基于ETM^+的遥感蚀变信息提取方法对比研究

A Comparative Study of Extraction Methods for Alteration Information Based on ETM^+

为从微观到宏观了解、掌握和运用围岩蚀变的特征与原岩特征,寻找大型矿床或大型矿集区,在总结常用遥感蚀变信息提取模型的基础上,结合主成分分析(PCA)和支持向量机(SVM)算法,建立了基于"掩模处理-主成分分析-支持向量机(MASK-PCA-SVM,MPS)"的遥感蚀变信息提取模型。在新疆哈图地区进行的基于ETM+的不同蚀变信息提取模型应用对比研究表明,使用常用模型提取出的蚀变信息和矿点吻合效果较差,而基于MPS模型提取出的蚀变信息与已知矿点的吻合率可达86.51%;结合地质专家知识、地质图和一个已知异常点的分析结果表明,基于MPS模型的蚀变信息提取精度较高,在研究区地质应用中效果较好,为今后该区的遥感蚀变信息提取工作提供了一种新思路。

In order to understand,grasp and apply the characteristics of altered wall rock and the original rock from microcosmic and macroscopic angles and look for large deposits or large ore districts,the authors summarized the common extraction models of alteration information,combined the principal component analysis（PCA） with support vector machines（SVM） to establish the MASK process-PCA-SVM（MPS） model,applied and comparatively studied these models in Hatu area of Xinjiang,and made use of the two methods to verify the extraction results.The results show that the extracting results of the common models are rather poorly consistent with the known deposits,but the alteration information extracted by MPS model is better concordant with the known mineral deposits and altered belts,with the coincidence rate arriving at 86.51%.The combination of geological experts＇ knowledge with geological maps and a known anomaly spot reveals that the result of MPS model is in accord with the geological fact.Therefore,the extracting precision based on MPS model is higher than that of the common models,and the effect of application is fairly good in the study area.The model proposed in this paper provides a new idea for extracting remote sensing alteration information in the future.