High quality underwater computational ghost imaging based on speckle decomposition and fusion of reconstructed images

Sheng Lv,Tianlong Man,Wenxue Zhang,Yuhong Wan
DOI: https://doi.org/10.1016/j.optcom.2024.130460
IF: 2.4
2024-03-21
Optics Communications
Abstract:Ghost imaging (GI) is an imaging method that reconstructs object information via light-intensity correlation measurements. The unconventional imaging mechanism of GI makes it suitable for imaging in scattering medium, for instance, turbid liquid, biological tissues, etc. However, the strong background noise in the reconstructed images is still the bottleneck problems. In this paper, we focus on the improvements of the reconstructions in underwater GI and propose a speckle decomposition and fusion method for suppressing the noise. The approach utilized computational ghost imaging framework and dictionary learning to acquire high- and low-frequency components that are better suited for image feature-based reconstruction. Furthermore, once the suitable threshold has been established, the speckles captured from the reference path are decomposed using the Nonsubsampled Contourlet Transform (NSCT). The acquired high- and low-frequency are then correlated with the overall optical intensity that captured by the bucket detectors. Ultimately, a multi-scale inverse NSCT transformation is employed to produce the final fusion-reconstructed image. The quantitative evaluation of the reconstructed image quality was conducted. In contrast to previous methodologies, the proposed method offers significant enhancements of the underwater GI imaging quality.
optics
What problem does this paper attempt to address?