【作者】 李盛丹；

【导师】 王双维；

【作者基本信息】 东北师范大学 ， 电路与系统， 2009， 硕士

【摘要】 随着社会经济发展,交通系统日益复杂,为了给交通系统的管理提供各种实时交通信息,以方便、高效的利用和管理现有的交通系统,我们研究开发了车型分类及车辆计数系统。本系统的核心算法是基于汽车噪音信号,具体实现思路是:首先在线采集车辆声音信号,通过端点检测截取出有效部分,再对截取后的声音采用Brug法提取其AR模型参数,然后把参数送进训练好的BP神经网络进行车型分类并统计出每种车型的数量。整个系统采用VC++和MATLAB混合编程的方法实现。由于本系统的核心算法是由MATLAB实现,所以便于系统算法的更新。相对于传统的车型分类方法,本系统具有成本低、开发时间短、易于实现等优点。本系统目前适用于单车道、小流量的公路,能够较准确的识别出公路上的基本车型。更多还原

【Abstract】 With the development of society and economy, the traffic system became more and more complex. In the interest of providing the management of traffic system with real-time traffic information to achieve efficient management and make full use of the traffic system. We investigate a vehicle classification and counting system.In this system we have designed a core algorithm based on vehicles noise, the concrete realization thought is: firstly, we on-line collect the vehicle noise, intercept the effective part of the noise by using endpoint detect method, and then extracting model parameter from AR spectrum analysis with Burg method, sending the parameter to the trained BP Neural network for classification, at last, calculating out the number of each vehicle model. The software system was constructed by VC++ and MATLAB. Because the core algorithm of the system was completed with MATLAB, it is facile to update the algorithm. Compared with the traditional vehicle classification method, this system has the advantages of shorter developing cycle, lower cost and easier to achieve.At present, the system is suitable for single-lane and small flow of highway, can be more accurate to identify the basic models on the road.更多还原