Artificial Intelligence in Geosciences最新文献_第6页

Intelligent identification of fractures and holes in ultrasonic logging images based on the improved YOLOv8 model 基于改进YOLOv8模型的超声测井图像缝孔智能识别

IF 4.2

Artificial Intelligence in Geosciences

Pub Date : 2025-12-01 Epub Date: 2025-11-11 DOI: 10.1016/j.aiig.2025.100167

Jingyi Han , Xiumei Zhang , Yujuan Qi , Lin Liu

Aiming to address the demand for intelligent recognition of geological features in whole-wellbore ultrasonic images, this paper integrates the YOLOv8 model with the Convolution Block Attention Module (CBAM). It proposes an intelligent method for detecting fractures and holes, as well as segmenting whole-wellbore images. Firstly, we develop a dataset sample of effective reservoir sections by integrating logging data and conducting data augmentation on fracture and hole samples in ultrasonic logging images. A standardized process procedure for the generation of new samples and model training has been proposed effectively. Subsequently, the improved YOLOv8 model undergoes a process of training and validation. The results indicate that the model achieves average accuracies of 0.910 and 0.884 in target detection and image segmentation tasks, respectively. These findings demonstrate a notable performance improvement compared to the original model. Furthermore, a sliding window strategy is proposed to tackle the challenges of high computational demands and insufficient accuracy in the intelligent processing of full-well ultrasonic images. To manage overlapping regions within the sliding window, we employ the Non-Maximum Suppression (NMS) principle for effective processing. Finally, the model has been tested on actual logging images and demonstrates an enhanced capability to identify irregular fractures and holes, which significantly improves the efficiency of geological feature recognition in the whole-well section ultrasonic logging images.

针对全井筒超声图像中地质特征的智能识别需求，本文将YOLOv8模型与卷积块注意模块（Convolution Block Attention Module， CBAM）相结合。提出了一种智能的裂缝、井眼检测及全井图像分割方法。首先，对测井资料进行整合，并对超声测井图像中的裂缝和孔样进行数据增强，得到有效储层剖面数据集样本；提出了一种用于新样本生成和模型训练的标准化流程。随后，改进的YOLOv8模型经历了一个训练和验证过程。结果表明，该模型在目标检测和图像分割任务中的平均准确率分别为0.910和0.884。这些发现表明，与原始模型相比，性能有了显著提高。此外，针对全井超声图像智能处理中计算量大、精度不高的问题，提出了滑动窗口策略。为了管理滑动窗口内的重叠区域，我们采用非最大抑制（NMS）原则进行有效的处理。最后，在实际测井图像上进行了验证，结果表明该模型对不规则裂缝和井眼的识别能力增强，显著提高了全井段超声测井图像的地质特征识别效率。

{"title":"Intelligent identification of fractures and holes in ultrasonic logging images based on the improved YOLOv8 model","authors":"Jingyi Han , Xiumei Zhang , Yujuan Qi , Lin Liu","doi":"10.1016/j.aiig.2025.100167","DOIUrl":"10.1016/j.aiig.2025.100167","url":null,"abstract":"<div><div>Aiming to address the demand for intelligent recognition of geological features in whole-wellbore ultrasonic images, this paper integrates the YOLOv8 model with the Convolution Block Attention Module (CBAM). It proposes an intelligent method for detecting fractures and holes, as well as segmenting whole-wellbore images. Firstly, we develop a dataset sample of effective reservoir sections by integrating logging data and conducting data augmentation on fracture and hole samples in ultrasonic logging images. A standardized process procedure for the generation of new samples and model training has been proposed effectively. Subsequently, the improved YOLOv8 model undergoes a process of training and validation. The results indicate that the model achieves average accuracies of 0.910 and 0.884 in target detection and image segmentation tasks, respectively. These findings demonstrate a notable performance improvement compared to the original model. Furthermore, a sliding window strategy is proposed to tackle the challenges of high computational demands and insufficient accuracy in the intelligent processing of full-well ultrasonic images. To manage overlapping regions within the sliding window, we employ the Non-Maximum Suppression (NMS) principle for effective processing. Finally, the model has been tested on actual logging images and demonstrates an enhanced capability to identify irregular fractures and holes, which significantly improves the efficiency of geological feature recognition in the whole-well section ultrasonic logging images.</div></div>","PeriodicalId":100124,"journal":{"name":"Artificial Intelligence in Geosciences","volume":"6 2","pages":"Article 100167"},"PeriodicalIF":4.2,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145519713","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

AI-based approaches for wetland mapping and classification: A review of current practices and future perspectives 基于人工智能的湿地制图与分类方法：现状与展望

IF 4.2

Artificial Intelligence in Geosciences

Pub Date : 2025-12-01 Epub Date: 2025-11-05 DOI: 10.1016/j.aiig.2025.100165

Kai Cheng , Cong Zhang , Yaocheng Fan , Hongli Diao , Shibin Xia

Wetlands are critical ecosystems that provide essential ecological, hydrological, and socio-economic services, such as water purification, climate regulation, and biodiversity conservation. However, effective wetland management faces significant challenges, particularly in the analysis and classification of complex wetland environments. Traditional methods of wetland monitoring often suffer from limitations in spatial coverage, temporal resolution, and data processing efficiency. Recent advancements in artificial intelligence (AI), particularly machine learning and deep learning techniques, have been increasingly integrated with remote sensing technologies, offering a powerful solution to these challenges. AI has demonstrated significant potential in automating large-scale remote sensing data analysis, enabling the extraction of detailed spatial information, and enhancing the accuracy and efficiency of wetland mapping and classification. Bibliometric analysis indicates a growing body of research, with notable contributions from China and the United States, though regional disparities and a lack of diverse datasets remain key issues. Despite the success of AI in wetland monitoring, challenges persist in addressing environmental heterogeneity, mixed pixels, and data quality. This review synthesizes the current state of AI-based approaches in wetland mapping and classification, identifies trends and gaps, and outlines future research directions, emphasizing the need for interdisciplinary collaboration and integration of multi-source data to advance AI applications in wetland conservation.

