由社会公众通过网络资源上传的开放性“众源地理信息”日益成为GIS的重要数据来源，众源地理信息在揭示社会行为时空规律、发现空间模式特征、诠释地理过程机理、预测时空演变趋势上具有重要作用。然而，众源地理信息的开放性、公众化与非结构性特征使得该类信息难于直接开发应用，面临着融合匹配、抽象概括、重分类等尺度变换处理的需求。本课题针对该问题开展网络众源地理信息在线式尺度变换的研究，适于网络众源数据特点，建立比例尺等元数据缺失条件下的尺度特征分析方法；面向网络渐进式传输需求研究空间数据连续尺度变换的原理，提出空间数据表达的尺度生命期模型；在大容量众源空间数据上通过尺度抽象化变换过滤细节从而获得时空主体特征与空间知识；技术方法上提出在线式与离线式两种综合在地理信息服务中的协同控制机制。本研究是在新一代GIS技术驱动下对传统地图综合原理与方法的扩展，为大数据背景下的网络空间数据处理与知识发现提供支持。

The multi-scale representation of spatial data aims at the cognition and model analysis of geographic phenomena from the point of multi-level The crowdsourcing geographic information which is uploaded by public users in open web context increasingly becomes more and more important for GIS applications. It plays a significant role in discovering spatial-temporal principles, mining spatial pattern, exploring the geo-process and forecasting spatial development trends. However the characteristics of open, unstructured and invalid accuracy make the crowdsourcing data difficult to directly utilize, requiring the matching, abstracting, reclassifying and other scale-transforms to enhance the data. This proposal aims at the question above trying to conduct the study of on-line scale transformation on crowdsourcing geographic information, to extend the tradition map generalization and scale transformation. To meet the requirement from the web crowdsourcing data, this study will build the scale transformation method without scale metadata. The continuous map generalization will be studied for the progressively transfer under web application, and a representation model, namely the scale life-span model, will be examined. The scale transformation will be applied in big geo-data to reduce the data volume by scale abstracting and further detect the main characteristics and spatial knowledge. In technology, this study will distinguish the on-line generalization and off-line generalization and develop the consistent control scheme in real applications of geo-information service. Driven by the modern web technologies, the proposal study will extend the traditional map generalization in both theories and technologies, to support the spatial data handling and knowledge discovering in the era of big data.