Self-adaptive target threshold determination method and device
Through the adaptive target threshold determination method, multi-layer wavelet decomposition and threshold processing technology are used to solve the problem of periodic noise interference in the cavity sonar signal, and the efficient denoising of the signal and improving the signal-to-noise ratio are achieved.
Patent Information
- Application Number
- CN202411661167.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-20
- Publication Date
- 2025-02-07
AI Technical Summary
The periodic noise generated by vibration in the chamber sonar signal interferes with the echo signal, and the existing filtering methods are difficult to completely remove, resulting in a degradation of signal performance.
Adaptive target threshold determination method is adopted to obtain low-frequency coefficients and high-frequency coefficients through multi-layer wavelet decomposition, and use different initial thresholds for threshold processing and signal reconstruction. The target threshold is determined based on the noise level of the reconstruction signal, which is used for wavelet filtering to remove noise. .
The periodic noise in the cavity sonar signal is effectively removed, the signal-to-noise ratio of the signal is improved, and the wavelet filtering effect is optimized.
Smart Images
Figure CN119395673A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the field of signal processing technology, and in particular to a method and device for determining an adaptive target threshold. Background Art
[0002] Cavity sonar is an instrument that measures the shape and volume of caves through acoustic echo technology. It is widely used in the monitoring and maintenance of salt caverns and underground reservoirs. Cavity sonar relies on the propagation of sound waves and echo signals to determine distance and shape. In practical applications, cavity sonar signals are usually interfered by various noises. In order to ensure that the cavity sonar system can accurately measure the geometric shape in complex environments such as salt caverns, the echo signals collected by the cavity sonar must be subjected to noise reduction processing.
[0003] For cavity sonar signals, the noise mainly comes from the vibration generated by the mechanical moving parts (such as the drive motor) inside the cavity sonar system. This vibration is coupled to the sonar system through the structure, thereby generating periodic noise that interferes with the received useful echo signal. Periodic noise can significantly reduce the performance of the cavity sonar system. Since this noise is concentrated on a specific frequency in the frequency domain, it may overlap with the spectrum of the useful signal. Commonly used Kalman and wavelet filtering methods are difficult to completely filter out. Therefore, it is urgent to develop a noise reduction method suitable for removing the periodic noise generated by vibration in cavity sonar signals. Summary of the invention
[0004] In view of this, embodiments of the present disclosure provide at least one method and device for determining an adaptive target threshold.
[0005] Specifically, the embodiments of the present disclosure are implemented through the following technical solutions:
[0006] In a first aspect, a method for determining an adaptive target threshold is provided, the method comprising:
[0007] Acquire a cavity measuring sonar signal, wherein the cavity measuring sonar signal is collected by the cavity measuring sonar;
[0008] Performing multi-layer wavelet decomposition on the cavity-measuring sonar signal to obtain low-frequency coefficients and high-frequency coefficients of each layer after decomposition;
[0009] According to the initial threshold, threshold processing is performed on the high frequency coefficients of each layer, and signal reconstruction is performed based on the low frequency coefficients of each layer and the high frequency coefficients of each layer after the threshold processing to obtain a reconstructed signal corresponding to the initial threshold;
[0010] According to the noise levels of the reconstructed signals respectively corresponding to the multiple initial thresholds, a target threshold is determined from the multiple initial thresholds, and the target threshold is used to perform wavelet filtering on the target signal collected by the cavity sonar.
[0011] In combination with any embodiment provided in the present disclosure, before performing multi-layer wavelet decomposition on the cavity measuring sonar signal, the method further includes:
[0012] Kalman filtering is performed on the cavity measuring sonar signal.
[0013] In combination with any embodiment provided in the present disclosure, the wavelet basis function used when performing multi-layer wavelet decomposition on the cavity measuring sonar signal is Dobesie 4 wavelet.
[0014] In combination with any implementation manner provided in the present disclosure, soft threshold processing is adopted when performing threshold processing on the high frequency coefficients of each layer.
[0015] In combination with any of the embodiments provided in the present disclosure, the noise level of the reconstructed signal is represented by a signal-to-noise ratio.
[0016] In combination with any embodiment provided in the present disclosure, the step of determining a target threshold from a plurality of initial thresholds according to the noise levels of the reconstructed signals respectively corresponding to the plurality of initial thresholds includes:
[0017] For each reconstructed signal corresponding to the initial threshold, estimating the signal power and the noise power of the reconstructed signal based on the power spectral density of the reconstructed signal;
[0018] Determining a signal-to-noise ratio of the reconstructed signal according to a ratio of a signal power to a noise power of the reconstructed signal;
[0019] Among the signal-to-noise ratios of the reconstructed signals corresponding to the multiple initial thresholds, a maximum signal-to-noise ratio is determined, and the initial threshold corresponding to the maximum signal-to-noise ratio is determined as the target threshold.
[0020] In combination with any embodiment provided in the present disclosure, after determining the target threshold from the multiple initial thresholds according to the noise levels of the reconstructed signal corresponding to the multiple initial thresholds, the method further includes:
[0021] Performing multi-layer wavelet decomposition on the target signal collected by the cavity sonar to determine the target low-frequency coefficient of each layer after decomposition and the to-be-determined high-frequency coefficient of each layer;
[0022] According to the target threshold, threshold processing is performed on the undetermined high-frequency coefficients of each layer to obtain the target high-frequency coefficients of each layer;
[0023] Signal reconstruction is performed based on the target low-frequency coefficients of each layer and the target high-frequency coefficients of each layer to obtain a reconstructed signal corresponding to the target signal.
