机器学习与成矿预测:以闽西南铁多金属矿预测为例

作为近年来爆炸式发展的方法模型,机器学习为地质找矿提供了新的思维和研究方法.本文探讨矿产预测研究的理论方法体系,总结机器学习在矿产预测领域的特征信息提取和信息综合集成两个方面的应用现状,并讨论机器学习在矿产资源定量预测领域面临的训练样本稀少且不均衡、模型训练中缺乏不确定性评估、缺少反哺研究、方法选择等困难和挑战.进一步...     

基于深度学习的智能矿物识别方法研究

矿物识别在许多研究领域都有着重要作用,基于深度学习技术的智能矿物识别为这些领域带来了新的发展方向,不仅能有效节省人工成本,还能减小识别错误。针对石英、角闪石、黑云母、石榴石和橄榄石共5种矿物进行实验,提出了一种准确高效的智能矿物识别方法。实验采用图像分析常用的卷积神经网络建立模型,设计出一套基于残差神经网络的矿物识别方法。本实验独立采集了5种矿物的偏光显微图像数据集,用于模型的训练、验证和测试,并通过合理的数据增强策略来扩充训练数据集。在卷积神经网络的结构设计上,选取了ResNet-18作为框架,最终于模型测试中取得89%的准确率,成功训练出一个较为精准的矿物识别模型,实现了基于深度学习的智能矿物识别方法。     

Determination and application of the weights for landslide susceptibility mapping using an artificial neural network

The purpose of this study is the development, application, and assessment of probability and artificial neural network methods for assessing landslide susceptibility in a chosen study area. As the basic analysis tool, a Geographic Information System (GIS) was used for spatial data management and manipulation. Landslide locations and landslide-related factors such as slope, curvature, soil texture, soil drainage, effective thickness, wood type, and wood diameter were used for analyzing landslide susceptibility. A probability method was used for calculating the rating of the relative importance of each factor class to landslide occurrence. For calculating the weight of the relative importance of each factor to landslide occurrence, an artificial neural network method was developed. Using these methods, the landslide susceptibility index (LSI) was calculated using the rating and weight, and a landslide susceptibility map was produced using the index. The results of the landslide susceptibility analysis, with and without weights, were confirmed from comparison with the landslide location data. The comparison result with weighting was better than the results without weighting. The calculated weight and rating can be used to landslide susceptibility mapping.     

Controls on mineral deposit occurrence inferred from analysis of their spatial pattern and spatial association with geological features

In this paper, point pattern analysis, fractal analysis and Fry analysis were employed to study the spatial pattern of known occurrences of mineral deposits of the type sought, whereas distance distribution method was applied to study the spatial associations between various geological features and known occurrences of mineral deposits of the type sought. In the Aroroy district (Philippines), the results of the applications of these spatial analytical techniques support a conceptual model of district-scale mechanism of geologic controls on low-sulphidation epithermal Au mineralization, which involves a more-or-less regular mesh of interlinked zones of extension faults/fractures at and/or around intersections of NNW- and NW-trending strike-slip faults/fractures. Integration of spatial evidential data layers representing these structural controls and surficial geochemical anomalies, via knowledge-guided application of data-driven evidential belief functions, results in delineation of prospective areas occupying about 25% of the district, in which there is about 70% likelihood of undiscovered occurrences of low-sulphidation epithermal Au deposits.     

Unsupervised ore/waste classification on open-cut mine faces using close-range hyperspectral data

The remote mapping of minerals and discrimination of ore and waste on surfaces are important tasks for geological applications such as those in mining. Such tasks have become possible using ground-based, close-range hyperspectral sensors which can remotely measure the reflectance properties of the environment with high spatial and spectral resolution. However, autonomous mapping of mineral spectra measured on an open-cut mine face remains a challenging problem due to the subtleness of differences in spectral absorption features between mineral and rock classes as well as variability in the illumination of the scene. An additional layer of difficulty arises when there is no annotated data available to train a supervised learning algorithm. A pipeline for unsupervised mapping of spectra on a mine face is proposed which draws from several recent advances in the hyperspectral machine learning literature. The proposed pipeline brings together unsupervised and self-supervised algorithms in a unified system to map minerals on a mine face without the need for human-annotated training data. The pipeline is evaluated with a hyperspectral image dataset of an open-cut mine face comprising mineral ore martite and non-mineralised shale. The combined system is shown to produce a superior map to its constituent algorithms, and the consistency of its mapping capability is demonstrated using data acquired at two different times of day.     