湿地是重要的生态系统，提供必要的生态、水文和社会经济服务，如水净化、气候调节和生物多样性保护。然而，有效的湿地管理面临着重大挑战，特别是在复杂湿地环境的分析和分类方面。传统的湿地监测方法在空间覆盖、时间分辨率和数据处理效率等方面存在局限性。人工智能（AI）的最新进展，特别是机器学习和深度学习技术，已越来越多地与遥感技术相结合，为应对这些挑战提供了有力的解决方案。人工智能在自动化大规模遥感数据分析、提取详细空间信息以及提高湿地制图和分类的准确性和效率方面显示出巨大的潜力。文献计量分析表明，尽管区域差异和缺乏多样化的数据集仍然是关键问题，但中国和美国的研究成果正在不断增加。尽管人工智能在湿地监测方面取得了成功，但在解决环境异质性、混合像素和数据质量方面仍然存在挑战。本文综述了基于人工智能的湿地制图和分类方法的现状，指出了趋势和差距，并概述了未来的研究方向，强调需要跨学科合作和多源数据的整合来推进人工智能在湿地保护中的应用。

{"title":"AI-based approaches for wetland mapping and classification: A review of current practices and future perspectives","authors":"Kai Cheng , Cong Zhang , Yaocheng Fan , Hongli Diao , Shibin Xia","doi":"10.1016/j.aiig.2025.100165","DOIUrl":"10.1016/j.aiig.2025.100165","url":null,"abstract":"<div><div>Wetlands are critical ecosystems that provide essential ecological, hydrological, and socio-economic services, such as water purification, climate regulation, and biodiversity conservation. However, effective wetland management faces significant challenges, particularly in the analysis and classification of complex wetland environments. Traditional methods of wetland monitoring often suffer from limitations in spatial coverage, temporal resolution, and data processing efficiency. Recent advancements in artificial intelligence (AI), particularly machine learning and deep learning techniques, have been increasingly integrated with remote sensing technologies, offering a powerful solution to these challenges. AI has demonstrated significant potential in automating large-scale remote sensing data analysis, enabling the extraction of detailed spatial information, and enhancing the accuracy and efficiency of wetland mapping and classification. Bibliometric analysis indicates a growing body of research, with notable contributions from China and the United States, though regional disparities and a lack of diverse datasets remain key issues. Despite the success of AI in wetland monitoring, challenges persist in addressing environmental heterogeneity, mixed pixels, and data quality. This review synthesizes the current state of AI-based approaches in wetland mapping and classification, identifies trends and gaps, and outlines future research directions, emphasizing the need for interdisciplinary collaboration and integration of multi-source data to advance AI applications in wetland conservation.</div></div>","PeriodicalId":100124,"journal":{"name":"Artificial Intelligence in Geosciences","volume":"6 2","pages":"Article 100165"},"PeriodicalIF":4.2,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145465152","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Machine-learning seismic damage assessment model for building structures 基于机器学习的建筑结构震害评估模型

IF 4.2

Artificial Intelligence in Geosciences

Pub Date : 2025-12-01 Epub Date: 2025-09-13 DOI: 10.1016/j.aiig.2025.100155

Fatma Zohra Belhadj , Ahmed Fouad Belhadj , Mohamed Chabaat

Buildings in seismic-prone regions are highly vulnerable to structural damage, necessitating meticulous Seismic Damage Assessment (SDA) for accurate design and mitigation strategies. The intricate nature of Seismic Damage Assessment (SDA) poses challenges, particularly when employing Finite Element Analysis (FE) for individual structures, as simulation techniques are time-intensive due to the inherent complexity of the models. Computational methods combining Soil-Structure Interaction (SSI) for earthquake damage assessment further compound the challenge, requiring substantial computational efforts to construct a comprehensive database for area-based prediction models. This study introduces such challenges via a novel Artificial Neural Network (ANN) approaches-based model as an alternative for prompt building Seismic Damage Assessment evaluation. The proposed ANN model leverages three key inputs—seismic, building, and soil parameters—incorporating a multi-step analysis process to generate seismic responses with soil-structure interaction. The findings underscore the remarkable accuracy of the SDA-Net model, positioning it as an effective predictive tool and rapid decision support system for structures affected by SSI impacts. This innovative approach not only serves as a proactive pre-disaster management tool for assessing potential damage but also emerges as a practical asset for ensuring the safety and durability of structures in the face of natural disasters. The study's contribution lies in its potential application as a valuable tool in structural engineering, aligning with the objectives and scope of the Research Journal of The Institution of Structural Engineers.

