CN117782403A - Loose bolt positioning method, device and medium based on separation network - Google Patents
Loose bolt positioning method, device and medium based on separation network Download PDFInfo
- Publication number
- CN117782403A CN117782403A CN202410211781.2A CN202410211781A CN117782403A CN 117782403 A CN117782403 A CN 117782403A CN 202410211781 A CN202410211781 A CN 202410211781A CN 117782403 A CN117782403 A CN 117782403A
- Authority
- CN
- China
- Prior art keywords
- network
- information
- bolt
- target sound
- signal
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000926 separation method Methods 0.000 title claims abstract description 101
- 238000000034 method Methods 0.000 title claims abstract description 71
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 claims abstract description 126
- 238000001514 detection method Methods 0.000 claims abstract description 76
- 229910052742 iron Inorganic materials 0.000 claims abstract description 63
- 238000012423 maintenance Methods 0.000 claims abstract description 24
- 238000012545 processing Methods 0.000 claims description 49
- 238000001228 spectrum Methods 0.000 claims description 46
- 230000000873 masking effect Effects 0.000 claims description 32
- 230000006870 function Effects 0.000 claims description 26
- 230000005236 sound signal Effects 0.000 claims description 25
- 238000005314 correlation function Methods 0.000 claims description 21
- 238000005316 response function Methods 0.000 claims description 16
- 238000013527 convolutional neural network Methods 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 11
- 238000013507 mapping Methods 0.000 claims description 11
- 230000005284 excitation Effects 0.000 claims description 8
- 230000004913 activation Effects 0.000 claims description 5
- 230000008447 perception Effects 0.000 claims description 5
- 238000012549 training Methods 0.000 claims description 5
- 230000005540 biological transmission Effects 0.000 claims description 4
- 238000004088 simulation Methods 0.000 claims description 3
- 238000004422 calculation algorithm Methods 0.000 description 10
- 230000008569 process Effects 0.000 description 9
- 238000010586 diagram Methods 0.000 description 7
- 230000004044 response Effects 0.000 description 7
- 238000005070 sampling Methods 0.000 description 6
- 238000004891 communication Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 229910000831 Steel Inorganic materials 0.000 description 3
- 230000009471 action Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 239000010959 steel Substances 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000006835 compression Effects 0.000 description 2
- 238000007906 compression Methods 0.000 description 2
- 230000002452 interceptive effect Effects 0.000 description 2
- 230000004807 localization Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000009877 rendering Methods 0.000 description 2
- 230000003595 spectral effect Effects 0.000 description 2
- 238000010183 spectrum analysis Methods 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000009194 climbing Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 238000012821 model calculation Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000003647 oxidation Effects 0.000 description 1
- 238000007254 oxidation reaction Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
- 239000013598 vector Substances 0.000 description 1
Landscapes
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The application relates to the field of iron tower detection and discloses a loose bolt positioning method, a device and a medium based on a separation network. In addition, the acquired sound detection signals are separated through the signal separation network, so that noise interference is reduced, and the detection accuracy is further improved. Compared with the prior art, maintenance personnel are not required to detect one by one, and the detection efficiency and the safety of the iron tower are improved.
Description
Technical Field
The application relates to the field of iron tower detection, in particular to a loose bolt positioning method, a loose bolt positioning device and a loose bolt positioning medium based on a separation network.
Background
The iron tower is one of important components in a communication or power transmission system, mainly comprises angle steel and a flange connection structure and is formed by connecting a large number of bolts. However, due to long-term exposure to the wind, rain or snow in nature, the bolts linking the joints bear the effects of vibration or corrosion oxidation, loosening and even falling off can occur, and finally the iron tower collapse accident is caused.
In order to improve the safety of the iron tower, the loose bolts on the iron tower are required to be fastened. At present, most of the iron tower is manually climbed regularly by a detector and the tightness of the bolts is manually checked, the scheme needs to fasten all the iron tower bolts, the maintenance efficiency is low, the labor cost is increased, the personal safety and other problems can be brought, the manual check is often influenced by subjective factors, and the accuracy and reliability requirements of each detection are difficult to meet.
Therefore, how to provide a new method for detecting the loose bolt of the iron tower so as to more accurately and efficiently maintain the loose bolt on the iron tower is a problem which needs to be solved by the person skilled in the art.
Disclosure of Invention
The utility model provides a purpose is in order to solve among the prior art that the manual climbing iron tower of inspector and manual inspection bolt elasticity lead to detection efficiency low, and the accuracy and the low problem of reliability of detection, for this reason, this application provides a loose bolt positioning method, device, medium based on separation network to the position of accurate determination loose bolt.
In order to solve the technical problem, the application provides a loose bolt positioning method based on a separation network, which comprises the following steps:
Acquiring a sound detection signal of an iron tower to be detected;
the sound detection signals are separated by a signal separation network so as to obtain target sound information; wherein the signal separation network is a network comprising a feature masking unit and a decoder based on a convolutional neural network and a determination of a transposed convolutional structure;
weighting the target sound information by using a beam forming task network to acquire spatial spectrum information; the weighted value of the beam forming task network is a value determined according to a characteristic cross-correlation function, wherein the characteristic cross-correlation function is a function obtained by carrying out time sequence association on the target sound information;
and classifying the target sound information by using an event classifier, and determining the position information of the loosened bolt according to the spatial spectrum information when the bolt loosening event is detected.
In some embodiments, the separating the sound detection signal using a signal separation network includes:
performing feature encoding operation on the sound detection signal by using a convolutional neural network so as to acquire initial feature information;
masking the initial feature information by using the feature masking unit to obtain target feature information; the masking codes of the characteristic masking network are masking codes in time and characteristic dimension;
And processing the target characteristic information by using the decoder to acquire the reconstructed target sound information.
In some embodiments, the acquiring the sound detection signal of the iron tower to be tested includes:
acquiring an initial sound signal and an impact response function acquired by a sensor array after pulse excitation is applied to the iron tower to be tested; the impact response function is determined according to a sound transmission path and an acoustic propagation attenuation factor and is used for representing different positions of the iron tower to be tested and different sensors in the sensor array;
the sound detection signal is determined from the impulse response function and the initial sound signal.