岩矿高光谱遥感及其在青藏高原的应用前景

高光谱成像作为目前遥感领域最先进的技术,在地质应用中取得了巨大成功。岩石和矿物由于电子过程和分子振动可以产生特征的光谱吸收,因此可以利用高光谱技术进行岩矿填图,快速且准确地获取区域内岩石和矿物的分布情况,进而圈定有潜力的找矿靶区。岩矿高光谱通过对岩矿信息的提取可获得矿物类别和矿物丰度分布甚至矿物化学成分的填图结果,识别方法包括光谱匹配、模式识别和人工智能方法3大类。GF-5卫星上搭载的高光谱成像仪覆盖度宽、光谱分辨率和信噪比高,满足矿物精细识别和大比例尺、大面积岩性填图的要求,应用前景广阔。青藏高原地区矿产资源丰富、岩体裸露但地势偏远,有利于高光谱遥感蚀变矿物填图,开展区域找矿工作。同时,高光谱遥感凭借其空间尺度优势,可以方便快捷地获取区域整体信息,结合地球化学的时间信息,可为青藏高原关键地质科学问题的解决提供新的视角。提升岩矿高光谱遥感的定量化和智能化水平,将传统地质学方法获得的地下深部资料与遥感手段获取的地表数据结合,是促进地质遥感和地球系统科学发展的重要途径。     

矿产地质专题调查试点总体构想

矿产地质专题调查是服务于找矿勘查的基础性、公益性地质调查工作,是以问题和需求为导向的矿产地质填图,是矿产地质综合填图的升级版,是当前形势下,地质调查工作方式转变的新探索,不仅对实现地调科研一体化具有重要意义,而且对于地质工作,特别是填图工作可持续发展具有重大战略意义。如何开展专题填图,目前没有现成的指南或技术规范,因此需要试点、探索和创新,总结经验,开展示范,最终提出专题矿调工作细则。从分析目前综合性矿产地质工作存在的不足入手,对矿产地质专题调查的定位、目标、总体思路、技术路线等提出了初步思考。     

地质领域机器学习、深度学习及实现语言

地质大数据正在以指数形式增长。只有发展智能数据处理方法才有可能追上大数据的超常增长。机器学习是人工智能的核心,是使计算机具有智能的根本途径。机器学习已成为地质大数据研究的前沿热点,它将让地质大数据插上翅膀,并因此改变地质。机器学习是一个源于数据的模型的训练过程,最终给出一个面向某种性能度量的决策。深度学习是机器学习研究中的一个重要子类,它通过构建具有很多隐层的机器学习模型和海量的训练数据,来学习更有用的特征,从而最终提升分类或预测的准确性。卷积神经网络算法是最为常用的一种深度学习算法之一,它广泛用于图像识别和语音分析等。Python语言在科学领域的地位占据着越来越重要。其下的Scikit-Learn是一个机器学习相关的库,提供有数据预处理、分类、回归、聚类、预测、模型分析等算法。Keras是一个基于Theano/Tensorflow的深度学习库,可以应用来搭建简洁的人工神经网络。     

基于MatLab的神经网络在江苏矿山地质环境评估中的应用

神经网络作为一种新的方法体系,具有分布并行处理、非线性映射、自适应学习和鲁棒容错等特性,在模式识别、控制优化和智能信息处理等方面有着广泛的应用。利用MatLab的神经网络工具箱,建立了江苏矿山地质环境质量的评估模型,评估结果经过实际验证,具有较高的可信度和实用性。     

