelation and density feature vectors to obtain a basic aggregation vector; the feature aggregation vector, element aggregation vector, and basic aggregation vector are transformed to obtain a data point aggregation feature vector; the data point aggregation feature vector is then input into a preset quantity prediction model to predict the number of data points, thus obtaining the number of data points.

[0071] Specifically, in order to improve the prediction accuracy of the preset quantity prediction model, it is necessary to train and construct the preset quantity prediction model. Based on this, the method includes: constructing a preset initial quantity prediction model; obtaining a sample dataset, wherein the sample dataset includes the autocorrelation information and power spectral density of sample mine monitoring data with data point quantity labels; dividing the sample dataset into a training set and a test set, using the training set to train the preset initial quantity prediction model, and using the test set to test the trained preset initial quantity prediction model, and finally using the trained preset initial quantity prediction model that meets the test conditions as the preset quantity prediction model.

[0072] Further, the vector aggregation process is as follows: Feature-level aggregation: Multiply the elements at the same position in each vector, and concatenate each multiplication result horizontally according to the element position order to obtain an initial aggregated feature vector. Multiply the initial aggregated feature vector with a first weight coefficient determined according to actual needs to obtain a feature aggregated vector. Element-level aggregation: Multiply the elements at the same position in each vector, and assign a corresponding second weight coefficient to each multiplication result. Then, concatenate each multiplication result with assigned weight coefficients horizontally according to the element position to obtain an element aggregated vector. Basic-level aggregation: Concatenate each vector horizontally, and assign a third weight coefficient determined according to actual needs to the concatenation result to obtain a basic aggregated vector. Then, concatenate the feature aggregated vector, element aggregated vector, and basic aggregated vector horizontally to obtain a data point aggregated feature vector. Finally, directly input the data point aggregated feature vector into the preset quantity prediction model, and the preset quantity prediction model directly outputs the number of data points. The preset quantity prediction model includes an input layer, a hidden layer, and a data point prediction layer. The input layer aggregates the feature vectors of data points and inputs them into the hidden layer for feature enhancement. The data point prediction layer identifies the number of data points based on the enhanced features from the hidden layer.

[0073] Furthermore, within the current data denoising window length, a number of mine monitoring data points, equal to the number of data points, are selected as reference mine monitoring data. Then, the mean of each reference mine monitoring data point is calculated as the mean of the current mine monitoring data, and this mean is used to replace the current mine monitoring data to be processed, thereby achieving smoothing of the current mine monitoring data. This embodiment of the invention, by selecting a portion of the data within the current data denoising window length for smoothing mine monitoring data, can reduce computational resources and improve processing efficiency.

[0074] 203. Based on the variance of the smoothed mine monitoring data within the current data denoising window length, dynamically determine the current measurement noise matrix of the current smoothed mine monitoring data, and determine the process noise matrix corresponding to the mine monitoring data to be processed.

[0075] Specifically, if the mine monitoring data is in univariate form, the current measurement noise matrix is ​​the smoothed variance of the mine monitoring data within the current data denoising window. If the mine monitoring data is in multivariate form, the current measurement noise matrix is ​​the covariance matrix of the smoothed variance of the mine monitoring data. The current measurement noise matrix reflects the uncertainty between the mine monitoring data to be processed and the actual state. Simultaneously, process noise can be determined based on external interference information during the mine monitoring data acquisition process. The covariance matrix of the process noise can be used as the process noise matrix, reflecting the degree of random fluctuation of the mine monitoring data during time propagation.

[0076] 204. Based on the process noise matrix, determine the predicted initial mine monitoring data corresponding to the current smoothed mine monitoring data.

[0077] Specifically, based on the process noise matrix Q, the initial mine monitoring data are first predicted using the following formula. :

[0078]

[0079] Where A is the state transition matrix of the mine monitoring data, B represents the predicted mine monitoring data at the previous k-1 time step corresponding to the current smoothed mine monitoring data, and B is the control input matrix for the mine monitoring data. If there is no external control quantity, set B=0. This is the external control input at time k. If there is no external control input, then... =0.

[0080] 205. Determine the previous error covariance matrix of the previous smoothed mine monitoring data corresponding to the current smoothed mine monitoring data. Based on the process noise matrix and the previous error covariance matrix, determine the current error covariance matrix of the predicted initial mine monitoring data.

[0081] Among them, the current smooth mine monitoring data is the smooth mine monitoring data at the current time k, and the previous smooth mine monitoring data is the smooth mine monitoring data at the previous time k-1.

[0082] Specifically, if the previous error covariance matrix is Then, the current error covariance matrix is ​​first determined according to the following formula. :

[0083]

[0084] Where Q is the process noise matrix and A is the state transition matrix of the mine monitoring data. If the mine monitoring data changes slowly, A=1.

[0085] 206. Determine the measurement matrix of the current smoothed mine monitoring data. Based on the measurement matrix, the current measurement noise matrix, and the current error covariance matrix, determine the current correction coefficients for predicting the initial mine monitoring data.

[0086] Specifically, based on the current error covariance matrix The current correction coefficient is calculated using the following formula, along with the current measurement noise matrix R. :

[0087]

[0088] Here, H is the measurement matrix, which maps the state to the measurement space; H can be set to 1. Further, after determining the current correction coefficients for the initial mine monitoring data, the initial mine monitoring data needs to be corrected based on the current smoothed mine monitoring data and the current correction coefficients.

[0089] 207. Based on the current smoothed mine monitoring data and the current correction coefficient, the predicted initial mine monitoring data is corrected, and the corrected predicted initial mine monitoring data is used as the preprocessed current mine monitoring data to be processed.

[0090] In this embodiment of the invention, after determining the current correction coefficient for the predicted initial mine monitoring data, it is necessary to correct the predicted initial mine monitoring data based on the current smoothed mine monitoring data and the current correction coefficient. Therefore, step 207 specifically includes: determining the product between the measurement matrix and the predicted initial mine monitoring data; using the difference between the current smoothed mine monitoring data and the product as the increase value of the predicted initial mine monitoring data; correcting the increase value using the current correction coefficient; and adding the corrected increase value to the predicted initial mine monitoring data to obtain the corrected predicted initial mine monitoring data.

[0091] Specifically, the initial mine monitoring data is corrected according to the following formula:

[0092]

[0093] in, The revised initial mine monitoring data for the forecast. To predict initial mine monitoring data, H represents the current smoothed mine monitoring data, and H is the measurement matrix. The measurement matrix is ​​defined as how state variables are observed through sensor measurements. The measurement matrix is ​​determined by the mine monitoring data collected by the sensors.

[0094] Furthermore, after correcting the initial predicted mine monitoring data, the correction result needs to be verified. Therefore, the method includes: determining a preset number of pre-processed historical mine monitoring data corresponding to the pre-processed current mine monitoring data based on the data collection time; collectively referring to the pre-processed current mine monitoring data and each pre-processed historical mine monitoring data as pre-processed mine monitoring data; determining the data slope of any two adjacent mine monitoring data in the pre-processed mine monitoring data, and determining the average slope of each data slope; and verifying the pre-processing effect of the pre-processed current mine monitoring data based on the average slope. The method for verifying the pre-processing effect of the pre-processed current mine monitoring data based on the average slope includes: determining whether the average slope is less than a preset slope threshold; if so, determining that the verification result of the pre-processed current mine monitoring data is qualified; otherwise, determining that the verification result of the pre-processed current mine monitoring data is unqualified.

[0095] The preset quantity and the preset slope threshold are set according to actual needs. Specifically, if the current collection time of the mine monitoring data to be processed is time k, and the preset quantity is 4, then the preprocessed historical mine monitoring data to be processed at time k-1, k-2, k-3, and k-4 are determined respectively. Next, the data slope 'a' of the mine monitoring data corresponding to time k and time k-1, the data slope 'b' of the mine monitoring data corresponding to time k-1 and time k-2, the data slope 'c' of the mine monitoring data corresponding to time k-2 and time k-3, and the data slope 'd' of the mine monitoring data corresponding to time k-3 and time k-4 are determined. Then, the mean value of the slopes corresponding to slopes a, b, c, and d is determined. If the mean value of the slope is less than a preset slope threshold, the current mine monitoring data to be processed is determined to pass the verification. If the mean value of the slope is greater than or equal to the preset slope threshold, the current mine monitoring data to be processed is determined to fail the verification. In this case, the current mine monitoring data to be processed needs to be reprocessed through smoothing and noise reduction preprocessing until the preprocessed mine monitoring data passes the verification. This embodiment of the invention ensures the accuracy of mine monitoring data preprocessing by verifying the preprocessing results.

