A Color Cast Image Enhancement Method Based on Affine Transform in Poor Visible Conditions
Zheng Liang,Xueyan Ding,Jie Jin,Yafei Wang,Yulin Wang,Xianping Fu
DOI: https://doi.org/10.1109/lgrs.2022.3156264
IF: 5.343
2022-01-01
IEEE Geoscience and Remote Sensing Letters
Abstract:In this letter, a simple yet effective dehazing framework is proposed, which consists of a novel color correction and a contrast enhancement. Most of the existing dehazing works focus on enhancing the contrast of the degraded images, but rarely of them concern about the color cast, which is ubiquitous in the scattering medium. To address the color distortion, an affine transform model-based color correction method is first proposed to improve the appearance of the image while preserving the details, which is inspired by the traditional color transfer. The color transfer alters the color values of a source image by sharing the global color statistics of a reference image, which makes it unsuitable to address the locally variable color deviations encountered in highly color distorted images as in poor visibility conditions (sandstorms and underwater). To alter color correction locally, we add local color fidelity and gradient constraint to the proposed technique, which overcomes the limitation that the traditional method depends too much on the global color statistics of the reference image and encourages it to handle the degraded image with various color casts and light conditions. In addition, a multiscale gradient-domain processing is applied to enhance the contrast. In this procedure, by extracting the information of different layers, we can easily restore the contrast while limiting the significant amplification of noise. The extensive qualitative and quantitative experiments reveal that the color and the contrast can be significantly improved by the proposed technique.
imaging science & photographic technology,remote sensing,engineering, electrical & electronic,geochemistry & geophysics