Computers & Geosciences最新文献

Accurately identifying grain types in thin sections of sandy sediments or sandstones is crucial for understanding their provenance, depositional environments, and potential as natural resources. Although traditional computer vision methods and machine learning algorithms have been used for automatic grain identification, recent advancements in deep learning techniques have opened up new possibilities for achieving more reliable results with less manual labor. In this study, we present Trans-SedNet, a state-of-the-art dual-modal Vision-Transformer (ViT) model that uses both cross- (XPL) and plane-polarized light (PPL) images to achieve semantic segmentation of thin-section images. Our model classifies a total of ten grain types, including subtypes of quartz, feldspar, and lithic fragments, to emulate the manual identification process in sedimentary petrology. To optimize performance, we use SegFormer as the model backbone and add window- and mix-attention to the encoder to identify local information in the images and to best use XPL and PPL images. We also use a combination of focal and dice loss and a smoothing procedure to address imbalances and reduce over-segmentation. Our comparative analysis of several deep convolution neural networks and ViT models, including FCN, U-Net, DeepLabV3Plus, SegNeXT, and CMX, shows that Trans-SedNet outperforms the other models with a significant increase in evaluation metrics of mIoU and mPA. We also conduct an experiment to test the models' ability to handle dual-modal information, which reveals that the dual-modal models, including Trans-SedNet, achieve better results than single-modal models with the extra input of PPL images. Our study demonstrates the potential of ViT models in semantic segmentation of thin-section images and highlights the importance of dual-modal models for handling complex input in various geoscience disciplines. By improving data quality and quantity, our model has the potential to enhance the efficiency and reliability of grain identification in sedimentary petrology and relevant subjects.

In Naturally Fractured Reservoirs (NFR) diffuse fractures arrangement results from mechanical stratigraphy and tectonic history during failure. Thus, modelling Discrete Fracture Network (DFN) requires to understand and to account for fracture relationships at bed-interface (abutment or crosscutting) in 3D through time (loading path). However, sampling fractures data meaningfully in subsurface has always been a challenge for geologist due to data scarcity.

To better understand and forecast fracture networks in stratified rocks, we study outcrops with a focus on geometric relationships between stratigraphic interfaces and fractures. This paper presents an original python toolbox called FracAbut. It is composed of 1 main and 2 auxiliary codes that quantify the geometric relation between fractures and stratigraphic interfaces from 1D (wells, scan-line) and 2D (digital image, photographs data). We calculate the Interface Impedance (II) that accounts for fracture abutment (crossing or not), persistence (single- or multi-bed) and propagation polarity (upward or downward). For each stratigraphic interface FracAbut provides information on fractures (type, number) and interface sensitivity (coupling strength).

First, we apply FracAbut on synthetic case studies, then, on naturally fractured and stratified carbonates in Berat, Albania. Using both 1D scan-line and 2D outcrop photograph, we show that i) a mechanical interface can have different coupling above and below based on propagation polarity, ii) FracAbut results can give useful insight on fracture transmissivity, iii) FracAbut is fast and efficient to quantify fracture patterns and classify mechanical interface impact; iv) they are no relation between bed thickness and fracture propagation.

Machine learning (ML) and deep learning (DL) techniques have recently shown encouraging performance in recognizing metal-vectoring geochemical anomalies within complex Earth systems. However, the generalization of these techniques to detect subtle anomalies may be precluded due to overlooking non-stationary spatial structures and intra-pattern local dependencies contained in geochemical exploration data. Motivated by this, we conceptualize in this paper an innovative algorithm connecting a DL architecture to a spatial ML processor to account for local neighborhood information and spatial non-stationarities in support of spatially aware anomaly detection. A deep autoencoder network (DAN) is trained to abstract deep feature codings (DFCs) of multi-element input data. The encoded DFCs represent the typical performance of a nonlinear Earth system, i.e., multi-element signatures of geochemical background populations developed by different geo-processes. A local version of the random forest algorithm, geographical random forest (GRF), is then connected to the input and code layers of the DAN processor to establish nonlinear and spatially aware regressions between original geochemical signals (dependent variables) and DFCs (independent variables). After contributions of the latter on the former are determined, residuals of GRF regressions are quantified and interpreted as spatially aware anomaly scores related to mineralization. The proposed algorithm (i.e., DAN‒GRF) is implemented in the R language environment and examined in a case study with stream sediment geochemical data pertaining to the Takht-e-Soleyman district, Iran. The high-scored anomalies mapped by DAN‒GRF, compared to those by the stand-alone DAN technique, indicated a stronger spatial correlation with locations of known metal occurrences, which was statistically confirmed by success-rate curves, Student's $t$‒statistic method, and prediction-area plots. The findings suggested that the proposed algorithm has an enhanced capability to recognize subtle multi-element geochemical anomalies and extract reliable insights into metal exploration targeting.

Digital watermarking technology plays a crucial role in securing trajectory data. However, as trajectory data usage scenarios continue to expand, the security requirements for it have changed from a single copyright protection to one that takes into account data integrity. Existing digital watermarking algorithms for trajectory data can only choose between implementing copyright protection or ensuring integrity, unable to simultaneously achieve both functionalities. This limitation impedes the sharing and utilization of trajectory data. A dual watermarking algorithm that combines robust and fragile watermarking was innovatively proposed to solve this problem based on the geometric domain. Firstly, a set of feature points is extracted from the trajectory, and the farthest point pair of the minimum convex hull of the feature points is set as fixed points. The robust watermark is then embedded in the angles constructed by the feature points and the fixed points using quantization index modulation. Meanwhile, the trajectory points are grouped based on the angle and distance ratio constructed from the trajectory points to the fixed points. In each group, the spatiotemporal attributes of the trajectory points are mapped to the fragile watermark, which is then embedded into the distance ratios constructed by the trajectory points. Experimental results show that the proposed algorithm achieves both copyright protection and integrity verification for trajectory data and exhibits stronger robustness and tampering localization ability. This research can provide security and privacy protection for trajectory data and contribute positively to the application of trajectory data.

Passive seismic exploration is an environmentally friendly, economical, and highly accessible exploration method that is widely used in different-scale subsurface imaging. Retrieving reflections from ambient noise is a newly developed passive seismic method that can be used in many fields such as mineral exploration and near-surface imaging, but the interferometry calculation is time-consuming because it requires a longer period of data acquisition to improve the signal-to-noise ratio of the retrieved reflections. In this study, we introduced a graphical processing unit (GPU)- based implementation of seismic interferometry for retrieving reflection responses from passive source seismic recordings. Because all traces are involved in the computation process, but the size of passive source data often exceeds available memory, repeated disk reads lead to a decrease in computational efficiency. We design a strategy of grouping computations followed by stacking to minimize disk input and output, simultaneously keeping the memory requirements low. Passive source data is read and written only once, and there is no requirement for the memory size to be greater than the data size. Additionally, acceleration technologies such as asynchronous execution, asynchronous memory transfer, and GPU-accelerated libraries are used. We test the efficiency using short data where the number of sampling points per trace is on the order of 30,000, and long data, where the number of sampling points per trace is on the order of 30,000,000. For short data, the average speedup is 4 compared with a multi-core central processing unit (CPU); for long data, the speedup can reach 24. The GPU-based implementation of interferometry greatly reduces the calculation time for retrieving reflections from passive source seismic recordings, providing a solution to the problem of large calculation in three-dimensional (3D) passive source reflection exploration.

Due to the detectors being distributed in the target medium, vertical seismic profile (VSP) is a seismic observation method that has the advantages of high resolution and a high signal-to-noise ratio. Separating the mixed wavefields into upgoing and downgoing waves can obtain more obvious dynamic and kinematic characteristics of seismic waves, guiding subsequent imaging and interpretation. The traditional method is mainly based on the pickup of first breaks. Flattening the global seismic data by first breaks can enhance the seismic events through techniques like median filter and singular value decomposition (SVD). However, this method relies on high-precision pickup of first breaks and is limited by zero offset. To address this limitation, we introduce an improved median filtering separation method. This method employs separations of the local dip angles into positive and negative through multi-window scanning (MWS). Due to the high accuracy and robustness of this method in 2-D VSP data, we propose to use the local dip angle of the wavefields to median filter the wavefield through positive and negative angles to obtain upgoing and downgoing waves. This method for wavefield separation is optimized by iteratively identifying the directions of the seismic data. However, this optimization comes at the cost of increased computational requirements, especially under high-precision conditions. To help alleviate this problem, we use multi-thread parallel computing technique on a multi-core central processing unit (CPU) to improve computational efficiency. Finally, we validate the proposed method by testing it on synthetic seismic data and field VSP data, respectively. The results show that this wavefield separation method has advantages in terms of accuracy and robustness compared to the median filtering method based on the first break picking.

To enable flexible and accurate seismic wave simulations at continental scales ($10°-60°$) based on the spectral-element method using the open-source SPECFEM3D_Cartesian package, we develop a toolkit, Cube2sph, that allows the generation of customized spherical meshes that account for the Earth’s curvature. This toolkit enables the usage of the perfectly matched layer (PML) absorbing boundary condition even when the artificial boundaries do not align with the coordinate axes. A series of numerical experiments are presented to validate the effectiveness of this toolkit. From these numerical experiments, we conclude that (1) continental-scale seismic wave simulations, especially surface wave simulations, can be more efficiently performed without the loss of accuracy by truncating the mesh at an appropriate depth, (2) curvilinear-grid PML can be used to effectively suppress artificial reflections for seismic wave simulations at continental scales, and (3) the Earth’s spherical geometry needs to be accurately meshed in order to obtain accurate simulation results for study regions larger than 8°.

Both abiotic and biotic natural spheres benefit from the high reactivity of natural waters, which are ubiquitous on planet Earth. The use of speciation-solubility codes like PHREEQC can provide a deeper understanding of aqueous equilibria and water-rock interactions. A significant number of newly derived variables are produced by these computations, which may significantly increase compared to input data. It is crucial to process vast amounts of data efficiently, particularly when dealing with datasets that contain thousands of water analyses.

To tackle this problem, we present PHREESQL, a software package designed to efficiently store and manage extensive data generated by geochemical speciation computations performed by PHREEQC. High efficiency in data extraction and filtering of the entire output from a single run can be achieved through a well-designed relational SQL database structure. The PHREESQL can be used as a stand-alone package or embedded in third-party applications. Thanks to the SQL structure, it is possible to create links with unstructured meshes by developers and experts in reaction-transport problems. Real-time data management from multiparameter devices in field and laboratory settings is made possible and efficient by parallel computation options and software integration. The toolkit encompasses both a C++ library and a command-line interface, facilitating its use by geochemists with limited programming skills.

Improving the accuracy and efficiency of seismic wavefield simulation aids geophysical problem-solving. The finite difference (FD) is widely used, but efficiency drops with increasing grids and higher order of difference formats. We propose an attention mechanism-based deep learning method called SeismicTransformer. Compared with theory-driven methods, such as the second-order central difference method, SeismicTransformer offers at least a tenfold improvement in speed. Compared with the networks without the attention mechanism, the SeismicTransformer achieves better results by utilizing global information. The proposed SeismicTransformer offers a promising solution for seismic wavefield simulation.

The feature extraction of sand grain size, color and texture is a necessary step to identify clastic components. The sand grain features are complex and various, which brings difficulties to geological identification. Aiming at these images, a cosine-enhanced tuna swarm optimized exponential entropy segmentation method is proposed, which can effectively preserve the texture features of various sand grains. Firstly, for the tuna swarm optimization (TSO) algorithm, three improvement strategies are proposed: cosine spiral movement, cosine parabolic movement and Gauss-Cauchy mutation, which improve the TSO's global and local search. This algorithm is called cosine-enhanced TSO (CETSO). Benchmark function experiments showed that the convergence accuracy and stability of CETSO are greatly improved, and the convergence speed is also slightly increased. Secondly, CETSO optimized the exponential entropy to automatically determine the segmentation thresholds, and the feasibility of the method is verified by taking the information content of the segmented image as the standard. Finally, segmentation experiments were carried out on the Yarlung Zangbo River sand microscopic image dataset, and the results show that the method has high segmentation accuracy and stability for images with high contrast, rich texture, or significant differences in the size of sand debris. Compared with TSO, the CETSO optimized exponential entropy segmented images achieved an improvement of 21% and 93% in the evaluation of the average and standard deviation of the peak signal-to-noise ratio on thirty experiments. And this method has a fast processing speed, and it only takes about 0.85s to divide an image on average, which meets the needs of engineering applications.