基于多源数据的高精地图生产技术及三维可视化表达

随着自动驾驶汽车的量产及逐步普及,对高精地图的需求也越来越高,汽车通过高精地图实现驾驶环境的精确感知,从而提升驾驶的安全性和可靠性。本文提出了综合利用倾斜模型和车载移动扫描点云等多源异构数据提取高精地图道路要素矢量线划的方法,分析不同源数据对数据提取精度的影响因素,探讨高精地图数据模型的空间表达、属性分层及关联规则,研究构建高精地图三维可视化的实现方式,并提出了可行的技术方法,为高精地图的生产提供可借鉴的参考价值。

High-precision map production technology and 3D visual expression based on multi-source data

With the mass production and gradual popularity of self-driving cars, the demand for high-precision maps is also increasing. Cars can realize the accurate perception of the driving environment through the high-precision map, so as to realize the safety and reliability of driving. The method of extracting the vector line drawing of high-precision map from multi-source heterogeneous data such as tilt model and vehicle mobile scanning point cloud is proposed, and analysing the different source data on the accuracy of data extraction, and discussing the spatial expression, attribute stratification and association rules of the high-precision map data model. We study the implementation of 3D visualization of high precision map, and propose feasible technical methods to provide a reference value for the production of high precision map.