应用ASTER遥感图像的岩矿信息提取研究

Research on Rock and Mineral Information Extraction Based on ASTER Remote Sensing Image

高级星载热发射反射辐射计(Advanced Spaceborne Thermal Emission and Reflection Radiometer,ASTER)的数据具有较高的空间分辨率、光谱分辨率和辐射分辨率,利用其波段设置对比其他多光谱数据具有更好的波谱连续性这一特点,文章提出一种应用ASTER遥感图像的岩矿信息提取方法,用于帕米尔高原地区的岩矿信息提取。利用ASTER图像中的5个热红外波段数值来计算矿物学指数:石英指数、碳酸盐指数、镁铁质指数,将以上3个指数作为特征,通过随机森林分类方法实现岩矿信息提取,成功提取了5种重要的岩性信息(硅质、硫酸盐、超镁铁质、石英以及变质硅酸盐)。通过对帕米尔高原地区的遥感图像分析,绘制相关矿藏分布图,与经过实地勘测的矿藏分布图以及经过实验测试得到的数据点进行对比,充分验证了方法的有效性。

ASTER data has high resolution in spatial,spectral and radiation resolution.With its band setting compared with other multi-spectral data,it has better spectral continuity.This paper proposes a rock and mineral information extraction method based on ASTER data for the Pamirs.The mineralogy index is calculated using the values of the five thermal infrared bands in the ASTER image:quartz index,carbonate index,and mafic index.This paper realizes rock and mineral information extraction through random forests,which uses the above three indexes.And it successfully extracts five important lithological information(siliceous,sulfatic,ultramafic,quartz and metamorphic silicate).Based on the remote sensing image analysis of the Pamirs,this paper draws the relevant mineral distribution map,compares with the mineral distribution map and the experimentally tested data points,which fully verifies the effectiveness of the method.