5. [stanbul 2004 
bove is pixel, and 
coordinate system 
en by the way of 
o. (9)- (12)). First 
joints to the four 
of calibration are 
ly, the multi-view 
done (the results 
art 1). 
  
90°) 
  
original image 
(270^) 
ibration test, the 
ble; however, the 
iibration test are 
iges 
9933 
436 
017 
  
le-view, and then 
rs calibrated with 
view is greatly 
1gth ete. Thereby, 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004 
  
we proposed a calibrating method of vanishing points based on 
multi orientations and multi views. Compared with single-view 
calibration, multi-view calibration has the characteristic as below: 
1) put interior and exterior orientation parameters of camera 
together directly into the model of calibration. Through 
vanishing points, direct functional model between calibrated 
parameters and observations (image lines) is formulated. 
Sequentially, interior and exterior parameters are united to adjust 
computation, and then exterior angle orientation parameters 
adjusted (namely camera pose) can be used to build 3D object 
model. 2) the result is not influenced by shooting angle. Under 
A-v-K , A,v 
- o o ne . . 
from 20" — 70" , the error of interior parameters changes little; 
the. system of when range 
3) the error of calibrated parameters has been improved greatly. 
Therefore, this camera calibration method with vanishing points 
of multi-image has the special advantage that other methods 
don't exist. In the case of no high-precision test range, this 
calibration method mentioned in this paper has good prospects 
aiming at applying in calibration of common zoom digital 
camera. 
7. REFERENCES 
[1] Wang Zhizhuo, 1980. The Theory of Photogrammetry. 
Surveying Press. 
[2] Ma Songde, Zhang Zhengyou., 1998. Computer Vision. 
Press of Science. 
[3] Feng Wenhao, 2002. Close-Range Photogrammetry. Press 
of Wuhan University. 
[4] Zhang Zuxun, Zhang Jianging, Zhang Shixing, 2001. 
Three-dimensional building model based on single image. 
Workshops on automatic engineering surveying of channel. 
[5] Frank A. Heuvel Vanishing point detection for architectural 
photogrammetry. International archives of 
photogrammetry and remote sensing. Vol.32 part5,ISSN 
0256-1840. 
[6] R.Cipolla and E.Boyer, 1998. 3D model acquisition from 
uncalibrated images. Proc.IAPR Workshop on Machine 
Vision Applications,Japan. 
[7] SchusterR, Ansari.N, Bani-Hashemi.A, 1993. Steering a 
robot with vanishing points. Robotics and Automation" 
IEEE Transactions on Rbotics and Automation, Vol.9,No. 4. 
IS]; Fengjun. Lv, Tao Nevatia, 2002. 
Self-Calibration of a camera from video of a walking 
Zhao, Ram 
human. /nternational Conference on Pattern Recognition, 
Quebec City, Canada, pp. III: 639-644 
[9] Ling-Ling Wang, Wen-Hsiang Tsai, 1991. 
Calibration by vanishing Lines for 3-D computer vision. 
Camera 
IEEE Transactions on Pattern Analysis and Machine 
Intelligence. Vol.13,No.4 
[10] Wu Qing, Xu Guangyou, Wang Lei, 2001. A Three-stage 
SPIE Conference 
Proceeding: Multispectral Image Processing and Pattern 
system for camera calibration. 
Recognition, p22-24 
[11] S.Barnard, 
[12] 
2 
] 
Interpreting perspective 
Intelligence, Vol.21,1983,435-462 
E.Lutton, H.Maitre,J. Lopez-Krabe, 1994. Contribution to 
the determination 
images. Artificial 
of vanishing points using hough 
transform. /EEE Transactions on Pattern Analysis and 
Machine Intelligence. Vol.16,No.4 
Leung.J.C.H, NcLean.G.F, 1996. Vanishing point matching. 
1996. Proceedings., 
Conference on, Vol.2,p305-308, Switzerland. 
Image Processing, International