Applied Computing and Geosciences最新文献_第5页

Development of a technique to identify μm-sized organic matter in asteroidal material: An approach using machine learning 在小行星材料中识别μm大小有机物的技术发展：一种使用机器学习的方法

IF 3.2 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Applied Computing and Geosciences

Pub Date : 2025-08-07 DOI: 10.1016/j.acags.2025.100277

Rahul Kumar, Katsura Kobayashi, Christian Potiszil, Tak Kunihiro

Asteroidal materials contain organic matter (OM), which records a number of extraterrestrial environments and thus provides a record of Solar System processes. OM contain essential compounds for the origin of life. To understand the origin and evolution of OM, systematic identification and detailed observation using in-situ techniques is required. While both nm- and μm-sized OM were studied previously, only a small portion of a given sample surface was investigated in each study. Here, a novel workflow was developed and applied to identify and classify μm-sized OM on mm-sized asteroidal materials. The workflow involved image processing and machine learning, enabling a comprehensive and non-biased way of identifying, classifying, and measuring the properties of OM. We found that identifying OM is more accurate by classification with machine learning than by clustering. On the approach of classification with machine learning, five algorithms were tested. The random forest algorithm was selected as it scored the highest in 4 out of 5 accuracy parameters during evaluation. The workflow gave modal OM abundances that were consistent with those identified manually, demonstrating that the workflow can accurately identify 1-15 μm-sized OM. The size distribution of OM was modeled using the power-law distribution, giving slope α values that were consistent with fragmentation processes. The shape of the OM was quantified using circularity and solidity, giving a positive correlation and indicating these properties are closely related. Overall, the workflow enabled identification of many OM quickly and accurately and the obtainment of chemical and petrographic information. Such information can help the selection of OM for further in-situ techniques, and elucidate the origin and evolution of OM preserved in asteroidal materials.

小行星物质中含有有机物（OM），它记录了许多地外环境，从而提供了太阳系过程的记录。OM含有生命起源所必需的化合物。为了了解有机质的起源和演化，需要利用原位技术进行系统的鉴定和详细的观察。虽然以前研究过nm和μm尺寸的OM，但每次研究只研究了给定样品表面的一小部分。本文提出了一种新的工作流程，并将其应用于mm级小行星材料上μm级OM的识别和分类。该工作流程涉及图像处理和机器学习，能够以一种全面、无偏见的方式识别、分类和测量OM的属性。我们发现，通过机器学习分类识别OM比通过聚类更准确。在机器学习的分类方法上，测试了五种算法。随机森林算法在评估的5个精度参数中有4个得分最高，因此选择了随机森林算法。该工作流给出的模态OM丰度与人工识别的丰度一致，表明该工作流可以准确识别1-15 μm大小的OM。OM的大小分布采用幂律分布建模，斜率α值与破碎过程一致。OM的形状用圆度和固体度来量化，给出了正相关关系，表明这些性质密切相关。总的来说，该工作流程能够快速准确地识别许多OM，并获得化学和岩石学信息。这些信息可以帮助进一步的原位技术选择OM，并阐明保存在小行星材料中的OM的起源和演化。

{"title":"Development of a technique to identify μm-sized organic matter in asteroidal material: An approach using machine learning","authors":"Rahul Kumar, Katsura Kobayashi, Christian Potiszil, Tak Kunihiro","doi":"10.1016/j.acags.2025.100277","DOIUrl":"10.1016/j.acags.2025.100277","url":null,"abstract":"<div><div>Asteroidal materials contain organic matter (OM), which records a number of extraterrestrial environments and thus provides a record of Solar System processes. OM contain essential compounds for the origin of life. To understand the origin and evolution of OM, systematic identification and detailed observation using in-situ techniques is required. While both nm- and μm-sized OM were studied previously, only a small portion of a given sample surface was investigated in each study. Here, a novel workflow was developed and applied to identify and classify μm-sized OM on mm-sized asteroidal materials. The workflow involved image processing and machine learning, enabling a comprehensive and non-biased way of identifying, classifying, and measuring the properties of OM. We found that identifying OM is more accurate by classification with machine learning than by clustering. On the approach of classification with machine learning, five algorithms were tested. The random forest algorithm was selected as it scored the highest in 4 out of 5 accuracy parameters during evaluation. The workflow gave modal OM abundances that were consistent with those identified manually, demonstrating that the workflow can accurately identify 1-15 μm-sized OM. The size distribution of OM was modeled using the power-law distribution, giving slope α values that were consistent with fragmentation processes. The shape of the OM was quantified using circularity and solidity, giving a positive correlation and indicating these properties are closely related. Overall, the workflow enabled identification of many OM quickly and accurately and the obtainment of chemical and petrographic information. Such information can help the selection of OM for further in-situ techniques, and elucidate the origin and evolution of OM preserved in asteroidal materials.</div></div>","PeriodicalId":33804,"journal":{"name":"Applied Computing and Geosciences","volume":"27 ","pages":"Article 100277"},"PeriodicalIF":3.2,"publicationDate":"2025-08-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144826846","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

On-board camera-based automatic zoning method for heading face by using computerized rock drilling cart 基于车载摄像头的微机凿岩车掘进工作面自动分区方法

IF 3.2 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Applied Computing and Geosciences

Pub Date : 2025-08-07 DOI: 10.1016/j.acags.2025.100275

Yong-Feng Li , Huan Li , Jing Xiao , Weidong Ren , Mohammed Abdalla Elsharif Ibrahim

During construction, drilling parameters are manually adjusted by the operator, which can affect the blasting effect due to inappropriate initial parameters. To address this issue, an automatic optimal drilling method based on image partitioning of the heading face is proposed: i) Obtain images of the heading face using a suitable vehicle camera, and calculate pixel coordinates on the virtual heading face through rock drilling cart positioning and virtual heading face positioning; ii) Apply the region growth algorithm to extract the image region of the heading face, segment the image into several super-pixel units using the linear iterative clustering algorithm, followed by combining super-pixels based on the gray difference criterion. The resulting super-pixel blocks serve as the training sample set for the rock-partition method based on super-pixels and support vector machine (SVM); iii) Establish a database of drilling parameters. The results demonstrate that, compared to the region growth algorithm, the classification method based on super-pixels and SVM has higher accuracy. The algorithm has high accuracy of partition effect and good real-time performance, providing a reliable basis for optimizing the opening parameters.

