Modeling and Analyzing Terrain Data Acquired by Modern Mapping Techniques
Abstract
Modern remote sensing methods such as LIDAR readily generate high-resolution elevation data, which can be tens or hundreds of gigabytes in size. Several applications including erosion modeling, landslide risk assessment, stream mapping, and hydrologic modeling can benefit from this high-resolution data but elevation data point sets must first be transformed into a digital elevation models (DEMs) and derived products such a river networks or watersheds before users can conduct relevant studies. Processing these massive data sets poses a number of algorithmic challenges. The goal of this project is to provide enhanced terrain modeling and analysis capabilities by developing sophisticated algorithms that function with massive non-standard datasets, such as point clouds, and that produce a confidence level for the results. We are developing algorithmic techniques to overcome the computational challenges encountered when processing the massive, dynamic, and heterogeneous geospatial data acquired today: We utilize approximation techniques to trade efficiency with accuracy-use hierarchies to represent the data at varying levels of detail, and rely on approximation algorithms to solve various terrain-analysis problems efficiently. To handle the massive amounts of data efficiently, we utilize recent advances in memory-aware algorithms, that is, algorithms that are specifically designed to handle data sets that do not fit in main memory of underlying devices.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 22, 2009
- Accession Number
- ADA509693
Entities
People
- Helena Mitasova
- Lars Arge
- Pankaj Agarwal
Organizations
- Duke University