Geomagnetic compatibility evaluation model training, application methods, equipment, media and products

Abstract

The present application discloses a geomagnetic adaptability evaluation model training, application method, equipment, medium and product, and relates to the field of mine safety monitoring technology. The method comprises: obtaining geomagnetic reference map data of underground tunnels; then selecting frequency domain adaptability evaluation features by performing ergodic correlation analysis on the cepstral coefficients of the geomagnetic reference map data and the matching probability; calculating the spatial domain feature index set of the underground tunnels according to the geomagnetic reference map data; performing ergodic correlation analysis on the spatial domain feature index set and the matching probability, and selecting spatial domain adaptability evaluation features; finally, training an adaptability evaluation network with frequency domain weighted adaptability evaluation features and spatial domain weighted adaptability evaluation features as input and adaptability degree as output to obtain an adaptability evaluation model; the adaptability evaluation network is an RBF neural network, and the present application realizes the simultaneous consideration of adaptability evaluation features of spatial and frequency domain types, thereby improving the accuracy of adaptability evaluation.

Description

Technical Field

[0001] The present application relates to the technical field of mine safety monitoring, and in particular to a geomagnetic adaptability evaluation model training, application method, equipment, medium and product. Background Art

[0002] Underground geomagnetic positioning is a passive autonomous positioning method and an important guarantee for achieving safe production, personnel management and emergency avoidance in underground mines.

[0003] The selection of adaptation features and the adaptability evaluation method of the geomagnetic reference map are the key technologies for geomagnetic positioning. At present, the commonly used geomagnetic adaptation features include 13 basic adaptation features such as geomagnetic field mean, standard deviation, roughness, roughness variance ratio, geomagnetic entropy, correlation coefficient, fractal dimension, cumulative gradient, correlation distance, geomagnetic Fischer information, geomagnetic coding distortion, and slope standard deviation. They describe the characteristics of geomagnetic field information from different angles, but they are all spatial geomagnetic features. When the geomagnetic spatial domain changes are not large, the accuracy of adaptability evaluation is not high. Summary of the invention

[0004] The purpose of this application is to provide a geomagnetic compatibility evaluation model training, application method, equipment, medium and product, which can combine the adaptability evaluation features of frequency domain and spatial domain to perform geomagnetic adaptability evaluation, thereby improving the accuracy of autonomous adaptability evaluation.

[0005] To achieve the above objectives, this application provides the following solutions:

[0006] In a first aspect, the present application provides a geomagnetic adaptability evaluation model training method, comprising:

[0007] Obtain geomagnetic reference map data of underground tunnels;

[0008] Performing ergodic correlation analysis on the cepstral coefficients and matching probabilities of the geomagnetic reference map data, and obtaining three indices with the largest correlation coefficients as frequency domain adaptability evaluation features; the matching probability is the probability of successful matching between each geomagnetic sequence in the geomagnetic reference map data and the sequence to be matched;

[0009] Calculate the adaptability evaluation weight of the frequency domain adaptability evaluation feature according to the first regression coefficient to obtain a first weight; the first regression coefficient is obtained by performing a multivariate linear regression analysis on the frequency domain adaptability evaluation feature and the matching probability;

[0010] Calculate the spatial characteristic index set of the underground tunnel according to the geomagnetic reference map data; the spatial characteristic index set includes: average geomagnetic field value, geomagnetic standard deviation, kurtosis coefficient, geomagnetic roughness, roughness variance ratio, autocorrelation coefficient and coefficient of variation;

[0011] Perform ergodic correlation analysis on the spatial feature index set and the matching probability, and obtain feature indexes with correlation coefficients greater than a preset threshold as spatial adaptability evaluation features;

[0012] Calculating the adaptability evaluation weight of the spatial adaptability evaluation feature according to the second regression coefficient to obtain a second weight; the second regression coefficient is obtained by performing a multivariate linear regression analysis on the spatial adaptability evaluation feature and the matching probability;

[0013] Taking the frequency domain weight adaptability evaluation feature and the spatial domain weight adaptability evaluation feature as input and the adaptability degree as output, the adaptability evaluation network is trained to obtain the adaptability evaluation model; the frequency domain weight adaptability evaluation feature is the product of the frequency domain adaptability evaluation feature and the first weight; the spatial domain weight adaptability evaluation feature is the product of the spatial domain adaptability evaluation feature and the second weight; the adaptability evaluation network is an RBF neural network.

