Abstract:The phenomenon of atmospheric haze arises due to the scattering of light by minute particles suspended in the atmosphere. This optical effect gives rise to visual degradation in images and videos. The degradation is primarily influenced by two key factors: atmospheric attenuation and scattered light. Scattered light causes an image to be veiled in a whitish veil, while attenuation diminishes the image inherent contrast. Efforts to enhance image and video quality necessitate the development of dehazing techniques capable of mitigating the adverse impact of haze. This scholarly endeavor presents a comprehensive survey of recent advancements in the domain of dehazing techniques, encompassing both conventional methodologies and those founded on machine learning principles. Traditional dehazing techniques leverage a haze model to deduce a dehazed rendition of an image or frame. In contrast, learning-based techniques employ sophisticated mechanisms such as Convolutional Neural Networks (CNNs) and different deep Generative Adversarial Networks (GANs) to create models that can discern dehazed representations by learning intricate parameters like transmission maps, atmospheric light conditions, or their combined effects. Furthermore, some learning-based approaches facilitate the direct generation of dehazed outputs from hazy inputs by assimilating the non-linear mapping between the two. This review study delves into a comprehensive examination of datasets utilized within learning-based dehazing methodologies, elucidating their characteristics and relevance. Furthermore, a systematic exposition of the merits and demerits inherent in distinct dehazing techniques is presented. The discourse culminates in the synthesis of the primary quandaries and challenges confronted by prevailing dehazing techniques. The assessment of dehazed image and frame quality is facilitated through the application of rigorous evaluation metrics, a discussion of which is incorporated. To provide empirical insights, the study meticulously elucidates simulation results and presents an in-depth analysis of prominent dehazing techniques. By orchestrating a seamless integration of theoretical exposition and practical evaluation, this research contributes to the advancement of the field, paving the way for improved image and video fidelity in the presence of atmospheric haze.

Image dehazing via enhancement, restoration, and fusion: A survey

A Comprehensive Survey on Image Dehazing Based on Deep Learning.

A Comprehensive Survey and Taxonomy on Single Image Dehazing Based on Deep Learning

From Coarse to Fine: A Two-Stage Network for Dense Haze removal

A comprehensive qualitative and quantitative survey on image dehazing based on deep neural networks

Image Dehazing: Improved Techniques

Visual-quality-driven unsupervised image dehazing

Single Image Dehazing Based on Two-Stream Convolutional Neural Network

Advancing Real-World Image Dehazing: Perspective, Modules, and Training

Review of dehazing techniques: challenges and future trends

Single image mixed dehazing method based on numerical iterative model and DehazeNet

Hierarchical deconvolution dehazing method based on transmission map segmentation

Joint Contrast Enhancement and Exposure Fusion for Real-World Image Dehazing

Image Dehazing Technique Based on DenseNet and the Denoising Self-Encoder

An Enhancement in Single-Image Dehazing Employing Contrastive Attention over Variational Auto-Encoder (CA-VAE) Method

Model-Based Single Image Deep Dehazing

A new end-to-end image dehazing algorithm based on residual attention mechanism

Dehazing Remote Sensing and UAV Imagery: A Review of Deep Learning, Prior-based, and Hybrid Approaches

Single image dehazing via cycle-consistent adversarial networks with a multi-scale hybrid encoder-decoder and global correlation loss

An End-to-End Image Dehazing Method Based on Deep Learning

Adaptive feature fusion network based on boosted attention mechanism for single image dehazing