亮度归一化在图像处理中的应用  

Application of Brightness Normalization in Image Processing

作　　者：蔡建新[1] 汪仁煌[1] 黄颖怡[1] 

机构地区：[1]广东工业大学自动化学院,广东广州510090

出　　处：《广东工业大学学报》2008年第4期65-68,共4页Journal of Guangdong University of Technology

摘　　要：针对工业现场图像识别系统中光照不均匀、不稳定的问题,提出了一种基于二次曲线拟合和直方图统计的图像亮度归一化方法.利用二次曲线拟合的方法将同一幅图像的各个部位均衡化,再应用直方图统计的方法将不同光照图像归一化到相同的亮度状态,从而使得光照强度不再影响检测结果.Regardirer the problem of varying lighting in industrial Image processing pipeline system, this paper introduces a brightness normalization method of image by using conics fitting and statistical moments of histogram. First, it used conics fitting method to equalize the image＇ s brightness of each section, and then adjusted different illumination images to the same brightness level by applying the histogram statistics method. Thus, the strength of illumination no longer affects the test result.

关 键 词：图像识别 光照不稳定 二次曲线拟合 直方图统计 亮度归一化 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]