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
Real-Time 3D Reconstruction in Dynamic Outdoor Environments: A Mobile Perspective with ARCore
Jönköping University, School of Engineering.
Jönköping University, School of Engineering.
2024 (English)Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
Abstract [en]

This research established a real-time 3D reconstruction pipeline for a dynamic outdoor environment, focusing on the criteria efficiency, accuracy, and applicability on later generation of Android smartphones. This research aimed to determine an effective data capture density through key metrics such as the number of stored vertices in a point cloud, confidence threshold for estimated depth data, image resolution, memory consumption, and computational time. With the concluded data capture density, the accuracy and consistency of the real-time 3D reconstruction pipeline was evaluated by point cloud registration. The implementation of this pipeline was assessed through experimental research. The methodology leveraged the Record API, Playback API, and Raw Depth API provided the ARCore library. Image and sensor data was recorded and played backed systematically with varying key metrics values. The reconstruction process involved requiring depth data from depth images. The concluded data capture density was a point cloud with 500 000 stored vertices, filtered with a high confidence threshold and created from 160x90 pixel depth images. The pipeline demonstrated suitable memory consumption with an average computational time under 16 milliseconds, confirming real-time operation. Point cloud registration demonstrated an average RMSE of roughly ± 2 cm and a fitness score of ± 0.6 between point clouds generated from the different datasets, and an RMSE ± 0,9 cm and a fitness score over 0.84 from same dataset. The findings demonstrate the possibility to achieve real-time 3D reconstruction with high spatial accuracy and reasonable consistency by utilising the ARCore library and an Android smartphone without additional equipment or active sensors.

Place, publisher, year, edition, pages
2024.
National Category
Computer Engineering
Identifiers
URN: urn:nbn:se:hj:diva-65486OAI: oai:DiVA.org:hj-65486DiVA, id: diva2:1879976
External cooperation
Husqvarna Group AB
Subject / course
JTH, Computer Engineering
Supervisors
Examiners
Available from: 2024-08-27 Created: 2024-06-29 Last updated: 2024-08-27Bibliographically approved

Open Access in DiVA

No full text in DiVA

By organisation
School of Engineering
Computer Engineering

Search outside of DiVA

GoogleGoogle Scholar

urn-nbn

Altmetric score

urn-nbn
Total: 106 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