TerraSAR-X/TanDEM-X升降轨双基干涉模式获取DEM方法研究  

A High Precision DEM Generation Method Based on Ascending and Descending Pass TerraSAR-X/TanDEM-X BiSAR Data

作　　者：秦小芳 张华春[1] 张衡 王宇[1] 刘华有 Qin Xiaofang;Zhang Huachun;Zhang Heng;Wang Yu;Liu Huayou(Institute of Electronics,Chinese Academy of Sciences,Beijing 100190,China;University of Chinese Academy of Sciences,Beijing 100049,China)

机构地区：[1]中国科学院电子学研究所,北京100190 [2]中国科学院大学,北京100049

出　　处：《雷达学报（中英文）》2018年第4期487-497,共11页Journal of Radars

基　　金：国家重点研发计划(2017YFB0502700)~~

摘　　要：该文基于TerraSAR-X/TanDEM-X(TSX/TDX)双基升降轨数据,首先采用非局部干涉(Non Local Interferometric SAR,NL-InSAR)相位滤波分别得到单航过升轨和降轨模式下的高分辨率DEM。在此基础上,基于NL-InSAR估计得到的较准确相干系数,提出一种升降轨DEM融合方法,恢复SAR侧视成像造成的几何畸变,提高DEM重建精度。该文采用两幅北京地区的TSX/TDX升降轨干涉对进行融合处理,结果表明,在地形复杂地区的叠掩和阴影等无效区域,融合之后的DEM无效点数明显减少。经统计,融合后无效点数比例由升轨、降轨的4.93%和4.52%降低到1.34%。同时,融合DEM的精度相比于升轨的6.74 m提高了8.7%、相比于降轨的6.67 m提高了9.6%,融合后高程精度达到6.09 m。A method for fine resolution and high precision Digital Elevation Model （DEM） generation using ascending and descending pass TerraSAR-X/TanDEM-X （TSX/TDX） datasets is proposed in this study. First, the NonLocal Interferometric SAR （NL-InSAR） can effectively generate ascending and descending pass raw DEMs. On this basis, the coherence well recovered by NL-InSAR is used to fusion the raw DEMs to further improve the accuracy and reduce the invalids caused by layover and shadows. This method was used to process the TSX/TDX data obtained in Beijing. The number of invalid points of DEM after fusion decreased significantly. Statistics result shows that it decreased from 4.93% in ascending raw DEM and 4.52% in descending raw DEM to 1.34% in the fusion DEM. At the same time, the accuracy of the fusion DEM increased by 9.6% compared to 6.67 m in the descending raw DEM and 8.7% compared to 6.74 m in the ascending raw DEM, reaching 6.09 m.

关 键 词：数字高程模型 非局部干涉参数估计 升降轨融合 高分辨率 高精度 

分 类 号：P237[天文地球—摄影测量与遥感]