快速图文分离算法  

Fast Algorithm for Text/Photo Separation

作　　者：王琴琴[1] 徐家恺[1] 

机构地区：[1]南京大学电子科学与工程系,南京210093

出　　处：《科学技术与工程》2008年第8期2089-2092,2100,共5页Science Technology and Engineering

摘　　要：地图中的道路和文字具有不同的特征，在很多情况下，如果一起处理便会产生一些问题。基于对数学形态学各基本运算的研究，提出了一种简易快速灵活的图文分离算法。数学形态学被广泛地运用于数字图像处理，基于数学形态学的图文分离算法的研究已有先例。但是传统的算法要用到近50个结构元素，运行起来将消耗系统很多资源，原理也很复杂。就算法只需要根据栅格地图的大概特征选择1～3个结构元素和合适的参数，运用数学形态学最基本简单的膨胀腐蚀操作便能达到很好的分离效果，计算快速便捷，适应范围广，实用性非常强。There are differences problems if they are treated equally between the roads and characters in the maps. In many conditions, there will be A simple fast flexible algorithm for text/photo separation based on the study on mathematical morphology is come up. Mathematical morphology has been widely used to process digital image. Someone has made the research of the Text/Photo separation based on mathematica morphology, but the national algorithm needs almost 50 structural elements to proceed, so it＇ s quite a waste of the system resources, and the principle is complex. This algorithm only needs to choose one to three structural elements and suitable parameters via the general characters of the map, a satisfied result of text/photo separation can get after using the most simple operation of mathematical morphology. It＇s easy and fast to culcalate and the application can be widly used and practical.

关 键 词：图文分离 数学形态学 结构元素 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]