摘要
针对车牌字符光照和污损的复杂特征,对车牌图象的二值化算法进行了深入的分析和研究,提出了一种基于差分边缘检测算法的改进图象二值化算法.实验结果表明,改进算法能够有效地将字符清晰地从背景中分割出来,分割速度快并且有助于提高字符识别的正确率.
According to the characteristic of license plate with complicated-background image, a new binarization algorithm based on improved LLT is presented here. Compared with other binarization methods, the result shows that this algorithm can make the character segmented from the background distinctly, effectively approve the speed and accuracy of vehicle license plate recognition.
出处
《四川大学学报(自然科学版)》
CAS
CSCD
北大核心
2006年第6期1253-1258,共6页
Journal of Sichuan University(Natural Science Edition)
关键词
图象分割
阈值
二值化
差分边缘检测
image segmentation
threshold
binarization
difference edge deetion