地震易发地区的建筑物极易受到结构破坏，因此需要进行细致的地震损害评估（SDA），以实现准确的设计和减灾策略。地震损伤评估（SDA）的复杂性带来了挑战，特别是当对单个结构使用有限元分析（FE）时，由于模型固有的复杂性，模拟技术需要耗费大量时间。结合土-结构相互作用（SSI）进行震害评估的计算方法进一步加剧了这一挑战，需要大量的计算工作来构建基于区域的预测模型的综合数据库。本研究通过一种新颖的基于人工神经网络（ANN）方法的模型来引入这些挑战，作为快速评估建筑物震害的替代方法。提出的人工神经网络模型利用三个关键输入-地震，建筑和土壤参数-结合多步骤分析过程来生成具有土壤-结构相互作用的地震响应。研究结果强调了SDA-Net模型的显著准确性，将其定位为受SSI影响的结构的有效预测工具和快速决策支持系统。这种创新的方法不仅可以作为评估潜在损害的一种主动的灾前管理工具，而且还可以作为确保建筑物在面对自然灾害时的安全性和耐久性的实用资产。该研究的贡献在于其作为结构工程中有价值的工具的潜在应用，与结构工程师学会研究期刊的目标和范围一致。

{"title":"Machine-learning seismic damage assessment model for building structures","authors":"Fatma Zohra Belhadj , Ahmed Fouad Belhadj , Mohamed Chabaat","doi":"10.1016/j.aiig.2025.100155","DOIUrl":"10.1016/j.aiig.2025.100155","url":null,"abstract":"<div><div>Buildings in seismic-prone regions are highly vulnerable to structural damage, necessitating meticulous Seismic Damage Assessment (SDA) for accurate design and mitigation strategies. The intricate nature of Seismic Damage Assessment (SDA) poses challenges, particularly when employing Finite Element Analysis (FE) for individual structures, as simulation techniques are time-intensive due to the inherent complexity of the models. Computational methods combining Soil-Structure Interaction (SSI) for earthquake damage assessment further compound the challenge, requiring substantial computational efforts to construct a comprehensive database for area-based prediction models. This study introduces such challenges via a novel Artificial Neural Network (ANN) approaches-based model as an alternative for prompt building Seismic Damage Assessment evaluation. The proposed ANN model leverages three key inputs—seismic, building, and soil parameters—incorporating a multi-step analysis process to generate seismic responses with soil-structure interaction. The findings underscore the remarkable accuracy of the SDA-Net model, positioning it as an effective predictive tool and rapid decision support system for structures affected by SSI impacts. This innovative approach not only serves as a proactive pre-disaster management tool for assessing potential damage but also emerges as a practical asset for ensuring the safety and durability of structures in the face of natural disasters. The study's contribution lies in its potential application as a valuable tool in structural engineering, aligning with the objectives and scope of the Research Journal of The Institution of Structural Engineers.</div></div>","PeriodicalId":100124,"journal":{"name":"Artificial Intelligence in Geosciences","volume":"6 2","pages":"Article 100155"},"PeriodicalIF":4.2,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145157171","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Deciphering influential features in the seismic catalog for large earthquake occurrence from a machine learning perspective 从机器学习的角度解读大地震发生的地震目录中的影响特征

IF 4.2

Artificial Intelligence in Geosciences

Pub Date : 2025-12-01 Epub Date: 2025-10-26 DOI: 10.1016/j.aiig.2025.100161

Jinsu Jang , Byung-Dal So , David A. Yuen , Sung-Joon Chang

The spatiotemporal distribution and magnitude of seismicity collected over decades are crucial for understanding the stress interactions underlying large earthquakes. In this study, machine learning (ML) explainers identify and rank the features that distinguish Large Earthquake Occurrence (LEO) from non-LEO spatiotemporal windows. Seventy-eight statistics related to time, latitude, longitude, depth, and magnitude were extracted from the earthquake catalog (Global Centroid Moment Tensor) to produce 202,706 spatiotemporally discretized windows. ML explainers trained on these windows revealed the maximum magnitude (M_max) as the most influential feature. Classification performance improved when the maximum inter-event time, the average inter-event time, and the minimum ratio of focal depth to magnitude were jointly trained with M_max. The top five features showed weak-to-moderate correlations, providing complementary information to the ML explainers. Our explainable ML framework can be extended to different earthquake catalogs, including those with focal mechanisms and small-magnitude events.

几十年来收集的地震活动的时空分布和震级对于理解大地震背后的应力相互作用至关重要。在本研究中，机器学习（ML）解释器识别并排序区分大地震发生（LEO）和非LEO时空窗口的特征。从地震目录（全球质心矩张量）中提取78个与时间、纬度、经度、深度和震级相关的统计量，产生202,706个时空离散窗口。在这些窗口上训练的ML解释器显示最大幅度（Mmax）是最具影响力的特征。当最大事件间隔时间、平均事件间隔时间和最小震级比与Mmax联合训练时，分类性能得到提高。前五个特征显示出弱到中等的相关性，为ML解释器提供了补充信息。我们的可解释的ML框架可以扩展到不同的地震目录，包括那些具有震源机制和小震级事件。

{"title":"Deciphering influential features in the seismic catalog for large earthquake occurrence from a machine learning perspective","authors":"Jinsu Jang , Byung-Dal So , David A. Yuen , Sung-Joon Chang","doi":"10.1016/j.aiig.2025.100161","DOIUrl":"10.1016/j.aiig.2025.100161","url":null,"abstract":"<div><div>The spatiotemporal distribution and magnitude of seismicity collected over decades are crucial for understanding the stress interactions underlying large earthquakes. In this study, machine learning (ML) explainers identify and rank the features that distinguish Large Earthquake Occurrence (LEO) from non-LEO spatiotemporal windows. Seventy-eight statistics related to time, latitude, longitude, depth, and magnitude were extracted from the earthquake catalog (Global Centroid Moment Tensor) to produce 202,706 spatiotemporally discretized windows. ML explainers trained on these windows revealed the maximum magnitude (<em>M</em><sub><em>max</em></sub>) as the most influential feature. Classification performance improved when the maximum inter-event time, the average inter-event time, and the minimum ratio of focal depth to magnitude were jointly trained with <em>M</em><sub><em>max</em></sub>. The top five features showed weak-to-moderate correlations, providing complementary information to the ML explainers. Our explainable ML framework can be extended to different earthquake catalogs, including those with focal mechanisms and small-magnitude events.</div></div>","PeriodicalId":100124,"journal":{"name":"Artificial Intelligence in Geosciences","volume":"6 2","pages":"Article 100161"},"PeriodicalIF":4.2,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145465151","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Loosening rocks detection at Draa Sfar deep underground mine in Morocco using infrared thermal imaging and image segmentation models 基于红外热成像和图像分割模型的摩洛哥Draa Sfar深部地下矿松动岩探测

