A Fast SRR Algorithm Based on Recursive Least Square Estimation and Simultaneous Image Registration

A Fast SRR Algorithm Based on Recursive Least Square Estimation and Simultaneous Image Registration by Stephane Kirchner This thesis proposed a Super Resolution Reconstruction (SRR) algorithm based on recursive least square estimation and simultaneous image registration. This recursive least square estimation algorithm is computationally fast and effective, and the simultaneous image registration algorithm is more efficient for real practical use. So this thesis will try to present the advantages by combining these two algorithms into a single framework. The thesis is structured into five main parts: • The first chapter introduces the super resolution reconstruction (SRR). • Later, the chapter 2 examines the super resolution reconstruction algorithm using the stochastic regularization approach. • The chapter 3 reviews and examines three SRR algorithms using the stochastic regularization approach (Classical SRR algorithm, Fast SRR algorithm based on Recursive Least Square and SRR algorithm based on Simultaneous Image Registration) and also the proposed algorithm based on recursive least square estimation and simultaneous image registration. • The chapter 4 gives some experimental comparisons among different algorithms. • And finally, the chapter 5 will give the conclusion and the future works directions.