Distributed sound wave sensing data noise reduction method based on high-quality sparse prior

CN118884533AActive Publication Date: 2024-11-01UNIV OF SCI & TECH OF CHINA
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202411137687.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-19
Publication Date
2024-11-01
Estimated Expiration
2044-08-19

Smart Images

  • Figure CN118884533A_ABST
    Figure CN118884533A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of distributed optical fiber sound wave sensing, and discloses a high-quality sparse prior-based distributed sound wave sensing data noise reduction method, which comprises the following steps of: selecting a plurality of space point locations on an optical fiber laying path, and arranging a seismic detector on each space point location; according to the method, a small amount of traditional geophone high-quality signals are added in the deep learning algorithm modeling process to serve as input, and the noise reduction performance of the model can be further improved on the basis of an existing seismic data intelligent noise reduction algorithm. The method can effectively improve the signal quality of the seismic data collected by the DAS, and can play a role in a plurality of application scenes needing DAS monitoring, such as building structure health assessment, geological disaster early warning, underground structure exploration and the like.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of distributed optical fiber acoustic wave sensing, and in particular to a distributed acoustic wave sensing data denoising method based on high-quality sparse priors. Background Art

[0002] Distributed acoustic sensing (DAS) is a new type of sensing technology. Since the phase of scattered light at a specific part of the optical fiber will change when it is subjected to micro-deformation, the demodulator can be used to collect sensor data such as temperature and strain at different parts of the optical fiber (Hartog, 2017). The collected strain data records the characteristics of the seismic wave field in the area where the optical fiber is located, and since DAS can achieve more intensive spatial sampling of seismic data at a lower cost (Madsen et al., 2013), it is widely used in seismic data acquisition (Fernández-Ruiz et al., 2018). However, due to the complex coupling between the optical fiber and the ground, the unstable state of the light source processed by the demodulator, and the influence of human activities in the deployment environment, the seismic data collected by the optical fiber has more serious noise pollution than traditional detectors (Feng and Li, 2019).

[0003] Since DAS data has high spatial resolution but is susceptible to noise interference, noise suppression for seismic data has become an important research topic. The noise generated by DAS is diverse, resulting in local structural correlation, which brings certain challenges to noise suppression (Feng and Li, 2021). In recent years, with the rise of artificial intelligence technology, noise suppression algorithms based on deep neural networks have shown better denoising effects than traditional seismic data noise suppression methods such as traditional FX transform (Harris and White, 1997) and empirical mode decomposition (Liu et al., 2013) (Zhao et al., 2018). Denoising algorithms based on deep learning are mainly divided into two categories: supervised learning and unsupervised learning. The main difference between these two methods is whether noise-free seismic data is provided as labels for neural network learning. However, since noise-free seismic data is unknown, supervised learning methods that require labels usually use near-real seismic data simulated based on physical laws as labels (Zhao et al., 2020; Ma et al., 2023), which makes such methods often unstable when applied to actual seismic data denoising (Zhao et al., 2023).

[0004] The noise suppression algorithm based on self-supervised learning is mainly based on the mathematical modeling of the characteristics that the effective signal of the current pixel can be predicted by the signal of the adjacent pixel, while the noise of the current pixel cannot be predicted by the noise of the adjacent pixel. It is widely used in the field of image noise suppression (Lehtinen et al., 2018; Batson and Royer, 2019). By replacing the concepts of the current pixel and the adjacent pixel in the image field with the waveform sampled at the current spatial position and the waveform sampled at the adjacent spatial position in the seismic data acquisition process, the noise suppression algorithm based on self-supervised learning can be applied to the seismic data collected by DAS, and obtain a better noise suppression effect (van den Ende et al., 2021).

[0005] In order to further improve the denoising performance of the noise suppression algorithm based on self-supervised learning, the present invention provides a distributed acoustic sensor data denoising method based on high-quality sparse prior. By taking advantage of the high signal-to-noise ratio quality of traditional detectors, the waveforms recorded at certain spatial points of the DAS are sparsely constrained through the signals recorded by traditional detectors, thereby further improving the noise suppression effect. Summary of the invention

[0006] In order to solve the above technical problems, the present invention provides a distributed acoustic sensor data denoising method based on high-quality sparse prior. By adding a small amount of traditional seismic detector high-quality signals as input in the process of deep learning algorithm modeling, the denoising performance of the model can be further improved on the basis of the existing seismic data intelligent denoising algorithm.

