Tunnel void disease identification method and system based on audio frequency analysis

Abstract

This invention relates to the field of tunnel caving identification technology using audio frequency analysis, specifically to a method and system for identifying tunnel caving based on audio frequency analysis. This invention obtains the signal resonance of the audio signal corresponding to each measuring point in the time and frequency domains; obtains multiple measuring point clusters; obtains the resonance response of each measuring point cluster based on the information resonance distribution of different measuring points within each cluster, thus filtering out caving areas; obtains the caving area conformity of each measuring point based on different caving areas, the number of measuring points in each measuring point's cluster, and the resonance response; obtains the signal resonance gradient of each measuring point based on the signal resonance of different measuring points within its neighborhood; and obtains the caving probability of each measuring point for caving identification. This invention improves the accuracy of caving identification by accurately obtaining the caving probability of each measuring point.

Description

Technical Field

[0001] This invention relates to the field of tunnel caving disease identification using audio frequency analysis, and specifically to a method and system for identifying tunnel caving disease based on audio frequency analysis. Background Technology

[0002] Due to factors such as surrounding rock deformation, insufficient construction quality, or groundwater infiltration, voids can easily occur between the tunnel lining and the surrounding rock, leading to a decrease in the lining's load-bearing capacity, cracks, leakage, or even structural instability, which seriously threatens tunnel safety. Therefore, it is necessary to identify voids.

[0003] In existing technologies, the propagation characteristics of sound waves or shock waves in the lining are measured and analyzed to extract audio signal features and establish a classifier model algorithm to effectively identify tunnel lining cavities. However, under real-world conditions, due to the complex semi-adhesive and semi-detached state between the tunnel lining concrete and the surrounding rock, there may be complex situations such as friction, cracks, and partial detachment. This causes the audio signal to exhibit mixed characteristics near the cavitation boundary, that is, some frequency components are attenuated or enhanced, and the energy distribution is uneven. The ambiguity of the audio signal makes it difficult to accurately determine the tunnel cavitation boundary, resulting in poor accuracy in identifying cavitation defects. Summary of the Invention

[0004] To address the technical problem of poor accuracy in identifying air defense defects due to the complex semi-adhesive and semi-detached state between tunnel lining concrete and surrounding rock, this invention aims to provide a method and system for identifying tunnel detachment defects based on audio frequency analysis. The specific technical solution adopted is as follows:

[0005] This invention proposes a method for identifying tunnel delamination defects based on audio frequency analysis, the method comprising:

[0006] Acquire the audio signal at each measuring point during tunnel void detection;

[0007] Based on the distribution characteristics of the audio signal at each measurement point in the time and frequency domains, the signal resonance of each measurement point is obtained; based on the signal resonance of different measurement points, all measurement points are clustered to obtain multiple measurement point cluster regions; based on the information resonance distribution of different measurement points in each measurement point cluster region, the resonance response of each measurement point cluster region is obtained, and the void regions are screened out.

[0008] Based on the number of measurement points in different gap regions and the measurement point cluster region, as well as the resonance responsivity, the gap region conformity of the measurement point cluster region is obtained; based on the signal resonance of different measurement points in the neighborhood of each measurement point, the signal complexity of each measurement point is obtained; based on the location characteristics and signal complexity of different measurement points, the signal resonance gradient of each measurement point is obtained.

[0009] Based on the degree of conformity of the void region, the signal resonance gradient, and the signal resonance of each measuring point in the cluster region, the void probability of each measuring point is obtained, and void disease identification is performed.

[0010] Furthermore, the method for obtaining the signal resonance includes:

[0011] Obtain any peak value of the audio signal at each measurement point in the frequency domain, acquire the frequencies at the half-peak values ​​on the left and right sides adjacent to the peak value, and select the maximum value of the ratio between different peak values ​​and their corresponding frequencies.

[0012] The fitting line is obtained by fitting the amplitude of the audio signal at each measurement point at different times in the time domain;

[0013] The signal resonance of each measurement point is obtained by the slope and the maximum ratio of the fitted straight line. The slope and the maximum ratio are both positively correlated with the signal resonance.

