基于图论的参数曲线集目标区域识别方法  

Method for target region identification based on graph theory from parameterized curve set

作　　者：谭正华[1,2] 王李管[1,2] 毕林[1] 陈建宏[1] 

机构地区：[1]中南大学资源与安全工程学院,长沙410083 [2]中南大学数字矿山研究中心,长沙410083

出　　处：《计算机工程与应用》2010年第32期192-196,共5页Computer Engineering and Applications

基　　金：国家自然科学基金(No.50774092)~~

摘　　要：参数曲线集将平面划分为多个区域,区域的准确识别和提取是矢量图案填充的基础。目前,已经有多种光栅图形学区域识别的成熟算法,但准确、快速地识别任意复杂参数曲线集目标区域还比较困难,其关键问题是:如何利用计算机有效地描述参数曲线集划分的多个区域(内含孔、岛)以及准确地识别目标区域。从空间认知的需求出发,采用图论中的多叉树形式化表达区域的空间组织结构,在建立定向闭环的基础上,给出了目标区域的自动识别方法。实验结果表明:该方法准确、有效,具有普遍适用性,在Dimine数字矿山软件中得到了良好的应用。Identification for region generated by parameterized curve set in plane is the base of pattern filling.At present,several algorithms for area-filling in raster graphics field are proposed,but there are difficulties of detection for region generated by parameterized curve set.Based on the spatial theory,some issues about the region identification are studied.The main researches concentrate on representation and relations of regions.Based on the concepts and property of directional closed loop,this paper proposes an algorithm for region identification and plane parameterized curve set.Experiments show this algorithm has wide adaptability and is easy to implement,and has been applied in Dimine software.

关 键 词：参数曲线 区域识别 多叉树 定向闭环 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]