Abstract: For the high requirements of the processing speed and recognition rate of the paper currency number recognition in practical engineering applications, a method based on structural features of the characters is proposed in this paper. Firstly, it preprocesses the image, including zoom, binarization and character segmentation. Then, it classifies and recognizes the characters according to the structural features. The method does not require denoising, skew correction and normalization, so it can save much time. Experimental results show that the recognition rate can reach 98.13%, while the recognition time is only 3.8 ms. So the method has advantages of high recognition rate and fast recognition speed, it can meet actual needs, and the method has been applied to the sorter successfully.

Key words: paper currency number, number recognition, structural features, character segmentation

摘要： 针对实际工程应用中对纸币号码识别在处理速度和识别率方面的高标准要求，提出了一种基于结构特征的纸币号码识别方法。该方法对采集到的图像先进行图像放大、二值化和字符分割等预处理操作，然后根据字符的结构特征，对字符进行分类、识别。该方法不需要图像去噪、倾斜校正和归一化等操作，节省大量时间。实验结果表明，该方法识别率可达98.13%，而识别时间仅为3.8 ms，显示出该方法具备识别率高和识别速度快的优势，能够满足实际需要，已经成功应用于清分机中。

关键词: 纸币号码, 号码识别, 结构特征, 字符分割