[0014] Optionally, the method for obtaining the cepstral coefficients of the geomagnetic reference map data specifically includes:

[0015] Performing frame-by-frame windowing on the geomagnetic reference map data to obtain a windowed signal;

[0016] Performing Fourier transform on the windowed signal to obtain a spectrum signal;

[0017] The logarithm of the frequency spectrum signal is taken to obtain the cepstrum coefficients of the geomagnetic reference map data.

[0018] Optionally, the calculation formula of the first weight is:

[0019] ;

[0020] in,

[0021] , =1, 2, 3;

[0022] in, Indicates The first weight of the frequency domain adaptability evaluation feature, Indicates The regression coefficient of the frequency domain adaptability evaluation feature.

[0023] Optionally, before taking the frequency domain weight adaptability evaluation feature and the spatial domain weight adaptability evaluation feature as input and the adaptability degree as output, training the adaptability evaluation network and obtaining the adaptability evaluation model, the method further includes:

[0024] Establish an evaluation experience database corresponding to different types of mines; the types of mines include metal and non-metallic mines; the metal mines include: iron ore, gold mine, silver mine and rare metal mines; the non-metallic mines include: coal mine, phosphate mine and silica mine.

[0025] Optionally, after acquiring the underground tunnel geomagnetic reference map data, the method further includes:

[0026] The geomagnetic reference map data is converted into a one-dimensional matrix form and subjected to pre-emphasis noise reduction processing.

[0027] Optionally, the performing ergodic correlation analysis on the cepstral coefficients and matching probabilities of the geomagnetic reference map data specifically includes:

[0028] The Pearson correlation coefficient method is used to perform ergodic correlation analysis on the cepstral coefficients and matching probabilities of the geomagnetic reference map data.

[0029] In a second aspect, the present application provides a method for applying a geomagnetic compatibility evaluation model, including:

[0030] Obtaining geomagnetic reference map data of underground tunnels at the location to be evaluated;

[0031] The underground tunnel geomagnetic reference map data of the position to be evaluated is input into the geomagnetic adaptability evaluation model to obtain the degree of adaptability; the geomagnetic adaptability evaluation model is trained by any of the geomagnetic adaptability evaluation model training methods described above.

[0032] In a third aspect, the present application provides a computer device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any one of the above-described geomagnetic suitability evaluation model training methods or the above-described geomagnetic suitability evaluation model application methods.

[0033] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-mentioned geomagnetic suitability evaluation model training methods or the above-mentioned geomagnetic suitability evaluation model application methods.

[0034] In a fifth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned geomagnetic suitability evaluation model training methods or the above-mentioned geomagnetic suitability evaluation model application methods.

[0035] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0036] The present application provides a geomagnetic adaptability evaluation model training, application method, equipment, medium and product. First, the geomagnetic reference map data of the underground tunnel is obtained; then, the three indicators with the largest correlation coefficients are obtained as frequency domain adaptability evaluation features by performing ergodic correlation analysis on the cepstral coefficients and matching probabilities of the geomagnetic reference map data; the matching probability is the probability of successful matching between each geomagnetic sequence in the geomagnetic reference map data and the sequence to be matched; the adaptability evaluation weight of the frequency domain adaptability evaluation feature is calculated according to the first regression coefficient to obtain the first weight; the first regression coefficient is obtained by performing multivariate linear regression analysis on the frequency domain adaptability evaluation feature and the matching probability; the acquisition of the frequency domain adaptability evaluation feature and its weight is realized; the spatial domain feature indicator set of the underground tunnel is calculated according to the geomagnetic reference map data; the spatial domain feature indicator set includes: average geomagnetic field value, geomagnetic standard deviation, kurtosis coefficient, geomagnetic roughness, roughness variance ratio, autocorrelation coefficient and coefficient of variation ; Perform ergodic correlation analysis on the spatial feature index set and the matching probability, and obtain feature indicators with correlation coefficients greater than a preset threshold as spatial adaptability evaluation features; calculate the adaptability evaluation weight of the spatial adaptability evaluation feature according to the second regression coefficient to obtain the second weight; the second regression coefficient is obtained by performing multivariate linear regression analysis on the spatial adaptability evaluation feature and the matching probability; the acquisition of spatial adaptability evaluation features and their weights is realized, and finally, the frequency domain weight adaptability evaluation features and the spatial domain weight adaptability evaluation features are used as inputs, and the degree of adaptation is used as output to train the adaptability evaluation network and obtain the adaptability evaluation model; the frequency domain weight adaptability evaluation feature is the product of the frequency domain adaptability evaluation feature and the first weight; the spatial domain weight adaptability evaluation feature is the product of the spatial domain adaptability evaluation feature and the second weight; the adaptability evaluation network is an RBF neural network, which realizes the simultaneous consideration of spatial and frequency domain features, thereby improving the accuracy of adaptability evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0038] Figure 1 This is an application environment diagram of a geomagnetic adaptability evaluation model training method in one embodiment of the present application;

