Rotation Distortion Invariant Image Recognition Based on Fractional Correlation

Abstract:

In this paper, a fast algorithm for fractional Fourier transform was proposed based on the theory of fractional Fourier transform, the fast numerical calculation of the fractional Fourier transform and fractional correlation was realized using MATLAB language. The characteristics of rotation distortion invariant image recognition based on fractional correlation were analyzed and compared with the traditional correlation. The results show that the quality of the fractional correlation peak is obviously improved. When fractional order P = 0.7, the fractional correlator can realize rotation distortion invariant image recognition within the scope of 15°.

• Periodical: Applied Mechanics and Materials (Volumes 220-223)
• Pages: 2899-2902
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.220-223.2899
• Online since: November 2012
• Authors: Hong Xia Wang, Pan Shi Li, Zhan Rong Zhou, You Zhang Zhu
• Keywords: Fractional Correlation, Fractional Fourier Transform (FrFT), Image Recognition, Rotation Distortion Invariant Recognition

References

[1] Mendlovic D, Ozaktas H M: J.Opt.Soc.Am.A, Vol. 10(1993), p.1875

[2] A.W. Lohmann: J.Opt.Soc.Am.A, Vol. 10(1993), p.2181

[3] B H Zhu, S T Liu, L X Chen: Appl.Opt., Vol. 40(2001) , p.6474

[4] Weimin Jin, Caijie Yan, Lihohg Ma, Hui Ye, Hui Wang: Opt Commun, Vol. 268(2006), p.34

[5] Qu Wang, Li Chen, Jinyun Zhou, Qinghua Lin: Opt. Eng, Vol. 50(2011), p.18201

[6] Namias V: J .Inst .Maths Applies, Vol. 25(1980), p.241