改进二值化算法在铁路路况图像处理中的应用  

Improved image binarization in railway image processing

作　　者：崔建荣[1] 阮双琛[1] 关惠玲[1] 

机构地区：[1]深圳大学工程技术学院,广东深圳518060

出　　处：《中国工程机械学报》2006年第2期233-237,共5页Chinese Journal of Construction Machinery

基　　金：广东省科技基金(2005B10201009)

摘　　要：针对铁路路况识别系统实时性要求高,图像受太阳光照射、阴影遮挡等外界条件影响大的特点,结合了两种对CTSU法的改进,提出了一种新的改进算法。改进二值化算法首先利用快速二值化算法找到一个初始的阈值,然后利用类内方差和图像灰像素的变化率的关系来判断是否对图像进行继续分割,直到找到最佳阈值为止。此法可在提高算法的适应性的同时使计算速度得到改善,实验结果表明这一算法效果良好。The railway image are affected by sunlight and shadow.In order to achieve good real-time of the system of automatic railway condition recognition and get perfect railway binarization image,we proposed a new algorithm by combining two improved OTSU method.The new algorithm find a initial threshold value through the fast binarization method.Then,using the difference between maximization of interclass variance and gray level to determine whether segment continue or not.The new method may get good adaptability and high efficiency.The method were tested with a lot of image,and result indicated that the algorithm is valid.

关 键 词：图像分割 改进二值算法 快速二值算法 铁路图像 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论] TP391[自动化与计算机技术—计算机科学与技术]