Rock mass fracture risk early warning method, system and device based on acousto-optic data fusion and storage medium

Abstract

The invention discloses a rock mass fracture risk early warning method, system and device based on acousto-optic data fusion and a storage medium, and the method comprises the steps: collecting acoustic emission energy, acoustic emission counting, DIC data and other acousto-optic monitoring data of a rock mass test piece, and calculating to obtain an acoustic emission b value, an acoustic emission energy variance, strain field differentiation data and other multi-dimensional precursor indexes; cUSUM-Pettitt is used for cooperatively checking and recognizing mutation points, confidence and physical mutation amplitude are combined to divide multiple early warning levels, a classification model is trained through a LightGBM algorithm, and accurate early warning of the rock mass fracture risk is achieved. Multi-source physical field information is fused, the limitation of single data monitoring is overcome, the early warning reliability and timeliness are improved, and the method is suitable for disaster prevention and control in the fields of mine safety, rock mass engineering and the like.

Description

Technical Field

[0001] The present application relates to the field of rock mass monitoring technology, and in particular to a rock mass fracture risk early warning method, system, device and storage medium based on acoustic and optical data fusion. Background Art

[0002] As mining projects advance deeper underground, mine safety challenges are becoming increasingly prominent. Disasters such as water inrush, rock bursts, coal and gas outbursts, and other dynamic coal and rock disasters are becoming increasingly common. The root cause is the exacerbation of internal rock damage caused by mining activities. Rock damage is accompanied by changes in multiple physical fields, including acoustic emission, strain energy, and temperature. Scientific monitoring and early warning of early damage are crucial for disaster control and safety assurance.

[0003] Existing research uses DIC (Digital Image Correlation) or acoustic emission monitoring to analyze single-field physical data, such as optical or acoustic data, to assess rock mass fracture risk. While DIC can capture surface strain (i.e., optical data), it struggles to capture internal fractures. Acoustic emission monitoring, while able to locate internal fractures by analyzing acoustic source data, is susceptible to environmental interference.

[0004] It can be seen from this that the existing single data monitoring and early warning methods have problems such as low reliability, poor correlation, and low accuracy. Therefore, fusing multi-source physical field information to identify rock fracture precursors has great engineering significance. Summary of the Invention

[0005] In order to solve the above technical problems, the present application provides a rock fracture risk warning method, system, device and storage medium based on acoustic and optical data fusion. By fusing acoustic emission and DIC data, combined with trend analysis, mutation point detection and physical mutation amplitude calculation, accurate judgment of rock fracture risk can be achieved, thereby improving the reliability and timeliness of the warning.

[0006] The first purpose of this application is to provide a rock fracture risk warning method based on acoustic and optical data fusion.

[0007] The above-mentioned application objective of this application is achieved through the following technical solutions:

[0008] A rock fracture risk early warning method based on acoustic and optical data fusion, the method comprising the following steps:

[0009] S1, obtaining acoustic and optical monitoring data of a rock specimen taken from a target rock mass during a monitoring period through an experiment, and calculating a multi-indicator data set of multi-dimensional fracture precursor indicators reflecting a load development process of the rock mass, wherein the acoustic and optical monitoring data includes acoustic emission energy, acoustic emission counts, and DIC data;

[0010] S2, normalizing and weighting the multi-index dataset to obtain an acoustic-optical fusion dataset;

[0011] S3, using the CUSUM-Pettitt collaborative test method to identify the mutation point in the acousto-optic fusion data set, and determine the warning level of the mutation point;

[0012] S4, marking the sound and light monitoring data according to the determined warning level, and constructing a training data set;

[0013] S5, performing classification training using the LightGBM algorithm based on the training data set to obtain a trained LightGBM model;

[0014] S6, performing a rock fracture risk warning on the target rock mass based on the trained LightGBM model to obtain a warning result, wherein the warning result is used to characterize the level of rock fracture risk in the target rock mass.

[0015] Preferably, step S1 includes:

[0016] S11, performing a simulated loading test on a rock specimen collected from the target rock mass, and collecting the acoustic and optical monitoring data of the rock specimen during a monitoring period;

[0017] S12, using a fixed step-size sliding window to perform segmented calculations on the acoustic and optical monitoring data within the monitoring period to obtain a multi-indicator data set of multidimensional fracture precursor indicators reflecting the load development process of the rock mass, wherein the multidimensional fracture precursor indicators include acoustic emission b-value, acoustic emission energy variance and strain field differentiation data.

[0018] Preferably, step S12 includes:

[0019] S121, dividing the monitoring period into multiple intervals according to a preset step size and sliding window length;

[0020] S122, calculating the acoustic emission b-value of the corresponding interval based on the acoustic emission counts of each interval, calculating the acoustic emission energy variance of the corresponding interval based on the acoustic emission energy of each interval, and calculating the strain field differentiation data of the corresponding interval based on the DIC data of each interval;

[0021] S123 , combining the acoustic emission b-value, acoustic emission energy variance, and strain field differentiation data of each interval according to a time series to obtain the multi-index data set.

[0022] Preferably, step S2 includes:

[0023] S21, normalizing the acoustic emission b-value, acoustic emission energy variance, and strain field differentiation data in the multi-index data set using a deviation normalization method;

[0024] S22, using an entropy weight method to weight the normalized acoustic emission b-value, acoustic emission energy variance, and strain field differentiation data to obtain a weighted data set, which is the acoustic-optical fusion data set.

[0025] Preferably, step S3 includes:

[0026] S31, using a CUSUM algorithm to detect an acoustic-optical fusion data time series having a change trend in the acoustic-optical fusion data time series of the acoustic-optical fusion data set;

[0027] S32, using the Pettitt algorithm to perform mutation point detection on the time series of the acoustic-optical fusion data with a change trend, so as to identify the mutation point in the time series of the acoustic-optical fusion data;

[0028] S33, calculating the confidence of the mutation point;

[0029] S34, calculating the physical mutation amplitude of the mutation point;

[0030] S35 , determining the warning level of the mutation point according to a preset warning level-trigger condition correspondence table based on the confidence level and the physical mutation amplitude of the mutation point.

[0031] Preferably, in the warning level-trigger condition correspondence table, the warning levels include level one warning, level two warning and level three warning, wherein:

[0032] Level 1 warning: indicates a potential trend abnormality and requires continuous monitoring. The corresponding trigger condition is: |C q |>4σ and Pettitt value>0.05;

[0033] Level 2 warning: indicates a small mutation, and the corresponding triggering condition is: |C q |>4σ and Pettitt value ≤0.05, M ≤ θ;

[0034] Level 3 warning: indicates that a large mutation has occurred. The corresponding triggering conditions are: |C q ∣>4σ and Pettitt value ≤ 0.05, M>θ;

[0035] Among them, C q is the standardized cumulative sum, σ represents the standard deviation of the time series of the acoustic-optic fusion data in the acoustic-optic fusion dataset, the Pettitt value represents the confidence of the mutation point, M represents the physical mutation amplitude of the mutation point, θ represents the amplitude threshold for mutation, and the value of θ is related to the rock type.

[0036] Preferably, step S4 includes:

[0037] S41, marking the warning level of the sound and light monitoring data according to the determined warning level to obtain a training data set;

[0038] S42, dividing the training data set into a training set, a validation set, and a test set in a ratio of 7:2:1;

[0039] S43, using a preset sample weight adjustment algorithm to perform oversampling processing on the sample categories in the training set, validation set and test set to obtain processed training set, validation set and test set.

[0040] The second purpose of this application is to provide a rock fracture risk early warning system based on acoustic and optical data fusion.

[0041] The second object of the present application is achieved through the following technical solutions:

[0042] A rock fracture risk early warning system based on acoustic and optical data fusion, the system comprising:

[0043] A monitoring data acquisition module is used to obtain acoustic and optical monitoring data of a rock specimen taken from a target rock mass during a monitoring period through experiments, and calculate a multi-indicator data set of multi-dimensional fracture precursor indicators reflecting the load development process of the rock mass, wherein the acoustic and optical monitoring data includes acoustic emission energy, acoustic emission counts, and DIC data;

[0044] A data preprocessing module is used to normalize and weight the multi-index data set to obtain an acoustic-optical fusion data set;

[0045] An early warning level determination module is used to identify a mutation point in the acousto-optic fusion data set using a CUSUM-Pettitt collaborative test method and determine an early warning level of the mutation point;

[0046] A training data set construction module is used to mark the sound and light monitoring data according to the determined warning level and construct a training data set;

[0047] A model training module is used to perform classification training using the LightGBM algorithm based on the training data set to obtain a trained LightGBM model;

[0048] The rock mass fracture risk warning module is used to perform rock mass fracture risk warning on the target rock mass based on the trained LightGBM model to obtain a warning result, wherein the warning result is used to characterize the level of fracture risk of the rock mass in the target rock mass.

[0049] Preferably, the monitoring data acquisition module is specifically used to:

[0050] Performing a simulated loading test on a rock specimen collected from the target rock mass, and collecting the acoustic and optical monitoring data of the rock specimen within a monitoring period;

[0051] The acoustic and optical monitoring data within the monitoring period are segmented and calculated using a fixed-step sliding window to obtain a multi-indicator data set of multidimensional fracture precursor indicators reflecting the load development process of the rock mass, wherein the multidimensional fracture precursor indicators include acoustic emission b-value, acoustic emission energy variance and strain field differentiation data.

[0052] Preferably, the step of performing segmented calculations on the acoustic and optical monitoring data within the monitoring period using a fixed-step sliding window to obtain a multi-indicator dataset of multi-dimensional fracture precursor indicators reflecting the load development process of the rock mass includes:

[0053] Dividing the monitoring period into multiple intervals according to a preset step size and sliding window length;

[0054] The acoustic emission b-value of the corresponding interval is calculated based on the acoustic emission counts of each interval, the acoustic emission energy variance of the corresponding interval is calculated based on the acoustic emission energy of each interval, and the strain field differentiation data of the corresponding interval is calculated based on the DIC data of each interval;

[0055] The acoustic emission b-value, acoustic emission energy variance and strain field differentiation data of each interval are combined according to a time series to obtain the multi-index data set.

[0056] Preferably, the data preprocessing module is specifically used to:

[0057] Normalizing the acoustic emission b-value, acoustic emission energy variance, and strain field differentiation data in the multi-index data set using a deviation normalization method;

[0058] The entropy weight method is used to weight the normalized acoustic emission b-value, acoustic emission energy variance and strain field differentiation data to obtain a weighted data set, which is the acoustic-optical fusion data set.

[0059] Preferably, the warning level determination module is specifically used to:

[0060] Using a CUSUM algorithm to detect the acoustic-optical fusion data time series with a change trend in the acoustic-optical fusion data time series of the acoustic-optical fusion data set;

[0061] Using the Pettitt algorithm to perform mutation point detection on the time series of the acoustic-optical fusion data with a change trend, so as to identify the mutation points in the time series of the acoustic-optical fusion data;

[0062] Calculating the confidence of the mutation point;

[0063] Calculating the physical mutation amplitude of the mutation point;

[0064] Based on the confidence level and the physical mutation amplitude of the mutation point, the warning level of the mutation point is determined according to a preset warning level-trigger condition correspondence table.

[0065] Preferably, in the warning level-trigger condition correspondence table, the warning levels include level one warning, level two warning and level three warning, wherein:

[0066] Level 1 warning: indicates a potential trend abnormality and requires continuous monitoring. The corresponding trigger condition is: |C q |>4σ and Pettitt value>0.05;

[0067] Level 2 warning: indicates a small mutation, and the corresponding triggering condition is: |C q |>4σ and Pettitt value ≤0.05, M ≤ θ;

[0068] Level 3 warning: indicates that a large mutation has occurred. The corresponding triggering conditions are: |C q ∣>4σ and Pettitt value ≤ 0.05, M>θ;

[0069] Among them, C q is the standardized cumulative sum, σ represents the standard deviation of the time series of the acoustic-optic fusion data in the acoustic-optic fusion dataset, the Pettitt value represents the confidence of the mutation point, M represents the physical mutation amplitude of the mutation point, θ represents the amplitude threshold for mutation, and the value of θ is related to the rock type.

