摘要
因摄像机角度而造成的机动车牌图像倾斜会对其后继的字符分割与识别带来不利的影响。本文在分析了车牌倾斜模式的基础上,提出了一种基于最小二乘支持向量机(LS-SVM)的车牌图像倾斜校正新方法。通过LS-SVM线性回归算法求取坐标变换矩阵并对畸变图像进行旋转校正。主要方法:首先,将二值倾斜车牌图像中的像素转换为二维坐标样本,并构造图像数据集;再通过LS-SVM线性回归算法对该数据集进行回归,求取主要参数;最后,再由该参数转换为能反映图像倾斜方向的2维坐标变换矩阵。实验结果表明,该方法简便实用,对光照、污迹等不敏感,抗干扰能力强。
Slant vehicle license plate induced by camera location had the bad effect on its character segmentation and recognnition. After slant models of license plate were analyzed , a new method to remedy the effect based on least squares support vector machine （LS-SVM） was presented. The geometrical transform matrix to correct the image was acquired by the linear regression algorithm of LS-SVM. Firstly, the pixels in binary slant license image were arranged into two-dimension coordinate samplings as image data set. Then, the main regression parameter of the image data set was obtained by the linear regression algorithm of LS-SVM. Finally, the main parameter was refold to two-dimension coordinate transform matrix, which was consistent with the main slant direction of the license image. Experimental results show that the method can be implemented easily and offers robustness when dealing with dirty number plates and license plates in variant lighting conditions.
出处
《微计算机信息》
北大核心
2007年第05X期310-312,共3页
Control & Automation
关键词
车牌
倾斜校正
最小二乘支持向量机
回归算法
Vehicle license plate,slant correction,least squares support vector machine （LS-SVM）,regression algorithm
