基于光学GCP图像切片的SAR图像自动匹配方法

地面控制点是进行星载SAR图像几何精校正的重要条件之一.通过光学GCP切片和SAR原始图像进行匹配可以实现控制点的自动提取.提出了一种在图像梯度域,实现光学图像和SAR图像异源匹配的方法.首先根据SAR的轨道参数等信息,实现光学GCP切片和SAR图像尺度和旋转的归一化,解决异源图像间的分辨率和旋转角度差异.然后分别提取SAR图像和光学图像的梯度强度信息,为了消除SAR图像的斑点噪声,采用ROA算法准确提取SAR图像的梯度强度.SAR图像和光学图像的成像机理差别很大,直接在图像域匹配较为困难.但是它们的梯度强度图像具有相似性,为此可以通过在梯度强度图像域,应用归一化互相关匹配算法,实现同名点的自动匹配.最后采用实验证明了方法的可行性.

SAR image registration algorithmic based on optics GCP slices

Ground control point （GCP） plays an important role in the processing of spaceborne SAR images geometric registration,matching GCP image slices and SAR images can extract control points automatically,we propose a method of matching optical image and SAR image in gradient field.First of all,according to the SAR＇ orbital parameters,we normalize the scale and rotation of GCP slices and SAR images,solving differences in resolution and rotation between different source images.Then the gradient intensity information of SAR and optical images are extracted respectively.Here,we pick ROA algorithm to calculate SAR images＇ gradient for purpose of eliminating the influence of speckle noise.It is difficult to match SAR and optical image because of obvious different imaging mechanisms.But their gradient maps are similar,so we apply using normalized cross-correlation matching method in gradient intensity filed to realize corresponding points matching.Finally,experiments show the feasibility of this approach.