[0007] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0008] A distributed acoustic sensor data denoising method based on high-quality sparse priors guides a deep neural network to denoise seismic data based on distributed optical fiber acoustic sensor signals through the waveform received by the seismic detector; specifically, it includes:

[0009] Select multiple spatial points on the optical fiber laying path, and deploy seismic detectors at each spatial point;

[0010] Based on the J-invariance denoising theory, we define a function g with J-invariance. θ :

[0011] g θ (·)=∑ k∈K′ ∏ k (f θ (Π kc (·)));

[0012] Where K′ represents a set of multiple spatial points selected on the optical fiber laying path, k∈K′ represents any spatial point in K′, and Π k represents the operation of setting all dimensions except the dimension represented by k to zero, f θ represents a deep neural network; represents the spatial points near the spatial point k, where N is the total number of spatial points selected on the optical fiber laying path, and s is the interval between the spatial points;

[0013] The parameters θ of the deep neural network are optimized in the following way to achieve denoising of seismic data:

[0014] argmin θ ‖g θ (x)-x k ‖ 2 ;

[0015] where x k Represents the waveform received by the seismic detector at spatial point k.

[0016] Furthermore, the high-quality waveforms received by the seismic detectors are introduced into the deep neural network as prior knowledge to guide the deep neural network to denoise the seismic data based on distributed fiber optic acoustic sensor signals. The deep neural network adopts a U-Net network.

[0017] Compared with the prior art, the beneficial technical effects of the present invention are:

[0018] The distributed acoustic sensor data denoising method based on high-quality sparse prior proposed in the present invention combines the high signal-to-noise ratio of data collected by traditional seismic detectors and the high spatial sampling rate of data collected by DAS, and obtains low-noise seismic data with high spatial sampling rate through an intelligent denoising algorithm based on deep learning. Figure 4 It is demonstrated that the present invention can further suppress coherent noise compared with the existing self-supervised learning scheme. The method proposed in the present invention can effectively improve the signal quality of seismic data collected by DAS, and can play a role in multiple application scenarios that require DAS monitoring, such as building structure health assessment, geological disaster warning, and underground structure exploration. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 Schematic diagram of the implementation of the noise reduction method of the present invention. Figure 1 The black dots in the figure represent the equivalent spatial points in the DAS data acquisition process, and the inverted triangles represent the traditional node-type seismic detectors, which overlap with some points acquired by optical fiber in spatial position. The black waveform represents the low signal-to-noise ratio dense sampling signal acquired by DAS, and the red waveform represents the signal acquired by the traditional detector.

[0020] Figure 2 The figure is a comparison diagram of the principles of the method of the present invention and the existing algorithms. The upper part of the figure is a schematic diagram of the existing unsupervised distributed fiber optic acoustic sensor data noise suppression algorithm, which uses the characteristics that the waveform to be denoised can be restored by the waveform sampled in the nearby space, while the noise cannot be predicted by the surrounding space waveform. The lower part is the noise reduction method proposed by the present invention, which converts the target predicted by the deep neural network from the noisy distributed fiber optic acoustic sensor data into high signal-to-noise ratio data received by the traditional seismic detector, and the noise with spatial correlation can be further suppressed.

[0021] Figure 3 Schematic diagram of a distributed fiber optic acoustic wave sensor data noise suppression network model based on a U-Net structure in an embodiment of the present invention.

[0022] Figure 4 The three sub-figures are, from left to right, noisy data, the denoising result of the existing unsupervised learning denoising algorithm, and the denoising result of the denoising method of the present invention.

[0023] Figure 5 The three sub-pictures from left to right are respectively the synthesized seismic waveform, the noise extracted from the real acquired data, and the noisy data obtained by adding the above two.

[0024] Figure 6 Schematic diagram of bandpass filtering; a is the mean of all seismic data, and b is the mean of the noise spectrum.

