Urban Modeling Group Assistant Research Scientist, Dr. Anh Vu Vo presented a talk entitled “Combining 2D And 3D Indexing for Efficient Airborne LiDAR Data Management” at the Open Geospatial Consortium (OGC) Technical Committee Meeting in St. Johns, Newfoundland. The OGC is a key player in the creation of standards for geospatial data across the globe and the Urban Modeling Group is delighted to have the opportunity to help shape those standards.
Abstract
This presentation introduced a hybrid spatial index solution for efficient management of large aerial, point cloud data. The proposed approach uses multiple data indexing layers: a top layer with a two-dimensional, Hilbert-coded, rectangular grid and a bottom layer with multiple, separate, in-memory, three-dimensional octrees. This approach is designed to push more of the workload into the main memory, thereby reducing disk scanning and improving the query resolving speed. Scalability is addressed by using the lightweight Hilbert-coded grid to rapidly localize the searching domain to only a few blocks, which can then be processed independently, irrespective of the dataset’s overall size. When tested on datasets from 90 million to 1.15 billion points, the new hybrid index was 1.7 to 9.1 times faster compared to traditional indexing. Additionally, the proposed approach offered in-base functionality not previously available in the form of incremental nearest neighbor searching and planar surface selection.