基于最小二乘平差的全极化SAR配准偏移量估计方法

An offset-tracking method based on least square adjustment of full polarimetric SAR

全极化合成孔径雷达影像(PolSAR)相对单极化SAR影像有更加丰富的地表信息。为了提高SAR影像偏移量跟踪技术估算偏移量的精度,提出一种基于最小二乘平差法的全极化SAR数据偏移量估计方法。首先利用全极化SAR不同极化通道数据估算偏移量得到多个观测值,然后通过最小二乘平差法对多余观测值循环剔除粗差得到最优偏移量。文中选取美国科罗拉多州湖城(Lake city)区域的两景JPL无人机UAVSAR全极化SAR影像进行实验,结果表明,文中新方法具有良好的粗差探测和去除功能,相对于已有研究结果,在方位向和距离向的偏移量估算精度都有明显提高,达到15%~25%。新方法提高了偏移量跟踪估算偏移量的精度,这对利用偏移量跟踪技术监测滑坡、地震及冰川等有重要的意义。

Comparing with single-channel SAR image, the full polarimetric synthetic aperture radar image （PolSAR） has more abundant terrain information. To improve the offset estimation accuracy of SAR images, the paper proposes an offset-tracking method based on least square adjustment of full polarimetric SAR data. Firstly, the multiple offset observes are obtained of different polarimetric channel SAR images. Then the least square adjustment is used to eliminate the errors and get the optimal results. Finally, two JPL UAVSAR full polarimetric SAR images are selected to cover Lake City, Colorado, USA for validating the proposed method. The experimental results show that, compared with the single-channel SAR image, the offset errors of the proposed method decrease significantly. The precision of offsets in range and azimuth improves by 15%~25%. The proposed method improves the accuracy of offset tracking. It is significant to use the offset-tracking method for monitoring deformation of landslide, earthquake and glacier.