基于灰度跳变的车牌定位算法及其实现

摘要

Location algorithm of the license plate recognition was studied. The recognition accuracy of traditional license plate location algorithm can not meet the requirement of real traffic management and monitoring, and the li cense plate image positioning accuracy is not high. Based on image preprocessing, the paper presented a transition based on properties of gray vehicle license location algorithm. It mainly included two location algorithms, the rough location and fine positioning. VC++ was used to design and implement the localization algorithm. The processing ex periment of the license plate images of the actual traffic used the proposed method to locate the license plates in differ ent images. Experimental results show that the algorithm can carry out high-precision positioning of license plate im age and the algorithm has strong robustness for the license plate character recognition and lays the basis for segmenta tion.%研究车牌识别定位算法问题.传统的车牌设识别定位算法的识别精确度难以满足现实在交通管理和流量监测中的应用,存在车牌图像定位的精度的不高等问题.为解决上述问题,在对车辆图像预处理基础上,提出了一种基于图像灰度跳变特性的车辆牌照定位算法.主要给出了粗定位和细定位两种定位算法,并用VC++设计实现定位算法.通过实验对实际交通中多幅车牌图像进行处理,对不同的车牌图像进行定位.实验结果表明,新算法能够对车牌图像进行高精度定位,并且具有较强的鲁棒性,为车牌字符的识别与分割奠定了一定的基础.