Abstract

Removing fog components from maritime images is more challenging than removing them from land images because of the complicated scattering effects. In this paper, we propose a single maritime image defogging algorithm based on illumination decomposition. First, we decompose a foggy input image into glow and glow-free layers by formulating an optimization problem and then solving it effectively, where the optimization variables are updated by closed-form solutions. Then, the fog components in the glow-free layer are removed using a defogging algorithm. Finally, the natural illumination in the glow layer is recovered using a compensation scheme to obtain the final fog-free image. Experimental results on real test images demonstrate that the proposed algorithm outperforms state-of-the-art algorithms in terms of both subjective and objective evaluation criteria.

Highlights

  • T HE IMAGES captured in adverse weather conditions, such as fog and haze, generally show low contrast and altered natural colors because of the scattering of atmospheric particles [1]

  • We propose a maritime defogging algorithm to address the aforementioned limitations of the conventional algorithms based on illumination decomposition inspired by the glow component analysis in [26]

  • We develop an effective illumination decomposition to separate an image into the glow and fog components for maritime image defogging

Read more

Summary

INTRODUCTION

T HE IMAGES captured in adverse weather conditions, such as fog and haze, generally show low contrast and altered natural colors because of the scattering of atmospheric particles [1]. A generative adversarial network and its variants [23], [24] have been shown to be highly effective for defogging while preserving textural details and natural colors These aforementioned defogging algorithms effectively remove fog components in the images, they focus primarily on images captured on land. Ma et al [27] fused high-boost and linear stretched images to enhance input foggy maritime images Their algorithm may yield unnatural results and fail to remove fog components effectively because it ignores the scattering effect of sea fog. Hu et al [26] developed an illumination decomposition technique to eliminate the glow components from maritime images caused by the multiple scattering effect.

PROPOSED ALGORITHM
ILLUMINATION DECOMPOSITION
FOG REMOVAL
ILLUMINATION COMPENSATION
DEFOGGED IMAGE SYNTHESIS
SUBJECTIVE ASSESSMENT
EFFECTS OF ILLUMINATION COMPENSATION ON DEFOGGING PERFORMANCE
APPLICATION TO DEFOGGING LAND IMAGES
CONCLUSIONS
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call