一种基于特征加权模板匹配方法在纸币字符识别中的应用

Application of Paper Currency Character Recognition Based on Feature Weight Template Matching

纸币字符是纸币的重要特征之一,一组字符能够唯一标识纸币的身份.能够快速地识别纸币中的字符直接影响到纸币字符识别的精度和速度.本文运用特征加权模板匹配算法,设计了一个纸币字符识别算法.该算法从当前应用最广泛的模板匹配法入手,对标准模板匹配算法加以改进,从而提高了纸币字符实别的效率和精度.该算法能够充分区分开字符笔画和非笔画部分对字符识别的影响大小,从而有效地提升了识别率和鲁棒性.从实验结果来看,该算法性能较优.

Paper currency is one of the important features of character notes, a group of characters to uniquely identify the paper currency identification. Able to quickly identify the paper characters directly affect character recognition accuracy and speed of paper money. A character recognition method was designed. Based on template matching algorithm which is widely used, the designed method firstly improved the standard template matching algorithm and employed fuzzy theory as identification criterion to recognize character. The proposed method can fully separate the influence on character recognition which was caused character stroke and character tailing, so as to effectively improve the recognition rate and robustness. The experiment results show the designed method has better performance.