摘要
为了提高车牌定位效率,提出一种混合简化脉冲耦合神经网络(PCNN)和快速连通域标记的车牌定位算法。基于简化PCNN进行图像增强,利用车牌字符的连通域特征、纹理特征和结构特征对增强后的二值图像进行过滤、筛选,得到图像中大致车牌区域,再对所得区域左边界起始的左扩展区域做垂直投影,确定车牌中汉字区域,从而定位车牌。实验结果表明,该算法性能优于其他车牌定位算法,其定位准确率为97.5%。
This paper proposes a hybrid method of simplified Pulse Coupled Neural Network（PCNN） and quick region identification for localizing car plate, and the algorithm can locate the car plate precisely and segment the characters simultaneously. The simplified PCNN is used to enhance the car plate image. Based on the car plate characters features of the region, texture and structure, the enhanced binary image is filtered to get the rough location of the car plate region, which is located precisely on the borders except the region of Chinese character. Afterwards, it expands the rough region to the left and fixes on the Chinese character region from the vertical projection of the expansion area, and the ear plate region is located Experimental results show that the precision can get 97.5%, which is higher than other localization methods on the same image database.
出处
《计算机工程》
CAS
CSCD
北大核心
2010年第24期178-179,182,共3页
Computer Engineering
基金
广东省工业重点攻关基金资助项目(2004B10101032)
关键词
简化脉冲耦合神经网络
车牌图像增强
车牌定位
连通域标记
simplified Pulse Coupled Neural Network（PCNN）
car plate image enhancement
car plate localization
connected region identification
