"De-ghosting" artifact in scene-based nonuniformity correction of infrared image sequences

Abstract

In this paper we present a new technique to improve the convergence and to reduce the ghosting artifacts based on constant statistics (CS) method. We propose to reduce ghosting artifacts and to speed up the convergence by using enhanced constant statistics method with the motion threshold. The key advantage of the method is based in its capacity for estimate detectors parameters, and then compensate for fixed-pattern noise in a frame by frame basics. The ability of the method to compensate for nonuniformity and reducing ghosting artifacts is demonstrated by employing video sequences of simulated and several infrared video sequences obtained using two infrared cameras. © 2011 Springer-Verlag.