[0070] Preferably, the training data set construction module is specifically used to:

[0071] Marking the warning level of the sound and light monitoring data according to the determined warning level to obtain a training data set;

[0072] The training data set is divided into a training set, a validation set, and a test set in a ratio of 7:2:1;

[0073] A preset sample weight adjustment algorithm is used to perform sampling processing on the sample categories in the training set, validation set and test set to obtain processed training set, validation set and test set.

[0074] The third purpose of this application is to provide a rock fracture risk warning device based on acoustic and optical data fusion.

[0075] The third object of the present application is achieved through the following technical solutions:

[0076] A rock fracture risk early warning device based on acoustic and optical data fusion, comprising:

[0077] A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of the rock fracture risk warning method based on acoustic and optical data fusion as described in any one of the first objectives of the present application are implemented.

[0078] The fourth object of this application is to provide a computer-readable storage medium.

[0079] The fourth object of the present application is achieved through the following technical solutions:

[0080] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the rock fracture risk warning method based on acoustic and optical data fusion as described in any one of the first objectives of the present application.

[0081] The above technical solution of this application has the following beneficial effects:

[0082] 1. Fusion of acoustic emission and DIC data covers the multi-physical field characteristics of rock fracture, overcoming the limitations of single data representation and improving the comprehensiveness of early warning;

[0083] 2. The CUSUM-Pettitt synergy test combines trend accumulation with statistical verification of mutation points to reduce noise interference and improve the sensitivity of early anomaly identification;

[0084] 3. Based on the magnitude of mutation and confidence, a three-level warning system is established to provide a basis for differentiated decision-making and achieve progressive risk identification from trend anomaly to mutation rupture;

[0085] 4. The LightGBM algorithm is combined with sample weight adjustment to solve the problem of class imbalance, improve the recall rate of high-risk mutation signals, and ensure the generalization ability and real-time performance of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments recorded in this application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0087] Figure 1 This is a flow chart of a rock fracture risk early warning method based on acoustic and optical data fusion in one embodiment of the present application;

[0088] Figure 2This is a structural diagram of a rock fracture risk early warning system based on acoustic and optical data fusion in one embodiment of the present application;

[0089] Figure 3 This is a structural schematic diagram of a rock fracture risk warning device based on acoustic and optical data fusion in one embodiment of the present application. DETAILED DESCRIPTION

[0090] In order to enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be described in detail and completely below. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of this application.

[0091] In the embodiments provided in this application, it should be understood that the disclosed methods and systems can be implemented in other ways. The system embodiments described below are merely illustrative. For example, the division of units and modules is merely a logical function division. In actual implementation, there may be other division methods, such as: multiple units or modules can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or modules, which can be electrical, mechanical or other forms.

[0092] In addition, all functional units in the embodiments of the present application may be integrated into one processor, or each unit may be a separate device, or two or more units may be integrated into one device; each functional unit in the embodiments of the present application may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0093] Those skilled in the art will understand that all or part of the steps of the following method embodiments can be implemented by program instructions and related hardware. The aforementioned program instructions can be stored in a computer-readable storage medium. When the program instructions are executed, the steps of the following method embodiments are executed; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM), magnetic disks or optical disks, and other media that can store program codes.

[0094] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of technical features indicated. Thus, a feature specified as "first" or "second" may explicitly or implicitly include one or more of such features. Throughout the description of this application, "plurality" or "several" means two or more, unless otherwise specifically defined.

[0095] The present application provides a rock fracture risk warning method based on acoustic and optical data fusion, such as Figure 1 As shown, the method may include the following steps:

[0096] S1, obtaining acoustic and optical monitoring data of rock specimens taken from the target rock mass during the monitoring period through experiments, and calculating a multi-indicator data set of multi-dimensional fracture precursor indicators reflecting the load development process of the rock mass, wherein the acoustic and optical monitoring data include acoustic emission energy, acoustic emission counts and DIC data;

[0097] Specifically, the target rock mass refers to the rock mass in the target environment that requires fracture risk monitoring and early warning, and the DIC data includes full-field strain data of the rock mass specimen. The acoustic emission energy, acoustic emission counts, and DIC data in the acoustic and optical monitoring data can be collected during the compression test of the rock mass specimen taken from the target rock mass through energy monitoring equipment, acoustic emission sensors, and digital image correlation (DIC) systems, respectively. After the acoustic and optical monitoring data of the rock mass specimen during the monitoring period are obtained through experimental collection, the collected data needs to be further calculated to obtain a multi-indicator data set of multi-dimensional fracture precursor indicators reflecting the load development process of the rock mass, so that an early warning model can be constructed subsequently based on the multi-indicator data set.

[0098] S2, normalize and weight the multi-index data set to obtain the sound and light fusion data set;

[0099] In order to eliminate the influence of dimensional differences in the modeling process and improve the accuracy of the early warning results, after calculating the multi-indicator data set that reflects the multi-dimensional fracture precursor indicators of the rock mass load development process, it is necessary to perform data preprocessing operations such as normalization and weighting on the multi-indicator data set to obtain the acoustic-optical fusion data set.

[0100] S3, using the CUSUM-Pettitt collaborative test method to identify the mutation points in the acousto-optic fusion dataset and determine the warning level of the mutation points;

[0101] After normalization and weighted preprocessing of the multi-index data set, the CUSUM-Pettitt collaborative test method is used to identify the mutation points in the preprocessed acoustic-optical fusion data set, and further determine the warning level of the mutation points.

[0102] In this embodiment, the CUSUM-Pettitt collaborative test method is used to identify mutation points in the acoustic-optical fusion data set, realizing dual-driven detection of trend evolution and mutation events. CUSUM suppresses short-term noise interference, and Pettitt provides statistical verification of mutation points, reducing the probability of false alarms and improving the sensitivity of early anomaly detection.

[0103] S4, labeling the sound and light monitoring data according to the determined warning level and constructing a training data set;

[0104] Next, the sound and light monitoring data are marked according to the determined warning level, and a training data set for model training is constructed based on the sound and light monitoring data marked with the warning level.