[0024] In a second aspect, a device for determining an adaptive target threshold is provided, the device comprising:
[0025] A signal acquisition module, used to acquire a cavity measuring sonar signal, wherein the cavity measuring sonar signal is collected by the cavity measuring sonar;
[0026] A wavelet decomposition module is used to perform multi-layer wavelet decomposition on the cavity sonar signal to obtain low-frequency coefficients and high-frequency coefficients of each layer after decomposition;
[0027] A signal reconstruction module, used to perform threshold processing on the high-frequency coefficients of each layer according to an initial threshold, and perform signal reconstruction based on the low-frequency coefficients of each layer and the high-frequency coefficients of each layer after threshold processing, to obtain a reconstructed signal corresponding to the initial threshold;
[0028] The threshold determination module is used to determine a target threshold from the multiple initial thresholds according to the noise levels of the reconstructed signals respectively corresponding to the multiple initial thresholds, and the target threshold is used to perform wavelet filtering on the target signal collected by the cavity sonar.
[0029] According to a third aspect, an electronic device is provided, comprising a memory and a processor, wherein the memory is used to store computer instructions executable on the processor, and the processor is used to implement the adaptive target threshold determination method described in any embodiment of the present disclosure when executing the computer instructions.
[0030] In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored, and when the program is executed by a processor, the method for determining an adaptive target threshold value described in any embodiment of the present disclosure is implemented.
[0031] The adaptive target threshold determination method provided by the embodiment of the present disclosure performs multi-layer wavelet decomposition on the cavity measuring sonar signal to obtain low-frequency coefficients and high-frequency coefficients of each layer, applies different initial thresholds to perform threshold processing on the high-frequency coefficients, and then restores the time domain signal in combination with the low-frequency coefficients of each layer to obtain reconstructed signals corresponding to different initial thresholds, and finally adaptively uses the noise level of the reconstructed signal as an evaluation index to select the optimal target threshold from different initial thresholds. The target threshold can be used to optimize the wavelet filtering effect, effectively remove the periodic noise in the cavity measuring sonar signal, and improve the signal-to-noise ratio of the signal collected by the cavity measuring sonar. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] In order to more clearly illustrate the technical solutions in one or more embodiments of the present disclosure or related technologies, the drawings required for use in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in one or more embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0033] Figure 1is a flow chart of a method for determining an adaptive target threshold value according to at least one embodiment of the present disclosure;
[0034] Figure 2 is a flowchart of another method for determining an adaptive target threshold value according to at least one embodiment of the present disclosure;
[0035] Figure 3 is a time domain diagram of a cavity measuring sonar signal shown in at least one embodiment of the present disclosure;
[0036] Figure 4 is a time domain diagram of a cavity-measuring sonar signal after Kalman filtering shown in at least one embodiment of the present disclosure;
[0037] Figure 5 The diagram is a schematic diagram of an adaptive target threshold selection method according to at least one embodiment of the present disclosure.
[0038] Figure 6 is a time domain diagram of a target signal after noise reduction shown in at least one embodiment of the present disclosure;
[0039] Figure 7 is a schematic diagram of an original omnidirectional signal shown in at least one embodiment of the present disclosure;
[0040] Figure 8 is a schematic diagram of an omnidirectional signal after Kalman filtering processing, shown in at least one embodiment of the present disclosure;
[0041] Fig. 9 is a schematic diagram of an omnidirectional signal after adaptive target threshold processing shown in at least one embodiment of the present disclosure;
[0042] Fig.10 is a block diagram of an adaptive target threshold determination device shown in at least one embodiment of the present disclosure;
[0043] Fig.11 A hardware structure diagram of an electronic device shown in at least one embodiment of the present disclosure DETAILED DESCRIPTION
[0044] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with this specification. Instead, they are merely examples of devices and methods consistent with some aspects of this specification as detailed in the appended claims.
[0045] The terms used in this specification are for the purpose of describing specific embodiments only and are not intended to limit this specification. The singular forms "a", "the" and "the" used in this specification and the appended claims are also intended to include plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.
[0046] It should be understood that although the terms first, second, third, etc. may be used in this specification to describe various information, this information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of this specification, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".
[0047] The noise in the echo signal collected by the cavity sonar mainly comes from the vibration generated by the mechanical moving parts (such as the drive motor) inside the cavity sonar system. The vibration is coupled into the sonar system through the structure, thereby generating periodic noise and interfering with the received useful echo signal.
[0048] Since this noise is usually low-frequency noise and is concentrated at a specific frequency in the frequency domain, it may overlap with the spectrum of the useful echo signal. Commonly used filtering methods such as wavelet filtering are difficult to distinguish between signals and noise, which may result in the loss of signal components or residual noise.
[0049] Wavelet filtering is a process of decomposing signals, removing noise and reconstructing signals through wavelet transform. The main steps are as follows:
[0050] 1. Select the wavelet basis and decomposition level
[0051] Select the appropriate wavelet basis and decomposition level according to the characteristics of the signal. The number of decomposition levels determines the resolution of the signal at different scales. The larger the number of decomposition levels, the easier it is to separate the noise in the signal, but the reconstructed signal will also be more distorted.
[0052] 2. Wavelet decomposition
[0053] The wavelet transform is used to decompose the signal layer by layer into low-frequency coefficients and high-frequency coefficients of different levels. In the first layer of decomposition, the signal is decomposed into low-frequency coefficients and high-frequency coefficients of the first layer. In each subsequent layer of decomposition, the wavelet transform only further decomposes the low-frequency coefficients of the signal, and no longer decomposes the high-frequency coefficients. The result of each layer of decomposition is a low-frequency coefficient and a high-frequency coefficient. The low-frequency coefficient (also called the approximate coefficient) represents the smoothness or overall trend of the signal, while the high-frequency coefficient (also called the detail coefficient) represents the rapid changes or detail information of the signal.
[0054] 3. Threshold Processing
[0055] Threshold processing is performed on the high-frequency coefficients obtained from each layer of decomposition to remove the noise component. Threshold processing can be achieved through hard thresholding or soft thresholding: hard thresholding directly sets the coefficients below the threshold to 0; soft thresholding sets each coefficient to zero if its absolute value is less than the threshold; if its absolute value is greater than the threshold, the threshold is subtracted and the direction remains unchanged.
[0056] 4. Signal reconstruction
[0057] Wavelet reconstruction uses the low-frequency coefficients of each layer and the processed high-frequency coefficients to perform inverse wavelet transform, thereby gradually restoring the signal.
[0058] Starting from the highest decomposition layer, the inverse transform is performed layer by layer, and each layer uses the low-frequency coefficients of the current layer and the processed high-frequency coefficients to reconstruct the denoised signal.
[0059] Through the above-mentioned multi-layer decomposition and threshold processing, wavelet filtering can effectively remove noise in the signal while retaining the main features of the signal.
[0060] For the wavelet filtering method, it is difficult to filter out the periodic noise in the cavity sonar signal mainly because:
[0061] If the frequency band of the periodic low-frequency noise overlaps with the low-frequency component of the echo signal, the decomposed low-frequency coefficients may contain both noise and signal. In this case, a threshold needs to be set to filter out the noise component. Determining the appropriate threshold is key, but also difficult: if the threshold is set too low, the noise removal may be insufficient; if it is set too high, the signal details may be lost, causing the reconstructed signal to be distorted. In addition, for periodic noise, the frequency characteristics of the signal and noise need to be weighed when selecting the threshold, which increases the difficulty of denoising parameter adjustment.
[0062] In addition, periodic low-frequency noise is sometimes not completely stable and may be affected by the environment or the operating status of the device, resulting in intensity changes or phase fluctuations, which makes it difficult to stably remove the noise using a fixed threshold.
[0063] Based on this, the present disclosure provides an adaptive target threshold determination method, which analyzes the cavity-measuring sonar signal collected by the currently used cavity-measuring sonar, takes the noise level as an evaluation index, and adaptively determines the target threshold used when performing wavelet filtering on the signal collected by the cavity-measuring sonar. The target threshold can well distinguish between the echo signal and the noise in the signal collected by the cavity-measuring sonar, thereby removing the periodic noise in the signal through the target threshold and improving the signal-to-noise ratio of the cavity-measuring sonar signal.
[0064] The following is an explanation of the adaptive target threshold determination method provided by the present disclosure.
[0065] like Figure 1 As shown, Figure 1 is a flowchart of a method for determining an adaptive target threshold value according to at least one embodiment of the present disclosure, and the method may include the following steps:
[0066] In step 102, a cavity measuring sonar signal is obtained.
[0067] Wherein, the cavity measuring sonar signal is collected by the cavity measuring sonar.
[0068] In this step, the signal collected by the cavity sonar may be sampled to obtain the cavity sonar signal. For example, the echo collected by the cavity sonar may be continuously sampled with a sampling frequency of 100 kHz and a sampling time of 0.1 s.
[0069] In one embodiment, after the cavity-measuring sonar signal is acquired, the cavity-measuring sonar signal is subjected to Kalman filtering. The Kalman filtering preprocessing of the cavity-measuring sonar signal after sampling can significantly reduce the high-frequency noise in the signal, so that the target threshold determined in the subsequent processing can focus on removing the low-frequency periodic noise.
[0070] In other embodiments, other methods may be used to pre-process the cavity-measuring sonar signal to filter out high-frequency noise, for example, using a Savitzky-Golay filtering algorithm to process the cavity-measuring sonar signal.
[0071] In step 104, multi-layer wavelet decomposition is performed on the cavity measuring sonar signal to obtain low-frequency coefficients and high-frequency coefficients of each layer after decomposition.
[0072] In this step, appropriate wavelet basis functions and decomposition layers are selected to perform multi-layer wavelet decomposition on the cavity measurement sonar signal, thereby obtaining low-frequency coefficients and high-frequency coefficients of each layer.
