摘要
随着人类工业化的加剧,污染物排放量也随之加大,因此雾霾天气在一年中的比例越来越大。雾霾天气造成大面积能见度低,在雾霾天采集的图像,对比度和色彩都存在很大的问题,因此图像的去雾研究有很大的应用价值和研究价值。该文介绍了一种基于最大类间方差法的加权直方图均衡化图像去雾算法,第一步,使用最大类间方差法进行直方图阈值分割,第二步,综合考虑子直方图的灰度级和像素点个数,对阈值进行加权再调整,最后对每个直方图进行单独的直方图均衡化。实验证明,基于最大类间方差法的加权直方图均衡化图像去雾算法有较好的去雾作用,且信息熵得以保证。
The intensify of human industrialization and the increase of pollutants emissions,would easily result in smoggy days.The smog caused low visibility widespreadly,the images collected in smoggy days have great problems in the contrast and color,so the study of the image defog has great application value and research value.This paper proposes a image defog algorithm which use Otsu and weighted histogram equalization, the first step is to use the Otsu method for the threshold segmentation of histogram, secondly,considering the grayscale range and pixel number of both sub-histogram, the threshold is weighted and resized,finally, the sub-histogram is equalizing independently. Experimental results prove that the algorithm has better defog effect and the information entropy can be guaranteed.
作者
季洁
JI Jie (College of Computer and Information Engineering, Hunan University of Commerce,Changsha 410205,China)
出处
《电脑知识与技术》
2018年第2期164-167,共4页
Computer Knowledge and Technology
基金
湖南省自然科学基金项目,面向身份认证的掌纹识别方法研究(2016JJ2070)
湖南省教育厅科学研究重点项目,多类型特征协同的主题模型多粒度掌纹识别研究(16A114)
关键词
最大类间方差法
直方图均衡化
图像去雾
图像增强
累积分布函数
信息熵
Otsu
histogram equalization
inaage defog
innge enhancement
cumulative distribution function
information entropy