基于图像处理的道路车辆信息提取与识别算法的研究与实现
发布时间：2019-03-08 10:52
【摘要】：智能交通系统(ITS)是将先进的科学技术相融合并有效地运用于在交通管理、交通信息服务和车辆控制等方面的综合交通管理系统。它增强了车辆和道路与使用者之间的联系,是一种大规模、全方位、实时、准确、高效的综合智能交通管理系统,对交通管理和控制有着重要的意义和作用。本论文设计的总体目标是通过图像处理算法提取车队的纵深长度或车辆的个数,并将其作为系统控制的参数来控制调整交通信号显示方式。全部的图像处理算法程序将移植到以ARM为核心的交通控制板上,从而实现一种自适应的节点智能交通控制方法。本课题主要研究一种简单有效的交通图像采集和处理算法,通过图像处理算法提取车队排队信息。在算法上,首先选择可以突显车辆并去除大量噪声的预处理方法,选择简单、快捷的背景差分法作为提取车辆的方法。在背景差分法中增添背景更新模型和分割算法,并对分割算法进行了改进,使算法运行效率更高。考虑到实时性的问题,提出了车道提取方法,从而提高了系统的实时性。对于车队长度和车辆个数的统计,建立了行坐标模型,实现了图像坐标系和实际坐标系的距离转换,从而估算出车队长度,为后续硬件系统实现提供了重要参数
[Abstract]:Intelligent Transportation system (ITS) is an integrated traffic management system, which integrates advanced science and technology and is effectively used in traffic management, traffic information service and vehicle control. It enhances the connection between vehicles, roads and users. It is a large-scale, omni-directional, real-time, accurate and efficient integrated intelligent traffic management system, which is of great significance and effect to traffic management and control. The overall goal of this paper is to extract the longitudinal length or the number of vehicles by image processing algorithm, and use it as the parameters of system control to control and adjust the display mode of traffic signals. All the image processing algorithm programs will be transplanted to the traffic control board with ARM as the core so as to realize an adaptive node intelligent traffic control method. In this paper, a simple and effective traffic image acquisition and processing algorithm is studied, and the queue information is extracted by image processing algorithm. In the algorithm, firstly, the pretreatment method which can highlight the vehicle and remove a lot of noise is selected, and the simple and fast background difference method is chosen as the method to extract the vehicle. The background update model and segmentation algorithm are added to the background difference method, and the segmentation algorithm is improved to make the algorithm run more efficiently. Considering the real-time problem, the lane extraction method is proposed to improve the real-time performance of the system. For the statistics of the fleet length and the number of vehicles, the row coordinate model is established, and the distance conversion between the image coordinate system and the actual coordinate system is realized. The length of the vehicle fleet is estimated, which provides an important parameter for the follow-up hardware system realization.
【学位授予单位】：内蒙古大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP391.41;U495
本文编号：2436731
[Abstract]:Intelligent Transportation system (ITS) is an integrated traffic management system, which integrates advanced science and technology and is effectively used in traffic management, traffic information service and vehicle control. It enhances the connection between vehicles, roads and users. It is a large-scale, omni-directional, real-time, accurate and efficient integrated intelligent traffic management system, which is of great significance and effect to traffic management and control. The overall goal of this paper is to extract the longitudinal length or the number of vehicles by image processing algorithm, and use it as the parameters of system control to control and adjust the display mode of traffic signals. All the image processing algorithm programs will be transplanted to the traffic control board with ARM as the core so as to realize an adaptive node intelligent traffic control method. In this paper, a simple and effective traffic image acquisition and processing algorithm is studied, and the queue information is extracted by image processing algorithm. In the algorithm, firstly, the pretreatment method which can highlight the vehicle and remove a lot of noise is selected, and the simple and fast background difference method is chosen as the method to extract the vehicle. The background update model and segmentation algorithm are added to the background difference method, and the segmentation algorithm is improved to make the algorithm run more efficiently. Considering the real-time problem, the lane extraction method is proposed to improve the real-time performance of the system. For the statistics of the fleet length and the number of vehicles, the row coordinate model is established, and the distance conversion between the image coordinate system and the actual coordinate system is realized. The length of the vehicle fleet is estimated, which provides an important parameter for the follow-up hardware system realization.
【学位授予单位】：内蒙古大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP391.41;U495
【参考文献】
相关期刊论文 前1条
1 靳涛;张红星;;基于动态图像识别的智能交通灯控制[J];科技传播;2012年20期
，本文编号：2436731
本文链接：https://www.wllwen.com/kejilunwen/daoluqiaoliang/2436731.html
