一种基于重心的快速细化算法  

A barycenter based fast thinning algorithm

作　　者：陈若珠[1] 田守军[1] 李战明[1] 吕强[2] 

机构地区：[1]兰州理工大学电气工程与信息工程学院,甘肃兰州730050 [2]长庆油田公司通信处,陕西西安710021

出　　处：《兰州理工大学学报》2009年第2期80-83,共4页Journal of Lanzhou University of Technology

基　　金：甘肃省科技支撑计划项目(0804GKCA046)

摘　　要：分析快速细化算法和OPTA细化算法不足产生的内在原因，提出一种新的基于重心的快速细化算法．该算法根据被细化图像的特点，用密度重心快速将纹线细化到3个像素宽度内，计算4邻域拓扑实现彻底细化．仿真结果表明，在细化效率方面，该算法一次遍历删除超过一半的大量冗余像素，是快速细化算法的3～7倍；在细化要求方面，该算法可达到绝对单像素、光滑无毛刺，并能保持端点不被吞噬，能够很好地满足图像细化的要求．The intrinsic origin of deficiencies in both the fast thinning algorithm （FTA） and one-pass thinning algorithm （OPTA） was analyzed, and a barycenter based fast thinning algorithm was presented. According to the image＇s feature to be fined, the veins of the image were fined to be of three pixel in width by using density barycenter and then the 4-neighborhood topology was computed to make complete thinning. Simulation result showed that, in connection with the thinning efficiency, the proposed algorithm exhibited so perfectly that a great amount （more than a half） of redundant pixels was eliminated after single overall checking, being 3 to 7 times faster than that with fast thinning algorithm. As to the requirement of thinning, this algorithm could implement absolute single-pixel thinning without blurred edges and keep the ends un-swallowed, so that the requirement of image thinning was well met.

关 键 词：密度重心 图像细化 单像素 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]