摘要： 字符识别是计算机模式识别中较为常见的问题。对于印刷体的字符来说，经常采用二值化和模板匹配方法进行识别。在实际应用中，为了提高识别率，减少计算量和运行时间，解决正确识别二值化后的正反不同图象的问题，提出一种改进的自适应的模板匹配方法，理论与实验结果表明具有较高的识别率和坚韧性。

关键词: 字符识别, 模板匹配, 二值化

Abstract: Character identification is an issue that computer pattern recognition often meets. We usually use binarization and template matching method for the printed character. In order to improve recognition-rate, decrease computation and operation time and solve the problem there has to match two kinds of reverse image after binarization in application, we describe a modified adaptive template matching method in this paper. The modified method has been proved and showed it has better recognition-rate and robustness.

中图分类号: 

• TP391.41