低照度液晶屏幕图像增强算法  

Image enhancement algorithm for low illumination LCD screen

作　　者：张薇薇 王彦 张庆伟 付龙明 黄辉 ZHANG Weiwei;WANG Yan;ZHANG Qingwei;FU Longming;HUANG Hui(Guiyang Power Supply Bureau of Guizhou Power Grid Co.,Ltd.,Guiyang 550004 Guizhou,China;Beijing Join Bright Digital Power Technology Co.Ltd.,Beijing 100085,China)

机构地区：[1]贵州电网有限责任公司贵阳供电局,贵州贵阳550004 [2]中恒博瑞数字电力科技有限公司,北京海淀100085

出　　处：《电力大数据》2020年第5期23-29,共7页Power Systems and Big Data

摘　　要：为解决目前自然场景文本识别系统文字识别率较低的问题,从识别图像的质量角度分析,发现用相机拍摄的带有液晶屏幕的图像,易受到光照或屏幕本身显示问题影响,成像质量低,难以辨别屏幕内信息。传统的解决方法如使用Retinex算法进行图像增强时,由于Retinex算法本身的局限性,对于光照不均的图像,会出现光晕现象。针对此问题,本文提出基于同态滤波的去光晕Retinex算法来进行光照的校正,首先使用改进的同态滤波算法对图像进行不同程度的滤波处理,校正光照,处理后图像再使用改进的Retinex算法进行图像增强,则可以避免光晕现象。实验结果表明,本文方法可以有效提高低照度液晶屏幕图像的视觉效果,优于传统的Retinex算法,无光晕现象产生,进而提高自然场景文本识别系统的识别准确率。in order to solve the problem of low recognition rate of natural scene text recognition system at present,from the perspective of recognition image quality,it is found that images with LCD screen taken by camera are easily affected by lighting or display problems of the screen itself,and the image quality is low,so it is difficult to distinguish information on the screen.Traditional methods such as Retinex algorithm for image enhancement,due to the limitations of the Retinex algorithm itself,halo phenomenon will appear for images with uneven illumination.In response to this problem,this paper proposes Retinex algorithm based on homomorphic filtering to correct the illumination.Firstly,the improved homomorphic filtering algorithm is used to filter the image in different degrees,correct the illumination,and then the improved Retinex algorithm is used to enhance the image after processing,which can avoid the halo phenomenon.The experimental results show that this method can effectively improve the visual effect of low illumination LCD screen image,which is better than the traditional retinex algorithm,without halo phenomenon,and then improve the recognition accuracy of natural scene text recognition system.

关 键 词：低照度 屏幕图像 光晕现象 同态滤波 图像增强 

分 类 号：TM744[电气工程—电力系统及自动化]