基于USM的改进Retinex算法安检图像增强  

USM-based Improved Retinex Algorithm for Security Image Enhancement

作　　者：孙星星 郑俊褒[1] 曹志玲 SUN Xingxing;ZHENG Junbao;CAO Zhiling(School of Information,Zhejiang University of Science and Technology,Hangzhou 310018;Huangpu Customs Technology Center,Guangzhou 510000)

机构地区：[1]浙江理工大学信息学院,杭州310018 [2]黄埔海关技术中心,广州510000

出　　处：《计算机与数字工程》2021年第11期2358-2361,共4页Computer & Digital Engineering

摘　　要：为了提高安检图像质量,提出用Retinex算法对其进行增强处理,并对Retinex算法做了改进。改进算法首先结合反锐化掩模(Unsharp Masking,USM)算法中提高高频成分的思想,增强了安检图像的轮廓部分;然后用限制对比度自适应直方图均衡化(Contrast Limited Adaptive Histogram Equalization,CLAHE)算法增强图像对比度。实验结果和数据表明,相比传统Retinex算法增强后的安检图像,论文算法能在有效增强安检图像的同时保留图像局部信息。In order to improve the quality of security inspection images,it is proposed to use the Retinex algorithm to enhance it and improve the Retinex algorithm.The improved algorithm first combines the idea of increasing high-frequency components in the unsharp masking algorithm to enhance the contour portion of the security check image.Then,contrast limited adaptive histogram equalization algorithm is used to enhance the image contrast.The experimental results and data show that compared with the security image enhanced by the traditional Retinex algorithm,the algorithm in this paper can effectively enhance the security inspection im⁃age while retaining the image local information.

关 键 词：X射线图像 RETINEX 图像增强 USM CLAHE 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]