In some embodiments, the weighting the target sound information with the beamforming task network to obtain spatial spectrum information includes:
performing time sequence association processing on the target sound information by using a gating circulation unit to acquire time dimension statistical characteristics;
and carrying out weighting processing on the time dimension statistical characteristics, and determining the spatial spectrum information of the weighted time dimension statistical characteristics.
In some embodiments, the event classifier is a multi-layer perceptron consisting of a two-layer fully connected network, the activation function of the output layer of the event classifier being a Softmax function;
The classifying the target sound information by using the event classifier comprises the following steps:
processing the target sound information by using the multi-layer perceptron to obtain a multi-layer perception network mapping value of the target sound information;
processing the multi-layer sensing network mapping value by using the Softmax function to obtain bolt loosening probability;
and determining whether the bolt corresponding to the target sound information is loosened according to the bolt loosening probability.
In some embodiments, before the step of obtaining the sound detection signal of the iron tower to be tested, the method further includes:
and pre-training the signal separation network based on the artificial marked simulation data set to determine the optimal parameters of the signal separation network.
In some embodiments, after the step of determining the position information of the loosened bolt according to the spatial spectrum information when the bolt loosening event is detected, the method further includes:
generating a maintenance task according to the number information and the position information of the iron tower to be tested;
and sending the maintenance task to a manager, and generating a maintenance log according to the feedback information of the maintenance task and the manager.
In order to solve the technical problem, the application also provides a loose bolt positioning device based on a separation network, which comprises:
The acquisition module is used for acquiring a sound detection signal of the iron tower to be detected;
the separation module is used for separating the sound detection signals by utilizing a signal separation network so as to acquire target sound information; wherein the signal separation network is a network comprising a feature masking unit and a decoder based on a convolutional neural network and a determination of a transposed convolutional structure;
the weighting processing module is used for carrying out weighting processing on the target sound information by utilizing a beam forming task network so as to acquire spatial spectrum information; the weighted value of the beam forming task network is a value determined according to a characteristic cross-correlation function, wherein the characteristic cross-correlation function is a function obtained by carrying out time sequence association on the target sound information;
and the classification processing module is used for classifying the target sound information by using an event classifier, and determining the position information of the loosened bolt according to the spatial spectrum information when the bolt loosening event is detected.
In order to solve the technical problem, the application also provides a loose bolt positioning device based on a separation network, which comprises a memory for storing a computer program;
and the processor is used for realizing the step of the loose bolt positioning method based on the separation network when executing the computer program.
In order to solve the above technical problem, the present application further provides a computer readable storage medium, where a computer program is stored, and the computer program when executed by a processor implements the steps of the loose bolt positioning method based on a separation network.
The application provides a loose bolt positioning method based on a separation network, which comprises the following steps: acquiring a sound detection signal of an iron tower to be detected; the method comprises the steps of performing separation processing on a sound detection signal by using a signal separation network so as to obtain target sound information; wherein the signal separation network is a network comprising a feature masking unit and a decoder based on a convolutional neural network and a determination of a transposed convolutional structure; weighting the target sound information by using a beam forming task network to acquire spatial spectrum information; the weighted value of the beam forming task network is a value determined according to a characteristic cross-correlation function, wherein the characteristic cross-correlation function is a function obtained by carrying out time sequence association on target sound information; and classifying the target sound information by using an event classifier, and determining the position information of the loosened bolt according to the spatial spectrum information when the bolt loosening event is detected. Therefore, according to the technical scheme provided by the application, the separated target sound information is weighted through the beam forming task network to acquire the spatial spectrum information of the target sound information, so that the generation position of the target sound information is judged, and when the bolt loosening event is determined according to the target sound information, the specific position of the loosened bolt is determined by combining the spatial spectrum information. In addition, the acquired sound detection signals are separated through the signal separation network, so that noise interference is reduced, and the detection accuracy is further improved. Compared with the prior art, maintenance personnel are not required to detect one by one, and the detection efficiency and the safety of the iron tower are improved.
In addition, the application also provides a loose bolt positioning device and medium based on a separation network, which correspond to the method and have the same effects.
Drawings
For a clearer description of the embodiments of the present application, the drawings that are needed in the embodiments will be briefly described, it being apparent that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a diagram of a scene of detecting loose bolts of an iron tower according to an embodiment of the present application;
FIG. 2 is a flowchart of a loose bolt positioning method based on a separation network according to an embodiment of the present application;
fig. 3 is a block diagram of a signal separation network and a beam forming task network according to an embodiment of the present application;
fig. 4 is a block diagram of a loose bolt positioning device based on a separation network according to an embodiment of the present application;
FIG. 5 is a block diagram of another loose bolt positioning apparatus based on a separation network according to an embodiment of the present application;
the reference numerals are as follows: the device comprises an iron tower to be tested, wherein the iron tower is 1, a bolt loosening point is 2, a bolt loosening point response signal is 3, a steel structure vibration response signal is 4, a sound sensor array is 5, and a pulse excitation source is 6.
Detailed Description
The following description of the technical solutions in the embodiments of the present application will be made clearly and completely with reference to the drawings in the embodiments of the present application, and it is apparent that the described embodiments are only some embodiments of the present application, but not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments herein without making any inventive effort are intended to fall within the scope of the present application.
The core of the application is to provide a loose bolt positioning method, a loose bolt positioning device and a loose bolt positioning medium based on a separation network so as to accurately determine the position of a loose bolt.
In order to ensure the safety of the iron tower, the bolt loosening condition of the iron tower 1 to be detected needs to be detected, and the existing iron tower bolt loosening detection method based on acoustic sensing is mainly used for detecting an iron tower acoustic response signal caused by a pulse excitation source through acoustic sensing deployed on the iron tower, for example, a sound sensor fixed at the bottom of the iron tower is used for recording a received sound signal, and then the frequency of the received signal is counted to judge a bolt loosening event. However, the response acoustic signal of bolt loosening is more complex, and the detection accuracy of bolt loosening event is low by judging only a few resonance frequencies. The loosening position of the bolt can be estimated through the strength of signals received by a plurality of sensors, but the method can be positioned in a near-field model, and if the loosening point of the bolt is far away from the sound sensor or accords with a far-field sound propagation model, the loosening point is difficult to position by a positioning method based on the strength of the sound signal.
Fig. 1 is a view of a detection scenario of loose bolts of an iron tower provided in an embodiment of the present application, as shown in fig. 1, a method for positioning loose bolts based on a separation network according to the present invention may use a manual or electric device to perform pulse excitation, and a vertical sound sensor array 5 is placed near 1 meter from a pulse excitation source 6 to obtain response sound signals of components of each part of the iron tower 1 to be detected (including a bolt loosening point response signal 3 generated by a bolt loosening point 2 and a steel structure vibration response signal 4 generated by other parts of the iron tower 1 to be detected). According to the technical scheme, the separated target sound information is weighted through the beam forming task network to acquire the spatial spectrum information of the target sound information, so that the generation position of the target sound information is judged, and when a bolt loosening event is determined according to the target sound information, the specific position of a loosening bolt is determined by combining the spatial spectrum information. In addition, the acquired sound detection signals are separated through the signal separation network, so that noise interference is reduced, and the detection accuracy is further improved. Compared with the prior art, maintenance personnel are not required to detect one by one, and the detection efficiency and the safety of the iron tower are improved.
In order to provide a better understanding of the present application, those skilled in the art will now make further details of the present application with reference to the drawings and detailed description.
Fig. 2 is a flowchart of a loose bolt positioning method based on a separation network according to an embodiment of the present application, as shown in fig. 2, the loose bolt positioning method based on a separation network provided in the present application includes:
s10: acquiring a sound detection signal of an iron tower to be detected;
s11: the method comprises the steps of performing separation processing on a sound detection signal by using a signal separation network so as to obtain target sound information; wherein the signal separation network is a network comprising a feature masking unit and a decoder based on a convolutional neural network and a determination of a transposed convolutional structure.
It will be appreciated that since more sound signals are collected by the sensors provided at the pylon, which may interfere with subsequent processing, it is necessary to separate the sound detection signals using a signal separation network to obtain sound signals therein for reflecting loosening of the bolts.
Furthermore, in order to improve the positioning accuracy of the loose bolt and reduce the noise signal in the sound detection signal, in a specific implementation, the sound detection signal can be obtained by adopting a non-contact pickup mode (namely, the sound sensor is not contacted with the iron tower to be detected).
In implementations, the separation of sound signals may be accomplished using Fourier transforms, and signal separation may also be performed using other neural network models (e.g., a full convolution time domain audio separation network). A separation network as used herein is a network comprising feature masking units and decoders based on a convolutional neural network and a determination of a transposed convolutional structure.
It should be noted that in the technical scheme provided by the application, a plurality of sound sensors are arranged at the bottom of the iron tower to be tested so as to collect sound signals. In the signal separation process, the signals input by each path of sound sensor are required to be separated separately, and the sound detection signals of different sound sensors cannot be mixed.
S12: weighting the target sound information by using a beam forming task network to acquire spatial spectrum information; the weighted value of the beam forming task network is a value determined according to a characteristic cross-correlation function, wherein the characteristic cross-correlation function is a function obtained by carrying out time sequence association on target sound information;
s13: and classifying the target sound information by using an event classifier, and determining the position information of the loosened bolt according to the spatial spectrum information when the bolt loosening event is detected.
The beam forming device is an important component in digital signal processing and is widely applied to the fields of radar, sonar, wireless communication, medical imaging and the like. The main function of the method is to carry out weighting and phase adjustment on signals from different directions, thereby achieving the purpose of enhancing the signal strength or suppressing interference signals in a specific direction. The process of acquiring spatial spectrum using a beamformer includes: each sound detection signal is weighted and phase-adjusted. These weights and phases are typically determined based on the direction of the target signal and the location of the interfering signal. This process may use different beamforming algorithms, such as a minimum mean square error (Mean Squared Error, MSE) algorithm, a maximum Signal-to-Noise Ratio (SNR) algorithm, or an adaptive beamforming algorithm. By adjusting the weights and phases, the beamformer can form a "beam" directed in the direction of the target signal. This beam will increase the signal strength in the target direction and suppress the interfering signal in the other direction. Finally, by further processing the output of the beamformer, an estimate of the spatial spectrum can be obtained. This typically involves spectral analysis of the beamformer output, for example by calculating the power spectral density or performing a correlation function analysis. This spatial spectrum estimate may provide information about signal direction, signal strength, and interference source location.
Fig. 3 is a block diagram of a signal separation network and a beam forming task network according to an embodiment of the present application, as shown in fig. 3, delay compensation is performed on a sound detection signal acquired by a sound sensor, and separation processing is performed on the sound detection signal by using a full convolution time domain audio separation network (Conv-TasNet) to acquire separated target sound information. And then respectively transmitting the separated target sound information to a classifier so as to classify the separated target sound information by the classifier, thereby determining whether a bolt loosening event exists. Meanwhile, the separated target sound information is subjected to time sequence association and multi-layer perception mapping to obtain a beam forming weighted value of the array, and then the weighted value is multiplied with an input signal to obtain the bolt loosening sound signal in the designated direction. And finally, obtaining a spatial spectrum according to the output bolt loosening signals at different positions, so that a manager can determine the specific position of the loosened bolt according to the spatial spectrum.
