摘要 高速公路出入口是常见的事故多发区域。该区域除了直行交通外，还存在转弯交通进行加减速、变换车道、分合流等复杂交通行为，容易引起交通混乱。本文从车辆的微观驾驶行为着手，提出了行车风险的概念并对各种风险进行量化，借助图像处理技术，建立了基于行车风险的高速公路出入口区域安全评价方法。 本文首先提出了三种行车风险指数的计算方法，包括跟驰风险指数、变换车道风险指数、以及高速行车风险指数。用车辆的实时车速以及车距等信息中提取出后车临界加速度，结合向后加速度对驾驶员的心理和生理反应，建立行车风险指数。然后提出了基于图像处理的交通信息采集方法，通过图像识别、跟踪、简化、网格转换等手段，获取车辆的实时的坐标，从而得到车辆的车速和车距等重要信息。最后，本文建立了行车风险的计算流程，介绍该方法的应用，以平均风险水平作为评价出入口区域安全性的标准，并用实例进行了说明，对该评价方法在不良天气条件下的应用有待修正。 基于行车风险的安全评价方法是以车辆的运行特征为基础的，在缺乏道路事故资料的情况下，它能实现安全评价的目的，是对传统安全评价方法的一种提高。 关键词：高速公路出入口，安全评价，行车风险，图像处理。 ABSTRACT Freeway on-ramp and off-ramp areas are general accident-prone sections. In those areas, vehicles frequently accelerate and decelerate, change lanes and merge from different directions and locations, so it is used to be in disordered conditions. Regarding to vehicle’s behavior, this paper presents driving risk as a surrogate measure to describe the level of safety, and quantitates three different driving risks. Then, using image-processing technique, a safety appraisal method based on driving risks on freeway on ramp and off-ramp areas is advanced. First, this paper defines three different risks and quantitates those risks, the three risks include car-following risk, lane-changing risk, and high-speed risks. Critical acceleration can be calculated with the real-time speed and the spaces between vehicles, which is a rule to quantitate driving risk. Then, this paper present traffic information gathering method based on image-processing technique, this method can discern, track and simplify moving vehicle in an image. Through grid conversion, real-time coordinate of vehicle can be obtained and also real-time speed and spaces between vehicles. At last, the flow to calculate driving risk are presented, and average risk level is used as a criterion to evaluate the safety of freeway on-ramp and off-ramp areas. An example is given to introduce this method. The safety appraisal method use driving risk not traffic accident so it’s a certain kind of improvement, which can be used to evaluate safety of freeway on-ramp and off-ramp areas when it is short of traffic accident data. |
硕士论文