一种位图矢量化新方法  

A New Approach for the Vectorization of Raster Imagei

作　　者：严素蓉[1] 朱桂林[1] 徐从富[1] 

机构地区：[1]浙江大学计算机科学与技术学院,杭州310027

出　　处：《计算机工程与应用》2005年第14期85-87,共3页Computer Engineering and Applications

基　　金：浙江省自然科学基金项目(编号:602045)资助

摘　　要：在图像的矢量化过程中,是否精确有效获取图像的几何特征点直接影响到矢量化效果,而已有的研究中并没有十分有效的方法。文章通过对位图边界轮廓进行合理分解,找出主要轮廓特征,得到轮廓关键特征点;并且用原始轮廓点对特征点进行插值,不但能够很好地保留图像的原始信息特征,失真小,还能适应多种形式的图像矢量化算法。在CA D/CA M中的图像运用中取得较好效果。A novel algorithm to vectorize raster image is approached.Based on some conventional contour detecting methods,an effective contour locating and tracing approach is employed to get the outline points on the image contour.In the succeeding steps,key points are chosen from the outline points by optimized selection scheme considering the special information of the contour.The final vector graph is constructed of lines and curves using key points.Tests have shown that the vector graph constructed by this algorithm has little distortion,and the vectorized process can obtain accurate geometric feature of images.When applied to CAD/CAM,this algorithm is also proved to be efficient and robust.

关 键 词：图像轮廓 径向增量同向点 插值 矢量化 

分 类 号：TP391.7[自动化与计算机技术—计算机应用技术]