Polarimetric image restoration with adaptive backscatter light estimation and angle of polarization-based color correction.

Authors

Luoen Niu, Jinqin Zhong, Zheng Liang

Published in

Applied optics. Volume 65. Issue 32. Pages 11259-11272. Nov 10, 2026. Epub Oct 12, 2026.

Abstract

Optical electronic devices often exhibit certain limitations when attempting to capture underwater images, such as poor contrast, color casts, and blurry details. To improve the image quality of underwater scenes, polarization-based methods are introduced to enhance the underwater image. Nevertheless, the enhancement performance of most current methods is limited by their manual backscatter region selection processes and their disregard for color and brightness restoration. In this paper, we propose an underwater polarimetric image restoration method based on adaptive backscatter light estimation and angle of polarization (AoP)-driven color correction. Specifically, we construct an adaptive standard deviation of Gaussian filter-based on the entropy of the input image to capture the complete backscatter light. Afterward, we correct the color casts based on a reference channel and two gain coefficients, fully considering the color attenuation introduced by the difference in the AoP, and simultaneously utilize the AoP map and the gray-world assumption to spatially modulate the correction process. Finally, we comprehensively consider the brightness information of HSV and normalized backscatter light estimation to design an adaptive gamma revision function, thus naturally adjusting the enhancement results. The enhancement results on the underwater polarimetric image dataset show that our method achieves a good enhancement ability with high contrast, reasonable appearance, and natural brightness. Moreover, the extensive histogram distribution and image segmentation experiments demonstrate that our enhancement results have almost the same histogram distribution for the r, g, and b channels and can yield more accurate segmentation effects.

PMID:
43008152
Bibliographic data and abstract were imported from PubMed on 11 Nov 2026.

Read full publication at:
Please sign in to see all details.

Stats

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.