[0025] Figure 7 This is a comparison chart of the noise suppression effect on different networks. From top to bottom, each row represents the effect comparison when the signal-to-noise ratio is -10dB, 0dB and 10dB, and from left to right, each column represents the noisy data, the denoising result of the existing unsupervised learning denoising algorithm, and the denoising result of the algorithm proposed in this invention.

[0026] Figure 8 It is a schematic diagram for comparing parameter statistics, i.e., the comparison of mean square error (MSE), signal-to-noise ratio (SNR) and correlation coefficient (PCC) for data denoising under different signal-to-noise ratios. The black line represents the data before denoising, and the green line and red line represent the denoising effect of the existing unsupervised learning denoising algorithm and the denoising effect of the algorithm proposed in the present invention, respectively. DETAILED DESCRIPTION

[0027] A preferred embodiment of the present invention is described in detail below with reference to the accompanying drawings.

[0028] The distributed acoustic sensor data denoising method based on high-quality sparse prior proposed in the present invention can constrain the denoising results output by the deep neural network by introducing traditional seismic detector signals with a higher signal-to-noise ratio into the self-supervised deep neural network training process, thereby further improving the noise suppression performance on the existing self-supervised learning method.

[0029] The core idea of ​​the present invention is to improve the denoising effect of distributed fiber optic acoustic wave sensing data by monitoring waveforms of a small number of traditional node-type seismic detectors with a high signal-to-noise ratio. Figure 1 This is a schematic diagram of the method, which requires the deployment of a small number of sparsely distributed traditional node-type seismic detectors at the location of the optical fiber when laying out the instrument. The waveforms collected by these traditional detectors are of higher quality than the DAS waveforms, but each detector can only receive data from a single spatial point, and it is very costly to deploy traditional detectors at all spatial points collected by the optical fiber. Therefore, traditional detectors only collect data at some points and have spatial sparsity. The present invention adds these high-quality sparse data as prior information to the training process of the denoising neural network, thereby improving the noise suppression effect of the neural network. Below we will introduce the technical content of the present invention in detail from three aspects: method principle, model training and denoising evaluation indicators.

[0030] 1. Principle of the method

[0031] The theoretical basis of this invention is the J-invariance denoising theory proposed by Batson and Royer (2019), which is defined as:

[0032] definition are noisy data samples and non-noisy data samples respectively, assuming is a set of dimensions {1,…,m}, and dimension J is A subset of If g(x) is represented in dimension J as g(x) J The value of does not depend on x J If true, then define the function g: It has J invariance.

[0033] Assume J c is J relative to The complement of , for a function g with J invariance, if the noise of y in dimension J and the noise of dimension J c The noises in are independent of each other, then:

[0034]

[0035] The above formula shows that when a function g with J invariance is used to fit the representation x of the noisy sample in dimension J, J When , it is equivalent to removing x JThe remaining dimensions cannot be sampled The predicted noise.

[0036] Based on this, Van den Ende et al. (2021) applied this theory to the noise suppression of seismic data based on distributed fiber optic acoustic sensor signals and defined a function g with J invariance. θ :

[0037]

[0038] Where K is the set of all spatial points collected by DAS, k∈K represents any spatial point collected by DAS, Π k represents the operation of setting all dimensions except the dimension represented by k to zero, f θ Represents a deep neural network. Represents the spatial points near the spatial point k. Where N is the number of spatial points taken (usually an odd number), and s is the interval of the spatial points taken. Next, the denoising of seismic data is achieved by optimizing the parameter θ of the deep neural network:

[0039] argmin θ ‖g θ (x)-x‖ 2 (3)

[0040] The present invention further innovates on the above unsupervised deep learning denoising theory. The denoising process of the above theory can be Figure 2 As summarized in the upper part of , the gray waveform is the waveform that needs to be denoised. The deep neural network attempts to predict the current waveform by sampling waveforms near the waveform. Since the seismic signal can be predicted by these waveforms, and the noise is unpredictable, the output of the deep neural network theoretically only has the signal part. However, not all noise is unpredictable. Some noise shows spatial correlation and can be predicted by the surrounding waveforms. Therefore, the present invention further guides the denoising process of the deep neural network through the signal received by the traditional seismic detector. Define g θ for:

[0041]

[0042] The present invention replaces all the spatial points collected by DAS in formula (2) with the set of spatial points arranged by traditional seismic detectors. In this case, g θ It still has J invariance, and the optimization objective becomes:

[0043] argmin θ ‖g θ (x)-x k ‖ 2 ; (5)

[0044] where x k Represents the waveform received by the seismic detector at spatial point k.