[0073] For example, based on the wavelet basis function, the cavity sonar signal is low-pass and high-pass filtered at the same time to obtain the low-frequency coefficients and high-frequency coefficients of the first layer after decomposition, and then the low-frequency coefficients of the first layer are repeatedly filtered to obtain the low-frequency coefficients and high-frequency coefficients of the second layer after decomposition, and so on, to obtain the low-frequency coefficients and high-frequency coefficients of each layer after decomposition. This multi-layer decomposition gradually decomposes the signal into low-frequency coefficients and high-frequency coefficients at different scales, forming a pyramid structure.
[0074] This embodiment does not limit the wavelet basis function used for wavelet decomposition, for example, it may be a Haar wavelet, db (Daubechies) wavelet, Coiflets wavelet or other wavelet basis functions.
[0075] Different wavelet bases have different sensitivities to different characteristics of signals. The choice of wavelet base can directly affect the decomposition and reconstruction effects of signals.
[0076] In one embodiment, the wavelet basis function used in the multi-layer wavelet decomposition of the cavity sonar signal is Dobesi 4 wavelet (DB4 wavelet for short). The DB wavelet suitable for signal denoising and feature extraction can play a good role in signal processing application scenarios. In this embodiment, the selection of DB4 wavelet as the wavelet basis function can better achieve the separation of noise and signal in multi-layer decomposition.
[0077] This embodiment does not limit the number of decomposition layers of wavelet decomposition, for example, the number of decomposition layers can be 5 to 8. The number of decomposition layers determines the representation and analysis depth of the signal at different scales. The more decomposition layers there are, the easier it is to separate the noise in the signal, but the reconstructed signal will also be more distorted.
[0078] In one embodiment, 7 layers are selected as the number of decomposition layers, which can remove noise well without losing signal details.
[0079] In step 106, the high frequency coefficients of each layer are threshold processed according to the initial threshold, and signal reconstruction is performed based on the low frequency coefficients of each layer and the high frequency coefficients of each layer after threshold processing to obtain a reconstructed signal corresponding to the initial threshold.
[0080] In this step, different initial thresholds are used to perform threshold processing on the high frequency coefficients and perform signal reconstruction.
[0081] This embodiment does not limit the specific selection of the initial threshold, which may be pre-set or randomly generated.
[0082] Considering that the purpose of this scheme is to select the initial threshold with the best denoising effect from multiple initial thresholds, the value range of the initial threshold can be set by technical personnel in this field based on experience, the step size can be set to take multiple values within the value range as the initial threshold, or multiple initial thresholds can be randomly initialized and generated within the value range.
[0083] Exemplarily, this step may be to randomly initialize and generate an initial threshold value A, use the initial threshold value A to perform threshold processing on the high-frequency coefficients of each layer through a threshold function, perform wavelet reconstruction using the low-frequency coefficients of each layer decomposed by wavelet and the high-frequency coefficients after threshold processing, and convert the processed wavelet coefficients into a new time domain signal, i.e., the reconstructed signal corresponding to the initial threshold value A.
[0084] An initial threshold B is generated by random initialization, and the above steps are also performed using the initial threshold B to obtain a reconstructed signal corresponding to the initial threshold B.
[0085] The above processing of the initial threshold A and the initial threshold B can be performed in parallel or sequentially. Similarly, the above steps are performed for other initial thresholds to obtain reconstructed signals corresponding to different initial thresholds.
[0086] It should be noted that, in this embodiment, the initial threshold is a global threshold, that is, for the high-frequency coefficients obtained by multi-layer decomposition, the same initial threshold is used for threshold processing to speed up signal processing, and finally a target threshold is determined for the cavity sonar signal.
[0087] In other embodiments, the initial threshold may not be a global threshold, that is, for the high-frequency coefficients obtained by multi-layer decomposition, different initial thresholds are used for threshold processing for the high-frequency coefficients of each layer, and finally multiple target thresholds corresponding to the number of decomposition layers are determined for the cavity sonar signal. For example, when the number of decomposition layers is 7, the determined target threshold is 7. This method is more complicated to calculate and has a better denoising effect.
[0088] In one embodiment, soft threshold processing is used when threshold processing is performed on the high frequency coefficients of each layer.
[0089] For example, a soft threshold function is selected, the threshold in the soft threshold function is randomly initialized, and the high-frequency coefficients obtained by wavelet transform are threshold processed to weaken the high-frequency noise. Among them, the soft threshold function sets the wavelet coefficients whose absolute value is less than the threshold to zero, and conversely reduces the wavelet coefficients whose absolute value is greater than the threshold. Then, the high-frequency coefficients after threshold processing are combined with the low-frequency coefficients initially obtained by wavelet transform, and the time domain signal is reconstructed layer by layer.
[0090] Using soft threshold processing can effectively remove smaller noise components while retaining the main signal, making the signal smoother. Compared with the hard threshold (all coefficients less than the threshold are directly set to zero), the soft threshold has a better smoothing effect.
[0091] In step 108, a target threshold is determined from the multiple initial thresholds according to the noise levels of the reconstructed signal corresponding to the multiple initial thresholds respectively.
[0092] Among them, the noise level of the reconstructed signal corresponding to the initial threshold reflects the denoising effect of the initial threshold. This step can be to determine the initial threshold with the best denoising effect (that is, the lowest noise level) as the target threshold, or to determine the initial threshold with the denoising effect reaching a preset standard as the target threshold.