[0039] Figure 2 A flowchart of a geomagnetic adaptability evaluation model training method provided in one embodiment of the present application;

[0040] Figure 3Error stacking histograms of different data sets provided in one embodiment of the present application;

[0041] Figure 4 The error graph of different neuron numbers provided by an embodiment of the present application;

[0042] Figure 5 A point-line diagram of different data sets provided in an embodiment of the present application; wherein, Figure 5 (a) is the training set fitting point line graph, Figure 5 (b) is the test set fitting point line graph, Figure 5 (c) is the fitted point-line graph of the validation set. Figure 5 (d) is the fitted point-line graph of the entire data;

[0043] Figure 6 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0045] The purpose of this application is to provide a geomagnetic compatibility evaluation model training, application method, equipment, medium and product, which can improve the accuracy of compatibility evaluation.

[0046] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0047] The geomagnetic adaptability evaluation model training method provided in the embodiment of the present application can be applied to Figure 1In the application environment shown, the terminal 102 communicates with the server 104 through a network. The data storage system can store data that the server 104 needs to process. The data storage system can be set up separately, integrated on the server 104, or placed on the cloud or other servers. The terminal 102 can send the acquired geomagnetic reference map data of the underground tunnel to the server 104. After receiving the geomagnetic reference map data of the underground tunnel, the server 104 performs traversal correlation analysis on the cepstral coefficients and matching probabilities of the geomagnetic reference map data, and obtains three indicators with the largest correlation coefficients as frequency domain adaptability evaluation features; the matching probability is the probability of successful matching of each geomagnetic sequence in the geomagnetic reference map data with the sequence to be matched; the adaptability evaluation weight of the frequency domain adaptability evaluation feature is calculated according to the first regression coefficient to obtain the first weight; the first regression coefficient is obtained by performing multivariate linear regression analysis on the frequency domain adaptability evaluation feature and the matching probability; the spatial domain feature indicator set of the underground tunnel is calculated according to the geomagnetic reference map data; the spatial domain feature indicator set includes: average geomagnetic field value, geomagnetic standard deviation, kurtosis coefficient, geomagnetic roughness, roughness variance ratio , autocorrelation coefficient and coefficient of variation; perform traversal correlation analysis on the spatial feature index set and the matching probability, and obtain the feature index with a correlation coefficient greater than a preset threshold as the spatial adaptability evaluation feature; calculate the adaptability evaluation weight of the spatial adaptability evaluation feature according to the second regression coefficient to obtain the second weight; the second regression coefficient is obtained by performing a multivariate linear regression analysis on the spatial adaptability evaluation feature and the matching probability; take the frequency domain weight adaptability evaluation feature and the spatial domain weight adaptability evaluation feature as input, and the adaptability degree as output, train the adaptability evaluation network, and obtain the adaptability evaluation model; the frequency domain weight adaptability evaluation feature is the product of the frequency domain adaptability evaluation feature and the first weight; the spatial domain weight adaptability evaluation feature is the product of the spatial domain adaptability evaluation feature and the second weight; the adaptability evaluation network is an RBF neural network. The server 104 can feed back the obtained adaptability evaluation model to the terminal 102. In addition, in some embodiments, the geomagnetic suitability evaluation model training method can also be implemented independently by the server 104 or the terminal 102. For example, the terminal 102 can directly process the underground tunnel geomagnetic reference map data, or the server 104 can obtain the underground tunnel geomagnetic reference map data from the data storage system and process it.

[0048] At the same time, the geomagnetic adaptability evaluation model application method provided in the embodiment of the present application can also be applied to Figure 1In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be set up separately, integrated on the server 104, or placed on the cloud or other servers. The terminal 102 can send the acquired frequency domain adaptability evaluation features and airspace adaptability evaluation features of the position to be evaluated to the server 104. After the server 104 receives the frequency domain adaptability evaluation features and airspace adaptability evaluation features of the position to be evaluated, the frequency domain weight adaptability evaluation features and airspace weight adaptability evaluation features are input into the geomagnetic adaptability evaluation model to obtain the degree of adaptation. The server 104 can feedback the obtained degree of adaptation to the terminal 102. In addition, in some embodiments, the geomagnetic adaptability evaluation model training method can also be implemented independently by the server 104 or the terminal 102. For example, the terminal 102 can directly process the frequency domain adaptability evaluation features and the spatial domain adaptability evaluation features of the position to be evaluated, or the server 104 can obtain the frequency domain adaptability evaluation features and the spatial domain adaptability evaluation features of the position to be evaluated from the data storage system and process them.

