铁路综合视频图像去雾算法研究与探讨
|
摘要
受雾霾等复杂介质环境影响,铁路视频监控系统获得的视频图像降质严重,使得雾霾天图像复原方法研究成为亟待解决的关键性问题。铁路雾霾视频监控图像具有分辨率低、灰度分布集中等主要特点,深入研究分析直方图均衡算法、Retinex图像增强算法和暗通道先验去雾算法的图像处理原理,分析图像处理效果。利用3种算法对铁路室外图像进行分析处理,结果表明3种算法均可以实现去雾,直方图均衡算法存在颜色失真和光晕现象; Retinex图像增强算法清晰度最好,但处理后的图像存在部分失真;暗通道先验去雾算法处理图像较为自然。
Affected by the complex medium environment such as fog and haze, the video image obtained by the railway video surveillance system is seriously degraded, which makes it a key issue to restore the degraded images, and the study on restoration method very urgent. This paper discusses the main features such as low-resolution and gray-scale distribution of the railway video surveillance image in haze day, analyses the image processing principle and the effects of the three algorithms, i. e., histogram equalization, Retinex improvement for image enhancement, and image haze removal using dark channel prior. Outdoor images of the railway are processed by the three algorithms, and the results show that all the three algorithms are effective in haze removal. Histogram equalization features color distortion and halo. Though the enhanced images using Retinex improvement have the best resolution, they are partially distorted. The images processed by haze removal using dark channel prior are more natural.
Affected by the complex medium environment such as fog and haze, the video image obtained by the railway video surveillance system is seriously degraded, which makes it a key issue to restore the degraded images, and the study on restoration method very urgent. This paper discusses the main features such as low-resolution and gray-scale distribution of the railway video surveillance image in haze day, analyses the image processing principle and the effects of the three algorithms, i. e., histogram equalization, Retinex improvement for image enhancement, and image haze removal using dark channel prior. Outdoor images of the railway are processed by the three algorithms, and the results show that all the three algorithms are effective in haze removal. Histogram equalization features color distortion and halo. Though the enhanced images using Retinex improvement have the best resolution, they are partially distorted. The images processed by haze removal using dark channel prior are more natural.
引文
[1]王海东.夜晚铁路图像增强技术的研究[D].北京:北京交通大学,2017:1-6.
[2]李正强,陈兴峰,马龙天,等.光学遥感卫星大气校正研究综述[J].南京信息工程大学学报(自然科学版),2018,10(1):6-15.
[3]刘吉.铁路综合视频监控系统的应用及发展趋势探讨[J].自动化与仪器仪表,2016(6):110-111.
[4]成晋军,张晓娟.雾霾天气图像去雾霾方法比较研究[J].山西电子技术,2018(1):82-83,87.
[5]何旭升.实时视频监控去雾算法研究与系统实现[D].西安:西安电子科技大学,2017:1-2.
[6]禹晶,徐东彬,廖庆敏.图像去雾技术研究进展[J].中国图象图形学报,2011,16(9):1561-1576.
[7]汪杰君,杨杰,张文涛,等.雾天偏振成像影响分析及复原方法研究[J].激光技术,2016,40(4):521-525.
[8]谢娜.基于图像增强的图像去雾算法研究[J].机械设计与制造工程,2017,46(12):31-33.
[9]Li Yi,Zhang Yunfeng,Geng Aihui,et al.Infrared image enhancement based on atmospheric scattering model and histogram equalization[J].Optics and laser technology,2016,83(3):99-107.
[10]张宝山,杨燕,陈高科,等.结合直方图均衡化和暗通道先验的去雾算法[J].传感器与微系统,2018(3):148-152.
[11]Lin Haoning,Shi Zhenwei.Multi-scale retinex improvement for nighttime image enhancement[J].Optik-International Journal for Light and Electron Optics,2014,125(24):7143-7148.
[12]He Kaiming,Sun Jian,Tang Xiaoou.Single image haze removal using dark channel prior[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2011,33(12):2341-2353.
[13]解书凯,赵红军,李莉娟.基于暗原色先验的雾天图像清晰度复原[J].激光杂志,2018(7):100-104.
[14]刘衍琦,詹福宇.Matlab图像与视频处理实用案例详解[M].北京:电子工业出版社,2015.
[15]汪秦峰.基于直方图均衡化和Retinex的图像去雾算法研究[D].西安:西北大学,2016:25-35.
[16]曹永妹,张尤赛.图像去雾的小波域Retinex算法[J].江苏科技大学学报(自然科学版),2014,28(1):50-55,62.
[17]甘玉泉,汶德胜,王乐,等.暗通道自然灾害遥感图像去雾[J].光子学报,2015,44(6):53-57.
[18]胡伟,袁国栋,董朝,等.基于暗通道优先的单幅图像去雾新方法[J].计算机研究与发展,2010,47(12):2132-2140.
[19]杜以清.基于暗原色先验与Retinex的图像去雾算法及改进[D].西安:西北师范大学,2015:3-4.
[20]张驰宇,贾银亮,梁康武.基于暗通道的单幅图像快速去雾算法[J].电子测量技术,2017,40(10):143-147.
[21]韩宇,王旭,阳曲.雾霾降质图像去雾算法的研究[J].科技与创新,2017(4):28-29.
[2]李正强,陈兴峰,马龙天,等.光学遥感卫星大气校正研究综述[J].南京信息工程大学学报(自然科学版),2018,10(1):6-15.
[3]刘吉.铁路综合视频监控系统的应用及发展趋势探讨[J].自动化与仪器仪表,2016(6):110-111.
[4]成晋军,张晓娟.雾霾天气图像去雾霾方法比较研究[J].山西电子技术,2018(1):82-83,87.
[5]何旭升.实时视频监控去雾算法研究与系统实现[D].西安:西安电子科技大学,2017:1-2.
[6]禹晶,徐东彬,廖庆敏.图像去雾技术研究进展[J].中国图象图形学报,2011,16(9):1561-1576.
[7]汪杰君,杨杰,张文涛,等.雾天偏振成像影响分析及复原方法研究[J].激光技术,2016,40(4):521-525.
[8]谢娜.基于图像增强的图像去雾算法研究[J].机械设计与制造工程,2017,46(12):31-33.
[9]Li Yi,Zhang Yunfeng,Geng Aihui,et al.Infrared image enhancement based on atmospheric scattering model and histogram equalization[J].Optics and laser technology,2016,83(3):99-107.
[10]张宝山,杨燕,陈高科,等.结合直方图均衡化和暗通道先验的去雾算法[J].传感器与微系统,2018(3):148-152.
[11]Lin Haoning,Shi Zhenwei.Multi-scale retinex improvement for nighttime image enhancement[J].Optik-International Journal for Light and Electron Optics,2014,125(24):7143-7148.
[12]He Kaiming,Sun Jian,Tang Xiaoou.Single image haze removal using dark channel prior[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2011,33(12):2341-2353.
[13]解书凯,赵红军,李莉娟.基于暗原色先验的雾天图像清晰度复原[J].激光杂志,2018(7):100-104.
[14]刘衍琦,詹福宇.Matlab图像与视频处理实用案例详解[M].北京:电子工业出版社,2015.
[15]汪秦峰.基于直方图均衡化和Retinex的图像去雾算法研究[D].西安:西北大学,2016:25-35.
[16]曹永妹,张尤赛.图像去雾的小波域Retinex算法[J].江苏科技大学学报(自然科学版),2014,28(1):50-55,62.
[17]甘玉泉,汶德胜,王乐,等.暗通道自然灾害遥感图像去雾[J].光子学报,2015,44(6):53-57.
[18]胡伟,袁国栋,董朝,等.基于暗通道优先的单幅图像去雾新方法[J].计算机研究与发展,2010,47(12):2132-2140.
[19]杜以清.基于暗原色先验与Retinex的图像去雾算法及改进[D].西安:西北师范大学,2015:3-4.
[20]张驰宇,贾银亮,梁康武.基于暗通道的单幅图像快速去雾算法[J].电子测量技术,2017,40(10):143-147.
[21]韩宇,王旭,阳曲.雾霾降质图像去雾算法的研究[J].科技与创新,2017(4):28-29.