基于超声波测距及BP神经网络的车型识别  

Vehicle Identification Based on Ultrasonic Ranging and BP Neural Network

作　　者：李长琦[1] 曲仕茹[1] 

机构地区：[1]西北工业大学自动化学院,陕西西安710068

出　　处：《计算机测量与控制》2009年第9期1804-1806,共3页Computer Measurement &Control

基　　金：陕西省工业攻关项目(2008K07-14)

摘　　要：车型识别在高速公路收费、停车收费、城市道路监控等诸多领域有着广泛的应用。针对传统车型识别系统存在的识别效率低、结构复杂、安装难度大等问题,提出并实现了一种基于超声波测距及BP神经网络的车型识别系统。该系统以嵌入式微处理器、485总线为核心,运用超声波测臣实现对车辆外形尺寸的检测,并从中提取表征车型的四个参数,通过BP神经网络对所获取的参数进行自动分类识别。实验结果表明,该方法车型识别的正确率在95%以上。Automatic Vehicle Identification （AVI） has a wide application in highway toll system, parking charge, surveillance of city road and so on. In order to resolve problems of traditional AVI system such as low efficiency, complex structure, difficult installation, we discussed and realized a vehicle identification system based on ultrasonic ranging and BP neural network. The system is grounded on MCU, 485 BUS, gets vehicle datas and characteristic parameters from the ultrasonic ranging module, and gives the final result after analyzing the four parameter by the BP neural network. Experimentation proves that the accuracy of the system can reach as high as 95%.

关 键 词：超声波测距 485总线 BP神经网络 车型识别 特征参数 

分 类 号：TB553[理学—物理]