摘要
设计了一种基于相位相关与归一化积相关的联合图像校正算法。该算法针对卫星姿态变化对LASIS(大孔径静态干涉成像光谱仪)图像造成的畸变,根据傅里叶变换的旋转平移特性,实现图像序列旋转失真的高精度校正,校正精度达到0.01 ;同时,基于归一化积相关方法实现图像序列平移失真的校正。针对LASIS数据量大的特点提出了改进的快速自适应模板选取法,使算法的运算复杂度按4n递减。仿真结果表明,该算法能快速有效地校正LASIS图像。
A joint image correction algorithm based on phase correlation and normalized product correlation is designed. In view of the LASIS (Large-Aperture Static Imaging Spectrometer) image distortion induced by satellite attitude variation, the high-accuracy correction for rotation distortion of image sequence is performed based on the algorithm according to rotation translation property of Fourier transform, with correction accuracy 0.01. The correction for translation distortion of image sequence is performed based on normalized product correlation method. An improved fast adaptive template match method is proposed against large data quantity of LASIS, thus the complexity of the algorithm operation decreases progressively according to 4n. The simulation results show that the algorithm can effectively correct distortion of LASIS image.
出处
《光电工程》
CAS
CSCD
北大核心
2004年第8期8-11,共4页
Opto-Electronic Engineering
关键词
成像光谱仪
干涉光谱仪
图像匹配
相位相关
图像校正
Imaging Spectrometers
Interference spectrometers
Image matching
Phase correlation
Image correction