[0014] Furthermore, the method for obtaining the clustering region of the measurement points includes:

[0015] Based on the signal resonance of different measurement points, DBSCAN clustering is performed on all measurement points to obtain multiple measurement point clustering regions.

[0016] Furthermore, the method for obtaining the resonant responsivity includes:

[0017] The mean signal resonance of all measurement points within each measurement point cluster region is obtained as the average resonance level of each measurement point cluster region;

[0018] Based on the resonance difference between the signal resonance and the average resonance level of different measurement points within each measurement point cluster region, as well as the signal resonance, the resonance responsivity of each measurement point cluster region is obtained. The resonance difference is negatively correlated with the resonance responsivity, while the signal resonance is positively correlated with the resonance responsivity.

[0019] Furthermore, the method for obtaining the vacancy region includes:

[0020] If the resonance response of the clustered region of the measuring points is greater than the preset response threshold, the corresponding clustered region of the measuring points will be regarded as the empty region.

[0021] Furthermore, the method for obtaining the conformity of the void region includes:

[0022] The absolute value of the correlation coefficient between the number of measuring points and the resonance response in different void regions is obtained as the overall correlation strength;

[0023] The two regions with the smallest difference between the resonant responsivity of other measurement point clusters and the measurement point clusters of each measurement point are selected as the resonant correlation regions of the measurement point clusters of each measurement point; the absolute value of the correlation coefficient of the sequence formed by the number of measurement points in different resonant correlation regions and the resonant responsivity is obtained as the local correlation strength;

[0024] The difference between the local correlation strength and the overall correlation strength of the cluster region of each measuring point is obtained, and negative correlation mapping is performed as the degree of conformity of the empty region of the cluster region of each measuring point.

[0025] Furthermore, the method for obtaining the signal complexity includes:

[0026] The mean value of the signal resonance difference between adjacent measurement points within the neighborhood of each measurement point is obtained as the signal complexity of each measurement point.

[0027] Furthermore, the method for obtaining the signal resonance gradient includes:

[0028] For each measurement point, in all neighboring measurement point clusters within its measurement point cluster, the measurement point with the highest signal resonance within the cluster is selected as a reference point. The average signal complexity of all reference points within the cluster is obtained as the average relative signal complexity of each measurement point. The first difference between the average signal complexity and the average relative signal complexity of each measurement point is then obtained.

[0029] The second difference in signal complexity between each measurement point and the reference point of each adjacent measurement point cluster region is obtained; the mean ratio of the first difference and the second difference corresponding to each measurement point and different adjacent measurement point cluster regions is obtained, the difference between the mean ratio and the positive integer 1 is calculated, and negative correlation mapping is performed as the signal resonance gradient of each measurement point.

[0030] Furthermore, the method for obtaining the empty probability includes:

[0031] The product between the signal resonance gradient of each measurement point and the degree of conformity of the empty region of the clustering region of the measurement point is obtained, and the group is used as the signal resonance transition of each measurement point.

[0032] Obtain the first sum between the positive integer 1 and the signal resonant transition value, and calculate the product of the first sum value and the signal resonant transition value as the probability of escaping from the point of measurement.

[0033] The present invention also proposes a tunnel caving defect identification system based on audio frequency analysis, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any of the steps of the tunnel caving defect identification method based on audio frequency analysis.

