Desert Noise Research Articles

Desert Seismic Low-Frequency Noise Attenuation Using Low-Rank Decomposition-Based Denoising Convolutional Neural Network

Desert seismic data are often characterized by low signal-to-noise ratio (SNR) due to the fickle surface conditions and desert random noise with nonstationarity, nonlinearity, spatial directivity, and low-frequency characteristics. This low SNR is likely to affect the following inversion and interpretation. Therefore, robust noise attenuation is crucial to improve the SNR of desert seismic data. We propose a novel method alternating direction method of multipliers-based denoising convolutional neural network (ADMM-CNN) by combining low-rank decomposition with feed-forward denoising convolutional neural network (DnCNN). DnCNN is a deep-learning-based method for noise removal, which can make good noise attenuation performance through training. However, there is no feature extraction procedure before model learning in its structure, so DnCNN cannot make full use of the prior information of signals. Combining low-rank decomposition just addresses this issue. Compared with DnCNN, our method has two main novelties. First, we use the synthetic seismic data that resemble the characteristics of field desert seismic data and desert noise to train the ADMM-CNN. Second, we use ADMM to decompose the data into three layers (low-rank, sparse, and perturbation) which are used as the inputs of three channels neural network. Through this decomposition, the neural network can capture more features and prior information of desert seismic data. Both synthetic field data tests demonstrate the robuster performance of ADMM-CNN compared to traditional methods and DnCNN, also that the ADMM-CNN method suppresses the desert noise more thoroughly and greatly improves SNR of desert seismic data at the same time.

Noise suppression method based on multi-scale Dilated Convolution Network in desert seismic data

Seismic data denoising is an important mean to extract useful information from seismic data, remove interference, and improve the SNR(signal-to-noise ratio) of seismic data. Therefore, the research on denoising methods of seismic data has always been a hot topic. At present, most convolutional neural networks for desert seismic data denoising use single scale convolutional kernels to extract feature information, which is prone to cause missing details. Therefore, we propose the Multi-scale Dilated Convolution Network (MDCN) to remove desert seismic noise. Dilational convolution operators of different sizes are used to autocratically extract features of different scales from seismic data. The extracted features are then connected in series and fused into multi-scale information used for denoising. Moreover, using dilated convolutions can increase the receptive field, so that the output of each convolution would contain a larger range of information than single scale convolutional neural networks, which means they have access to a larger window and as a result can use temporal information. In order to increase the receiving range of the network and obtain more context information, we cascade multiple modules to form a deep network. In this way, we can extract as much detailed information as possible from the desert seismic data. The results of the experiment show that our method effectively suppresses the desert noise and also better retains the effective signal.

Parameter-shared variational auto-encoding adversarial network for desert seismic data denoising in Northwest China

One of the key and difficult points in seismic data processing is seismic data denoising. Under the influence of the acquisition environment, the collected seismic data usually have low signal-to-noise ratio (SNR) and low resolution. In desert region of western China, different from other regions, the random noise there has complex characteristics of non-Gaussian, non-stationary and non-linearity. Its main frequency is quite low and its spectral overlap with those of effective signals. Moreover, large amount of data and intelligent requirements make traditional denoising methods encounter difficulties. In addition, the existing deep learning denoising methods also reveal two problems: first, they can only effectively suppress simple seismic noise such as Gaussian noise. But for the case of desert noise, they usually mistakenly judge noise as effective signals and destroy signal characteristic structure. Second, what is more important, in supervised-based methods paired noisy and pure data are usually used as training set. However, there is no noise-free data in actual desert seismic data, which severely limits the processing performance of supervised-based networks. Thus, a new parameter-shared variational auto-encoding adversarial network (PS-VAAN) is proposed for desert seismic data denoising in this paper. This new method includes two encoders, two generators and two discriminators. It can realize domain conversion from noisy data domain to pure data domain. We construct relatively complete unpaired pure and noisy data training sets for training, and design reconstruction loss function and adversarial loss function to optimize network parameters. Moreover, cycle-consistency loss is introduced to make sure two domains mapping into the same space. Compared with the state-of-art denoising methods on synthetic and actual seismic records, the proposed method not only has superior denoising ability, but also has strong feature extraction and fitting ability to recover effective signals with almost no energy loss.

