基于区域网平差的多景SAR影像正射纠正

Ortho-rectification of Multi-SAR Images Based on Block Adjustment

感兴趣区域若处于两景或多景SAR影像的交叠覆盖区,那么要得到该区域具有同等精度的正射影像产品,则需对覆盖该区域的各影像进行区域网平差处理。本文对这一问题进行了深入分析,建立了基于G.Konecny方程的区域网平差模型,通过岭估计和最小二乘迭代实现了平差模型的解算。为检验本文提出的区域网平差方法,选取了两景相邻轨道的Radarsat-1SAR影像进行了区域网平差纠正处理,结果表明区域网平差所得产品的误差可在1.5个像素之内,显示了本文方法的可行性。

This paper proposes an ortho-rectification method for multi-SAR images captured for one limited area,which is based on block adjustment and is expected to get the same accuracy in this area.The ortho-rectification model is built from the G.Konecny equations,which solves the adjustment model by Ridge Estimate and least squares adjustment.To validate this method,two Radarsat-1SAR images at adjacent tracks were used to produce ortho-rectification product.The results demonstrate that this method is available to keep the error of ortho-rectification product within one and a half pixels.