施工过程中，钻孔参数由作业人员手动调整，由于初始参数不合适，会影响爆破效果。针对这一问题，提出了一种基于掘进工作面图像分割的自动优化钻进方法：i)利用合适的车载摄像头获取掘进工作面图像，通过凿岩车定位和虚拟掘进工作面定位计算虚拟掘进工作面像素坐标；ii)应用区域增长算法提取标题面图像区域，使用线性迭代聚类算法将图像分割成多个超像素单元，然后基于灰度差准则对超像素进行组合。得到的超像素块作为基于超像素和支持向量机（SVM）的岩石划分方法的训练样本集；iii)建立钻井参数数据库。结果表明，与区域增长算法相比，基于超像素和支持向量机的分类方法具有更高的准确率。该算法分割效果精度高，实时性好，为优化开孔参数提供了可靠的依据。

{"title":"On-board camera-based automatic zoning method for heading face by using computerized rock drilling cart","authors":"Yong-Feng Li , Huan Li , Jing Xiao , Weidong Ren , Mohammed Abdalla Elsharif Ibrahim","doi":"10.1016/j.acags.2025.100275","DOIUrl":"10.1016/j.acags.2025.100275","url":null,"abstract":"<div><div>During construction, drilling parameters are manually adjusted by the operator, which can affect the blasting effect due to inappropriate initial parameters. To address this issue, an automatic optimal drilling method based on image partitioning of the heading face is proposed: i) Obtain images of the heading face using a suitable vehicle camera, and calculate pixel coordinates on the virtual heading face through rock drilling cart positioning and virtual heading face positioning; ii) Apply the region growth algorithm to extract the image region of the heading face, segment the image into several super-pixel units using the linear iterative clustering algorithm, followed by combining super-pixels based on the gray difference criterion. The resulting super-pixel blocks serve as the training sample set for the rock-partition method based on super-pixels and support vector machine (SVM); iii) Establish a database of drilling parameters. The results demonstrate that, compared to the region growth algorithm, the classification method based on super-pixels and SVM has higher accuracy. The algorithm has high accuracy of partition effect and good real-time performance, providing a reliable basis for optimizing the opening parameters.</div></div>","PeriodicalId":33804,"journal":{"name":"Applied Computing and Geosciences","volume":"27 ","pages":"Article 100275"},"PeriodicalIF":3.2,"publicationDate":"2025-08-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144827362","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

Orange-Volcanoes: A new open and collaborative platform to perform data-driven investigations and machine learning analyses in petrology and volcanology Orange-Volcanoes：一个新的开放和协作平台，用于在岩石学和火山学中进行数据驱动的调查和机器学习分析

IF 3.2 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Applied Computing and Geosciences

Pub Date : 2025-08-05 DOI: 10.1016/j.acags.2025.100270

Alessandro Musu , Valerio Parodi , Marko Toplak , Alessandro Carfì , Mónica Ágreda-López , Fulvio Mastrogiovanni , J. ZhangZhou , Diego Perugini , Donato Belmonte , Penny E. Wieser , Blaž Zupan , Maurizio Petrelli

Orange-Volcanoes is an extension of the open-source Orange data mining platform specifically tailored for geochemical, petrological, and volcanological investigations. Orange-Volcanoes enhances the original platform by incorporating specialized tools to enable interactive data-driven investigations in geochemistry, such as performing Compositional Data Analysis (CoDA). Applying CoDA transformations enables the use of many standard and multivariate statistical methods like principal component analysis, discriminant analysis, and hierarchical clustering on compositional data. In this way, Orange-Volcanoes allows for the application of a wide range of data mining and statistical methods implemented in Orange using geochemical data. Moreover, Orange allows the use of advanced methods in the field of explainable artificial intelligence, such as feature importance and Shapley additive explanations. Also, within Orange-Volcanoes, we demonstrate the flexibility of the Orange platform by developing visual tools that allow for conducting mineral-liquid equilibrium tests and calculating thermo-barometric estimates. The Orange-Volcanoes supports collaborative efforts and reproducibility by offering a visual programming interface that requires no coding experience, making it accessible to a wide range of users, including scientists, educators, and students. We provide a series of case studies, including interactive petrological data exploration and clustering in tephra studies to highlight Orange-Volcanoes’ potential and versatility in volcanological applications. Orange-Volcanoes can be downloaded using pip, and its documentation is available at https://orange3-volcanoes.readthedocs.io/en/latest/.

Orange- volcanoes是开源Orange数据挖掘平台的扩展，专门为地球化学、岩石学和火山学调查量身定制。Orange-Volcanoes通过整合专门的工具来增强原始平台，使交互式数据驱动的地球化学调查成为可能，例如执行成分数据分析（CoDA）。应用CoDA转换可以使用许多标准和多元统计方法，如主成分分析、判别分析和组合数据的分层聚类。通过这种方式，Orange- volcanoes允许应用在Orange中使用地球化学数据实施的广泛的数据挖掘和统计方法。此外，Orange允许使用可解释人工智能领域的先进方法，例如特征重要性和Shapley加法解释。此外，在Orange- volcanoes中，我们通过开发可视化工具来展示Orange平台的灵活性，这些工具允许进行矿物-液体平衡测试和计算热气压估算。Orange-Volcanoes通过提供不需要编码经验的可视化编程界面来支持协作和再现性，使包括科学家、教育工作者和学生在内的广泛用户都可以访问它。我们提供了一系列的案例研究，包括交互式岩石学数据勘探和火山研究中的聚类，以突出橙色火山在火山学应用中的潜力和多功能性。可以使用pip下载Orange-Volcanoes，其文档可从https://orange3-volcanoes.readthedocs.io/en/latest/获得。

