Evidence of Walls in Oblique Images for Automatic Verification of Buildings
Year: 2010
Type: conference-paper
Abstract: With the availability of airborne sensor platforms capable of capturing multiple overlapping images, oblique images have become interesting not only for visualization but also for acquisition, verification and update of vector data. In recent years, techniques for automatic verification of buildings vector data have used building characteristics associated with roof colour, texture and height while oblique images contain wall facade information that can also be used to identify buildings. This paper presents a methodology to verify 2D vector datasets of buildings by searching in oblique images clues such as building edges, wall facade edges and texture. The 2D evidence in images taken from different perspective are transformed to 3D and then matched. Results from experiments clearly distinguish existing buildings from demolished ones. The assessment of a building is done by combining results from individual walls. Thus, besides overall building verification, the results are useful for discriminating demolished walls of modified buildings.
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2181714579