Optics Temperature-Dependent Nonuniformity Correction Via $\ell _{0}$-Regularized Prior for Airborne Infrared Imaging Systems

In this paper, we propose a new
$\ell _{0}$
-regularized approach to remove the temperature-dependent nonuniformity effects induced by the infrared (IR) imaging optics in an aerothermal environment. The
$\ell _{0}$
image prior is inspired by observing distinct characteristics of IR images with small targets. Based on this effective prior, we present a variational framework where we optimize an energy functional to estimate the optics-related fixed pattern noise (FPN) and the latent image. A computationally efficient numerical algorithm based on half-quadratic regularization is used to solve the optimization problem. The proposed method is fundamentally different from the existing nonuniformity correction techniques developed for infrared focal plane arrays and simultaneously suppresses the optics-related FPN and random noise. Both quantitative and qualitative comparisons to specialized state-of-art algorithms demonstrate its superiority.