Artificial Intelligence in Geosciences

Pub Date : 2025-06-01 Epub Date: 2025-01-27 DOI: 10.1016/j.aiig.2025.100106

Kaoutar Clero , Said Ed-Diny , Mohammed Achalhi , Mouhamed Cherkaoui , Imad El Harraki , Sanaa El Fkihi , Intissar Benzakour , Tarik Soror , Said Rziki , Hamd Ait Abdelali , Hicham Tagemouati , François Bourzeix

Rockfalls are among the frequent hazards in underground mines worldwide, requiring effective methods for detecting unstable rock blocks to ensure miners' and equipment's safety. This study proposes a novel approach for identifying potential rockfall zones using infrared thermal imaging and image segmentation techniques. Infrared images of rock blocks were captured at the Draa Sfar deep underground mine in Morocco using the FLUKE TI401 PRO thermal camera. Two segmentation methods were applied to locate the potential unstable areas: the classical thresholding and the K-means clustering model. The results show that while thresholding allows a binary distinction between stable and unstable areas, K-means clustering is more accurate, especially when using multiple clusters to show different risk levels. The close match between the clustering masks of unstable blocks and their corresponding visible light images further validated this. The findings confirm that thermal image segmentation can serve as an alternative method for predicting rockfalls and monitoring geotechnical issues in underground mines. Underground operators worldwide can apply this approach to monitor rock mass stability. However, further research is recommended to enhance these results, particularly through deep learning-based segmentation and object detection models.

岩崩是世界范围内地下矿山频发的灾害之一，为保证矿工和设备的安全，需要有效的检测不稳定岩块的方法。本研究提出了一种利用红外热成像和图像分割技术识别潜在岩崩带的新方法。利用FLUKE TI401 PRO热像仪在摩洛哥Draa Sfar深部地下矿山拍摄了岩石块的红外图像。采用经典阈值分割和k均值聚类两种分割方法定位潜在的不稳定区域。结果表明，虽然阈值允许对稳定和不稳定区域进行二元区分，但K-means聚类更准确，特别是当使用多个聚类来显示不同的风险水平时。不稳定块的聚类掩模与其对应的可见光图像的紧密匹配进一步验证了这一点。研究结果证实，热图像分割可以作为地下矿山岩崩预测和岩土工程问题监测的替代方法。世界各地的地下运营商都可以应用这种方法来监测岩体的稳定性。然而，建议进一步研究以增强这些结果，特别是通过基于深度学习的分割和目标检测模型。

{"title":"Loosening rocks detection at Draa Sfar deep underground mine in Morocco using infrared thermal imaging and image segmentation models","authors":"Kaoutar Clero , Said Ed-Diny , Mohammed Achalhi , Mouhamed Cherkaoui , Imad El Harraki , Sanaa El Fkihi , Intissar Benzakour , Tarik Soror , Said Rziki , Hamd Ait Abdelali , Hicham Tagemouati , François Bourzeix","doi":"10.1016/j.aiig.2025.100106","DOIUrl":"10.1016/j.aiig.2025.100106","url":null,"abstract":"<div><div>Rockfalls are among the frequent hazards in underground mines worldwide, requiring effective methods for detecting unstable rock blocks to ensure miners' and equipment's safety. This study proposes a novel approach for identifying potential rockfall zones using infrared thermal imaging and image segmentation techniques. Infrared images of rock blocks were captured at the Draa Sfar deep underground mine in Morocco using the FLUKE TI401 PRO thermal camera. Two segmentation methods were applied to locate the potential unstable areas: the classical thresholding and the K-means clustering model. The results show that while thresholding allows a binary distinction between stable and unstable areas, K-means clustering is more accurate, especially when using multiple clusters to show different risk levels. The close match between the clustering masks of unstable blocks and their corresponding visible light images further validated this. The findings confirm that thermal image segmentation can serve as an alternative method for predicting rockfalls and monitoring geotechnical issues in underground mines. Underground operators worldwide can apply this approach to monitor rock mass stability. However, further research is recommended to enhance these results, particularly through deep learning-based segmentation and object detection models.</div></div>","PeriodicalId":100124,"journal":{"name":"Artificial Intelligence in Geosciences","volume":"6 1","pages":"Article 100106"},"PeriodicalIF":0.0,"publicationDate":"2025-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143143632","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Fast 2D forward modeling of electromagnetic propagation well logs using finite element method and data-driven deep learning 利用有限元方法和数据驱动的深度学习快速二维电磁传播测井曲线正演建模

Artificial Intelligence in Geosciences

Pub Date : 2025-06-01 Epub Date: 2025-03-28 DOI: 10.1016/j.aiig.2025.100112

A.M. Petrov, A.R. Leonenko, K.N. Danilovskiy, O.V. Nechaev

We propose a novel workflow for fast forward modeling of well logs in axially symmetric 2D models of the near-wellbore environment. The approach integrates the finite element method with deep residual neural networks to achieve exceptional computational efficiency and accuracy. The workflow is demonstrated through the modeling of wireline electromagnetic propagation resistivity logs, where the measured responses exhibit a highly nonlinear relationship with formation properties. The motivation for this research is the need for advanced modeling algorithms that are fast enough for use in modern quantitative interpretation tools, where thousands of simulations may be required in iterative inversion processes. The proposed algorithm achieves a remarkable enhancement in performance, being up to 3000 times faster than the finite element method alone when utilizing a GPU. While still ensuring high accuracy, this makes it well-suited for practical applications when reliable payzone assessment is needed in complex environmental scenarios. Furthermore, the algorithm's efficiency positions it as a promising tool for stochastic Bayesian inversion, facilitating reliable uncertainty quantification in subsurface property estimation.

