基于改进Loop细分的二维图像矢量化方法  

A Vectorization Method of Two-dimensional Image Based on Improved Loop Subdivision

作　　者：邓重阳[1] 熊玉丹 郑乐豪 曹晨妤 朱明超 DENG Chongyang;XIONG Yudan;ZHENG Lehao;CAO Chenyu;ZHU Mingchao(School of Sciences,Hangzhou Dianzi University,Hangzhou Zhejiang 310018,China)

机构地区：[1]杭州电子科技大学理学院,浙江杭州310018

出　　处：《杭州电子科技大学学报（自然科学版）》2020年第2期52-55,共4页Journal of Hangzhou Dianzi University：Natural Sciences

基　　金：国家自然科学基金资助项目(61872121);浙江省大学生科技创新活动计划(新苗计划)资助项目(2018R407035)。

摘　　要：现有的矢量化方法得到的矢量图缩放不方便,为此,用改进的Loop细分方法实现二维图像矢量化。首先用Douglas-Peucker算法提取图像边界,然后用Delaunay三角剖分算法得到三角网格,最后用改进的Loop细分算法细分三角网格得到矢量图。细分时,对三角网格顶点进行线性细分,对顶点的像素值进行Loop细分,同时,在原有Loop细分规则的基础上,修改网格边界上的三角形内部新点生成规则。实验结果表明:提出的矢量化方法得到的矢量图缩放较为方便。The existing vectorization methods are inconvenient to scale the vector image. An improved Loop subdivision method is proposed to realize the vectorization of two-dimensional image. Firstly, Douglas-Peucher algorithm is used to extract the boundary of the image, then Delaunay triangulation algorithm is used to obtain the triangular mesh, and finally, the improved Loop subdivision algorithm is used to subdivide the triangular mesh to get the vector image. In order to facilitate the scaling of vector graph, linear subdivision is carried out for the vertices of triangular mesh, and Loop subdivision is carried out for pixel values attached to the vertices. At the same time, based on the original Loop subdivision rule, the new point generation rule of triangle on the mesh boundary is modified. Experimental results show that the vectorization method is convenient to scale the vector image.

关 键 词：图像矢量化 DOUGLAS-PEUCKER算法 DELAUNAY三角剖分 线性细分 LOOP细分 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]