[0049] The terminal 102 may be, but is not limited to, various desktop computers, laptop computers, smart phones, tablet computers, IoT devices, and portable wearable devices. The IoT devices may be smart speakers, smart TVs, smart air conditioners, smart vehicle-mounted devices, etc. The portable wearable devices may be smart watches, smart bracelets, head-mounted devices, etc. The server 104 may be implemented as an independent server or a server cluster consisting of multiple servers, or may be a cloud server.

[0050] In an exemplary embodiment, Figure 2 As shown, a method for training a geomagnetic adaptability evaluation model is provided. The method is executed by a computer device, and can be executed by a computer device such as a terminal or a server alone, or can be executed by a terminal and a server together. In the embodiment of the present application, the method is applied to Figure 1 The server 104 in the example is used as an example to illustrate the method, which includes the following steps S1 to S7. Among them:

[0051] S1. Obtaining geomagnetic reference map data of underground tunnels.

[0052] The acquisition device in this embodiment is a geomagnetic combined positioning device. The device has built-in sensors such as fluxgate and accelerometer, which can measure the magnetic information, acceleration and angular velocity of pedestrian paths in real time (wherein the magnetic information refers to the total geomagnetic field and the three components of the total magnetic field. Acceleration and angular velocity are not used in constructing the geomagnetic reference map, but are used in other links of geomagnetic positioning), and has a certain built-in noise reduction algorithm function.

[0053] The signal collection location is the long underground tunnel (except for the transportation tunnel and the mining face, and does not include the tunnel for large electromechanical equipment). Pedestrians carry positioning positions and walk at a constant speed based on the average speed of miners walking underground (1.5m / s). The reference setting sampling frequency should be no less than 5Hz.

[0054] According to the above requirements, the geomagnetic information of the tunnel is continuously collected (geomagnetic information includes: magnetic information, acceleration and angular velocity, among which the magnetic information has three components of the total magnetic field and the magnetic field. However, only the total magnetic field is used to establish the geomagnetic reference map). After preprocessing the data, a strip geomagnetic reference map is produced.

[0055] Repeat the above steps to collect geomagnetic reference map data of multiple underground tunnels (this embodiment is to study the strip geomagnetic reference map in combination with the characteristics of the underground strip area and study its adaptation characteristics), and then extract the geomagnetic reference map data sequence Mgm, Mg It is a one-dimensional matrix, and the data length is related to the length of the underground tunnel to be matched and the sampling frequency. Mgm=[Mg1,Mg2...Mggn ┅MgN] ,in, Mgn Represents the total value of the magnetic field at the corresponding position, in nanoseconds / nt ; Mgn The corresponding coordinates are Xn , Mgn The representative coordinates are Xn The total value of the magnetic field at the position, where N is used to represent the maximum value of n.

[0056] Finally, the obtained geomagnetic reference map data sequence Mg Perform decimal calibration and normalization, and the conversion formula is as follows:

[0057] ;

[0058] The value of q is the logarithm of the largest element Mgn in the sequence Mgm with base 10 and then rounded (for example, if the largest element is 10000, the value of q is 4; if the largest element is 100000, the value of q is 5). Represents the result after conversion.

[0059] After normalization, Pre-emphasis noise reduction processing: First, Wavelet transform is performed to decompose the low-frequency component and the high-frequency component; the high-frequency component is then processed according to the threshold judgment, and finally, the signal of the de-noised high-frequency component and the low-frequency component is reconstructed to obtain the reconstructed geomagnetic reference sequence. In this embodiment, the geomagnetic reference sequence is reconstructed Proceed to the next step as the geomagnetic reference map data.

[0060] S2. Perform ergodic correlation analysis on the cepstral coefficients and matching probabilities of the geomagnetic reference map data, and obtain three indicators with the largest correlation coefficients as frequency domain adaptability evaluation features; the matching probability is the probability that each geomagnetic sequence in the geomagnetic reference map data successfully matches the sequence to be matched. The three indicators form a vector.

[0061] In this embodiment, The MFCC cepstrum coefficients are obtained after frame division, windowing, Fourier transform and logarithm taking.

[0062] Frame Windowing: Add a window function to each frame of magnetic signal, that is, and Multiply to get the windowed signal : ,in, n is the magnetic serial number variable, The position of the window.