Multi-scale DCFF network: a new desert low-frequency noise suppression method

Seismic exploration is an essential way for stratigraphic information acquisition and resource exploitation. However, the unique near surface conditions of the desert region in northwest China pose a special problem. On the one hand, compared with Gaussian noise, desert noise concentrates in the low-frequency band, which is seriously overlapped with those of the effective signals; On the other hand, its nonstationary, nonlinear, and non-Gaussian characteristics seriously affect the accuracy of weak signal recovery and increase the difficulty of noise suppression. In the collected field data, effective signals are often submerged by the intense low-frequency noise. Considering that traditional denoising methods have some limitations on this kind of noise, a new multi-scale dense connection feature fusion (MS-DCFF) denoising convolution neural network is presented in this paper. This denoising network can adaptively learn the potential features of effective signals through multi-scale feature fusion techniques and increase the degree of information exchange by utilizing the dense connections between different blocks. Moreover, we construct relatively complete training dataset containing an effective signal dataset and a noise dataset for desert low-frequency noise suppression, thereby boosting the feasibility of the network for desert noise suppression. Both simulation and field data prove that the superior performance of the proposed MS-DCFF in the aspect of low-frequency noise (mainly includes random noise and surface waves) suppression and effective signal recovery, compared with band-pass filtering, f-x filtering, and DnCNNs.

Intelligent random noise modeling by the improved variational autoencoding method and its application to data augmentation

Deep-learning methods facilitate the development of seismic data processing methods; however, they also offer some challenges. The primary challenges are the lack of labeled samples for training, due to heterogeneity in seismic data, expensive acquisition apparatus, and data confidentiality. These problems limit the acquisition of high-quality training data. To solve this problem, we have developed variational autoencoding (VAE) to generate synthetic noise for data augmentation; however, the simplified Kullback-Leibler (KL) distance definition and parameter learning result in the outputs of the original VAE being blurry. To optimize VAE for simulating random desert noise and improve its simulation capability, here we have developed an improved VAE based on KL redefinition and learning parameter replacement. Specifically, we (1) build a training set containing desert random noise samples, (2) redefine the KL distance calculated between two Gaussian mixture densities (rather than two simple Gaussians) because the KL distance plays an important role in the learning accuracy of VAE, and (3) use [Formula: see text] rather than [Formula: see text] to improve the learning efficiency. Statistical analysis indicates that the simulated random noise is statistically indistinguishable from real noise, indicating that our improved VAE is suitable for noise modeling. We also trained a denoising convolutional neural network (DnCNN) using the simulated noise. Data augmentation conducted using the simulated noise improved the effect of DnCNN, proving that our method contributes to data augmentation.

DnResNeXt Network for Desert Seismic Data Denoising

In recent years, the denoising of low-frequency desert noise has been the significant and difficult point in processing seismic data. Traditional random noise suppression methods could not get a good result in processing seismic data in desert areas. Moreover, convolutional neural network (CNN) has made notable achievements in many fields recently. In order to denoise seismic data in desert areas and improve the signal-to-noise ratio (SNR), CNN is introduced to process seismic data. According to the characteristics of desert seismic data, we designed a new network suitable for desert seismic data training and denoising, which is named DnResNeXt. Then, to form a mapping from the noisy data to the pure desert noise, we build a mass of training sets to train the denoising network. Thus, the network can predict the noise, then by subtracting the predicted noise from the noisy data, the denoised data are obtained. Consequently, compared with the traditional methods in suppressing random noise, DnResNeXt network has obvious advantages in both simulation and actual experiments.

