摘要
分析了车辆的大小、形状和边缘特征,将粒子群算法和云模型理论具体运用到车型识别问题中.通过粒子群编码,设计适应值函数,然后训练样本,通过5类车型的云模型参数和属性相似度计算各特征的在分类中的权值,然后通过单分类器和多分类器进行车型识别,车型分类效果较好.
Vehicle classes are recognized according to Particle Swarm Optimization and cloud mode theory .Vehicle features ,such as vehicle size ,shape information and edge feature are extracted for PSO training .Through cloud model parameters and the attribute similarity of five kind of vehicle calculate various characteristics weight ,then carries vehicle recognition by the single sorter and the multi-sorters ,the vehicle classification effect is good .
出处
《微电子学与计算机》
CSCD
北大核心
2013年第11期80-83,共4页
Microelectronics & Computer
基金
江苏省教育厅项目(2012SJD870001)
江苏省科技型企业创新项目(BC2011441)
关键词
车型识别
粒子群算法
云模型
vehicle recognition
particle swarm optimization
cloud mode
