一种字符识别算法在自动识别系统中的应用

光学字符识别是模式识别领域中最经典也是得到最广泛应用的方向之一,而车牌识别系统是应用光学字符识别技术的典型系统。文章基于车牌识别系统的开发,提出一种基于二值图像的字符识别算法。在该算法中,提取字符的点阵特征、特征线和网格特征,分类器采用神经网络。为充分利用各组特征向量的互补作用,采取层次结构来获得系统的最佳性能。在车牌号图片库中测试其算法。实验表明,算法非常有效。

A character recognition algorithm in the application of vehicle license plate recognition system

Optical character recognition(OCR) is one of classical fields in pattern recognition and becomes one of most successful applications in auto recognition.Vehicle license plate recognition system(VLPRS) is one of classical systems using OCR.In the paper,we extract the features including the lattice structure,key line and grid features.The network is used as the classification.To make full of use of all kinds of feature vectors,the hierarchical algorithm is adapted for optimal performance.We test our algorithms in car number plate database.As the experimental results show,our algorithm is very efficient.