[0093] This embodiment does not limit the method of characterizing the noise level of the reconstructed signal, for example, it can be a signal-to-noise ratio, noise power, noise coefficient, etc. When the noise level is represented by the signal-to-noise ratio, the larger the signal-to-noise ratio of the reconstructed signal, the lower the noise level, and the better the denoising effect; when the noise level is represented by the noise power, the lower the noise power of the reconstructed signal, the lower the noise level, and the better the denoising effect; when the noise level is represented by the noise coefficient, the smaller the noise coefficient, the lower the noise level, and the better the denoising effect. In one embodiment, the noise level of the reconstructed signal is represented by the signal-to-noise ratio.
[0094] Exemplarily, the optimal target threshold may be adaptively selected based on the maximum signal-to-noise ratio of the reconstructed signal estimated by PSD (Power Spectral Density).
[0095] This step can be specifically processed as follows:
[0096] For each reconstructed signal corresponding to the initial threshold, estimating the signal power and the noise power of the reconstructed signal based on the power spectral density of the reconstructed signal;
[0097] The power spectral density of the reconstructed signal describes the power distribution of the reconstructed signal in the frequency domain. The reconstructed signal can be Fourier transformed to obtain the spectrum, and then the power spectral density is calculated. The power spectral density within the signal bandwidth is integrated to obtain the signal power, and the power spectral density within the noise bandwidth is integrated to obtain the noise power (or the power spectral density within the entire bandwidth can be integrated to obtain the total power, and the total power minus the signal power is used to obtain the noise power);
[0098] Determining a signal-to-noise ratio of the reconstructed signal according to a ratio of a signal power to a noise power of the reconstructed signal;
[0099] The signal-to-noise ratio is calculated using the ratio of signal power to noise power;
[0100] Among the signal-to-noise ratios of the reconstructed signals corresponding to the multiple initial thresholds, the maximum signal-to-noise ratio is determined, and the initial threshold corresponding to the maximum signal-to-noise ratio is determined as the target threshold. For the reconstructed signal corresponding to each initial threshold, the signal power of the useful signal and the noise power of the noise in the reconstructed signal are calculated based on the PSD estimation, so as to determine the signal-to-noise ratio of the reconstructed signal, the signal-to-noise ratio with the largest ratio among the multiple signal-to-noise ratios is determined as the maximum signal-to-noise ratio, and the initial threshold corresponding to the maximum signal-to-noise ratio is determined as the target threshold.
[0101] The PSD estimation method can be used to analyze the spectral characteristics of the reconstructed signal, identify the distribution of the useful signal and noise in the reconstructed signal in the frequency domain, and thus estimate the signal-to-noise ratio of the reconstructed signal. The optimal target threshold can be adaptively selected with the maximum signal-to-noise ratio as the goal.
[0102] In this embodiment, the target threshold is used to perform wavelet filtering on the target signal collected by the cavity sonar.
[0103] Since the target threshold has achieved good results in denoising the cavity-measuring sonar signal acquired in step 102, it can be used for wavelet filtering of the signal acquired by the cavity-measuring sonar.
[0104] In one embodiment, after this step, the method further includes:
[0105] Performing multi-layer wavelet decomposition on the target signal collected by the cavity sonar to determine the target low-frequency coefficient of each layer after decomposition and the to-be-determined high-frequency coefficient of each layer;
[0106] According to the target threshold, threshold processing is performed on the undetermined high-frequency coefficients of each layer to obtain the target high-frequency coefficients of each layer;
[0107] Signal reconstruction is performed based on the target low-frequency coefficients of each layer and the target high-frequency coefficients of each layer to obtain a reconstructed signal corresponding to the target signal.
[0108] For example, a large number of target signals may be collected during cave measurement of a cave by a cavity sonar, and a segment of the signals may be selected to perform steps 102 to 108 to obtain a target threshold, and the target threshold may be used to replace the initial threshold. Steps 104 to 106 may be performed on the previously collected target signal to obtain a reconstructed signal corresponding to the target signal.
[0109] Since the target threshold is adaptively determined based on a section of the target signal, it can well separate the useful signal and noise in the cavity sonar measurement environment, thereby effectively removing the periodic noise in the target signal and optimizing the wavelet filtering effect.
[0110] The adaptive target threshold determination method provided by the embodiment of the present disclosure performs multi-layer wavelet decomposition on the cavity measuring sonar signal to obtain low-frequency coefficients and high-frequency coefficients of each layer, applies different initial thresholds to perform threshold processing on the high-frequency coefficients, and then restores the time domain signal in combination with the low-frequency coefficients of each layer to obtain reconstructed signals corresponding to different initial thresholds, and finally adaptively uses the noise level of the reconstructed signal as an evaluation index to select the optimal target threshold from different initial thresholds. The target threshold can be used to optimize the wavelet filtering effect, effectively remove the periodic noise in the cavity measuring sonar signal, and improve the signal-to-noise ratio of the signal collected by the cavity measuring sonar.
[0111] The following is a detailed description of the adaptive target threshold determination method provided by the present disclosure in conjunction with a specific embodiment. Figure 2 As shown, this embodiment may include the following steps:
[0112] In step 202, a cavity measuring sonar signal is obtained.
