Extraction of Three-Dimensional Architectural Data from QuickBird Images
- PDF / 613,004 Bytes
- 8 Pages / 595.276 x 790.866 pts Page_size
- 34 Downloads / 213 Views
RESEARCH ARTICLE
Extraction of Three-Dimensional Architectural Data from QuickBird Images Peifeng Zhang & Yuanman Hu & Zaiping Xiong
Received: 15 April 2013 / Accepted: 12 August 2013 # Indian Society of Remote Sensing 2013
Abstract Extraction of accurate spatial information from high-resolution satellite imagery is becoming increasingly important for a variety of tasks. In this study, threedimensional architectural data were extracted from QuickBird images using Barista’s monoplotting function. We evaluated the accuracy of the Rational Polynomial Coefficients bundle adjustment and extracted building heights. We obtained accuracies of one-pixel in geo-positioning and 2.66 m in building height. The height accuracy is 0.16 m greater than the estimated error for a one-pixel image measurement. The presence of roof overhangs is one primary factors affecting height accuracy. The application of three-dimensional architectural data represents well the vertical extension of urban growth in Tiexi District from 2002 to 2008.
Keywords Quickbird . Barista . Monoplotting . Architectural landscape . Variation
This work was funded by the Projects of National Natural Science Foundation of China (41301198) and (41171155), the Fundamental Research Funds for the Central Universities(14CX02081A). P. Zhang : Y. Hu (*) : Z. Xiong State Key Laboratory of Forest and Soil Ecology, Institute of Applied Ecology, Chinese Academy of Sciences, 72 Wenhua Road, Shenyang, Liaoning 110164, China e-mail: [email protected] P. Zhang e-mail: [email protected] P. Zhang College of Pipeline and Civil Engineering, Chinese University of Petroleum, 66 West of Changjiang Road, Qingdao, Shandong 266580, China
Introduction Three-dimensional architectural reconstruction has considerable practical application in fields as diverse as municipal planning, real estate, and environmental management (Suveg and Vosselman 2004). Unfortunately, quantitative information about urban three-dimensional architectural data is frequently unavailable, incomplete, or out-of-date. Although original building plans contain detailed information, they are often poorly filed, stored, and maintained. It is extremely difficult to establish an accurate and full inventory of urban buildings and structures by assembling building design plans in a piecemeal fashion (Yang et al. 2006). Field surveys can be conducted to measure the footprint and height of buildings, while, they are often labor-intensive, time-consuming and error prone. Moreover, only limited urban areas can be covered by conventional ground surveys. Various remote sensing images and methods have been used to efficiently extract urban building information, such as automatic three-dimensional building reconstruction from aerial images (Suveg and Vosselman 2002), LIDAR data and multispectral imagery (Cici et al. 2009; Awrangjeb et al. 2010; Yu et al. 2010; Kim and Muller 2011), high-resolution satellite imagery (Wang and Wang 2009; Lu et al. 2011) and synthetic aperture radar (SAR) images (Tupin and Roux 2003; Soergel et al. 2009)
Data Loading...
