Single Image Dehazing Based on Generic Regularity

TL;DR

This paper introduces a hybrid single image dehazing technique combining dark channel prior and color line models, improving robustness and performance, especially in sky regions, with less computational demand than deep learning methods.

Contribution

It proposes a novel combination of DCP and color line models with NN regularization for effective, robust dehazing without deep learning.

Findings

Outperforms recent deep learning methods on benchmark datasets.

Effective in removing haze from sky regions.

Requires less computational resources than deep learning approaches.

Abstract

This paper proposes a novel technique for single image dehazing. Most of the state-of-the-art methods for single image dehazing relies either on Dark Channel Prior (DCP) or on Color line. The proposed method combines the two different approaches. We initially compute the dark channel prior and then apply a Nearest-Neighbor (NN) based regularization technique to obtain a smooth transmission map of the hazy image. We consider the effect of airlight on the image by using the color line model to assess the commitment of airlight in each patch of the image and interpolate at the local neighborhood where the estimate is unreliable. The NN based regularization of the DCP can remove the haze, whereas, the color line based interpolation of airlight effect makes the proposed system robust against the variation of haze within an image due to multiple light sources. The proposed method is tested on benchmark datasets and shows promising results compared to the state-of-the-art, including the deep learning based methods, which require a huge computational setup. Moreover, the proposed method outperforms the recent deep learning based methods when applied on images with sky regions.