重庆理工大学学报（自然科学版）Issue(6)：100-106,7.DOI:10.3969/j.issn.1674-8425(z).2015.06.019
基于超像素分割的图像去雾算法
Singie Image Dehazing Based on Superpixei Segmentation
摘要
Abstract
According to the halo artifacts and the long time computation when using the dark channel prior image dehazing algorithm,we proposed a new dehazing method from a single image using a seg-mentation algorithm based on superpixel. Firstly,we divided an image into many small parts. Second-ly,we got the more accurate transmission map in every small region based on the dark channel prior algorithm. Thirdly,we used the guiding filter in place of the soft matting algorithm in dark channel prior. Lastly,we recovered the haze image based on the atmosphere attenuation model. Experiments show that the presented method processes the haze image fast and removes the halo artifacts effective-ly,and it not only enhances the contrast,but also keeps the edges sharp.关键词
大气散射模型/暗原色先验/超像素/图像分割/去雾Key words
atmosphere scattering model/dark channel prior/superpixel/image segmentation/de-haze分类
电子信息工程引用本文复制引用
郑良缘,王平,高颖慧..基于超像素分割的图像去雾算法[J].重庆理工大学学报（自然科学版）,2015,(6):100-106,7.基金项目
国家自然科学基金资助项目 （）