我们提出了一种新的工作流程，用于在近井筒环境的轴对称二维模型中对测井曲线进行快速正演建模。该方法将有限元法与深度残差神经网络相结合，实现了极高的计算效率和精度。工作流程通过有线电磁传播电阻率测井建模进行了演示，测得的响应与地层属性呈现高度非线性关系。这项研究的动机是现代定量解释工具需要足够快的先进建模算法，在迭代反演过程中可能需要进行数千次模拟。所提出的算法性能显著提高，在使用 GPU 时比单独使用有限元方法快 3000 倍。在确保高精度的同时，该算法非常适合在复杂环境场景中需要进行可靠的薪区评估的实际应用。此外，该算法的高效性使其成为随机贝叶斯反演的理想工具，有助于在地下属性评估中进行可靠的不确定性量化。

{"title":"Fast 2D forward modeling of electromagnetic propagation well logs using finite element method and data-driven deep learning","authors":"A.M. Petrov, A.R. Leonenko, K.N. Danilovskiy, O.V. Nechaev","doi":"10.1016/j.aiig.2025.100112","DOIUrl":"10.1016/j.aiig.2025.100112","url":null,"abstract":"<div><div>We propose a novel workflow for fast forward modeling of well logs in axially symmetric 2D models of the near-wellbore environment. The approach integrates the finite element method with deep residual neural networks to achieve exceptional computational efficiency and accuracy. The workflow is demonstrated through the modeling of wireline electromagnetic propagation resistivity logs, where the measured responses exhibit a highly nonlinear relationship with formation properties. The motivation for this research is the need for advanced modeling algorithms that are fast enough for use in modern quantitative interpretation tools, where thousands of simulations may be required in iterative inversion processes. The proposed algorithm achieves a remarkable enhancement in performance, being up to 3000 times faster than the finite element method alone when utilizing a GPU. While still ensuring high accuracy, this makes it well-suited for practical applications when reliable payzone assessment is needed in complex environmental scenarios. Furthermore, the algorithm's efficiency positions it as a promising tool for stochastic Bayesian inversion, facilitating reliable uncertainty quantification in subsurface property estimation.</div></div>","PeriodicalId":100124,"journal":{"name":"Artificial Intelligence in Geosciences","volume":"6 1","pages":"Article 100112"},"PeriodicalIF":0.0,"publicationDate":"2025-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143776649","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

An intelligent recognition method of deep shale gas reservoir laminaset based on laminaset clustering and R-L-M algorithm 基于层状集聚类和R-L-M算法的深层页岩气藏层状集智能识别方法

Artificial Intelligence in Geosciences

Pub Date : 2025-06-01 Epub Date: 2025-04-07 DOI: 10.1016/j.aiig.2025.100113

Yu Zeng , Fuqiang Lai , Haijie Zhang , Yi Jiang , Junwei Pu , Tongtong Luo , Xiaoxia Zhao

Lamina structures, as typical sedimentary features in shale formations, determine both the quality of shale reservoirs and fracturing effects. In this study, through electric imaging logging, based on core scanning photos, thin sections, and other data from the Wufeng-Longmaxi Formation shale reservoirs in the western Sichuan Block, the characteristics and classification scheme of deep shale gas reservoir laminaset were clarified. In addition, with core scale electrical images, the electrical imaging logging response characteristics of different types of laminaset were identified. Based on electrical imaging logging images, a laminaset clustering algorithm was designed to segment the laminaset and then Levenberg-Marquardt (L-M) algorithm was improved by introducing a random forest to obtain the R-L-M algorithm, which was used to extract key parameters of laminaset such as attitude, type, density, and thickness. The average accuracy, recall rate, and F1 score of laminaset recognition results of this algorithm were 14.82 % higher than those of a well-known international commercial software (T). This method was used to evaluate the Longmaxi Formation shale gas reservoir in the western Sichuan Block. The development density of clay-siliceous (organic-lean) laminaset from the Longyi 1–4 small layer to the lower Wufeng Formation firstly decreased and then increased and the minimum value was found in Longyi 1-1 small layer. In contrast, the development density of siliceous-clay laminaset (organic-rich) first increased and then gradually decreased and the maximum value was found in Longyi 1-1 small layer. The clay-siliceous laminaset (organic matters-contained) and the calcareous-clay laminaset (organic matters-contained) showed a stable developmental trend.