[0045] The noise reduction method of the present invention is shown as follows: Figure 2 The deep neural network no longer predicts all noisy data, but only predicts the waveform received by the traditional seismic detector. Since the waveform received by the traditional seismic detector lacks noise components, it can avoid the interference of spatial coherent noise on the denoising result.

[0046] 2. Model training

[0047] According to the above introduction to the principle of the method, the present invention does not specify a specific deep neural network structure. Theoretically, all Figure 2 Any deep neural network structure that meets the input and output data shape requirements can be used in the present invention. In a preferred embodiment, the present invention adopts the U-Net network structure (Ronneberger et al., 2015), and the specific structure is as follows: Figure 3 As shown in the figure. This structure can extract multi-scale features from the input data, effectively transmit high-frequency information of the data, improve the convergence of the model training, and avoid gradient explosion and gradient disappearance problems. The Adam algorithm is used for U-Net network optimization, and the initial learning rate is set to 0.001, the number of training cycles is 100, and the learning rate is reduced by half every 20 cycles.

[0048] 3. Denoising evaluation indicators

[0049] In order to compare the denoising effect of the method proposed in the present invention and the existing deep learning self-supervised algorithm, three comparison indicators are introduced: Mean Square Error (MSE), Signal-to-Noise Ratio (SNR) and Pearson Correlation Coefficient (PCC). Mean Square Error (MSE) is a commonly used indicator to measure the difference between the predicted value and the actual value. It is the average value of the residual (square of the difference) between the actual value and the model predicted value. Assuming that the total number of sampling points of DAS acquisition data is N, the noisy seismic data and the noise-free seismic data are represented by K and I respectively, the mean square error MSE is defined as:

[0050]

[0051] The signal-to-noise ratio (SNR) is an indicator of signal quality, which is used to describe the ratio between the signal strength and the background noise strength. The higher the signal-to-noise ratio, the greater the ratio of useful information to noise in the signal, and the better the signal quality. The signal-to-noise ratio is usually expressed in decibels (dB). The calculation formula for SNR is as follows:

[0052]

[0053] Among them, P signal and P noise is the power of the signal and noise. For a discrete digital signal x(n), assuming its length is N, its power P can be expressed as:

[0054]

[0055] The Pearson correlation coefficient (PCC) is an indicator to measure the similarity between two signals. It is a value between -1 and 1, which is used to measure the strength of the linear association between two continuous variables. The larger the absolute value, the stronger the linear relationship. The calculation formula of PCC is:

[0056]

[0057] in are the mean of the noisy seismic data and the mean of the noise-free seismic data, respectively.

[0058] The distributed acoustic sensor data denoising method based on high-quality sparse prior proposed in the present invention is applied to the noise suppression work of actual DAS recording. Due to the complex DAS monitoring deployment environment and the inherent noise of the instrument, it is difficult to collect noise-free seismic data in actual data collection. However, the evaluation indicators of the noise suppression algorithm, mean square error, signal-to-noise ratio and correlation coefficient, all require noise-free seismic data for evaluation. Therefore, the present invention obtains noise-free seismic data by physical simulation, and adds it to the real noise data collected by DAS to obtain the noisy seismic data to be denoised. Figure 5 This is an example of data set synthesis. The leftmost sub-image is the seismic waveform obtained by physical simulation through the spectral element method, the middle sub-image is the noise in the real DAS record, and the right one is the noisy seismic data of the two added together, which is the target to be denoised.

[0059] Before the noise reduction method proposed in the present invention is applied to noise suppression, the input data is preprocessed in an appropriate manner, including bandpass filtering, resampling and data standardization. Figure 6 The frequency range of seismic data and noise is shown. The filter range is selected as 5-40Hz based on the dominant frequency range of seismic data. The data is then resampled to 200Hz, and the data is normalized by subtracting the mean and dividing by the standard deviation. At this time, the waveform data input to the network are all data with a mean of 0 and a standard deviation of 1. This operation can simplify the data features and increase the convergence speed of the network.

