In: Paparoditis N., Pierrot-Deseilligny M.. Mallel C.. Tournaire O. (Eds), IAPRS. Vol. XXXVIII. Part ЗА - Saint-Mandé, France, September 1-3, 2010 
Position of S (Z) 
1 ; . 
.*♦♦!• f! ! ♦ ♦ ♦ ; 
■ :. ! 
к 
[ ♦ 0,3 pixel of noise *0,5 pixel of noise 1,0 pixel of noise 
Figure 9: Projection center of images in Z after calibration and 
noise on measures 
Positions of S (X,Y) 
-0,25 -0,2 -0,15 -0,1 -0,05 0 0,05 0,1 0,15 0,2 0,25 
A 
1 
1 
'— 
* ■ 
♦ 1 
♦ 
♦ 
.♦ ; 
♦ 
a 1 
■ 
À 
■ 
■ 
♦ " 
♦ 
■ 
■ 
Xjm] 
♦ 0,3 pixel of noise ■ 0,5 pixel of noise 1,0 pixel of noise 
Figure 10: Centres of projection of images in plan (X.Y) after 
calibration and noise on measures 
Position of S (Z) 
♦ 0,3 pixel of noise ■ 0,5 pixel of noise 1,0 pixel of noise 
Figure 11: Centres of projection of images in Z after calibration 
and noise on measures 
3.3 Conclusion 
In this section, we exhibited the limits of the polygon based cal 
ibration and the difficulty to estimate the intrinsic and extrinsic 
parameters in the same process. The geometry of the panoramic 
avoids calculating the position of each image. Furthermore, ge 
ometry of panorama constraint naturally and geometrically intrin 
sic parameters. 
4 EXPERIMENTS WITH REAL DATA 
2 models can be more than 10 pixels (in the corners). It means 
an difference on field of camera of 0.7123° in column and per 
0.5989° in line. 
Figure 12: Sample of panorama 
5 CONCLUSIONS AND FUTURE WORK 
This paper has presented a method to calibrate a camera's intrin 
sic parameters but also to estimate a distortion pattern. We have 
shown that this method is not sensitive to noise and it is applica 
ble with both short and long focal. Future work will consist on 
looking for more complex and accurate model. We also look for 
a framework to estimate the mechanical residual parallax. 
REFERENCES 
Agapito. L.. Hayman. E. and Reid, I., 2001. Self-calibration of 
rotating and zooming cameras. International Journal of Computer 
Vision 45(2), pp. 107-127. 
Coorg, S. and Teller, S.. 2000. Spherical mosaics with quater 
nions and dense correlation. In: International Journal of Com 
puter Vision 37(3), pp. 259-273. 
Craciun. D.. Paparoditis, N. and Schmitt, F., 2009. Automatic 
gigapixel mosaicing in large scale unstructured underground en 
vironments. In: IAPR Machine Vision Applications, Yokohama, 
Japan. 
Hartley, R. I.. 1994. Self-calibration from multiple views with 
a rotating camera. In: ECCV ’94: Proceedings of the third Eu 
ropean conference on Computer vision (vol. 1), Secaucus, NJ, 
USA. pp. 471-478. 
Lowe, D. G., 2004. Distinctive image features from scale- 
invariant keypoints. International Journal of Computer Vision 
60(2) pp. 91-110. 
Ramalingam. S., Sturm, P. and Lodha. S. K., 2010. Generic self 
calibration of central cameras. Computer Vision and Image Un 
derstanding 114(2), pp. 210-219. Special issue on Omnidirec 
tional Vision, Camera Networks and Non-conventional Cameras. 
Tordoff, B. and Murray, D. W., 2004. The impact of radial dis 
tortion on the self-calibration of rotating cameras. International 
Journal of Computer Vision 96(1). pp. 17-34. 
After working with simulated data, we have calibrated a real cam 
era. The interest of this acquisition is to validate the convergence 
of the process on real data. We also compare our results with a 
calibration on a topometrically surveyed network targets. 
After a bundle adjustment by estimating only rotation and the 
camera calibration obtained on targets network, we have a value 
of RMS equal to 0.021. With our method, we have 0.015 RMS. 
i.e. 28% better. The focal length difference between two tech 
niques is more important than 18 pixels. The difference between 
Tsai. R., 1986. An efficient and accurate camera calibration tech 
nique for 3d machine vision. In: IEEE Computer Vision and 
Pattern Recognition, Miami. FL. USA, pp. 364-374. 
Zhang, Z., 2000. A flexible new technique for camera calibration. 
In: IEEE Transactionson Pattern Analysis and Machine Intelli 
gence 22(11), pp. 1330-1334.