Show simple item record

dc.contributor.advisorHendseth, Sverre
dc.contributor.authorMysen, Mats Meier
dc.date.accessioned2019-10-31T15:00:12Z
dc.date.issued2019
dc.identifierno.ntnu:inspera:35771502:18350752
dc.identifier.urihttp://hdl.handle.net/11250/2625654
dc.descriptionFull text not available
dc.description.abstractDenne rapporten dekker arbeid på et 3D object detection rammeverk basert på Vote3Deep arkitekturen. Nevrale nettverk basert på dette rammeverket er trent opp til å detektere objekter i genererte punktskyer. Dette gjøres for å kunne plassere 3D-modeller av objektene i en augmented reality setting. Formålet med arbeidet er å bistå augmented reality visualisering av sensordata fra non-destructive testing.
dc.description.abstractThis report covers work on a 3D object detection framework based on a 3D convolutional network architecture, Vote3Deep. A collection of neural networks using this framework are trained to detect objects within point clouds of varying quality, with the intention of placing digital copies of the detected objects in an augmented reality setting. The end goal of the work is to facilitate augmented reality based visualization of non-destructive testing data.
dc.languageeng
dc.publisherNTNU
dc.titleObject Detection in Point Clouds for Augmented Reality
dc.typeMaster thesis


Files in this item

FilesSizeFormatView

This item appears in the following Collection(s)

Show simple item record