Research on the Geometric Distortion Auto-Correction Algorithm for Image Scanned

Abstract:

In the process of gaining digital images whether scanning or photographing, the rugged manuscript or the nonlinear characteristic of the camera lens, etc. may lead to severe geometric distortions of the captured images. Therefore, geometric rectification of the distorted images is necessary before they can be further processed. In this paper, A careful analysis was done on the cases of geometric distortion, and then an auto-correction method is deduced. In the correction process for image scanned, global thresholding was used to extract the feature region which was needed, use Sobel operator to detect edge, then use Hough transform to auto-extract the boundary. After Hough transform, some parameters associated with the control point were obtained for correcting images. Extensive experiments proved that it can recognize the outline fast and correct geometric distortions effectively. The results show that the method was applicable and had high accuracy.

• Periodical: Applied Mechanics and Materials (Volumes 644-650)
• Pages: 4477-4481
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.644-650.4477
• Online since: September 2014
• Authors: Jing Zhang, Guang Xue Chen*
• Keywords: Auto-Correction, Edge Detection, Geometric Distortion, Image Processing

References

[1] R.C. Gonzalez, R.E. Woods: Digital Image Processing (Publishing House of Electronics Industry, Beijing 2004).

[2] Q. Zhang, Z.L. Wang: Master MATLAB Image Processing (Publishing House of Electronics Industry, Beijing 2012).

[3] S. Zhang, Q.F. Zhao and J.K. Ye: Mechatronics Vol. 3 (2007), p.60.

[4] Y. Lu, G.M. Chen and S. Cheng: Computing Engineering and Design Vol. 17 (2007), p.4290.

[5] Information on http: /en. wikipedia. org/wiki/Perspective_control.