摘要
针对传统去雾算法处理图像后存在颜色不均衡、能见度较低等问题,提出一种基于雾线先验的双边滤波优化透射率算法。首先,将像素在RGB空间中聚类成雾线并引入自适应模块对大气光值进行预估。将大气散射模型和上下文正则化原理相结合,对图像透射率进行初步优化,同时基于最小通道对透射率进行更正,使得传输率图更加平滑,防止相邻景深区域透射率差距过大;再经过双边滤波对透射率进行二次优化,使其变得更加精准。然后,将大气光值和透射率输入到大气散射模型进行去雾处理得到无雾图像。最后,将去雾后的图像进行色彩增强,以提升图像的色彩真实性和亮度。实验结果表明,所提方法在主观上提升了人眼的视觉效果,在客观评价指标结构相似性(SSIM)、峰值信噪比(PSNR)、角点检测数、通用质量指数(UQI)、自然图像质量评估(NIQE)和处理时长上均有着显著的优越性。
For most traditional defogging algorithms,there are problems such as color imbalance and low visibility after processing images.In this paper,a bilateral filtering transmttance optimization algorithm based on fog line priori is proposed.Firstly,the pixels are clustered into fog lines in RGB space,and the adaptive module is introduced to predict the atmospheric light value.Secondly,the transmittance of the image is preliminarily optimized according to the combination of atmospheric scattering model with context regularization,At the same time,the transmittance is corrected based on the minimum channel,which makes the transmittance map smoother and prevents the transmitance difference between adjacent depth-of-field areas from being too large.Then,the transmittance is optimized for the second time through bilateral filtering to make it more accurate.Moreover,the atmospheric light value and transmittance are input into the atmospheric scattering model for defogging to obtain a clear image.Finally,the defogged image is color-enhanced to improve the color authenticity and brightness of the image.The experimental results show that this method improves the subjective visual effect to the human eye.This method has significant advantages in the objective evaluation indicators such as Structural Similarity(SSIM),Peak Signal-to-NoiseRatio(PSNR),corner detection number,Universal Quality Index(UQI),Natural Image Quality Evaluator(NIQE)and processing time.
作者
任晓奎
乐明杰
赵振禹
陶志勇
REN Xiaokui;LE Mingjie;ZHAO Zhenyu;TAO Zhiyong(School of Electronic and Infomation Engineering,Liaoning Technical University,Huludao 125000,China)
出处
《电光与控制》
CSCD
北大核心
2023年第6期47-54,共8页
Electronics Optics & Control
基金
国家重点研发计划项目(2018YFB1403303)。
关键词
图像去雾
雾线先验
透射率优化
自适应
image defogging
fog line priori
transmittance optimization
self-adaption
作者简介
任晓奎(1965-),男,辽宁阜新人,硕士,副教授,硕导;通讯作者:乐明杰(1998-),男,浙江舟山人,硕士生。
