单幅图像去雾方法研究  

A Research on Defogging Methods with Single Image

作　　者：毕笃彦[1] 葛渊[2] 李权合[1] 任志河 南栋[1] 陈剑鹏[1] 

机构地区：[1]空军工程大学航空航天工程学院,陕西西安710038 [2]西北有色金属研究院金属材料多孔材料国家重点实验室,陕西西安710016 [3]空军后勤部司令部,北京100720

出　　处：《空军工程大学学报（自然科学版）》2013年第6期46-53,共8页Journal of Air Force Engineering University(Natural Science Edition)

基　　金：国家自然科学基金资助项目(61372167;61379104;61203268;61202339)

摘　　要：介绍了单幅图像去雾方法的研究现状、分析了基于增强方法和基于复复方法的一些经典图像去雾算法,指出了各种算法的优缺点。综合评价得出基于复原的图像去雾方法优于基于增强的图像去雾方法。针对现有的基于图像复原去雾方法提出了仍需要深入研究的问题,并从建立全面物理模型、探索模型求解的先验知识、设计基于人眼视觉机制的模型求解方法和图像去雾质量评价等几个方面分析如何突破图像去雾的关键技术。最后,对现有技术的发展趋势进行了分析,指出了去雾技术的研究方向。Current situation of single image defogging method are introduced,some classic algorithms based on image enhancement and image restoration are analyzed in detail,and advantages and disadvantages are listed in this paper.Comprehensive evaluation results show that the image restoration method is superior to the image enhancement method in the aspect of defogging.Aimed at the existing image restoration defog-ging method,thorough investigation and study are needed,and some problems about how to break through key technologies in defogging are analyzed from the establishment of a comprehensive physical model,the exploration of a prior knowledge of the model solution,the design of model solution methods based on human visual mechanism,and the evaluation of image defogging quality and some other aspects, etc..Finally,the development trend of the existing technology is analyzed,and the orientation of studying defogging technology is given.

关 键 词：雾霾 大气退化 图像去雾 图像恢复 图像增强 人眼视觉 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]