Spectral reflectance reconstruction in non-ideal broadband spectral encoding systems using commercial off-the-shelf filters

We investigate the spectral reflectance reconstruction in non-ideal broadband spectral encoding systems constructed using commercial off-the-shelf (COTS) broadband filters. Based on the measured spectral responses of ten COTS broadband filters, two reconstruction methods incorporating natural reflectance priors and regularization constraints are evaluated: a principal component analysis–ridge method and a physics-constrained U-Net method. The results show that despite the strong inter-channel correlations from broad, overlapping filter responses, stable reconstruction is achievable with appropriate priors and reconstruction strategies. Further analysis shows that reducing the number of filter channels decreases the Frobenius norm of the system matrix and weakens the complementary spectral constraints provided by the filter set. When too few channels are used, the measurements become less capable of distinguishing spectrally similar reflectance curves, resulting in degraded reconstruction performance. Overall, COTS broadband filters, owing to their availability, low cost, and ease of implementation, can serve as practical components for compact spectral reflectance reconstruction systems, provided that filter channel selection jointly considers system complexity, spectral coverage, and channel complementarity.