面向高分三号全球正射影像生成的无控定位精度提升方法

Geometric positioning accuracy improvement method without ground control points for global orthorectification of GF-3

高分三号(GF-3)是我国首颗C频段多极化高分辨率合成孔径雷达(synthetic aperture radar,SAR)卫星,由于GF-3卫星影像初始几何定位精度不高,难以直接满足大区域高精度制图需求。本文提出了一套面向全球正射影像生成的无控定位精度提升方法,首先通过几何再定标提升单景SAR影像的初始定位精度,其次利用SAR-SGFM匹配算法,结合自适应区域网平差方法精化影像几何模型参数,显著提升大规模SAR正射影像生成中的几何精度和处理效率。通过大规模数据生产验证,GF-3正射影像绝对几何精度优于10 m(1个像素),相对几何精度优于1个像素。结果表明,本文方法在超大规模星载SAR正射影像生产中具有有效性和可行性。

The GF-3 satellite is the first C-band multi-polarization high-resolution synthetic aperture radar(SAR)satellite.However,the initial positioning accuracy of GF-3 satellite images is unsatisfaction,it is difficult to directly meet the application requirements of large-area mapping.This paper proposes a large-area SAR orthoimage production method without ground control points(GCP),which significantly improves the geometric positioning accuracy and processing efficiency for SAR image processing.Firstly,the initial positioning accuracy of single-scene SAR images is improved by geometric recalibration.Secondly,the SAR-SGFM matching algorithm is used combined with the multi-node parallel matching strategy to automatically extract the tie points.The image geometric model is refined by the large-area adaptive block adjustment method and the radiometric equalization of adjacent images is realized by using the radiometric consistency processing algorithm.Based on the production and experimental verification of global SAR orthoimages based on the method in this paper,the absolute positioning accuracy of SAR orthoimage is better than 10 meters(1 pixel),the relative positioning accuracy is better than 1 pixel and the radiation of adjacent images is excessively smooth.The results show that the proposed method is effective and feasible in the production of large-scale SAR orthorectification image.