Journal of Applied Geophysics最新文献_第7页

ME-YOLO: A novel real-time detection network for pavement interlayer distress using ground-penetrating radar ME-YOLO：一种基于探地雷达的路面夹层损伤实时检测网络

IF 2.1 3区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Journal of Applied Geophysics

Pub Date : 2025-12-03 DOI: 10.1016/j.jappgeo.2025.106057

Senguo Cao , Congde Lu , Xiao Wang , Peng Zhang , Guanglai Jin , Wenlong Cai

Interlayer distress detection in asphalt pavement is critical for highway maintenance, as timely identification of pavement distress can ensure operational safety, reliability, and extended service life. However, the problems of feature information loss and the substantial confusable backgrounds significantly hinder detection accuracy. To address these limitations, we propose an enhanced network specifically designed for automated interlayer distress detection named ME-YOLO. Firstly, we design a Multiscale Adaptive Feature Fusion (MAFF) module, which aggregates more scale information by Adaptive Spatial Feature Fusion (ASFF). This design links all feature scales to make discriminative features in each scale propagate directly to subsequent modules, enriching semantic representations and mitigating the risk of feature loss, while leveraging shallow-layer features to strengthen spatial localization. Furthermore, the Efficient Partial Self-Attention (EPSA) module is introduced to suppress background interference in complex environments. Unlike conventional transformers, EPSA adopts partial self-attention operations with multi-path fusion, which can enable the network to acquire global representation capability with low computational overhead. Extensive experiments indicate that the ME-YOLO network outperforms the given state-of-the-art models, including Faster-RCNN, RT-DETR, YOLOv8s, and YOLOv11s, on the interlayer distress dataset. Compared to YOLOv5s, ME-YOLO achieves improvements of 2.2% in mAP_0.5 and 3.5% in mAP_0.5:0.95, while maintaining an inference speed of 6.7 ms per image. The source code will be available at https://github.com/caosenguo/ME-YOLO.

沥青路面夹层损伤检测对公路养护至关重要，及时识别路面损伤可以保证路面运行的安全性、可靠性和延长使用寿命。然而，特征信息的丢失和大量的背景混淆问题严重影响了检测的准确性。为了解决这些限制，我们提出了一个专门为层间自动遇险检测设计的增强网络，名为ME-YOLO。首先，设计了多尺度自适应特征融合（MAFF）模块，通过自适应空间特征融合（ASFF）聚合更多尺度信息；本设计将所有特征尺度联系起来，使每个尺度中的判别特征直接传播到后续模块，丰富语义表示，降低特征丢失的风险，同时利用浅层特征加强空间定位。此外，本文还引入了EPSA （Efficient Partial Self-Attention）模块来抑制复杂环境下的背景干扰。与传统的变压器不同，EPSA采用部分自关注的多路径融合运算，使网络能够以较低的计算开销获得全局表示能力。大量实验表明，在层间压力数据集上，ME-YOLO网络优于现有的最先进模型，包括Faster-RCNN、RT-DETR、YOLOv8s和YOLOv11s。与YOLOv5s相比，ME-YOLO在mAP0.5和mAP0.5:0.95中分别提高了2.2%和3.5%，同时保持了6.7 ms /张图像的推理速度。源代码可从https://github.com/caosenguo/ME-YOLO获得。

{"title":"ME-YOLO: A novel real-time detection network for pavement interlayer distress using ground-penetrating radar","authors":"Senguo Cao , Congde Lu , Xiao Wang , Peng Zhang , Guanglai Jin , Wenlong Cai","doi":"10.1016/j.jappgeo.2025.106057","DOIUrl":"10.1016/j.jappgeo.2025.106057","url":null,"abstract":"<div><div>Interlayer distress detection in asphalt pavement is critical for highway maintenance, as timely identification of pavement distress can ensure operational safety, reliability, and extended service life. However, the problems of feature information loss and the substantial confusable backgrounds significantly hinder detection accuracy. To address these limitations, we propose an enhanced network specifically designed for automated interlayer distress detection named ME-YOLO. Firstly, we design a Multiscale Adaptive Feature Fusion (MAFF) module, which aggregates more scale information by Adaptive Spatial Feature Fusion (ASFF). This design links all feature scales to make discriminative features in each scale propagate directly to subsequent modules, enriching semantic representations and mitigating the risk of feature loss, while leveraging shallow-layer features to strengthen spatial localization. Furthermore, the Efficient Partial Self-Attention (EPSA) module is introduced to suppress background interference in complex environments. Unlike conventional transformers, EPSA adopts partial self-attention operations with multi-path fusion, which can enable the network to acquire global representation capability with low computational overhead. Extensive experiments indicate that the ME-YOLO network outperforms the given state-of-the-art models, including Faster-RCNN, RT-DETR, YOLOv8s, and YOLOv11s, on the interlayer distress dataset. Compared to YOLOv5s, ME-YOLO achieves improvements of 2.2% in mAP<sub>0.5</sub> and 3.5% in mAP<sub>0.5:0.95</sub>, while maintaining an inference speed of 6.7 ms per image. The source code will be available at <span><span>https://github.com/caosenguo/ME-YOLO</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"245 ","pages":"Article 106057"},"PeriodicalIF":2.1,"publicationDate":"2025-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145694472","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Formation mechanism of low signal-to-noise ratio seismic data in complex Karst areas 复杂岩溶地区低信噪比地震资料形成机制

