Adaptive simultaneous algebraic reconstruction technique for retrieving refractive index profiles of optical fiber

Abstract

An efficient adaptive simultaneous algebraic reconstruction technique (ASART) to calculate optical fiber refractive index profiles is proposed based on phase difference curves obtained by digital holography technique. We develop adaptive relaxation parameter (ARP) on simultaneous algebraic reconstruction technique (SART) to increase the convergence speed and improve the reconstruction accuracy. A formula of ARP is derived mathematically and multilevel scheme (MLS) is used to increase convergence speed in the first iteration. Experimental results show the proposed ASART convergences over 40% faster than SART and achieve significantly higher reconstruction accuracy. Experimental verification shows that ASART is more efficient than SART and filtered back projection in image reconstruction, especially with few projection views. The running time of ASART is much shorter than that of SART, and ASART needs fewer iteration numbers to obtain the same reconstruction effects. In addition, it can be used to measure optical fibers with various diameters that cannot be measured with S14 refractive index profiler (S14).