基于主成分分析的植被掩模与模式滤波方法在中等植被区蚀变信息提取中的应用  

A New Method for Alteration Extraction in Mid-vegetated Areas Using TM Data -Combined Technique of Vegetation Masking and Mode Filtering (CTVMMF) Based on Principal Component Analysis(PCA)

作　　者：张晋开[1] 崔承禹[1] 支毅乔[1] 

机构地区：[1]中国科学院遥感应用研究所

出　　处：《中国图象图形学报（A辑）》1996年第2期108-114,共7页Journal of Image and Graphics

摘　　要：植被干扰的消除对于应用ＴＭ图象进行蚀变信息提取是非常重要的。本文从矿物、岩石和植被在ＴＭ波段内的波谱响应分析入手，选择了ＴＭ５和ＴＭ１，ＴＭ５和ＴＭ７，ＴＭ４和ＴＭ３这三个组合分别来增强含铁矿物（主要是铁氧化物）、含羟基矿物和碳酸盐矿物及植被信息。通过对比值、比值合成以及主成分分析等方法的比较，提出了一种适合于中等植被覆盖区蚀变信息提取的新方法，即基于主成分分析的植被掩模和模式滤波相结合的方法。此方法在试验区进行蚀变信息提取中取得良好效果，它不仅识别出了与地表金矿化区一致的蚀变矿物富集区，而且对于土石矿区的一些矿脉也有较好反映。Reducing vegetation effects is very important for alteration extraction using TM Image Based on the spectral responds of minerals and vegetation in TM bands, TM5 and TM7 are selected for OH-bearing mineral spectral feature enhancement, TM5 and TM1 for iron oxides, TM4 and TM3 for vegetation in this paper. After comparing ratio composition, various PCA methods and other techniques for alteration extraction, CTVMMF is proposed This method has been successfully employed for the extration of alteration information in test areas. Most of alteration belts derived form TM images are identified with gold mineralization areas and some linear alterations extended on the ground surface along the gold veins.

关 键 词：矿床勘探 TM图象 蚀变信息提取 图象处理 主成分分析 植被掩模 遥感 模式滤波 围岩蚀变 中等植被覆盖区 

分 类 号：P627[天文地球—地质矿产勘探] TP751[天文地球—地质学]