IF 2.1 3区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Journal of Applied Geophysics

Pub Date : 2025-12-03 DOI: 10.1016/j.jappgeo.2025.106056

Feng Gao , Zhangqing Sun , Xiunan Fan , Zihao Li , Fuxing Han , Jipu Lu , Wenpan Cen , Mingchen Liu , Zhenghui Gao , Jiawei Xie

Karst formations near the surface in complex geological settings scatter seismic waves, adversely affecting the signal-to-noise ratio (SNR) of seismic data. Accurately characterizing the formation mechanisms of low SNR seismic data is vital for enhancing the efficacy of seismic exploration. This study introduces a composite multi-scale random medium modeling technique that addresses the characteristics of random heterogeneous media in karst regions. The methodology superimposes various random perturbations of different scales in the same area. The elastic wave spectral element method (SEM) is employed to numerically simulate the seismic wave field in complex karst environments. A case study in Guangxi, China, demonstrates that the composite multi-scale random medium modeling approach effectively captures the characteristics of the medium. The simulated data generated using the elastic wave SEM closely resembling actual data. This paper offers insights into the formation mechanisms of low SNR seismic data in complex karst areas. These insights provide valuable references for advancing seismic data processing techniques.

在复杂的地质环境中，地表附近的岩溶地层会使地震波散射，对地震数据的信噪比（SNR）产生不利影响。准确刻画低信噪比地震资料的形成机理，对提高地震勘探效果至关重要。针对岩溶地区随机非均质介质的特点，提出了一种复合多尺度随机介质模拟技术。该方法将不同尺度的各种随机扰动叠加在同一区域。采用弹性波谱元法（SEM）对复杂岩溶环境下的地震波场进行了数值模拟。以广西为例，表明复合多尺度随机介质模拟方法能有效地捕捉介质的特征。用弹性波扫描电镜模拟得到的数据与实际数据接近。本文对复杂岩溶地区低信噪比地震资料的形成机制进行了探讨。这些见解为改进地震数据处理技术提供了有价值的参考。

{"title":"Formation mechanism of low signal-to-noise ratio seismic data in complex Karst areas","authors":"Feng Gao , Zhangqing Sun , Xiunan Fan , Zihao Li , Fuxing Han , Jipu Lu , Wenpan Cen , Mingchen Liu , Zhenghui Gao , Jiawei Xie","doi":"10.1016/j.jappgeo.2025.106056","DOIUrl":"10.1016/j.jappgeo.2025.106056","url":null,"abstract":"<div><div>Karst formations near the surface in complex geological settings scatter seismic waves, adversely affecting the signal-to-noise ratio (SNR) of seismic data. Accurately characterizing the formation mechanisms of low SNR seismic data is vital for enhancing the efficacy of seismic exploration. This study introduces a composite multi-scale random medium modeling technique that addresses the characteristics of random heterogeneous media in karst regions. The methodology superimposes various random perturbations of different scales in the same area. The elastic wave spectral element method (SEM) is employed to numerically simulate the seismic wave field in complex karst environments. A case study in Guangxi, China, demonstrates that the composite multi-scale random medium modeling approach effectively captures the characteristics of the medium. The simulated data generated using the elastic wave SEM closely resembling actual data. This paper offers insights into the formation mechanisms of low SNR seismic data in complex karst areas. These insights provide valuable references for advancing seismic data processing techniques.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"245 ","pages":"Article 106056"},"PeriodicalIF":2.1,"publicationDate":"2025-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145737524","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A multi-facies constrained geostatistical inversion for predicting fractured tight sandstone thin reservoirs: a case study in Western China 裂缝性致密砂岩薄储层的多相约束地质统计反演——以中国西部地区为例

IF 2.1 3区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Journal of Applied Geophysics

Pub Date : 2025-12-02 DOI: 10.1016/j.jappgeo.2025.106052

Sijia Li , Guangzhi Zhang , Zhenfeng Liu , Zhen Yang , Yuanyuan Tan , Jinghui Cui

This study tackles the persistent challenge of predicting thin, heterogeneous reservoirs within fault-fracture-modified tight sandstone formations by proposing a multi-facies constrained geostatistical inversion method. Traditional lithofacies-constrained geostatistical inversion methods fail to resolve high heterogeneity and accurately find low Poisson impedance “sweet spots”. To overcome these limitations, we redefine fault-fracture bodies as distinct geological facies and integrate them with lithofacies constraints within a Bayesian inversion framework to facilitate the building of the variogram with the aim of realizing partition modeling. Curvature and structural entropy are preferred as the tools to characterize the fault-fracture bodies to guide spatial modeling. Application in a western China basin demonstrates superior performance over conventional methods: the characterization accuracy of lateral heterogeneity is improved while preserving preferable vertical resolution, high-quality sandstone reservoirs are precisely positioned. The remarkable consistency between the predicted results and well log data serves as a robust validation of the proposed method's reliability. The success of this study reveals the potential of this method to conduct predictions on fractured tight sandstone thin reservoirs characterized by high lateral heterogeneity in China as well as other areas with similar geological backgrounds.

本研究通过提出一种多相约束的地质统计反演方法，解决了预测断层裂缝改造致密砂岩储层中薄而非均质储层的长期挑战。传统的岩相约束地球统计反演方法无法解决高非均质性问题，也无法准确找到低泊松阻抗的“甜点”。为了克服这些局限性，我们将断裂体重新定义为不同的地质相，并将其与岩相约束整合到贝叶斯反演框架中，以便于变异函数的建立，以实现分区建模。曲率和结构熵是表征断裂体的首选工具，可以指导空间建模。在中国西部盆地的应用表明，该方法优于常规方法：在保持较好垂向分辨率的同时，提高了横向非均质性表征精度，精确定位了优质砂岩储层。预测结果与测井数据之间的显著一致性，有力地验证了该方法的可靠性。该研究的成功表明，该方法在中国以及其他具有类似地质背景的地区具有高度横向非均质性的裂缝性致密砂岩薄储层预测中的潜力。

{"title":"A multi-facies constrained geostatistical inversion for predicting fractured tight sandstone thin reservoirs: a case study in Western China","authors":"Sijia Li , Guangzhi Zhang , Zhenfeng Liu , Zhen Yang , Yuanyuan Tan , Jinghui Cui","doi":"10.1016/j.jappgeo.2025.106052","DOIUrl":"10.1016/j.jappgeo.2025.106052","url":null,"abstract":"<div><div>This study tackles the persistent challenge of predicting thin, heterogeneous reservoirs within fault-fracture-modified tight sandstone formations by proposing a multi-facies constrained geostatistical inversion method. Traditional lithofacies-constrained geostatistical inversion methods fail to resolve high heterogeneity and accurately find low Poisson impedance “sweet spots”. To overcome these limitations, we redefine fault-fracture bodies as distinct geological facies and integrate them with lithofacies constraints within a Bayesian inversion framework to facilitate the building of the variogram with the aim of realizing partition modeling. Curvature and structural entropy are preferred as the tools to characterize the fault-fracture bodies to guide spatial modeling. Application in a western China basin demonstrates superior performance over conventional methods: the characterization accuracy of lateral heterogeneity is improved while preserving preferable vertical resolution, high-quality sandstone reservoirs are precisely positioned. The remarkable consistency between the predicted results and well log data serves as a robust validation of the proposed method's reliability. The success of this study reveals the potential of this method to conduct predictions on fractured tight sandstone thin reservoirs characterized by high lateral heterogeneity in China as well as other areas with similar geological backgrounds.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"245 ","pages":"Article 106052"},"PeriodicalIF":2.1,"publicationDate":"2025-12-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145737521","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Object-oriented algorithm for the extraction of diffracted/scattered waves in the data domain 面向对象的数据域中绕射/散射波提取算法

IF 2.1 3区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Journal of Applied Geophysics

Pub Date : 2025-12-01 DOI: 10.1016/j.jappgeo.2025.106054

M. Protasov

Diffracted and scattered waves are used to construct diffraction images and enhance seismic interpretation. This paper presents an efficient object-oriented algorithm for detecting and enhancing scattered waves in the seismic data domain, enabling to solve challenging problem i.e. the generation of high-quality diffraction images based on scattered wave energy. Scattered waves are detected through coherent summation over sufficiently large apertures, while excess reflected events are suppressed. A key innovation of the algorithm is its method for determining the kinematic parameters of scattered waves, it calculates the hodographs of scattered waves within a pre-constructed velocity model for the specific target area where diffraction images are to be generated. For the data summation using stage two approaches are implemented: the first is based on weights derived from semblance analysis; the second employs an operator-oriented summation approach. Tests on simple methodological data and realistic data for a fractured model from Eastern Siberia demonstrate that the operator-oriented approach effectively enhances the amplitudes of scattered waves originating in the target domain while providing an accurate enough approximation of these waves. Application of the algorithm to real data shows that it produces diffraction images with a higher signal-to-noise ratio and image focusing compared to conventional methods.

衍射波和散射波用于构造衍射图像，增强地震解释。本文提出了一种高效的面向对象的地震数据域散射波检测与增强算法，解决了基于散射波能量生成高质量衍射图像的难题。散射波通过在足够大的孔径上的相干求和来检测，而多余的反射事件被抑制。该算法的一个关键创新是其确定散射波运动学参数的方法，它在一个预先构建的速度模型中计算特定目标区域的散射波的全息图，其中衍射图像将被生成。对于使用阶段的数据求和，实现了两种方法：第一种方法是基于从相似性分析中得出的权重；第二种方法采用面向运算符的求和方法。对东西伯利亚裂缝模型的简单方法数据和实际数据的测试表明，面向操作员的方法有效地增强了源自目标域的散射波的振幅，同时提供了这些波的足够精确的近似。对实际数据的应用表明，与传统方法相比，该算法产生的衍射图像具有更高的信噪比和图像聚焦。

{"title":"Object-oriented algorithm for the extraction of diffracted/scattered waves in the data domain","authors":"M. Protasov","doi":"10.1016/j.jappgeo.2025.106054","DOIUrl":"10.1016/j.jappgeo.2025.106054","url":null,"abstract":"<div><div>Diffracted and scattered waves are used to construct diffraction images and enhance seismic interpretation. This paper presents an efficient object-oriented algorithm for detecting and enhancing scattered waves in the seismic data domain, enabling to solve challenging problem i.e. the generation of high-quality diffraction images based on scattered wave energy. Scattered waves are detected through coherent summation over sufficiently large apertures, while excess reflected events are suppressed. A key innovation of the algorithm is its method for determining the kinematic parameters of scattered waves, it calculates the hodographs of scattered waves within a pre-constructed velocity model for the specific target area where diffraction images are to be generated. For the data summation using stage two approaches are implemented: the first is based on weights derived from semblance analysis; the second employs an operator-oriented summation approach. Tests on simple methodological data and realistic data for a fractured model from Eastern Siberia demonstrate that the operator-oriented approach effectively enhances the amplitudes of scattered waves originating in the target domain while providing an accurate enough approximation of these waves. Application of the algorithm to real data shows that it produces diffraction images with a higher signal-to-noise ratio and image focusing compared to conventional methods.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"245 ","pages":"Article 106054"},"PeriodicalIF":2.1,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145694548","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

3D facies modeling of a carbonate reservoir in the Santos Basin using machine learning 利用机器学习技术对Santos盆地碳酸盐岩储层进行三维相建模

IF 2.1 3区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Journal of Applied Geophysics

Pub Date : 2025-11-28 DOI: 10.1016/j.jappgeo.2025.106049

Henrique M. Santana, Emilson P. Leite

Three-dimensional facies modeling is essential for optimizing hydrocarbon exploration and production, particularly in complex carbonate reservoirs. Traditional classification methods rely on manually delineating homologous zones in scatter plots of elastic parameters, introducing subjectivity and limiting predictive accuracy. This study implements a machine learning-based workflow to classify petroelastic facies, flow units, and geomechanical units, minimizing arbitrary thresholding and improving interwell predictions. A supervised learning model was trained using well log data and validated through cross-validation, incorporating pseudo-curves derived from seismic inversion products (e.g., Zp, Zs) to enhance resolution. Data resampling techniques were applied to address class imbalance within the Barra Velha Formation. The resulting classification achieved 70–95 % accuracy and generated 3D models that highlight distinct facies distributions, with reservoir facies concentrated in the top and bottom intervals, while the middle interval exhibited predominantly non-reservoir facies, consistent with petrographic analyses. The facies distribution also revealed elongated reservoirs aligned with NE-SW structural highs, compartmentalized by faults of the same orientation. This improved understanding of facies distribution supports better placement of wells and reduced development risk. By eliminating the need for manual delineation of facies zones, this study presents a novel approach to three-dimensional facies classification from seismic inversion data, offering an objective and efficient alternative to more traditional methods used for reservoir characterization.

三维相建模对于优化油气勘探和生产至关重要，特别是在复杂的碳酸盐岩储层中。传统的分类方法依赖于在弹性参数散点图中手工圈定同源区域，引入了主观性，限制了预测精度。该研究采用了一种基于机器学习的工作流程，对岩石弹性相、流动单元和地质力学单元进行分类，最大限度地减少了任意阈值，提高了井间预测。使用测井数据训练监督学习模型，并通过交叉验证进行验证，结合地震反演产品（如Zp， Zs）的伪曲线来提高分辨率。数据重采样技术应用于解决Barra Velha组的类不平衡问题。最终的分类准确率达到70 - 95%，生成的3D模型突出了不同的相分布，储层相集中在顶部和底部层段，而中间层段主要显示非储层相，与岩石学分析一致。相展布还显示出沿北东—西向构造高点排列的细长型储层，被同一方位的断裂分隔。这种对相分布的更好理解有助于更好地布置井并降低开发风险。通过消除人工圈定相带的需要，本研究提出了一种利用地震反演数据进行三维相分类的新方法，为更传统的储层表征方法提供了一种客观、有效的替代方法。

{"title":"3D facies modeling of a carbonate reservoir in the Santos Basin using machine learning","authors":"Henrique M. Santana, Emilson P. Leite","doi":"10.1016/j.jappgeo.2025.106049","DOIUrl":"10.1016/j.jappgeo.2025.106049","url":null,"abstract":"<div><div>Three-dimensional facies modeling is essential for optimizing hydrocarbon exploration and production, particularly in complex carbonate reservoirs. Traditional classification methods rely on manually delineating homologous zones in scatter plots of elastic parameters, introducing subjectivity and limiting predictive accuracy. This study implements a machine learning-based workflow to classify petroelastic facies, flow units, and geomechanical units, minimizing arbitrary thresholding and improving interwell predictions. A supervised learning model was trained using well log data and validated through cross-validation, incorporating pseudo-curves derived from seismic inversion products (e.g., Zp, Zs) to enhance resolution. Data resampling techniques were applied to address class imbalance within the Barra Velha Formation. The resulting classification achieved 70–95 % accuracy and generated 3D models that highlight distinct facies distributions, with reservoir facies concentrated in the top and bottom intervals, while the middle interval exhibited predominantly non-reservoir facies, consistent with petrographic analyses. The facies distribution also revealed elongated reservoirs aligned with NE-SW structural highs, compartmentalized by faults of the same orientation. This improved understanding of facies distribution supports better placement of wells and reduced development risk. By eliminating the need for manual delineation of facies zones, this study presents a novel approach to three-dimensional facies classification from seismic inversion data, offering an objective and efficient alternative to more traditional methods used for reservoir characterization.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"245 ","pages":"Article 106049"},"PeriodicalIF":2.1,"publicationDate":"2025-11-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145694552","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

2.75-D Global joint inversion of gravity and magnetic anomalies with appraisal of model reconstruction uncertainty 2.75-D全球重磁异常联合反演及模型重建不确定性评价

IF 2.1 3区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Journal of Applied Geophysics

Pub Date : 2025-11-27 DOI: 10.1016/j.jappgeo.2025.106036

Yunus Levent Ekinci , Hanbing Ai , Çağlayan Balkaya , Arka Roy

Inversion procedures are fundamental tools for reconstructing causative sources of gravity and magnetic anomalies. While 2-D polygonal and 3-D polyhedral models can represent irregularly shaped bodies, a more flexible framework is needed to bridge the gap between computational simplicity and geometric realism. To address this, we propose a 2.75-D global joint inversion scheme based on the nature-inspired Hunger Games Search (HGS) metaheuristic algorithm. Synthetic tests involving modal and sensitivity analyses were carried out to identify potential difficulties and uncertainties in the considered problem, revealing that the global optimizer must efficiently balance the critical trade-off between global exploration and local exploitation. The proposed scheme was applied to synthetic anomalies, with and without noise, and benchmarked against the probabilistic Hamiltonian Monte Carlo algorithm. Two field datasets were then analyzed, and the results were interpreted considering existing geological and geophysical knowledge. Post-inversion analyses confirmed the reliability of the estimated models, while compact inversion and correlation imaging techniques supported the HGS outcomes. Notably, joint inversion consistently improved convergence and reduced estimation errors compared to individual inversions by exploiting the complementary sensitivities of gravity and magnetic data. The 2.75-D approach enhances geometric flexibility while maintaining a parsimonious parameterization. HGS is an efficient and robust optimizer for joint inversion of gravity and magnetic anomalies, capable of producing geologically plausible models with rapid convergence and minimal uncertainty.

反演程序是重建重磁异常成因的基本工具。虽然二维多边形和三维多面体模型可以表示不规则形状的物体，但需要一个更灵活的框架来弥合计算简单性和几何真实感之间的差距。为了解决这个问题，我们提出了一种基于自然启发的饥饿游戏搜索（HGS）元启发式算法的2.75维全球联合反演方案。采用模态分析和敏感性分析进行综合测试，以识别所考虑问题中的潜在困难和不确定性，揭示全局优化器必须有效地平衡全局勘探和局部开采之间的关键权衡。将该方法应用于有噪声和无噪声的综合异常，并以概率哈密顿蒙特卡罗算法为基准进行了测试。然后分析了两个现场数据集，并根据现有的地质和地球物理知识对结果进行了解释。反演后分析证实了估计模型的可靠性，而紧凑的反演和相关成像技术支持了HGS的结果。值得注意的是，通过利用重磁数据的互补灵敏度，联合反演与单独反演相比，始终提高了收敛性，减少了估计误差。2.75-D方法增强了几何灵活性，同时保持了简洁的参数化。HGS是一种高效、稳健的重磁联合反演优化器，能够生成地质上合理的模型，具有快速收敛和最小的不确定性。

{"title":"2.75-D Global joint inversion of gravity and magnetic anomalies with appraisal of model reconstruction uncertainty","authors":"Yunus Levent Ekinci , Hanbing Ai , Çağlayan Balkaya , Arka Roy","doi":"10.1016/j.jappgeo.2025.106036","DOIUrl":"10.1016/j.jappgeo.2025.106036","url":null,"abstract":"<div><div>Inversion procedures are fundamental tools for reconstructing causative sources of gravity and magnetic anomalies. While 2-D polygonal and 3-D polyhedral models can represent irregularly shaped bodies, a more flexible framework is needed to bridge the gap between computational simplicity and geometric realism. To address this, we propose a 2.75-D global joint inversion scheme based on the nature-inspired Hunger Games Search (HGS) metaheuristic algorithm. Synthetic tests involving modal and sensitivity analyses were carried out to identify potential difficulties and uncertainties in the considered problem, revealing that the global optimizer must efficiently balance the critical trade-off between global exploration and local exploitation. The proposed scheme was applied to synthetic anomalies, with and without noise, and benchmarked against the probabilistic Hamiltonian Monte Carlo algorithm. Two field datasets were then analyzed, and the results were interpreted considering existing geological and geophysical knowledge. Post-inversion analyses confirmed the reliability of the estimated models, while compact inversion and correlation imaging techniques supported the HGS outcomes. Notably, joint inversion consistently improved convergence and reduced estimation errors compared to individual inversions by exploiting the complementary sensitivities of gravity and magnetic data. The 2.75-D approach enhances geometric flexibility while maintaining a parsimonious parameterization. HGS is an efficient and robust optimizer for joint inversion of gravity and magnetic anomalies, capable of producing geologically plausible models with rapid convergence and minimal uncertainty.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"245 ","pages":"Article 106036"},"PeriodicalIF":2.1,"publicationDate":"2025-11-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145694550","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Unsupervised learning for forecasting deep water slope reservoirs in the Offshore Nile Delta: A novel classification model 无监督学习预测尼罗河三角洲深水斜坡储层：一种新的分类模型

IF 2.1 3区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Journal of Applied Geophysics

Pub Date : 2025-11-27 DOI: 10.1016/j.jappgeo.2025.106038

Ramy Eid , Mohamed El-Anbaawy , Adel El-Tehiwy

This study introduces an unsupervised neural network classification approach utilizing multiple seismic attributes to enhance reservoir characterization in the slope channel systems of the Simian Field, offshore Nile Delta. Given the complex nature of these reservoirs marked by significant heterogeneity and anisotropy affecting porosity and permeability, advanced analytical techniques are essential. Principal Component Analysis (PCA) was employed to reduce dimensionality and identify the most influential seismic attributes, including acoustic impedance, Root Mean Square (RMS) amplitude, and variance. The classification revealed two distinct seismic facies patterns, providing insights into subsurface heterogeneity. Furthermore, probability occurrence and zonation maps derived from the classification results enabled the identification of promising drilling targets in the eastern sector of the field. This integrated methodology offers a novel and efficient framework for reservoir forecasting in the geologically complex settings.

本研究引入了一种无监督神经网络分类方法，利用多个地震属性来增强尼罗河三角洲海上Simian油田斜坡河道系统的储层特征。考虑到这些储层的复杂性质，显著的非均质性和各向异性影响着孔隙度和渗透率，先进的分析技术是必不可少的。采用主成分分析（PCA）降维并识别最具影响的地震属性，包括声阻抗、均方根（RMS）振幅和方差。分类揭示了两种不同的地震相模式，为深入了解地下非均质性提供了依据。此外，根据分类结果得出的概率产状图和分区图能够确定该油田东部地区有希望的钻探目标。这种综合方法为复杂地质条件下的储层预测提供了一种新颖有效的框架。

引用次数: 0

A variational inference inversion approach to electrical resistivity tomography 电阻率层析成像的变分推理反演方法

IF 2.1 3区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Journal of Applied Geophysics

Pub Date : 2025-11-27 DOI: 10.1016/j.jappgeo.2025.106041

Sean Berti, Mattia Aleardi, Felipe Rincón, Eusebio Stucchi

We present a particle-based variational inference approach for Electrical Resistivity Tomography (ERT) using the Annealed Stein Variational Gradient Descent (A-SVGD) algorithm. This approach enables efficient approximation of the posterior distribution while addressing common limitations of standard SVGD—such as mode collapse and underestimation of variance—through a temperature-based annealing strategy. To reduce both the dimensionality of the inverse problem and computational cost, we perform the inversion in a compressed model space using the Discrete Cosine Transform (DCT), aligning the number of particles with the number of parameters—an ideal setting for a SVGD algorithm. We validate our method on both synthetic and real datasets, showing improved convergence and lower data misfit compared to standard SVGD and conventional deterministic inversion, along with effective uncertainty quantification comparable to those provided by a well-established gradient-based Markov Chain Monte Carlo algorithm. These results underscore the potential of A-SVGD, especially when combined with model-space compression, as a scalable and robust framework for nonlinear geophysical inversion.

我们提出了一种基于粒子的变分推理方法，用于电阻率层析成像（ERT），该方法使用退火Stein变分梯度下降（a - svgd）算法。这种方法能够有效地逼近后验分布，同时通过基于温度的退火策略解决标准svgd的常见限制，如模态崩溃和方差低估。为了降低反问题的维数和计算成本，我们使用离散余弦变换（DCT）在压缩模型空间中执行反演，将粒子数量与参数数量对齐——这是SVGD算法的理想设置。我们在合成数据集和真实数据集上验证了我们的方法，与标准SVGD和传统确定性反演相比，显示出更好的收敛性和更低的数据不拟合，以及与基于梯度的马尔可夫链蒙特卡罗算法提供的有效的不确定性量化相比较。这些结果强调了a - svgd的潜力，特别是当与模型空间压缩相结合时，作为非线性地球物理反演的可扩展和鲁棒框架。

{"title":"A variational inference inversion approach to electrical resistivity tomography","authors":"Sean Berti, Mattia Aleardi, Felipe Rincón, Eusebio Stucchi","doi":"10.1016/j.jappgeo.2025.106041","DOIUrl":"10.1016/j.jappgeo.2025.106041","url":null,"abstract":"<div><div>We present a particle-based variational inference approach for Electrical Resistivity Tomography (ERT) using the Annealed Stein Variational Gradient Descent (A-SVGD) algorithm. This approach enables efficient approximation of the posterior distribution while addressing common limitations of standard SVGD—such as mode collapse and underestimation of variance—through a temperature-based annealing strategy. To reduce both the dimensionality of the inverse problem and computational cost, we perform the inversion in a compressed model space using the Discrete Cosine Transform (DCT), aligning the number of particles with the number of parameters—an ideal setting for a SVGD algorithm. We validate our method on both synthetic and real datasets, showing improved convergence and lower data misfit compared to standard SVGD and conventional deterministic inversion, along with effective uncertainty quantification comparable to those provided by a well-established gradient-based Markov Chain Monte Carlo algorithm. These results underscore the potential of A-SVGD, especially when combined with model-space compression, as a scalable and robust framework for nonlinear geophysical inversion.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"245 ","pages":"Article 106041"},"PeriodicalIF":2.1,"publicationDate":"2025-11-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145694547","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Electrical resistivity inversion based on the minimum support gradient functional for water-bearing structure in tunnel 基于最小支撑梯度函数的隧道含水构造电阻率反演

IF 2.1 3区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Journal of Applied Geophysics

Pub Date : 2025-11-25 DOI: 10.1016/j.jappgeo.2025.106039

Zhaoyang Deng , Shuo Zhang , Pengyu Jing , Lichao Nie , Shilei Zhang

Electrical resistivity tomography plays a critical role in the investigation of tunnel water and mud inrush, but its inversion is an ill-posed and non-unique problem that requires regularization. Traditional smooth regularization ensures the overall stability of the inversion and provides continuous, smooth results. However, when applied to tunneling, especially in areas with clear geological boundaries (e.g. faults, karst cave), it is limited in its ability to produce sharp boundaries and blocky features. To address this limitation, this study incorporates a focusing regularization technique—the Minimum Support Gradient (MSG) functional—into a tunnel resistivity inversion framework. We first analyze the composition of several regularization terms and highlight the advantages of MSG in preserving sharp structural boundaries. Numerical simulations of typical water-bearing structures demonstrate that the proposed method effectively delineates the geometry of water-rich features and achieves improved resistivity recovery. In addition, the method was successfully applied to the Central Yunnan Water Diversion Project, where it effectively identified water-bearing structures in front of the tunnel face.

电阻率层析成像在隧道涌水涌泥研究中起着至关重要的作用，但其反演是一个病态且非唯一的问题，需要正则化。传统的平滑正则化保证了反演的整体稳定性，并提供连续、平滑的结果。但是，在隧道掘进中，特别是在地质边界明确的地区（如断层、溶洞），其产生清晰边界和块状特征的能力有限。为了解决这一限制，本研究将聚焦正则化技术-最小支持梯度（MSG）函数-纳入隧道电阻率反演框架。我们首先分析了几个正则化项的组成，并强调了MSG在保持清晰结构边界方面的优势。典型含水构造数值模拟结果表明，该方法能有效圈定富水构造的几何形态，提高了电阻率恢复。此外，该方法还成功应用于滇中引水工程，有效地识别了隧洞工作面前方含水构造。

{"title":"Electrical resistivity inversion based on the minimum support gradient functional for water-bearing structure in tunnel","authors":"Zhaoyang Deng , Shuo Zhang , Pengyu Jing , Lichao Nie , Shilei Zhang","doi":"10.1016/j.jappgeo.2025.106039","DOIUrl":"10.1016/j.jappgeo.2025.106039","url":null,"abstract":"<div><div>Electrical resistivity tomography plays a critical role in the investigation of tunnel water and mud inrush, but its inversion is an ill-posed and non-unique problem that requires regularization. Traditional smooth regularization ensures the overall stability of the inversion and provides continuous, smooth results. However, when applied to tunneling, especially in areas with clear geological boundaries (e.g. faults, karst cave), it is limited in its ability to produce sharp boundaries and blocky features. To address this limitation, this study incorporates a focusing regularization technique—the Minimum Support Gradient (MSG) functional—into a tunnel resistivity inversion framework. We first analyze the composition of several regularization terms and highlight the advantages of MSG in preserving sharp structural boundaries. Numerical simulations of typical water-bearing structures demonstrate that the proposed method effectively delineates the geometry of water-rich features and achieves improved resistivity recovery. In addition, the method was successfully applied to the Central Yunnan Water Diversion Project, where it effectively identified water-bearing structures in front of the tunnel face.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"245 ","pages":"Article 106039"},"PeriodicalIF":2.1,"publicationDate":"2025-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145625387","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Automated fault interpretation from gravity and magnetic data in covered areas using machine learning: A case study of the Eastern Tianshan orogenic belt 基于机器学习的覆盖区重磁数据自动断层解释——以东天山造山带为例

