Computational optical time-domain reflectometry

Optical time-domain reflectometry (OTDR) is a well-developed technique detecting returned light signals which are scattered and reflected from the optical fiber when optical pulses are propagating forward along the fiber, providing useful information regarding the characteristics of the optical fiber, indicating loss, locations of faults or connections, etc. The returned signals need to be acquired at a sampling rate satisfying Nyquist–Shannon sampling theorem. In this paper, we propose a computational OTDR based on computational ghost imaging technique in the time domain by sending pre-known pulse sequences and collecting the “integrated” returned light signals, realizing a very different probing and acquiring approach compared to conventional OTDR, reducing sampling rate requirement significantly. Theoretical analysis and mathematical details are given for demonstrating the performance difference of the proposed method using probe sequences derived from a random matrix and an orthogonal matrix. Experiments are carried out to verify the theory indicating that significant improvement in signal-to-noise ratio (SNR) can be achieved using Walsh–Hadamard matrix which is suggested by the theoretical analysis. Finally, we give detailed comparison of the proposed computational OTDR with the conventional technique, showing that sampling rate can be reduced greatly using the computational technique, but it comes with a penalty of requiring more measurements for achieving the same SNR as the conventional technique.