基于天空分割的单幅交通标志图像去雾算法

doi:10.3778/j.issn.1002-8331.2006-0388

摘要/Abstract

摘要：

针对现有去雾算法应用于交通标志图像时容易产生信息丢失、色彩失真等问题，导致去雾后图像质量较低，不能很好地满足交通标志识别系统（TSRS）的实际应用需求，提出一种基于天空分割的单幅交通标志图像去雾算法。根据大津算法结合图像灰度特征得到自适应阈值实现天空区域和非天空区域的准确分割；非天空区域采取改进的暗通道先验算法去雾，引入自适应中值滤波和快速双边滤波联合的方法优化透射率，天空区域则采取直方图均衡化算法去雾；通过融合得到无雾图像；引入高斯滤波对严重降质图像进行去雾后清晰化处理。实验结果表明，去雾后图像在峰值信噪比等多个客观评价指标上的综合表现优于其他几种去雾方法，所提算法在保证较低的时间复杂度的同时，能有效地保留图像信息，还原出清晰的真实图像，满足TSRS的实际应用需求。

关键词: 交通标志图像去雾, 天空分割, 暗通道先验, 直方图均衡化, 高斯滤波

Abstract:

For the issues of information loss and color distortion when the existing defogging algorithms are applied to traffic sign images, which lead to the lower quality of the defogged images and can not meet the practical application requirements of Traffic Sign Recognition System（TSRS）, a single traffic sign image defogging method based on sky segmentation is proposed. Firstly, according to Otsu algorithm and image gray features, an adaptive threshold is obtained to realize accurate segmentation of sky region and non-sky region. Non-sky area adopts improved DCP approach to defog, introducing adaptive median filter and fast bilateral filter to optimize transmittance, while sky area adopts histogram equalization algorithm to defog. A fog-free image is acquired by fusion. Finally, the Gaussian filter is introduced to clear the severely degraded image after defogging. Experimental results reveal that the comprehensive performance of defogged images in several objective evaluation indexes such as peak signal-to-noise ratio is better than other defogged methods. The proposed method can effectively preserve image information and restore clear images while ensuring lower time complexity, and meet the practical application requirements of TSRS.

Key words: traffic sign image defogging, sky segmentation, dark channel prior, histogram equalization, Gaussian filtering

李文龙，李兴广，胡冉冉，崔炜. 基于天空分割的单幅交通标志图像去雾算法[J]. 计算机工程与应用, 2021, 57(20): 221-228.

LI Wenlong, LI Xingguang, HU Ranran, CUI Wei. Single Traffic Sign Image Defogging Algorithm Based on Sky Segmentation[J]. Computer Engineering and Applications, 2021, 57(20): 221-228.

