基于Hyperion与SASI数据的矿物填图方法对比研究

发布时间：2018-04-20 14:29

  本文选题：航天高光谱 + 航空高光谱　； 参考：《中国地质大学(北京)》2017年硕士论文

【摘要】：高光谱遥感技术诞生后的几十年里,已经在众多领域中投入实际应用。其中,地质勘查是应用最为成功的领域之一。这是因为矿物填图是地质勘查工作中极为重要的一环,而高光谱遥感超高的光谱分辨率对提高矿物填图精度意义重大。国内外已经进行了许多高光谱矿物填图技术及应用研究,但在不同尺度高光谱数据、不同填图方法的填图效果对比研究方面还很少,本文在此基础上,以新疆和布克赛尔县白杨河地区为试验区,开展基于Hyperion卫星高光谱和SASI航空高光谱数据的矿物填图方法对比研究,完成的主要工作量和取得的主要成果如下:1.通过对Hyperion数据与SASI数据的预处理,包括辐射校正、几何校正、大气校正、坏像元去除等,获取了矿物填图方法对比研究所需的反射率数据。2.综合利用PPI与SMACC两种算法对航天Hyperion数据与航空SASI数据进行端元提取,建立了研究区图像端元光谱库,为本文开展矿物填图对比研究奠定了基础。3.通过对SAM、MTMF、SFF等三种矿物填图方法的原理及基于Hyperion数据和基于SASI数据的高光谱矿物填图效果的对比研究并经野外验证与样品光谱测量,认为对于Hyperion数据,MTMFF填图效果好于SAM和SFF算法。对于SASI航空高光谱数据,SAM与MTMF矿物填图也好于SFF算法;其中,对于高铝绢云母、中铝绢云母、低铝绢云母等光谱特征十分相似的矿物,SAM与MTMF矿物效果相近,但在绿泥石与绿帘石矿物填图方面,SAM填图效果更好;在叶腊石与碳酸盐等分布范围都较小,且光谱吸收特征明显的矿物来说,MTMF方法填图效果更好,精度更高。4.通过对两种不同尺度的高光谱数据的矿物填图效果对比研究认为SASI航空高光谱数据矿物填图种类和精度明显高于Hyperion数据。但两种数据的矿物填图结果在矿物大类和空间分布上具有较好的一致性,表明Hyperion数据在实际应用中能满足对蚀变矿物大类的检出率要求。
[Abstract]:Hyperspectral remote sensing technology has been applied in many fields in recent decades. Among them, geological exploration is one of the most successful applications. This is because mineral mapping is a very important part of geological exploration, and hyperspectral remote sensing is of great significance to improve the precision of mineral mapping. Many hyperspectral mineral mapping techniques and applications have been carried out at home and abroad, but there are few comparative studies on the mapping effects of different scale hyperspectral data and different mapping methods. Taking the Baiyang River area of Xinjiang and Bukeser County as the experimental area, a comparative study of mineral mapping methods based on Hyperion satellite hyperspectral data and SASI aerial hyperspectral data is carried out. The main workload and the main results obtained are as follows: 1. By preprocessing Hyperion data and SASI data, including radiation correction, geometric correction, atmospheric correction and bad pixel removal, the reflectivity data of mineral mapping method is obtained. End-component extraction of aerospace Hyperion data and aeronautical SASI data was carried out by using PPI and SMACC algorithms, and the end-component spectral database of the study area was established, which laid a foundation for the comparative study of mineral mapping in this paper. Based on the principle of three mineral mapping methods, such as Hyperion data and SASI data, the comparative study on the effect of hyperspectral mineral mapping based on SASI data and field verification and sample spectral measurement, it is concluded that the mapping effect of Hyperion data is better than that of SAM and SFF algorithms. For SASI aerial hyperspectral data, the mineral mapping of SAM and MTMF is also better than that of SFF algorithm, among which, for the minerals with very similar spectral characteristics, such as high alumina sericite, medium aluminum sericite and low alumina sericite, the effect of MTMF mineral is similar. However, the mapping effect of SAM is better in the mineral mapping of chlorite and verdant, and the mapping effect of MTMF is better and the precision is higher for minerals with smaller distribution range of pyrophyllite and carbonate and obvious spectral absorption characteristics. By comparing the mineral mapping effects of two kinds of hyperspectral data of different scales, it is concluded that the types and accuracy of mineral mapping of SASI aerial hyperspectral data are obviously higher than those of Hyperion data. However, the mineral mapping results of the two kinds of data are in good agreement with each other in the mineral classification and spatial distribution, which indicates that the Hyperion data can meet the requirements for the detection rate of altered minerals in practical applications.
【学位授予单位】：中国地质大学(北京)
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P627
，

本文编号：1778150