We propose to use an active shape model for correcting a road map in a stable manner. Active shape model can deform itself preserving the given basic shape by restricting the deformation to an affine transformation. In order to consider the topological connections for road maps, we extend the active shape model to network active shape model. It allows to deform road network globally preserving a topology of road network. By iterating the deformation step, road network is refined gradually. Experimental results will show that proposed method can refine the existing road map correctly by fitting aerial image.
Network Active Shape Model for Updating Road Map from Aerial Images
2006 IEEE Intelligent Vehicles Symposium ; 325-330
2006-01-01
4130785 byte
Conference paper
Electronic Resource
English
Network Active Shape Model for Updating Road Map from Aerial Images
British Library Conference Proceedings | 2006
|Automatic extraction of road network from aerial images
IEEE | 2006
|Finding Road Seeds in Aerial Images
British Library Online Contents | 1993
|Updating road navigation model using non-semantic road feature points
European Patent Office | 2024
|Updating virtual aerial map using sensors associated with aerial vehicles
European Patent Office | 2023
|