海上图像通常存在亮度较高的天空区域，故直接应用暗通道先验算法进行去雾处理的图像会在天空区域出现色斑和光晕。为此，本文提出一种基于暗通道先验的海上图像改进去雾方法。利用四分法进行大气光值的优化，即对图像进行四等分区域分割，并通过构建的评价函数来确定每次分割后的候选区域，且将符合阈值要求的候选区域像素均值作为大气光值估计值。针对暗通道模型估计出的透射率存在块效应问题，采用导向滤波器对透射率进行细化并将立体匹配和图像增强交替迭代过程应用其中。实验结果表明，相比于原始暗通道及一些其他算法，本文算法对天空区域的去雾效果更好，图像质量评价指标更高，立体匹配对数的优秀率提高显著，去雾后的图像包含更多的图像细节，更有利于无人船的精确测距与避障等后续工作。
Because of the sky region with high brightness, the image processed by dark channel prior algorithm will produce color spots and halos in the sky region. Therefore, this paper proposes an improved sea image defogging method based on dark channel prior. Firstly, the atmospheric light value is optimized by the quadrature method, that is, the image is divided into four equal regions, and the candidate regions are determined by the evaluation function, at the same time, considering the block effect of the transmission map estimated by the dark channel model, the average of the pixels in the candidate region which meets the threshold value is taken as the estimated atmospheric light value, a guided filter is used to refine the transmission map and the iterative process of stereo matching and image enhancement is applied. The experimental results show that, compared with the original dark channel and other algorithms, this algorithm has better defogging effect in the sky area, higher image quality evaluation index, and higher excellence rate of stereo matching pairs, after defogging, the image contains more image details, which is more beneficial to the follow-up work such as accurate distance measurement and obstacle avoidance of unmanned ship.
2021,43(10): 163-168 收稿日期：2020-09-24
DOI：10.3404/j.issn.1672-7649.2021.10.033
分类号：TP391.41
基金项目：中央军委装备发展部装备预研重点实验室基金项目(6142215200106)；国家自然科学基金资助项目(51579024)；中央高校基本科研业务费专项资金资助项目(3132019318，3132019344)
作者简介：赵红(1967-),女,教授,研究方向为电气传动与控制、船舶电力推进以及智能控制在电气工程领域的应用等
参考文献：
[1] LIU R, FAN X, HOU M, et al. Learning aggregated transmission propagation networks for haze removal and beyond[J]. IEEE Transactions on Neural Networks and Learning Systems, 2019
[2] 刘兴瑞. 自适应直方图均衡处理下船舶视频监控图像去雾算法[J]. 舰船科学技术, 2020, 42(16): 70–72
[3] KIM J H, JANG W D, SIM J Y, et al. Optimized contrast enhancement for real-time image and video dehazing[J]. Journal of Visual Communication and Image Representation, 2013, 24(3): 410–425
[4] 王博. 无人艇光视觉感知研究发展综述[J]. 舰船科学技术, 2019, 41(12): 44–49
[5] SINGH G, JAGGI N, VASAMSETTI S, et al. Underwater image/video enhancement using wavelet based color correction (WBCC) method[J]. IEEE Underwater Technology, 2015: 1–5
[6] WANG Y, SONG W, FORTINO G, et al. An experimental-based review of image enhancement and image restoration methods for underwater imaging[J]. arXiv, 2019
[7] TAN R T. Visibility in bad weather from a single image [C]// 2008 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2008), 24−26 June 2008, Anchorage, Alaska, USA. IEEE, 2008.
[8] FATTAL R. Single image dehazing[J]. Proceedings of SIGGRAPH, 2008, 27(3): 1–9
[9] HE K, SUN J, et al. Single image haze removal using dark channel prior[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2011, 33(12): 2341–2353
[10] SONG Y, LUO H, HUI B, et al. An improved image dehazing and enhancing method using dark channel prior [C]//Qingdao: The 27th Chinese Control and Decision Conference, 2015: 5840−5845.
[11] HAN M, LYU Z, QIU T, et al. A review on intelligence dehazing and color restoration for underwater images[J]. IEEE Transactions on Systems Man & Cybernetics Systems, 2018: 1–13
[12] CHEN X, YU J, KONG S, et al. Towards real-time advancement of underwater visual quality with GAN [J]. 2017.
[13] HE K, SUN J, TANG X. Guided image filtering[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2013, 35(6): 1397–1409
[14] 陈书贞, 任占广, 练秋生. 基于改进暗通道和导向滤波的单幅图像去雾算法[J]. 自动化学报, 2016, 42(3): 455–464
[15] LI S, KANG X, HU J. Image fusion with guided filtering[J]. IEEE Transactions on Image Processing, 2013, 22(7): 2864–2875
[16] ZHANG Q, LI X. Fast image dehazing using guided filter[C]//Hangzhou: IEEE International Conference on Communication Technology. IEEE, 2016.
[17] RUBLEE E, RABAUD V, KONOLIGE K, et al. ORB: An efficient alternative to SIFT or SURF[C]//Barcelona: International Conference on Computer Vision. IEEE, 2012.
