中文摘要：

设计了一种能准确识别纸币号码的系统；该系统首先通过CCD摄像头采集图像，然后经过图像处理、扫描法定位、模板匹配几个过程输出号码串；为了提高识别率，采用基于图像轮廓提取和利用图像上、下矩对字符进行粗分类的模板匹配算法，显著减少模板匹配的计算量；实验结果表明，该方法取得了较高的识别率，并且已经在一个实际的系统中得到应用。

英文摘要：

An automatic recognition system for paper currency numbers is developed, Firstly, color image is gathered through CCD camera, then numbers are outputted through process of image processing, orientation, and template matching. In order to improve recognition rate, the system adopt the algorithm of template matching based on extracting of contour and rough classifying by top and bottom quadrature, accordingly, matching calculation was reduced remarkably. The experiment reveals that this method achieves a high recognition rate. The proposed method has been applied to a practical recognition system.