层状构造作为页岩层的典型沉积特征，决定着页岩储层的质量和压裂效果。本研究通过电成像测井，基于四川西部区块五峰-龙马溪地层页岩储层的岩心扫描照片、薄切片等资料，明确了深层页岩气储层层理的特征和分类方案。此外，通过岩心尺度电图像，确定了不同类型层系的电成像测井响应特征。基于电成像测井图像，设计了层丛聚类算法对层丛进行划分，然后通过引入随机森林对 Levenberg-Marquardt 算法（L-M）进行改进，得到 R-L-M 算法，用于提取层丛的姿态、类型、密度和厚度等关键参数。该算法的层集识别结果的平均准确率、召回率和 F1 分数比国际知名商业软件（T）高出 14.82%。该方法被用于评估四川西部区块龙马溪地层页岩气藏。结果表明，龙马溪地层页岩气储层的粘土-硅质（有机-鳞片）层状发育密度从龙一1-4小层到五峰地层下部先减小后增大，最小值出现在龙一1-1小层。而硅质粘土层组（富含有机质）的发育密度先增大后逐渐减小，最大值出现在龙宜 1-1 小层。粘土-硅质层组（含有机质）和石灰质-粘土层组（含有机质）呈稳定的发育趋势。

{"title":"An intelligent recognition method of deep shale gas reservoir laminaset based on laminaset clustering and R-L-M algorithm","authors":"Yu Zeng , Fuqiang Lai , Haijie Zhang , Yi Jiang , Junwei Pu , Tongtong Luo , Xiaoxia Zhao","doi":"10.1016/j.aiig.2025.100113","DOIUrl":"10.1016/j.aiig.2025.100113","url":null,"abstract":"<div><div>Lamina structures, as typical sedimentary features in shale formations, determine both the quality of shale reservoirs and fracturing effects. In this study, through electric imaging logging, based on core scanning photos, thin sections, and other data from the Wufeng-Longmaxi Formation shale reservoirs in the western Sichuan Block, the characteristics and classification scheme of deep shale gas reservoir laminaset were clarified. In addition, with core scale electrical images, the electrical imaging logging response characteristics of different types of laminaset were identified. Based on electrical imaging logging images, a laminaset clustering algorithm was designed to segment the laminaset and then Levenberg-Marquardt (L-M) algorithm was improved by introducing a random forest to obtain the R-L-M algorithm, which was used to extract key parameters of laminaset such as attitude, type, density, and thickness. The average accuracy, recall rate, and F1 score of laminaset recognition results of this algorithm were 14.82 % higher than those of a well-known international commercial software (T). This method was used to evaluate the Longmaxi Formation shale gas reservoir in the western Sichuan Block. The development density of clay-siliceous (organic-lean) laminaset from the Longyi 1–4 small layer to the lower Wufeng Formation firstly decreased and then increased and the minimum value was found in Longyi 1-1 small layer. In contrast, the development density of siliceous-clay laminaset (organic-rich) first increased and then gradually decreased and the maximum value was found in Longyi 1-1 small layer. The clay-siliceous laminaset (organic matters-contained) and the calcareous-clay laminaset (organic matters-contained) showed a stable developmental trend.</div></div>","PeriodicalId":100124,"journal":{"name":"Artificial Intelligence in Geosciences","volume":"6 1","pages":"Article 100113"},"PeriodicalIF":0.0,"publicationDate":"2025-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143828417","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Digital core reconstruction of tight carbonate rocks based on SliceGAN 基于SliceGAN的致密碳酸盐岩数字岩心重建

Artificial Intelligence in Geosciences

Pub Date : 2025-06-01 Epub Date: 2025-04-22 DOI: 10.1016/j.aiig.2025.100116

Ying Zhou , Taiping Zhao , Wenjing Zhang , Feiqi Teng , Xin Nie

The pore structures of the Majiagou Formation in the Ordos Basin are complex, featuring micro- and nano-scale intra-crystalline and inter-crystalline pores that significantly impact hydrocarbon storage and flow. Precisely characterizing the rock internal structures is crucial for reservoir exploration and development. However, it is difficult to accurately characterize the pore structure of rock using traditional imaging methods to meet the simulation requirements. In this context, this study focuses on high-resolution 3D digital core reconstruction using the SliceGAN model. Specifically, the Modular Automated Processing System (MAPS) image and Quantitative Evaluation of Minerals by Scanning Electron Microscopy (QEMSCAN) image were combined to divide MAPS into three categories: pore, dolomite, and calcite. Then, through the SliceGAN algorithm, the 3D digital core was reconstructed. To evaluate the reconstruction, the auto-correlation function, two-point probability function, porosity, mineral content, and specific surface area were employed. The results show that the SliceGAN can effectively capture the micro-features in the core, and the internal structure of the generated core was consistent with that of the original core. This study provided a new sight for reconstructing cores with complex pore structures and strong heterogeneity and innovatively supports tight carbonate reservoir characterization and evaluation.

鄂尔多斯盆地马家沟组孔隙结构复杂，具有微纳米级的晶内孔和晶间孔，对油气的储集和流动具有重要影响。准确表征岩石内部构造对储层勘探开发至关重要。然而，传统的成像方法难以准确表征岩石孔隙结构，难以满足模拟要求。在此背景下，本研究的重点是使用SliceGAN模型进行高分辨率3D数字岩心重建。具体而言，将模块化自动化处理系统（MAPS）图像与扫描电子显微镜矿物定量评价（QEMSCAN）图像相结合，将MAPS分为孔隙、白云石和方解石三类。然后，通过SliceGAN算法对三维数字核进行重构。利用自相关函数、两点概率函数、孔隙度、矿物含量和比表面积对重建结果进行评价。结果表明，SliceGAN能够有效捕获岩心内部的微观特征，生成的岩心内部结构与原始岩心基本一致。该研究为孔隙结构复杂、非均质性强的岩心重建提供了新的思路，为致密碳酸盐岩储层的表征和评价提供了创新的支持。

