基于DSP的车型检测器设计及其算法的研究

车型分类是高速公路自动收费和交通流量统计的重要依据,它是智能交通(ITS)的一个重要组成部分。本文针对车型检测器硬件结构和处理算法两方面提出一些具有创新性的设计方案,介绍了由LC振荡电路和TMS320F2812处理芯片构成的车型检测器的硬件结构,为克服LC振荡电路频率不稳定的固有缺陷提出了基频更新算法。并提出了利用一维数学形态滤波方法过滤实际应用中的噪声信号的方法,最后简单介绍了基于粗糙集BP神经网络车型分类算法。

Vehicle Detector Design Based on TMS320F2812 and Its Algorithms Research

Vehicle classification is an important criterion for highway automatic toll collection and traffic flow statistic. And it' s an important part of ITS as well. This paper brings forward some innovative ideas about hardware design and algorithms on vehicle de-tector. In this paper, the design of vehicle detector based on TMS320F2812 and LC oscillation circuit is proposed. To solve the fre-quency instability of LC oscillation circuit, a fundamental frequency update algorithm is introduced. And a one-dimension morphologi-cal filter method is applied into filtering the noise signal in practice. Besides, this paper is introduced a BP neural network vehicle classification approach based on rough set.