一种改进的全局Retinex图像增强算法及其仿真研究.pdfVIP

第 21 卷第 4 期 系 统 仿 真 学 报© Vol. 21 No. 4 2009 年 2 月 Journal of System Simulation Feb., 2009 一种改进的全局 Retinex 图像增强算法及其仿真研究 1 1 2 乔小燕 ，姬光荣 ，陈 雾 （1.中国海洋大学信息科学与工程学院，青岛 266100 ；2.南京理工大学计算机科学与技术学院，南京 210094 ） 摘 要：Retinex 理论认为物体对光线的反射能力决定其颜色，故该算法就是从图像中获得物体的 反射性质来得到物体本来面貌。为了消除传统Retinex 算法假定两像素点入射光分量相同而导致图 像颜色失真的缺陷，提出了一种改进的Retinex 方法，该方法使用原像素灰度值与以它为中心高斯 平滑后的灰度值间的差异来校正原像素点。仿真结果及两种评价标准表明，该方法更具准确性， 增强效果优于传统Retinex 算法。 关键词：Retinex；图像增强；高斯滤波；像素校正 中图分类号：TP751 文献标识码：A 文章编号：1004-731X (2009) 04-1195-03 Improved Retinex Image Enhancement Algorithm and Simulation Study 1 1 2 QIAO Xiao-yan , JI Guang-rong , CHEN Wu (1. College of Information Science and Technology, Ocean University of China, Qingdao 266100, China; 2. College of Information Science and Technology, Nanjing University of Science and Technology, Nanjing 210094, China) Abstract: Retinex theory points that the color of each object is determined by its reflectance, and thus several Retinex-based algorithms are developed to uncover the real appearance of objects by estimating the reflectance of them in image. However, since they assume that components of the incidental lights on two different points are the same, the color distortion problem cannot be avoided. Now an improved Retinex algorithm was proposed to adjust each pixel by utilizing the difference between its original gray value and the one after Gaussian filtering. The simulation experiment and the two known evaluation criteria show that this a