Recently, the three-dimensional (3D) sensing technique that uses multiple cameras has been applied to various areas, such as visualization, motion capturing, and so on. However, improvement of camera model calibration is required for higher precision of measurements. In this study, we propose a practicable fuzzy modeling approach for 3D sensing that utilizes stereo vision configuration. The distance between a sensing target and the camera is used to construct a camera fuzzy model that considers optical projection characteristics. In our approach, the weighted least squares method is successfully applied considering a fuzzy partition to formulate the fuzzy model. Then, iterative calculations for solving the inverse problem of the camera fuzzy model are performed to obtain measured coordinates. Through sensing experiments of stereo vision measurement based on the proposed approach, we show that performance of the model is drastically improved compared with the conventional modeling approach.

1. [1] Y. Saito and T. Watanabe, “A 3D sensing technique using fuzzy modeling based on stereo vision,” Proc. of the 6th Int. Conf. on Soft Computing and Intelligent Systems, and the 13th Int. Symposium on Advanced Intelligent Systems (SCIS&ISIS2012), pp. 2138-2142, 2012.
2. [2] R. Hartly and A. Zisserman, “Multiple View Geometry in Computer Vision,” Cambridge university press, 2000.
3. [3] K. Deguchi, “Foundation of Robot Vision,” Corona Publishing, 2000.
4. [4] J. Sato, “Computer Vision – Geometry of Vision –,” Corona Publishing, 1999.
5. [5] Z. Zhang, “A flexible new technique for camera calibration,” IEEE Trans. on PAMI, pp. 1330-1334, 2000.
6. [6] Z. Zhang, “A flexible new technique for camera calibration,” Microsoft Technical Report, MSR-TR-98-71, 1998.
7. [7] M. Fujigaki, “Whole-space tabulation method for real-time shape measurement and compact strain distribution measurement system,” Proc. of ICCES’09, pp. 566-566, 2009.
8. [8] Y. J. Xing, J. Xing, J. Sun, and L. Hu, “An improved neural networks for stereo-camera calibration,” J. of Achievements in Materials and Manufacturing Engineering, Vol.20, Issues 1-2, pp. 315-318, 2007.
9. [9] T. Xinxing, X. Miao, and H. Lipeng, “RBF neural network-based calibration of wide angle fovea lens for active vision tracking,” J. of Theoretical and Applied Information Technology, Vol.50, No.3, pp. 689-695, 2013.
10. [10] T. Takagi and M. Sugeno, “Fuzzy identification of systems and its applications to modeling and control,” IEEE Trans. on Syst., Man, and Cybern., Vol.15, pp. 116-132, 1985.
11. [11] T. Watanabe and H. Seki, “Modeling approach based on modular fuzzy model,” J. of Advanced Computational Intelligence and Intelligent Informatics, Vol.16, No.5, pp. 653-661, 2012.