{"title":"Orange-Volcanoes: A new open and collaborative platform to perform data-driven investigations and machine learning analyses in petrology and volcanology","authors":"Alessandro Musu , Valerio Parodi , Marko Toplak , Alessandro Carfì , Mónica Ágreda-López , Fulvio Mastrogiovanni , J. ZhangZhou , Diego Perugini , Donato Belmonte , Penny E. Wieser , Blaž Zupan , Maurizio Petrelli","doi":"10.1016/j.acags.2025.100270","DOIUrl":"10.1016/j.acags.2025.100270","url":null,"abstract":"<div><div>Orange-Volcanoes is an extension of the open-source Orange data mining platform specifically tailored for geochemical, petrological, and volcanological investigations. Orange-Volcanoes enhances the original platform by incorporating specialized tools to enable interactive data-driven investigations in geochemistry, such as performing Compositional Data Analysis (CoDA). Applying CoDA transformations enables the use of many standard and multivariate statistical methods like principal component analysis, discriminant analysis, and hierarchical clustering on compositional data. In this way, Orange-Volcanoes allows for the application of a wide range of data mining and statistical methods implemented in Orange using geochemical data. Moreover, Orange allows the use of advanced methods in the field of explainable artificial intelligence, such as feature importance and Shapley additive explanations. Also, within Orange-Volcanoes, we demonstrate the flexibility of the Orange platform by developing visual tools that allow for conducting mineral-liquid equilibrium tests and calculating thermo-barometric estimates. The Orange-Volcanoes supports collaborative efforts and reproducibility by offering a visual programming interface that requires no coding experience, making it accessible to a wide range of users, including scientists, educators, and students. We provide a series of case studies, including interactive petrological data exploration and clustering in tephra studies to highlight Orange-Volcanoes’ potential and versatility in volcanological applications. Orange-Volcanoes can be downloaded using pip, and its documentation is available at <span><span>https://orange3-volcanoes.readthedocs.io/en/latest/</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":33804,"journal":{"name":"Applied Computing and Geosciences","volume":"27 ","pages":"Article 100270"},"PeriodicalIF":3.2,"publicationDate":"2025-08-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144780986","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

Recent advances in explainable Machine Learning models for wildfire prediction 用于野火预测的可解释机器学习模型的最新进展

IF 3.2 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Applied Computing and Geosciences

Pub Date : 2025-07-30 DOI: 10.1016/j.acags.2025.100266

Abira Sengupta, Brendon J. Woodford

Climate change has caused increasingly frequent occurrences of forest fires around the world. Machine Learning (ML) and Artificial Intelligence models have emerged to predict both the onset of wildfires and evaluate the extent of damage a wildfire would cause. However, understanding what factors lead to generating models that exhibit optimal performance and providing insight into the importance of features on model outcomes is the subject of ongoing research. To help answer these questions, we propose a framework which adopts recent advances in methods for obtaining optimal models along with the application of SHAP (SHapley Additive exPlanations) values to obtain the most important features which affect the performance of wildfire prediction models. We use this framework as a classification task to predict the likelihood of wildfire occurrence based on environmental conditions, using a data set which represents instances of forest fires in Algerian, and as a regression task to predict the burned area once a wildfire has begun, using a data set from Portugal that recorded the area burned after a fire event. Insights provided by this framework allow us to assess the efficacy of specific ML models for wildfire prediction, ultimately making recommendations as to which ML models are more suited towards these challenging tasks.

气候变化导致世界范围内森林火灾的发生越来越频繁。机器学习（ML）和人工智能模型已经出现，既可以预测野火的发生，也可以评估野火可能造成的破坏程度。然而，了解哪些因素导致生成表现出最佳性能的模型，并深入了解特征对模型结果的重要性，是正在进行的研究的主题。为了帮助回答这些问题，我们提出了一个框架，该框架采用了获得最优模型的最新进展以及SHapley加性解释（SHapley Additive explanation）值的应用，以获得影响野火预测模型性能的最重要特征。我们使用此框架作为分类任务，根据环境条件预测野火发生的可能性，使用代表阿尔及利亚森林火灾实例的数据集，并使用来自葡萄牙的记录火灾事件后燃烧区域的数据集，作为回归任务，预测野火开始后燃烧区域。该框架提供的见解使我们能够评估特定ML模型对野火预测的功效，最终就哪种ML模型更适合这些具有挑战性的任务提出建议。

{"title":"Recent advances in explainable Machine Learning models for wildfire prediction","authors":"Abira Sengupta, Brendon J. Woodford","doi":"10.1016/j.acags.2025.100266","DOIUrl":"10.1016/j.acags.2025.100266","url":null,"abstract":"<div><div>Climate change has caused increasingly frequent occurrences of forest fires around the world. Machine Learning (ML) and Artificial Intelligence models have emerged to predict both the onset of wildfires and evaluate the extent of damage a wildfire would cause. However, understanding what factors lead to generating models that exhibit optimal performance and providing insight into the importance of features on model outcomes is the subject of ongoing research. To help answer these questions, we propose a framework which adopts recent advances in methods for obtaining optimal models along with the application of SHAP (SHapley Additive exPlanations) values to obtain the most important features which affect the performance of wildfire prediction models. We use this framework as a classification task to predict the likelihood of wildfire occurrence based on environmental conditions, using a data set which represents instances of forest fires in Algerian, and as a regression task to predict the burned area once a wildfire has begun, using a data set from Portugal that recorded the area burned after a fire event. Insights provided by this framework allow us to assess the efficacy of specific ML models for wildfire prediction, ultimately making recommendations as to which ML models are more suited towards these challenging tasks.</div></div>","PeriodicalId":33804,"journal":{"name":"Applied Computing and Geosciences","volume":"27 ","pages":"Article 100266"},"PeriodicalIF":3.2,"publicationDate":"2025-07-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144780987","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

Classification of microscopic images of rock thin sections based on TLCA-ResNet34 基于TLCA-ResNet34的岩石薄片显微图像分类

IF 3.2 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Applied Computing and Geosciences

Pub Date : 2025-07-29 DOI: 10.1016/j.acags.2025.100272

Zhenyu Zhao , Shucheng Tan , Hui Chen , Pengwei Wang , Qinghua Zhang , Haoyu Wei , Zhenlin Zhang

Identifying microscopic images of rocks is a crucial method for rock identification, playing a vital role in geological exploration and mineral mining. To facilitate the quick classification and identification of rock thin sections under a microscope, a dataset with 3116 microscopic images of 9 types of rock thin sections was developed using publicly accessible network datasets. By adopting the transfer learning method, a context-aware residual block was designed using the coordinate attention(CA) mechanism, and a targeted TLCA-ResNet34 neural network model was developed. This model is capable of extracting deep-layer feature information from entire rock thin section images, thus achieving the classification and identification of microscopic images. The experimental results show that, compared with several other common models, TLCA-ResNet34, while maintaining the light weight of the model, has the best recognition accuracy, recall rate, and Matthews correlation coefficient (MCC) for the microscopy image test set. It can efficiently and accurately identify microscopic images of rocks.

岩石显微图像识别是岩石识别的重要方法，在地质勘查和矿产开采中起着至关重要的作用。为了方便显微镜下岩石薄片的快速分类和识别，利用可公开访问的网络数据集，开发了包含9种岩石薄片3116张显微图像的数据集。采用迁移学习方法，利用坐标注意（CA）机制设计了上下文感知残差块，建立了TLCA-ResNet34靶向神经网络模型。该模型能够从整个岩石薄片图像中提取深层特征信息，从而实现微观图像的分类识别。实验结果表明，与其他几种常用模型相比，TLCA-ResNet34在保持模型轻量化的同时，对显微镜图像测试集具有最好的识别准确率、召回率和马修斯相关系数（MCC）。该方法能够高效、准确地识别岩石显微图像。

{"title":"Classification of microscopic images of rock thin sections based on TLCA-ResNet34","authors":"Zhenyu Zhao , Shucheng Tan , Hui Chen , Pengwei Wang , Qinghua Zhang , Haoyu Wei , Zhenlin Zhang","doi":"10.1016/j.acags.2025.100272","DOIUrl":"10.1016/j.acags.2025.100272","url":null,"abstract":"<div><div>Identifying microscopic images of rocks is a crucial method for rock identification, playing a vital role in geological exploration and mineral mining. To facilitate the quick classification and identification of rock thin sections under a microscope, a dataset with 3116 microscopic images of 9 types of rock thin sections was developed using publicly accessible network datasets. By adopting the transfer learning method, a context-aware residual block was designed using the coordinate attention(CA) mechanism, and a targeted TLCA-ResNet34 neural network model was developed. This model is capable of extracting deep-layer feature information from entire rock thin section images, thus achieving the classification and identification of microscopic images. The experimental results show that, compared with several other common models, TLCA-ResNet34, while maintaining the light weight of the model, has the best recognition accuracy, recall rate, and Matthews correlation coefficient (MCC) for the microscopy image test set. It can efficiently and accurately identify microscopic images of rocks.</div></div>","PeriodicalId":33804,"journal":{"name":"Applied Computing and Geosciences","volume":"27 ","pages":"Article 100272"},"PeriodicalIF":3.2,"publicationDate":"2025-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144757227","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

SRT-Ai: Identifying seismic reflection terminations using deep learning SRT-Ai：利用深度学习识别地震反射终端

IF 3.2 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Applied Computing and Geosciences

Pub Date : 2025-07-26 DOI: 10.1016/j.acags.2025.100271

Waleed M. AlGharbi , Rebecca E. Bell , Cédric M. John

Seismic stratigraphy entails a regional scanning (reconnaissance) of seismic data to identify and annotate seismic reflection terminations. To identify these terminations in modern 3D seismic datasets, interpreters have to examine thousands of inlines and crosslines, which is a time-consuming process. Furthermore, accurate identification of these features relies heavily on human visual observation along with individual expertise.

A growing number of studies have shown promising results applying machine learning techniques to identify geological features from seismic data such as salt bodies and faults. However, the identification of seismic reflection terminations has not received the same level of interest and remains a manual process. One of the barriers to utilizing machine learning techniques in seismic interpretation is the lack of “labelled” data. In this study, we evaluate the ability of deep learning Convolutional Neural Networks (CNN) trained on synthetic seismic images to identify seismic reflection terminations.

A dataset comprising 160 000 synthetic seismic images that represent conformable and four types of seismic reflection terminations (truncation, toplap, onlap, and downlap) were created using geometric geological modelling and 1D convolution seismic modelling. The dataset was then split into two classes (“Contains Termination” and “No Termination”). A new CNN model architecture named “Seismic Reflection Terminations Attribute (SRT-Ai)” was trained on 80 % of the synthetic seismic dataset. SRT-Ai predicted the test set (remaining 20 %) with an accuracy and precision of 99.9 %. To test its generalization, SRT-Ai was also evaluated on real seismic images, achieving 91 % accuracy and 96 % precision against published interpretations used as reference labels. Qualitative analysis of predictions along seismic sections shows a strong correspondence between the model predictions and manual regional interpretations.

SRT-Ai is proposed as a screening tool that will assist seismic interpreters with the identification of major seismic terminations, minimise seismic interpretation uncertainties, reduce the time taken for seismic reconnaissance, and limit the reliance on human visual observation at the early stage of seismic interpretation process.

