摘要
提出一种简单易行的二值图像并行细化算法 PABIT,该算法通过模板匹配的方式 ,层层剥离原始图像的边缘像素 ,使具有像素点宽度为偶数的笔划保留双像素的中心骨架 ,使像素点宽度为奇数的笔划只保留单像素的中心骨架线 ,最终在此准骨架的基础上通过进一步处理得到最后的笔划宽度为 1的细化图案 .由于该算法采用并行处理的模式 ,使得该算法较通常的串行算法效果更好 .经该算法细化后的结果具有保形 ,连通 。
This article presents a simple and facile parallel algorithm of binary image thinning PABIT which gets rid of the edge points step by step and turns the lines with even pixels to double point lines, turns the lines with odd pixels to single point lines. Based on the figure, the algorithm wipes off some redundant black points by modes and turns the figure into one thinned figure whose lines have only one pixel. For the algorithm being parallel, it makes thinning more effective and accurate. The thinning procedure keeps the basic shape of the original figure. The generated figure is adjoined and non distorted.
出处
《武汉理工大学学报(交通科学与工程版)》
北大核心
2004年第2期270-273,共4页
Journal of Wuhan University of Technology(Transportation Science & Engineering)
基金
公安部重点科研项目资助 (批准号 :2 0 0 14 2 2 710 1
9982 2 2 5 60 )