[0060] The noise suppression algorithm for distributed fiber acoustic sensor data based on unsupervised learning proposed by van den Ende et al. (2021) was selected as the comparison object of the noise reduction method proposed in the present invention. In order to ensure that the unsupervised denoising algorithm compared with the noise reduction method proposed in the present invention has the best effect, we conducted a comparison test of different parameters on the parameters that mainly affect the effect in the algorithm (the number of waveforms N input to the deep neural network and the interval s between the waveforms), and the comparison results are shown in Table 1. According to the comparison results, the best parameters (N=11, s=3) were selected to compare the effect with the noise reduction method proposed in the present invention.

[0061] Table 1

[0062]

[0063] Figure 7 The comparison of noise suppression effects under different signal-to-noise ratios is shown. The three rows in the figure represent denoising experiments under different signal-to-noise ratios (-10dB, 0dB and 10dB). The leftmost column in the figure is the noisy data, the middle column is the noise suppression result of the existing unsupervised learning, and the right column is the noise suppression result of the noise reduction method proposed in the present invention. The results in the figure can intuitively show that under different signal-to-noise ratios, the noise reduction method proposed in the present invention can further suppress noise and improve signal quality. Figure 8 The noise suppression results were quantitatively analyzed from three statistical indicators: mean square error, signal-to-noise ratio and correlation coefficient. The results show that the method proposed in the present invention is superior in different signal-to-noise ratios and different evaluation indicators.

[0064] By comparing and analyzing the results of example applications, it is proved that the distributed acoustic sensor data denoising method based on high-quality sparse prior proposed in the present invention can effectively improve the noise suppression effect of distributed optical fiber acoustic sensor data. Given that DAS is now widely used in a variety of application scenarios such as traffic monitoring, resource exploration, and underground structure exploration, the denoising method proposed in the present invention has significant reference value for the actual engineering application of DAS.

[0065] It is obvious to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential features of the present invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting, and the scope of the present invention is defined by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention, and any reference numerals in the claims should not be regarded as limiting the claims involved.

[0066] In addition, it should be understood that although this specification is described in terms of implementation methods, not every implementation method contains only one independent technical solution. This narrative method of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

Claims

1. A distributed acoustic sensor data denoising method based on high-quality sparse prior, characterized in that: The waveform received by the seismic detector guides the deep neural network to denoise the seismic data based on the distributed fiber optic acoustic sensor signal; specifically, it includes: Select multiple spatial points on the optical fiber laying path, and deploy seismic detectors at each spatial point; Based on the J-invariance denoising theory, we define a function g with J-invariance. θ : Where K′ represents a set of multiple spatial points selected on the optical fiber laying path, k∈K′ represents any spatial point in K′, and Π k represents the operation of setting all dimensions except the dimension represented by k to zero, f θ represents a deep neural network; represents the spatial points near the spatial point k, where N is the total number of spatial points selected on the optical fiber laying path, and s is the interval between the spatial points; The parameters θ of the deep neural network are optimized in the following way to achieve denoising of seismic data: argmin θ ‖g θ (x)-x k ‖ 2 ; where x k Represents the waveform received by the seismic detector at spatial point k.

2. The method for denoising distributed acoustic wave sensing data based on high-quality sparse prior according to claim 1 is characterized in that: The waveform received by the seismic detector is introduced into the deep neural network as prior knowledge to guide the deep neural network to denoise the seismic data based on distributed optical fiber acoustic sensor signals. The deep neural network adopts the U-Net network.

Citation Information

Patent Citations

  • Unsupervised one-pot multi-frame image denoising method based on mutual information loss function

    CN114529479A

  • Strong VSP coupling noise suppression method based on unsupervised iterative robust deep learning

    CN115980855A

  • Polyester polyol for manufacturing polyurethane, method for manufacturing the same and polyurethane composition comprising the same

    KR1020240081181A

  • Device and method for texture-aware self-supervised blind denoising using self-residual learning

    US20240202883A1