烟雾会降低能见度，为船舶火灾消防带来极大不便。研究图像实时去烟算法，并应用到视频监控或消防设备中，为消防行动提供视觉支持，对提高行动效率具有重大意义。本文提出一种图像实时去烟算法，在分析了船舶舱室火灾图像特点的基础上，基于传统Retinex理论，引入滚动导向滤波代替高斯滤波，取消对数计算，并结合图像金字塔设计了多尺度处理，通过对比度拉伸优化视觉效果。主、客观评价的结果表明，该算法去烟效果良好，优于直方图均衡化、多尺度Retinex、暗通道先验3种经典图像清晰化算法，且对640×480×3像素RGB图像的处理时间仅为95 ms。
Smoke will reduce visibility and bring great inconvenience to fire-fighting in the cabin. It is of great significance to study the real-time image de-smoke algorithm and apply it to video monitoring or fire-fighting devices to provide visual support for fire-fighting and improve the operation efficiency. Based on the analysis of the image characteristics of ship cabin fire, and the traditional Retinex theory, this paper proposed a real-time image de-smoke algorithm, which introduces rolling guidance filter instead of Gaussian filter, cancels logarithmic calculation, and designs multi-scale processing combined with image pyramid, and optimizes the visual effect by contrast stretching. The results of subjective and objective evaluation show that the algorithm has a good effect on smoke removal, which is better than histogram equalization, multi-scale Retinex and dark channel prior. The processing time of a 640×480×3 pixel RGB image is only 95 ms.
2022,44(22): 148-153 收稿日期：2021-09-06
DOI：10.3404/j.issn.1672-7649.2022.22.029
分类号：TP391.41
作者简介：何谦(1998-),男,硕士研究生,研究方向为图像处理
参考文献：
[1] 郑曦, 郑航. 无人机在消防灭火救援中的应用分析[J]. 低碳世界, 2020, 10(6): 209–210
[2] 张成. 基于多源信息融合的智能消防头盔关键技术研究[D]. 上海: 东华大学, 2016.
[3] 许骏. 面向火灾场景的图像去烟雾系统研究[D]. 上海: 东华大学, 2016.
[4] 李森. 火灾初期建筑内图像清晰化及人员检测技术研究[D]. 合肥: 中国科学技术大学, 2014.
[5] LI S, WANG S, ZHANG D, et al. Real-time smoke removal for the surveillance images under fire scenario[J]. Signal, Image and Video Processing, 2019.
[6] 马悦. 基于深度学习的火场灰度图像去烟算法[J]. 计算机与现代化, 2020(10): 64–68
[7] CHEN W T, YUAN S Y, TSAI G C, et al. Color channel-based smoke removal algorithm using machine learning for static images[C]// Proceedings of the 2018 25th IEEE International Conference on Image Processing. 2018: 2855-2859.
[8] ZHANG Q, SHEN X, XU L, et al. Rolling Guidance Filter[C]// European Conference on Computer Vision. Springer, Cham, 2014.
[9] 黄楠华. 恶劣环境下图像清晰化处理研究及应用[D]. 西安: 西安理工大学, 2017.
[10] 陈本豪, 高涛, 卢玮, 等. 基于雾天图像退化模型的自适应参数优化的去雾算法[J]. 科学技术与工程, 2019, 19(21): 219–227
[11] CHOI Kwon, JAEHEE Y, BOVIK A C. Referenceless prediction of perceptual fog density and perceptual image defogging. [J]. IEEE Transactions on Image Processing : a Publication of the IEEE Signal Processing Society, 2015, 24(11).
[12] LARSON G W, RUSHMEIER H, PIATKO C. A visibility matching tone reproduction operator for high dynamic range scenes[J]. IEEE Transactions on Visualization & Computer Graphics, 1997, 3(4): 0–306
[13] KIM J Y, KIM L S, HWANG S H. An advanced contrast enhancement using partially overlapped sub-block histogram equalization[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2001, 11(4): 475–484
[14] JUN Y C, HUANG K C Y, BRONGERSMA M L. Plasmonic beaming and active control over fluorescent emission[J]. Nature Communications, 2011, 2: 283
[15] JOBSON D J, RAHMAN Z U, WOODELL G A. Properties and Performance of a Center/Surround Retinex[J]. IEEE Transactions on Image Processing, 1997, 6(3): 451–462
[16] MALLAT S. A Wavelet Tour of Signal Processing, Third Edition: The Sparse Way[M]. Academic Press, 2008.
[17] 杨溪. 基于深度学习的图像增强技术研究[D]. 大连: 大连海事大学, 2020.
[18] JOBSON D J, RAHMAN Z, WOODELL G A. A multiscale retinex for bridging the gap between color images and the human observation of scenes[J]. IEEE Transactions on Image Processing, 1997, 6(7): 965–976
[19] 崔芮. 基于金字塔结构的人脸识别算法研究[D]. 西安: 西安电子科技大学, 2014.
[20] 郑婷. 基于人眼视觉感知图像对比度增强算法的研究[D]. 成都: 电子科技大学, 2016.
[21] 孙瑶. 基于FPGA图像显示的双线性插值算法的设计与实现[D]. 南京: 东南大学, 2017.
[22] HAUTIERE N, TAREL J P, AUBERT D, et al. Blind contrast enhancement assessment by gradient ratioing at visible edges. (Report)[J]. Image Analysis & Stereology, 2011, 27(2): 87–95
[23] ZIAEI NAFCHI H, CHERIET M. Efficient no-reference quality assessment and classification model for contrast distorted images[J]. IEEE Transactions on Broadcasting, 2018: 1–6
