Storage of big, dynamic, regular 3D grid models with RasDaMan.
Georg Semmler and Jan Gietzel and Helmut Schaeben. ( 2015 )
in: 35th Gocad Meeting - 2015 RING Meeting, ASGA
Abstract
Historically there are two common representations of two-dimensional geometry data: the vector and the pixel mode. Both representations can be generalized in three dimensions. The decision for one of the two modes depends heavily on the problem to solve. More and detailed methods in geophysical and laboratory measurements of rock properties and moreover geological classification leads to an increasing amount of data for a certain underground area and different scales and resolutions. Typically big data analysis like statistical based predictions of mineral deposits in three dimensions are preformed on voxel mode. Commonly the structural geological modeling is done in vector mode, which demands a geometrical algorithm for the conversion into voxel mode (Wieczoreck et al. [2015]) to include this data for analysis. To keep solutions for analysis feasible for large amounts of data it is necessary to develop systems to store and work with large multi-dimensional grids. In this presentation a Database Management System is introduced based on RasDaMan (Raster Data Manager, http://rasdaman.org/). The systems handles Gocad Voxet objects as generic big data grids and delivers functions for model extent changes, manipulation queries and all functionality of RasDaMan while being flexible to develop extensions. Rasdaman features an efficient storage layout and achieves a good performance when accessing and manipulating the grid data using a well defined query language. Furthermore RasDaMan implements several OGC interfaces like WCS, WMS and WCPS to represent and distribute the data on the internet for presentation or for being harvested by data infrastructures for further analysis.
Download / Links
BibTeX Reference
@inproceedings{SemmlerGM2015,
abstract = { Historically there are two common representations of two-dimensional geometry data: the vector and the pixel mode. Both representations can be generalized in three dimensions. The decision for one of the two modes depends heavily on the problem to solve. More and detailed methods in geophysical and laboratory measurements of rock properties and moreover geological classification leads to an increasing amount of data for a certain underground area and different scales and resolutions. Typically big data analysis like statistical based predictions of mineral deposits in three dimensions are preformed on voxel mode. Commonly the structural geological modeling is done in vector mode, which demands a geometrical algorithm for the conversion into voxel mode (Wieczoreck et al. [2015]) to include this data for analysis. To keep solutions for analysis feasible for large amounts of data it is necessary to develop systems to store and work with large multi-dimensional grids. In this presentation a Database Management System is introduced based on RasDaMan (Raster Data Manager, http://rasdaman.org/). The systems handles Gocad Voxet objects as generic big data grids and delivers functions for model extent changes, manipulation queries and all functionality of RasDaMan while being flexible to develop extensions. Rasdaman features an efficient storage layout and achieves a good performance when accessing and manipulating the grid data using a well defined query language. Furthermore RasDaMan implements several OGC interfaces like WCS, WMS and WCPS to represent and distribute the data on the internet for presentation or for being harvested by data infrastructures for further analysis. },
author = { Semmler, Georg AND Gietzel, Jan AND Schaeben, Helmut },
booktitle = { 35th Gocad Meeting - 2015 RING Meeting },
publisher = { ASGA },
title = { Storage of big, dynamic, regular 3D grid models with RasDaMan. },
year = { 2015 }
}