[0034] The present invention has the following beneficial effects:

[0035] This invention obtains the signal resonance of each measuring point based on the distribution characteristics of the audio signal in the time and frequency domains, and preliminarily analyzes the resonance characteristics of each measuring point. Based on the signal resonance of different measuring points, all measuring points are clustered to obtain multiple measuring point cluster regions. Measuring points with similar resonance modes are further clustered into regional clusters for targeted analysis of local areas. Since the voiding phenomenon in tunnel lining usually exhibits spatial continuity, a comprehensive analysis of the resonance characteristics of all measuring points within a cluster region can eliminate interference from local anomalies and more reliably reflect the voiding state. Based on the information resonance distribution of different measuring points within each measuring point cluster region, the resonance response of each measuring point cluster region is obtained. The process involves several steps: First, identifying voided regions. Based on the number of measurement points within each voided region and the number of measurement points in their respective clusters, along with their resonant responsivity, the degree of conformity to the voided region criteria for each measurement point's cluster is determined, assessing the likelihood that each cluster conforms to voided region characteristics. Second, determining the signal complexity of each measurement point based on the signal resonance of different measurement points within its neighborhood, quantifying the complexity of each measurement point's audio response. Third, determining the signal resonant gradient of each measurement point based on its location characteristics and signal complexity. Finally, determining the voiding probability of each measurement point based on its voiding region conformity, signal resonant gradient, and signal resonance, thus enabling voiding disease identification. This invention improves the accuracy of voiding disease identification by accurately obtaining the voiding probability of each measurement point. Attached Figure Description

[0036] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 A flowchart illustrating a method for identifying tunnel delamination defects based on audio frequency analysis, provided in one embodiment of the present invention;

[0038] Figure 2 This is a flowchart of a method for obtaining signal resonance gradient according to an embodiment of the present invention;

[0039] Figure 3 This is a flowchart illustrating a method for obtaining the compliance of a void region according to an embodiment of the present invention. Detailed Implementation

[0040] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a tunnel delamination identification method and system based on audio frequency analysis proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0041] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0042] The following description, in conjunction with the accompanying drawings, details the specific scheme of the tunnel delamination identification method and system based on audio frequency analysis provided by this invention.

[0043] Please see Figure 1 The diagram illustrates a flowchart of a method for identifying tunnel delamination defects based on audio frequency analysis, according to an embodiment of the present invention. The method specifically includes:

[0044] Step S1: Acquire the audio signal at each measuring point during tunnel delamination detection.

[0045] In an embodiment of the invention, the detection route is rationally planned based on the tunnel's total length, cross-sectional characteristics, and traffic conditions. The detection speed range and the required lane positions are determined. The excitation device, sensor array, data acquisition instrument, and GPS / IMU positioning system are integrated and installed on the detection vehicle. The detection vehicle travels at a constant speed, and the automatic tapping device continuously taps the lining surface at a fixed frequency. The sensor array synchronously collects the sound generated by each tap. The sound frequency signal obtained from each tap is bound in real time with the spatial information such as mileage, ring number, and horizontal and vertical coordinates output by the positioning system to form a one-to-one corresponding data block. The sound frequency response is accurately mapped to the actual spatial position of the tunnel lining, that is, the sound frequency signal of each measuring point is obtained and analyzed during the tunnel void detection.

[0046] Step S2: Based on the distribution characteristics of the audio signal at each measurement point in the time and frequency domains, obtain the signal resonance of each measurement point; based on the signal resonance of different measurement points, cluster all measurement points to obtain multiple measurement point cluster regions; based on the information resonance distribution of different measurement points in each measurement point cluster region, obtain the resonance response of each measurement point cluster region and screen out the void regions.

[0047] When a tunnel experiences voids, the lining loses support, forming localized cavities. Sound waves generated by vibration propagate within the concrete slab. Upon encountering the boundary of the void, the sound waves are confined and repeatedly reflected because air cannot transmit vibrations as effectively as a solid. This results in energy superposition at specific frequencies, manifesting as specific resonance peaks or energy enhancement phenomena in the audio signal. The greater the resonance, the more pronounced the resonance. Considering that the audio signal at each measuring point reveals different physical information in the time and frequency domains, in the time domain, if voids exist, sound waves will repeatedly reflect between the lining and the cavity, leading to a slower signal attenuation rate and exhibiting characteristics of longer duration and slower energy decay. In dense regions, the signal attenuates rapidly. In the frequency domain, voided regions correspond to obvious resonance phenomena, manifesting as sharp and high peaks in the spectrum, while the spectrum of dense regions is relatively flat. Therefore, based on the distribution characteristics of the audio signal at each measuring point in both the time and frequency domains, the signal resonance of each measuring point can be obtained.

