本研究提出一個新內插方法，進行低階車牌影像強化。本研究認為影像的灰階並非單一點的灰階值而是一小塊區域的平均灰階值，內插時將原像素區域分割為較高階之像素區域，並且考慮原像素與兩側之灰階值，當三者的關係為遞增或遞減時，計算原像素與兩側灰階值差距，令靠近差距較小的高階區域與其側邊之灰階值相同，而另一邊之高階區域灰階值則依照區域平均灰階值來決定。內插的作法分為水平、垂直與左右45度四個方向進行，最後將各方向所得的灰階值平均。
本研究中將比較新內插方法與數種靜態影像重建法的結果，規劃兩部份測試方法，分別透過影像退化模型所得的測試影像與實際車牌影像進行實驗，由實驗結果可知，新內插法的確可改善模糊的情形，使影像的輪廓更加明顯。

In this study, a new interpolation method was proposed to enhance low resolution license plate images. We believed that the gray value of the image is close to the average gray value of an area, instead of the gray value of a certain point. With the novel approach, a pixel is divided into a 2x2 area first. The gray value of the original pixel and neighbors of the original pixel are all taken into account. When the gray level in a certain direction increased or decreased gradually, the difference in gray level between original pixel and both sides of the original pixel are calculated. If the difference at a certain side is smaller, the new pixel at side is assigned to the gray level of its neighbor, while the average of the gray level is equal to that of the original pixel. After considering horizontal, vertical, and orthogonal directions, the gray levels for the new 2x2 area are determined.
In this study, the new approach and several static reconstruction methods are tested. Images from image degradation model and those taken from camera directly were used to test these approaches. Based on the test results, the new method shows promising potential for improving the license plate recognition.

[1]W. B. Horng, C. L. Lee, C. H. Fan, “A Study and Implementation on Automatic Intelligent Vehicle License Plate Recognition Systems,” 第二屆台灣智慧型運輸系統ITS國際研討暨展覽會, 臺北, Apr. 2000.

[2]M. J. Ahmed, M. Sarfraz, A. Zidouri, W. G. Al-Khatib, “License Plate Recognition System,” Electronics, Circuits and Systems, 2003. ICECS 2003. Proceedings of the 2003 10th IEEE International Conference on Volume 2, 14-17 Dec. 2003 Page(s):898 - 901 Vol.2

[3]S. Z. Wang, H. J. Lee, “Detection and Recognition of License Plate Characters with Different Appearances,” Intelligent Transportation Systems, 2003. Proceedings. 2003 IEEE Volume 2, 12-15 Oct. 2003 Page(s):979 - 984 Vol.2

[4]T. Naito, T. Tsukada, K. Yamada, K. Kozuka, S. Yamamoto, “Robust License-Plate Recognition Method for Passing Vehicles Under Outside Environment,” Vehicular Technology, IEEE Transactions on Volume 49, Issue 6, Nov. 2000 Page(s):2309 - 2319 Digital Object Identifier 10.1109/25.901900

[5]R. R. Schultz, R. L. Stevenson, “A Bayesian approach to image expansion for improved definition,” Image Processing, IEEE Transactions on Volume 3, Issue 3, May 1994 Page(s):233 - 242 Digital Object Identifier 10.1109/83.287017

[6]R. Y. Tsai, T. S. Huang, “Multiframe image restoration and registration,” Advances in Computer Vision and Image Processing (R. Y. Tsai, T. S. Huang, Eds.), vol. 1, pp.317-339, JAI Press, London, 1984.

[7]S. C. Park, M. K. Park, M. G. Kang, “Super-resolution image reconstruction: a technical overview,” Signal Processing Magazine, IEEE Volume 20, Issue 3, pp. 21 – 36, Digital Object Identifier 10.1109/MSP.2003.1203207, May 2003.

[8]R. G. Keys, ”Cubic Convolution Interpolation for Digital Image Processing,” IEEE Trans. vol. ASSP-29, no. 6, pp. 1153-1160, 1981.

[9]H. Greenspan, C. H. Anderson, and S. Akber, “Image Enhancement by Nonlinear Extrapolation in Frequency Space,” IEEE Trans. on Image Processing, vol.9, no.6, pp. 1035-1048, 2000.

[10]J. G. Leu, “Edge sharpening through ramp width reduction,” Image and Vision Computing, vol. 18, no. 6-7, pp. 501-514, 2000.

[11]X. Li and M. T. Orchard, “New Edge-Directed Interpolation,” IEEE Trans. on Image Processing, vol. 10, pp. 1521-1527, 2001.

[12]G. Ramponi, ”Warped Distance for Space-Variant Linear Image Interpolation,” IEEE Trans. on Image Processing, vol. 8, no. 5, pp. 629-639, 1999.

[13]W. T. Freeman, E. C. Pasztor, and O. T. Carmichael, “Learning Low-Level Vision,” International Journal of Computer Vision, vol. 40, no. 1, pp. 25-47, 2000.

[14]W. T. Freeman, T. R. Jones, and E. C. Pasztor, “Example-Based Super-Resolution,” IEEE Trans. Computer Graphics and Applications, vol. 22, no. 2, pp. 55-65, 2002.

[15]S. A. Nene, S. K. Nayar, “A Simple Algorithm for Nearest Neighbor Search in High Dimensions,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 19, no. 9, pp.989-1003, Sept. 1997.

[16]S. Baker and T. Kanade, “Limits on Super-Resolution and How to Break Them,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 24, no. 9, pp. 1167-1183, 2002.

[17]J. Sun, N. N. Zheng, H. Tao, and H. Y. Shum, “Image Hallucination with Primal Sketch Priors,” Proceedings of CVPR ’03, Madison, Wisconsin, pp.729-736, 2003.

[18]C. Su, Y. Zhuang, L. Huang, and F. Wu, “Steerable Pyramid-based Face Hallucination,” Pattern Recognition, vol. 38, pp. 813-824, 2005.

[19]M. V. Joshi, S. Chaudhuri, and R. Panuganti, “A Learning-Based Method for Image Super-Resolution From Zoomed Observations,” IEEE Trans. on System Man and Cybernetics, vol.35, no.3, pp. 527-537, 2005.

[20]Z. Wang, F. Qi, “On Ambiguities in Super-Resolution Modeling,” Signal Processing Letters, IEEE Volume 11, Issue 8, Aug. 2004 Page(s):678 - 681 Digital Object Identifier 10.1109/LSP.2004.831674

[21]S. Chaudhuri, D. R. Taur, “High-Resolution Slow-Motion Sequencing: How to Generate a Slow-Motion Sequence from a Bit Stream,” Signal Processing Magazine, IEEE Volume 22, Issue 2, Mar 2005 Page(s):16 - 24 Digital Object Identifier 10.1109/MSP.2005.1406471

[22]R. C. Gonzalez, R. E. Woods, “Digital Image Processing 2/E,” Prentice Hall Publication, pp. 57-58, pp. 220-262, 2002.

[23]E. D. Castro, C. Morandi, “Registration of translated and rotated images using finite Fourier transform,” IEEE Transactions on Pattern Analysis and Machine Intelligence 9 (1987) 700–703.