[0063] Fourier transform: Perform fast Fourier transform (FFT) on each frame of the preprocessed magnetic sequence to obtain the spectrum value of each frame, and convert the spatial domain signal (i.e. the geomagnetic reference map data mentioned above) into a frequency domain signal:

[0064] ;

[0065] in, represents the time domain signal before transformation, represents the transformed frequency domain signal, N Indicates the length of the magnetic signal. Then the results of each frame are stacked along another dimension (meaning that the results of each frame are connected in sequence, and there is no specific dimension here) to convert the spatial domain signal into a spectrum signal.

[0066] Cepstral coefficients MFCC Calculation: Perform cepstrum analysis on the generated spectrum signal and take the logarithm of the spectrum signal to obtain MFCC . MFCC The 13 coefficients of f i , corresponding rights W ai ，i= 1-13.

[0067] S3. Calculate the adaptability evaluation weight of the frequency domain adaptability evaluation feature according to the first regression coefficient to obtain a first weight; the first regression coefficient is obtained by performing a multivariate linear regression analysis on the frequency domain adaptability evaluation feature and the matching probability.

[0068] In this embodiment, the Pearson correlation coefficient method is used to analyze the 13 coefficients of MFCC. fi A traversal correlation analysis is performed with the matching probability P, and the top three indicators with the largest correlation coefficients are taken to form the frequency domain adaptability evaluation feature vector for geomagnetic adaptability evaluation.

[0069] The matching probability is a quantitative indicator for evaluating the adaptability of the geomagnetic reference map, which is defined as follows: select each point in the candidate matching area as a reference point p, and extend point p in a certain direction by a length of l A sequence is used as a template, and the reference point p is used as the point to be matched. The measured sequence corresponding to the template sequence is searched and matched in the candidate matching area through the relevant matching algorithm. If the distance between the registration point (i.e., the real positioning point) pm and the point to be matched p is less than the given error ε, the match is considered successful; otherwise, the match is considered unsuccessful. The number of successful matches divided by the total number of matches is the matching probability.

[0070] Among them, the frequency domain adaptability evaluation feature vector is:

[0071] ;

[0072] in, f m1 , f m2 and f m3 Respectively represent the three selected indicators (i.e. the three cepstral coefficients with the highest correlation with the matching probability mentioned above), F m Represents the frequency domain adaptability evaluation feature vector.

[0073] The spectrum characteristic index and matching probability are subjected to multivariate linear regression analysis, and the regression coefficient is β mj , the corresponding adaptability evaluation weight is:

[0074] ;

[0075] in,

[0076] , =1, 2, 3;

[0077] in, Indicates The first weight of the frequency domain adaptability evaluation feature, Indicates The regression coefficient of the frequency domain adaptability evaluation feature.

[0078] S4. Calculate the spatial characteristic index set of the underground tunnel according to the geomagnetic reference map data; the spatial characteristic index set includes: average geomagnetic field value, geomagnetic standard deviation, kurtosis coefficient, geomagnetic roughness, roughness variance ratio, autocorrelation coefficient and coefficient of variation.

[0079] The calculation formula is shown in Table 1:

[0080] Table 1 Calculation table of airspace characteristic index set

[0081]

[0082] in, M gn represents the nth total magnetic field (total geomagnetic field) of a linear magnetic sequence, M gn and M gn+1 Represents the magnetic field of two adjacent points; N represents the number of measuring points in a magnetic sequence.

[0083] Constructing a set of spatial feature indicators F a =[ , ,r, , ,K], the weight of each airspace adaptability evaluation feature is: w’ aj , j =1,2,3,4,5,6.

[0084] The spatial feature index set F a Perform normalization.

[0085] S5. Perform ergodic correlation analysis on the spatial feature index set and the matching probability, and obtain feature indexes with correlation coefficients greater than a preset threshold as spatial adaptability evaluation features.

[0086] The Pearson correlation coefficient method is used to perform correlation analysis on each spatial feature index and the matching probability. The feature index with a correlation coefficient greater than 0.5 is selected as the geomagnetic matching spatial domain suitability evaluation feature for suitability evaluation.

[0087] S6. Calculate the adaptability evaluation weight of the spatial domain adaptability evaluation feature according to the second regression coefficient to obtain a second weight; the second regression coefficient is obtained by performing a multivariate linear regression analysis on the spatial domain adaptability evaluation feature and the matching probability.

[0088] The selected spatial adaptability evaluation features F a The regression analysis is performed with the matching probability P, as shown below:

[0089] ;

[0090] Among them, s≤6, which means that s spatial domain adaptability evaluation features are selected. F aj Indicates the first j A specific set of airspace adaptability evaluation features, β aj Indicates j The regression coefficient of each airspace adaptability evaluation feature is used as the weight set of each airspace adaptability evaluation feature. W’ a :

[0091] ;

[0092] in, ; Indicates j The second weight of the spatial adaptability evaluation feature, j =1,2,3……s.