[0105] S5, based on the training data set, use the LightGBM algorithm to perform classification training to obtain a trained LightGBM model;

[0106] After constructing the training data set, the LightGBM algorithm is used for classification training to obtain a trained LightGBM model, which is used for rock fracture risk warning in the target environment.

[0107] S6, based on the trained LightGBM model, performs rock fracture risk warning on the target rock mass to obtain a warning result, wherein the warning result is used to characterize the level of rock fracture risk in the target rock mass.

[0108] When providing rock fracture risk warning for the target rock mass based on the trained LightGBM model, the acoustic and optical monitoring data such as acoustic emission energy, acoustic emission counts and DIC data of the rock mass in the target monitoring environment are collected, and the monitored acoustic emission energy, acoustic emission counts and DIC data are input into the trained LightGBM model. The rock fracture risk warning result that can characterize the fracture risk level of the rock mass in the target rock mass is obtained, thereby realizing real-time warning of rock fracture risk.

[0109] From the above, it can be seen that the rock fracture risk warning method based on the fusion of acoustic and optical data in this embodiment first obtains the acoustic and optical monitoring data of the rock specimen taken from the target rock during the monitoring period through experiments, and calculates a multi-index data set of multi-dimensional fracture precursor indicators reflecting the load development process of the rock; then the multi-index data set is normalized and weighted to obtain an acoustic and optical fusion data set; then the CUSUM-Pettitt collaborative test method is used to identify the mutation points in the acoustic and optical fusion data set, and the warning level of the mutation point is determined; then the acoustic and optical monitoring data are marked according to the determined warning level, and a training data set is constructed; then the LightGBM algorithm is used to perform classification training based on the training data set to obtain a trained LightGBM model; finally, the target rock is warned of rock fracture risk based on the trained LightGBM model to obtain a warning result. The embodiment of the present application achieves accurate identification of rock fracture risk and improves the reliability and timeliness of the warning by fusing acoustic emission and DIC data, combining trend analysis, mutation point test and physical mutation amplitude calculation.

[0110] In one embodiment, step S1 includes:

[0111] S11, performing a simulated loading test on a rock specimen collected from the target rock mass, and collecting acoustic and optical monitoring data of the rock specimen during a monitoring period;

[0112] S12, using a fixed-step sliding window to perform segmented calculations on the acoustic and optical monitoring data within the monitoring period, to obtain a multi-indicator data set of multidimensional fracture precursor indicators reflecting the load development process of the rock mass, wherein the multidimensional fracture precursor indicators include acoustic emission b-value, acoustic emission energy variance and strain field differentiation data.

[0113] Specifically, step S12 includes:

[0114] S121, dividing the monitoring period into multiple intervals according to a preset step size and sliding window length;

[0115] S122, calculating the acoustic emission b-value of the corresponding interval based on the acoustic emission counts of each interval, calculating the acoustic emission energy variance of the corresponding interval based on the acoustic emission energy of each interval, and calculating the strain field differentiation data of the corresponding interval based on the DIC data of each interval;

[0116] S123 , combining the acoustic emission b-value, acoustic emission energy variance, and strain field differentiation data of each interval according to a time series to obtain a multi-index data set.

[0117] The dynamic response of rocks before failure can be characterized through the coordinated evolution of multiple physical fields. The activity of acoustic emission signals and the heterogeneity of strain fields can provide a multi-dimensional depiction of the evolution of damage accumulation to unstable fracture. Quantitative analysis of the intensity distribution of acoustic emission events, the temporal discreteness of parameters, and the spatial differentiation of strain fields reveals the entire process of rock fracture, from microcrack initiation and localized propagation to macroscopic fracture.

[0118] To ensure the accuracy of the subsequent early warning model, this embodiment selects appropriately correlated indicators as input features. The acoustic emission b-value reflects the energy transition characteristics of crack activity within the rock by statistically analyzing the proportional changes in acoustic emission events of varying intensities. Low b-values ​​indicate a clustering of high-intensity events, foreshadowing a critical fracture state. The acoustic emission energy variance dynamically assesses the degree of dispersion of acoustic emission parameters. Its sudden increase captures the heterogeneous evolution of microfracture activity, indicating the stage of accelerated damage development. The strain field differentiation rate quantifies the dynamic evolution rate of localized strain field regions. The spatial expansion of high-value regions is directly associated with the formation precursors of macroscopic fracture surfaces.

[0119] In addition, through sliding calculations in a fixed time window, the above parameters can simultaneously extract the coupling characteristics of acoustic force and deformation during the loading process of the rock specimen, breaking through the one-sidedness of single physical field monitoring, providing highly sensitive multi-source input parameters for subsequent models, and accurately identifying the cumulative trend of fracture precursors.

[0120] In this embodiment, acoustic emission energy, acoustic emission counts, and DIC data are obtained through simulated loading tests and preprocessed to calculate a multi-indicator data set that reflects the multidimensional fracture precursor indicators of the rock mass under load development process. Before calculating the data, a fixed step size and sliding window length are set to divide the time. The step size is set to Δt to divide the monitoring period into r intervals to ensure that each interval has sufficient data. The acoustic emission data (acoustic emission counts and emission energy) obtained in the laboratory are calculated to obtain the acoustic emission b value and acoustic emission energy variance, and the DIC data are calculated to obtain strain field differentiation data. The specific calculation process is as follows:

[0121] Assume the window length is ξ, the step size is Δt, and the time point covered by the λth window is:

[0122] t∈[λΔt,λΔt+ξ](λ=0,1,...r)

[0123] Where: r is the total number of windows;

[0124] The calculation formula of the acoustic emission b value is:

[0125] logN=a-bM

[0126] Where: N is the number of events with acoustic emission signal intensity greater than M within a certain period of time; M is a constant, representing the size or intensity of the acoustic emission signal; a is a constant, where b represents the ratio of high-intensity events to low-intensity events. The values ​​of a and b are obtained by least squares fitting.

