大量目标识别的光学相关与神经网络融合方法

Fusion of OCPR and neural network for recognition of large number targets

本文基于光学相关与神经网络学习算法和编码聚类原理的融合 ,提出一种大数量目标的高效率分类识别方法 ,并以 67个字符 ,其中包括2 6个英文字母、1 0个数字和 31个省和直辖市的简称汉字的分类识别作为实例 ,给出详细的原理说明和结果。

An effective method of recognition of targets with large number is proposed based on fusion of optical correlation pattern recognition(OCPR) and neural network learning algorithm and clustering encoding.Detailed principle and recognition results on 67 characters,including 26 English capitals,10 digits and 31 Chinese characters which are the abbreviations of provinces and municipalities,as a case study,are presented in this paper.