基于Fourier-Meliin算法的干涉图像配准

提出了采用Fourier-Mellin算法对大孔径静态干涉成像光谱仪的原始干涉图像进行配准的方法,用于校正由于推扫平台系统姿态不稳所造成的失真.采用Fourier-Mellin算法和相位相关算法求取干涉图像的旋转角度和缩放及平移参数,并通过多帧未校正的图像和校正后的图像分别拼接成大面积地域图像来验证算法.实验结果表明,通过人眼判断,可以实现对图像的配准,配准精度达到1Pixel,基本满足将LASIS原始干涉图校正为不失真图像的要求.

Interference image registration based on Fourier-Mellin algorithm

In order to align the distortion caused by the jitter of push-broom flat, the Fourier-Mellin algorithm is proposed to recover the interference image of a large aperture static interference imaging spectrometer. The rotation, scaling and translation parameters are acquired by Fourier-Mellin algorithm and phase correlation algorithm. The registration effect is testified by comparing the mosaic images before or after registration. The experiment results indicate that the registration of the original interference image can be testified through vision observation. The registration precision of the algo- rithm reaches pixel level,which can meet the requirement of aligning the interference image to non-dis tortion image.