[0096] According to another mine monitoring data preprocessing method provided by the present invention, compared with the current method of simply replacing the maximum and minimum values ​​of mine monitoring data with the mean value during the preprocessing process, the present invention performs pre-processing smoothing on the mine monitoring data based on the length of the data denoising window. This smooths short-term fluctuations in the mine monitoring data and removes high-frequency noise and random interference from the data. Based on the variance of the smoothed mine monitoring data within the current data denoising window, the measurement noise matrix is ​​dynamically determined, and the mine monitoring data is preprocessed based on the dynamically determined measurement noise matrix. The method of dynamically determining the measurement noise matrix can adapt to the complex and ever-changing situation of the mining environment in real time, thereby more effectively eliminating noise in the data and improving the accuracy of data preprocessing. By combining smoothing and dynamic noise processing to preprocess the mine monitoring data, multi-level smoothing from coarse to fine can be achieved, thereby improving the accuracy of data preprocessing and enhancing data quality.

[0097] Furthermore, as Figure 1 In specific implementation, embodiments of the present invention provide a mine monitoring data preprocessing device, such as... Figure 3 As shown, the device includes: an acquisition unit 31, a smoothing unit 32, a determination unit 33, and a correction unit 34.

[0098] The acquisition unit 31 can be used to acquire mine monitoring data to be processed.

[0099] The smoothing unit 32 can be used to determine the current data denoising window length, perform smoothing processing on the current mine monitoring data to be processed in the mine monitoring data to be processed based on the current data denoising window length, and use the smoothed current mine monitoring data to be processed as the current smoothed mine monitoring data.

[0100] The determining unit 33 can be used to dynamically determine the current measurement noise matrix of the current smoothed mine monitoring data based on the variance of the smoothed mine monitoring data within the current data denoising window length, and to determine the process noise matrix corresponding to the mine monitoring data to be processed.

[0101] The correction unit 34 can be used to determine the predicted initial mine monitoring data corresponding to the current smoothed mine monitoring data based on the process noise matrix, and to correct the predicted initial mine monitoring data based on the current measurement noise matrix, and use the corrected predicted initial mine monitoring data as the preprocessed current mine monitoring data to be processed.

[0102] In specific application scenarios, in order to correct the initial mine monitoring data for prediction, such as Figure 4 As shown, the correction unit 34 includes a first determining module 341 and a correction module 342.

[0103] The first determining module 341 can be used to determine the previous error covariance matrix of the previous smoothed mine monitoring data corresponding to the current smoothed mine monitoring data, and to determine the current error covariance matrix of the predicted initial mine monitoring data based on the process noise matrix and the previous error covariance matrix.

[0104] The first determining module 341 can also be used to determine the measurement matrix of the current smoothed mine monitoring data, and based on the measurement matrix, the current measurement noise matrix, and the current error covariance matrix, determine the current correction coefficient of the predicted initial mine monitoring data.

[0105] The correction module 342 can be used to correct the predicted initial mine monitoring data based on the current smoothed mine monitoring data and the current correction coefficient.

[0106] In specific application scenarios, in order to correct the predicted initial mine monitoring data, the correction module 342 can be used to determine the product between the measurement matrix and the predicted initial mine monitoring data, take the difference between the current smoothed mine monitoring data and the product as the increase value of the predicted initial mine monitoring data, correct the increase value using the current correction coefficient, and add the corrected increase value to the predicted initial mine monitoring data to obtain the corrected predicted initial mine monitoring data.

[0107] In specific application scenarios, in order to verify the preprocessing effect of monitoring data, the device also includes a verification unit 35.

