| Large aperture static imaging spectroscopy (LASIS) is an imaging spectroscopy with new principle which may effectively compromise flux output and device stability. However, satellite attitude changes, such as yaws, pitches and rolls, will cause data distortion during the imaging process. This limits the application of LASIS seriously. As a result, image correction becomes a key technology of LASIS data processing which possesses not only important theoretical research value but also practical application significance. In this thesis, the image correction technology of LASIS based on satellite attitude changes is studied, taking the imaging characteristics into account.The effect of satellite attitude changes upon LASIS images is analyzed firstly. Main achievements include: (1) The mathematical relationship between attitude angle and pixel offset is established by a discussion of the respective effect and a distortion model is derived accordingly; (2) According to the imaging mechanism of LASIS, the difference of correction technology between LASIS images and normal remote sensed images is summarized so as to provide the theory basis for a rational correction algorithm design. A set of algorithms are then proposed to realize the correction algorithm for LASIS images: (1) A correction system is constructed which includes image matching technique and resampling technique; (2) A joint image correction algorithm based on correlation is designed which can match the images with an accuracy of 1/10 pixel; (3) A fast adaptive template matching algorithm is proposed. This algorithm adopts pyramid structure and takes the advantage of edge feature, which increases the accuracy of matching; (4) Sliding window is used in the matching of image sequence which is proved to possess better error-resilience and higher precision.
|
PDF Full Text Request