基于层级聚合的多源三维瓦片构建方法

Hierarchical Aggregation-Based Method for Multi-Source 3D Tile Construction

3Dtiles规范因其大规模流式传输、面向GPU等特性是Web端大规模三维模型构建的首选标准.现有3Dtiles构建方式依赖于单一数据源且难以满足多源数据的LOD(Level of Detail)层级关系,使得GIS(Geographic Information System)三维数据在传输、加载过程中受到限制.针对该问题,本文提出了一种基于层级聚合的多源三维瓦片构建方法,该方法首先按照3Dtiles规范实现多源数据到三维瓦片的映射转换,然后按照LOD层级关系对多源瓦片进行编排并使用HR树索引重组瓦片集文件.最后,使用Bing Maps建筑轮数据集进行3Dtiles瓦片构建实验,并分析了瓦片插入时延、加载性能.结果表明该策略可以在保证瓦片基本插入性能情况下有效提升瓦片加载性能.

The 3D Tiles specification is the preferred standard for large-scale 3Dmodel construction on the web due to its characteristics such as massive streaming transmission and GPU optimization.Existing 3D Tiles construction methods rely on a single data source and face challenges in satisfying the level of detail(LOD)hierarchy for multisource data,restricting the transmission and loading processes of GIS 3D data.To address this issue,this paper proposes a hierarchical aggregation-based method for constructing 3D tiles from multiple sources.This method initially implements the mapping transformation of multi-source data to 3D tiles according to the 3D Tiles specification.Subsequently,it organizes multi-source tiles based on LOD hierarchy and utilizes an HR-tree index to reorganize the tileset files.Finally,experiments are conducted using the Bing Maps building footprint dataset for 3D Tiles tileset construction.The paper analyzes tile insertion latency and loading performance.The results demonstrate that this strategy can effectively improve tile loading performance while ensuring basic tile insertion performance.