A closer look at a spectrographic wavelength calibration

Abstract

Methods to spectrally calibrate imaging spectrographs, also known as Hyperspectral Imagers (HSIs), are common and known, but few look into how well the wavelength calibration performs. When building hyperspectral instruments in-house, performing sufficient wavelength calibration is also necessary. This report therefore investigates how well different methods describe the pixel to wavelength relationship, using data from a simple set-up with argon and mercury spectral lamps. The methods are evaluated using Root Mean Square Error (RMSE). A linear fit between pixel index and wavelength, which is commonly assumed, is shown to describe the relationship well. Estimating the diffraction angle with an arctan function and calculating wavelength using the grating equation shows promising results with even lower RMSE values than the linear fit. This report is meant to aid others in calibrating their own hyperspectral imagers with limited resources by showing how wavelength calibration accuracy varies with different methods.