Distributed acoustic sensing (DAS) is a new tool with low cost, sensitive signal capture, and complete coverage for vertical seismic profile (VSP) acquisition. Although DAS has obvious advantages over geophones, some weaknesses may limit its application. The main challenge that polluted by various types of noise, including optical abnormal random background fading so on. To suppress these novel noises, we developed denoising neural network based on singular spectrum analysis—multichannel...

In recent years, the denoising of low-frequency desert noise has been significant and difficult point in processing seismic data. Traditional random suppression methods could not get a good result data areas. Moreover, convolutional neural network (CNN) made notable achievements many fields recently. order to denoise areas improve signal-to-noise ratio (SNR), CNN is introduced process According characteristics data, we designed new suitable for training denoising, which named DnResNeXt....

Deep learning (DL) exhibits excellent performance in seismic noise suppression, and DL successes are attributed to its ability learn rich representations from a large amount of data. However, obtaining numerous high-quality labeled data is challenging owing confidentiality, regional sensitivity, manual labeling, which limits the capability DL. To reduce dependency improve network generalization, this study proposes novel denoising architecture based on small-sample transfer learning. The...

In seismic exploration, random noise is an obstacle to the extraction of effective signals, so investigation aimed at basis signal processing. It great significance analyze properties and establish accurate models. Since complex changes actual medium seriously affect propagation characteristics, it necessary a model in more realistic medium. this letter, we suppose weakly heterogeneous whose vary with position. And link between Lam constants established. Therefore, wave equation deduced that...

Deep-learning methods facilitate the development of seismic data processing methods; however, they also offer some challenges. The primary challenges are lack labeled samples for training, due to heterogeneity in data, expensive acquisition apparatus, and confidentiality. These problems limit high-quality training data. To solve this problem, we have developed variational autoencoding (VAE) generate synthetic noise augmentation; simplified Kullback-Leibler (KL) distance definition parameter...

Abstract. In contemporary urban environments, there has been a notable increase in the construction of high-rise edifices with glass curtain walls. While walls contribute to aesthetic appeal buildings, they also give rise number challenges. recent decades, cleaning robots have developed address issue maintaining However, challenge efficiently avoiding obstacles and gaps on remains. The research aims optimize trajectory wall robots. A series simulations were conducted using MATLAB 2024b...

Distributed acoustic sensing (DAS) is an emerging seismic acquisition technique with great practical potential. However, various types of noise seriously corrupt DAS signals, making it difficult to recover particularly in low signal-to-noise ratio (S/N) regions. Existing deep-learning methods address this challenge by augmenting data sets or strengthening the complex architecture, which can cause overdenoising and a computational power burden. Hence, heterogeneous knowledge distillation...

Summary The passive seismic interferometry, harnessing ambient noise or unconventional sources, has garnered widespread attention in the fields of Earth science and resource exploration. Conventional interferometry requires several assumptions to be satisfied, including uniform distribution subsurface an adequate number long recording periods. However, these often fall short real-world scenarios, leading suboptimal reconstruction quality subsequently impacting imaging results. Therefore, we...

Summary The pivotal role of seismic velocity inversion in oil and gas exploration geological research has been widely acknowledged. However, conventional methods face challenges such as strong reliance on initial models high computational costs. Velocity based deep learning primarily relies active source data, training neural networks to learn the mapping between records subsurface velocities. In contrast, signals passive data typically exhibit a relatively broad frequency band, encompassing...

The modelling technique contributes to understanding noise nature and properties. Prior work has established a primary random model in the homogeneous medium, however, this strict assumption of medium may be not valid under actual environment so that will reduce accuracy. Therefore, paper, is mixed heterogeneous improve accuracy, scattered mechanism used describe field medium. Since perturbation method always applied solve scattering problem, wave can regarded as superposition unperturbation...

Summary In vertical seismic profile (VSP) exploration, distributed optical fiber acoustic sensing (DAS) technology uses to simultaneously complete formation strain detection and signal transmission. Compared with traditional electronic detectors, this has the advantages of higher spatiotemporal accuracy, smaller sampling interval lower layout cost. However, borehole recordings explored by DAS are usually contaminated multi-types noise strong energy, such as ringing noise, fading background...

Summary Contamination of seismic data by background noise causes difficulties for the following inversion, imaging, stratigraphic interpretation, etc. Desert records pose a particular problem because strong energy desert random and its serious spectrum overlapping with effective signals. Thus, robust principle component analysis (RPCA) is introduced into denoising data. RPCA classical method low-rank matrix recovery. By kernel norm optimization, it can decompose noisy optimal (LM) sparse...

Noise reduction is an essential step in seismic data processing. In this paper, we are studying the combination of conventional methods and emerging deep learning for denoising. Based on residual strategy, use a convolutional neural network to denoise shearlet domain, avoiding complicated tedious process threshold selection. Moreover, Shearlet transform has characteristics sparse representation, scale division, direction which conducive extraction signal noise features by reduces training...

The seismic background noise seriously decreases the quality of data. Effective denoising methods can significantly increase signal-to-noise ratio (SNR) and resolution Recently, deep-learning-based have shown more remarkable performance than traditional methods. In this paper, we propose a novel convolutional neural network (CNN) based on gradual strategy, call it CNN (GD-CNN). proposed GD-CNN contains feature extraction sub-network prediction sub-network. former extracts high-dimensional...