面向国外地理信息提取的地图图像智能化处理方法研究

Quickly acquiring foreign geographic information is essential to construct global geospatial information infrustructure. This project aims at extracting geographic information using foreign topographic maps as data sources. Research works will be done by applying theories and methods of intelligent map image processing to solve the bottleneck problem of manual map data collection with slow speed and low efficiency. Through the analysis of characterics of color, symbol, text and graphics in foreign topographic maps, the following methods will be focused on: (1) Color map image pre-processing and color segmentation methos.This study is expected to improve image quality, and reduce interference between different map features; (2) Map symbol recognition and text extraction methods.This study is expected to quickly obtain some attribute information and toponymic information of geographic features; (3) Linear feature vectorization method.This study is expected to quickly acquire geometrical information of geographic features. We will try to make breakthrough on the three aspects: (1) Color image separation with low image quality; (2) Separation of text and symbols from intersecting linear features; (3) Directive vectorization of linear features in complex background. The probject is of great importance both theoretically and practically in improving the automation and intelligence level of acquiring foreign geographic information data, and accelerate the construction of global geospatial database.

国外地理信息的快速获取是构建全球地理空间信息基础框架迫切需要解决的重大问题。本项目针对以国外地形图为资料源的地理信息提取所涉及的关键科学问题，应用地图图像智能化处理的理论与方法进行研究，旨在解决靠人工方法采集地图数据速度慢、效率低的瓶颈。针对国外地形图图像质量差、颜色种类多、符号多变形、文字与图形大量粘连所带来的困难，研究彩色地图图像预处理与颜色分割方法，以改善图像质量，减少地图要素间的相互干扰；研究地图符号识别和文字注记自动提取的方法，以快速获取地理要素的属性和地名信息；研究彩色地图线划要素自动矢量化方法，实现对地理要素几何信息的快速提取。突破低质量彩色地图图像预处理与颜色分割、图文粘连的文字注记与线划图形的自动分离、复杂背景下彩色地图线划要素直接矢量化等关键技术。该项目的研究，对于提高国外地理信息数据获取的自动化和智能化水平，加快全球地理空间数据库建设具有重要的理论意义和应用价值。

国外地理信息的快速获取是构建全球地理空间信息基础框架迫切需要解决的重大问题。本项目针对以国外地形图为资料源的地理信息提取所涉及的关键科学问题，应用地图图像智能化处理的理论与方法进行研究，旨在解决靠人工方法采集地图数据速度慢、效率低的瓶颈。针对国外地形图图像质量差、颜色种类多、符号多变形、文字与图形大量粘连所带来的困难，研究了彩色地图图像预处理与颜色分割方法，以改善图像质量，减少地图要素间的相互干扰；研究了地图符号识别和文字注记自动提取的方法，以快速获取地理要素的属性和地名信息；研究了彩色地图线划要素自动矢量化方法，实现对地理要素几何信息的快速提取。突破了低质量彩色地图图像预处理与颜色分割、图文粘连的文字注记与线划图形的自动分离、复杂背景下彩色地图线划要素直接矢量化等关键技术。该项目的研究，对于提高国外地理信息数据获取的自动化和智能化水平具有重要的理论意义和应用价值。