基于天桥的几何与属性特征,提出了一种志愿者地理信息中自动识别天桥的方法。天桥从几何结构上可分为主桥和附属设施两个部分,主桥部分特征鲜明,可以视为两类分类问题,依据其几何特征和属性特征构建特征空间,利用支持向量机的方法进行识别;附属设施部分可依据已识别的天桥主桥,按照路段的长度、属性等判定规则进行识别,从而完成整个天桥的自动识别。以北京市OpenStreetMap(osm)数据进行试验验证的结果表明,本文提出的方法能有效地识别出志愿者地理信息中的典型天桥结构,可以为志愿者地理信息道路网的多尺度建模与化简、步行导航等提供帮助。
The paper presents an overpass recognition method in volunteered geographic information based on the geometry and attribute characteristics. The structure of the overpass is divided into the main bridge parts and the affiliated facilities. The main bridge parts with distinctive characters could be treated as a two-class classification problem. The characteristic vectors could build on the foundation of analysis and quantization the geometry and attribute characteristics. Then, the main bridge is recognized automatically through the support vector machine. The affiliated facilities of the overpass are recognized based on the main bridge with some relevant judgment rules. The OpenStreetMap(osm) is selected for the experiment. The results show that the method could effectively recognize the overpass and could provide help for the road simplification and walking guidance.