Research on the sequential images registration of the temporally and spatially modulated Fourier transform imaging spectrometer

Abstract

Temporally and Spatially Modulated Fourier Transform Imaging Spectrometer (TSMFTIS) is a new imaging spectrometer without moving mirror and slit. Through scanning, it can acquire sequential images superposed with interference fringes. The interferogram can be acquired by orderly arranging the extracted interference information of the same spatial point from the sequential images, and the spectrum can be recovered by using FFT. Therefore, the attitude of the bearing platform will affect the images so as to reduce the accuracy of the recovered spectrums. Since current attitude measurement accuracy can not meet the needs of error correction, in this paper, the image registration method is applied to acquire the accurate translations for the future correction between two sequential images. The single-step DFT registration method is applied to register the selected window areas away from the null optical path difference position in sequential images. That is full utilizing of common information meanwhile reducing impact of interference fringes and improving registration accuracy and efficiency. In the simulation experiment, a common large remote sensing image is used as ground object. The Fourier shift principle is applied to acquire simulation scanning images with sub-pixel displacement. Artificial spectral data cube produced with the RGB values of each image is utilized as the input data of the TSMFTIS, and sequential images superposed with interference fringes are acquired. Registration according to the method mentioned above is performed and the results are compared with the accurate values. It shows that the method is feasible and can achieve sub-pixel level accuracy.