A Novel Point Matching Method for Stereovision Measurement Using RANSAC Affine Transformation

Abstract:

A binocular point matching method using affine transformation is presented in this paper to deal with matching ambiguities. The epipolar geometry is used to determine all the possible matching pairs to form an initial correspondence data set. Then, an affine registration model with four parameters that is invariant to scaling, rotation and translation is built using the Random Sample Consensus (RANSAC) method to describe the coordinate transformation between the two members of correspondences. Finally, correspondences are picked out using minimal nearest neighbor distances based on the geometric similarity between the right image points and the transformed left ones. The proposed method is applied to measure the profile of a 3.5m parabolic reflector of an inflatable antenna and proved to be able to handle the extra or missing point matching problems coursed by occlusion and sheltering. Satisfactory results are obtained with high correct rate for matching most image pairs despite the significantly different viewpoints, which indicates its validity and application feasibility in automatic stereovision measurement field.

• Periodical: Key Engineering Materials (Volume 437)
• Pages: 184-188
• DOI: https://doi.org/10.4028/www.scientific.net/KEM.437.184
• Online since: May 2010
• Authors: Nai Guang Lu, Peng Sun, Wen Yi Deng, Lian Qing Zhu, Xiao Ping Lou
• Keywords: Affine Transformation, Point Matching, RANSAC, Stereo Vision Measurement

References

[1] S.H. Chang, F.H. Cheng, W.H. Hsu and G.Z. Wu: Fast Algorithm for point pattern matching: invariant to translations, rotations and scale changes, Pattern Recognition, Vol. 30 (1997), p.311.

DOI: 10.1016/s0031-3203(96)00076-3

[2] Andrea Fusiello, Emanuele Trucco and Alessandro Verri: A Compact Algorithm for Rectification of Stereo Pairs, Machine Vision and Applications, Vol. 12 (2000), p.16.

DOI: 10.1007/s001380050120

[3] Peng Sun and Naiguang Lu: A robust and efficient point matching method for stereovision measurement using dynamic programming, Proc. SPIE, Vol. 6834 (2007), p.6834 2W.

[4] M.A. Fischler and R.C. Bolles: Random sample consensus: A paradigm for model fitting with applications to image analysis and automated cartography, Comm. of the ACM, 24 (1981), p.381.

DOI: 10.1145/358669.358692

[5] Naiguang Lu, Wenyi Deng, Yongqiang Wang and Mingli Dong: Profile measurement of microwave antenna using close range photogrammetry, Proc. SPIE, Vol. 5852 (2005), p.508.

[6] Manolis I. A. Lourakis, Stavros V. Tzurbakis, Antonis A. Argyros and Stelios C. Orphanoudakis: Feature Transfer and Matching in Disparate Stereo Views through the Use of Plane Homographies, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 25 (2003).

DOI: 10.1109/tpami.2003.1177157