BROWSE
Detail
Conference Paper
A Texel-based Texture Extraction of the Building Facade from Aerial Images
Cited - time in 
Share 
Share
- Authors
- Young Jae Lim, Ji Sang Park
- Issue Date
- 2009-10
- Citation
- International Symposium on Remote Sensing (ISRS) 2009, pp.1-4
- Language
- English
- Type
- Conference Paper
- Abstract
- To build the three-dimensional building models, several image-based methods using building's shadows and/or relief displacement on the single aerial imagery or photometric survey techniques for two or more images have been employed. These methods have an advantage that realistic building modeling is available more rapidly because texture images of the building's roof and wall can be obtained directly from the images. However, it is difficult to get good quality textures when buildings are located near the principal point of the imagery or the direction between building walls and camera are not aligned properly because image are normally taken directly down to the ground. In this paper, a method is proposed to estimate the texture quality that helps the selection of the optimal texture of the buildings wall from the several aerial images. The proposed algorithm estimates the texture quality from the relationship between the principal point of each image and the wall element. Moreover, the texture quality may get worse due to the adjacent buildings, trees or shadows. In order to eliminate these noise factors and acquire the better texture, a new algorithm is proposed, which restores the occluded part with texels that are the repetitive patterns of each floor.
- KSP Keywords
- Adjacent buildings, Building facade, Building modeling, Building walls, Image-based method, Relief displacement, Texture extraction, Texture image, Three dimensional(3D), aerial imagery, new algorithm