[0113] In this embodiment, the signal collected by the actual measurement cavity sonar of the salt cave can be continuously sampled to obtain the cavity sonar signal. The sampling frequency is 100kHz, the sampling time is 0.1s, and the sampling signal is shown in FIG. Figure 3 .
[0114] During the sampling process, some high-frequency noise can be filtered out, the signal quality can be enhanced, and useful signals can be extracted for more complex signal processing.
[0115] In step 204, Kalman filtering is performed on the cavity detection sonar signal.
[0116] The Kalman filter is performed on the cavity-measuring sonar signal, and the formula thereof can be as follows:
[0117] y(k)=Ay(k-1)+w(k-1) (1)
[0118] z(k)=Hy(k)+v(k) (2)
[0119] Among them, formula (1) is the state equation in Kalman filtering, and formula (2) is the observation equation, which are used to represent the dynamic model and measurement model of the system respectively; y(k) is the state vector of the system at time k; A is the state transfer matrix; z(k) is the observation vector of the system at time k; H is the observation matrix; w(k-1) and v(k) are process noise with a mean of 0 and observation noise, respectively, and their corresponding noise covariance matrices are Q(k-1) and R(k), respectively.
[0120] Kalman filtering is widely used in dynamic systems. The cavity sonar signal is filtered through Kalman filtering to remove some significant noise and further smooth the sampled signal. The noise reduction and smoothing effects are shown in Figure 4 , it can be seen that Kalman filtering removes Figure 3 Random oscillating noise of medium to high frequency.
[0121] In step 206, multi-layer wavelet decomposition is performed on the cavity-measuring sonar signal to obtain low-frequency coefficients and high-frequency coefficients of each layer after decomposition.
[0122] In this step, wavelet basis functions and decomposition levels are selected, and wavelet transform is performed on the Kalman filtered signal to obtain low-frequency coefficients and high-frequency coefficients.
[0123] Specifically, the wavelet base selected in this embodiment is db4, the number of decomposition layers is 7, and the wavelet transform formula is as follows:
[0124]
[0125] Among them, a j (n) is the low frequency coefficient, d j (n) is the high frequency coefficient. h(k) and g(k) are the low-pass and high-pass filter coefficients respectively, and j represents the jth layer.
[0126] In step 208, threshold processing is performed on the high frequency coefficients of each layer according to the initial threshold.
[0127] In this embodiment, the high frequency coefficients of each layer are randomly initialized and threshold processing is performed, and the threshold function selects a soft threshold, and its formula is as follows:
[0128]
[0129] Among them, d thr is the high-frequency coefficient after threshold processing, thr is the initial threshold, and d is the high-frequency coefficient of each layer in wavelet decomposition.
[0130] In step 210, signal reconstruction is performed based on the low-frequency coefficients of each layer and the high-frequency coefficients of each layer after threshold processing to obtain a reconstructed signal corresponding to the initial threshold.
[0131] The low-frequency coefficients of each layer and the processed high-frequency coefficients are used for wavelet reconstruction. Specifically, the low-frequency coefficients of each layer of wavelet decomposition and the high-frequency coefficients after threshold processing are used for wavelet reconstruction, and the processed wavelet coefficients are converted into a new time domain signal, namely, the reconstructed signal.
[0132] In step 212, a target threshold is determined from the multiple initial thresholds according to the noise levels of the reconstructed signal corresponding to the multiple initial thresholds respectively.
[0133] In this step, the maximum signal-to-noise ratio of the reconstructed signal can be estimated based on PSD to adaptively select the optimal target threshold from multiple initial thresholds. The formula for estimating SNR (Signal-to-noise Ratio) based on PSD is as follows:
[0134]
[0135] P noise =P total -P signal (8)
[0136]
[0137] The reconstructed signal of length N is divided into L segments, each of which is of length M. After windowing each segment, a fast Fourier transform is performed to calculate the corresponding power spectrum. P signal is the signal power obtained by integrating PSD within the signal bandwidth [f1, f2], [f1, f2] is the effective frequency range of the signal, P total is the total power obtained by integrating PSD over the entire frequency range, P noise is the total noise power.
[0138] Based on PSD estimation, the optimal threshold is adaptively selected from multiple initial thresholds. The PSD estimation method can analyze the spectral characteristics of the reconstructed signal, identify the distribution of the reconstructed signal and noise in the frequency domain, and thus estimate the signal-to-noise ratio of the reconstructed signal. The optimal threshold is adaptively selected with the signal-to-noise ratio as the target. The process of adaptively selecting the optimal target threshold is shown in Figure 5 .
[0139] exist Figure 5 In the figure, the horizontal axis represents the initial threshold, the vertical axis represents the signal-to-noise ratio, and the unit of the signal-to-noise ratio is decibel. It can be seen from the figure that for different initial thresholds in the range of [0,1], the signal-to-noise ratio of the reconstructed signal is different. When the initial threshold is 0.64, the corresponding signal-to-noise ratio of 0.453 is the highest value of the vertical axis, which is determined as the maximum signal-to-noise ratio, and 0.64 is determined as the target threshold.
[0140] The optimal target threshold is selected to replace the randomly initialized threshold, and the target signal after Kalman filtering is processed by wavelet filtering. The noise reduction effect is shown in Figure 6 .