地震地层学需要对地震数据进行区域扫描（侦察），以识别和注释地震反射终端。为了在现代三维地震数据集中识别这些端点，解释人员必须检查数千条内线和交线，这是一个耗时的过程。此外，这些特征的准确识别在很大程度上依赖于人类的视觉观察以及个人的专业知识。越来越多的研究表明，应用机器学习技术从地震数据中识别地质特征（如盐体和断层）取得了可喜的结果。然而，地震反射终端的识别并没有受到同等程度的关注，仍然是一个人工过程。在地震解释中使用机器学习技术的障碍之一是缺乏“标记”数据。在这项研究中，我们评估了在合成地震图像上训练的深度学习卷积神经网络（CNN）识别地震反射终止的能力。使用几何地质建模和一维卷积地震建模，创建了一个包含16万张合成地震图像的数据集，这些图像代表了四种类型的地震反射终端（截断、上覆、上覆和下覆）。然后将数据集分成两类（“包含终止”和“无终止”）。在80%的合成地震数据集上训练了一种新的CNN模型架构“地震反射终止属性（SRT-Ai）”。SRT-Ai预测测试集（剩余20%）的准确度和精密度为99.9%。为了验证其泛化性，SRT-Ai还在真实地震图像上进行了评估，与用作参考标签的已发表解释相比，准确率达到91%，精度达到96%。沿地震剖面预测的定性分析表明，模式预测与人工区域解释之间有很强的对应关系。SRT-Ai被认为是一种筛选工具，可以帮助地震解释人员识别主要的地震终端，最大限度地减少地震解释的不确定性，减少地震侦察所需的时间，并限制在地震解释过程的早期阶段对人类视觉观测的依赖。

{"title":"SRT-Ai: Identifying seismic reflection terminations using deep learning","authors":"Waleed M. AlGharbi , Rebecca E. Bell , Cédric M. John","doi":"10.1016/j.acags.2025.100271","DOIUrl":"10.1016/j.acags.2025.100271","url":null,"abstract":"<div><div>Seismic stratigraphy entails a regional scanning (reconnaissance) of seismic data to identify and annotate seismic reflection terminations. To identify these terminations in modern 3D seismic datasets, interpreters have to examine thousands of inlines and crosslines, which is a time-consuming process. Furthermore, accurate identification of these features relies heavily on human visual observation along with individual expertise.</div><div>A growing number of studies have shown promising results applying machine learning techniques to identify geological features from seismic data such as salt bodies and faults. However, the identification of seismic reflection terminations has not received the same level of interest and remains a manual process. One of the barriers to utilizing machine learning techniques in seismic interpretation is the lack of “labelled” data. In this study, we evaluate the ability of deep learning Convolutional Neural Networks (CNN) trained on synthetic seismic images to identify seismic reflection terminations.</div><div>A dataset comprising 160 000 synthetic seismic images that represent conformable and four types of seismic reflection terminations (truncation, toplap, onlap, and downlap) were created using geometric geological modelling and 1D convolution seismic modelling. The dataset was then split into two classes (“Contains Termination” and “No Termination”). A new CNN model architecture named “Seismic Reflection Terminations Attribute (SRT-Ai)” was trained on 80 % of the synthetic seismic dataset. SRT-Ai predicted the test set (remaining 20 %) with an accuracy and precision of 99.9 %. To test its generalization, SRT-Ai was also evaluated on real seismic images, achieving 91 % accuracy and 96 % precision against published interpretations used as reference labels. Qualitative analysis of predictions along seismic sections shows a strong correspondence between the model predictions and manual regional interpretations.</div><div>SRT-Ai is proposed as a screening tool that will assist seismic interpreters with the identification of major seismic terminations, minimise seismic interpretation uncertainties, reduce the time taken for seismic reconnaissance, and limit the reliance on human visual observation at the early stage of seismic interpretation process.</div></div>","PeriodicalId":33804,"journal":{"name":"Applied Computing and Geosciences","volume":"27 ","pages":"Article 100271"},"PeriodicalIF":3.2,"publicationDate":"2025-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144721214","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

Electrical anisotropy calculation of the continental crust by resistor network-based circuit simulations 基于电阻网络的大陆地壳电性各向异性计算电路模拟

IF 2.6 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Applied Computing and Geosciences

Pub Date : 2025-07-24 DOI: 10.1016/j.acags.2025.100265

Song Luo , Haiying Hu , Lidong Dai

Electrical anisotropy has been broadly observed by magnetotelluric (MT) surveys in the continental crust. It is proposed to be caused by rock microfabrics, lithologic layering, or oriented alignment of fluid or melt in rocks, whereas the validity of these mechanisms has not yet been verified due to the lack of experimental and computational evidence. Laboratory measurements on the electrical anisotropy of crustal rocks are extremely challenging when considering microfabrics and oriented microcracks filled with fluid. In contrast, numerical modeling, being an efficient approach, can be used to compute the anisotropic physical properties of rocks. In this study, the electrical anisotropy of crustal rocks was first modeled by circuit simulation techniques using a random resistor network model, based on the lattice-preferred orientation, modal compositions, and mineral electrical conductivity. The results indicate that the conversion from single crystals to the corresponding aggregates leads to a great reduction in electrical anisotropy, particularly for quartz single crystal with high anisotropy. Moreover, the electrical anisotropy of two-phase aggregates decreases with the increasing proportion of the second low-anisotropy minerals (e.g., plagioclase), such as from quartzite to granite. For layered lithology, the lower-crustal gabbro has higher electrical anisotropy compared to middle-crustal quartz-bearing rocks. The modeled electrical anisotropy from the middle to lower crust matches well with the geophysical observations in the Central Great Basin.