The application provides a loose bolt positioning method based on a separation network, which comprises the following steps: acquiring a sound detection signal of an iron tower to be detected; the method comprises the steps of performing separation processing on a sound detection signal by using a signal separation network so as to obtain target sound information; wherein the signal separation network is a network comprising a feature masking unit and a decoder based on a convolutional neural network and a determination of a transposed convolutional structure; weighting the target sound information by using a beam forming task network to acquire spatial spectrum information; the weighted value of the beam forming task network is a value determined according to a characteristic cross-correlation function, wherein the characteristic cross-correlation function is a function obtained by carrying out time sequence association on target sound information; and classifying the target sound information by using an event classifier, and determining the position information of the loosened bolt according to the spatial spectrum information when the bolt loosening event is detected. Therefore, according to the technical scheme provided by the application, the separated target sound information is weighted through the beam forming task network to acquire the spatial spectrum information of the target sound information, so that the generation position of the target sound information is judged, and when the bolt loosening event is determined according to the target sound information, the specific position of the loosened bolt is determined by combining the spatial spectrum information. In addition, the acquired sound detection signals are separated through the signal separation network, so that noise interference is reduced, and the detection accuracy is further improved. Compared with the prior art, maintenance personnel are not required to detect one by one, and the detection efficiency and the safety of the iron tower are improved.
As a preferred embodiment, the separation processing of the sound detection signal using the signal separation network includes: performing feature encoding operation on the sound detection signal by using a convolutional neural network so as to acquire initial feature information; masking the initial feature information by using a feature masking unit to obtain target feature information; the masking codes of the feature masking network are masking codes in time and feature dimension; and processing the target characteristic information by using a decoder to acquire reconstructed target sound information.
The full convolution time domain audio separation network (Conv-TasNet) consists of three processing stages: encoder, separator and decoder. First, the encoder module converts short segments of the mixed signal into corresponding representations in the intermediate feature space. The representation is then used to estimate the multiplication function (mask) for each source. Finally, the decoder module reconstructs the source waveform by converting the masked encoder characteristics.
The full convolution time domain audio separation network directly adopts the sound signals picked up by the sound sensor as the input of the model, and the end-to-end sound signal classification is realized. The full convolution time domain audio frequency separation network mainly comprises a convolution neural network, firstly, input signals are subjected to convolution coding, then masking codes in time and characteristic dimension are generated through multi-jump linkage of a multi-layer convolution network, and finally, signals are reconstructed through a decoder. The full convolution time domain audio separation network can be simplified to the following steps:
1. Convolution encoding operation:
；
first, convolutionally encoding an input sound detection signal according to the above, whereinFor the product network operation, ++>For linear mapping weighting matrix +.>To activate the function, thereby achieving a code compression of the input signal,/->The result is output for convolutional encoding. The encoder is similar to a fourier transform, but since the model is derived by data learning, its compression encoding is more efficient.
2. Feature masking operation:
；
generating masking codes in time and feature dimensions via multi-layer convolutional network multi-hop chainingWhereinIs a time convolution network for acquiring features in the time dimension. Finally pass->The activation function obtains feature masking weights. Thereby realizing the separation and enhancement of the loosening signal of the bolt.
3. Decoding operation:
；
performing a decoding operation on the output signal, whereinFor deconvolution networks, the masked features are reconstructed into decoded signals. />Providing more accurate characteristics for classifying and positioning subsequent bolt loosening events for separated target sound information>Is the real number domain>Is->Go->Real number field matrix of columns,/->For the number of sampling points, +.>Is the total number of sound sensors.
In some embodiments, obtaining the sound detection signal of the iron tower to be tested includes: acquiring an initial sound signal and an impact response function acquired by a sensor array after pulse excitation is applied to an iron tower to be tested; the impact response function is determined according to the sound transmission path and the sound propagation attenuation factor and is used for representing different positions of the iron tower to be tested and different sensors in the sensor array; a sound detection signal is determined from the impulse response function and the initial sound signal.
The initial sound signal is sound which is collected by the sound sensor and sent out by the iron tower to be tested. Because the sound that awaits measuring the iron tower sent needs to pass through the transfer path and just can transmit to sensor department, external factor can cause the influence to the sound signal in the transfer path, the impact response function that mentions in this scheme is used for simulating these external factor's influence promptly. The sound detection signal is information obtained after the sound sensor processes the initial sound signal by using an impulse response function.
The sound detection signals acquired by each sound sensor can be modeled as a group of sound detection signals arranged in the setSound sources at different sampling points, for example: the sound sensor with serial number 1 is +.>Sound detection signal acquired at the moment +.>The method comprises the following steps: the sound sensor with serial number 1 is +.>Initial sound signal acquired at the moment +.>The signal set obtained after convolution with the impulse response function is therefore in +.>Time->Sound detection signal set received by sound sensor，/>For the number of sampling points, +.>Serial number of sound sensor, +.>Characterization->Real number domain matrix of row 1 column, +.>For the total number of sound sensors, +.>Is->Time number->The sound detection signals acquired by the sensor of (2), wherein the sound detection signal set +. >The method comprises the following steps:
；
wherein the method comprises the steps ofCharacterization of the->Sound source propagation to the +.>An impulse response function corresponding to the individual sound sensor,/->Is the first part of the iron tower>Sound source propagation to the +.>An impulse response function corresponding to the individual sound sensor,/->The specific formula is as follows:
；
wherein the method comprises the steps ofIs iron tower->The path of the sampling points to the sound sensor, < >>For the corresponding acoustic propagation attenuation factor +.>For time (I)>Is the speed of sound,ifor the number of the path of the sampling point to the sound sensor,/->Is the total number of paths from the sampling point to the sound sensor.
In a specific implementation, weighting the target sound information by using the beam forming task network to obtain spatial spectrum information includes: performing time sequence association processing on the target sound information by using a gating circulation unit to acquire time dimension statistical characteristics; and carrying out weighting processing on the time dimension statistical characteristics, and determining spatial spectrum information of the weighted time dimension statistical characteristics.
Beamforming task network forProcessing the target sound information separated by each channel, performing time series association by a gating circulation unit (gate recurrent unit, GRU), and then counting the characteristics of the target sound information in the time dimension to obtain the time dimension statistical characteristics ，
；
The process is equivalent to the inverse calculation of the autocorrelation matrix, and has better nonlinear adaptability and convergence. Then, the weighting values of the beam formerCan be simplified as a Multi-Layer Perceptron (MLP) model:
；
Lthe signals of the loosening bolts at different positions are weighted and overlapped to obtain weighted signals according to the following formulaThe method comprises the following steps:
；
finally, the spatial spectral distribution of the bolt loosening acoustic signalThe method comprises the following steps:
；
wherein,is->First->Column vectors.
