星载高分辨率合成孔径雷达影像纠正技术

Geo-rectification Method of High Spatial Resolution SAR Data

为检测TERRASAR、COSMO SkyMed、RADARSAT-2等星载高分辨率合成孔径雷达影像(SAR)在土地利用调查监测中的适用性,该文针对高分辨率SAR数据和产品特性,提出了控制点选取方法,分析了不同纠正模型的应用效果。试验表明高分辨率SAR几何纠正一般需要10～15个控制点,1m聚束模式纠正中误差约3m～5m,3m条带模式纠正中误差约5m～8m,分别满足1∶1万和1∶2.5万土地调查监测几何精度要求。研究结果为构建基于高分辨率SAR数据土地利用调查监测应用技术流程和促进高分辨率极化SAR数据业务化应用奠定了基础。

In order to test suitability of the high spatial resolution SAR（1m/3m） for land dynamic change monitoring,in this paper the TERRASAR、COSMO SkyMed and RADARSAT-2 data were used for deep study.The critic problems of geo-rectification of the high spatial resolution SAR data are analyzed;the principle and method of GCP collection are proposed;different SAR geometric correction models are used for comparison and their suitability is discussed.The results show that 10-15 GCP is required to obtain a good result of geometric correction.Suggested geometric correction method can obtain good result,that RMSE for 1m spot-line mode SAR data is 3-5m and 5-8m for 3m strip mode SAR data which are suitable for 1∶10,000 and 1∶25,000 mapping scales respectively.The results are very important to make good use of SAR data in land use dynamic change monitoring project.