【摘要】 车型识别已成为智能交通系统研究中的一个重要内容。根据同类车型尺寸特征如长、宽、高特征值都具有相似性特点,利用已有车型数据库,分别使用逆传播(BP)神经网络、支持向量机(SVM)网络、有监督Kohonen网络这三种神经网络分类方法对车型尺寸特征进行分类,得到三个车型识别准确率并进行比较验证。实验结果表明:有监督Kohonen网络的分类方法具有较高的车型识别精度,实验效果明显,车型识别准确率高。更多还原

【Abstract】 Vehicle type identification has become an important research content in intelligent transportation systems. According to the characteristics of size similar vehicle models such as long,width,height features value are similar characteristics,using the existing vehicle models database,respectively using three neural network classification methods,such as BP neural network,SVM network,supervised Kohonen network,to classify features of vehicle type size,obtain identification accuracy rate of three vehicle type and comparison and verification are carried out. Experimental results show that the classification method of supervised Kohonen network has high precision of vehicle identification,the experimental effect is obvious,and the accuracy rate of vehicle type identification is high.更多还原