Automated model-based calibration of short-wavelength infrared (SWIR) imaging spectrographs.

Among the variety of available hyperspectral imaging systems, the line-scan technique stands out for its short acquisition time and good signal-to-noise ratio. However, due to imperfections in the camera lens and, in particular, optical components of the imaging spectrograph, the acquired images are spatially and spectrally distorted, which can significantly degrade the accuracy of the subsequent hyperspectral image analysis. In this work, we propose and evaluate an automated method for correction of spatial and spectral distortions introduced by a line-scan hyperspectral imaging system operating in the short wavelength infrared (SWIR) spectral range from 1000 nm to 2500 nm. The proposed method is based on non-rigid registration of the distorted and reference images corresponding to two passive calibration objects. The results of the validation show that the proposed method is accurate, efficient, and applicable for calibration of line-scan hyperspectral imaging systems. Moreover, the design of the method and of the calibration objects allows integration with systems operating in diffuse reflectance or transmittance modes.