DONATE

IAHR Document Library


« Back to Library Homepage « Proceedings of the 37th IAHR World Congress (Kuala Lumpur, 2...

Efficient Handling of Big Data for Topobathymetric Applications: Examples From Lake Constance and Bavaria in Europe

Author(s): Frank Steinbacher, Werner Benger, Ramona Baran, Wolfgang Dobler, Markus Aufleger

Linked Author(s): Markus Aufleger

Keywords: Topobathymetry, software, big data, HDF5, visualization

Abstract: The demand on topobathymetric data is growing quickly due to availability of newly developed airborne LiDAR sensors capturing high quality and resolution data. The data amount acquired is thereby increasing drastically. If the area of interest covers several hundred km�, the data amount can quickly reach up to several terabytes, which is on the edge of storage device capacities and efficient data use with available software packages. For example, the topobathymetric point cloud of Lake Constance consists of approximately 10 billion points, which is equivalent to about 700 gigabytes in classical las format. Moreover, the digital surface model for Bavaria with a grid size of 40 cm (data from Bavarian mapping agency) comprises about 460 billion grid points equivalent to approximately 3 terabytes in las format. These examples illustrate the requirement of an appropriate file format and software framework to store, visualize, process and analyze topobathymetric data efficiently. We employ a block-structured hierarchy to organize arbitrarily large unsorted point clouds and place them in a spatially ordered level-of-detail scheme allowing for recursive on-demand queries on data sections of interest. As data are much larger than available RAM, our out-of-core technique only loads the minimum amount needed for visualization to achieve interactive rendering rates of 30 frames/sec regardless of zoom level or placement within dataset during 3D camera navigation. The Hierarchical Data Format V5, designed for processing and archival of massive data generated high performance computing, is well suited to describe the complex relationships between data blocks and their meta-data, and to handle arbitrarily large files or file sets transparently and efficiently, optionally providing a multitude of compression schemes crucial for such large data. Interactive rendering is performed by the HydroVISHTM Visualization Shell based on OpenGL Shaders allowing displaying various point-based attributes combined with cartographic information like contour lines

DOI:

Year: 2017

Copyright © 2024 International Association for Hydro-Environment Engineering and Research. All rights reserved. | Terms and Conditions