摘要
复图像配准是干涉SAR处理中的关键步骤之一,要求配准精度达到0.1个像元的水平,而控制点的精确定位是复图像配准中的核心环节。尽管前人提出采用复相关系数法、信噪比法、平均扰动函数法等算法,但他们都是基于滑动窗口的算法,运算效率低。在分析前人算法的基础上,引入基于谱运算的复相关函数算法实现控制点的亚像元级精确定位,分析比较了算法的计算复杂度。应用该算法对我国某地区ERS 2的SLC数据进行了实验,生成了干涉图像,并进行了相干性分析,证明了该算法的可行性和稳健性。实验研究证明即使在主从图像时间间隔较大、相干性差的情况下,应用该算法也能获得干涉条纹图。
Complex image corregistration is one of the key steps in SAR interfereometric processing. While tie points accurate positioning is a core procedure in the complex image corregistration. The algorithms based on shifting window are time consuming. Introduces the FFT-based complex correlation function method which can realize integeral pixel-level control points positioning. This method can efficiently avoid complicated computation resulted in by shifting window-based algorithms. Chooses China Taiwan region's ERS-2 SLC data to perform above mentioned algorithms, and meanwhile analyze the coherence of coregistered images. The results prove that the algorithms are feasible,robust and efficient. Namely, even though the two images' acquisition time partition was relatively large, concerning with bad circumstance of coherence, Interfereogram can also be generated.
出处
《测绘学报》
EI
CSCD
北大核心
2004年第2期127-131,共5页
Acta Geodaetica et Cartographica Sinica
基金
国家自然科学基金资助项目(40242020)
国家863计划资助项目(2001AA135100)
关键词
复图像配准
复相关函数
二次曲面拟合
SINC函数插值
干涉条纹图
谱运算
complex image corregistration
complex correlation function
quadric curved surface fitting
SINC function interpolation
interferogram
