摘要
对车牌识别中的背景噪声、边框和铆钉干扰、字符内噪声等提出了行之有效的去除方法。背景噪声去除方法包含了颜色增强、杂色抑制的思想,通过简单的投影分析即可以消除绝大部分的噪声。介绍了利用边框、铆钉与字符的跳变特征和汉字字符自身特征去除边框和铆钉的方法,该方法去噪高效快捷,误切率极底,正确去除率可达98%以上。
Some valuable approach to eliminate the background noise , the influence caused by frame of license plate & its rivets ,and the noise inside the characters are put forward.The approach to eliminate the background noise contains the idea of color enhancement & noise constraining.It can easily eliminate most noise in background.The approach, which uses the transition information that the frames and the rivets have and the information that the Chinese characters have to eliminate the frames & the rivets, is also introduced.The experimental result shows that it can quickly and effectively eliminate the noise in low error rate and can be put into practical use.The proper noise eliminating rate is above 98%.
出处
《计算机工程》
CAS
CSCD
北大核心
2004年第24期112-115,共4页
Computer Engineering
关键词
噪声抑制
跳变
连通区域
Noise constrain
Binary transition
Connected components