摘要
针对水下图像出现的颜色失真、对比度低、雾化现象等问题,提出双透射率成像模型与Retinex融合的水下图像清晰化方法。首先,采用基于改进双透射率成像模型的复原算法,用以解决图像雾化以及亮度失衡;其次,在带色彩恢复的多尺度Retinex增强算法中引入引导滤波,解决图像色偏问题;此外,引用自动色彩增强算法,有效提升对比度;最后,将三个输入图像与结合对比度、显著性、饱和性得到的对应权重图采用多尺度融合框架得到清晰化水下图像。实验结果表明,与现有新颖算法相比,所提方法可以最大程度地将多种单一算法的优势有效结合起来,水下彩色图像质量评价指标(underwater color image quality evaluation,UCIQE)均值高于各比较算法6.03%且加速鲁棒特征(speeded up robust features,SURF)特征匹配点明显提升,算法能在保留图像细节的同时有效校正色偏现象、提升图像对比度及清晰度,更符合人眼的视觉效果。
Aiming at problems of color distortion,low contrast and atomization phenomenon in underwater images,a underwater image sharpening based on fusion of double transmission imaging model and Retinex was proposed.Firstly,the restoration algorithm based on improved double transmission underwater imaging model was used to solve the image atomization and brightness imbalance.Secondly,the guidedfilter was introduced into the multi-scale Retinex enhancement algorithm with color recovery to solve the image color deviation.In addition,the automatic color enhancement algorithm was used to effectively improve the contrast.Finally,each input images was fused with the corresponding weight map that obtained by combining contrast,significance and saturation to obtain a clear underwater image.Experimental results show that compared with existing novel algorithms,the proposed method can effectively combine the advantages of multiple single algorithms with the largest scale.The average value of underwater image quality evaluation(UCIQE)index is 6.03%higher than the comparison algorithms and the speeded up robust features(SURF)index feature matching points are significantly improved.It can effectively solve the problems of color,sharpness and contrast while preserving the details of the image,which is more in line with the visual effect of human eye.
作者
林森
周天飞
LIN Sen;ZHOU Tian-fei(College of Automation and Electrical Engineering, Shenyang Ligong University, Shenyang 110159, China;College of Electronic and Information Engineering, Liaoning Technical University, Huludao 125105, China)
出处
《科学技术与工程》
北大核心
2021年第18期7627-7634,共8页
Science Technology and Engineering
基金
国家重点研发计划(2018YFB1403303)
沈阳理工大学引进高层次人才科研支持计划(1010147000915)。
