Návrh akcelerace časově náročných operací při tvorbě 3D modelu povrchu z lidarovych dat za pomocí distribuovaných výpočtů

Hovad, Jan

Digitální knihovna UPCE
→
Univerzita Pardubice
→
Publikace vydané Univerzitou Pardubice / Published by the University of Pardubice
→
Odborné časopisy / Journals
→
Scientific papers of the University of Pardubice
→
Series D, Faculty of Economics and Administration
→
2014/32 Scientific papers, Series D
→
Zobrazit záznam

dc.contributor.author	Hovad, Jan
dc.date.accessioned	2015-02-12T09:16:09Z
dc.date.available	2015-02-12T09:16:09Z
dc.date.issued	2014
dc.identifier.issn	1211-555X (Print)
dc.identifier.issn	1804-8048 (Online)
dc.identifier.uri	http://hdl.handle.net/10195/58658
dc.format	p. 18-27	eng
dc.language.iso	cze
dc.publisher	Univerzita Pardubice
dc.relation.ispartof	Scientific papers of the University of Pardubice. Series D, Faculty of Economics and Administration. 32 (3/2014)	eng
dc.rights	open access	eng
dc.subject	GFS	eng
dc.subject	distributed computing	eng
dc.subject	functional approach	eng
dc.subject	LIDAR	eng
dc.title	Návrh akcelerace časově náročných operací při tvorbě 3D modelu povrchu z lidarovych dat za pomocí distribuovaných výpočtů	cze
dc.title.alternative	Acceleration of time-consuming operations in case of 3D model creation by utilization of lidar technology and distributed computations	eng
dc.type	Article
dc.description.abstract-translated	This article is focused on optimization of the processing of large volumes of data sets obtained by LIDAR technology. Data are previously cleansed, transformed into a square grid, which resolution is adaptable to the terrain slope factor. Intermediate interpolated data sets are qualitatively compared by statistical methods and transformed into the form of realistic model of the terrain. In case that the entire process is performed by a single PC, its implementation is limited to the relatively small spatial areas. The size of growing input data gradually reduces the possibility to create 3D terrain model and to be successful, must be handled by means of distributed computing. Implementation of this process is difficult and involves some critical tasks. The most common issues are directed into the area of jobs planning, thread management, optimization of the available hardware resources, solving of critical situations and controlling the data flow throughout the network. Simplification lies in the abstraction of the above mentioned problems. The focus is directed into the programming phase and into the explanation of distributed principles. This article is aimed to prepare the implementation of computationally intensive tasks that are computed by the distributed computations. Proposed distributed solution is based on the principles of Google File System.	eng
dc.peerreviewed	yes	eng
dc.publicationstatus	published	eng
dc.identifier.scopus	2-s2.0-84929456911