大陆地壳电性各向异性在大地电磁测量中得到了广泛的观察。它被认为是由岩石微组构、岩性分层或岩石中流体或熔体的定向排列引起的，然而由于缺乏实验和计算证据，这些机制的有效性尚未得到验证。当考虑到充满流体的微结构和定向微裂缝时，对地壳岩石电性各向异性的实验室测量极具挑战性。而数值模拟是计算岩石各向异性物理性质的有效方法。在这项研究中，首先通过电路模拟技术，利用随机电阻网络模型，基于晶格优选取向、模态组成和矿物电导率，模拟了地壳岩石的电各向异性。结果表明，从单晶到相应的聚集体的转变导致电各向异性的显著降低，特别是对于具有高各向异性的石英单晶。两相团聚体的电性各向异性随第二低各向异性矿物（如斜长石）的比例增加而降低，如从石英岩到花岗岩。对于层状岩性，下地壳辉长岩电性各向异性高于中地壳含石英岩。模拟的中下地壳电性各向异性与中央大盆地的地球物理观测结果吻合较好。

{"title":"Electrical anisotropy calculation of the continental crust by resistor network-based circuit simulations","authors":"Song Luo , Haiying Hu , Lidong Dai","doi":"10.1016/j.acags.2025.100265","DOIUrl":"10.1016/j.acags.2025.100265","url":null,"abstract":"<div><div>Electrical anisotropy has been broadly observed by magnetotelluric (MT) surveys in the continental crust. It is proposed to be caused by rock microfabrics, lithologic layering, or oriented alignment of fluid or melt in rocks, whereas the validity of these mechanisms has not yet been verified due to the lack of experimental and computational evidence. Laboratory measurements on the electrical anisotropy of crustal rocks are extremely challenging when considering microfabrics and oriented microcracks filled with fluid. In contrast, numerical modeling, being an efficient approach, can be used to compute the anisotropic physical properties of rocks. In this study, the electrical anisotropy of crustal rocks was first modeled by circuit simulation techniques using a random resistor network model, based on the lattice-preferred orientation, modal compositions, and mineral electrical conductivity. The results indicate that the conversion from single crystals to the corresponding aggregates leads to a great reduction in electrical anisotropy, particularly for quartz single crystal with high anisotropy. Moreover, the electrical anisotropy of two-phase aggregates decreases with the increasing proportion of the second low-anisotropy minerals (e.g., plagioclase), such as from quartzite to granite. For layered lithology, the lower-crustal gabbro has higher electrical anisotropy compared to middle-crustal quartz-bearing rocks. The modeled electrical anisotropy from the middle to lower crust matches well with the geophysical observations in the Central Great Basin.</div></div>","PeriodicalId":33804,"journal":{"name":"Applied Computing and Geosciences","volume":"27 ","pages":"Article 100265"},"PeriodicalIF":2.6,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144714167","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

Attention and deformable convolution-based dual-task high-precision fault recognition 基于注意力和可变形卷积的双任务高精度故障识别

IF 3.2 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Applied Computing and Geosciences

Pub Date : 2025-07-23 DOI: 10.1016/j.acags.2025.100267

Zhen Peng , Danping Cao , Huiqun Xu , Dan Zhu

Deep learning has been widely applied in fault recognition task. However, current two-dimensional (2D) deep learning-based training methods fail to adequately consider the overall spatial characteristics of faults, resulting in discontinuous fault recognition results and unable to achieve the effect of three-dimensional (3D) deep learning training methods. To address this issue, we propose an attention and deformable convolution-based dual-task high-precision fault recognition method (ADTFM), which introduces a dual-task deep learning network architecture within the 2D training framework, effectively improving the fault recognition accuracy and reliability. ADTFM consists of two tasks (Main task and Auxiliary task) with the same network structure based on the deformable convolution operators and U-Net. The main task uses the Inline direction for training, and uses the deformable convolution operator to capture more accurate fault feature. At the same time, the auxiliary task is trained in the Time-slice direction, and the features generated by auxiliary task direction are transferred to the main task in training process. The two tasks are connected through the attention mechanism, so as to increase the spatial characteristics of faults in 2D training process, and effectively compensate for the spatial limitations of 2D training. By testing the public 3D datasets and the field 3D datasets, and comparing with the current high-precision FaultSeg3D fault recognition method, the results show that our method can improve the accuracy of fault recognition. Moreover, through the quantitative evaluation of computing consumption time and memory, it is shown that the proposed method effectively reduces the computational complexity and decreases the consumption of computational resources, and provide a more efficient solution for fault recognition task.

深度学习在故障识别中得到了广泛的应用。然而，目前基于二维（2D）深度学习的训练方法没有充分考虑故障的整体空间特征，导致故障识别结果不连续，无法达到三维（3D）深度学习训练方法的效果。针对这一问题，我们提出了一种基于注意力和可变形卷积的双任务高精度故障识别方法（ADTFM），该方法在二维训练框架内引入了双任务深度学习网络架构，有效提高了故障识别的准确性和可靠性。ADTFM由两个具有相同网络结构的任务（主任务和辅助任务）组成，它们基于可变形卷积算子和U-Net。主要任务使用内联方向进行训练，并使用可变形卷积算子捕获更准确的故障特征。同时在时间片方向上对辅助任务进行训练，并在训练过程中将辅助任务方向产生的特征传递给主任务。通过注意机制将两个任务连接起来，增加了二维训练过程中故障的空间特征，有效弥补了二维训练的空间局限性。通过对公共三维数据集和现场三维数据集的测试，并与现有的高精度FaultSeg3D故障识别方法进行比较，结果表明本文方法可以提高故障识别的精度。此外，通过对计算消耗时间和内存的定量评价，表明所提方法有效地降低了计算复杂度，减少了计算资源的消耗，为故障识别任务提供了更高效的解决方案。

