The Research On The Algorithms Of Super-resolution Remote Sensing Image Reconstruction In Spectral Domain

In 1960s, Remote Sensing technology of which Remote Sensing Image Processing is an important part,as a new all-around detection technology, is developed greatly. But limited by the manufacturing process of imagery equipment,these images are the under-sampled of scene .Also, the motion between the earth and the platform, atmosphere disturbance and so on make the images blurred and degraded. It is the aim for many people to improve the image resolution.In this paper, central researches on the algorithms of super-resolution in spectral domain are done. The spectral de-aliasing algorithm in frequency domain and the wavelet and fractal interpolated algorithm are implemented in this paper. The role of spectral aliasing and fractal in resolution is also analyzed in-depth.The dissertation first describes the development history and significance of super-resolution and the fundamentals and feasibility of resolution by analyzing the image processing. Then the cause of spectral aliasing is analyzed deeply. And based on the fundamentals of super-resolution in spectral domain, spectral de- aliasing algorithm is implemented. The model of image resolution is established and validated by experiment. After introducing wavelet theory and fractal conception and displaying the advantages of them in image analyzing domain, the image sequence with displacement is reconstructed by wavelet and fractal interpolated algorithm. The results of experiment and computer simulations show that the algorithm is feasible and efficient both on subjective way and objective way. Finally, the summary and prospects of the thesis are given.