SAR影像配准中控制点粗差剔除方法研究

The Gross Error Removing Methods for Control-point Selection in SAR Registration

复数影像配准是雷达干涉测量中的关键步骤之一,而配准中选取的控制点往往存在粗差,影响配准精度。分析了SAR影像配准时控制点粗差产生的原因,提出了方差因子检验和Baarda数据探测法剔除控制点粗差的方法,建立了基于可靠控制点计算影像纠正的几何变换模型。利用ENVISAT卫星ASAR图像对所提方法进行了验证,结果表明,影像配准的精度达到了0.1像素,粗差剔除后SAR影像配准精度有明显提高。

The registration of the complex SAR images is a key step for interferometric synthetic aperture radar （InSAR）. The gross error of control-points in the registration of the complex SAR images will decrease registration accuracy. This paper investigates the reasons of the control-point gross errors. And a new method,which integrates overall test and Baarda data snooping, is proposed to detect and eliminate the gross errors. Two pairs of ENVISAT ASAR images from BAM, Iran were used to test the proposed method. The testing results show that the error from registration of the complex SAR images is less than 0. 1 pixel （RMS）. Thus, the SAR image registration accuracy is improved significantly.