基于模糊分类的地图整体识别方法

A Global Recognition Method of Maps Based on Fuzzy Classification

1 引言在地理信息系统(GIS)飞速发展的今天,对地图扫描图像的识别已成为图像识别技术的重要分支.如何运用已有的图像识别技术,尽快发展地图全自动识别的理论和方法,对于我国GIS的应用和发展将发挥重要作用.在城市地图中,图形一般是具有实际地理意义的对象,因此在地图识别中,不仅要将线段矢量化以便保存,而且要对各种具有实际地理意义的对象进行分类,以便修改、查找、赋予属性和进行分析,并且应把不同类的对象归入各相应图层,以尽量减少后期的工作.本文对目前城市应用极广的1:500地图进行了研究,提出了相应地图中各类地物的分类识别方法.

This paper proposes a automatic classification and recognition method of maps. On the basis of patch structure and connected graph,the graph theory is used for the rough classification of graphics,after vectorizing the graphics,a fuzzy classification method is described for the careful classification of objects on maps. Experiment shows that the performance of the method is very good.