基于AlexNet网络的二维找矿预测——以松桃—花垣地区沉积型锰矿为例

在大数据的时代背景下,地质大数据逐渐趋于复杂化的模式与其间的空间关联性为基于机器学习算法的矿产资源定量预测带来了更大的挑战。利用深度卷积网络算法优异的分析性能来提取不同成矿条件下多种二维要素图层的空间分布特征与关联性是一项非常有意义的探索性实验。以松桃—花垣地区沉积型锰矿为例,利用深度卷积神经网络模型AlexNet挖掘Mn元素、沉积相、大塘坡组出露、断裂及水系的空间分布与锰矿矿床的就位空间的耦合相关性,以及不同的控矿要素之间的相关性,以此训练出二维矿产预测分类模型。经过训练后,可以得到验证准确率88.89%,召回率为66.67%,损失值0.08的深度卷积神经网络分类模型。应用该模型对未知区进行二维成矿预测,共圈定出91、96、154、184号4个成矿远景区,其中91号和154号的区域含矿概率为1,96号含矿概率为0.5。由此可见,预测区具有很大概率存在尚未发现的矿床。     

人工神经网络在徐州岩溶地面塌陷评价中的应用

岩溶地面塌陷的影响因素很多,发展过程也复杂。在众多的对岩溶地面塌陷的评价方法中,神经网络具有自学习、自适应与高度非线性映射的特点,是一种非常有效的评价手段。在徐州岩溶石地面塌陷的评价中,成功地运用了人工神经网络技术,它具有的强大非线性映射能力,能够建立评价因子和评价对象之间的关系,正确选取评价因子,避免主观判断取值,从而得出可靠的预测模型和岩溶塌陷危险性分区图。     

新疆西天山式可布台铁矿地质、矿物化学和S同位素特征及其对矿床成因的约束

新疆西天山式可布台铁矿发育于伊犁裂谷内,赋存于上石炭统中酸性火山碎屑岩、浅变质片岩、千枚岩中,矿体呈层状、似层状以及透镜状顺层产出。金属矿物以赤铁矿、镜铁矿为主,含少量黄铁矿、菱铁矿;脉石矿物主要为碧玉、重晶石、石英以及少量方解石。矿石构造以条带状、纹层状和块状为主,矿物结构多为隐晶质结构、半自形结构以及充填结构。矿床分为4个成矿阶段,即黄铁矿-赤铁矿-铁碧玉-重晶石阶段、菱铁矿-软锰矿阶段、石英-镜铁矿阶段、氧化物阶段。矿体顺层产出和发育纹层状矿石构造指示矿床为沉积成因。电子探针分析显示:(1)块状赤铁矿Al2O3、Na2O、MgO、SiO2含量相对分散,推测这可能与块状矿石快速沉淀结晶有关,暗示了剧烈的流体喷流活动,而纹层状和条带状赤铁矿Al2O3、Na2O、MgO、Si O2含量相对集中则反映平静的沉积环境以及微弱的喷流活动,两者的比较可能暗示了成矿过程中流体喷溢速率以及沉积环境都不断改变;(2)黄铁矿中含有较高的Co、Ni比,显示其形成与火山作用关系密切;(3)菱铁矿的FeOT与Mn O+MgO含量呈负相关关系,并形成两个聚集区,与镜下其具有不同特征相吻合,可能暗示了成矿后期菱铁矿随热液析出时候发生了分异作用。黄铁矿(δ34S=-6.1‰~6.5‰)和重晶石(δ34S=12.9‰)硫同位素组成显示曾发生过硫酸盐和硫化物之间的硫同位素分馏作用,成矿热液的硫可能来源于岩浆硫。综合分析认为,式可布台铁矿可能为海相火山喷流沉积型铁矿床。     

对数字测绘数据传输的网络安全与数据文件自动更新的思考

本文结合实例介绍了数字测绘科学数据在局域网与互联网上的网络安全问题,包括加密、传输、防病毒、防黑以及局域网内基于数据存储服务器的数据文件自动备份更新的方法.     

