基于众源地理数据的人车混合路网一体化构建方法

行人和行车道路网络的一体化表达是精准位置服务、交通规划管理与智慧城市建设的重要前提。近年来，众源地理数据为行人和行车道路网的提取与构建提供了一种极具潜力和价值的数据源，但众源地理数据同时存在时空尺度多变、语义标签模糊、拓扑关系错误等问题。因此，本项目以行人和行车道路网络的时间-空间-语义一致性保持为研究目标，旨在发展一套多模式人车混合路网一体化构建的方法模型，通过融合信息互补、质量各异的众源地理数据，重点研究顾及时空演变模式的行车道路网增量生成方法，基于多源数据融合的行人道路全要素分治提取算法，以及空间-语义模式保持的人车混合路网一体化构建模型，从而达到城市导航地图信息完整性和更新时效性的有效提升，为快速、高效的出行服务提供关键数据支撑。

Congruent representation of pedestrian and vehicle road networks is a crucial precondition of accurate Location-based Services (LBS), transportation planning management and smart city construction. Presently, crowdsourcing geospatial data provides a potential and valuable data source for extracting pedestrian and vehicle road networks. However, there may be some special problems existing in crowdsourcing geospatial data, such as various spatio-temporal scales, ambiguous semantic descriptions and incorrect topological relations. Hence, the project tries to construct a congruent space-time-semantic representation for pedestrian-vehicle hybrid road networks by utilizing crowdsourcing geospatial data. The project considers the spatio-temporal changing patterns of vehicle road networks and proposes a multi-temporal incremental method for generating vehicle road networks. Meanwhile, we adopt a divide-and-conquer strategy to extract all kinds of pedestrian paths and walking micro-structures by fusing various geospatial data sources. After that, the project proposes a data conflation framework for achieving congruent space-time-semantic representation of multi-model pedestrian-vehicle hybrid road networks, so as to improve the information richness and data freshness of navigation map and provide a significant Spatial Data Infrastructure (SDI) for real-time and efficient travelling services.