Determining Smile And Keystone Of Lab Hyperspectral Line Cameras

Modern imaging spectrometers are compact, low weight, reasonably priced, and can be used in the laboratory. Accurate hyperspectral images require low noise, high calibration accuracy, and high response uniformity. Due to ageing of the optical and electronic components a periodic check is recommended. This paper presents a method for measuring the smile and keystone geometric distortions from test images. Measurements are obtained by using simple models for the smile and keystone, and existing SciPy functions for quadratic spline interpolation, normalized cross-correlation, and optimization to quickly find a solution for the models. Results show that the smile and keystone of a laboratory imaging spectrometer can be obtained with sub-pixel accuracy.