Abstract:Haze can seriously affect the visible and visual quality of outdoor images. As a challenge in practice, image dehazing techniques are always used to remove haze from the captured images. Existing image dehazing algorithms focus on enhancing both global image contrast and saturation, but ignore the local enhancement. So the dehazed images do not often have good performance in the visual quality of local details. This paper proposes a new single-image dehazing solution based on the adaptive structure decomposition integrated multi-exposure image fusion (PADMEF). A set of underexposed image sequences are extracted from a single blurred image first by a series of gamma correction and the spatial linear adjustment of saturation. Then different exposure-level images are fused into a haze-free image by applying a multi-exposure image fusion (MEF) scheme based adaptive structure decomposition to each image patch. The proposed image dehazing scheme can effectively eliminate the visual degradation caused by haze without the physical model inversion of haze formation. Both apriori estimation of scene depth and the expensive refinement process of depth mapping can be avoided. The entropy of image texture named as texture energy is used to measure the image energy and obtain the information size contained in an image. Meanwhile, a texture energy based method is presented to adaptively select the corresponding patch size for the decomposition of image structure. In addition, this paper verifies that the dehazed images obtained by the patch based MEF algorithm always meet the requirements of intensity decrease. The comparative experiment results are evaluated in both qualitative and quantitative aspects, which confirm the effectiveness of the proposed solution in haze removal.

Single Image Numerical Iterative Dehazing Method Based on Local Physical Features

Nighttime Image Dehazing by Render.

Accelerated Haze Removal for a Single Image by Dark Channel Prior

Single image dehazing based on dark channel prior

Color-line and Dark Channel Based Dehazing for Remote Sensing Images

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

Single image dehazing algorithm based on sky segmentation and optimal transmission maps

Image Dehazing Based on Accurate Estimation of Transmission in the Atmospheric Scattering Model

Hierarchical deconvolution dehazing method based on transmission map segmentation

Single Image Dehazing Using Scene Depth Ordering

Multi-Scale Attention Feature Enhancement Network for Single Image Dehazing

A Self-Adaption Single Image Dehaze Method Based on Clarity-evaluation-function of Image

Single Image Dehazing in Inhomogeneous Atmosphere

Single Image Dehazing Method Via Sky-Regions Segmentation and Dark Channel Prior

Fast Single-Image Dehazing Method Based On Luminance Dark Prior

A Defogging Method of Single Foggy Image Based on Physical Model

Image Haze Removal Using Airlight White Correction, Local Light Filter, and Aerial Perspective Prior

Image Dehazing by an Artificial Image Fusion Method Based on Adaptive Structure Decomposition

High-quality Image Dehazing with Diffusion Model

Improved Single Image Dehazing Using Segmentation

A Haze Density Aware Adaptive Perceptual Single Image Haze Removal Algorithm.