Effects of strong ground motion on the susceptibility of gully type debris flows

Typhoon Herb in 1996 caused widespread debris flows in central Taiwan. The 7.3 Chi-Chi earthquake on September 21, 1999, which also took place in central Taiwan, induced many landslides in the region. These landslides turned into debris flows when Typhoon Toraji struck Taiwan in 2001. This research selects three regions which suffered a ground motion class of 5, 6, and 7 on the Richter scale during the Chi-Chi earthquake as study areas. Air photos from 1997 and 2001 of these regions are used to map the gully-type debris flows that took place after Typhoons Herb and Toraji, respectively. The gullies adjacent to the debris flow, but without a trace of debris flows, are also mapped as the non-debris flow data. The topography, hydrogeology, and rainfall factors – where debris flow occurred and in which there was no occurrence of debris flows in these gullies were retrieved from DTM, geological maps, and iso-countour maps, and of rainfall through GIS processing. These characteristic are introduced into a probabilistic neural network to build a predicting model for the probability of the occurrence of debris flows. Three series of cross analyses are conducted to compare the probability of the occurrence of debris flows of the same dataset predicted by different prediction models. The results reveal that the susceptibility of debris flows was elevated after the Chi-Chi earthquake struck. The upsurge of susceptibility was more obvious for the regions that received a higher class of ground motion.     

基于BP神经网络与数学形态学的彩色地质图面要素信息智能提取

地质图是一个区域地质研究的重要成果,也是前人留下的宝贵资料,更融合了地质专家的丰富知识。本研究的目的是通过新的思路将彩色地质图信息提取出来,使其结果能直接进行数据分析,并用于决策和分析。以机器学习为指导,在分析半结构化（栅格）地质图特征的基础上,根据图例信息,提出一种彩色地质图信息提取新思路,对彩色地质图进行分层信息提取,并结合数学形态学和多层前向反馈式神经网络,探索半结构化数据转换为结构数据的有效技术方法。利用图像信息提取技术将半结构化地质图转化为结构化数据,可用于成矿预测等研究。这一变化将改变传统地质数据的结构,地质研究的信息基础和来源将会增加,对于获取更多的数据源和信息源,进一步开展地质分析研究具有重要意义。     

基于多尺度区域卷积神经网络的露头孔洞自动提取

碳酸盐岩储层解释与评价的一个重要内容是确定其孔隙空间特征,野外露头是地下储层的真实刻画,对野外露头进行孔洞的自动化提取与其参数定量表征具有重要意义。在深度学习Mask-RCNN模型基础上进行改进,提出一种基于多尺度的区域卷积神经网络孔洞检测新方法,并通过两种方式进行准确度分析,第一种是将该方法的孔洞提取结果与OSTU分割法、分水岭分割法、BP神经网络法、支持向量机法以及Mask-RCNN的孔洞提取结果进行比较,其结果显示该方法有更高的检测准确度;第二种是通过该方法提取的孔洞结果计算洞数量、面孔率和洞面积均值三个孔洞特征参数,以人工提取结果为参照,比较得到洞数量准确度在88%以上,面孔率准确度在93%以上,洞面积均值准确度在93%以上。最后将提出的方法应用于峨边先锋灯二段的数字露头剖面孔洞自动识别,并分层计算孔洞参数,定量分析其分布特征,为该露头碳酸盐岩储层评价提供了依据。     

深度学习在地质学上的应用

自Hinton等使用基于卷积神经网络的深度学习模型赢得Image Net分类比赛以来,深度学习的研究席卷了各个行业。通过介绍深度学习的历史,探索国内地质行业中深度学习模型的使用情况,并介绍深度学习的基础概念(如神经元、神经网络、监督学习和无监督学习等)以及深度学习基础模型中的2个重要网络:深度信念网络(DBN)和卷积神经网络(CNN)。在此基础上,类比深度学习在医学等相关领域的应用,提出了深度学习在地质上的几点应用:利用深度学习在计算机视觉上表现出的强大能力,可以对遥感图像进行聚类、对岩石样品图像进行分类、对岩石薄片数据进行描述;利用深度学习对原始数据表现出的强大识别能力,处理地质异常数据,从而确定成矿靶区的可能位置;利用深度学习的特点,对地震前的声信号数据进行处理,从而判断出地震发生前的剩余时间。     