It should be noted that, in order to improve the accuracy of the loose bolt positioning method provided by the application, the sound signal separation network and the beam forming task network need to be pre-trained through the manually marked public data set and the manually marked simulation data set so as to determine the optimal parameters of the signal separation network, so that the generalization and anti-interference capability of the model can be further improved.
In some embodiments, the event classifier is a multi-layer perceptron consisting of two layers of fully connected networks, the activation function of the output layer of the event classifier is a Softmax function; the classifying processing of the target sound information by using the event classifier comprises: processing the target sound information by using a multi-layer sensor to obtain a multi-layer sensing network mapping value of the target sound information; processing the multi-layer sensing network mapping value by using a Softmax function to obtain bolt loosening probability; and determining whether the bolt corresponding to the target sound information is loosened according to the bolt loosening probability.
Finally selecting Softmax activation function at the output layer of the neural network to estimate the probability of different events and the loosening probability of boltsCan be expressed as:
；
wherein the method comprises the steps ofThe method is a result obtained by processing the separated target sound information by using the multi-layer perception network mapping function.
In particular implementation, when the bolt loosens the probabilityGreater than a predetermined probability thresholdAt this value, a bolt loosening event is considered to be present. For example: if the preset probability threshold is 0.8, if the bolt loosening probability is 0.7, the bolt is considered not to be loosened. When the bolt loosening probability is 0.9, the bolt is considered to have loosened.
When it is determined that a bolt loosening event has occurred, it is also necessary to determine the specific location of the loosened bolt in combination with spatial spectrum information. Specifically, the acquired target sound information is first analyzed to determine frequency information and phase information of the target sound information in different directions.
The spatial spectrum analysis results are then processed using a source localization algorithm to determine the source direction and position of the noise signal, which is determined as the specific position of the loose bolt. Common source localization algorithms include, among others, high resolution based algorithms and signal arrival time difference based algorithms. The high resolution based algorithm receives phase difference information of signals through the array and utilizes the phase delay of signal propagation and the geometric relationship of the array to determine the position of the signal source. The algorithm based on the signal arrival time difference is to determine the position of the signal source by comparing the time delay differences of the signal reaching different sensors of the array and using the methods such as triangulation or Doppler shift.
In addition, in the training process of the separation network and the beam forming task network, in order to further improve the reliability of the network and further improve the accuracy of the loose bolt position determined according to the target sound signal in the scheme of the application, the calculated loose bolt position can be compared in a modeling mode. For example: an environment model (such as a sound propagation model and the like) can be built aiming at the environment of the iron tower to be tested so as to predict and simulate the propagation process and scattering condition of noise signals, and the result obtained by source positioning is compared with the result obtained by model calculation so as to detect the accuracy of the positioning result of the loose bolt, and a manager is not required to collect data on site, so that the labor cost is reduced. Meanwhile, the data volume of the training data set is increased, and the training effect is effectively improved.
Further, when the bolt loosening event is detected, after the step of determining the position information of the loosened bolt according to the spatial spectrum information, the method further comprises the following steps: and generating maintenance tasks according to the number information and the position information of the iron tower to be tested, so that management staff can maintain the loosened bolts in time. Meanwhile, the maintenance task is sent to the manager, and a maintenance log is generated according to the feedback information of the maintenance task and the manager.
In specific implementation, pulse excitation is firstly applied to the iron tower to be detected, then an initial sound signal and an impulse response function acquired by the sensor array are acquired, and a sound detection signal is determined according to the impulse response function and the initial sound signal. The sound sensor array is in non-contact with the iron tower to be detected, and sound detection signals are obtained in a non-contact pickup mode.
Then, performing feature coding operation on the sound detection signals by utilizing a pre-trained signal separation network so as to acquire initial feature information; masking the initial feature information by using a feature masking unit to obtain target feature information; and then processing the target characteristic information by using a decoder to acquire reconstructed target sound information so as to acquire a sound signal capable of reflecting the loosening condition of the bolt.
Performing time sequence association processing on the target sound information by using a gating circulation unit to acquire time dimension statistical characteristics; and carrying out weighting processing on the time dimension statistical characteristics, and determining spatial spectrum information of the weighted time dimension statistical characteristics. And processing the target sound information by using the multi-layer perceptron to obtain a multi-layer perception network mapping value of the target sound information. Then, the multi-layer sensing network mapping value is processed by using a Softmax function so as to obtain bolt loosening probability; and judging whether the bolt loosening probability value is larger than a preset probability threshold value, so as to determine whether a bolt loosening event occurs. If a bolt loosening event occurs, determining a bolt of the iron tower to be tested corresponding to the target sound information by using the acquired spatial spectrum information. And then, acquiring the number information of the iron tower with the bolt loosening event and the corresponding position information of the loosened bolt so as to generate a maintenance task. And sending the maintenance task to a manager so as to facilitate the manager to maintain in time.
In addition, after the manager executes the maintenance task, a maintenance log is generated according to the feedback information (including the completion condition of the maintenance task, whether the bolt loosening event is accurate, whether the position of the loosened bolt is accurate, etc.) of the maintenance task and the manager, so that the subsequent system is updated, and the reliability of the system is further improved.
In the above embodiments, the detailed description is given to the loose bolt positioning method based on the separation network, and the application also provides a corresponding embodiment of the loose bolt positioning device based on the separation network. It should be noted that the present application describes an embodiment of the device portion from two angles, one based on the angle of the functional module and the other based on the angle of the hardware.