{"title":"Attention and deformable convolution-based dual-task high-precision fault recognition","authors":"Zhen Peng , Danping Cao , Huiqun Xu , Dan Zhu","doi":"10.1016/j.acags.2025.100267","DOIUrl":"10.1016/j.acags.2025.100267","url":null,"abstract":"<div><div>Deep learning has been widely applied in fault recognition task. However, current two-dimensional (2D) deep learning-based training methods fail to adequately consider the overall spatial characteristics of faults, resulting in discontinuous fault recognition results and unable to achieve the effect of three-dimensional (3D) deep learning training methods. To address this issue, we propose an attention and deformable convolution-based dual-task high-precision fault recognition method (ADTFM), which introduces a dual-task deep learning network architecture within the 2D training framework, effectively improving the fault recognition accuracy and reliability. ADTFM consists of two tasks (Main task and Auxiliary task) with the same network structure based on the deformable convolution operators and U-Net. The main task uses the Inline direction for training, and uses the deformable convolution operator to capture more accurate fault feature. At the same time, the auxiliary task is trained in the Time-slice direction, and the features generated by auxiliary task direction are transferred to the main task in training process. The two tasks are connected through the attention mechanism, so as to increase the spatial characteristics of faults in 2D training process, and effectively compensate for the spatial limitations of 2D training. By testing the public 3D datasets and the field 3D datasets, and comparing with the current high-precision FaultSeg3D fault recognition method, the results show that our method can improve the accuracy of fault recognition. Moreover, through the quantitative evaluation of computing consumption time and memory, it is shown that the proposed method effectively reduces the computational complexity and decreases the consumption of computational resources, and provide a more efficient solution for fault recognition task.</div></div>","PeriodicalId":33804,"journal":{"name":"Applied Computing and Geosciences","volume":"27 ","pages":"Article 100267"},"PeriodicalIF":3.2,"publicationDate":"2025-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144721213","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

Lithological mapping and spectroscopic studies of carbonatite and clinopyroxenite from Hogenakkal carbonatite complex, India 印度Hogenakkal碳酸岩杂岩中碳酸岩和斜辉石岩的岩性填图和光谱研究

IF 2.6 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Applied Computing and Geosciences

Pub Date : 2025-07-23 DOI: 10.1016/j.acags.2025.100269

Saraah Imran , Sourav Bhattacharjee , Ajanta Goswami , Aniket Chakrabarty

The Paleoproterozoic Hogenakkal carbonatite complex, situated within the Mettur shear zone, Southern Granulite Terrain, Tamil Nadu, India, is known for its enigmatic carbonatite-clinopyroxenite association and lithology specific rare earth elements (REE) mineralization. The complex comprises two types of carbonatites (silicate-rich carbonatite-I, and silicate-poor carbonatite-II), intruding the clinopyroxenite as isolated pods or ovoid bodies, and are together emplaced within the granulite country rocks. This study employs Landsat 8 multispectral data to map the spatial distribution and extent of the clinopyroxenite dykes. These dykes serve as geological tracers for locating the spatially associated carbonatite bodies. In addition, the present work investigate the spectroscopic properties of REE-bearing carbonatites and clinopyroxenite. Petrography, Raman spectroscopy of minerals, and spectroradiometric measurements of rock samples support the interpretations derived from Principal Component Analysis (PCA), Spectral Angle Mapper (SAM), Support Vector Machine (SVM), Decision Tree, and Random Forest algorithms, thereby aiding in the identification of lithological variations and potential clinopyroxenite occurrences. Carbonatite-II shows more prominent REE absorption features compared to carbonatite-I. This is consistent with petrographic observations and Raman spectroscopy, which show that the REE mineralization in carbonatite-II is dominated by monazite-(Ce) and hydroxylbastnäsite-(Ce), whereas carbonatite-I contains allanite-(Ce) as the primary REE-bearing phase. This study exhibits the efficacy of Landsat series data and non-destructive spectroscopic methods for preliminary mineral exploration and evaluating REE potential before detailed field investigation.

古元古代Hogenakkal碳酸盐岩杂岩位于印度泰米尔纳德邦南部麻粒岩地的Mettur剪切带内，以其神秘的碳酸岩-斜辉石岩组合和岩性特异的稀土元素成矿作用而闻名。杂岩包括两种类型的碳酸盐（富硅酸盐碳酸盐i型和贫硅酸盐碳酸盐ii型），以孤立的荚状或卵状体侵入斜辉石岩，并共同侵位于麻粒岩围岩中。利用Landsat 8多光谱数据绘制斜辉石岩岩脉的空间分布和范围。这些岩脉可作为地质示踪剂，用于定位空间关联的碳酸盐岩。此外，本文还研究了含稀土碳酸盐和斜辉石岩的光谱性质。岩石学、矿物拉曼光谱和岩石样品的光谱辐射测量支持主成分分析（PCA）、光谱角度映射器（SAM）、支持向量机（SVM）、决策树和随机森林算法的解释，从而有助于识别岩性变化和潜在斜辉石岩的产状。碳酸盐- ii比碳酸盐- i具有更明显的稀土吸收特征。这与岩石学和拉曼光谱观测结果一致，表明碳酸盐岩ii型中稀土矿化以独居石-(Ce)和hydroxylbastnäsite-(Ce)为主，而碳酸盐岩i型中则以allanite-(Ce)为主要含矿相。本研究展示了Landsat系列数据和非破坏性光谱方法在详细野外调查前的初步矿产勘探和稀土潜力评估中的有效性。

