摘要
高光谱遥感探测技术已成为探测油气藏的前沿新技术之一.研究以油气微渗漏地表共生异常理论为基础,采用基于小波主成份分析(principal component analysis,PCA)最大似然分类、端元提取分类、光谱库典型蚀变光谱分类和植被指数决策树分类方法,对榆林典型稀疏植被地区的进行油气勘探,提取了与烃异常相关的粘土、碳酸盐、植被异常等相关的专题信息产品,得出综合异常区图.对照分析已知气井与油气异常区分布,证明了油气微渗漏信息的提取与识别方法的有效性.
Hyperspectral remote sensing technology is new in oil and gas exploration. Based on the theory of abnormal surface symbiosis with oil and gas microseepage, four typical classifications, maximum likelihood classification in the wavelet-based principal component analysis, endmember extraction, typical alteration classification with spectral libraries and decision tree classification based on vegetation indices are chosen to carry a case study of Hyperion images in Yulin, to obtain the related thematic maps such as clay, carbonate, vegetation and determine six comprehensive anomalous areas. A comprehensive analysis of the distribution of existing gas well and oil-gas anomalous areas show that the information extraction method of oil and gas microseepage is valid.
出处
《地球科学（中国地质大学学报）》
EI
CAS
CSCD
北大核心
2015年第8期1301-1309,共9页
Earth Science-Journal of China University of Geosciences
基金
国家高技术研究发展计划(863计划)项目(Nos.2008AA121100
2012AA12A308)
国家自然科学基金项目(No.41402293)
关键词
高光谱遥感
油气探测
小波PCA
端元提取
光谱库
植被指数
遥感
决策树分类.
hyperspectral remote sensing
oil and gas exploration
wavelet-based PCA
endmember extraction
spectral library
vegetation index
remote sensing
decision tree classification.
