摘要
车牌定位是车牌识别(LPR)系统中的关键技术之一,通过对车牌区域的各种固有特征的研究和探讨,提出了一种高效快速的车牌字符定位算法,该算法基于二值边缘图像的频率变化特征,在整个车辆图像中搜索到车牌的精确位置,得到牌照图像,为车牌的进一步识别打下良好的基础.现场实验表明,该算法快速准确,对提高整个识别系统的正确率起到了关键的作用.
License plate location is one of the most significant subjects in automatic license plate recognition. The intrinsic characteristics of license plates are studied and a quick and efficient license plate location algorithm is presented. The algorithm can search the accurate position of car plate from the whole vehicle picture based on the changing characters of binary edge detection and thus lay a good foundation for the next step of recognition. Experimental results show that this algorithm is quick and accurate. In this way, the veracity of the identifying system has been improved dramatically.
出处
《甘肃科学学报》
2006年第4期88-91,共4页
Journal of Gansu Sciences
基金
甘肃省科技攻关计划项目(ZGS047-A52-004-01)
关键词
车辆识别
边缘检测
二值化
先验知识
license plate recognition
edge detection
binary processing
prior knowledge
