Efficient massive spatial data search with version control

In large scale earth observation systems, data is gathered in short duration surveys returning massive numbers of point data observations (often millions) covering significantly sized regions. To maintain up-to-date products (e.g. navigation charts in harbours with continuous silt build-up), successive surveys are performed that often cover overlapping areas. Searches to find all the data in a specific region are expected to return only the most recent data in any given area, and must ignore any older, overlapping versions. In some cases, the data has errors that are difficult to correct; we would like to ignore or exclude this data during retrieval so that products derived from the data (e.g. contours on navigational charts) are accurate. This MITACS project investigates how to build a spatial data structure that efficiently supports this use case, termed “exclusion persistence range search”, over multiple versions of data covering the same area. CARIS makes software that is used to create navigational charts used by ships, aircraft and other moving vehicles. CARIS will benefit from this research by having a prototype multi-version index system that increases the efficiency and flexibility of its software, making it easier to track and change the data sources used to create navigation charts.

Intern: 
Stuart MacGillivray
Faculty Supervisor: 
Dr. Bradford Nickerson
Partner: 
Discipline: 
Program: