测绘学报

• 学术论文 • 上一篇    下一篇

扫描细化算法的地图自动矢量化研究

刘人午1,李燕2   

  1. 1. 资中县国土资源局
    2. 昆明理工大学
  • 收稿日期:2011-03-29 修回日期:2011-08-09 出版日期:2012-04-25 发布日期:2012-04-25
  • 通讯作者: 刘人午

Study of Auto-vectorization based on Scan-thinning-algorithm

  • Received:2011-03-29 Revised:2011-08-09 Online:2012-04-25 Published:2012-04-25

摘要: 栅格地图自动矢量化是地理信息采集的重要方式,但目前的主流自动矢量化软件的处理结果却不尽如人意,需要进行大量的后期编辑处理。本文根据栅格线体的形态特征设计了通过对栅格线体的序贯推扫获取待定像元,再从待定像元处进行截断扫描获取截断处的中心像元,最后再反连该中心像元及上一中心像元的扫描细化算法。该算法的细化结果受线体边缘条件的影响小、中心特征像元较少、细化精度较高。设计了针对于扫描细化算法计算结果中细化线体像元被分为了特征点与普通点的自动追踪矢量化算法,减少了坐标转换量的同时还使矢量线体结点较少、拟合光滑后更加趋于栅格线体中心、在交点处线体不相互打断,大大提高了自动矢量化及其后处理工作的效率。文章还分析了栅格区划图的自动矢量化及后处理方法。

Abstract: The auto-vectorization of raster map is the important way of the collection of geographic information. But at present, the result of the main software for auto-vectorization is bad with some aspects which make the great computation of disposing the vector lines. It’s designed the Scan-thinning-algorithm based on the morphological characteristics of the raster lines. This algorithm though 3 steps to thinning the raster lines, first, tracking the undetermined pixel by pushing with the sequential of the line; the second, obtaining the Center-pixel by scanning truncation of the line; finally, connection the present Center-pixel and the last Center-pixel with the ordinary Thinning-pixel. The result of the Scan-thinning-algorithm has some characteristics, less influence of the edge conditions, less Center-pixels, and grate precision. It’s designed the auto-tracking-vectorization algorithm dispose the result of the Scan-thinning-algorithm. By the way, reduced the working of coordinate transfer, at the same time the result of this algorithm has some characteristics, less nodes, the smooth result approach with the center of line, do not mutual interrupted at the intersection, so that , which greatly improved the efficiency of auto-vectorization and post-processing. Finally, this paper introduces the auto-vectorization and the post-processing method of the raster area map.