Fig. 4 is a block diagram of a loose bolt positioning device based on a separation network according to an embodiment of the present application, as shown in fig. 4, the loose bolt positioning device based on a separation network provided in the present application includes:
the acquisition module 10 is used for acquiring a sound detection signal of the iron tower to be detected;
a separation module 11, configured to perform separation processing on the sound detection signal by using a signal separation network, so as to obtain target sound information; wherein the signal separation network is a network comprising a feature masking unit and a decoder based on a convolutional neural network and a determination of a transposed convolutional structure;
A weighting processing module 12, configured to perform weighting processing on the target sound information by using the beam forming task network, so as to obtain spatial spectrum information; the weighted value of the beam forming task network is a value determined according to a characteristic cross-correlation function, wherein the characteristic cross-correlation function is a function obtained by carrying out time sequence association on target sound information;
the classification processing module 13 is configured to perform classification processing on the target sound information by using an event classifier, and determine the position information of the loosened bolt according to the spatial spectrum information when the bolt loosening event is detected.
Since the embodiments of the apparatus portion and the embodiments of the method portion correspond to each other, the embodiments of the apparatus portion are referred to the description of the embodiments of the method portion, and are not repeated herein.
The application provides a loose bolt positioner based on separation network, include: acquiring a sound detection signal of an iron tower to be detected; the method comprises the steps of performing separation processing on a sound detection signal by using a signal separation network so as to obtain target sound information; wherein the signal separation network is a network comprising a feature masking unit and a decoder based on a convolutional neural network and a determination of a transposed convolutional structure; weighting the target sound information by using a beam forming task network to acquire spatial spectrum information; the weighted value of the beam forming task network is a value determined according to a characteristic cross-correlation function, wherein the characteristic cross-correlation function is a function obtained by carrying out time sequence association on target sound information; and classifying the target sound information by using an event classifier, and determining the position information of the loosened bolt according to the spatial spectrum information when the bolt loosening event is detected. Therefore, according to the technical scheme provided by the application, the separated target sound information is weighted through the beam forming task network to acquire the spatial spectrum information of the target sound information, so that the generation position of the target sound information is judged, and when the bolt loosening event is determined according to the target sound information, the specific position of the loosened bolt is determined by combining the spatial spectrum information. In addition, the acquired sound detection signals are separated through the signal separation network, so that noise interference is reduced, and the detection accuracy is further improved. Compared with the prior art, maintenance personnel are not required to detect one by one, and the detection efficiency and the safety of the iron tower are improved.
Fig. 5 is a block diagram of another loose bolt positioning device based on a separation network according to an embodiment of the present application, as shown in fig. 5, the loose bolt positioning device based on the separation network includes: a memory 20 for storing a computer program;
a processor 21 for implementing the steps of the loose bolt positioning method based on the separation network according to the above embodiment when executing a computer program.
The loose bolt positioning device based on the separation network provided by the embodiment can include, but is not limited to, a smart phone, a tablet computer, a notebook computer, a desktop computer or the like.
Processor 21 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 21 may be implemented in hardware in at least one of a digital signal processor (Digital Signal Processor, DSP), a Field programmable gate array (Field-Programmable Gate Array, FPGA), a programmable logic array (Programmable Logic Array, PLA). The processor 21 may also comprise a main processor, which is a processor for processing data in an awake state, also called central processor (Central Processing Unit, CPU), and a coprocessor; a coprocessor is a low-power processor for processing data in a standby state. In some embodiments, the processor 21 may be integrated with an image processor (Graphics Processing Unit, GPU) for taking care of rendering and rendering of the content that the display screen is required to display. In some embodiments, the processor 21 may also include an artificial intelligence (Artificial Intelligence, AI) processor for processing computing operations related to machine learning.
Memory 20 may include one or more computer-readable storage media, which may be non-transitory. Memory 20 may also include high-speed random access memory, as well as non-volatile memory, such as one or more magnetic disk storage devices, flash memory storage devices. In this embodiment, the memory 20 is at least used for storing a computer program 201, which, when loaded and executed by the processor 21, enables the implementation of the relevant steps of the loose bolt positioning method based on the split network disclosed in any of the foregoing embodiments. In addition, the resources stored in the memory 20 may further include an operating system 202, data 203, and the like, where the storage manner may be transient storage or permanent storage. The operating system 202 may include Windows, unix, linux, among others. The data 203 may include, but is not limited to, sound detection signals, target sound information, spatial spectrum information, and the like.
In some embodiments, the loose bolt positioning device based on the separation network can further comprise a display screen 22, an input/output interface 23, a communication interface 24, a power supply 25 and a communication bus 26.
It will be appreciated by those skilled in the art that the structure shown in fig. 5 is not limiting of a split network based loose bolt locating apparatus and may include more or fewer components than illustrated.
The loose bolt positioning device based on the separation network, provided by the embodiment of the application, comprises a memory and a processor, wherein when the processor executes a program stored in the memory, the processor can realize the following method: acquiring a sound detection signal of an iron tower to be detected; the method comprises the steps of performing separation processing on a sound detection signal by using a signal separation network so as to obtain target sound information; wherein the signal separation network is a network comprising a feature masking unit and a decoder based on a convolutional neural network and a determination of a transposed convolutional structure; weighting the target sound information by using a beam forming task network to acquire spatial spectrum information; the weighted value of the beam forming task network is a value determined according to a characteristic cross-correlation function, wherein the characteristic cross-correlation function is a function obtained by carrying out time sequence association on target sound information; and classifying the target sound information by using an event classifier, and determining the position information of the loosened bolt according to the spatial spectrum information when the bolt loosening event is detected.
Finally, the present application also provides a corresponding embodiment of the computer readable storage medium. The computer-readable storage medium has stored thereon a computer program which, when executed by a processor, performs the steps as described in the method embodiments above.
It will be appreciated that the methods of the above embodiments, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored on a computer readable storage medium. Based on such understanding, the technical solution of the present application may be embodied essentially or in part or all of the technical solution contributing to the prior art, or in a software product stored in a storage medium, performing all or part of the steps of the methods of the various embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
The loose bolt positioning method, the loose bolt positioning device and the loose bolt positioning medium based on the separation network are described in detail. In the description, each embodiment is described in a progressive manner, and each embodiment is mainly described by the differences from other embodiments, so that the same similar parts among the embodiments are mutually referred. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section. It should be noted that it would be obvious to those skilled in the art that various improvements and modifications can be made to the present application without departing from the principles of the present application, and such improvements and modifications fall within the scope of the claims of the present application.
