摘要
针对遥感蚀变矿物填图方法多样、影响因素复杂、效果参差不齐等问题,对TM或ETM+数据遥感蚀变矿物填图方法进行剖析,采用比值法、主成分分析法、Crosta法以及基于已知矿点和光谱角法的监督分类等方法,结合在新疆西天山等地的应用实例,认为植被、地形、土壤覆盖等是影响应用效果的主要因素;矿化目标大小与遥感数据像元大小之间的关系很可能导致混合反射率起主导作用,加之矿化蚀变中常存在石英、长石等高反射率矿物,会影响到含羟基与含碳酸根蚀变矿物之特征性诊断谱带的正常体现,从而大幅降低蚀变矿物填图效果;一些数据处理层面上的方法,例如借助"纯净物质"反射率的子像元分类、混合光谱分析、模糊分类等,能够一定程度地缓解问题,但对找矿来说也仅限于地表线索的二维定位。因此,在现有技术水平下,遥感还主要为战略找矿手段,只有综合运用地质、物探、化探、遥感等多种方法,才能避免片面性并增大发现新矿产的可能。
Several methods exist in the alteration mineral mapping using TM or ETM data. Those mainly comprise the ratio imagery, principal component analyses or Crosta technique and the supervised classification based on known altered mineral localities and spectral angle methods. Detailed analyses and lots of applications related to this study show certain problems in the methodology. Firstly, all of them are notably hampered by certain natural factors like vegetation, topography and soil cover. Secondly, mixture pixels and the related mixture reflectance can be readily caused by smaller altered mineral targets and larger remote sensing data pixels, in addition to the common presence of some high reflectance minerals like quartz and feldspar in the altered mineral belt, which may seriously influence normal occurrence of the so-called diagnostic spectral absorption that acts as the basis of all the alteration mineral mapping methods. Lastly, certain data processing-based methods using the reflectance of relevant ＂pure materials＂, such as the suh-pixel classifying, spectral mixture analysis and fuzzy classification, can solve the problem to some extent and even so they are confined to locating the surface altered mineral targets only. Therefore, it is important to employ multi-methods including geology, geophysical and geochemical exploration and remote sensing in mineral resources prospecting, and that will largely reduce a lopsided view resulted from a single method.
出处
《地球科学与环境学报》
CAS
2008年第3期254-259,共6页
Journal of Earth Sciences and Environment
关键词
遥感
蚀变矿物填图
混合像元
多元信息
找矿
remote sensing
alteration mineral mapping
mixture pixel
multi-information
mineral resources prospecting
