基于模板匹配的手写字符识别技术实现

在传统的图像处理程序中，由于没有标准的API，往往会造成编码量大，效率又不高的结果．而OpenCV提供了统一的API，为后续的图像处理提供了极大的便利，在这里首先介绍了在VC＋＋6．0下安装、配置OpenCV（这里主要指的是OpenCV1．0版本）环境，以及建立OpenCV下的VC＋＋6．0应用程序框架．在模板数据库建立的过程中，首先将模板图片进行二值化，将特征值提取出来，存入模板库，从而完成模板数据库建立，再根据模板数据库对目标图片进行比对的过程中，将该目标图片的特征值与模板数据库中的所有特征值进行比对并计算相似度，进而针对相同字符的不同模板计算该相似度的期望和方差，根据计算所得的期望和方差得到最匹配的字符，即最终结果．

Design of Handwritten Character Recognition Based on Template Matching

In conventional image processing programs,the code compilation is tedious and inefficient because of lacking the standard API.As a consequent,OpenCV provides us with API.It is convenient for the following image processing.There is an introduction of installing and building OpenCV compilation environment based on VC++6.0 and of the framework establishment of OpenCV application program.During the process of building the template,the image should be preprocessed firstly,then the algorithm need extract characteristics and put the them into the template database.And then with comparing the character of the object image with the template,there is a result based on the index of the similarity.