摘要
由于广泛的实用价值与理论价值,高动态范围成像(HDRI)技术成为图像处理领域的一个研究热点,如何检测及去除成像过程中产生的干扰影像(即"鬼影")也引起了研究者的广泛关注。将鬼影检测与去除方法按照运动物体是否属于拍摄目标及算法操作域进行系统分类,将鬼影检测算法分为无参考图像、有参考图像两大类并将每一大类分为辐射域检测、图像域检测两类,将鬼影去除算法分为辐射域去除、图像域去除两大类,对其中每一类算法分别全面综述了其特点及最新研究成果,并对不同方法的优缺点进行了比较分析。最后总结了不同方法的适用条件并给出了关于这一课题未来发展的思考。
Because of its extensive practical and theoretical values, the high dynamic range imaging （HDRI） technique has become a hot topic in the areas of image processing. How to deal with the ghost produced in the process has attrac- ted many researchers＇ attentions. This paper respectively categorized ghost detection and removal methods according to whether the moving objects belong to the shooting target and operation domain into two categories. The ghost detection was classified into radiation field detection and image domain detection. The removal methods was classified into radia- tion field removal and image domain removal. For each category, the state-of-the-art research achievements were re- viewed comprehensively. We also analyzed the characteristics of different algorithms. Thoughts and foresights of this field were given at the end of this paper.
作者
胡胜男
张伟
刘侃
顾建军
HU Sheng-nan ZHANG Wei LIU Kan GU Jian-jun(School of Control Science and Engineering, Shandong University,Jinan 250061, China)
出处
《计算机科学》
CSCD
北大核心
2016年第8期13-18,共6页
Computer Science
基金
国家自然科学基金项目(61573222
61233014)
山东省科技重大专项(2015ZDXX0801A02)
江苏省重点实验室开放项目(3DL201502)
山东大学基本科研业务费专项资金(2016JC014)资助
关键词
高动态范围成像
鬼影
参考图像
辐射域
图像域
High dynamic range imaging, Ghost, Reference image, Radiance domain, Image domain
