Octree-based Indexing for 3D Point Clouds within an Oracle Spatial DBMS

A large proportion of today’s digital data has a spatial component. The storage and effective 10 management of such data poses particular challenges. Nowhere is this truer than with Light De-11 tection and Ranging (LiDAR), where datasets of even small geographic areas may contain sever-12 al hundred millions points. Currently available spatial information systems do not provide suita-13 ble support for 3D data. As a consequence, while a specific system can be used to store the data, 14 another has to be used to process it. Typically several software applications are used to analyze 15 and process LiDAR data, requiring multiple format transformations. Our work aims at providing 16 a more cost-effective way for managing LiDAR data that allows for the storage and manipulation 17 of these data within a single system. We achieve this by exploiting current Spatial Database 18 Management technology. In order to provide an efficient solution, suitable 3D indexing mecha-19 nisms are essential. In this paper we describe the implementation of an octree index for 3D Li-20 DAR data atop Oracle Spatial 11g and present a comprehensive evalu