本文论述并设计实现了一个脱机自由手写体数字识别系统。文中首先对待识别数字的预处理进行了介绍，包括二值化、平滑滤波、规范化、细化等图像处理方法；其次，探讨了如何提取数字字符的结构特征和笔划特征，并详细地描述了知识库的构造方法；最后采用了以知识库为基础的模板匹配识别方法,并以MATLAB作为编程工具实现了具有友好的图形用户界面的自由手写体数字识别系统。实验结果表明，本方法具有较高的识别率，并具有较好的抗噪性能。-In this paper, and design implementation of a freely Offline handwritten numeral recognition system. First, identify the figure of the pre-processing treatment were introduced, including binarization, smoothing filtering, standardization, refinement, such as image processing method　Secondly, to explore how to extract the structural features of characters and stroke characteristics, and described in detail Knowledge of the construction method　Finally the use of a Knowledge-based template matching to identify methods and MATLAB as a programming tool to achieve with a friendly graphical user interface of freely handwritten numeral recognition system. The experimental results show that this method has higher recognition rate, and has better anti-noise performance.