基于Hyperion的新疆罗布泊地区蚀变提取与应用

发布时间：2018-02-12 15:24

本文关键词： 罗布泊 Hyperion影像蚀变提取矿物填图　出处：《东华理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：利用星载Hyperion遥感数据找矿主要是利用了图像处理等技术挖掘出有关矿化的蚀变信息,从而圈定或预测有可能成矿的地区和范围,这是遥感技术与地质学的完美结合。本研究区位于若羌县的罗布泊镇东部地区,岩矿出露程度好、切割深度大、植被覆盖度少和大型矿床多等特点,使其成为很好的Hyperion遥感数据找矿示范区域。本文的研究内容主要分为三个部分:基于主成分分析法(PCA)和波段比值法(Band Ratio)提取铁染和羟基蚀变异常信息;利用光谱角填图法(SAM)获取研究区内的成矿分布图;结合研究区内的成矿分布图分别与铁染、羟基蚀变信息和研究区内的地质图进行分析对比。首先分析了罗布泊研究区域内的历史、地理和地质背景,概括了研究区的地层、构造、矿产和蚀变情况,大致确定了蚀变矿物的种类。根据岩石和矿物在地质热液的作用下产生矿化蚀变的原理,结合矿物蚀变后产生的粘土矿物、铁离子矿物和碳酸根离子矿物的光谱特性对研究区内的Hyperion影像数据进行矿物蚀变异常信息的提取。在此基础上对Hyperion L1R级原始影像进行预处理如未定标及水汽影响波段去除、绝对辐射值转换、坏线修复、条纹去除、Smile效应修复、大气校正和几何校正等。利用主成分分析法和波段比值法对预处理后的Hyperion影像进行铁染矿物和羟基矿物蚀变异常信息的提取。获取研究区内矿物分布图需对预处理后的Hyperion影像进行光谱角填图(SAM),在N维可视化的数据结果中获得了五类端元矿物,最后利用光谱角填图法(SAM)通过分析矿物组合,最终确定了绢云母、赤铁矿、中长石、金云母和电气石五种矿物并最终生成研究区的成矿信息图。研究区曾多次出现大型或超大型矿床,根据研究区内这一地质情况,在查阅相关地质文献资料,尤其是研究区内的成矿分布状况,结合本文提取的蚀变信息和成矿填图。实验结果表明,提取的铁染矿物、含羟基离子矿物和成矿填图信息与该研究区地层构造以及矿点有比较好的吻合,对利用Hyperion影像数据预测蚀变信息和成矿填图有一定的实用价值,能有效的预测其它地区有利成矿靶区。
[Abstract]:The use of space-borne Hyperion remote sensing data for ore prospecting is mainly to mine alteration information about mineralization by using image processing and other techniques, so as to delineate or predict the area and range of possible mineralization. This is a perfect combination of remote sensing and geology. The study area is located in the eastern part of Luoppo Town, Ruoqiang County. The research content of this paper is divided into three parts: extracting abnormal information of iron stain and hydroxyl alteration based on principal component analysis (PCA) and band ratio method (Ratio); Using spectral angle mapping (SAM) to obtain the metallogenic distribution map of the study area, combining the metallogenic distribution map of the study area with iron staining, The hydroxy alteration information is analyzed and compared with the geological map of the study area. Firstly, the historical, geographical and geological background of the Lobop study area is analyzed, and the stratigraphic, structural, mineral and altered conditions in the study area are summarized. The types of altered minerals have been roughly determined. According to the principle of mineralized alteration of rocks and minerals under the action of geological hydrothermal solution, combined with clay minerals produced by mineral alteration, The spectral characteristics of iron ion minerals and carbonate ion minerals are used to extract the abnormal information of mineral alteration from Hyperion image data in the study area. On this basis, the original Hyperion L1R images are preprocessed such as uncalibrated and water vapor wave bands removed. Conversion of absolute radiation value, repair of bad lines, restoration of stripe removal effect, Atmospheric correction and geometric correction, etc. Using principal component analysis and band ratio method to extract iron-stained mineral and hydroxyl mineral alteration anomaly information from pretreated Hyperion image, the distribution map of minerals in the study area needs to be prepositioned. Spectral angle mapping was carried out on the processed Hyperion image, and five kinds of end-element minerals were obtained from the N-dimensional visualization data. Finally, through the analysis of mineral assemblage, the sericite, hematite and meso-feldspar were determined by using the spectral angle mapping method (SAM). Five kinds of minerals, phlogopite and tourmaline, were formed and formed in the study area. Large or super large ore deposits were found in the study area many times. According to this geological situation in the study area, we are consulting the relevant geological documents. In particular, the metallogenic distribution in the study area, combined with the alteration information and metallogenic map extracted in this paper, shows that the extracted iron stained minerals are extracted. The information of hydroxyl ion mineral and ore-forming mapping is in good agreement with the stratigraphic structure and ore spot of the study area, and it is of certain practical value for predicting alteration information and metallogenic mapping by using Hyperion image data. It can effectively predict the favorable ore-forming target in other areas.
【学位授予单位】：东华理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P627

【相似文献】