[0127] The calculation formula of the acoustic emission energy variance is:

[0128]

[0129] Where: D represents the acoustic emission energy variance, Represents the i1th reference point of the acoustic emission energy in the system, represents the average value of the acoustic energy data set, and S represents the total number of data points monitored by the system.

[0130] The calculation formula of the strain field differentiation data is:

[0131]

[0132] Where: m is the total number of strain field calculation points, Expressed as the strain of the j1th calculation point in the calculation area at the i2th moment, is the mean strain value of the strain field sub-region, Expressed as the strain field differentiation rate at the i2th moment.

[0133] In one embodiment, step S2 includes:

[0134] S21, the deviation normalization method is used to normalize the acoustic emission b value, acoustic emission energy variance and strain field differentiation data in the multi-index data set;

[0135] S22, the entropy weight method is used to weight the normalized acoustic emission b-value, acoustic emission energy variance and strain field differentiation data to obtain a weighted data set, which is the acoustic-optical fusion data set.

[0136] In this embodiment, the entropy weight method is used to dynamically assign weights based on the discrete degree of characteristic parameters, thereby avoiding subjective weighting bias, enhancing the model's sensitivity to key precursor indicators, improving the accuracy and comprehensiveness of rupture risk warnings, and reducing false alarms and missed alarms.

[0137] Specifically, the calculation formula of the deviation standardization method is:

[0138]

[0139] Where: x' pq is the normalized sample value; x pqis the sample value, which represents the value of the qth data of the pth input feature, where p = 1, 2, 3 (i.e., the three input features of acoustic emission b value, acoustic emission energy variance and strain field differentiation data), and q = 1, 2, 3…n (n is the data point in the multi-index dataset).

[0140] The entropy weight method weighting process is as follows: first, the original data is standardized to obtain a standardized matrix:

[0141]

[0142] Where: τ pq is a standardized matrix.

[0143] Then, calculate the entropy value of the p-th indicator:

[0144]

[0145] Where: e p is the information entropy of the pth indicator; α is a constant; n is the total number of samples.

[0146] Calculate the information utility value of each indicator:

[0147] d p =1-e p

[0148] Where: d p is the information utility value of each indicator.

[0149] According to the information utility value, the weight coefficient of each indicator is determined:

[0150]

[0151] Where: w p is the weight of the p-th input feature.

[0152] The acoustic emission b value, acoustic emission energy variance and strain field differentiation data are weighted to obtain a data set:

[0153]

[0154] Wherein: X is the dataset obtained after weighting, that is, the acoustic-optical fusion dataset.

[0155] In one embodiment, step S3 includes:

[0156] S31, using the CUSUM algorithm to detect the acoustic-optical fusion data time series of the acoustic-optical fusion data set with a change trend;

[0157] S32, using the Pettitt algorithm to perform mutation point detection on the time series of the acoustic-optical fusion data with a changing trend, so as to identify the mutation points in the time series of the acoustic-optical fusion data;

[0158] S33, calculate the confidence of the mutation point;

[0159] S34, calculating the physical mutation amplitude of the mutation point;

[0160] S35 , based on the confidence level and the physical mutation amplitude of the mutation point, determine the warning level of the mutation point according to a preset warning level-trigger condition correspondence table.

[0161] Specifically, in the warning level-trigger condition correspondence table, the warning levels include level one warning, level two warning, and level three warning, among which,

[0162] Level 1 warning: indicates a potential trend abnormality and requires continuous monitoring. The corresponding trigger condition is: |C q |>4σ and Pettitt value>0.05;

[0163] Level 2 warning: indicates a small mutation, and the corresponding triggering condition is: |C q |>4σ and Pettitt value ≤0.05, M ≤ θ;

[0164] Level 3 warning: indicates that a large mutation has occurred. The corresponding triggering conditions are: |C q |>4σ and Pettitt value ≤0.05, M>0;

[0165] Among them, C q is the standardized cumulative sum, σ represents the standard deviation of the time series of the acoustic-optic fusion data in the acoustic-optic fusion dataset, the Pettitt value represents the confidence of the mutation point, M represents the physical mutation amplitude of the mutation point, and θ represents the amplitude threshold for mutation. The value of θ is related to the rock mass type.

[0166] In this embodiment, the CUSUM algorithm is used to detect the long-term trend of the multi-source monitoring data in the obtained acoustic-optical fusion dataset to discover the trend of change. The Pettitt test is performed on the data with a trend of change to identify the mutation points in the time series. The specific detection process is as follows:

[0167] Calculate the mean μ and standard deviation σ of the data in the preprocessed acousto-optic fusion dataset:

[0168]

[0169] Calculate the standardized CUSUM statistic (i.e. cumulative sum):

[0170]

[0171] Where: C q is the standardized cumulative sum, representing the cumulative standardized deviation of the time series.

[0172] Set the control limit h, and detect and calculate the cumulative sum C through the CUSUM algorithm q Compare it with the control limit h to determine the significance of the change trend, where the calculation formula of h is:

[0173] h = 4σ

[0174] If |C q | > h, it indicates that the data has a significant change trend;

[0175] If |C q | ≤ h, it indicates that the data has no obvious trend change.

[0176] The inspection process of the Pettitt algorithm is as follows:

[0177] Inspect the time series dataset X' in the acousto-optic fusion dataset with a change trend to check if there is a mutation point τ (1 ≤ τ < n). The specific process is as follows:

[0178] Calculate the statistic U for each point in the time series dataset with a change trend t :

[0179]

[0180] Then solve the optimal position K of the mutation point t :

[0181] K t = max|U t |

[0182]

[0183] Where: K t is the optimal position of the mutation point, and the corresponding t* is used as the estimated value of the mutation point.

[0184] After determining the mutation point, verify it and approximately calculate the confidence level of the mutation point

[0185]

[0186] If then it is considered that there is a mutation point at t*.

[0187] If then it is considered that there is no mutation.

[0188] In order to further determine the rupture risk level, the standardized mutation amplitude is used to give early warning of the rupture scale:

[0189]

[0190] Where: M is the physical mutation amplitude, μ pre is the mean before the mutation point, μ post Mean after mutation point, σ global Global standard deviation.

[0191] If M≤θ, there is a significant small mutation; if M>θ, there is a significant large mutation. The value of θ is adjusted according to the existing research in this field and the corresponding relationship between common rock types and θ values ​​is shown in the following table:

[0192] rock mass type θ value range Applicable Scenarios granite 1.2-1.5 Hard rock tunnels, mine tunnels sandstone 0.8-1.0 Oil and gas reservoirs, slope engineering shale 0.5-0.7 Shale gas extraction, weak surrounding rock

[0193] Specifically, in this embodiment, a three-level warning discriminant function is designed. Based on the combined criteria of trend accumulation, physical mutation amplitude and mutation point confidence, a progressive risk warning chain is formed to provide a differentiated decision-making basis for engineering safety.

[0194] In one embodiment, step S4 includes:

[0195] S41, marking the warning level of the sound and light monitoring data according to the determined warning level to obtain a training data set;

[0196] Specifically, the data for which no warning was issued is marked as 0, and the rest are marked as 1, 2, and 3 according to the warning levels from level 1 to level 3.

[0197] S42, divide the training data set into training set, validation set and test set according to the ratio of 7:2:1;

[0198] The labeled dataset is divided into training set, validation set and test set with a ratio of 7:2:1. Define the input feature matrix X and label vector Y:

[0199] X=[x1,x2,...,x n ] T ,Y=[y1,y2,...,y n ] T

[0200] Where: y i ∈{0,1,2,3} represents the warning level.

[0201] S43, using a preset sample weight adjustment algorithm to perform oversampling processing on the sample categories in the training set, validation set and test set to obtain processed training set, validation set and test set.

[0202] Since normal samples far outnumber warning samples, the model may tend to predict the majority class, resulting in missed high-risk events. Therefore, a sample weight adjustment algorithm is needed to balance the classes or oversample them.

[0203] In this embodiment, there is no limitation on the specific algorithm of the sample weight adjustment algorithm.

[0204] Specifically, in step S5, the specific process of classification training using the LightGBM algorithm based on the training data set is as follows:

[0205] 1. Initialize the LightGBM model: Use the LGBMClassifier or LGBMRegressor class in the lightgbm library to initialize the LightGBM model and set the corresponding parameters, such as objective, metric, boosting_type, etc.

[0206] 2. Training model: Use the training set in the training set data to call the fit method to train the LightGBM model. During the training process, you can use the validation set data in the training set data to evaluate the performance and adjust the model parameters based on the evaluation results. You can use the early_stopping_rounds parameter to control the early stopping strategy to avoid overfitting;

[0207] 3. Model evaluation: After training is completed, the generalization ability of the LightGBM model is evaluated using the test set data in the training set data. Various evaluation indicators can be calculated, such as precision, recall, weighted average F1 value, geometric mean score and Matthews correlation coefficient;

[0208] 4. Model saving and loading: The trained LightGBM model can be saved to disk using the lgb.save_model function, and then loaded using the lgb.load_model function for prediction.

[0209] In step S6, when the target rock mass is given a rock fracture risk warning based on the trained LightGBM model, the acoustic emission energy, acoustic emission counts and DIC data obtained by monitoring the target rock mass are used as monitoring data X new , input the trained LightGBM model:

[0210]

[0211] Where: Indicates the output warning level, which has four types: 0, 1, 2, and 3. 0 indicates normal, 1 corresponds to the first-level warning, indicating an abnormal trend, 2 corresponds to the second-level warning, indicating a small mutation, and 3 corresponds to the third-level warning, indicating a large mutation. c is the category label of the warning, and the value range of c is 0, 1, 2, and 3. k represents the kth classifier, K represents the total number of classifiers, and f k (·) represents the output of the k-th classifier.

[0212] like Figure 2 As shown, the embodiment of the present application also provides a rock fracture risk early warning system based on acoustic and optical data fusion, which may include:

[0213] The monitoring data acquisition module 201 is used to obtain acoustic and optical monitoring data of a rock specimen taken from a target rock mass during a monitoring period through experiments, and calculate a multi-indicator data set of multi-dimensional fracture precursor indicators reflecting the load development process of the rock mass, wherein the acoustic and optical monitoring data includes acoustic emission energy, acoustic emission counts, and DIC data;

[0214] The data preprocessing module 202 is used to normalize and weight the multi-index data set to obtain an acoustic-optical fusion data set;

[0215] The warning level determination module 203 is used to identify the mutation point in the acousto-optic fusion data set using the CUSUM-Pettitt collaborative test method and determine the warning level of the mutation point;

[0216] A training data set construction module 204 is used to mark the sound and light monitoring data according to the determined warning level and construct a training data set;

[0217] The model training module 205 is used to perform classification training using the LightGBM algorithm based on the training data set to obtain a trained LightGBM model;

[0218] The rock mass fracture risk warning module 206 is used to perform rock mass fracture risk warning on the target rock mass based on the trained LightGBM model to obtain a warning result, wherein the warning result is used to characterize the level of fracture risk in the target rock mass.

[0219] In one embodiment, the monitoring data acquisition module 201 is specifically configured to:

[0220] Conduct simulated loading tests on rock specimens collected from the target rock mass, and collect acoustic and optical monitoring data of the rock specimens during the monitoring period;

[0221] A fixed-step sliding window is used to perform segmented calculations on the acoustic and optical monitoring data within the monitoring period, and a multi-indicator dataset of multidimensional fracture precursor indicators reflecting the load development process of the rock mass is obtained. Among them, the multidimensional fracture precursor indicators include acoustic emission b-value, acoustic emission energy variance and strain field differentiation data.

[0222] In one embodiment, the acoustic and optical monitoring data within the monitoring period are segmented and calculated using a fixed-step sliding window to obtain a multi-indicator dataset of multi-dimensional fracture precursor indicators reflecting the load development process of the rock mass, including:

[0223] Divide the monitoring period into multiple intervals according to the preset step size and sliding window length;

[0224] The acoustic emission b-value of the corresponding interval is calculated based on the acoustic emission counts of each interval, the acoustic emission energy variance of the corresponding interval is calculated based on the acoustic emission energy of each interval, and the strain field differentiation data of the corresponding interval is calculated based on the DIC data of each interval;

[0225] The acoustic emission b-value, acoustic emission energy variance and strain field differentiation data of each interval are combined according to the time series to obtain a multi-index data set.

