摘要
对交叉路口行驶车辆进行有效分类,是进行流量统计、交通状况分析,以及交通事故分析的基础。在运动目标检测和图像标定技术的基础上,通过引入多类支持矢量机方法,提出了一种利用车辆的多个特征对交叉路口车辆进行分类的方法,解决了传统分类方法存在的"误分"问题,实验结果表明,该方法分类精确度高,满足交叉路口车型分类的要求。另外,又给出了一种"车流速度"的提取方法,为交通管理部门分析交通状况提供了依据。
To classify correctly the tuning vehicles in the crossing is the basis of traffic flow statistics,traffic situation analysis, and accident analysis. Based on the detection for running objects and the calibration for representation, we draw into the multi-support vector machine and propose a method to classify the vehicles in crossing, which needs several features and can overcome "wrong-classification'. The test shows that the method has high accuracy and can meet the requirements to classify the vehicles at cross-roads. Furthermore,based on the former study,we also offer a method to obtain" flowing speed" and provide a basis for the traffic management.
出处
《中国图象图形学报》
CSCD
北大核心
2008年第4期801-807,共7页
Journal of Image and Graphics
基金
江苏省高校自然科学计划项目(06KJD520037)
关键词
车型分类
多类支持矢量机
图像标定
特征提取
车流速度
vehicle classification, multi-class support vector machine, representation calibration, feature extracting,flowing speed