It should also be noted that in this specification, relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises an element.
Claims (10)
1. The loose bolt positioning method based on the separation network is characterized by comprising the following steps of:
acquiring a sound detection signal of an iron tower to be detected;
the sound detection signals are separated by a signal separation network so as to obtain target sound information; wherein the signal separation network is a network comprising a feature masking unit and a decoder based on a convolutional neural network and a determination of a transposed convolutional structure;
Weighting the target sound information by using a beam forming task network to acquire spatial spectrum information; the weighted value of the beam forming task network is a value determined according to a characteristic cross-correlation function, wherein the characteristic cross-correlation function is a function obtained by carrying out time sequence association on the target sound information;
and classifying the target sound information by using an event classifier, and determining the position information of the loosened bolt according to the spatial spectrum information when the bolt loosening event is detected.
2. The loose bolt positioning method based on a separation network according to claim 1, wherein the separating the sound detection signal using a signal separation network comprises:
performing feature encoding operation on the sound detection signal by using a convolutional neural network so as to acquire initial feature information;
masking the initial feature information by using the feature masking unit to obtain target feature information; the masking codes of the characteristic masking network are masking codes in time and characteristic dimension;
and processing the target characteristic information by using the decoder to acquire the reconstructed target sound information.
3. The method for positioning loose bolts based on a separation network according to claim 1, wherein the step of obtaining the sound detection signal of the iron tower to be tested comprises the steps of:
acquiring an initial sound signal and an impact response function acquired by a sensor array after pulse excitation is applied to the iron tower to be tested; the impact response function is determined according to a sound transmission path and an acoustic propagation attenuation factor and is used for representing different positions of the iron tower to be tested and different sensors in the sensor array;
the sound detection signal is determined from the impulse response function and the initial sound signal.
4. The method of claim 2, wherein weighting the target sound information with a beam forming task network to obtain spatial spectrum information comprises:
performing time sequence association processing on the target sound information by using a gating circulation unit to acquire time dimension statistical characteristics;
and carrying out weighting processing on the time dimension statistical characteristics, and determining the spatial spectrum information of the weighted time dimension statistical characteristics.
5. The loose bolt positioning method based on the separation network according to claim 2, wherein the event classifier is a multi-layer sensor consisting of two layers of fully connected networks, and the activation function of the output layer of the event classifier is a Softmax function;
the classifying the target sound information by using the event classifier comprises the following steps:
processing the target sound information by using the multi-layer perceptron to obtain a multi-layer perception network mapping value of the target sound information;
processing the multi-layer sensing network mapping value by using the Softmax function to obtain bolt loosening probability;
and determining whether the bolt corresponding to the target sound information is loosened according to the bolt loosening probability.
6. The method for positioning loose bolts based on a separation network according to claim 1, wherein before the step of obtaining the sound detection signal of the iron tower to be tested, further comprises:
and pre-training the signal separation network based on the artificial marked simulation data set to determine the optimal parameters of the signal separation network.
7. The method for positioning loose bolts based on a separation network according to any one of claims 1 to 6, wherein after the step of determining the position information of the loose bolts according to the spatial spectrum information when the bolt loosening event is detected, further comprising:
Generating a maintenance task according to the number information and the position information of the iron tower to be tested;
and sending the maintenance task to a manager, and generating a maintenance log according to the feedback information of the maintenance task and the manager.
8. Loose bolt positioning device based on separation network, characterized by comprising:
the acquisition module is used for acquiring a sound detection signal of the iron tower to be detected;
the separation module is used for separating the sound detection signals by utilizing a signal separation network so as to acquire target sound information; wherein the signal separation network is a network comprising a feature masking unit and a decoder based on a convolutional neural network and a determination of a transposed convolutional structure;
the weighting processing module is used for carrying out weighting processing on the target sound information by utilizing a beam forming task network so as to acquire spatial spectrum information; the weighted value of the beam forming task network is a value determined according to a characteristic cross-correlation function, wherein the characteristic cross-correlation function is a function obtained by carrying out time sequence association on the target sound information;
and the classification processing module is used for classifying the target sound information by using an event classifier, and determining the position information of the loosened bolt according to the spatial spectrum information when the bolt loosening event is detected.
9. A loose bolt positioning device based on a separation network, which is characterized by comprising a memory for storing a computer program;
a processor for implementing the steps of the split network based loose bolt locating method according to any one of claims 1 to 7 when executing the computer program.
