Enhancement of fish-eye imaging quality based on compressive sensing

Abstract

Compressive sensing is a novel theory in signal acquisition and reconstruction, which states that the original signal can be reconstructed with few data. Limited by the computer memory, the image must be partitioned into small regional blocks for the reconstruction, especially when dealing with fish-eye images. However, the reconstruction has a major shortcoming, that is, the image block is more blurred at the edge of the reconstructed image than at the center. This phenomenon is caused by the adoption of similar sampling rate to different image blocks in the traditional process. Given that each image block has different spatial resolutions, the sampling process is not reasonable. This paper proposes the fish-eye camera spatial resolution first. Afterward, the sampling rate of different image blocks is improved according to spatial resolution. Finally, some numerical experiments are performed. Results of the numerical experiments show that the method proposed in this study is effective.