关于手写汉字的切分算法的探讨

汉字识别是模式识别的一个重要分支。识别过程包括光电扫描，模数转换，预处理，模式表达，判别。其中预处理阶段的字切分是整个识别过程的一个重要组成部分，也是难点之一，使如污点的处理，粘连问题，字误切后的重新合并问题等，都给我们的准确切分带来许多麻烦。本文以确定的竖写字贴为处理对象，对手写汉字的字切分做了比较细致的工作，特别就以上提高的问题都是提出了比较实际的解决算法，整体上取得了良好的切分效果。

The Separation Algorith of Handwritten Chinese Characters

Chinese character recognition is an important subdivision of pattern recognition. It comprise photoelectic scanning,modulus conversion,pretreatment,pattern expression and discrimination. The Chinese character separation in the phase of pretreatment is a key point and a hard nut in the whole recognition course. Troubles often impede the correct separation,such as stain-blurring,blending of illegible scribbling,recom position after a wrong separation,etc. In this article the Chinese character separation is dealt with meticulously with the definite vertical-lined calligraphy book as treatment subject. Problems mentioned above are all solved by means of practical algorithms made in the article, achieving good sparation results on the whole.