10. A computer readable storage medium, characterized in that the computer readable storage medium has stored thereon a computer program which, when executed by a processor, implements the steps of the loose bolt positioning method based on a split network according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410211781.2A CN117782403B (en) | 2024-02-27 | 2024-02-27 | Loose bolt positioning method, device and medium based on separation network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410211781.2A CN117782403B (en) | 2024-02-27 | 2024-02-27 | Loose bolt positioning method, device and medium based on separation network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117782403A true CN117782403A (en) | 2024-03-29 |
| CN117782403B CN117782403B (en) | 2024-05-10 |
Family
ID=90396703
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410211781.2A Active CN117782403B (en) | 2024-02-27 | 2024-02-27 | Loose bolt positioning method, device and medium based on separation network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117782403B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118032937A (en) * | 2024-04-12 | 2024-05-14 | 南京土星信息科技有限公司 | System for detecting looseness voiceprint of bolt of power transmission tower |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103529471A (en) * | 2013-10-10 | 2014-01-22 | 国家电网公司 | Method and device for rapidly detecting loosening of bolt of iron tower |
| CN109839612A (en) * | 2018-08-31 | 2019-06-04 | 大象声科（深圳）科技有限公司 | Sounnd source direction estimation method based on time-frequency masking and deep neural network |
| CN211784228U (en) * | 2020-02-28 | 2020-10-27 | 西安工程大学 | Power transmission tower bolt loosening fault monitoring device |
| WO2021159775A1 (en) * | 2020-02-11 | 2021-08-19 | 腾讯科技（深圳）有限公司 | Training method and device for audio separation network, audio separation method and device, and medium |
| CN114777986A (en) * | 2022-03-11 | 2022-07-22 | 国网河北能源技术服务有限公司 | Transmission tower bolt looseness detection method and system based on correlation coefficient |
| CN114972205A (en) * | 2022-05-05 | 2022-08-30 | 湖南中登科技有限公司 | Track bolt looseness detection method and device |
| CN115127797A (en) * | 2022-07-06 | 2022-09-30 | 郑州铁路职业技术学院 | Acoustic-based steel rail bolt looseness detection device |
| CN115565548A (en) * | 2022-09-20 | 2023-01-03 | 中国电信股份有限公司 | Abnormal sound detection method, abnormal sound detection device, storage medium and electronic equipment |
| US20230388737A1 (en) * | 2022-05-27 | 2023-11-30 | Shure Acquisition Holdings, Inc. | Inferring characteristics of physical enclosures using a plurality of audio signals |
| CN117373477A (en) * | 2023-10-20 | 2024-01-09 | 东南大学 | Voice separation method and equipment based on multichannel full convolution time domain network |
| CN117388383A (en) * | 2023-10-07 | 2024-01-12 | 北京理工大学唐山研究院 | Method and device for detecting loosening state of threaded connection structure based on knocking method |
-
2024
- 2024-02-27 CN CN202410211781.2A patent/CN117782403B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103529471A (en) * | 2013-10-10 | 2014-01-22 | 国家电网公司 | Method and device for rapidly detecting loosening of bolt of iron tower |
| CN109839612A (en) * | 2018-08-31 | 2019-06-04 | 大象声科（深圳）科技有限公司 | Sounnd source direction estimation method based on time-frequency masking and deep neural network |
| WO2021159775A1 (en) * | 2020-02-11 | 2021-08-19 | 腾讯科技（深圳）有限公司 | Training method and device for audio separation network, audio separation method and device, and medium |
| CN211784228U (en) * | 2020-02-28 | 2020-10-27 | 西安工程大学 | Power transmission tower bolt loosening fault monitoring device |
| CN114777986A (en) * | 2022-03-11 | 2022-07-22 | 国网河北能源技术服务有限公司 | Transmission tower bolt looseness detection method and system based on correlation coefficient |
| CN114972205A (en) * | 2022-05-05 | 2022-08-30 | 湖南中登科技有限公司 | Track bolt looseness detection method and device |
| US20230388737A1 (en) * | 2022-05-27 | 2023-11-30 | Shure Acquisition Holdings, Inc. | Inferring characteristics of physical enclosures using a plurality of audio signals |
| CN115127797A (en) * | 2022-07-06 | 2022-09-30 | 郑州铁路职业技术学院 | Acoustic-based steel rail bolt looseness detection device |
| CN115565548A (en) * | 2022-09-20 | 2023-01-03 | 中国电信股份有限公司 | Abnormal sound detection method, abnormal sound detection device, storage medium and electronic equipment |
| CN117388383A (en) * | 2023-10-07 | 2024-01-12 | 北京理工大学唐山研究院 | Method and device for detecting loosening state of threaded connection structure based on knocking method |
| CN117373477A (en) * | 2023-10-20 | 2024-01-09 | 东南大学 | Voice separation method and equipment based on multichannel full convolution time domain network |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118032937A (en) * | 2024-04-12 | 2024-05-14 | 南京土星信息科技有限公司 | System for detecting looseness voiceprint of bolt of power transmission tower |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117782403B (en) | 2024-05-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117782403B (en) | Loose bolt positioning method, device and medium based on separation network | |
| CN112526451B (en) | Compressed beam forming and system based on microphone array imaging | |
| CN111442827B (en) | Optical fiber passive online monitoring system for transformer winding vibration | |
| Mrówczyńska et al. | Compression of results of geodetic displacement measurements using the PCA method and neural networks | |
| Gaidai et al. | A novel bio-system reliability approach for multi-state COVID-19 epidemic forecast | |
| Chen et al. | Spatial–temporal convolutional gated recurrent unit network for significant wave height estimation from shipborne marine radar data | |
| Avanzato et al. | A convolutional neural networks approach to audio classification for rainfall estimation | |
| Wang et al. | Prediction model of natural gas pipeline crack evolution based on optimized DCNN-LSTM | |
| Peso Parada et al. | Using Gaussian mixture models to detect and classify dolphin whistles and pulses | |
| KR20150044703A (en) | Apparatus and method for measuring maritime information using radar | |
| KR102314824B1 (en) | Acoustic event detection method based on deep learning | |
| CN111695461B (en) | Radar clutter intelligent classification method based on image frequency characteristics | |
| CN116609440A (en) | Intelligent acceptance management method and system for building engineering quality based on cloud edge cooperation | |
| Tian et al. | Underwater multi-target passive detection based on transient signals using adaptive empirical mode decomposition | |
| Vorobeva et al. | Benchmarking microbarom radiation and propagation model against infrasound recordings: a vespagram-based approach | |
| CN109901021B (en) | Multi-scale SVD (singular value decomposition) transformation method based on two-dimensional Hankel matrix | |
| Robinson et al. | Terminology for ocean ambient noise monitoring | |
| CN112380949B (en) | Method and system for detecting arrival time of micro-seismic waves | |
| Sha et al. | Bayesian sonar detection performance prediction with source position uncertainty using SWellEx-96 vertical array data | |
| CN118032120A (en) | Method, device and medium for positioning loose bolt | |
| Huang et al. | Research on structural sound source localization method by neural network | |
| Raffa et al. | Coupling of Wave Data and Underwater Acoustic Measurements in a Maritime High-Traffic Coastal Area: A Case Study in the Strait of Sicily | |
| CN117825898B (en) | GIS distributed vibration and sound combined monitoring method, device and medium | |
| Wang et al. | Research on Underwater Sound Source Localization Based on NLM‐EEMD and FCM‐Generalized Quadratic Correlation | |
| Herselman et al. | Evaluation and performance comparison of detection algorithms in a maritime environment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |