Computers & Geosciences最新文献

{"title":"Linking geo-models for geomorphological classification using knowledge graphs","authors":"Yanmin Qi ,&nbsp;Yunqiang Zhu ,&nbsp;Shu Wang ,&nbsp;Yutao Zhong ,&nbsp;Stuart Marsh ,&nbsp;Amin Farjudian ,&nbsp;Heshan Du","doi":"10.1016/j.cageo.2025.105873","DOIUrl":"10.1016/j.cageo.2025.105873","url":null,"abstract":"<div><div>Geographic computation is an important process in geographic information systems to detect, predict, and simulate geographic entities, events, and phenomena, which is performed through a series of geographic models over geographic data. However, selecting and sequencing appropriate models is challenging for users with limited knowledge. To automate the process of linking models into workflows, a knowledge graph-based approach is proposed. In this approach, the first part is to construct a knowledge graph that integrates knowledge from geographic models and domain experts. Then, an algorithm is designed to assist the constructed knowledge graph in automating model linking. This paper takes the geomorphological classification of the Hengduan Mountains in China as a case study, which geomorphological classification maps are generated by performing querying and computing through the geomorphological classification knowledge graph. Experimental results demonstrate that the proposed knowledge graph-based approach links the models into workflows automatically and generates reliable classification results.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"196 ","pages":"Article 105873"},"PeriodicalIF":4.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143093386","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"VIF-Net: Interface completion in full waveform inversion using fusion networks","authors":"Zixuan Deng ,&nbsp;Qiong Xu ,&nbsp;Fan Min ,&nbsp;Yanping Xiang","doi":"10.1016/j.cageo.2024.105834","DOIUrl":"10.1016/j.cageo.2024.105834","url":null,"abstract":"<div><div>Deep learning full waveform inversion (DL-FWI) distinguishes itself from traditional physics-based methods for its robust nonlinear fitting, rapid prediction, and reduced reliance on initial velocity models. However, existing end-to-end deep learning approaches often neglect the reconstruction of layer interfaces and faults. In this article, we propose a two-stage DL-FWI approach named Velocity Interface Fusion (VIF). The first stage comprises two subnetworks: VIF-Velocity (VIF-V) generates the intermediate velocity model, and VIF-Interface (VIF-I) predicts velocity model interfaces. They have the same UNet++ architecture and an optional Fourier transform-based preprocessing module. Their main difference lies in the binary class-balanced cross-entropy loss tailored for VIF-I. The second stage is fulfilled by a fusion subnetwork with a limited downsampling encoder–decoder structure. This network refines the intermediate velocity model using the predicted interfaces to reconstruct the final model. A dynamic learning strategy combining warm-up and cosine annealing is employed to train all three subnetworks jointly. Our method is evaluated on two SEG salt and four OpenFWI datasets using four metrics in comparison with three popular DL-FWI methods. Results demonstrate its superior performance in interface completion and reconstruction. The source code is available at <span><span>https://github.com/FanSmale/VIF-dev</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"196 ","pages":"Article 105834"},"PeriodicalIF":4.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143093819","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Thin layer identification method of prestack seismic data based on an improved 2D Teager-Huang transform","authors":"Xudong Jiang,&nbsp;ChuPeng You,&nbsp;ZeTao Zhang,&nbsp;XiaoHui Qi,&nbsp;Junxing Cao","doi":"10.1016/j.cageo.2024.105827","DOIUrl":"10.1016/j.cageo.2024.105827","url":null,"abstract":"<div><div>Thin layer identification is significant in reservoir prediction, especially in deep and ultradeep reservoirs. Due to the weak effective signals of deep reservoirs, the effectiveness of the current thin layer positioning method is significantly reduced. Therefore, an improved 2D Teager-Huang transform (2D-THT) operator based on prestack seismic data is developed in this paper to identify and locate deep thin reservoirs. First, we combine bidimensional empirical mode decomposition (BEMD) and ensemble empirical mode decomposition (EEMD) to form the bidimensional ensemble empirical mode decomposition (BEEMD) algorithm. This BEEMD algorithm can effectively avoid the mode mixing problem of BEMD and efficiently decompose gathers to separate and extract the critical information. Then, we introduce a 2D Teager-Kaiser energy operator (2D-TKEO) to enhance the lateral and longitudinal instantaneous change characteristics of the selected effective components and focus on the weak reflection signal. Finally, the Hilbert transform is used to obtain the prestack spectral decomposition results, and the time–frequency superposition spectrum is obtained through superposition. The positions of the thin layer top and bottom are obtained by locating the frequency change point. This method uses the intrinsic signal change to obtain a strong sensitivity to thin layers to achieve thin reservoir positioning research. Synthetic data and actual data application examples of deep carbonate rocks in western Sichuan show that a method complementary to the existing thin layer detection methods is proposed in this paper and that good results on low signal-to-noise ratio data are achieved.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"196 ","pages":"Article 105827"},"PeriodicalIF":4.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143093825","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Capturing exposed bedrock in the upland regions of Great Britain: A geomorphometric focused random forest approach","authors":"Chris Williams ,&nbsp;Katie Whitbread ,&nbsp;Alex Hall ,&nbsp;Sam Roberson ,&nbsp;Andrew Finlayson ,&nbsp;Romesh N. Palamakumbura ,&nbsp;Andrew Hulbert ,&nbsp;Matthew Paice","doi":"10.1016/j.cageo.2024.105814","DOIUrl":"10.1016/j.cageo.2024.105814","url":null,"abstract":"<div><div>Rock exposure distribution maps provide invaluable information for a range of applications from geohazard assessment through to aggregate resource potential assessments. Despite the usefulness of such information, it is only available to a limited extent across Great Britain (GB). Recent developments in the application of machine learning approaches to map exposed rock distribution rely on existing geological and land cover maps as the key input data for model training. We present a catchment-scale approach for delivering high-resolution rock exposure maps for GB mountain terrains. Our application has two objectives: establish a consistent and cross-applicable approach enabling feature identification from elevation datasets; use the results and diagnostics of the application to assist in further environmental process understanding. We utilize manual aerial image interpretation, and a suite of geomorphic terrain variables generated from a 5 m Digital Terrain Model as inputs to a distributed random forest model. Eight separate catchment models were derived from the training datasets using a leave-one-out approach. Aggregated results indicate a model accuracy of 79%, with a relatively high model sensitivity (78%) at the cost of relatively low precision (20%). Variable importance assessment highlighted patterns consistent with expected geomorphic controls on rock exposure related to gravity-driven slope processes in mountain landscapes. These results highlight the potential of multi-variant approaches for high-resolution rock exposure mapping, and lay a foundation for further development, particularly in relation to opportunities for further training data capture to ensure model accuracy. The ability to associate features based on geomorphological variables - indicative of landscape processes including erosion and deposition - presents opportunities that go beyond rock exposure such as for critical mineral and resource assessment. This approach will be applied for initial site characterisation as part of future onshore and offshore geological survey activities where high-resolution terrain and bathymetric data are available.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"196 ","pages":"Article 105814"},"PeriodicalIF":4.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143102304","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Object segmentation of near surface magnetic field data based on deep convolutional neural networks","authors":"Qiang Li ,&nbsp;Zhiqiao Wang ,&nbsp;Wenyu Li ,&nbsp;Jie Hu ,&nbsp;Xiaoli Rong ,&nbsp;Leixiang Bian ,&nbsp;Weitao Wu","doi":"10.1016/j.cageo.2024.105847","DOIUrl":"10.1016/j.cageo.2024.105847","url":null,"abstract":"<div><div>Near surface magnetic data contain valuable information on buried structures. Magnetic gradient tensor (MGT) can provide more detailed information than total magnetic intensity (TMI) and magnetic vectors. Traditional methods extract target-related features from magnetic data by identifying manually designed edge-features. These methods are either sensitive to changes in inclination and declination angles or have limited spatial resolution. This highlights the need for new approaches to accurately capture the complex details inherent in magnetic data. Although some studies have introduced machine learning techniques, their models often focus narrowly on identifying simple object shapes. Additionally, their models often overlook the potential of using MGT data directly as input, thereby missing the opportunity to fully utilize neural networks to interpret magnetic data. In this study, a deep convolutional neural network-based object segmentation method on MGT data was developed. We introduced a U-net-like architecture neural network model combined with attention modules to collect detailed information about buried objects from magnetic data. Attention modules refine feature maps by inferring attention maps along two dimensions: channel and spatial. The channel dimension refers to the axis along which different 2D feature maps are organized. The segmentation performance of our model was evaluated on both magnetic data with simple magnetic sources and complicated magnetic data. Compared to traditional approaches, our model provides more satisfactory segmentation results. The perspective for this work is to create a more detailed database to better reflect the intricate patterns typically observed in magnetic data. This effort aims to make our model more accurate and useful in interpreting diverse magnetic signatures.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"196 ","pages":"Article 105847"},"PeriodicalIF":4.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143093103","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Automatic recognition of debris rock lithology based on unsupervised semantic segmentation","authors":"Shengda Qin ,&nbsp;Qing Wang ,&nbsp;Qihong Zeng ,&nbsp;Maolin Ye ,&nbsp;Anqi Fu ,&nbsp;Guanzhou Chen","doi":"10.1016/j.cageo.2024.105790","DOIUrl":"10.1016/j.cageo.2024.105790","url":null,"abstract":"<div><div>Accurate identification of lithology in debris rock is crucial for optimizing resource development in geological exploration and the oil and gas industry. The traditional approach, which depends on experts manually analyzing remote sensing images, is not only laborious but also vulnerable to subjectivity. In contrast, supervised learning, although highly automated, is limited by the need for large-scale annotated data and sample imbalance issues. In our proposed unsupervised semantic segmentation method, automatic segmentation of rock images not only improves the efficiency and accuracy of lithology recognition but also reduces human errors, providing an effective solution for automated lithology analysis. We collected a large amount of debris rock data from the Qingshuihe-Karazha using remote sensing satellites and used an improved FCN network combined with super-pixel segmentation to generate pseudo labels instead of manual labeling, achieving unsupervised segmentation. We compared this method with traditional K-Means, ISODATA, and CNN + K-Means pseudo-label generation methods. By calculating evaluation metrics named ARE, AMI, and FMI, which are used for unsupervised semantic segmentation methods, we found that our method maintains high consistency and robustness in various image sizes, especially when the size of debris rock images is large, and its stability is superior. At the same time, we addressed the boundary issues caused by the need for block division in the lithology image of ultra-large debris rocks, as well as the problem of a large number of similar blocks after block division. The efficiency and accuracy of this method in lithology identification were determined, providing more convenient and efficient data processing methods for geological researchers.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"196 ","pages":"Article 105790"},"PeriodicalIF":4.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143093165","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Laplacian deep ensembles: Methodology and application in predicting dUT1 considering geophysical fluids","authors":"Mostafa Kiani Shahvandi ,&nbsp;Siddhartha Mishra ,&nbsp;Benedikt Soja","doi":"10.1016/j.cageo.2024.105818","DOIUrl":"10.1016/j.cageo.2024.105818","url":null,"abstract":"<div><div>Increasing the accuracy and reliability of deep learning models is a crucial yet challenging task. The Bayesian approach is typically inefficient in achieving this goal, because of its daunting computational complexity. A promising alternative approach is based on ensembling of models with disparate initial parameters, which result in different model predictions. However, this approach is mainly based on the assumption of Gaussian distribution for data, which might suffer from the presence of outliers, out-of-distribution data, and the lack of diversity among ensemble members. Here we propose to consider Laplacian distribution for data, and introduce Laplacian Deep Ensembles (LDE). We present the formulation of LDE and show that it is akin to the familiar L1 norm minimization, thus being more resilient to outliers and out-of-distribution data. We also introduce the repulsive form of the LDE that enhances the diversity among ensemble members and is aysmptotically convergent to the Bayesian approach. We present an application in the field of geodesy, for the short-term prediction of dUT1, which represents the deviation of universal time (tied to Earth’s rotation) from the coordinated universal time (based on atomic clocks) due to the effect of geophysical fluids (namely atmosphere, ocean, land hydrology, and sea-level variations). We show that dUT1 is predictable with high accuracy up to 10 days ahead. We demonstrate that not only is LDE more accurate than its Gaussian counterpart, but also the repulsive LDE represents on average <span><math><mo>∼</mo></math></span>12% improvement compared to alternative, state-of-the-art predictions. This improvement has considerable importance for various applications that rely on precise timekeeping, such as satellite navigation.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"196 ","pages":"Article 105818"},"PeriodicalIF":4.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143093175","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A combined deep learning and morphology approach for DFS identification and parameter extraction","authors":"Maolin Ye ,&nbsp;Qing Wang ,&nbsp;Changmin Zhang ,&nbsp;Shengda Qin ,&nbsp;Shuoyue Yan","doi":"10.1016/j.cageo.2025.105856","DOIUrl":"10.1016/j.cageo.2025.105856","url":null,"abstract":"<div><div>Since the concept of the Distributive Fluvial System (DFS) was introduced, understanding DFS river parameters has been vital for oil and gas reservoirs. Traditional measurement methods are often time-consuming and labour-intensive. This paper presents a deep learning and morphology-based method for the automatic extraction of DFS river parameters. We propose an optimized model, Seg_ASPP, which integrates Segformer and ASPP (Atrous Spatial Pyramid Pooling) to generate river network masks. The river centerline is then extracted via accumulation cost and polynomial fitting algorithms, allowing for length, width, and sinuosity calculations. Using the Geermu DFS area in the Qaidam Basin for evaluation, we compare the parameters extracted via our method against manual measurements. The average relative errors for length, width, and curvature are 10.22%, 13.57%, and 5.41%, respectively, demonstrating the strong performance of the model. Our experiments show that the DFS parameter extraction method proposed in this paper has great potential for practical applications.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"196 ","pages":"Article 105856"},"PeriodicalIF":4.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143093389","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Enhancing 3D migration images resolution: A fast and robust implicit point spread function deconvolution method based on wavenumber domain representation","authors":"Zhongfan Jia ,&nbsp;Cewen Liu ,&nbsp;Haohuan Fu ,&nbsp;Kaisheng Ma","doi":"10.1016/j.cageo.2024.105837","DOIUrl":"10.1016/j.cageo.2024.105837","url":null,"abstract":"<div><div>Image-domain least-squares migration has been proven effective in enhancing the spatial resolution of migration images. However, due to the intrinsic complexity of the 3D problem, the conventional local-stationary deblurring method cannot achieve a good balance between computational efficiency and deblurring accuracy. The explicit point spread function (PSF) deconvolution method has been proposed to improve the resolution of migration images on a point-wise bias with substantial computational and storage costs, especially for the 3D PSF deconvolution. To achieve high-resolution imaging with reduced costs, we introduce an implicit PSF deconvolution method based on wavenumber domain representation to enhance the resolution of 3D migration images. Using a deep-learning optimizer, coupled with a corresponding modified hyperparameter, we can achieve a fast convergence for the solution of the 3D PSF deconvolution operator. A loss interpolation technique is introduced to obtain non-sensitive PSF deconvolution operators for dense interpolation to reduce high-frequency artifacts. 3D synthetic and large-scale field dataset results demonstrate that our approach can produce high-resolution images with reduced migration artifacts and balanced amplitude. Additionally, our proposed method outperforms existing methods with at least three times computational cost reduction and two orders of magnitude reduction in storage cost for the 3D PSF deconvolution. It demonstrates that the proposed approach can be a cost-effective and practical high-resolution imaging tool for large-scale 3D datasets.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"196 ","pages":"Article 105837"},"PeriodicalIF":4.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143093817","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Zero watermarking for vector geographic point data based on convex layers","authors":"Qifei Zhou ,&nbsp;Qiang Zhao ,&nbsp;Yingchen She ,&nbsp;Wen Yang ,&nbsp;Luanyun Hu ,&nbsp;Weitong Chen ,&nbsp;Na Ren ,&nbsp;Changqing Zhu","doi":"10.1016/j.cageo.2025.105861","DOIUrl":"10.1016/j.cageo.2025.105861","url":null,"abstract":"<div><div>Zero watermarking plays a critical role in the copyright protection of vector geographic data in geographic information industries, as it does not distort the data. However, how to extract stable features from vector geographic point data is still a challenging problem. A zero watermarking method based on convex layers is proposed to address the issue. First, the vector geographic point data is simplified as a point set. Next, construct convex layers from the point set through convex hull peeling or onion peeling. Then, calculate the density of vertices wrapped within each convex hull of the convex layers to obtain an integer number sequence. Finally, the sequence of vertex density is binarized to the watermark by comparing each value with its adjacent counterparts. Experiments substantiate the theoretical achievements of the proposed method. Its normalized correlation coefficient always keeps at 1.00 when suffering rotation, uniform scaling, non-uniform scaling, translation attacks, and map projection transformation attacks, demonstrating superior robustness compared to the existing methodologies. The watermark uniqueness has also been verified through the analysis of a set of 100 randomly generated point datasets.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"196 ","pages":"Article 105861"},"PeriodicalIF":4.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143093826","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}