Borehole lithology discrimination is the foundation for formation evaluation and reservoir characterization. Due to limitation of costing or accuracy, direct methods, such as borehole core drilling cutting analysis, are unable widely apply, while logging interpretation provides an alternative solution this task. To mitigate influence subjective bias, several machine learning algorithms, neural network, support vector machine, decision tree, random forest (RF), have already been applied...

Full waveform inversion (FWI) is a powerful tool for estimating the underground velocity model. However, it computationally expensive and resulting models tend to be not accurate enough. Thus, improve efficiency accuracy of FWI, we propose super-resolution (SR) method based on deep learning enhance resolution seismic Since edge images model are also widely used in geophysics, multitask (MTL) network with hard parameter sharing applied perform SR its images. The proposed MTL dubbed M-RUDSR...

Accurate estimation of volumetric seismic dip is great significance for subsequent processing and interpretation works. Recently, with the development deep learning techniques, convolutional networks are also applied estimation. Compared traditional approaches, estimating dips not only more efficient but shows promise in accuracy robustness. However, if we take estimated by approaches as labels train on field data directly, robustness learned influenced due to error labels. An alternative...

Structural curvatures are widely used seismic attributes that help interpreters to understand both structural and stratigraphic features. Traditional curvature extractions mainly calculated from dip estimations through lateral scanning of events, which is not only a very time-costing approach but also influenced by parameter settings, frequency, data quality. In this article, we propose deep learning-based volumetric extraction directly derives volumes the response. To realize above...

Seismic inversion is aimed at building a mapping from low-resolution seismic data to high-resolution impedance data. Most of the traditional methods have satisfactory interpretability, and most parameters tend specific physical definitions. On other hand, deep learning-based present poor interpretability as their prediction performance not always clearly explainable. One significant challenges quantify uncertainty model. The includes aleatoric epistemic uncertainty, can be used evaluate...

Recently, a multitask learning framework named M: multitask, R: global residual skip connection structure, U: encoder–decoder structure of U-Net, D: dense and SR: super-resolution (M-RUDSR) has successfully improved the accuracy full-waveform inversion (FWI) results by enhancing resolution seismic velocity model. However, M-RUDSR does not make full use data even though it contains high wavenumber information, which can help enhance Moreover, effects employing realized simply increasing...

Deep learning has been applied to tackle the seismic inversion problem, bringing more efficiency and accuracy. However, bad spatial continuity poor generalizability limit practical application. To solve these problems, we propose a 2D end-to-end method based on domain adaption. Firstly, proposed network learns mapping of data under constraint adaption layer, which can reduce difference between features real synthetic data, improving generalization ability data. Then, trained model is...

Seismic stratigraphic interpretation plays an important role in geophysics and geosciences. Recently, deep learning has been explored for seismic interpretation. However, learning-based methods usually require sufficient labeled samples. This is often too hard to be satisfied field In this paper, we propose a active method address issue. Active typically exploits prediction uncertainty reduce labeling effort. We found that of easily obtained the Since adjacent images are very similar, they...

Transverse relaxation T2 spectrum obtained by nuclear magnetic resonance (NMR) logging tools is an intuitive reflection of the pore size distribution for subsurface formation, which valuable petroleum reservoir characterization. However, deployment NMR constrained financial and operational factors, while data are only available in very limited wells. This seriously limits its application practices. Therefore, researchers try to synthesize spectra from more widely measured conventional with...

Spatial variable estimation is a basic application of geostatistics. In general, this task performed based on observations limited points. For some cases, intensive observed data obtained from other sources are also available as the auxiliary variables. To utilize information in these data, methods such regression kriging (RK) or cokriging proposed. However, all assume that variables keep linear correlation with target implicitly, which not satisfied most cases. letter, through combination...

Essentially, post-stack seismic interpretation task is equivalent to a supervised learning in the category of machine which tries build an attribute map between responses and formation properties under constraint well data. For success learning, key point provide sufficient number reliable training samples. Unfortunately, due expensive construction cost, only limited samples can be provided based on data at points, cannot meet requirements effective learning. To solve this problem, article...

Abstract. Panoramic images are widely used in many scenes, especially virtual reality and street view capture. However, they new for furniture identification which is usually based on mobile laser scanning point cloud data or conventional 2D images. This study proposes to perform semantic segmentation panoramic transformed separate light poles traffic signs from background implemented by pre-trained Fully Convolutional Networks (FCN). FCN the most important model deep learning applied its...

Ground-roll is a typical Rayleigh-type interference noise in field seismic data, which characterized by low frequency, velocity and high amplitude. Since it will interfere effective signals severely degrade the signal-to-noise ratio of observed records, many approaches have been developed for ground-roll attenuation or separation. In this letter, we proposed an improved separation algorithm through combination deep learning based low-frequency generation dictionary reconstruction. Moreover,...

Detecting lane automatically from the IP camera is an important component of intelligent vision-based traffic big data system. Many previous studies focus on main lanes detection task based clustering algorithm. However, some left-turning or right-turning are ignored in these methods due to their seldom happening real scene. This paper attempts address this issue. We try detect those which appear trajectory lines by combining and curve complexity computing method. Firstly, vehicles detected...

For petroleum exploration and development, inter-well formation property estimation is very important since it the foundation of further reservoir modeling simulation. most cases, this task performed based on observations at well-points, while seismic data also provided as supplement. In essence, a spatial multi-source data. Even though various geo-statistical interpolation machine learning mapping algorithms have been proposed, they all limitations in accuracy, horizontal resolution or...