Remote Sensing Image Super-resolution Using Dual-Dictionary Pairs Based on Sparse Presentation and Multiple Features
Xiaomin Yang, Wei Wu, Weilong Chen, Gwanggil Jeon, Binyu Yan · 2014
In this paper, a super-resolution method based on sparse dictionary and multiple futures is proposed for remote sensing images. Super-resolution aims to reconstruct the high-frequency detail from the low resolution image. In this paper, high frequency is decomposed into two parts: primary high-frequency and residual high frequency. We proposed dual-dictionary pairs, i.e. primitive sparse dictionary pair and residual sparse dictionary pair to recover primary high-frequency and residual high frequency respectively. To describe the image more precise, we use multiple features to describe the structure of the image, and combine them together to present the image. Then use the combination futures to train the dictionary. The experimental results show that the proposed algorithm has a good performance, and the high-resolution image generated by the proposed method is with better subjective and objective quality compared with other methods.