IF 2.1 3区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Journal of Applied Geophysics

Pub Date : 2025-11-25 DOI: 10.1016/j.jappgeo.2025.106040

Huaqing Yang , Fan Xiao , Hao Jia , Yongzhang Zhou , Shu Jiang

Machine learning has emerged as a powerful tool for integrating multi-source geophysical data, such as gravity and magnetic measurements, to enhance geological mapping. The Eastern Tianshan region, characterized by its Gobi-desert landscape, poses significant challenges for understanding regional structures, thereby complicating geological mapping and exploration targeting. To address these challenges, this study employs machine learning methods to fuse gravity and magnetic data for automated fault interpretation. Predictor maps were generated by applying a suite of boundary enhancement techniques, including the total horizontal derivative, vertical derivative, tilt angle, analytic signal, theta map, total horizontal derivative tilt angle, horizontal tilt angle, enhanced analytic signal tilt angle, and enhanced tilt angle methods, to aerial gravity and magnetic data. Two machine learning algorithms, namely Random Forest (RF) and Deep Neural Networks (DNNs), were then applied to these predictor maps to automatically identify the spatial distribution of concealed faults beneath the Gobi-desert covered layer. The grid search method was used to optimize the parameter combinations for both models. The results indicate that the RF model achieved an average accuracy of 86.20 %, outperforming the DNNs model, which achieved an average accuracy of 78.22 %. Based on this comparative analysis, the RF model was selected for fault interpretation in the Eastern Tianshan region. This approach provides critical data-driven insights into the spatial distribution of concealed faults, offering valuable guidance for subsequent geological mapping and exploration efforts.

机器学习已经成为整合多源地球物理数据（如重力和磁测量）以增强地质测绘的强大工具。东天山地区以戈壁沙漠地貌为特征，对区域构造的认识提出了重大挑战，从而使地质填图和勘探目标复杂化。为了解决这些挑战，本研究采用机器学习方法融合重力和磁数据，以实现自动故障解释。通过对航空重磁数据应用一系列边界增强技术，包括总水平导数、垂直导数、倾斜角、分析信号、θ图、总水平导数倾斜角、水平倾斜角、增强分析信号倾斜角和增强倾斜角方法，生成预测图。然后将随机森林（Random Forest， RF）和深度神经网络（Deep Neural Networks, dnn）两种机器学习算法应用于这些预测图，自动识别戈壁沙漠覆盖层下隐伏断层的空间分布。采用网格搜索方法对两种模型的参数组合进行优化。结果表明，RF模型的平均准确率为86.20%，优于DNNs模型的平均准确率78.22%。在此基础上，选择RF模型进行东天山地区断层解释。这种方法为隐伏断层的空间分布提供了关键的数据驱动见解，为后续的地质填图和勘探工作提供了有价值的指导。

{"title":"Automated fault interpretation from gravity and magnetic data in covered areas using machine learning: A case study of the Eastern Tianshan orogenic belt","authors":"Huaqing Yang , Fan Xiao , Hao Jia , Yongzhang Zhou , Shu Jiang","doi":"10.1016/j.jappgeo.2025.106040","DOIUrl":"10.1016/j.jappgeo.2025.106040","url":null,"abstract":"<div><div>Machine learning has emerged as a powerful tool for integrating multi-source geophysical data, such as gravity and magnetic measurements, to enhance geological mapping. The Eastern Tianshan region, characterized by its Gobi-desert landscape, poses significant challenges for understanding regional structures, thereby complicating geological mapping and exploration targeting. To address these challenges, this study employs machine learning methods to fuse gravity and magnetic data for automated fault interpretation. Predictor maps were generated by applying a suite of boundary enhancement techniques, including the total horizontal derivative, vertical derivative, tilt angle, analytic signal, theta map, total horizontal derivative tilt angle, horizontal tilt angle, enhanced analytic signal tilt angle, and enhanced tilt angle methods, to aerial gravity and magnetic data. Two machine learning algorithms, namely Random Forest (RF) and Deep Neural Networks (DNNs), were then applied to these predictor maps to automatically identify the spatial distribution of concealed faults beneath the Gobi-desert covered layer. The grid search method was used to optimize the parameter combinations for both models. The results indicate that the RF model achieved an average accuracy of 86.20 %, outperforming the DNNs model, which achieved an average accuracy of 78.22 %. Based on this comparative analysis, the RF model was selected for fault interpretation in the Eastern Tianshan region. This approach provides critical data-driven insights into the spatial distribution of concealed faults, offering valuable guidance for subsequent geological mapping and exploration efforts.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"245 ","pages":"Article 106040"},"PeriodicalIF":2.1,"publicationDate":"2025-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145625386","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0