[0226] In one embodiment, the data pre-processing module 202 is specifically configured to:

[0227] The deviation normalization method was used to normalize the acoustic emission b value, acoustic emission energy variance and strain field differentiation data in the multi-index data set.

[0228] The entropy weight method is used to weight the normalized acoustic emission b-value, acoustic emission energy variance and strain field differentiation data to obtain the weighted data set, which is the acoustic-optical fusion data set.

[0229] In one embodiment, the warning level determination module 203 is specifically configured to:

[0230] The CUSUM algorithm is used to detect the time series of acoustic-optical fusion data with changing trends in the acoustic-optical fusion data set;

[0231] The Pettitt algorithm is used to detect the mutation points in the time series of acoustic-optical fusion data with a changing trend, so as to identify the mutation points in the time series of acoustic-optical fusion data.

[0232] Calculate the confidence of the mutation point;

[0233] Calculate the physical mutation amplitude of the mutation point;

[0234] Based on the confidence level and physical mutation amplitude of the mutation point, the warning level of the mutation point is determined according to the preset warning level-trigger condition correspondence table.

[0235] In one embodiment, in the warning level-trigger condition correspondence table, the warning levels include level one warning, level two warning, and level three warning, wherein:

[0236] Level 1 warning: indicates a potential trend abnormality and requires continuous monitoring. The corresponding trigger condition is: |C q |>4σ and Pettitt value>0.05;

[0237] Level 2 warning: indicates a small mutation, and the corresponding triggering condition is: |C q |>4σ and Pettitt value ≤0.05, M≤0;

[0238] Level 3 warning: indicates that a large mutation has occurred. The corresponding triggering conditions are: |C q |>4σ and Pettitt value ≤0.05, M>0;

[0239] Among them, C q is the standardized cumulative sum, σ represents the standard deviation of the time series of the acoustic-optic fusion data in the acoustic-optic fusion dataset, the Pettitt value represents the confidence of the mutation point, M represents the physical mutation amplitude of the mutation point, 0 represents the amplitude threshold for mutation, and the value of 0 is related to the rock mass type.

[0240] In one embodiment, the training data set construction module 204 is specifically configured to:

[0241] According to the determined warning level, the warning level is marked on the sound and light monitoring data to obtain a training data set;

[0242] The training dataset is divided into training set, validation set and test set in a ratio of 7:2:1;

[0243] The preset sample weight adjustment algorithm is used to oversample the sample categories in the training set, validation set and test set to obtain the processed training set, validation set and test set.

[0244] It should be noted that the rock fracture risk warning system based on sound and light data fusion and the various embodiments of the rock fracture risk warning method based on sound and light data fusion in the above-mentioned embodiments are implemented based on the same inventive concept, and have the same working principles and technical effects as the rock fracture risk warning method based on sound and light data fusion in the above-mentioned embodiments, which will not be repeated here.

[0245] like Figure 3 As shown, the embodiment of the present application further provides a rock fracture risk warning device 3 based on acoustic and optical data fusion, and the rock fracture risk warning device 3 may include:

[0246] The system includes a memory 301 , a processor 302 , and a computer program 303 stored in the memory 301 and executable on the processor 302 . The memory 301 and the processor 302 communicate with each other via a bus 304 .

[0247] The memory 301 of the rock fracture risk warning device 3 stores an operating system, a computer program 303 for the rock fracture risk warning method based on acoustic and optical data fusion of the present application, and related data. The processor 302 is a high-performance central processing unit (CPU) or other type of processor capable of quickly executing the computer program 303.

[0248] When the rock fracture risk warning device is started and running, the processor 302 loads and executes the computer program 303 in the memory 301, implements the various steps of the rock fracture risk warning method based on the fusion of sound and light data in the above embodiment, and completes the relevant functions of the rock fracture risk warning.

[0249] The rock fracture risk warning device 3, with its high-performance processor 302 and large-capacity memory 301, can quickly and stably run a computer program 303 for a rock fracture risk warning method based on acoustic and optical data fusion. Whether processing complex operations like photo classification and recognition or storing large amounts of photo data and label information, it can be completed efficiently, ensuring the smooth implementation of the rock fracture risk warning function and providing users with a stable and efficient user experience.

[0250] Specifically, the rock fracture risk warning device 3 can be an intelligent device with memory and processor, such as an industrial computer, a PC, or an intelligent mobile terminal, or a computer component with memory and processor, such as a CPU or GPU. In this embodiment, the rock fracture risk warning device 3 is an industrial computer.

[0251] The present application also provides a computer-readable storage medium, which can be a storage device such as a hard disk, solid-state drive, optical disk, or USB flash drive. During manufacture or use, a computer program for the rock fracture risk warning method based on acoustic and optical data fusion described in the above-mentioned embodiments of the present application is stored in the storage medium. When an electronic device needs to execute the method, the computer program is read from the storage medium and executed on a processor to implement the rock fracture risk warning function.

[0252] Computer-readable storage media provide a reliable storage medium for the computer program for the rock fracture risk early warning method based on acoustic and optical data fusion. Different types of storage devices are suitable for different usage scenarios and needs, and users can choose the appropriate storage medium based on their actual needs. Furthermore, the stability and reliability of the storage medium ensure the integrity of the computer program during storage and retrieval, ensuring that electronic devices can accurately execute the program and realize the various functions of the rock fracture risk early warning system.

[0253] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Reference can be made to the common and similar parts between the various embodiments. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method description.

[0254] Professionals will further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented using electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the above description generally describes the components and steps of each example according to their functions. Professionals and technicians may 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.

