基于知识的蚀变遥感异常信息快速提取及找矿应用实践

Method of Fast Extracting Alteration Remote Sensing Abnormal Information Based on Knowledge and Prospecting Application

本文以新疆卡拉先格尔成矿带内新发现的哈腊苏铜矿化区为例,进行ETM+矿化蚀变遥感异常信息快速提取研究。分别在ETM+1,ETM+2,ETM+3,ETM+4,ETM+5和ETM+7等6个波段图像和增强图像(比值图像、主成分分析图像)上已知蚀变岩分布区及地质背景进行像元光谱采样,并分析蚀变岩在ETM+的6个波段图像及铁染异常和羟基异常图像上的光谱特征,进而发现蚀变岩与矿区地层区分的光谱知识。利用矿区成矿地质特征和典型矿床成矿模型建立遥感地质找矿模型,发现遥感地质控矿特征。在此基础上建立了基于知识规则的在ETM+图像上提取蚀变遥感异常信息模型,利用ERDAS IMAGING V8.7图像处理系统中的专家分类器模块快速提取该区蚀变遥感异常信息,实现了地质与遥感在图像处理平台上的有机结合。经实际验证,其结果与实际情况基本吻合,并且新发现铜矿化带一处,可为与该区地质背景类似的新疆阿勒泰南缘开展铜矿普查勘探提供参考依据。

The new discovered copper mineralized zone in Halasu region, Xinjiang kalaxiange mineralization belt,was selected as the case study in order to fast extract ETM+ mineralization alteration remote sensing anomaly.First of all,the paper discussed the geological background,geophysical and geochemical anomaly characteristics of this area,aggregated metallogenic regularities,then,compared the features of remote sensing data.Though choosing ETM data sources,we achieved geological,geophysical,geochemical data processing and remote sensing image enhancement processing.Spectrum characteristics of known alteration distribution area and geologic setting,six(bands band 1—5 and 7) of alteration rock of ETM+ image and the enhanced image(ratio image,the principal component analysis image),iron-stained anomaly and hydroxyl anomaly were analyzed.Spectrum knowledge to differentiate alteration rock from mining region strata was discovered.By introducing the standard error based on TM4<120,we known that the brightness value of phyllic rock was 75.6<TM5/7<87.3,the propvlitization altered rock 64.2<PC4-H<73.42.By the use of the mineralization geological feature of mining region and typical ore deposit mineralization model,we could establish the remote sensing geology mineral prospecting model,and then,the remote sensing geology ore-control characteristic was investigated.The middle-bottom basalt and augitophyre,in the middle Devonian Beita mountain group,were the main rock of the porphyry mineralization within the copper(or molybdenum) belt.As a result,the abnormal information model of extracting alteration remote sensing based on ETM+ image has been developed.Additionaly,the alteration remote sensing abnormal information could be extracted by the expert classifier model of ERDAS IMAGING V8.7.Based on the result we can combine geology with remote sensing image processing.Although the establishment of the regulation was manual,whereas at present,the geological expert system theory was not yet ripe,such model in this field has opened up a new development path.Through actual verification,it proved that the results had good correspondence with the actual condition,and the copper mineralization belt had been discovered.This model provides reference to geological investigation and prospecting for copper mines in those area with the similar geologic background to the southern margin of Kalaxiange region,in Xinjiang.