遥感异常与地球化学指标间的映射关系模型探索

Exploration of the mapping relationship model between remote sensing anomaly and geochemical index

采用改进的遥感异常提取方法,以甘肃金川地区的LandSat 8遥感影像数据为基础,结合主成份分析方法及低通滤波后期处理,得到空间规律性较强的烃基及铁染异常;进而对测区的土壤地球化学取样分析数据,利用统计学方法寻找不同元素异常(或地球化学指标)的内在关联,并建立土壤地球化学指标比值与遥感异常之间的关系模型。继续构建出基于LandSat 8遥感影像的净反射率值与地球化学因子之间的统计模型,并以此预测模型对测区进行了遥感地球化学异常的圈定。预测结果表明:建立的遥感地球化学统计模型高异常区与已知矿区范围基本吻合,通过提取的遥感地球化学异常不仅能表达化探常规异常的信息,还可以把其空间分布规律明晰出来,让某些局部弱异常得到增强。遥感地球化学异常得到了野外验证,说明利用地质资料建立的基于LandSat 8遥感地球化学预测模型比单一的遥感或化探异常有更好的预测效果。

Through an improved remote sensing anomaly extraction method,on the basis of LandSat 8 remote sensing image data in Jinchuan area,Gansu,combined with the principal component analysis and low-pass filtering post-processing,the hydroxyl and iron-stained anomalies with strong spatial regularity are obtained.The statistical methods are used to process the analysis data of soil geochemical samples in the survey area in order to find out the internal correlation of different element anomalies(or geochemical index),and to establish the relationship model between soil geochemical index ratios and remote sensing anomalies.Then a statistical model is built on the basis of the net reflectance value of LandSat 8 remote sensing images and geochemical factors.Based on this prediction model,remote sensing geochemical anomalies are delineated in the survey area.The prediction result shows that the high anomaly area of the established remote sensing geochemical statistical model is basically consistent with the range of known mining area.The extracted remote sensing geochemical anomalies can not only express the information of conventional geochemical exploration anomaly,but also their spatial distribution regularity is clarified out to enhance some local weak anomalies.The remote sensing geochemical anomaly has been verified in the field.It is indicated that the LandSat 8 remote sensing geochemical prediction model established on the use of geological data has a better prediction effect than the single remote sensing anomaly or geochemical anomaly.