[0048] Preferably, in one embodiment of the present invention, the method for obtaining signal resonance is described in [reference needed]. Figure 2 It illustrates a flowchart of a method for obtaining signal resonance, including:

[0049] Step S201: Obtain any peak value of the audio signal at each measurement point in the frequency domain, obtain the frequencies at the half-peak values ​​on the left and right sides adjacent to the peak value, and select the maximum value of the ratio between different peak values ​​and their corresponding frequencies.

[0050] It should be noted that the difference represents the absolute value of the calculated difference; since the sharper and higher the peak in the spectrum, the more likely it is to be a void region showing obvious resonance phenomenon, that is, the larger the peak, the smaller the difference between the spectra, and the larger the maximum value of the ratio.

[0051] Step S202: Obtain the fitting line for fitting the amplitude of the audio signal at each measurement point at different times in the time domain.

[0052] It should be noted that, in the embodiments of the present invention, the least squares method or the polynomial fitting method can be used for fitting. The specific means are well known to those skilled in the art and will not be described in detail here.

[0053] Step S203: Based on the slope and maximum ratio of the fitted straight line, obtain the signal resonance of each measurement point. The slope and maximum ratio are both positively correlated with the signal resonance.

[0054] In the time domain, the signal is constantly decaying. The slower the decay, the more likely there is a gap, and the greater the resonance. The faster the decay, the smaller the resonance. The slope can represent the decay of the signal. In the embodiments of the present invention, the slope can be obtained by taking the derivative of the fitted straight line or by calculating the ratio of the difference between the vertical coordinate and the horizontal coordinate between the end time and the start time. The smaller the ratio, the faster the decay. The specific means are well known to those skilled in the art and will not be described in detail here.

[0055] In an embodiment of the present invention, the slope is mapped using an exponential function with the natural constant as the base, and the product of the slope mapping value and the maximum ratio value is obtained. This product is then normalized and mapped to represent the signal resonance of each measurement point. Therefore, based on the above basic mathematical operations, a correlation is constructed between the slope and the maximum ratio value and the signal resonance. That is, the larger the slope, the smaller the attenuation, the larger the mapping value, the larger the maximum ratio value, and the greater the signal resonance.

[0056] It should be noted that in some embodiments of the present invention, normalization can be performed by linear normalization or a normalization function. The specific means are well known to those skilled in the art and will not be described in detail here.

[0057] Clustering is used to group similar signal resonances into one category for targeted analysis, eliminating interference from local anomalies. Based on the signal resonances of different measurement points, all measurement points are clustered to obtain multiple measurement point cluster regions.

[0058] Preferably, in one embodiment of the present invention, the method for obtaining the clustering region of measurement points includes:

[0059] Based on the signal resonance of different measurement points, DBSCAN clustering is performed on all measurement points to obtain multiple measurement point clustering regions.

[0060] It should be noted that DBSCAN clustering is density-based clustering, which groups measurement points that are density-connected and have similar signal resonance into a class, forming multiple measurement point cluster regions; the specific clustering method is a well-known technique in the art and will not be elaborated here.

[0061] When the detection excitation is applied to the lining, the lining section will respond to the acoustic excitation with similar natural frequencies and vibration modes. The acoustic signals of adjacent measuring points in the void area show similar and more obvious resonance. Based on the information resonance distribution of different measuring points in the cluster area of ​​each measuring point, the resonance response of the cluster area of ​​each measuring point is obtained.

[0062] Preferably, in one embodiment of the present invention, the method for obtaining the resonant responsivity includes:

[0063] The mean signal resonance of all measurement points within each measurement point cluster region is obtained as the average resonance level of each measurement point cluster region;

[0064] Based on the resonance difference between the signal resonance and the average resonance level of different measurement points within each measurement point cluster region, as well as the signal resonance, the resonance responsivity of each measurement point cluster region is obtained. The resonance difference is negatively correlated with the resonance responsivity, while the signal resonance is positively correlated with the resonance responsivity.