[0141] Compared to Figure 4 , Figure 6 The continuous low-frequency noise and residual random noise in the signal are further removed, making the signal smoother, and the main signal is not significantly affected. It can be seen that wavelet filtering can effectively remove low-frequency periodic noise after selecting the optimal target threshold.
[0142] For example, the omnidirectional cavity sonar signal is selected for filtering effect comparison, and the original omnidirectional signal is shown in Figure 7 . Figure 7In the figure, the horizontal axis represents the distance between the salt cavern and the cavity sonar, with a distance range of [0,22], and the vertical axis represents the azimuth angle detected by the cavity sonar, which collects signals by scanning in all directions from 0° to 360°.
[0143] The omnidirectional signal after Kalman filtering is shown in Figure 8 , compared to Figure 7 , Figure 8 The signal in the image is smoother and the high-frequency background noise is removed.
[0144] The omnidirectional signal after adaptive target threshold processing is shown in Fig. 9 , compared to Figure 8 , Fig. 9 The smoothness of the signal in the image is improved, the high-frequency background noise is lower, the continuous small fluctuations of the low frequency are suppressed, and the amplitude of the main signal is retained.
[0145] Comparison of the above three figures shows that the denoising method based on adaptive target threshold is effective for cavity sonar signals.
[0146] The method in this embodiment takes the signal-to-noise ratio of PSD estimation as the target, and performs adaptive optimal threshold selection on the cavity measuring sonar signal of the wavelet transform with randomly initialized threshold, thereby effectively removing periodic noise and improving the signal-to-noise ratio of the cavity measuring sonar signal, providing better detection information for subsequent acquisition of salt cavity morphology, volume and other information, and providing technical support for the availability of salt cavities, optimization of cavity making technology and periodic operation management of cavities.
[0147] like Fig.10 As shown, Fig.10 is a block diagram of a device for determining an adaptive target threshold value according to at least one embodiment of the present disclosure, the device comprising:
[0148] A signal acquisition module 11 is used to acquire a cavity measuring sonar signal, wherein the cavity measuring sonar signal is collected by the cavity measuring sonar;
[0149] The wavelet decomposition module 12 is used to perform multi-layer wavelet decomposition on the cavity sonar signal to obtain low-frequency coefficients and high-frequency coefficients of each layer after decomposition;
[0150] A signal reconstruction module 13 is used to perform threshold processing on the high frequency coefficients of each layer according to an initial threshold, and to perform signal reconstruction based on the low frequency coefficients of each layer and the high frequency coefficients of each layer after the threshold processing, so as to obtain a reconstructed signal corresponding to the initial threshold;
[0151] The threshold determination module 14 is used to determine a target threshold from the multiple initial thresholds according to the noise levels of the reconstructed signals respectively corresponding to the multiple initial thresholds, and the target threshold is used to perform wavelet filtering on the target signal collected by the cavity sonar.
[0152] In some embodiments, the signal acquisition module 11 is further configured to perform Kalman filtering on the cavity measuring sonar signal before performing multi-layer wavelet decomposition on the cavity measuring sonar signal.
[0153] In some embodiments, the wavelet basis function used when performing multi-layer wavelet decomposition on the cavity measuring sonar signal is Dobesie 4 wavelet.
[0154] In some embodiments, soft threshold processing is used when threshold processing is performed on the high frequency coefficients of each layer.
[0155] In some embodiments, the noise level of the reconstructed signal is represented by a signal-to-noise ratio.
[0156] In some embodiments, the threshold determination module 14 is specifically configured to:
[0157] For each reconstructed signal corresponding to the initial threshold, estimating the signal power and the noise power of the reconstructed signal based on the power spectral density of the reconstructed signal;
[0158] Determining a signal-to-noise ratio of the reconstructed signal according to a ratio of a signal power to a noise power of the reconstructed signal;
[0159] Among the signal-to-noise ratios of the reconstructed signals corresponding to the multiple initial thresholds, a maximum signal-to-noise ratio is determined, and the initial threshold corresponding to the maximum signal-to-noise ratio is determined as the target threshold.
[0160] In some embodiments, the threshold determination module 14, after determining the target threshold from the plurality of initial thresholds, is further configured to:
[0161] Performing multi-layer wavelet decomposition on the target signal collected by the cavity sonar to determine the target low-frequency coefficient of each layer after decomposition and the to-be-determined high-frequency coefficient of each layer;
[0162] According to the target threshold, threshold processing is performed on the undetermined high-frequency coefficients of each layer to obtain the target high-frequency coefficients of each layer;
[0163] Signal reconstruction is performed based on the target low-frequency coefficients of each layer and the target high-frequency coefficients of each layer to obtain a reconstructed signal corresponding to the target signal.
[0164] The implementation process of the functions and effects of each module in the above-mentioned device is specifically described in the implementation process of the corresponding steps in the above-mentioned method, which will not be repeated here.
[0165] The present disclosure also provides an electronic device, such as Fig.11As shown, the electronic device includes a memory 111 and a processor 112, wherein the memory 111 is used to store computer instructions that can be executed on the processor, and the processor 112 is used to implement the adaptive target threshold determination method described in any embodiment of the present disclosure when executing the computer instructions.
[0166] The embodiments of the present disclosure also provide a computer program product, which includes a computer program / instruction. When the computer program / instruction is executed by a processor, the method for determining an adaptive target threshold value described in any embodiment of the present disclosure is implemented.