Desert mixed seismic noise suppression by using multiple forward models and a supervised deep-learning method

Desert seismic data pose a particular problem due to its low signal-to-noise ratio and severe spectral overlapping. Most of the existing denoising methods have negative suppression effect on the desert mixed seismic noise, which mainly includes random noise and surface waves. The convolutional neural network (CNN) is a classical supervised deep-learning method that has proven to be an effective tool for noise suppression. The drawback is that CNN usually needs a large amount of ideal noise-free data to complete network training, which largely limits its applications to seismic data denoising. In this paper, we combine forward modelling with CNN to overcome this technical problem. Specifically, we construct multiple forward models from different physical parameters to obtain a noise-free seismic dataset by finite difference method. After adding noise data taken from a real desert noise record, we use the above data as the input for CNN training. Then, the MSE loss function allows us to obtain an optimal denoising model for the mixed noise suppression of desert seismic data. Continuing with this optimal model, we can predict mixed noise, which, once subtracted from the desert original seismic data, provide the denoised desert seismic data. Experiments with synthetic and real data demonstrate that this CNN-based denoising method can suppress the desert mixed seismic noise successfully. Several tests using artificially generated models with a different number of signal patches prove that the forward modelling can significantly improve the continuity of events and prevent the appearance of false events. In addition, analysis of the effect of network depth supports the effectiveness of our method through a trade-off among higher signal-to-noise ratio, lower root mean square error and shorter training time.

Seismic exploration desert noise suppression based on complete ensemble empirical mode decomposition with adaptive noise

A foundation for the imaging and interpretation of stratigraphic structures is the high quality seismic data. The desert seismic events are contaminated by strong random noise. The random noise in desert seismic records has complex features, including nonlinear, non-Gaussian and low frequency, which is different from the random noise in people's previous cognition. In addition, the effective signal of desert seismic record and noise has the same frequency band. These situations have brought great difficulties to the denoising of conventional methods. While the conventional denoising algorithm for dealing with seismic signals is mainly for the suppression of random Gaussian noise. So the suppression result of random desert noise is not ideal. Based on these problems, we introduce the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) algorithm into the noise reduction process of desert seismic records. Compared with other time-frequency analysis methods, CEEMDAN algorithm has significant feature recognition effect, and remains the characteristics of adaptability. At the same time, it overcomes the problems of modal aliasing and false components. The experimental results show that denoising effect is obviously better than those of conventional denoising methods, and the suppression of the surface wave in the real seismic record is relatively thorough.

Low-frequency noise suppression of desert seismic data based on signal oscillation component decomposition

Exploration of oil and gas resources in complex areas has always been a key issue in seismic exploration. The Tarim Basin, covered by a large area of desert, contains abundant oil-gas reservoir resources. However, due to the complex exploration environment, the signal-to-noise ratio of seismic data collected in desert area is low and the resolution is poor. The frequency band overlaps between effective reflected signal and low-frequency desert noise. Aiming at the specific characteristics, a desert noise suppression method based on signal oscillation component decomposition had been proposed for desert seismic data. According to the oscillation characteristics of desert seismic data, tunable-Q wavelet transform was introduced to achieve oscillation component decomposition. We put forward the multi-Q (multi-quality factors) model to determine basis function, so as to achieve the better matching of signals and noise. It can better adapt to the decomposition of desert seismic data, and complete the separation of reflected events and low-frequency desert noise with overlapping frequency band under the condition of low signal-to-noise ratio. It can be seen from the processing results of synthetic records and real data that the proposed method has better advantages in desert noise suppression compared with the Shearlet transform, empirical mode decomposition and F-x deconvolution methods.

Denoising of desert seismic signal based on synchrosqueezing transform and Adaboost algorithm

Seismic data in desert area generally have low signal-to-noise ratio (SNR) due to special surface conditions. Desert noise is characterized as low-frequency, non-Gaussian and non-stationary noise, which makes the noise suppression in desert area more challenging by conventional methods. Conventional methods are effective for the signal with high SNR, but in desert seismic signal, the SNR is low and the signal can easily be obliterated in desert noise. In this paper, we propose an approach that operates in synchrosqueezing transform (SST) domain and use classification techniques obtained from supervised machine learning to identify the coefficients associated with signal and noise. First of all, we transform the real desert seismic data into time–frequency domain by SST. Secondly, we select features by calculating the SST coefficients of signal and noise. And then, we train them in the Adaboost classifier. Finally, when the training is completed, we can obtain the final classifier that can effectively separate the signal from noise. We perform tests on synthetic and field records, and the results show great advantages in suppressing random noise as well as retaining effective signal amplitude.