大数据挖掘与智能预测找矿靶区实验研究——卷积神经网络模型的应用

成矿预测需要通过一定的规则集合将专家观点、地质背景、成矿类型等因素进行综合考虑。但由于受到人类实际计算能力的生物条件限制,影响找矿预测成果的最大因素是找矿者的经验知识。随着大数据时代的到来,成矿预测可充分利用数学计算,即以特定规则对成矿系统进行计算,以概率表示成矿前景。依靠计算机的超级运算能力,结合机器学习的方法技术,可以对地质大数据进行成矿预测特征学习,实现对众多地质变量与矿体相关性之间的验证,从而进行预测。本文以安徽东至兆吉口铅锌矿床为例,示范如何通过机器学习的卷积神经网络方法,学习元素Zn在地表的分布特征与矿体在地下空间就位的耦合关系,并圈定靶区。经过450次训练后,得到了准确率95%,损失率14%的CNN模型,并成功实现智能圈定3块找矿靶区。这种神经网络模型可能表达了矿体在地下就位时元素在地表分布的响应,可以用来进行找矿勘查并圈定靶区。     

基于卷积神经网络的福建省区域滑坡灾害预警模型

福建省滑坡灾害频发,开展区域尺度上的滑坡灾害预警是防灾减灾的重要手段,但由于滑坡成灾机理复杂,传统的区域滑坡预警方法存在精度不足等问题。深度学习是指通过构建神经网络模型进行特征的提取、抽象、表示与学习的技术,是机器学习的一种。卷积神经网络作为一种经典的深度学习算法,具有比传统机器学习更强大的分类能力与表征能力。文章以福建省为研究区,将卷积神经网络引入滑坡灾害预警领域,构建福建省区域滑坡预警模型,过程及结果如下：（1）采用SMOTE优化算法对2010—2018年福建省滑坡灾害样本库进行优化,扩充正样本的个数,将正负样本比例从1∶3.4扩充到1∶2,样本总量达到18040个;（2）构建卷积神经网络模型结构,模型结构包括一个输入层、两个卷积层、两个最大池化层和一个全连接层以及一个输出层;（3）使用卷积神经网络对优化后的样本（2010—2018年样本的80%作为训练集）进行训练,并用贝叶斯优化算法优化模型超参数,得到福建省区域滑坡预警模型;（4）以2010—2018年样本的20%作为测试集对模型进行测试,采用混淆矩阵、ROC曲线进行模型测试,结果显示模型准确度为0.96～0.97,AUC值达到0.977,模型精度与泛化能力良好;（5）以2019年汛期滑坡灾害实况作为正样本,通过时空采样的方法采集负样本,构建2019年区域滑坡样本校验集（样本数603个）,对模型进行进一步实况校验,采用混淆矩阵、ROC曲线进行模型校验,结果显示模型准确度为0.75～0.85,AUC值为0.852。虽然仅用了2019年汛期的滑坡实况样本进行校验,但也达到较好的效果。将卷积神经网络算法应用到区域滑坡预警中,为建立区域滑坡预警模型提供了一种新的途径,初步校验表明,模型效果良好,今后将在福建省对模型进行进一步的应用与校验。     

Advanced prediction of tunnel boring machine performance based on big data

Predicting the performance of a tunneling boring machine is vitally important to avoid any possible accidents during tunneling boring.The prediction is not straightforward due to the uncertain geological conditions and the complex rock-machine interactions.Based on the big data obtained from the 72.1 km long tunnel in the Yin-Song Diversion Project in China,this study developed a machine learning model to predict the TBM performance in a real-time manner.The total thrust and the cutterhead torque during a stable period in a boring cycle was predicted in advance by using the machine-returned parameters in the rising period.A long short-term memory model was developed and its accuracy was evaluated.The results show that the variation in the total thrust and cutterhead torque with various geological conditions can be well reflected by the proposed model.This real-time predication shows superior performance than the classical theoretical model in which only a single value can be obtained based on the single measurement of the rock properties.To improve the accuracy of the model a filtering process was proposed.Results indicate that filtering the unnecessary parameters can enhance both the accuracy and the computational efficiency.Finally,the data deficiency was discussed by assuming a parameter was missing.It is found that the missing of a key parameter can significantly reduce the accuracy of the model,while the supplement of a parameter that highly-correlated with the missing one can improve the prediction.