Comparative Analysis of Single and Multi Frame Super Resolution in Satellite Imagery

摘要

Since last two decades there have been plenty of papers proposing a variety methods of single frame and multi-frame resolution enhancement. These methods are usually very sensitive to their assumed model of data and noise, which limits their utility. This paper analyses the contrast between single and multi-frame super resolution and addresses their shortcomings. The comparison is mainly visual and statistical. Where on one hand Multi-Frame technique is implemented using Ll norm minimization and robust regularization to deal with different data and noise models [1], Single Frame technique on the other is implemented through a tool developed in Java. The study area is Sitarganj city, Udham Singh Nagar district, Uttarakhand, India and the imagery is of LISS III and LISS IV sensors of ResourceSat-2 satellite. Both of these computationally inexpensive SR methods are robust to errors in motion, blur estimation, and result in sharp edges. Simulation results confirm the effectiveness of our method and demonstrate its superiority to other robust super-resolution methods.