大比例栅格地形图中城区道路目标的自动识别

Algorithems for Recognition of Urban Roads in Large scaled Topographical Maps Based on Integration of Segmentation and Recognition

城市GIS发展迅速，各种大比例地形图是其基础数据，流行的地图数据数字化方法效率很低，一定程度上已经成为城市GIS发展的障碍，扫描输入技术有望解决这个问题。扫描输入技术的核心是地图符号的自动识别，城区道路目标是城镇地区大比例尺地形图中的一类主要目标，但现有大多数识别方法采用的都是先分割后识别的策略，使得分割难以充分利用目标符号的有关知识约束，本文提出并实现了大比例尺地形图中城区道路目标自动识别的分割识别一体化方法。

Urban GIS is developing rapidly, and various current large-scaled topographical maps are its essential data. Urban roads are the main geographic objects in large-scaled urban topographical maps. In this paper, based on the integration of segmentation and recognition, an algorithem for automatic recognition of urban roads in 1:10000 topographical maps which are most commonly used in Urban GIS, is proposed.