摘要
道路交通流参数是交通安全管理、交通状况态势评估和决策的重要基础信息。提出一种改进的具有一定自适应功能的高斯背景建模法,并采用基于透视变换的计算模型,建立车辆的几何尺寸的视觉测量模型,通过CCD传感器感知道路环境并形成视频流记录,利用图像处理和分析方法,对视频流实时处理、跟踪和分析车辆状态参数及道路状况,实现道路交通流基础信息准确实时获取和辨识。实验和现场试验结果表明,平均测量精度达到工程应用需求,可为交通管理部门进行交通管理和交通规划决策提供科学的基础参数信息。
Road traffic flow parameters are important basic information for traffic safety management, traffic situation evaluation and decision-making. This paper proposes an improved Gaussian background modeling method with certain self-adaptability, which uses the calculation model based on perspective transformation and constructs a vision measurement model for vehicle geometric size. It first uses CCD vision sensors to acquire vehicle state parameters and road conditions based on video stream. Then the parameters are processed, tracked and analyzed in real-time with image processing and analysis methods to accurately implement the acquisition and identification of the basic information of road traffic flow parameters in real-time. In situ test results show that the average measurement precision meets the demands of engineering applications, and the system can provide basic parameter information for traffic management department to implement traffic management and transport planning.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2010年第11期2542-2548,共7页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金(60872096)
中央高校基本科研业务费专项(2009B31914)基金资助
关键词
交通流量参数
背景建模
视觉测量
透视变换
机器视觉
traffic flow parameter
background modeling
vision measurement
perspective transformation
machine vision