Desert seismic noise suppression based on multimodal residual convolutional neural network

Seismic exploration is an important means of oil and gas detection, but affected by complex surface and near-surface conditions, and the seismic records are polluted by noise seriously. Particularly in the desert areas, due to the influence of wind and human activities, the complex desert noise with low-frequency, nonstationary and non-Gaussian characteristics is produced. It is difficult to extract effective signals from strong noise using existing denoising methods. To address this issue, the paper proposes a new denoising method, called multimodal residual convolutional neural network (MRCNN). MRCNN combines convolutional neural network (CNN) with variational modal decomposition (VMD) and adopts residual learning method to suppress desert noise. Since CNN-based denoisers can extract data features based on massive training set, the impact of noise types and intensity on the denoised results can be ignored. In addition, VMD algorithm can sparsely decompose signal, which will facilitate the feature extraction of CNN. Therefore, using VMD algorithm to optimize the input data will conducive to the performance of the network denoising. Moreover, MRCNN adopts reversible downsampling operator to improve running speed, achieving a good trade-off between denoising results and efficiency. Extensive experiments on synthetic and real noisy records are conducted to evaluate MRCNN in comparison with existing denoisers. The extensive experiments demonstrate that the MRCNN can exhibit good effectiveness in seismic denoising tasks.

Application of adaptive iterative low-rank algorithm based on transform domain in desert seismic signal analysis

Elimination of random noise is crucial in seismic data processing. Especially in desert area, field record generally has problems of weak effective reflection wave and strong noise due to its special surface factors. Besides, desert noise has characteristics of low-frequency, non-stationary and non-Gaussian. Thus, it is difficult to separate the effective signal from desert noise in low-frequency band. In order to solve these problems, this paper proposes an iterative low-rank denoising method based on synchrosqueezed wavelet transform (SWT). The algorithm first transforms seismic signal into time–frequency domain by SWT, then the signal is decomposed by iterative low-rank decomposition. Different from a traditional low-rank algorithm, this paper performs an adaptive iterative convergence on low-rank decomposition algorithm. When the error of decomposition reaches the predetermined range, the effective low-rank component is extracted. In the end, the low-rank matrix is converted back to time domain by inverse SWT to achieve the denoising. The results of the synthetic and field records verify the effectiveness of the proposed method so that it can be applied to the denoising of desert seismic data. In addition, the surface waves in real desert seismic record have obvious suppression effects and the advantages of the algorithm are shown in the comparison experiments.

Application of SNMF-MSS in low-frequency seismic noise suppression in desert areas

Noise suppression is an important step in seismic data processing. Various geographical environments make random noise in the acquired seismic records different in characteristics. Random noise in desert areas has non-Gaussian, non-linearity, non-stationary and heterogeneity. What is more, its frequency range is generally below 15 Hz, showing distinct low-frequency characteristics. Therefore, it brings much difficulty to the suppression of desert noise. Most of the existing noise suppression methods perform well on the condition of high-frequency noise, but perform worse for low-frequency noise because the denoising process of some methods depends on the waveform of noisy signal. And in some transformed methods, signal and noise are mixed with each other in the transformed domain. Thus, to resolve this problem, this paper presents a novel multiband spectrum separation (MSS) method based upon sparse nonnegative matrix factorization (SNMF) algorithm. It firstly decomposes noisy signal into some sub-signals of different narrow frequencies bands. Then, signal and noise components of these sub-signals are further separated by spectral unmixing in time-frequency domain. Several groups of experiments were carried out on synthetic records and real seismic data in Tarim oilfield in the northwest of China. Moreover, we compare its denoising effect with time-frequency peak filtering (TFPF), empirical mode decomposing (EMD) and Radon transform methods. The experimental and comparison results demonstrate that the proposed SNMF-MSS method performs better than the other three methods in the aspect of low-frequency suppression and weak signal recovery.

