.png){kind=icon}
.png){kind=icon}
Return
Fusing neutral pixel prior and adaptive exposure for atmospheric light estimation under a decoupled atmospheric scattering model
.png){kind=icon}
PRE
AI
Z
Z
DOI：10.1117/1.JEI.35.2.023041.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Accurate atmospheric light (AL) estimation is crucial within the atmospheric scattering model (ASM) for dehazing, as it directly governs color fidelity and exposure quality of the dehazed results. Most ASM-based methods prioritize transmission estimation, often neglecting AL by assuming spatial uniformity and relying on simple distant-pixel approximations. This neglect of the distinct color and intensity properties of AL frequently causes residual color casts and exposure issues. To address these limitations, we propose: (1) a decoupled AL representation within a refined ASM, (2) a neutral pixel prior (NPP) for accurate AL color estimation to eliminate casts, and (3) global-local intensity estimation for balanced exposure. Finally, these components are integrated into the proposed AL estimation algorithm via NPP and adaptive exposure. Comparative evaluations demonstrate that our method outperforms existing AL recovery methods and enables high-quality dehazing.
Keywords:
atmospheric light
atmospheric scattering model
dehazing
exposure
color cast
Journal
J
IF:
1
Papers:
109
Citations:
2.7K
Organization
Cited Papers