[0093] S7. Taking the frequency domain weight adaptability evaluation feature and the spatial domain weight adaptability evaluation feature as input and the adaptability degree as output, train the adaptability evaluation network to obtain an adaptability evaluation model; the frequency domain weight adaptability evaluation feature is the product of the frequency domain adaptability evaluation feature and the first weight; the spatial domain weight adaptability evaluation feature is the product of the spatial domain adaptability evaluation feature and the second weight; the adaptability evaluation network is an RBF neural network.

[0094] Before training, you need to build a classification indicator data set. The adaptability evaluation level is 0 to 3, representing incompatible, weak, adaptable and strong, as shown in Table 2 below:

[0095] Table 2 Classification index dataset

[0096]

[0097] According to empirical values, P1, P2 and P3 are taken as 0.35, 0.65 and 0.85 respectively.

[0098] Construct frequency domain / spatial domain joint evaluation index dataset F and corresponding weight set W';

[0099] , ;

[0100] The initial weight is determined by regression analysis of the correlation between the frequency domain / spatial domain joint adaptation feature F and the matching probability P of the underground geomagnetic space. The equation is as follows:

[0101] ;

[0102] in P is the matching probability, To adapt the feature parameters (index data set F elements in, including spatial domain adaptability evaluation features and frequency domain adaptability evaluation features), W’ t is the element in the weight set, i.e., the initial weight, W’ 0 is a constant, .

[0103] Establishing the adaptability evaluation network:

[0104] The RBF neural network is used as the adaptability evaluation model. The model includes three layers: input layer, hidden layer and output layer, and its network structure is mh-1.

[0105] Determine the number of neurons m in the input layer, including the frequency domain / spatial domain joint evaluation feature vector set F.

[0106] The number of neurons in the output layer is determined to be 1, and the value range of the output parameter is [0, 1], [0, 0.35], [0.35, 0.65], [0.65, 0.85], and [0.85, 1], representing inappropriate, weak, appropriate, and strong adaptation, respectively.

[0107] The model is trained by using proportional samples of four adaptation zones, including unfit, weak fit, fit and strong fit, and then verified and evaluated, finally forming a geomagnetic reference map adaptability evaluation model with strong generalization ability and high classification accuracy. A total of three functions are involved in the model. The radial basis function uses the Gaussian function, the objective function is the product of the connection weight and the corresponding hidden layer node, and the root mean square error is used to evaluate the output result.

[0108] Establish RBF neural network evaluation experience database:

[0109] The empirical database should include various types of mines and underground air-raid shelters, etc. According to the existing mine types, they are divided into metal and non-metallic mines. Metal mines include iron ore, gold mine, silver mine, rare metals, etc.; non-metallic mines include coal mines, phosphate mines, silica mines, etc.

[0110] The adaptive feature indicators and weights selected for different types of mines may be different. We can establish an empirical database of the corresponding type based on a large number of samples, including different types and quantities of adaptive features. F CP , and the model parameters V m , connection right W , which can be expressed as:

[0111] ;

[0112] ;

[0113] Among them, CP is the type of area to be evaluated, F CP is the adaptation feature selected for the area to be evaluated, V m are model parameters, m is the number of input layers, h is the number of hidden layers, C h’ and σ h’ It is h’ The centers and widths of the hidden layer neurons, W’ is the connection weight from the hidden layer to the output layer, with a total of h A connection right.

[0114] Finally, the indicators and parameters of the evaluation model are determined: 100 sample data are selected for each tunnel type, 80 data samples are used to train the default parameters of the corresponding model, and the remaining 20 samples are used as validation data sets for verification and model evaluation. When the model is actually applied, any type of tunnel geomagnetic reference map is selected, and after input into the model, an excellent adaptability evaluation result should be output.

