华南地质2024，Vol.40Issue(2)：393-403,11.DOI:10.3969/j.issn.2097-0013.2024.02.014
白垩系古沙漠露头遥感特征及其识别方法
Remote Sensing Characteristics and Identification Methods of Cretaceous Palaeo-desert Outcrop:Taking the Red Bed Basin in Southeast Hunan as an Ex-ample
摘要
Abstract
The Cretaceous palaeo-desert aeolian sandstone is a new type of sedimentary discovered in Hunan Province in recent years.This rock is not only a characteristic tourism resource,but also an important target reservoir for energy and carbon storage of global concern,so it is of great significance to conduct an in-depth study on the spatial distribution pattern of palaeo-deserts in Hunan Province.In this paper,remote sensing technology is introduced to establish the recognition and extraction method of remote sensing features of pal-aeo-desert outcrops:With Google Earth Engine(GEE)as the main processing platform and Sentinel-2A multi-spectral image as the main data source,by analyzing the remote sensing characteristics of palaeo-desert outcrops from three aspects:spectral,remote sensing index and topography and geomorphology features,and the Random Forest Algorithm was used to identify the palaeo-desert outcrops by remote sensing.Different from the traditional steep and straight Danxia landforms,the remote sensing images of the palaeo-desert are uniform in colour tone,continuously distributed,with undeveloped vegetation and mountain shadows,and with the geomorphological characteristics of'rounded tops and gentle slopes'.At the same time,deeply ana-lysing the stratigraphic lithology characteristics of the palaeo-desert,the outcrop of the palaeo-desert is char-acterized accurately by increasing geological constraints.With the methods above,a number of palaeo-desert outcrops in Danxia landform area of various red bed basins in Hunan Province have been interpreted,and two large areas of palaeo-desert aeolian sandstones in the eastern margin of Hengyang Basin(Dukou area)and the southeastern margin of Chayong Basin(Feitianshan-Gaoyiling area of Chenzhou City)have been verified.This study provides an efficient way to study the spatial distribution law of palaeo-desert,which is an impor-tant application of remote sensing technology in the field of geology,aiming to provide technical and geosci-ence references.关键词
古沙漠/遥感特征/Sentinel-2A/谷歌地球引擎/随机森林算法/湖南省Key words
palaeo-desert/remote sensing feature/Sentinel-2A/Google Earth Engine/Random Forest Algo-rithm/Hunan province分类
大气科学引用本文复制引用
周丽芸,黄乐清..白垩系古沙漠露头遥感特征及其识别方法[J].华南地质,2024,40(2):393-403,11.基金项目
湖南省自然科学基金项目(2021JJ30388)、湖南省地质院科研项目(HNGSTP202322)、古生物与地质环境演化湖北省重点实验室开放基金(PEL-202203) （2021JJ30388）