[0255] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0256] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. The present application will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A rock fracture risk early warning method based on acoustic and optical data fusion, characterized in that: The method comprises the following steps: S1, obtaining acoustic and optical monitoring data of a rock specimen taken from a target rock mass during a monitoring period through an experiment, and calculating a multi-indicator data set of multi-dimensional fracture precursor indicators reflecting a load development process of the rock mass, wherein the acoustic and optical monitoring data includes acoustic emission energy, acoustic emission counts, and DIC data; S2, normalizing and weighting the multi-index dataset to obtain an acoustic-optical fusion dataset; S3, using the CUSUM-Pettitt collaborative test method to identify the mutation point in the acousto-optic fusion data set, and determine the warning level of the mutation point; S4, marking the sound and light monitoring data according to the determined warning level, and constructing a training data set; S5, performing classification training using the LightGBM algorithm based on the training data set to obtain a trained LightGBM model; S6, performing a rock fracture risk warning on the target rock mass based on the trained LightGBM model to obtain a warning result, wherein the warning result is used to characterize the level of rock fracture risk in the target rock mass.

2. The rock mass fracture risk early warning method based on acoustic and optical data fusion according to claim 1 is characterized in that: Step S1 includes: S11, performing a simulated loading test on a rock specimen collected from the target rock mass, and collecting the acoustic and optical monitoring data of the rock specimen during a monitoring period; S12, using a fixed step-size sliding window to perform segmented calculations on the acoustic and optical monitoring data within the monitoring period to obtain a multi-indicator data set of multidimensional fracture precursor indicators reflecting the load development process of the rock mass, wherein the multidimensional fracture precursor indicators include acoustic emission b-value, acoustic emission energy variance and strain field differentiation data.

3. The rock fracture risk early warning method based on acoustic and optical data fusion according to claim 2 is characterized in that: Step S12 includes: S121, dividing the monitoring period into multiple intervals according to a preset step size and sliding window length; S122, calculating the acoustic emission b-value of the corresponding interval based on the acoustic emission counts of each interval, calculating the acoustic emission energy variance of the corresponding interval based on the acoustic emission energy of each interval, and calculating the strain field differentiation data of the corresponding interval based on the DIC data of each interval; S123 , combining the acoustic emission b-value, acoustic emission energy variance, and strain field differentiation data of each interval according to a time series to obtain the multi-index data set.

4. The rock fracture risk early warning method based on acoustic and optical data fusion according to claim 2 is characterized in that: Step S2 includes: S21, normalizing the acoustic emission b-value, acoustic emission energy variance, and strain field differentiation data in the multi-index data set using a deviation normalization method; S22, using an entropy weight method to weight the normalized acoustic emission b-value, acoustic emission energy variance, and strain field differentiation data to obtain a weighted data set, which is the acoustic-optical fusion data set.

5. The rock fracture risk early warning method based on acoustic and optical data fusion according to claim 3 is characterized in that: Step S3 includes: S31, using a CUSUM algorithm to detect an acoustic-optical fusion data time series having a change trend in the acoustic-optical fusion data time series of the acoustic-optical fusion data set; S32, using the Pettitt algorithm to perform mutation point detection on the time series of the acoustic-optical fusion data with a change trend, so as to identify the mutation point in the time series of the acoustic-optical fusion data; S33, calculating the confidence of the mutation point; S34, calculating the physical mutation amplitude of the mutation point; S35 , determining the warning level of the mutation point according to a preset warning level-trigger condition correspondence table based on the confidence level and the physical mutation amplitude of the mutation point.

6. The rock mass fracture risk early warning method based on acoustic and optical data fusion according to claim 5 is characterized in that: In the warning level-trigger condition correspondence table, the warning levels include level 1 warning, level 2 warning and level 3 warning, where: Level 1 warning: indicates a potential trend abnormality and requires continuous monitoring. The corresponding trigger condition is: |C q |>4σ and Pettitt value>0.05; Level 2 warning: indicates a small mutation, and the corresponding triggering condition is: |C q |>4σ and Pettitt value ≤0.05, Level 3 warning: indicates that a large mutation has occurred. The corresponding triggering conditions are: |C q |>4σ and Pettitt value ≤0.05, Among them, C q is the standardized cumulative sum, σ represents the standard deviation of the time series of the acousto-optic fusion data in the acousto-optic fusion data set, the Pettitt value represents the confidence of the mutation point, and M represents the physical mutation amplitude of the mutation point. Indicates the amplitude threshold for mutation, The value of is related to the rock type.

7. The rock fracture risk early warning method based on acoustic and optical data fusion according to any one of claims 1 to 6, characterized in that: Step S4 includes: S41, marking the warning level of the sound and light monitoring data according to the determined warning level to obtain a training data set; S42, dividing the training data set into a training set, a validation set, and a test set in a ratio of 7:2:1; S43, using a preset sample weight adjustment algorithm to perform oversampling processing on the sample categories in the training set, validation set and test set to obtain processed training set, validation set and test set.

8. A rock fracture risk early warning system based on acoustic and optical data fusion, characterized in that: The system comprises: A monitoring data acquisition module is used to obtain acoustic and optical monitoring data of a rock specimen taken from a target rock mass during a monitoring period through experiments, and calculate a multi-indicator data set of multi-dimensional fracture precursor indicators reflecting the load development process of the rock mass, wherein the acoustic and optical monitoring data includes acoustic emission energy, acoustic emission counts, and DIC data; A data preprocessing module is used to normalize and weight the multi-index data set to obtain an acoustic-optical fusion data set; An early warning level determination module is used to identify a mutation point in the acousto-optic fusion data set using a CUSUM-Pettitt collaborative test method and determine an early warning level of the mutation point; A training data set construction module is used to mark the sound and light monitoring data according to the determined warning level and construct a training data set; A model training module is used to perform classification training using the LightGBM algorithm based on the training data set to obtain a trained LightGBM model; The rock mass fracture risk warning module is used to perform rock mass fracture risk warning on the target rock mass based on the trained LightGBM model to obtain a warning result, wherein the warning result is used to characterize the level of fracture risk of the rock mass in the target rock mass.

9. A rock fracture risk warning device based on acoustic and optical data fusion, characterized in that: The method comprises a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method implements the steps of the rock fracture risk warning method based on acoustic and optical data fusion as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the rock fracture risk early warning method based on acoustic and optical data fusion according to any one of claims 1 to 7 are implemented.

Images