[0115] The present embodiment provides a method for training a geomagnetic adaptability evaluation model, including three parts: optimizing and constructing a geomagnetic frequency domain feature adaptability evaluation index of an underground tunnel, optimizing and constructing a geomagnetic spatial domain adaptability evaluation index of an underground tunnel, and constructing a frequency domain and spatial domain feature index adaptability joint evaluation method; firstly, obtaining a geomagnetic sequence of an underground strip geomagnetic reference map; determining a frequency domain feature evaluation index construction module for extracting the frequency domain feature index of the geomagnetic reference map of the underground tunnel, and providing input frequency domain adaptability evaluation feature parameters for the adaptability evaluation of the neural network; determining a geomagnetic spatial domain evaluation index construction module for the underground geomagnetic reference map; and finally determining a frequency domain feature evaluation index construction module for extracting the frequency domain feature index of the geomagnetic reference map of the underground tunnel. The module for weighting and constructing price indicators is used to extract the spatial evaluation indicators of the geomagnetic reference map of the underground tunnel, and to weight different spatial adaptability evaluation features, so as to provide input spatial adaptability evaluation feature parameters for the adaptability evaluation of the neural network; a module for determining the frequency domain + spatial feature joint adaptability evaluation method is used to perform spatial and frequency domain joint evaluation on the adaptability of the geomagnetic matching positioning of the matching area through the RBF neural network, and determine the adaptability attributes of the matching area such as strong matching, adaptability, weak matching and incompatibility; the evaluation system evaluation method provided in this embodiment is simple and accurate.

[0116] In order to verify the effect of the present application, this embodiment has been experimentally verified, and the data used are shown in Table 3 below:

[0117] Table 3 Description of the original data used in the experimental verification

[0118]

[0119] The model accuracy obtained in this embodiment is shown in Table 4 below. According to the results after neural network training, the prediction data of three data sets, namely training set, test set and validation set, are obtained respectively, and the errors between the predicted values ​​and the true values ​​of different data sets are calculated. The same number of error data of three data sets are randomly extracted, and the errors are evenly divided by orgin at a scale of 0.02 to draw histograms, and the graphs of different data sets are superimposed together according to the frequency to obtain error stacking histograms of different data sets; the number of hidden layer neuron nodes is increased one by one starting from 1, and the root mean square error is used as the judgment criterion. After several trainings, the first 20 training results with gradually flattened errors are selected to obtain error graphs with different numbers of neurons; according to the results after neural network training, the predicted values ​​of different data sets are combined as X-axis data and the true values ​​as Y-axis data. Use orgin to draw scatter plots for different data sets respectively, and then use the fitting tool to fit the scatter plot data, calculate the determination coefficient, and draw fitting lines according to the fitting calculation results to obtain fitting point and line graphs of different data sets. The results are as follows Figure 3-Figure 5 As shown, Figure 3 is the error stacked histogram for different data sets, Figure 4 is the error diagram for different numbers of neurons, Figure 5 Fit point-line plots to different data sets, where Figure 5 (a) is the training set fitting point line graph, Figure 5 (b) is the test set fitting point line graph, Figure 5 (c) is the fitted point-line graph of the validation set. Figure 5 (d) is the fitted point-line graph of the entire data.

[0120] Table 4 Comparison of model accuracy

[0121]

[0122] In another exemplary embodiment of the present application, a method for applying a geomagnetic compatibility evaluation model is provided, the method comprising:

[0123] Step A1: Obtain geomagnetic reference map data of the underground tunnel at the location to be evaluated.

[0124] Step A2: input the underground tunnel geomagnetic reference map data of the position to be evaluated into the geomagnetic adaptability evaluation model to obtain the degree of adaptability; the geomagnetic adaptability evaluation model is trained by the geomagnetic adaptability evaluation model training method described in the previous embodiment. In this embodiment, the underground tunnel geomagnetic reference map data of the position to be evaluated is processed as above, and the degree of adaptability is obtained using the geomagnetic adaptability evaluation model.

[0125] The present application also provides an application scenario, which applies the above-mentioned geomagnetic adaptability evaluation model application method. Specifically: The geomagnetic adaptability evaluation model application method provided in this embodiment can be applied in a location selection scenario. The location selection scenario includes selecting a location, training an evaluation algorithm, and evaluating whether the selected location is suitable; selecting a suitable location from a plurality of selected locations. The geomagnetic adaptability evaluation model application method provided in this embodiment belongs to the link of evaluating whether the selected location is suitable.

[0126] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 6 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store underground tunnel geomagnetic reference map data or frequency domain adaptability evaluation features and airspace adaptability evaluation features of the position to be evaluated. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a geomagnetic adaptability evaluation model training method or a geomagnetic adaptability evaluation model application method is implemented.