[0108] The verification unit 35 can be used to determine, based on the acquisition time of the mine monitoring data, a preset number of preprocessed historical mine monitoring data corresponding to the preprocessed current mine monitoring data to be processed, and to collectively refer to the preprocessed current mine monitoring data and each preprocessed historical mine monitoring data as preprocessed mine monitoring data; to determine the data slope of any two adjacent mine monitoring data in the preprocessed mine monitoring data, and to determine the average slope of each data slope; and to verify the preprocessing effect of the preprocessed current mine monitoring data based on the average slope. The method for verifying the preprocessing effect of the preprocessed current mine monitoring data based on the average slope includes: determining whether the average slope is less than a preset slope threshold; if so, determining that the verification result of the preprocessed current mine monitoring data is qualified; otherwise, determining that the verification result of the preprocessed current mine monitoring data is unqualified.

[0109] In specific application scenarios, in order to smooth the current mine monitoring data to be processed, the smoothing unit 32 includes a second determining module 321 and a smoothing module 322.

[0110] The second determining module 321 can be used to determine multiple mine monitoring data within the current data denoising window based on the current data denoising window length, wherein the multiple mine monitoring data includes the current mine monitoring data to be processed and the smoothed historical mine monitoring data before the current mine monitoring data to be processed.

[0111] The smoothing module 322 can be used to determine the mean value of the current mine monitoring data after smoothing the historical mine monitoring data and the current mine monitoring data to be processed, and replace the current mine monitoring data to be processed with the mean value of the current mine monitoring data, and use the replaced current mine monitoring data to be processed as the current smoothed mine monitoring data.

[0112] In specific application scenarios, in order to identify abnormal data in the mine monitoring data to be processed, the device also includes an anomaly identification unit 36.

[0113] The anomaly identification unit 36 ​​can be used to take any one of the mine monitoring data to be processed as a target mine monitoring data, determine the surrounding mine monitoring data within a preset range corresponding to the target mine monitoring data, and determine the average distance between the target mine monitoring data and the surrounding mine monitoring data; based on the average distance, dynamically determine the area radius, determine the target data density of the mine monitoring data within the area radius corresponding to the target mine monitoring data; determine the density distribution information of the data density corresponding to each mine monitoring data, and dynamically determine a preset density threshold based on the density distribution information; determine whether the target data density is less than the preset density threshold, and if so, determine that the target mine monitoring data is an abnormal mine monitoring data, and perform cleaning processing on the abnormal mine monitoring data.

[0114] In specific application scenarios, in order to determine the mean of the current mine monitoring data, the smoothing module 322 can be used to determine historical mine monitoring data, determine the autocorrelation information of the mine monitoring data to be processed based on the correlation of the historical mine monitoring data under different time delays, determine the power spectral density of the mine monitoring data to be processed based on the distribution of the historical mine monitoring data on different frequency components, determine the number of data points based on the autocorrelation information and the power spectral density, select reference mine monitoring data for mean calculation based on the number of data points from the smoothed historical mine monitoring data and the current mine monitoring data to be processed, and use the mean of the reference mine monitoring data as the mean of the current mine monitoring data.

[0115] It should be noted that other corresponding descriptions of the functional modules involved in the mine monitoring data preprocessing device provided in this embodiment of the invention can be found in the following references. Figure 1 The corresponding description of the method shown will not be repeated here.

[0116] Based on the above, Figure 1 Accordingly, this embodiment of the invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the following steps: acquiring mine monitoring data to be processed; determining the current data denoising window length; smoothing the current mine monitoring data to be processed based on the current data denoising window length; using the smoothed current mine monitoring data as the current smoothed mine monitoring data; dynamically determining the current measurement noise matrix of the current smoothed mine monitoring data based on the variance of the smoothed mine monitoring data within the current data denoising window length; and determining the process noise matrix corresponding to the mine monitoring data to be processed; determining the predicted initial mine monitoring data corresponding to the current smoothed mine monitoring data based on the process noise matrix; correcting the predicted initial mine monitoring data based on the current measurement noise matrix; and using the corrected predicted initial mine monitoring data as the preprocessed current mine monitoring data to be processed.

