摘要： 提出一种基于视频跟踪和有限状态自动机的运动车辆行为表达与分析方法。采用减背景法得到前景运动车辆，基于快速归一化互相关理论，通过预测实现车辆跟踪，得到准确的车辆运动轨迹。利用有限状态自动机，将车辆的行为表达为连续的微观行为状态，从而在运动跟踪的基础上，结合时域与空域信息分析车辆行为模式。对交叉路口的运动车辆进行跟踪实验，结果表明，该方法能够准确得到车辆的状态信息。

关键词: 智能交通系统, 快速归一化互相关, 加和表, 有限状态自动机, 行为模式

Abstract: This paper proposes a method of expressing and analyzing behavior in moving vehicles based on video tracking and Finite State Automata(FSA). It gets foreground moving vehicles by using background subtraction method. In order to obtain accurate motion trail of vehicles, it predicts to achieve vehicle tracking based on Fast Normalized Cross-correlation(FNCC) theory. The state of vehicle behavior is transformed to that of continuous microscopic behavior by FSA. By combining time domain with spatial domain, it succeeds to analyze behavior of vehicles on the basis of motion tracking. Experimental results show that the method is able to gain status information of vehicles accurately.

Key words: Intelligent Transportation System(ITS), Fast Normalized Cross-correlation(FNCC), sum-table, Finite State Automata(FSA), behavior pattern

中图分类号: 

• TP391.41