[0065] It should be noted that the greater the resonance difference, the more the signal resonance at each measuring point deviates from the overall resonance, and the smaller the resonance responsivity; the greater the signal resonance at each measuring point, the more obvious the resonance response is; therefore, resonance difference is negatively correlated with resonance responsivity, and signal resonance is positively correlated with resonance responsivity.

[0066] In one embodiment of the invention, an exponential function with a base of the natural constant is used. The resonance difference is negatively correlated and mapped to the resonance response coefficient. The average product of the resonance response coefficient and the signal resonance of all measurement points in each measurement point cluster region is obtained as the resonance responsivity of each measurement point cluster region. Therefore, based on the above basic mathematical operations, the correlation between resonance difference, signal resonance, and resonance responsivity is constructed, that is, the greater the resonance difference, the smaller the resonance response coefficient, the smaller the signal resonance, and the smaller the resonance responsivity.

[0067] The greater the resonance response, the more likely it is to be a void region. Therefore, void regions are screened based on the resonance response.

[0068] Preferably, in one embodiment of the present invention, the method for obtaining the voided region includes:

[0069] If the resonance response of the clustered region of the measuring points is greater than the preset response threshold, the corresponding clustered region of the measuring points will be regarded as the empty region.

[0070] It should be noted that, in one embodiment of the present invention, the preset response threshold is set to 0.7; in other embodiments of the present invention, the size of the preset response threshold can be set according to specific circumstances, and will not be limited or elaborated here.

[0071] Step S3: Based on the number of measurement points in different void regions and the measurement point cluster region where the measurement point is located, as well as the resonance responsivity, obtain the void region conformity of the measurement point cluster region where each measurement point is located; based on the signal resonance of different measurement points in the neighborhood of each measurement point, obtain the signal complexity of each measurement point; based on the location characteristics and signal complexity of different measurement points, obtain the signal resonance gradient of each measurement point.

[0072] The number of measurement points reflects the size of the region. Analyzing the degree to which the resonance responsivity changes with the size of the region, we assess whether the cluster of measurement points where each measurement point is located conforms to the resonance law of the voided region. Based on the number of measurement points in different voided regions and the cluster of measurement points where each measurement point is located, as well as the resonance responsivity, we obtain the degree of conformity of the voided region of the cluster of measurement points where each measurement point is located.

[0073] Preferably, in one embodiment of the present invention, the method for obtaining the compliance of the void region is described in [reference needed]. Figure 3 It illustrates a flowchart of a method for obtaining the conformity of a void region, including:

[0074] Step S301: Obtain the absolute value of the correlation coefficient of the sequence formed by the number of measuring points and the resonance response in different void regions, as the overall correlation strength;

[0075] It should be noted that in the embodiments of the present invention, the correlation coefficient is the Pearson correlation coefficient, wherein the correlation coefficient ranges from -1 to 1. The larger or smaller the correlation coefficient, the greater the resonance law of the number of measurement points and resonance response in the region. Therefore, the absolute value of the correlation coefficient is taken to reflect the correlation strength. The larger the absolute value, the greater the correlation strength. The specific means are well known to those skilled in the art and will not be described in detail here.

[0076] Step S302: Select the two regions with the smallest difference between the resonance responsivity of other measurement point cluster regions and the measurement point cluster region of each measurement point, and take them as the resonance correlation regions of the measurement point cluster region of each measurement point; obtain the absolute value of the correlation coefficient of the sequence formed by the number of measurement points in different resonance correlation regions and the resonance responsivity, and take it as the local correlation strength.

[0077] Based on the overall correlation strength between the number of measurement points and the resonance response in the void region, the local correlation strength between the number of measurement points and the resonance response in the local area of ​​the measurement point cluster region where each measurement point is located is analyzed, which helps to compare and analyze the deviation between the correlation strength and the void region.

