汉字手写输入识别方法的研究.rar

摘要:在科技发展过程中，智能化成为设备发展的必然方向。智能设备的小型化，成本降低，使得日常生活中智能设备日益增多。

　　通常，智能设备，例如个人电脑和按键手机，使用传统全键盘和T9键盘作为人机输入通道。但伴随着智能设备的小型化、便携化，键盘的使用将不利于智能设备的体积简化。并且，键盘使用人必须对拼音或者五笔输入法能够进行基本掌握，才能进行基本输入操作。

　　在智能设备的体积简化要求下，手写输入必然成为在语音识别还未成熟的情况下，优先选择作为便携设备的输入机制。

　　手写输入在1995年由摩托罗拉公司率先应用之后，已经慢慢在字母类外文方面技术成熟，但是在中文手写输入方面，由于汉字本身结构、数量的特性，加之书写人书写习惯因人而异，造成汉字手写输入在识别方面存在很大困难。

　　考虑到嵌入式设备系统特点，在考虑汉字字库量大的问题时，就要考虑汉字识别速率的问题和正确率的而问题。

　　本文主要阐述压力预处理方法和基于笔段、结构、首末笔划的类识别方案的可行性，主要研究一下问题：

　　1、对压力电阻屏采集压力数据笔划划分法研究，旨在能够实现模仿正常书写时，以用力不同是造成的压力值变化来识别实际笔划和连笔笔划的区分，从而在预处理阶段完成笔划识别。

　　2、研究最小二乘法在拟合笔划时的实用性。

　　3、研究笔划中心点分布对汉字结构的区分实用性。

关键词：汉字；手写输入；类识别； 预处理

Abstract:In the process of the development of science and technology, intelligent equipment development is the inevitable direction. Intelligent device miniaturization, lower costs, making the smart devices in daily life is increasing.

　　　Usually, smart devices, such as mobile phone personal computer and buttons, using traditional qwerty keyboard and T9 keyboard as man-machine input channel. But with intelligent device miniaturization, portable, and the use of the keyboard will not conducive to intelligent devices and simplify. And keyboard users must to basic grasp pinyin or wubi input method can be basic input operation.

　　　Simplify in smart devices volume request, handwriting input must become under the condition of the voice recognition still immature, preference as input mechanism of portable devices.

　　　Handwritten input by MOTOROLA in 1995 after the first application, has mature technology slowly in the class of foreign language, but in terms of Chinese handwritten input, because the Chinese character itself structure, the number of features, combined with the writing people writing habits vary from person to person, result in recognition in handwritten Chinese characters input is very difficult.

　　　Considering the embedded equipment system characteristics, considering the characters ZiKuLiang bigger problem, consider Chinese character recognition rate of the problem and correct the problem.

　　　A pretreatment method, the purpose of this article is to study the main problem:

　　　1）the strokes of pressure resistance screen gathering data classification method research, aimed at to achieve imitation normal writing, with different hard change is caused by the pressure value to identify the actual difference between stroke and cursive stroke, and thereby complete stroke identification in preprocessing stage.

　　　2）study the least square method in fitting strokes practicality.

　　　3）the research to distinguish the practicability of the structure of Chinese characters stroke center of distribution.

Key words: Chinese  handwriting input   preprocessing