Segmented iterative compensation algorithm for phase noise mitigation in optical frequency domain reflectometry

In optical frequency-domain reflectometry (OFDR) systems, phase noise caused by the nonlinear frequency sweep of the light source leads to spectrum broadening and polarization fading. This problem becomes more significant in multi-target scenarios, where peak shifts and target overlap occur, which limits OFDR performance and reduces spatial resolution and localization accuracy. To improve the realism and robustness of the algorithm, Gaussian white noise is introduced during modeling to simulate the non-ideal behavior of the light source, making the simulation closer to practical conditions. On this basis, this work proposes a segmented polynomial iterative compensation method for suppressing nonlinear tuning noise. The auxiliary interferometer signal is divided into multiple segments, and the residual phase of each segment is fitted by a multi-order polynomial. An adaptive strategy is used to adjust the number of iterations to avoid overfitting or underfitting. The iterative complex exponential compensation factors are then applied to the main interferometer signal segment by segment, enabling effective suppression of nonlinear phase errors. Meanwhile, a differentiated frequency-domain processing strategy is designed for multi-target conditions: a narrow sliding window is used for closely spaced targets to enhance separation, while a reduced compensation strength is applied to distant targets to maintain peak stability. Simulation results show that two targets with a separation of less than 5m are clearly distinguished, with spatial resolution improved by 12.31 times and 3.63 times, respectively. For a distant target at 2220.0m, the spatial resolution improves by about 27.14 times, and the peak deviation is significantly reduced. Even with added Gaussian white noise, the proposed method maintains stable compensation performance and high localization accuracy. Compared with traditional compensation methods, the proposed approach shows stronger noise robustness and better multi-target discrimination across different distance ranges. It provides an efficient and practical solution for nonlinear tuning and multi-target interference, demonstrating strong potential for engineering applications.