摘要
设计了一种基于侧面轮廓的实时车型分类系统。该分类系统分为3个部分:侧面轮廓处理、车型模板建立和实时车型匹配。轮廓处理部分介绍了一种用于实时测量移动车辆轮廓数据的采集仪器以及数据处理的方法。车型模板建立部分使用拟合算法构建不同车型的模板。实时车型匹配部分则采用修正豪斯多夫距离作为距离算法,用于衡量待测车辆与模板间的匹配程度。该模式识别系统能够完成准确快速的车型分类。
This paper proposes a real- time vehicle classification system using side profiles of vehicles as characteristics. The system can be roughly divided into three stages： profile data processing,template establishment and real-time vehicle matching. In profile data processing,a profile data collection device is provided for measuring the side profile of moving vehicles and data processing method. We investigate the method to establish template library using fitting algorithm. Thus,the modified Hausdorff distance is proposed to determine the degree of resemblance between two side profiles. This pattern recognition system is capable of classifying vehicle types with high accuracy and low complexity.
出处
《微型机与应用》
2016年第10期50-51,54,共3页
Microcomputer & Its Applications
关键词
车型分类
修正豪斯多夫距离
模式识别
侧面轮廓
vehicle classification
modified Hausdorff distance
pattern recognition
side profile
