摘要
字符识别技术是用机器视觉进行模式识别的重要研究方向之一。烟草制造的工业现场环境恶劣,会引起所拍摄的图像失真、模糊、含噪声及字符产生缩放、平移旋转等情况。针对以上情况,首先对图像进行了前期预处理,并对字符特征进行了提取;之后,针对两种字符特征,用改进的BP算法设计了两个神经网络识别系统。
Character recognition technique is one of the most important research directions of identifying model by machine vision.For the bad circumstance in industrial field of tobacco manufacturing,the captured image may be distorted and fuzzy,and created characters with zoom,rotatation and so on.Considering the problems above,we preprocessed the image at first,then extracted the features of characters.In the end,by use of the improved BP arithmetic,two neural network recognition systems based on two kinds of character features were designed.
出处
《机械工程与自动化》
2010年第3期117-119,共3页
Mechanical Engineering & Automation
关键词
字符识别
二值化
BP网络
character recognition
binarization
BP network
