低空机载激光雷达支持的道路渠化交叉口重建研究

Channelized road intersection carries the key node information of urban road network. Reconstructing its complex movement pattern can fill the gap of existing road network database model with hubs information of intersections, improve road model’s cognitive performance, which provide a greater level of GIS_T system in analysis and management of emergency response, traffic monitoring and road construction. Aiming at the problem of high quality automatic reconstruction of movement pattern of channelized road intersection, this proposal researches on building the theory and methods of obtaining high completeness and accuracy laser point cloud based on UAV-borne LiDAR system, working through the key challenge of automatic searching of movement path from point cloud of channelized road intersection, and finally validating the result, thus providing technical support for road map key node information updating and road network cognition. In our method, the optimized observation mode of UAV-borne LiDAR system and enhancement model for the point cloud are built to form as core algorithm for the data enhancement of channelized road intersection and, consequently, to obtain high quality data for movement pattern reconstruction of channelized road intersection. The method for movement path searching from point cloud of channelized road intersection is developed for the purpose of both pattern and topology reconstruction. Finally, based on the existing platform, the overall evaluation for both laser surveying quality and reconstruction result are performed through the empirical research on typical area of channelized road intersection.

道路渠化交叉口承载了城市道路网络的枢纽信息，掌握其复杂的通行结构能够有效填补现有路网模型枢纽交叉口信息空白，提升路网整体认知，从而提高交通GIS系统在应急响应、交通监控、道路建设等方面的分析与管理水平。本项目面向道路渠化交叉口重建需求，研究低空机载激光雷达高质量点云获取的理论与方法，突破从渠化交叉口增强点云中进行通行路径自动识别的难题，解决渠化交叉口复杂结构的高质量自动重建问题，并对重建结果进行评价，为道路网络枢纽节点信息的更新与整体认知提供技术支撑。为此，本项目研究低空机载激光强化观测模式以及点云增强理论模型，形成针对道路渠化交叉口目标的点云获取与增强的核心算法，实现面向道路渠化交叉口重建的高质量数据获取；研究基于渠化交叉口点云的通行路线识别方法，实现渠化交叉口通行结构与拓扑关系重建；最后，基于已有观测平台，通过典型道路渠化交叉口的实证研究，评价观测与重建的综合能力。

本项目基于低空平台传感器开展面向道路渠化交叉口重建的研究工作，研究在高质量点云获取基础上，进行交叉口增强点云中进行通行路径自动识别的关键理论与方法，从而解决渠化交叉口复杂结构的高质量自动重建问题，并对重建结果进行评价。为此，本项目研究了低空机载激光强化观测模式以及点云增强理论模型，形成针对道路渠化交叉口目标的点云获取与增强的核心算法，实现了面向道路渠化交叉口重建的高质量数据获取；研究基于高阶张量模型，实现了渠化交叉口通行结构与拓扑关系重建，并对实验结果进行评价与验证。