Digitala Vetenskapliga Arkivet

Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
A Domain Adjusting Region Octree: Indexing a stream of unpredictable point cloud data for line-of-sight analysis
KTH, School of Electrical Engineering and Computer Science (EECS).
2024 (English)Independent thesis Advanced level (professional degree), 20 credits / 30 HE creditsStudent thesisAlternative title
Ett domänjusterande regionsokträd : Indexering av en ström av oförutsägbar punktmolnsdata för siktlinjeanalys (Swedish)
Abstract [en]

High resolution point clouds generated by Light Detection and Ranging (LiDAR) devices have great potential in the field of topography, representing terrain surfaces as numerous spatial points. It is desirable to analyse such point clouds, for instance in order to determine Line-of-Sight (LoS) between two locations in recorded areas, but point clouds can be so large that they are difficult to make effective use of without first turning the data into a more organised format. If LiDAR readings could be organised as quickly as they are generated, anyone or anything with access to a LiDAR device could collect and analyse point clouds in real-time. This thesis uses octrees to quickly index data points one by one, for the purpose of allowing a hypothetical real-time LoS application to exploit large point clouds. By letting the domain of the octree expand and shrink as needed to follow the most recent set of input, memory usage remains manageable, access speed stays fast, while the most relevant information is kept. This solution does not quite meet the real-time targets but works for the most part as performance does not degrade over time, thanks to the shrinking operations that reduce the depth of the octree. However, special attention must be given to memory management to avoid severe drops in performance during these same shrinking operations. A few concerns are also presented about how well the representation corresponds to reality. With some additional adjustments, features and optimisation, this approach could probably accomplish the set real-time targets.

Abstract [sv]

Punktmoln med hög upplösning genererade av Light Detection and Ranging (LiDAR)-enheter har stor potential inom topografi vilka representerar ytor av terräng som många punkter en rymd. Det är önskvärt att analysera sådana punktmoln, bland annat för att beräkna siktlinjer mellan två platser i avlästa områden, men punktmoln kan vara såpass stora att de är svåra att utnyttja effektivt utan att först omvandla sådan data till ett mer organiserat format. Om LiDAR-data kunde organiseras lika fort som det genereras, skulle allt och alla med tillgång till en LiDAR-enhet kunna samla och analysera punktmoln i realtid. Denna tes använder okträd för att snabbt indexera datapunkter en efter en, med avsikt att tillåta en hypotetisk siktlinjeapplikation att i realtid utnyttja stora punktmoln. Genom att låta okträdets domän expandera och krympa efter behov, för att följa den senaste uppsättningen av indata, hålls minnesanvändning hanterbart och tillgångshastighet förblir snabb medan den mest relevanta informationen behålls. Denna lösning uppnår knappt realtidsmålen men fungerar annars för det mesta då prestanda inte försämras över tiden, tack vare förkrympningsoperationerna som reducerar okträdets djup. Däremot kräver minneshantering särskild uppmärksamhet för att undvika drastiska prestandaproblem under förkrympningsoperationerna. Ett antal bekymmer presenteras angående hur pass väl representationerna motsvarar verkligheten. Trots detta är det sannolikt att målen kan uppnås givet ett par extra funktioner, optimeringar och justeringar.

Place, publisher, year, edition, pages
Stockholm: KTH Royal Institute of Technology , 2024. , p. 71
Series
TRITA-EECS-EX ; 2024:549
Keywords [en]
DATA STRUCTURES, Linked representations, Octrees, Point Cloud, Digitalelevation model, Spatial indexing
Keywords [sv]
Datastrukturer, Länkade representationer, Okträd, Punktmoln, Digital höjdmodell, Spatial indexering
National Category
Computer Sciences Computer Engineering
Identifiers
URN: urn:nbn:se:kth:diva-352837OAI: oai:DiVA.org:kth-352837DiVA, id: diva2:1895902
External cooperation
Carmenta Geospatial Technologies AB
Subject / course
Information and Communication Technology
Supervisors
Examiners
Available from: 2024-10-01 Created: 2024-09-08 Last updated: 2024-10-01Bibliographically approved

Open Access in DiVA

fulltext(6663 kB)206 downloads
File information
File name FULLTEXT01.pdfFile size 6663 kBChecksum SHA-512
f8faa061bf4e1dfc82e82ae4242f015b90ab81d27e7791fe1118a2ff0fa48237f718c10cfbfa48b3860f24c4e1aa31faea83a8c11743e944b60eed68d2f0615d
Type fulltextMimetype application/pdf

By organisation
School of Electrical Engineering and Computer Science (EECS)
Computer SciencesComputer Engineering

Search outside of DiVA

GoogleGoogle Scholar
Total: 206 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

urn-nbn

Altmetric score

urn-nbn
Total: 85 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf