光学与光电技术, 2019, 17 (6): 45, 网络出版: 2020-01-07  

一种基于暗通道理论的海上图像去雾算法

A Dehazing Algorithm of Marine Images Based on Improved Dark Channel Prior Theory
作者单位
1 中国人民解放军92001 部队,山东青岛266011
2 国防科技大学前沿交叉学科学院,湖南长沙410073
摘要
针对船载光电侦察设备在海上取证过程中,易受海上雾气影响导致图像清晰度下降的问题,提出了一种基于暗通道理论的海上图像去雾算法。海天背景图像中较大面积的天空区域不满足暗通道先验理论而产生颜色失真,通过Canny 算子检测图像边缘,结合自适应滤波及像素判定方法将图像分割为天空区域和非天空区域,对两部分透射率分别进行计算,并采用天空区域像素平均值估算大气光强度。针对软抠图算法复杂度高、实时性差的缺点,采用改进的快速导向滤波对透射率图进行修正。实验结果表明,当天空区域调节系数为0.2,非天空区域调节系数为0.9 时,该算法可有效克服天空区域颜色失真及光晕等问题,同时提高了图像处理速度,对海上图像的去雾效果较为理想。
Abstract
When the shipborne photoelectric reconnaissance equipment is used to take evidence of other ships,the image clarity always reduces influenced by haze over the sea. A dehazing algorithm aiming at marine images based on the improved dark channel prior is proposed. Considering the large area of sky in the image does not meet the dark channel prior theory and color distortion is produced. Canny operator is adopted to detect the edge. Combining with adaptive filtering and pixel determination,the sky area is divided by calculating the transmittance of the two parts separately,and the ambient light intensity is estimated using the average of the sky area pixels. Then,to conquer the shortcoming of the slow processing speed of the soft map algorithm,the improved steering filter is used to correct the transmittance map. The experimental results show that if the accommodation coefficient of sky aera equals 0.2 while other place equals 0.9,the image obtained by the algorithm effectively overcomes the problem of color distortion in the sky area. The image processing speed is improved simultaneously,and the image dehazing effect is good for marine images.

吴国鹏, 李盾, 何宇龙, 周纪元. 一种基于暗通道理论的海上图像去雾算法[J]. 光学与光电技术, 2019, 17(6): 45. WU Guo-peng, LI Dun, HE Yu-long, ZHOU Ji-yuan. A Dehazing Algorithm of Marine Images Based on Improved Dark Channel Prior Theory[J]. OPTICS & OPTOELECTRONIC TECHNOLOGY, 2019, 17(6): 45.

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!