Digital dispersion compensation for ultrabroad-bandwidth single-camera spectral-domain polarization-sensitive OCT

Abstract

Polarization-sensitive OCT is used to examine tissue microstructure by providing imaging of birefringent properties. Single-camera spectral-domain polarization-sensitive OCT has been of recent interest, whereby a custom spectrometer is employed to simultaneously measure orthogonal polarization states scattered from the sample. This avoids synchronization and triggering issues associated with multiple-camera setups. It also has the advantage that the optic axis can be extracted without polarization modulating the incident light. However, the disadvantage is that the line camera pixel-to-wavenumber nonlinearity requires either careful spectrometer alignment, or digital compensation. In fact, this problem is further exacerbated in high resolution PSOCT systems as they require compensation over larger bandwidths. Here we report the construction of an ultrabroad-bandwidth PSOCT system using a single camera spectrometer similar to Baumann <i>et al</i>. In order to enjoy the benefits of this instrument, we outline a method for digital dispersion compensation that removes the necessity for special camera alignment. We find that there are three non-negligible types of dispersion to consider: 1) the aforementioned camera pixel-to-wavenumber nonlinearity, 2) the refractive index dispersion in the sample itself, and 3) the dispersion imbalance between the arms of the OCT interferometer. The latter two were previously recognized for time-domain high-resolution OCT, where a digital dispersion compensation method was successfully employed to treat them both. For our SDOCT application, we find that dispersion types 1 and 2 have the same functional effect and can be combined into one compensation step, and as such, much of the previous compensation method can be used. However, we find that it is necessary to add two steps to the analysis technique whereby the relative scaling and positioning of the two polarization images is adjusted to align the scatterers. We also find that better results are achieved by fitting to larger polynomial orders. We show how our technique provides high-resolution PSOCT with precise alignment between the orthogonal polarization images.