Model-free self-calibration based high-speed spectrally encoded Mueller matrix polarimetry

To address the degradation of measurement accuracy caused by fabrication and alignment errors of optical components as well as random noise in spectrally encoded Mueller matrix polarimeters (MMP), this paper proposes a model-free self-calibration and reconstruction method for the system modulation matrix. The method dispenses with the complex procedures of conventional schemes that rely on external high-precision reference systems or independent calibration of component parameters. Using the spectra of only 28 standard samples with known Mueller matrices, the determination of the modulation matrix is transformed into a nonlinear optimization problem solved by a differential evolution algorithm, directly estimating the system modulation matrix that embodies all non-ideal errors of the system. Based on this matrix, high-precision reconstruction of Mueller matrices for both static and dynamic samples is achieved through a linear operator approach. Experimental results show that when the proposed method is applied to a fast measurement system incorporating time-stretch technology, the single-shot measurement time is 92 ns, and the average root mean square error (RMSE) of the reconstructed Mueller matrices is 0.038 for static samples and 0.048 for dynamic samples, outperforming the conventional Fourier-domain method and the results relying on dual rotating retarder calibration. The method features conceptual simplicity, no requirement for wavenumber equally spaced sampling, and better noise immunity, providing an effective self-calibration solution for high-speed, high-precision spectrally encoded MMP.