Single remote sensing datum is difficult to obtain both high spatial resolution and high spectral resolution characteristics at the same time in the field of remote sensing geology,so it is necessary to synthesize remote sensing data from different sources together to identify minerals and rocks.In this paper,lithological classification through the combination of multi-source remote sensing data of high spatial resolution of WorldView-2 data and high spectral resolution of Hyperion data is exercised in the identification of rocks and minerals and division of stratigraphy in the Wuqia County of Xinjiang.The basis of synergy is found and established by comparison of the detection ability of space and spectra between the two kinds of data source.Taking traditional source data fusion method as reference,the WorldView-2 multi-spectral data are sampled down to different spatial scale and the principal component is transformed.Then,synergy data are made by reconciling the first principal component and the Hyperion data.Based on the SAM method,the lithology in the study area is classified by using WorldView-2,Hyperion and synergy data.The experimental results show that the classification accuracy of all the synergy data are high,the Hyperion data comes next,and the accuracy of WorldView-2 data is the lowest.%在目前的技术条件下,一种遥感数据源很难同时具有高空间分辨率和高光谱分辨率特征,因此如何协同使用不同来源的遥感数据进行矿物岩石识别就成了遥感地质领域的重要研究内容之一.本文结合高空间分辨率WorldView-2数据和高光谱分辨率Hyperion数据,以新疆鸟恰县矿物岩石识别与地层划分为例探讨多源遥感数据协同的岩性分类.通过对比2种数据源的空间和光谱探测能力,找到其协同基础.借鉴传统的同源数据融合方法,把WorldView-2多光谱数据降采样到不同空间尺度,并进行主成分变换,采用第一主成分与Hyperion融合产生协同数据.基于SAM光谱角分类法,分别采用WorldView-2多光谱原始数据、Hyperion原始数据和各协同数据对研究区的岩性进行自动分类.结果表明,各协同数据岩性分类精度较高,Hyperion次之,WorldView-2数据岩性分类精度最低.
展开▼
