Presentation 2011-02-18
A study for segmentation method of buildings from bird's eye view and streetscape images based on regular patterns in buildings
Satoshi UENO, Shigeyuki SAKAZAWA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, we propose a segmentation method to distinguish a building area from a background based on bird' s eye view and streetscape images published on the web. Conventional methods use edge information on the buildings; however, it is difficult to detect the edge from those images on the web due to their low resolution. In our method, we utilize a periodical structure on the building wall such as windows as well as edges in order to improve robustness for poor image quality. Experimental results show that our proposed method can distinguish the building area correctly for 62% of the buildings in Shinjuku metropolitan area. The building area segmentation is followed by the estimation of window area ratio per wall area on the building, which is used in a radio wave propagation simulation for wireless communication systems.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) building detection / window detection / blob region / streetscape image / bird's eye view image
Paper # PRMU2010-236
Date of Issue

Conference Information
Committee PRMU
Conference Date 2011/2/10(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Pattern Recognition and Media Understanding (PRMU)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) A study for segmentation method of buildings from bird's eye view and streetscape images based on regular patterns in buildings
Sub Title (in English)
Keyword(1) building detection
Keyword(2) window detection
Keyword(3) blob region
Keyword(4) streetscape image
Keyword(5) bird's eye view image
1st Author's Name Satoshi UENO
1st Author's Affiliation KDDI R&D Laboratories Inc.()
2nd Author's Name Shigeyuki SAKAZAWA
2nd Author's Affiliation KDDI R&D Laboratories Inc.
Date 2011-02-18
Paper # PRMU2010-236
Volume (vol) vol.110
Number (no) 414
Page pp.pp.-
#Pages 6
Date of Issue