基于图段宏观特征的直线与圆弧矢量化方法

An Algorithm for Vectorization of lines and arcs Based on the Global Character of the Image Segment

本文所描述的工程图扫描图像中直线与圆弧的矢量化算法考虑了图段的宏观特征，引入了主方向码和平均方向码的概念，提出了图段跟踪分析的原理，它通过主方向码和平均方向码反映的图段的整体斜率特征，对不同斜率的图段，分别沿Ｘ轴方向或Ｙ轴方向进行跟踪分析，能够同时得到图段的拓扑信息与几何信息，跟踪的单元是垂直于跟踪方向的黑游程。该算法的基础是直线段的跟踪分析，图象中的圆弧先用折线近似，再按照一定准则由近似折线重建。该矢量化算法处理速度快。

Abstract The algorithm for vectorization described in the paper focuses on the global character of the image segment. It introduces the concept of major code and mean code, and gives a principle for image segments analysis. With different slanting, the image segments are followed in different orientations. The element is the black run length perpendicular to the segment following orientation. The foundation of the algorithm is the analysis of straight line segment. The arcs are approximated by polygon, and then reconstructed according to certain rules. The algorithm has such advantages as high speed, and especially performing well at processing scanned image of large engineering drawing.