利用Landsat-8和Worldview-2数据进行协同岩性分类

Lithological Mapping by Using the Synergestic Landsat-8 and Worldview-2 Images

提出了一种对Landsat-8和Worldview-2协同后的岩性分类方法。首先对Landsat-8和Worldview-2影像进行协同:在对Landsat-8全色波段与其多光谱进行自协同后,与Worldview-2多光谱第8波段数据协同,将协同后的Landsat-8中短波红外数据与Worldview-2数据进行叠加,得到最后协同结果。对协同后的数据进行岩性分类:利用基于最大似然法(maximum likelihood,ML)进行初始分类,由马尔科夫随机场法(Markov Random Field,MRF)对结果进行优化得到最终分类结果。采用新疆西昆仑地区遥感数据进行了实验,结果证实协同后数据的分类结果具有更高的分类精度。

A Lithologic Mapping method is proposed for the synergestic images of Landsat-8 and World- view-2. The synergestic images are generated by 4 processes. 1） Landsat-8 and Worldview-2 images are used for synergest; 2） The panchromatic and multispectral image of Landsat-8 are integrated to achieve a {used multispectral image; 3） The fused multi-spectral image of Landsat-8 is {used with the multispeetral image of Worldview-2; 4） The Worldview-2 multispectral and Landsat-8 shortwave infrared data are inte- grated to get the final synergestic result by optimization. Whereafter, synergestic images are used for li- thologic mapping. The Maximum Likelihood method is used to get the initial lithologic classification re- sults from the images, after that Markov Random Field is introduced to optimize the initial results to a- chieve the final lithologic classification results. The contrastive experiment is applied to the remote sensing images of west Kunlun mountains in Xinjiang. The experiment result indicates that the synergestic images are able to obtain higher accuracy of lithologic classification than the unsynergested images.