节点文献
复杂背景下的汽车牌照自动识别系统
Vehicle License Plate Recognition System Under Complex Background
【摘要】 分析了目前存在的车牌自动识别系统及其相应的问题,提出了一种在多分辨率下串行运算的金字塔结构的车牌自动识别系统。该系统由两个重要步骤组成:1)基于灰度方差的车牌定位算法,2)结合连通域分析和投影分析的字符切分算法。通过大量的实验证明该方法准确率高,鲁棒性好,速度快。
【Abstract】 In this paper, some existing automatic vehicle license plate recognition (LPR) systems and corresponding problems are discussed first. Then a pyramid architecture processing multi-resolution images serially is introduced to our LPR system. The proposed method is divided into two steps: 1) license plate location by the variance of gray image, and 2) license plate segmentation by connected components analysis and projection analysis. A set of experiments has been performed to prove the robustness, accuracy and speed.
【关键词】 金字塔结构;
车牌定位;
连通域分析;
投影分析;
模板匹配;
【Key words】 pyramid architecture; license plate location; connected components analysis; projection analysis; template match;
【Key words】 pyramid architecture; license plate location; connected components analysis; projection analysis; template match;
【基金】 国家自然科学基金项目(编号60472002)
- 【文献出处】 电视技术 ,TV Engineering , 编辑部邮箱 ,2005年06期
- 【分类号】TP391.4
- 【被引频次】16
- 【下载频次】420