基于BP人工神经网络的数字字符识别

Number character recognition based on BP artificial neural network

提出一种基于BP人工神经网络来识别含有噪声的数字字符的方法.通过对BP网络模型进行训练,并选择纯净和带有噪声的数字字符组成样本,借助Matlab7.0实现数字字符识别,识别率为96%,达到了设计要求.

Based on BP artificial neural network, a method for recognition of the number characters containing noise was proposed . Trained with pure samples and noise samples of the number characters, the digital characters were recognied, using with Matlab 7.0 recognition rate reaches 96% ,which meets the design requirement.