{"title":"Digital core reconstruction of tight carbonate rocks based on SliceGAN","authors":"Ying Zhou , Taiping Zhao , Wenjing Zhang , Feiqi Teng , Xin Nie","doi":"10.1016/j.aiig.2025.100116","DOIUrl":"10.1016/j.aiig.2025.100116","url":null,"abstract":"<div><div>The pore structures of the Majiagou Formation in the Ordos Basin are complex, featuring micro- and nano-scale intra-crystalline and inter-crystalline pores that significantly impact hydrocarbon storage and flow. Precisely characterizing the rock internal structures is crucial for reservoir exploration and development. However, it is difficult to accurately characterize the pore structure of rock using traditional imaging methods to meet the simulation requirements. In this context, this study focuses on high-resolution 3D digital core reconstruction using the SliceGAN model. Specifically, the Modular Automated Processing System (MAPS) image and Quantitative Evaluation of Minerals by Scanning Electron Microscopy (QEMSCAN) image were combined to divide MAPS into three categories: pore, dolomite, and calcite. Then, through the SliceGAN algorithm, the 3D digital core was reconstructed. To evaluate the reconstruction, the auto-correlation function, two-point probability function, porosity, mineral content, and specific surface area were employed. The results show that the SliceGAN can effectively capture the micro-features in the core, and the internal structure of the generated core was consistent with that of the original core. This study provided a new sight for reconstructing cores with complex pore structures and strong heterogeneity and innovatively supports tight carbonate reservoir characterization and evaluation.</div></div>","PeriodicalId":100124,"journal":{"name":"Artificial Intelligence in Geosciences","volume":"6 1","pages":"Article 100116"},"PeriodicalIF":0.0,"publicationDate":"2025-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143864930","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A staged deep learning approach to spatial refinement in 3D temporal atmospheric transport 基于阶段深度学习的三维时空大气传输空间细化方法

Artificial Intelligence in Geosciences

Pub Date : 2025-06-01 Epub Date: 2025-05-29 DOI: 10.1016/j.aiig.2025.100120

M. Giselle Fernández-Godino , Wai Tong Chung , Akshay A. Gowardhan , Matthias Ihme , Qingkai Kong , Donald D. Lucas , Stephen C. Myers

High-resolution spatiotemporal simulations effectively capture the complexities of atmospheric plume dispersion in complex terrain. However, their high computational cost makes them impractical for applications requiring rapid responses or iterative processes, such as optimization, uncertainty quantification, or inverse modeling. To address this challenge, this work introduces the Dual-Stage Temporal Three-dimensional UNet Super-resolution (DST3D-UNet-SR) model, a highly efficient deep learning model for plume dispersion predictions. DST3D-UNet-SR is composed of two sequential modules: the temporal module (TM), which predicts the transient evolution of a plume in complex terrain from low-resolution temporal data, and the spatial refinement module (SRM), which subsequently enhances the spatial resolution of the TM predictions. We train DST3D-UNet-SR using a comprehensive dataset derived from high-resolution large eddy simulations (LES) of plume transport. We propose the DST3D-UNet-SR model to significantly accelerate LES of three-dimensional (3D) plume dispersion by three orders of magnitude. Additionally, the model demonstrates the ability to dynamically adapt to evolving conditions through the incorporation of new observational data, substantially improving prediction accuracy in high-concentration regions near the source.

高分辨率时空模拟有效地捕捉了复杂地形下大气羽散的复杂性。然而，它们的高计算成本使得它们不适合需要快速响应或迭代过程的应用，例如优化、不确定性量化或逆建模。为了应对这一挑战，本工作引入了双阶段时间三维UNet超分辨率（DST3D-UNet-SR）模型，这是一种用于羽散预测的高效深度学习模型。DST3D-UNet-SR由两个连续模块组成：时间模块（TM）和空间细化模块（SRM），前者从低分辨率时间数据预测复杂地形中羽流的瞬态演变，后者提高了TM预测的空间分辨率。我们使用来自羽流传输的高分辨率大涡模拟（LES）的综合数据集来训练DST3D-UNet-SR。我们提出了DST3D-UNet-SR模型，将三维羽散的LES显著加速了三个数量级。此外，该模型通过纳入新的观测数据，显示出动态适应不断变化的条件的能力，大大提高了源附近高浓度区域的预测精度。

{"title":"A staged deep learning approach to spatial refinement in 3D temporal atmospheric transport","authors":"M. Giselle Fernández-Godino , Wai Tong Chung , Akshay A. Gowardhan , Matthias Ihme , Qingkai Kong , Donald D. Lucas , Stephen C. Myers","doi":"10.1016/j.aiig.2025.100120","DOIUrl":"10.1016/j.aiig.2025.100120","url":null,"abstract":"<div><div>High-resolution spatiotemporal simulations effectively capture the complexities of atmospheric plume dispersion in complex terrain. However, their high computational cost makes them impractical for applications requiring rapid responses or iterative processes, such as optimization, uncertainty quantification, or inverse modeling. To address this challenge, this work introduces the Dual-Stage Temporal Three-dimensional UNet Super-resolution (DST3D-UNet-SR) model, a highly efficient deep learning model for plume dispersion predictions. DST3D-UNet-SR is composed of two sequential modules: the temporal module (TM), which predicts the transient evolution of a plume in complex terrain from low-resolution temporal data, and the spatial refinement module (SRM), which subsequently enhances the spatial resolution of the TM predictions. We train DST3D-UNet-SR using a comprehensive dataset derived from high-resolution large eddy simulations (LES) of plume transport. We propose the DST3D-UNet-SR model to significantly accelerate LES of three-dimensional (3D) plume dispersion by three orders of magnitude. Additionally, the model demonstrates the ability to dynamically adapt to evolving conditions through the incorporation of new observational data, substantially improving prediction accuracy in high-concentration regions near the source.</div></div>","PeriodicalId":100124,"journal":{"name":"Artificial Intelligence in Geosciences","volume":"6 1","pages":"Article 100120"},"PeriodicalIF":0.0,"publicationDate":"2025-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144185115","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Variogram modelling optimisation using genetic algorithm and machine learning linear regression: application for Sequential Gaussian Simulations mapping 变异函数建模优化使用遗传算法和机器学习线性回归：应用顺序高斯模拟映射

