Super Resolution Imaging And It’s Applications

Images with higher resolution are required in almost all digital imaging applications. For past few decades considerable advancement has been realized in imaging devices. Modern digital cameras are equipped with high quality lenses, increased pixel resolution as well as small size, but still they are far from the perfection because of being costly and having physical limitations of hardware for example sensor, lens and optics. Using signal processing techniques to generate a high resolution image from a low resolution image or a set of low resolution images is a cheaper and effective alternative. Such kind of resolution enhancement is called super resolution image reconstruction. The resolution enhancement is achieved by fractional-pixel displacements between observed low resolution images. Super resolution methods tend to overcome the limitations of an imaging system where there is no need for additional hardware. This fact has made super resolution a hottest research area both scientifically and commercially. The research work in this thesis is mainly focused on four tasks.Firstly, we present an overview of well-known super resolution techniques for both multi-image super resolution and single-image super resolution. The performance comparison and analysis are the main concerns to see the advantages and disadvantages of multi-image super resolution techniques. After applying the super resolution techniques to our tested synthetic image data we critic the affect of the full reference quality metrics (peak signal-to-noise ratio and mean square error). Secondly, an optimized approach is introduced for the effective selection of low resolution images in the process of super resolution image reconstruction. We used minimum two low resolution images and performed experiments to validate our methodology.Thirdly, we investigate the non-parametric super resolutions for object recognition in Radar. We analyze how super resolution algorithm can be useful for resolving two or more targets in radar imaging system. The performed simulation results help us to evaluate the performance of non-parametric super resolution algorithms discussed. Finally, a little research has been carried out about a proposal for futuristic design of implementing super resolution image reconstruction in the environment of mobile cloud computing. The proposed framework is based on the idea of utilizing the cloud computing infrastructure through the mobile device+mobile web (or internet service provider) as an Educational Tool for super resolution image reconstruction.Throughout this research work, experiments on various real and synthetic image data are conducted to validate and evaluate the performance of the existing and proposed approaches.