[0078] Step S303: Obtain the difference between the local correlation strength and the overall correlation strength of the cluster region where each measuring point is located, and perform negative correlation mapping as the degree of conformity of the empty region of the cluster region where each measuring point is located.

[0079] It should be noted that, in one embodiment of the present invention, an exponential function with a base of the natural constant is used. When performing negative correlation mapping, the greater the difference, the smaller the result after negative correlation, and the less consistent it is with the overall correlation strength corresponding to the empty region, and the smaller the consistency of the empty region. In other embodiments of the present invention, negative correlation mapping can also be performed by taking the reciprocal. In order to avoid the denominator of the formula being 0 and the formula being meaningless, an artificially set threshold, such as 0.01, is added to the denominator. The specific means are well known to those skilled in the art and will not be described in detail here.

[0080] Due to the presence of randomly and unevenly distributed adhesion points in the tunnel detachment boundary region, the signal resonance becomes increasingly inconsistent, and the local boundary conditions become highly complex, resulting in significant differences in the audio frequency response of adjacent measuring points, exhibiting spatial inconsistency. The signal complexity of each measuring point is obtained based on the signal resonance of different measuring points within the neighborhood of each measuring point.

[0081] Preferably, in one embodiment of the present invention, the method for obtaining signal complexity includes:

[0082] The mean value of the signal resonance difference between adjacent measurement points within the neighborhood of each measurement point is obtained as the signal complexity of each measurement point.

[0083] It should be noted that, in one embodiment of the present invention, the method for obtaining the neighborhood range is to construct a circular range with a radius of 3 based on each measuring point, and analyze the measuring points within the range; in other embodiments of the present invention, the size of the neighborhood range can be set according to specific circumstances, and will not be limited or described here.

[0084] The center of the de-void region usually exhibits a significant resonant response, and the resonance is highly consistent. For measurement points located in the de-void boundary region, the audio frequency response differs significantly from that of different regions, resulting in significant spatial variations in resonance. When the response differences between each measurement point and its neighboring regions are large and similar, the measurement point exhibits obvious resonant gradient changes. Based on the location characteristics and signal complexity of different measurement points, the signal resonant gradient of each measurement point is obtained.

[0085] Preferably, in one embodiment of the present invention, the method for obtaining the signal resonance gradient includes:

[0086] For each measurement point, in all neighboring measurement point clusters within its measurement point cluster, the measurement point with the highest signal resonance within the cluster is selected as a reference point. The average signal complexity of all reference points within the cluster is obtained as the average relative signal complexity of each measurement point. The first difference between the average signal complexity and the average relative signal complexity of each measurement point is then obtained.

[0087] The second difference in signal complexity between each measurement point and the reference point of each adjacent measurement point cluster region is obtained; the mean ratio of the first difference and the second difference corresponding to each measurement point and different adjacent measurement point cluster regions is obtained, the difference between the mean ratio and the positive integer 1 is calculated, and negative correlation mapping is performed as the signal resonance gradient of each measurement point.

[0088] The average value is used to quantify the general level of signal complexity in clustered regions of adjacent measuring points, reflecting the relative overall signal complexity level of each measuring point. The first difference reflects the deviation between the signal complexity of each measuring point and the mean of relative signal complexity. The second difference reflects the deviation of signal complexity between each measuring point and the reference point in each adjacent region. The average ratio of the second difference to the first difference between each measuring point and the reference point in different adjacent regions is calculated to reflect the uniformity of the distribution of signal complexity deviation of each measuring point relative to the adjacent regions. The closer it is to 1, the more similar the signal complexity differences are and the greater the consistency of resonance characteristics.

[0089] The more the mean ratio deviates from the positive integer 1, the more uneven the complexity distribution between the cluster regions of each measuring point and the cluster regions of adjacent measuring points, the more different there are between adjacent regions, the more irregular the distribution of measuring points in the region, and the smaller the resonance gradient.

