An image defogging algorithm based on gaussian mixture model and transmittance compensation

Xiaohui He; Jinbo Lu

doi:10.1117/12.2689514

31 July 2023 An image defogging algorithm based on gaussian mixture model and transmittance compensation

Xiaohui He, Jinbo Lu

Proceedings Volume 12747, Third International Conference on Optics and Image Processing (ICOIP 2023); 127471O (2023) https://doi.org/10.1117/12.2689514
Event: Third International Conference on Optics and Image Processing (ICOIP 2023), 2023, Hangzhou, China

Abstract

An improved image defogging method is proposed to address the problem that the use of dark channel a priori principle for image defogging is likely to cause distortion in the sky region and the overall darkness of the recovered image. In this paper, based on the idea of sky segmentation, we use Gaussian mixture model and EM algorithm to divide the fogged image containing sky region into sky region and non-sky region, and take the mean value of the pixel value of the sky part as the atmospheric light value after getting the sky region. The transmittance is estimated by calculating the RAS color channel to eliminate the effect of atmospheric scattering, based on which the transmittance compensation function is constructed, and the transmittance is corrected by using the constructed transmittance compensation function to reduce the transmittance estimation error. The experimental results show that the algorithm proposed in this paper has a good defogging effect.

Citation Download Citation

Xiaohui He and Jinbo Lu "An image defogging algorithm based on gaussian mixture model and transmittance compensation", Proc. SPIE 12747, Third International Conference on Optics and Image Processing (ICOIP 2023), 127471O (31 July 2023); https://doi.org/10.1117/12.2689514

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available