基于控制点的SAR图像亚像素自动配准

Automatic Sub-pixel Registration for SAR Image Based on Control Point

为了解决同一SAR系统所拍摄的多幅图像配准问题,提出了一种基于控制点的自动配准方法。此方法在边缘检测的基础上用Harris算子对主图像提取角点作为控制点,再利用基于灰度相关的局部区域匹配寻找同名像点,最后通过控制点求解仿射变换参数。大量仿真结果表明,此方法的配准误差在0.2左右,达到了亚像素级水平。对控制点的选取和匹配不需要人工参与,计算速度快、精度高、有较好的稳定性。

To process images registration of the sample SAR system,an automatic sub-pixel registration algorithm based on control point for SAR image is presented in this paper.Firstly based on the edge detection,we chose corner points as control points by the Harris detector in the master image.The search for the corresponding CPs in the slave image is based on local region matching by means of grave correlation measure.Finally we calculate the parameter of affine transform according to the selected precise CPs.Lots of simulation experimental results demonstrate that the registration error of the proposed algorithm is about 0.2,and the precision of the algorithm is sub-pixel level.In this paper the process of choosing and searching for CPs is totally automatic,and the algorithm is efficient,accurate and robust.