[0090] Step S4: Based on the conformity of the void region, the signal resonance gradient, and the signal resonance of the cluster region of each measuring point, obtain the void probability of each measuring point and perform void disease identification.

[0091] The higher the degree of gap filling in the cluster region of the measurement point, the more significant the resonance gradient, and the greater the obvious signal resonance transition, the more likely it is to be a gap filling region. Therefore, the gap filling probability of each measurement point is obtained based on the degree of gap filling region conformity, signal resonance gradient, and signal resonance of the cluster region of the measurement point.

[0092] Preferably, in one embodiment of the present invention, the method for obtaining the slippage probability includes:

[0093] The signal resonance transition of each measurement point is obtained by multiplying the signal resonance gradient of each measurement point with the degree of conformity between the empty region of the clustering region of the measurement point.

[0094] Obtain the first sum between the positive integer 1 and the signal resonant transition value, and calculate the product of the first sum value and the signal resonant transition value as the probability of escaping from the point of measurement.

[0095] Based on this, in another embodiment of the present invention, tunnel defects are identified by accurately obtaining the voiding probability of each measuring point. This includes: using a two-dimensional unfolded diagram of the tunnel, flattening the lining surface into a planar schematic, performing DBSCAN clustering based on the final voiding probability of the measuring points, with the cluster having the highest voiding probability as the voiding region, and using color partitioning to intuitively mark the distribution of defects, for example, green represents dense areas and red represents voiding areas, facilitating rapid identification of the range and distribution pattern of defects; on the other hand, combining acoustic detection data, generating a spectrogram, unfolding the continuously detected spectral information along the tunnel mileage direction to form a time-frequency two-dimensional image, where voiding regions, due to differences in resonance characteristics, will exhibit obvious characteristic color bands, further enhancing the accuracy and intuitiveness of defect identification; the specific means are well known to those skilled in the art and will not be elaborated here.

[0096] In summary, this invention obtains the signal resonance of the audio signal corresponding to each measuring point in both the time and frequency domains; obtains multiple measuring point clusters; obtains the resonance responsivity of each measuring point cluster based on the information resonance distribution of different measuring points within each cluster, thus filtering out void regions; obtains the void region conformity of each measuring point based on different void regions, the number of measuring points in each measuring point's cluster, and the resonance responsivity; obtains the signal resonance gradient of each measuring point based on the signal resonance of different measuring points within its neighborhood; and obtains the void probability of each measuring point for void defect identification. This invention improves the accuracy of void defect identification by accurately obtaining the void probability of each measuring point.

[0097] The present invention also proposes a tunnel caving defect identification system based on audio frequency analysis, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any of the steps of a tunnel caving defect identification method based on audio frequency analysis.

