比值合成和特征主成分选择技术在提取油区信息中的应用

The Application of Band Ratio Compsite And Feature-oriented Principal Component Selection in the Extraction of Oil-gas Signature

根据烃类微渗漏蚀变物质的波谱特性，在ＴＭ影像上研究了三种计算机提取油气信息的方法：（１）比值合成法；（２）六波段特征主成分选择→ＩＨＳ逆变换→彩色合成；（３）四波段特征主成分选择→ＩＨＳ逆变换→彩色合成。利用这些方法有效地提取了油气信息，取得了较好的实际效果。

After having fully studied the spectral properties of typical hydrocarbon alternationminerals,and the physical geographical features of a certain oil area in East China,the au-thors present the research of direct oil exploration by extracting hydrocarbon microseepagesignatures with TM data,In this paper,three function-combined image process methodsare introduced,and oil-gas signatures have been successfully extracted with them,whichis also supported by the material of geochemical exploration in the same area.