参考文献

[1] LI W L，LI X G，QIN Y Y，et al.Application of improved LeNet-5 network in traffic sign recognition[C]//Proceedings of the 3rd International Conference on Video and Image Processing，Shanghai.New York：ACM，2019：13-18.
[2] ZHANG W D，DONGL L，PAN X P，et al.Single image defogging based on multi-channel convolutional MSRCR[J].IEEE Access，2019，7：72492-72504.
[3] ZHOU J C，ZHANG D H，ZOU P Y，et al.Retinex-based laplacian pyramid method for image defogging[J].IEEE Access，2019，7：122459-122472.
[4] LING Z G，GONG J W，FAN G L，et al.Optimal transmission estimation via fog density perception for efficient single image defogging[J].IEEE Transactions on Multimedia，2018，20（7）：1699-1711.
[5] RAIKWA S C，TAPASWI S.Adaptive dehazing control factor based fast single image dehazing[J].Multimedia Tools and Applications，2020，79（1）：891-918.
[6] 李加元，胡庆武，艾明耀，等.结合天空识别和暗通道原理的图像去雾[J].中国图象图形学报，2015，20（4）：514-519.
LI J Y，HU Q W，AI M Y，et al.Image haze removal based on sky region detection and dark channel prior[J].Journal of Image and Graphics，2015，20（4）：514-519.
[7] 李尧羿，杜宇超，顾振飞.结合天空区域识别的单幅图像去雾方法[J].计算机工程与应用，2018，54（19）：204-215.
LI Y Y，DU Y C，GU Z F，Single image dehazing method via sky region recognition[J].Computer Engineering and Applications，2018，54（19）：204-215.
[8] 谭龙江.天空识别的改进暗通道先验单幅图像去雾算法研究[J].小型微型计算机系统，2019，40（1）：210-214.
TAN L J.Single image dehazing with sky-identified improved dark channel prior[J].Journal of Chinese Computer Systems，2019，40（1）：210-214.
[9] TAN R T.Visibility in bad weather from a single image[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition（CVPR），2008：1-8.
[10] FATTAL R.Single image dehazing[J].ACM Transactions on Graphics，2008，27（3）：1-9.
[11] HE K M，SUN J，TANG X O.Single image haze removal using dark channel prior[C]//Proceedings of Computer Vision and Pattern Recognition（CVPR），2009：1956-1963.
[12] HE K，SUN J，TANG X.Guided image filtering[C]//Proceedings of European Conference on Computer Vision.Berlin：Springer-Verlag，2010：1-14.
[13] 徐喆，陈美竹.改进的雾霾天气交通标志图像去雾算法[J].计算机应用，2017，37（8）：2329-2333.
XU Z，CHEN M Z.Modified image dehazing algorithm of traffic sign image in fog and haze weather[J].Journal of Computer Applications，2017，37（8）：2329-2333.
[14] 王园园，黄丽琼.基于双边滤波器的雾霾天气交通标志图像去雾算法[J].舰船电子工程，2019，39（12）：126-128.
WANG Y Y，HUANG L Q，Defogging algorithm for haze weather traffic sign image based on bilateral filter[J].Ship Electronic Engineering，2019，39（12）：126-128.
[15] THFAIL Z，KHURSHID K，SALMAN A，et al.Optimisation of transmission map for improved image defogging[J].IET Image Processing，2019，13（7）：1161-1169.
[16] 寸鹤林.暗通道先验去雾算法的研究与改进及其在视频图像去雾中的应用[D].昆明：云南大学，2018.
CUN H L.Research and improvement of image haze removal algorithm based on dark channel priors and its application in video haze removal[D].Kunming：Yunnan University，2018.
[17] XIAO J S，ZHU L，ZHANG Y Q.Scene-aware image dehazing based on sky segmented dark channel prior[J].IET Image Processing，2017，11（12）：1163-1171.
[18] 杨红，崔艳.基于开运算暗通道和优化边界约束的图像去雾算法[J].光子学报，2018，47（6）：244-250.
YANG H，CUI Y.Image defogging algorithm based on opening dark channel and improved boundary constraint[J].Acta Photonica Sinica，2018，47（6）：244-250.
[19] 寇大磊，钱敏，权冀川，等.基于多尺度卷积网络的快速图像去雾算法[J].计算机工程与应用，2020，56（20）：191-198.
KOU D L，QIAN M，QUAN J C，et al.Fast image dehazing algorithm based on multi-scale convolutional network[J].Computer Engineering and Applications，2020，56（20）：191-198.
[20] 黄鹤，郭璐，王会峰，等.无人机航拍图像雾霾天空和白色区域自适应处理方法：CN109087254A[P].2018-12-25.
HUANG H，GUO L，WANG H F，et al.Adaptive processing method for haze sky and white area in aerial images of unmanned aerial vehicles：CN109087254A[P].
2018-12-25.
[21] CHAUDHURY K N，SAGE D，UNSER M.Fast O（1） bilateral filtering using trigonometric range kernels[J].IEEE Transactions on Image Processing，2011，20（12）：3376-3382.
[22] WANG Z Y，FU Y，HUANG T S.Deep learning through sparse and low-rank modeling[M].[S.l.]：Elsevier Inc，2019：253-255.