实用干涉合成孔径声纳复图像配准法

海底地貌测绘的最新技术是利用干涉合成孔径声纳(InSAS)进行三维立体成像,其信号处理的关键在于成像之后的图像配准、干涉成像等步骤。图像配准的精度直接影响相位图的质量,从而影响所求目标区域高度的准确性。在分析InSAS复图像数据特点的基础上给出一种实用的基于最大干涉频谱法的复图像配准方法,缩减了配准搜索范围,在一定程度上提高了配准速度和精度。利用该方法对湖试测得的湖底地貌数据进行配准,用复图像相干系数和干涉相位图的残余点数对配准效果进行评估,通过配准前后的参数比较,证实了最大干涉频谱法用于InSAS复图像配准能够取得较好的配准效果。

A practical co-registration method of InSAS

Interferometric Synthetic Aperture Sonar(InSAS) has been used to map seabed elevation in re-cent years.Registration of InSAS complex images is a critical steps,which requires sub-pixel precision.This paper presents an image registration method using a maximum spectra algorithm that reduces the search range,accelerates the search speed and improves precision.It is used to analyze the InSAS signal acquired in lake trials,and evaluated by correlation coefficient and the number of residues.The results show that the maximum spectra algorithm applied to the registration of InSAS complex images is valid.