[0167] The embodiments of the present disclosure further provide a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the method for determining an adaptive target threshold value described in any embodiment of the present disclosure is implemented.
[0168] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can refer to the partial description of the method embodiment. The device embodiment described above is only schematic, wherein the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, that is, they may be located in one place, or they may be distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this specification. A person of ordinary skill in the art can understand and implement it without paying creative labor.
[0169] The above is a description of a specific embodiment of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0170] Those skilled in the art will readily appreciate other embodiments of the specification after considering the specification and practicing the invention claimed herein. The specification is intended to cover any variations, uses or adaptations of the specification that follow the general principles of the specification and include common knowledge or customary techniques in the art that are not claimed in the specification. The specification and examples are to be considered exemplary only, and the true scope and spirit of the specification are indicated by the following claims.
[0171] It should be understood that the present description is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present description is limited only by the appended claims.
[0172] The above description is only a preferred embodiment of this specification and is not intended to limit this specification. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of this specification should be included in the scope of protection of this specification.
Claims
1. A method for determining an adaptive target threshold, characterized in that: The method comprises: Acquiring a cavity measuring sonar signal, wherein the cavity measuring sonar signal is collected by the cavity measuring sonar; Performing multi-layer wavelet decomposition on the cavity-measuring sonar signal to obtain low-frequency coefficients and high-frequency coefficients of each layer after decomposition; According to the initial threshold, threshold processing is performed on the high frequency coefficients of each layer, and signal reconstruction is performed based on the low frequency coefficients of each layer and the high frequency coefficients of each layer after the threshold processing to obtain a reconstructed signal corresponding to the initial threshold; According to the noise levels of the reconstructed signals respectively corresponding to the multiple initial thresholds, a target threshold is determined from the multiple initial thresholds, and the target threshold is used to perform wavelet filtering on the target signal collected by the cavity sonar.
2. The method according to claim 1, characterized in that Before performing multi-layer wavelet decomposition on the cavity measuring sonar signal, the method further includes: Kalman filtering is performed on the cavity measuring sonar signal.
3. The method according to claim 1, characterized in that The wavelet basis function used when performing multi-layer wavelet decomposition on the cavity measuring sonar signal is the Dobesie 4 wavelet.
4. The method according to claim 1, characterized in that: The soft threshold processing is adopted when performing threshold processing on the high frequency coefficients of each layer.
5. The method according to claim 1, characterized in that The noise level of the reconstructed signal is represented by the signal-to-noise ratio.
6. The method according to claim 5, characterized in that The step of determining a target threshold from the plurality of initial thresholds according to the noise levels of the reconstructed signals respectively corresponding to the plurality of initial thresholds comprises: For each reconstructed signal corresponding to the initial threshold, estimating the signal power and the noise power of the reconstructed signal based on the power spectral density of the reconstructed signal; Determining a signal-to-noise ratio of the reconstructed signal according to a ratio of a signal power to a noise power of the reconstructed signal; Among the signal-to-noise ratios of the reconstructed signals corresponding to the multiple initial thresholds, a maximum signal-to-noise ratio is determined, and the initial threshold corresponding to the maximum signal-to-noise ratio is determined as the target threshold.
7. The method according to claim 1, characterized in that After determining the target threshold from the multiple initial thresholds according to the noise levels of the reconstructed signals respectively corresponding to the multiple initial thresholds, the method further includes: Performing multi-layer wavelet decomposition on the target signal collected by the cavity sonar to determine the target low-frequency coefficient of each layer and the to-be-determined high-frequency coefficient of each layer after decomposition; According to the target threshold, threshold processing is performed on the undetermined high-frequency coefficients of each layer to obtain the target high-frequency coefficients of each layer; Signal reconstruction is performed based on the target low-frequency coefficients of each layer and the target high-frequency coefficients of each layer to obtain a reconstructed signal corresponding to the target signal.
8. An adaptive target threshold determination device, characterized in that: The device comprises: A signal acquisition module, used to acquire a cavity measuring sonar signal, wherein the cavity measuring sonar signal is collected by the cavity measuring sonar; A wavelet decomposition module is used to perform multi-layer wavelet decomposition on the cavity sonar signal to obtain low-frequency coefficients and high-frequency coefficients of each layer after decomposition; A signal reconstruction module, used to perform threshold processing on the high-frequency coefficients of each layer according to an initial threshold, and perform signal reconstruction based on the low-frequency coefficients of each layer and the high-frequency coefficients of each layer after threshold processing, to obtain a reconstructed signal corresponding to the initial threshold; The threshold determination module is used to determine a target threshold from the multiple initial thresholds according to the noise levels of the reconstructed signals respectively corresponding to the multiple initial thresholds, and the target threshold is used to perform wavelet filtering on the target signal collected by the cavity sonar.
9. An electronic device, characterized in that: The device includes a memory and a processor, wherein the memory is used to store computer instructions that can be executed on the processor, and the processor is used to implement the adaptive target threshold determination method described in any one of claims 1 to 7 when executing the computer instructions.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the adaptive target threshold determination method according to any one of claims 1 to 7 is implemented.
Citation Information
Cited By
Tiny precision bearing fatigue damage detection system and detection method
CN121453915A