Guided total variation approach based non-uniformity correction for infrared focal plane array

Abstract

Fixed pattern noise (FPN) in infrared images seriously degrades the imaging quality and visual effect of infrared focal plane arrays (FPAs). Although many scene-based non-uniformity correction (NUC) algorithms have been developed recent years, the convergence speed of the bias and gain correction parameters still need to be further improved. In this paper, we present a novel NUC approach for IR FPAs which minimizes the total variation of the estimated IR irradiance guided by a noise model image, and we name it guided total variation (GTV) NUC method. A temporal detection factor is introduced to NUC procedure to prevent NU parameters updating when scene movement stops. In the proposed scheme, the correction parameters of the FPN are estimated via an iterative optimization strategy, frame by frame. The experimental results of synthetic and real IR videos demonstrate that the proposed algorithm have better NUC performance in terms of fewer ghosting artifacts and faster convergence than the state-of-the-art methods.