Constructing accurate architectural models of interiors is a nontrivial task. Photogrammetry can be used to recover the large-scale features but has difficulty in modelling sufficient detail. Laser scanning on the other hand produces highly detailed depth data but is difficult to turn into a globally consistent model. We present a hybrid system for the creation of photorealistic and geometrically efficient and accurate models through the integration of photogrammetric modelling techniques with laser acquired range data. Our hybrid rendering software builds on the foundation of the Facade photogrammetric modelling system, but improves the degree of complexity of this representation by incorporating range data for those parts of the scene that are unmodelled. The salient planar features in the scan are extracted and used to align the models via a two-stage quaternion-based and ICP-based registration. The process allows for straightforward integration of multiple range scans. Rendering of novel views is achieved using view-dependent texture mapping techniques combined with point-based rendering.
No takes yet. Share an insight, caveat, or question.
Lamond et al. (2004) studied this question.
Synapse has enriched 4 closely related papers on similar clinical questions. Consider them for comparative context: