Optical frequency domain reflectometer based on spectral segmented normalized cross correlation

Cross-correlation is an algorithm that sums the product of sliding data, commonly used to determine wavelength shift in the local Rayleigh scattering spectrum of optical frequency domain reflectometers. Due to the fluctuation of Rayleigh scattering intensity, additional noise can be introduced during the cross-correlation process. This effect is particularly pronounced in the high spatial resolution, the intensity distribution of the local Rayleigh scattering spectrum no longer exhibits statistical properties, leading to an imbalance in its Rayleigh distribution. Certain high-amplitude intensity points can introduce additional noise, and it will cause a significant reduction in the algorithm’s resistance to intensity fluctuations, which may result in fake peaks affecting the demodulation outcomes. This work introduces a segmented normalization cross-correlation algorithm to address the significant errors in traditional cross-correlation result. The major technical way is to reduce the influence of the local Rayleigh scattering spectrum intensity fluctuation through segmented normalization. This approach aims to diminish the intensity of fake peaks in the cross-correlation process, thereby mitigating large errors. Using the proposed algorithm, measurements with a spatial resolution of 3.2 mm were achieved at 3000 με, and the spatial resolution has been improved by 5.6 times compared to traditional cross-correlation demodulation algorithms.