归一化植被指数在遗址探测中的应用  

Application of NDVI time series in site detection

作　　者：张文君[1,2] 于丽君[1,2] 聂跃平[1] 潘玉青 

机构地区：[1]中国科学院遥感与数字地球研究所教育部国家文物局遥感考古联合实验室,北京100101 [2]中国科学院大学,北京100049

出　　处：《测绘科学》2016年第3期62-68,共7页Science of Surveying and Mapping

基　　金：十二五国家科技支撑计划项目(2013BAK08B06);中国科学院遥感与数字地球所所长青年基金专项(Y3SG8300CX)

摘　　要：针对目前遗址探测研究多基于单时相遥感数据开展,存在偶然性,对最佳探测时间的研究较少等问题,该文以洛阳盆地为研究对象,利用多时相遥感数据,通过时间序列谐波分析算法(HANTS)重构时间序列植被指数数据集,去噪的同时对比分析出利用冬小麦长势信息进行地下遗址遥感监测的最佳时间区间。研究表明,受地下遗址的胁迫,在分蘖期,冬小麦长势明显比非遗址区的长势差,表明该时期是进行地下遗址探测的最佳时期,进而对最佳探测时期内NDVI积分,有效增强了遗址区和非遗址区之间差异,突出地下遗址的位置和轮廓信息。利用该研究成果,文章成功探测到汉魏洛阳故城以及古伊洛河异常区,与现有的考古资料相吻合。Aiming at the problem that current site detection usually uses RS data of single phase,which are random,and the study on optimal detection time is insufficient,the paper used multi-source RS images of Luoyang Basin and HANTS algorithm was used to construct NDVI time series data set,by which the noises were removed and the best time of detecting underground sites using NDVI time series of winter wheat was found. Result showed that the growth of winter wheat of site area would be worse than the non- site area obviously in tillering stage of winter wheat,which should be the best time to detect the sites;and then the difference between the site and the non-site was enhanced by calculating the integral of NDVI in this time,to highlight the position of the underground ruins and contour information. Based on this meth- od,the site of capital Luoyang of Han-wei period and the paleochannel of Yiluo River were effectively de- tected on the integral image.

关 键 词：遥感考古 NDVI时间序列 时间序列谐波分析(Hants) MODIS TM/ETM 

分 类 号：P237.9[天文地球—摄影测量与遥感]