This paper describes a process for the radiometric calibration of gray scale images and linear transformations of gray scale images. We begin by defining radiometric calibration and show that it can be posed as a linear least squares problem. We then consider radiometric calibration using linear transformations of images. When the transform coefficients are real valued we show that the calibration problem is still a linear least squares problem. When the transform coefficients are complex valued the calibration problem is a rather simple non-linear least squares problem. We briefly discuss existing optimization software for solving both linear and non-linear least squares problems.
J. Howse, J. Hogden, P. Fasel, P. Kelly, and R. Fortson. Radiometric Calibration of Gray Scale Images. Los Alamos National Laboratory Technical Report LA-UR-06-7109, 2006. [ Abstract | PDF (160 KB) ]
