| Summary: | 碩士 === 國立臺灣科技大學 === 電機工程系 === 99 === In many image processing applications, images with high resolution can provide more detailed information, and make images look close to the original scene. In order to satisfy human visual sense, using super-resolution image reconstruction technique to acquire a high-resolution image becomes an important research issue in image processing. Traditionally, image enlargement methods can only enlarge single image by interpolation. Thus, the quality of reconstructed image is constrained. Multiframe image super-resolution reconstruction approach is to obtain a high-resolution image using multiple low-resolution images. Therefore, there is more information that can be used to improve the image quality.
In this thesis, a super-resolution image reconstruction technique based on nonuniform interpolation method is proposed. This method consists of three major steps. First, the relative sub-pixel shifts between the multiple low-resolution images are estimated by image registration. Second, high-resolution image is reconstructed using an edge-oriented nonuniform interpolation method, which considers the edge information for image sharpness. Finally, blind deconvolution is used as image restoration process to remove blur and noise of the image since the system blur function is unknown. Experimental results show that the performance of the proposed method is better than that of other super- resolution approaches in terms of the subjective image quality and objective assessment.
|
|---|