{"title":"Lithological mapping and spectroscopic studies of carbonatite and clinopyroxenite from Hogenakkal carbonatite complex, India","authors":"Saraah Imran , Sourav Bhattacharjee , Ajanta Goswami , Aniket Chakrabarty","doi":"10.1016/j.acags.2025.100269","DOIUrl":"10.1016/j.acags.2025.100269","url":null,"abstract":"<div><div>The Paleoproterozoic Hogenakkal carbonatite complex, situated within the Mettur shear zone, Southern Granulite Terrain, Tamil Nadu, India, is known for its enigmatic carbonatite-clinopyroxenite association and lithology specific rare earth elements (REE) mineralization. The complex comprises two types of carbonatites (silicate-rich carbonatite-I, and silicate-poor carbonatite-II), intruding the clinopyroxenite as isolated pods or ovoid bodies, and are together emplaced within the granulite country rocks. This study employs Landsat 8 multispectral data to map the spatial distribution and extent of the clinopyroxenite dykes. These dykes serve as geological tracers for locating the spatially associated carbonatite bodies. In addition, the present work investigate the spectroscopic properties of REE-bearing carbonatites and clinopyroxenite. Petrography, Raman spectroscopy of minerals, and spectroradiometric measurements of rock samples support the interpretations derived from Principal Component Analysis (PCA), Spectral Angle Mapper (SAM), Support Vector Machine (SVM), Decision Tree, and Random Forest algorithms, thereby aiding in the identification of lithological variations and potential clinopyroxenite occurrences. Carbonatite-II shows more prominent REE absorption features compared to carbonatite-I. This is consistent with petrographic observations and Raman spectroscopy, which show that the REE mineralization in carbonatite-II is dominated by monazite-(Ce) and hydroxylbastnäsite-(Ce), whereas carbonatite-I contains allanite-(Ce) as the primary REE-bearing phase. This study exhibits the efficacy of Landsat series data and non-destructive spectroscopic methods for preliminary mineral exploration and evaluating REE potential before detailed field investigation.</div></div>","PeriodicalId":33804,"journal":{"name":"Applied Computing and Geosciences","volume":"27 ","pages":"Article 100269"},"PeriodicalIF":2.6,"publicationDate":"2025-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144711663","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 lightweight knowledge graph-driven question answering system for field-based mineral resource survey 面向矿产资源实地调查的轻量级知识图驱动问答系统

IF 2.6 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Applied Computing and Geosciences

Pub Date : 2025-07-23 DOI: 10.1016/j.acags.2025.100268

Mingguo Wang , Chengbin Wang , Jianguo Chen , Bo Wang , Wei Wang , Xiaogang Ma , Jiangtao Ren , Zichen Li , Yicai Ye , Jiakai Zhang , Yue Wang

Geoscience data associated with mineral resource surveys have become essential digital assets for governments and mining companies. The rapid increase in the volume of geoscience data makes it challenging to acquire knowledge quickly. In this study, we proposed and built a workflow that employs knowledge graph techniques, deep learning, question templates, and matching algorithms to provide a lightweight question-answering service for field-based geologists involved in mineral resource surveys. Initially, we utilized deep-learning-based geological entities and their semantic relation recognition, along with relational data mapping, to construct the mineral resource survey knowledge graph based on the ontology model. We then employed question template matching, a geological entity recognition model, and a sentence transformer to determine the optimal question template and generate a query statement for knowledge acquisition from a knowledge graph based on the Cypher language. Subsequently, we utilized a subgraph and a short abstract to express the results. The comparison with large language models and retrieval-augmented generation indicates that our solution is suitable for field-based mineral source surveys in a poor network environment with low-performance devices, data privacy concerns, and narrowly focused topics. The results also suggest that further studies on geoscience pre-trained models, an informative library of question templates, and multimodal knowledge graphs are necessary to improve the performance of the knowledge graph-driven question-answering system.

与矿产资源调查相关的地球科学数据已成为政府和矿业公司必不可少的数字资产。地球科学数据量的快速增长使快速获取知识成为一项挑战。在这项研究中，我们提出并构建了一个工作流，该工作流采用知识图谱技术、深度学习、问题模板和匹配算法，为参与矿产资源调查的现场地质学家提供轻量级的问答服务。首先，利用基于深度学习的地质实体及其语义关系识别，结合关系数据映射，构建基于本体模型的矿产资源调查知识图谱。然后，我们采用问题模板匹配、地质实体识别模型和句子转换来确定最优问题模板，并生成基于Cypher语言的知识图知识获取查询语句。随后，我们利用一个子图和一个简短的摘要来表达结果。与大型语言模型和检索增强生成的比较表明，我们的解决方案适用于在具有低性能设备、数据隐私问题和狭窄主题的恶劣网络环境中的基于现场的矿产资源调查。研究结果还表明，为了提高知识图驱动问答系统的性能，需要进一步研究地球科学预训练模型、信息丰富的问题模板库和多模态知识图。

{"title":"A lightweight knowledge graph-driven question answering system for field-based mineral resource survey","authors":"Mingguo Wang , Chengbin Wang , Jianguo Chen , Bo Wang , Wei Wang , Xiaogang Ma , Jiangtao Ren , Zichen Li , Yicai Ye , Jiakai Zhang , Yue Wang","doi":"10.1016/j.acags.2025.100268","DOIUrl":"10.1016/j.acags.2025.100268","url":null,"abstract":"<div><div>Geoscience data associated with mineral resource surveys have become essential digital assets for governments and mining companies. The rapid increase in the volume of geoscience data makes it challenging to acquire knowledge quickly. In this study, we proposed and built a workflow that employs knowledge graph techniques, deep learning, question templates, and matching algorithms to provide a lightweight question-answering service for field-based geologists involved in mineral resource surveys. Initially, we utilized deep-learning-based geological entities and their semantic relation recognition, along with relational data mapping, to construct the mineral resource survey knowledge graph based on the ontology model. We then employed question template matching, a geological entity recognition model, and a sentence transformer to determine the optimal question template and generate a query statement for knowledge acquisition from a knowledge graph based on the Cypher language. Subsequently, we utilized a subgraph and a short abstract to express the results. The comparison with large language models and retrieval-augmented generation indicates that our solution is suitable for field-based mineral source surveys in a poor network environment with low-performance devices, data privacy concerns, and narrowly focused topics. The results also suggest that further studies on geoscience pre-trained models, an informative library of question templates, and multimodal knowledge graphs are necessary to improve the performance of the knowledge graph-driven question-answering system.</div></div>","PeriodicalId":33804,"journal":{"name":"Applied Computing and Geosciences","volume":"27 ","pages":"Article 100268"},"PeriodicalIF":2.6,"publicationDate":"2025-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144711662","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