您的位置:首页  > 论文页面

基于暗通道先验的车载实时去雾算法

发表时间:2018-11-30  浏览量:1241  下载量:368
全部作者: 罗凡,吴友宇,梁昌珺,熊振华,高琳钧,姜春言
作者单位: 武汉理工大学信息工程学院
摘 要: 暗通道先验理论的去雾算法有着良好的去雾效果,是诸多雾霾图像恢复算法的典型,但其在车载等方面的实际运用中存在诸多缺陷。为减小由于算法本身造成的白场现象和色调失衡等问题,研究提出一种针对此算法的优化和改进。该算法首先改良了透射率曲线,使图像亮度较大的区域恢复效果趋于平滑;再通过HSV色彩空间对色度的分割来替换原图的色彩比例,以此矫正失衡的亮度与色度;改进后的算法还利用各种快速算法简化了计算,从而实现了实时去雾。在OpenCV软件平台下,得到了良好的实验结果,证明了该算法的有效性。
关 键 词: 信息处理技术;图像恢复;视频去雾;暗原色先验;透射率;色度校正
Title: Vehicle-mounted real-time defogging algorithm based on dark channel priori
Author: LUO Fan, WU Youyu, LIANG Changjun, XIONG Zhenhua, GAO Linjun, JIANG Chunyan
Organization: School of Information Engineering, Wuhan University of Technology
Abstract: The defogging algorithm based on the dark channel prior theory has a good defogging effect and is typical in many haze image restoration algorithms. But it has many defects in the practical application of vehicle-mounted and other aspects. In order to reduce the white field phenomenon and tonal imbalance caused by the algorithm itself, this paper proposes an optimization and improvement for this algorithm. The algorithm firstly improves the transmissivity curve, which makes the recovery effect of the region with larger brightness of the image tend to be smooth. Then, the color ratio of the original image is replaced by the chromaticity segmentation through HSV color space to rectify the unbalanced brightness and chromaticity. The improved algorithm also utilizes a variety of fast algorithms to simplify calculations and to achieve real-time defogging. Under the Open CV software platform, good experimental results are obtained, which proves the effectiveness of the algorithm.
Key words: information processing technology; image restoration; video defogging; dark color prior; transmissivity; chroma correction
发表期数: 2018年11月第22期
引用格式: 罗凡,吴友宇,梁昌珺,等. 基于暗通道先验的车载实时去雾算法[J]. 中国科技论文在线精品论文,2018,11(22):2295-2303.
 
2 评论数 0
暂无评论
友情链接