Artificial Intelligence in Geosciences

Pub Date : 2025-06-01 Epub Date: 2025-05-26 DOI: 10.1016/j.aiig.2025.100124

André William Boroh , Alpha Baster Kenfack Fokem , Martin Luther Mfenjou , Firmin Dimitry Hamat , Fritz Mbounja Besseme

The objective of this study is to develop an advanced approach to variogram modelling by integrating genetic algorithms (GA) with machine learning-based linear regression, aiming to improve the accuracy and efficiency of geostatistical analysis, particularly in mineral exploration. The study combines GA and machine learning to optimise variogram parameters, including range, sill, and nugget, by minimising the root mean square error (RMSE) and maximising the coefficient of determination (R²). The experimental variograms were computed and modelled using theoretical models, followed by optimisation via evolutionary algorithms. The method was applied to gravity data from the Ngoura-Batouri-Kette mining district in Eastern Cameroon, covering 141 data points. Sequential Gaussian Simulations (SGS) were employed for predictive mapping to validate simulated results against true values. Key findings show variograms with ranges between 24.71 km and 49.77 km, optimised RMSE and R² values of 11.21 mGal² and 0.969, respectively, after 42 generations of GA optimisation. Predictive mapping using SGS demonstrated that simulated values closely matched true values, with the simulated mean at 21.75 mGal compared to the true mean of 25.16 mGal, and variances of 465.70 mGal² and 555.28 mGal², respectively. The results confirmed spatial variability and anisotropies in the N170-N210 directions, consistent with prior studies. This work presents a novel integration of GA and machine learning for variogram modelling, offering an automated, efficient approach to parameter estimation. The methodology significantly enhances predictive geostatistical models, contributing to the advancement of mineral exploration and improving the precision and speed of decision-making in the petroleum and mining industries.

本研究的目的是通过将遗传算法（GA）与基于机器学习的线性回归相结合，开发一种先进的变异函数建模方法，旨在提高地质统计分析的准确性和效率，特别是在矿产勘探方面。该研究结合了遗传算法和机器学习，通过最小化均方根误差（RMSE）和最大化决定系数（R2）来优化变异函数参数，包括范围、基差和块金。使用理论模型对实验变差进行计算和建模，然后通过进化算法进行优化。该方法应用于喀麦隆东部Ngoura-Batouri-Kette矿区的141个数据点的重力数据。采用序贯高斯模拟（SGS）进行预测映射，根据真实值验证模拟结果。结果表明，42代遗传优化后的变异区间为24.71 ~ 49.77 km，优化后的RMSE和R2分别为11.21 mGal2和0.969。使用SGS进行预测映射表明，模拟值与真实值非常匹配，模拟平均值为21.75 mGal，而真实平均值为25.16 mGal，方差分别为465.70 mGal2和555.28 mGal2。结果证实了n170 ~ n210方向的空间变异性和各向异性，与前人的研究结果一致。这项工作提出了一种新的遗传算法和变异函数建模机器学习的集成，提供了一种自动化，有效的参数估计方法。该方法大大增强了预测地质统计模型，有助于推进矿产勘探，提高石油和采矿业决策的精度和速度。

{"title":"Variogram modelling optimisation using genetic algorithm and machine learning linear regression: application for Sequential Gaussian Simulations mapping","authors":"André William Boroh , Alpha Baster Kenfack Fokem , Martin Luther Mfenjou , Firmin Dimitry Hamat , Fritz Mbounja Besseme","doi":"10.1016/j.aiig.2025.100124","DOIUrl":"10.1016/j.aiig.2025.100124","url":null,"abstract":"<div><div>The objective of this study is to develop an advanced approach to variogram modelling by integrating genetic algorithms (GA) with machine learning-based linear regression, aiming to improve the accuracy and efficiency of geostatistical analysis, particularly in mineral exploration. The study combines GA and machine learning to optimise variogram parameters, including range, sill, and nugget, by minimising the root mean square error (RMSE) and maximising the coefficient of determination (R<sup>2</sup>). The experimental variograms were computed and modelled using theoretical models, followed by optimisation via evolutionary algorithms. The method was applied to gravity data from the Ngoura-Batouri-Kette mining district in Eastern Cameroon, covering 141 data points. Sequential Gaussian Simulations (SGS) were employed for predictive mapping to validate simulated results against true values. Key findings show variograms with ranges between 24.71 km and 49.77 km, optimised RMSE and R<sup>2</sup> values of 11.21 mGal<sup>2</sup> and 0.969, respectively, after 42 generations of GA optimisation. Predictive mapping using SGS demonstrated that simulated values closely matched true values, with the simulated mean at 21.75 mGal compared to the true mean of 25.16 mGal, and variances of 465.70 mGal<sup>2</sup> and 555.28 mGal<sup>2</sup>, respectively. The results confirmed spatial variability and anisotropies in the N170-N210 directions, consistent with prior studies. This work presents a novel integration of GA and machine learning for variogram modelling, offering an automated, efficient approach to parameter estimation. The methodology significantly enhances predictive geostatistical models, contributing to the advancement of mineral exploration and improving the precision and speed of decision-making in the petroleum and mining industries.</div></div>","PeriodicalId":100124,"journal":{"name":"Artificial Intelligence in Geosciences","volume":"6 1","pages":"Article 100124"},"PeriodicalIF":0.0,"publicationDate":"2025-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144168286","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0