3D modelling of building interiors has gained a lot of interest recently, specifically since the rise of Building Information Modeling (BIM). A number of methods have been developed in the past, however most of them are limited to modelling non-complex interiors. 3D laser scanners are the preferred sensor to collect the 3D data, however the cost of state-of-the-art laser scanners are prohibitive to many. Other types of sensors could also be used to generate the 3D data but they have limitations especially when dealing with clutter and occlusions. This research has developed a platform to produce 3D modelling of building interiors while adapting a low-cost, low-level laser scanner to generate the 3D interior data. The PreSuRe algorithm developed here, which introduces a new pipeline in modelling building interiors, combines both novel methods and adapts existing approaches to produce the 3D modelling of various interiors, from sparse room to complex interiors with non-ideal geometrical structure, highly cluttered and occluded. This approach has successfully reconstructed the structure of interiors, with above 96% accuracy, even with high amount of noise data and clutter. The time taken to produce the resulting model is almost real-time, compared to existing techniques which may take hours to generate the reconstruction. The produced model is also equipped with semantic information which differentiates the model from a regular 3D CAD drawing and can be use to assist professionals and experts in related fields.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:582355 |
| Date | January 2013 |
| Creators | Abdul Shukor, Shazmin |
| Publisher | University of Warwick |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://wrap.warwick.ac.uk/55722/ |
Page generated in 0.002 seconds