Dispersive hyperspectral VNIR (visible and near-infrared) imagers using back-illuminated CCDs will suffer from interference fringes in near-infrared bands, which can cause a sensitivity modulation as high as 40% or more when the spectral resolution gets higher than 5 nm. In addition to the interference fringes that will change with time, there is fixed-pattern non-uniformity between pixels in the spatial dimension due to the small-scale roughness of the imager’s entrance slit, creating a much more complicated problem. A two-step method to remove fringes for dispersive hyperspectral VNIR imagers is proposed and evaluated. It first uses a ridge regression model to suppress the spectral fringes, and then computes spatial correction coefficients from the object data to correct the spatial fringes. In order to evaluate its effectiveness, the method was used to remove fringes for both the calibration data and object data collected from two VNIR grating-based hyperspectral imagers. Results show that the proposed method can preserve the original spectral shape, improve the image quality, and reduce the fringe amplitude in the 700–1000 nm region from about ±23% (10.7% RMSE) to about ±4% (1.9% RMSE). This method is particularly useful for spectra taken through a slit with a grating and shows flexible adaptability to object data, which suffer from time-varying interference fringes.
This is an open access article distributed under the Creative Commons Attribution License
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited