中文摘要：

文中提出了一种基于结构特征的识别方法,该方法通过去模糊和图像清晰化及目标框定进行预处理,然后通过定位切分和多次分层以及对模糊字符＂穿孔＂进行特征提取.选取具有描述字符特性能力强、字符间区分度高的特征,通过特征比对,进行钢坯端面字符的实时自动识别,由此降低计算复杂性.实验证明,该方法可有效的降低字符的拒识率和误识率,提高了识别率.

英文摘要：

For the current methods dissatisfied with practical demands,an identification method using features with good description and discrimination of character structures is presented in this paper.Firstly the image is preprocessed by deblurring and object framing,the features are extracted using multi-layering and artificial hole after character location and segmentation.The features with high performance are extracted.The steel ends characters real-time recognition is performed using features contrast.The experiments show that this method can reduce the complexity greatly,lower error rate and rejection rate effectively,and enhance the recognition rate.