[0117] Based on the above, Figure 1 The method shown and as Figure 3 The embodiment of the device shown in the invention also provides a physical structure diagram of a computer device, such as... Figure 5As shown, the computer device includes: a processor 41, a memory 42, and a computer program stored in the memory 42 and executable on the processor. Both the memory 42 and the processor 41 are mounted on a bus 43. When the processor 41 executes the program, it performs the following steps: acquiring mine monitoring data to be processed; determining the current data denoising window length; smoothing the current mine monitoring data to be processed based on the current data denoising window length; using the smoothed current mine monitoring data as the current smoothed mine monitoring data; dynamically determining the current measurement noise matrix of the current smoothed mine monitoring data based on the variance of the smoothed mine monitoring data within the current data denoising window length; and determining the process noise matrix corresponding to the mine monitoring data to be processed; determining the predicted initial mine monitoring data corresponding to the current smoothed mine monitoring data based on the process noise matrix; correcting the predicted initial mine monitoring data based on the current measurement noise matrix; and using the corrected predicted initial mine monitoring data as the preprocessed current mine monitoring data to be processed.

[0118] Through the technical solution of this invention, the present invention smooths mine monitoring data in advance based on the length of the data denoising window, which can smooth short-term fluctuations in mine monitoring data and remove high-frequency noise and random interference from the data; based on the variance of the smoothed mine monitoring data within the current data denoising window, the measurement noise matrix is ​​dynamically determined, and based on the dynamically determined measurement noise matrix, the mine monitoring data is preprocessed. The method of dynamically determining the measurement noise matrix can adapt to the complex and ever-changing situation of the mining environment in real time, thereby more effectively eliminating noise in the data and improving the accuracy of data preprocessing; by combining smoothing processing and dynamic noise processing to preprocess mine monitoring data, multi-level smoothing from coarse to fine can be achieved, thereby improving the accuracy of data preprocessing and enhancing data quality.

[0119] It is obvious to those skilled in the art that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby storing them in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those presented herein, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0120] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for preprocessing mine monitoring data, characterized in that, include: Acquire mine monitoring data to be processed; Determine the current data denoising window length, and perform smoothing processing on the current mine monitoring data to be processed in the mine monitoring data to be processed based on the current data denoising window length. Use the smoothed current mine monitoring data to be processed as the current smoothed mine monitoring data. Based on the variance of the smoothed mine monitoring data within the current data denoising window length, the current measurement noise matrix of the current smoothed mine monitoring data is dynamically determined, and the process noise matrix corresponding to the mine monitoring data to be processed is determined. Based on the process noise matrix, the predicted initial mine monitoring data corresponding to the current smoothed mine monitoring data is determined, and based on the current measurement noise matrix, the predicted initial mine monitoring data is corrected, and the corrected predicted initial mine monitoring data is used as the preprocessed current mine monitoring data to be processed. The correction of the predicted initial mine monitoring data based on the current measurement noise matrix includes: The following steps are performed: First, determine the previous error covariance matrix of the previous smoothed mine monitoring data corresponding to the current smoothed mine monitoring data. Second, determine the current error covariance matrix of the predicted initial mine monitoring data based on the process noise matrix and the previous error covariance matrix. Third, determine the measurement matrix of the current smoothed mine monitoring data. Fourth, determine the current correction coefficient of the predicted initial mine monitoring data based on the measurement matrix, the current measurement noise matrix, and the current error covariance matrix. Fifth, correct the predicted initial mine monitoring data based on the current smoothed mine monitoring data and the current correction coefficient.

2. The method according to claim 1, characterized in that, The step of correcting the initial predicted mine monitoring data based on the current smoothed mine monitoring data and the current correction coefficient includes: Determine the product between the measurement matrix and the predicted initial mine monitoring data, and use the difference between the current smoothed mine monitoring data and the product as the increment value of the predicted initial mine monitoring data; The current correction factor is used to correct the increase value, and the corrected increase value is added to the predicted initial mine monitoring data to obtain the corrected predicted initial mine monitoring data.

3. The method according to claim 1, characterized in that, After correcting the predicted initial mine monitoring data based on the current measurement noise matrix, the method further includes: Based on the collection time of the mine monitoring data, a preset number of preprocessed historical mine monitoring data to be processed are determined corresponding to the preprocessed current mine monitoring data to be processed. The preprocessed current mine monitoring data to be processed and each preprocessed historical mine monitoring data to be processed are collectively referred to as preprocessed mine monitoring data. The slope of any two adjacent mine monitoring data in the preprocessed mine monitoring data is determined, and the mean slope of each slope is determined. Based on the mean slope, the preprocessing effect of the current mine monitoring data to be processed is verified. The method for verifying the preprocessing effect of the current mine monitoring data to be processed based on the mean slope includes: If the average slope is less than a preset slope threshold, the verification result of the preprocessed current mine monitoring data is deemed qualified; otherwise, the verification result of the preprocessed current mine monitoring data is deemed unqualified.

