Strength Pareto Evolutionary Algorithm II Based Gradient Channel Prior To Restore Hazy Images

Abstract: Images obtain in poor environmental circumstances has poor visibility. These images limit the performance of many imaging systems. Many techniques have been implemented in the literature to handle this issue. However, designing an efficient channel prior to restore hazy images is still an open area of research. The comprehensive review of the existing techniques has shown following gaps in the literature: The hyper-parameter tuning of Gradient channel prior has been ignored in the literature. An efficient tuning has an ability to improve the results further. Most of existing techniques still suffer from texture distortion issue. Therefore, a suitable gradient aware channel prior is proposed to handle these issues. Extensive experimental results show that the proposed technique has an ability to remove the limitations of existing techniques.

Keywords: Dehazing, Hazy images, Gradient, Channel prior.