运用仿生模式识别方法构建提取基本笔段的神经元序列覆盖手写体汉字图像, 分析笔段神经元间的拓扑性质, 将手写体汉字图像转化为具有容错表征方式的种汉字笔划类型组成的几何图形模仿人类汉字形码输人法统计具有冗余容错形状的笔划神经元类型、数量、位置、相合和相交点数量, 建立手写体汉字特征知识的数据结构表对一手写体汉字库中手写体汉字识别进行仿真实验。方法具有较强的“ 认知”手写体汉字的能力-Construction of the use of pattern recognition methods of extracting essential Strokes neurons handwritten Chinese character image sequence coverage, analysis Strokes topological properties of neurons, the handwritten Chinese character image into a fault-tolerant species Representation type of character strokes mimic the human form of geometry Statistics character image code input method has the shape of the strokes Redundancy neuron type, quantity, location, consistency, and the intersection points, and build knowledge of handwritten Chinese character feature table data structure of a handwritten Chinese character recognition of handwritten Chinese characters library of simulation experiments. The results proved that this method has a strong "cognitive" ability of handwritten Chinese characters.