4. The method according to claim 1, characterized in that, The smoothing process of the current mine monitoring data to be processed based on the current data denoising window length includes: Based on the current data denoising window length, multiple mine monitoring data within the current data denoising window are determined, wherein the multiple mine monitoring data include the current mine monitoring data to be processed and the smoothed historical mine monitoring data preceding the current mine monitoring data to be processed; Determine the mean value of the current mine monitoring data after smoothing the historical mine monitoring data and the current mine monitoring data to be processed, and replace the current mine monitoring data to be processed with the mean value of the current mine monitoring data, and use the replaced current mine monitoring data to be processed as the current smoothed mine monitoring data.

5. The method according to claim 1, characterized in that, Before smoothing the current mine monitoring data in the mine monitoring data to be processed based on the current data denoising window length, the method further includes: Take any one of the mine monitoring data to be processed as a target mine monitoring data, determine the surrounding mine monitoring data within a preset range corresponding to the target mine monitoring data, and determine the average distance between the target mine monitoring data and the surrounding mine monitoring data. Based on the average distance, the field radius is dynamically determined, and the target data density of the mine monitoring data to be processed within the field radius corresponding to the target mine monitoring data to be processed is determined. Determine the density distribution information of the data density corresponding to each mine monitoring data to be processed, and dynamically determine the preset density threshold based on the density distribution information; If the target data density is less than the preset density threshold, then the target mine monitoring data to be processed is determined to be abnormal mine monitoring data, and the abnormal mine monitoring data is cleaned.

6. The method according to claim 4, characterized in that, The determination of the mean value of the smoothed historical mine monitoring data and the current mine monitoring data to be processed includes: Historical mine monitoring data is determined, and based on the correlation of the historical mine monitoring data under different time delays, the autocorrelation information of the mine monitoring data to be processed is determined. Based on the distribution of the historical mine monitoring data on different frequency components, the power spectral density of the mine monitoring data to be processed is determined. Based on the autocorrelation information and the power spectral density, the number of data points is determined. Based on the number of data points, reference mine monitoring data for mean calculation is selected from the smoothed historical mine monitoring data and the current mine monitoring data to be processed, and the mean of the reference mine monitoring data is used as the mean of the current mine monitoring data.

7. A mine monitoring data preprocessing device, characterized in that, include: The acquisition unit is used to acquire mine monitoring data to be processed. A smoothing unit is used to determine the current data denoising window length, perform smoothing processing on the current mine monitoring data to be processed in the mine monitoring data to be processed based on the current data denoising window length, and use the smoothed current mine monitoring data to be processed as the current smoothed mine monitoring data. The determining unit is used to dynamically determine the current measurement noise matrix of the current smoothed mine monitoring data based on the variance of the smoothed mine monitoring data within the current data denoising window length, and to determine the process noise matrix corresponding to the mine monitoring data to be processed. The correction unit is configured to determine the predicted initial mine monitoring data corresponding to the current smoothed mine monitoring data based on the process noise matrix, and to correct the predicted initial mine monitoring data based on the current measurement noise matrix, using the corrected predicted initial mine monitoring data as the preprocessed current mine monitoring data to be processed. The correction of the predicted initial mine monitoring data based on the current measurement noise matrix includes: determining the previous error covariance matrix of the previous smoothed mine monitoring data corresponding to the current smoothed mine monitoring data; determining the current error covariance matrix of the predicted initial mine monitoring data based on the process noise matrix and the previous error covariance matrix; determining the measurement matrix of the current smoothed mine monitoring data; determining the current correction coefficient of the predicted initial mine monitoring data based on the measurement matrix, the current measurement noise matrix, and the current error covariance matrix; and correcting the predicted initial mine monitoring data based on the current smoothed mine monitoring data and the current correction coefficient.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images