Low-Frequency Desert Noise Intelligent Suppression in Seismic Data Based on Multiscale Geometric Analysis Convolutional Neural Network

Existing denoising algorithms often need to meet some premise assumptions and applicable conditions, such as the signal-to-noise ratio (SNR) cannot be too low, and the noise needs to obey a specific distribution (such as Gaussian distribution) and to satisfy some properties (such as stationarity). For the desert noise that shares the same frequency band with the effective signal and has complex characteristics (nonlinear, nonstationary, and non-Gaussian), it is difficult to find a universally applicable method. In response to this problem, a multiscale geometric analysis (MGA) convolutional neural network (CNN) is proposed in this article. One of the most important features of the CNN is that it can extract data-rich intrinsic information from the training set without relying on a priori assumption. By introducing the CNN into the MGA, a new kind of denoising method can be created, which can achieve good results even under a low SNR. This article takes the nonsubsampled contourlet transform as an example to create a denoising network named NC-CNN for high-efficiency and intelligent denoising of desert seismic data. The processing results of synthetic seismic records and field seismic records prove that NC-CNN can effectively suppress the low-frequency noise (random noise and surface wave), and the effective signal almost has no energy loss. In addition, the reconstruction ability of the missing signals is also an advantage of this method.

Desert Seismic Low-Frequency Noise Attenuation Based on Approximate-Message-Passing-Based Complex Diffusion

High-quality seismic exploration data are the basis of geological exploration. However, due to the increasingly complex environment in the exploration area, the signal-to-noise ratio (SNR) of seismic data is getting lower and lower. Eliminating noise is the most direct mean to improve the SNR and quality of seismic data. The Tarim Basin with a desert surface in China is rich in energy, but due to the variability of surface conditions, the desert noise has nonstationary, nonlinear, spatial directivity, low frequency, and other characteristics, which make noise attenuation very difficult. Aiming at this problem, we develop an extension of complex diffusion filter, namely approximate-message-passing-based complex diffusion, which is used to eliminate noise in desert seismic exploration data and enhance effective signals. Compared with the traditional complex diffusion, this letter adds a noise estimate in its diffusion processing, which improves the algorithm’s ability to eliminate desert noise. The experimental results of synthetic and field seismic data show that the proposed method is more effective than other methods. Not only it can remove the noise and surface wave in the desert seismic record but also can preserve the effective signal. Therefore, we believe that the proposed method can provide a new idea for desert seismic signal processing.

Low-Frequency Noise Suppression Method Based on Improved DnCNN in Desert Seismic Data

High-quality seismic data are the basis for stratigraphic imaging and interpretation, but the existence of random noise can greatly affect the quality of seismic data. At present, most understanding and processing of random noise still stay at the level of Gaussian white noise. With the reduction of resource, the acquired seismic data have lower signal-to-noise ratio and more complex noise natures. In particular, the random noise in the desert area has the characteristics of low frequency, non-Gaussian, nonstationary, high energy, and serious aliasing between effective signal and random noise in the frequency domain, which has brought great difficulties to the recovery of seismic events by conventional denoising methods. To solve this problem, an improved feed-forward denoising convolution neural network (DnCNN) is proposed to suppress random noise in desert seismic data. DnCNN has the characteristics of automatic feature extraction and blind denoising. According to the characteristics of desert noise, we modify the original DnCNN from the aspects of patch size, convolution kernel size, network depth, and training set to make it suitable for low-frequency and non-Gaussian desert noise suppression. Both simulation and practical experiments prove that the improved DnCNN has obvious advantages in terms of desert noise and surface wave suppression as well as effective signal amplitude preservation. In addition, the improved DnCNN, in contrast to existing methods, has considerable potential to benefit from large data sets. Therefore, we believe that it can open a new direction in the area of seismic data processing.