摘要
传统的单尺度Retinex算法无法同时兼顾图像细节和颜色保真度,也无法凸显景深不同的雾天图像中远处景物的细节信息.针对此问题,提出了一种改进的单尺度Retinex算法对以上问题进行改善.首先,对原始雾天图像取反后进行改进的单尺度Retinex处理;然后将处理后的图像再取反后在HSI空间中对其饱和度分量进行拉伸.其中,改进的单尺度Retinex算法中运用可自适应调节的S型函数Sigmoid代替其中的log函数改善图像处理过程中的展宽效果.将输出结果图与原图相比,其对比度、信息熵、标准差都有所提高.仿真实验结果表明,所提算法对于雾天浓度不均匀的图像细节及颜色保真度尤其是图像远景的细节信息取得了较好的效果.
Since the traditional single-scale Retinex algorithm can neither retain the image detail and color fidelity,nor highlight the detail information of the distant foggy image of different depth of the field.In order to solve the above problems,firstly,the paper used the single-scale Retinex to process the original foggy negated image,and then stretch the saturation component in the HSI color space of the negated processed image.To improve its broadening effect,the improved single-scale Retinex algorithm used an adaptive function Sigmoid to substitute the log function.The simulated experiment shows that the algorithm has achieved good effects on details of the image with uneven concentrations in fog density and color fidelity,especially on the detail information of the distant foggy image.
作者
郭瑞
党建武
沈瑜
刘成
GUO Rui;DANG Jian-wu;SHEN Yu;LIU Cheng(School of Electronic and Information Engineering,Lanzhou Jiaotong University,Lanzhou 730070,China)
出处
《兰州交通大学学报》
CAS
2018年第6期69-75,共7页
Journal of Lanzhou Jiaotong University
基金
国家自然科学基金(61562057
61761027
51541902
51669010
61202314
61663021)
甘肃省自然科学基金(17JR5RA101)
长江学者和创新团队发展计划资助(IRT_16R36)
甘肃省"十三五"教育科学规划课题(GS[2016]GHB0217)