[0127] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0128] In an exemplary embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0129] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0130] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0131] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0132] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0133] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0134] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0135] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A geomagnetic adaptability evaluation model training method, It is characterized in that The geomagnetic adaptability evaluation model training method comprises: Obtain geomagnetic reference map data of underground tunnels; Performing ergodic correlation analysis on the cepstral coefficients and matching probabilities of the geomagnetic reference map data, and obtaining three indices with the largest correlation coefficients as frequency domain adaptability evaluation features; the matching probability is the probability of successful matching between each geomagnetic sequence in the geomagnetic reference map data and the sequence to be matched; Calculate the adaptability evaluation weight of the frequency domain adaptability evaluation feature according to the first regression coefficient to obtain a first weight; the first regression coefficient is obtained by performing a multivariate linear regression analysis on the frequency domain adaptability evaluation feature and the matching probability; Calculate the spatial characteristic index set of the underground tunnel according to the geomagnetic reference map data; the spatial characteristic index set includes: average geomagnetic field value, geomagnetic standard deviation, kurtosis coefficient, geomagnetic roughness, roughness variance ratio, autocorrelation coefficient and coefficient of variation; Perform ergodic correlation analysis on the spatial feature index set and the matching probability, and obtain feature indexes with correlation coefficients greater than a preset threshold as spatial adaptability evaluation features; Calculating the adaptability evaluation weight of the spatial adaptability evaluation feature according to the second regression coefficient to obtain a second weight; the second regression coefficient is obtained by performing a multivariate linear regression analysis on the spatial adaptability evaluation feature and the matching probability; Taking the frequency domain weight adaptability evaluation feature and the spatial domain weight adaptability evaluation feature as input and the adaptability degree as output, the adaptability evaluation network is trained to obtain the adaptability evaluation model; the frequency domain weight adaptability evaluation feature is the product of the frequency domain adaptability evaluation feature and the first weight; the spatial domain weight adaptability evaluation feature is the product of the spatial domain adaptability evaluation feature and the second weight; the adaptability evaluation network is an RBF neural network; The method for obtaining the cepstral coefficients of the geomagnetic reference map data specifically includes: Performing frame-by-frame windowing on the geomagnetic reference map data to obtain a windowed signal; Performing Fourier transform on the windowed signal to obtain a spectrum signal; Taking the logarithm of the frequency spectrum signal to obtain the cepstrum coefficients of the geomagnetic reference map data; Fourier Transform: ； in, represents the time domain signal before transformation, represents the transformed frequency domain signal, N Indicates the length of the magnetic signal; Stack the results of each frame along another dimension to convert the geomagnetic reference map data into a spectrum signal; The calculation formula of the first weight is: ； in, ， =1，2，3； in, Indicates The first weight of the frequency domain adaptability evaluation feature, Indicates The regression coefficient of the frequency domain adaptability evaluation feature.

2. The geomagnetic adaptability evaluation model training method according to claim 1, It is characterized in that Before taking the frequency domain weight adaptability evaluation feature and the spatial domain weight adaptability evaluation feature as input and the adaptability degree as output, training the adaptability evaluation network and obtaining the adaptability evaluation model, the following is also included: Establish an evaluation experience database corresponding to different types of mines; the types of mines include metal and non-metallic mines; the metal mines include: iron ore, gold mine, silver mine and rare metal mines; the non-metallic mines include: coal mine, phosphate mine and silica mine.

3. The geomagnetic adaptability evaluation model training method according to claim 1, It is characterized in that After obtaining the underground tunnel geomagnetic reference map data, the method further includes: The geomagnetic reference map data is converted into a one-dimensional matrix form and subjected to pre-emphasis noise reduction processing.

4. The geomagnetic adaptability evaluation model training method according to claim 1, It is characterized in that The performing ergodic correlation analysis on the cepstral coefficients and matching probabilities of the geomagnetic reference map data specifically includes: The Pearson correlation coefficient method is used to perform ergodic correlation analysis on the cepstral coefficients and matching probabilities of the geomagnetic reference map data.

5. An application method of geomagnetic compatibility evaluation model, It is characterized in that The geomagnetic adaptability evaluation model application method comprises: Obtaining geomagnetic reference map data of underground tunnels at the location to be evaluated; The underground tunnel geomagnetic reference map data of the position to be evaluated is input into the geomagnetic adaptability evaluation model to obtain the degree of adaptability; the geomagnetic adaptability evaluation model is trained by the geomagnetic adaptability evaluation model training method described in any one of claims 1-4.

6. A computer device, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the geomagnetic suitability evaluation model training method described in any one of claims 1 to 4 or the geomagnetic suitability evaluation model application method described in claim 5.

7. A computer-readable storage medium having a computer program stored thereon, It is characterized in that When the computer program is executed by a processor, the method for training a geomagnetic adaptability evaluation model described in any one of claims 1 to 4 or the method for applying a geomagnetic adaptability evaluation model described in claim 5 is implemented.

8. A computer program product comprising a computer program, It is characterized in that When the computer program is executed by a processor, the method for training a geomagnetic adaptability evaluation model described in any one of claims 1 to 4 or the method for applying a geomagnetic adaptability evaluation model described in claim 5 is implemented.

Images