[0098] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0099] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A tunnel void disease identification method based on audio frequency analysis, characterized in that, The method comprises: acquiring acoustic frequency signals of each measuring point during tunnel void detection; obtaining signal resonance of each measuring point according to distribution characteristics of the acoustic frequency signals of each measuring point in time domain and frequency domain; clustering all measuring points according to the signal resonance of different measuring points to obtain a plurality of measuring point clustering regions; obtaining resonance response degree of each measuring point clustering region according to information resonance distribution of different measuring points in each measuring point clustering region; and screening out void regions; obtaining void region coincidence degree of the measuring point clustering region where each measuring point is located according to the measuring point quantity and resonance response degree of different void regions and measuring point clustering regions; obtaining signal complexity of each measuring point according to signal resonance of different measuring points in a neighborhood range of each measuring point; and obtaining signal resonance gradient of each measuring point according to position characteristics and signal complexity of different measuring points; obtaining void probability of each measuring point according to the void region coincidence degree, signal resonance gradient and signal resonance of the measuring point clustering region where each measuring point is located, and performing void disease identification; The signal resonance comprises: obtaining any peak value of the acoustic frequency signals of each measuring point in frequency domain, obtaining frequencies at half-peak values on left and right sides of the peak value, and selecting a maximum value of a difference ratio between different peak values and corresponding frequencies; obtaining a fitting straight line of amplitudes of the acoustic frequency signals of each measuring point at different time instants; obtaining signal resonance of each measuring point according to a slope of the fitting straight line and the maximum value of the difference ratio, wherein the slope and the maximum value of the difference ratio are positively correlated with the signal resonance; The void region coincidence degree comprises: obtaining an absolute value of a correlation coefficient of a sequence composed of the measuring point quantity and the resonance response degree in different void regions as an overall correlation strength; selecting two other measuring point clustering regions with the smallest difference between the resonance response degrees of the two other measuring point clustering regions and the measuring point clustering region where each measuring point is located as resonance correlation regions of the measuring point clustering region where each measuring point is located; obtaining an absolute value of a correlation coefficient of a sequence composed of the measuring point quantity and the resonance response degree in different resonance correlation regions as a local correlation strength; obtaining a difference between the local correlation strength and the overall correlation strength of the measuring point clustering region where each measuring point is located, and performing negative correlation mapping as the void region coincidence degree of the measuring point clustering region where each measuring point is located; the negative correlation mapping comprises negative correlation mapping by an exponential function with a natural constant as a base or negative correlation mapping by taking an inverse; The signal resonance gradient comprises: for all adjacent measuring point clustering regions of the measuring point clustering region where each measuring point is located, obtaining a measuring point with the maximum signal resonance in the region as a reference point; obtaining a mean value of signal complexity of all reference points in the region as a relative signal complexity mean value of each measuring point; and obtaining a first difference between the signal complexity and the relative signal complexity mean value of each measuring point. Obtaining the second difference of signal complexity between each measuring point and the reference point of the clustering area of each adjacent measuring point; obtaining the average of the ratio of the first difference and the second difference corresponding to the different adjacent measuring point clustering areas of each measuring point, calculating the difference between the average of the ratio and the positive integer 1, and performing negative correlation mapping as the signal resonance gradient of each measuring point; the negative correlation mapping acquisition method includes negative correlation mapping by an exponential function with a natural constant as the base or negative correlation mapping by taking the reciprocal; The acquisition method of the void probability includes: Obtaining the product between the signal resonance gradient of each measuring point and the void area coincidence degree of the measuring point clustering area where the measuring point is located, and grouping as the signal resonance transition of each measuring point; Obtaining the first sum value between the positive integer 1 and the signal resonance transition, calculating the product of the first sum value and the signal resonance, and taking the product as the void probability of each measuring point.

2. The tunnel void disease identification method based on audio analysis according to claim 1, characterized in that, The acquisition method of the measuring point clustering area includes: According to the signal resonance of different measuring points, DBSCAN clustering is performed on all measuring points to obtain a plurality of measuring point clustering areas.

3. The tunnel void disease identification method based on audio analysis according to claim 1, characterized in that, The acquisition method of the resonance response degree includes: Obtaining the average of the signal resonance of all measuring points in each measuring point clustering area as the average resonance level of each measuring point clustering area; According to the resonance difference between the signal resonance of different measuring points in each measuring point clustering area and the average resonance level, and the signal resonance, the resonance response degree of each measuring point clustering area is obtained, the resonance difference and the resonance response degree are negatively correlated, and the signal resonance and the resonance response degree are positively correlated.

4. The tunnel void disease identification method based on audio analysis according to claim 1, wherein, The acquisition method of the void area includes: If the resonance response degree of the measuring point clustering area is greater than the preset response threshold, the corresponding measuring point clustering area is taken as the void area.

5. The tunnel void disease identification method based on audio analysis according to claim 1, wherein, The acquisition method of the signal complexity includes: Obtaining the average of the signal resonance difference between adjacent measuring points in the neighborhood range of each measuring point as the signal complexity of each measuring point.

6. A tunnel void disease identification system based on audio analysis, the system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor executes the computer program to realize the steps of the tunnel void disease identification method based on audio analysis according to any one of claims 1-5.

Images