Features and Properties of Single-Pixel Imaging Using Speckle Patterns Generated by Multi-Core Fiber

Single-pixel imaging (SPI) has recently drawn considerable attention as a new imaging technique. An SPI system using a multi-core fiber (MCF-SPI system) that we proposed has the potential to make the system very compact. This study is concerned with the features and reconstruction properties of MCF-SPI system. In this system, the reconstruction quality varies widely depending on the number and placement of cores. It is necessary to use reconstruction algorithms suitable for the system, considering the performance limitations of the patterns, to improve the output reconstruction quality. The features and properties of speckle patterns generated by MCFs with different core layouts and algorithms were investigated to improve the reconstruction performance based on numerical simulation. Four existing algorithms were compared under several conditions to evaluate the algorithms that improve reconstruction quality. Compressive sensing based on total variation is the most compatible algorithm for MCF-SPI. It was confirmed that the MCF-SPI system performs well in terms of imaging quality if a suitable core layout and algorithm for the application are set.