on. If we stitch the 
eometrical exterior 
ded that the internal 
enough. 
t between the sub- 
g image can be too 
hoto in a metric 
one possible way to 
bhotos, as shown in 
thin a limited depth 
vithin that depth of 
nages for different 
can be seen in the 
ire 10. The closer 
y or smaller target 
nge increases if the 
aller. The leftmost 
set allows seamless 
1isition, we can use 
tion if we use only 
| of range in focus 
. On the other hand, 
ding recommended 
ature matching and 
to full panoramic 
  
itched with nearly 
ligible perspective 
photos with offset 
the points in close 
te the proper usage 
ch way to best use 
-ogram in our own 
measure a specific 
| is oblique aerial 
ate the perspective 
  
error caused by the movement of the plane between the image 
capture when we know the speed of the plane and the time 
between the acquisitions of two images. In addition, simulation 
can also be used to illustrate the panoramic imaging geometry for 
educational purposes. 
One drawback of our simulation is that it becomes inaccurate if 
we are not able to correctly estimate the eccentricity of rotating 
perspective centres. The future research includes how to 
accurately estimate the eccentricity of real panoramic image 
sequences. 
5. CONCLUSIONS 
The aim of this work was to quantify the perspective error 
introduced by an offset of projection centre in panoramic 
imaging process. This examination was limited to frame 
cameras. We developed simulation software for understanding 
and predicting the effect of such errors. Using our simulation 
software, we illustrated geometrical conditions in both concentric 
and eccentric image acquisition cases. Simulation allows us to 
calculate the largest allowable projection centre offset for any 
particular shooting distance to depth ratio. 
This paper shows that if we increase the shooting distance or 
only consider some objects that are within relatively small depth 
of range, we can accept larger projection centre offsets and still 
maintain relatively high accuracy in the object space. 
The results are applicable in real cases when planning panoramic 
image acquisition. Even if it is advisable to use properly 
calibrated camera rig to get uniform perspective to all images, in 
some cases this is not possible or desired. Therefore, it is 
advantageous to be able to estimate accuracies beforehand using 
simulation. 
6. REFERENCES 
Briggs, A., Li, Y., Scharstein, D., and Wilder, M., 2006. Robot 
navigation using 1D panoramic images. Proc. ICRA 2, 7 pages. 
Brown, D. C., 1971. Close-range camera calibration, Photo- 
grammetric Engineering, vol. 37, no. 8, pp. 855-866. 
Deng, Y. and Zhang, T., 2003. Generating panorama photos. 
Proceedings of the SPIE: Internet Multimedia Management 
Systems IV, vol. 5242, pp. 270-279. 
Fan, Z. and Qi-dan, Z., 2009. Super-Resolution Image 
Reconstruction for Omni-Vision based on POCS. IEEE 
Conference publications, Control and Decision Conference, 
2009. CCDC '09, pp. 5045-5049. 
Fernandes, J. C. A. and Neves, J. A. B. C., "Using Conical and 
Spherical Mirrors with Conventional Cameras for 360° 
Panorama Views in a Single Image," accepted for ICM 2006 - 
IEEE International Conference on Mechatronics, July 3-5, 2006, 
Budapest, Hungary, pp. 157-160. 
Gaspar, J. and Victor, J., 1999. Visual Path Following with a 
Catadioptric Panoramic Camera. International Symposium on 
Intelligent Robotic Systems — SIRS'99, Coimbra, Portugal, pp. 
139-147. 
    
    
    
   
    
    
      
    
   
   
   
   
   
    
   
  
  
   
   
   
   
   
   
  
  
  
  
  
   
  
  
  
  
   
  
   
    
   
    
     
   
  
  
   
  
   
  
    
   
   
     
Hotten, R. and Diprose, P., 2000. From Dreamtime to 
QuickTime: The resurgence of the 360-degree panoramic view as 
a form of computer-synthesised architectural 
representation. Proceedings of ACADIA Eternity, Infinity and 
Virtuality in Architecture, Washington, pp. 1-11. 
Huang, F., Wei, S.K. and Klette, R. 2003. Comparative Studies 
of Line-based Panoramic Camera Calibration. IEEE Int. Conf. 
Computer Vision and Pattern Recognition, vol. 7, pp. 16-22. 
Kauhanen, H, Heiska, N., and Kurkela, M., 2009. Long focal 
length imaging for photogrammetric reconstruction. The 
International Archives of the Photogrammetry, Remote Sensing 
and Spatial Information Sciences, 38 (5/W1), 5 pages. 
Kopf, J., Uyttendaele, M., Deussen, O., and Cohen, M., 2007. 
Capturing and viewing gigapixel images. ACM Transactions on 
Graphics (SIGGRAPH ’07), 10 pages. 
Luhmann, T. and Tecklenburg, W., 2004. 3-D Object 
Reconstruction from Multiplestation Panorama Imagery. In 
Proceedings of ISPRS Workshop on Panorama Photogrammetry, 
Dresden, Germany, Volume 34, Part 5/W16., 8 pages. 
Nakao, T. and Kashitani, A.,2001. In Proceedings of the IEEE 
International Conference on Image Processing, pp. 1045-1048. 
Póntinen, P., 2004. On the Geometrical Quality of Panoramic 
Images. IAPRS 35(Part 5), pp. 82-87. 
Svoboda, T., Pajdla, T. and Hlavac, V., 1997. Central Panoramic 
Cameras: Design and Geometry. Third Computer Vision Winter 
Workshop, Gozd Martuljek, 14 pages. 
Wei, S. K., Huang, Y. F. and Klette, R., 1998. Colour Anaglyphs 
for Panorama Visualizations. Technical Report, CITR-TR-19, 
Computer Science Department of the University of Auckland 
(New Zealand), 15 pages. 
Yan, Z., Ma, X. and Tao, S., 2009. The Virtual Display Study on 
3D Panorama in Tourist Areas - take Shilin World Geopark as 
an example. Proceedings of the 2009 International Symposium 
on Information Processing (ISIP'09), pp. 229-232. 
Yuen, D. C. K. and MacDonald, B. A., 2002. Natural landmark 
based localisation system using panoramic images. Proc. ICRA, 
pp. 915-920. 
Zara, J., 2004. Virtual reality and cultural heritage on the Web. 
In Proceedings of the 7th International Conference on Computer 
Graphics and Artificial Intelligence (31A 2004), pp. 101-112. 
7. ACKNOWLEDGEMENTS 
The article is supported by the Finnish Society of 
Photogrammetry and Remote Sensing.