基于透射率归一化的弱光图像增强-天津大学期刊中心.PDFVIP

第 50 卷 第 9 期 天津大学学报 (自然科学与工程技术版) Vol. 50 No. 9 2017 年 9 月 Journal of Tianjin University (Science and Technology) Sep. 2017 DOI:10.11784/tdxbz201607033 基于透射率归一化的弱光图像增强 杨爱萍，宋曹春洋，张莉云，白煌煌，卜令勇 (天津大学电气自动化与信息工程学院，天津 300072) 摘 要 ：针对阴天或夜晚等弱光条件下拍摄的图像具有低信噪比、低对比度、强噪声等问题，在暗原色先验理论去 雾框架下，提出了一种基于透射率归一化的弱光图像增强方法．首先，针对弱光图像的特点，简化弱光图像增强模 型；然后，用点暗原色取代块暗原色求取透射率，并利用局部维纳滤波进行细化，以保留更多图像的细节；最后 由弱光图像直接对透射率进行归一化即可得到增强图像．实验结果表明，该方法运算简单，可有效增强弱光图像 且能保留更多的图像细节． 关键词 ：图像增强；点暗原色；透射率；归一化；维纳滤波 中图分类号：TP751.1 文献标志码 ：A 文章编号 ：0493-2137 (2017)09-0997-07 Low-Light Image Enhancement Based on Transmission Normalization Yang Aiping ，Song Caochunyang ，Zhang Liyun ，Bai Huanghuang ，Bu Lingyong (School of Electrical and Information Engineering ，Tianjin University ，Tianjin 300072 ，China) The pictures taken in low-light environment such as on a cloudy day or at night have very low SNR ，low Abstract ： contrast and strong noise．To solve the problem ，under the haze removal framework of dark channel prior theory ，a low-light image enhancement method based on transmission normalization is proposed．First ，according to the char- acteristics of low-light image ，the model of low-light image enhancement is simplified．Then the transmission is cal- culated by point dark channel instead of block dark channel and the transmission is further detailed by local Wiener filter in order to retain more image details．Finally ，The transmission is directly normalized by the low-light image to get the enhanced image．Experimental results demonstrate that the proposed method is effective not only in needing little computation to get the enhanced im