Geographic information systems were introduced commercially in the late 1970s as two-dimensional (2D), electronic mapping. A decade ago, the technology was pushed into 2.5D (only 1 point supportable in the elevation direction), and today there is limited 3D support, but in what is essentially a 2D system with all of its inherent 2D limitations. Meanwhile, millions of dollars are invested annually to collect 3D urban remote sensing data, without a spatio-temporal database system that fully supports these large 3D datasets. This project aims to develop just such a storage, indexing, and querying system for 3D data. The system is being built in a distributed environment using HBASE as a foundation, in order to best take advantage of the ever-increasing density of urban datasets while still preserving the inherent three dimensionality and spatio-temporality of the data and addressing the performance and storage size challenges inherent with such massive datasets.
This project has been awarded a high-performance computing allocation from XSEDE which we are currently using for test deployments of various data models and indexing paradigms.