基于旋转变换和鲁棒主成分分析的车牌校正方法

摘要

本文研究了受到非高斯噪声污染及边框信息不完整的车牌图像校正的问题.利用鲁棒主成分分析与旋转变换结合的方法,获得了更具普适性的车牌矫正方法.并通过与主成分分析法、旋转投影法的矫正结果相比较,推广了本文方法具有更好的鲁棒性和普适性的结果.%In this paper, we study the problem of correcting the license plate image which are polluted by non-gaussian noise or with incomplete border information. By using the method of robust principal component analysis and rotation transformation, a more general applicability method of license plate correction is obtained. Through the comparison of the correction results with the method of principal component analysis and rotation projection, we popularizes the conclusion that the proposed mthod has better robustness and general applicability than other methods.