基于OTSU和轮廓追踪的地图线条识别方法

纸质地图或扫描地图是传统地图保存方式。这种方式使得空间查询与分析困难，需要将这些地图矢量化才有助于对历史地理空间数据的查询和分析。然而，传统矢量化方式耗时耗力，不能满足实际需求。为此，提出了一种基于边界追踪的地图自动矢量化方法。首先，将地图灰度化和二值化；然后，对二值化的地图应用栅格转矢量方法矢量化。最后，基于4幅地图，测试了本文方法的提取效果。实验结果表明，本文方法能够较为完整地提取出全部地图的线条，且可以有效地将地图中的文字分离出来。

Map Line Recognition Method Based on OTSU and Contour Tracking

Paper map and scanned map are the traditional way of map preservation. These methods make spatial query and analysis difficult. It is necessary to vectorize these maps to facilitate the query and analysis of historical geospatial data. However, the traditional vectorization method is time-consuming and labor-consuming, and cannot meet the actual needs. Therefore, this paper proposes an automatic map vectorization method based on boundary tracking. Firstly, the map is grayed and binarized; then the raster to vector method is applied to the binary map; finally, based on four maps, the extraction effect of this method is compared with common methods. The experimental results show that this method can extract all map lines completely, and can effectively separate the text in the map.