用图像处理技术提取交通车辆移动轨迹  

Extracting the Trace of Moving Vehicles Using Imaging Processing Techniques

作　　者：张建华[1] 杨建国[1] 尹旭全[1] 肖永剑[1] 

机构地区：[1]西安交通大学电气工程学院,西安710049

出　　处：《计算机工程与应用》2004年第21期213-215,225,共4页Computer Engineering and Applications

摘　　要：交通仿真的不断发展要求能为其提供现实中的车辆运动过程,从而使得模拟更加逼真,同时交通仿真模型评估也需要用到现实的车辆运动轨迹来比对仿真模型模拟出来的车辆运动轨迹。该文提出了一种用图像处理技术提取交通车辆移动轨迹的方法。该文介绍的这一方法采用了动态目标区域检测的技术来追踪运动车辆,所谓动态目标区域检测就是指被检测区域是动态更新的,是随着被检测对象位置的变化而改变的。这样一种思路有效地提高了系统运行的速度和系统检测的精度。避免了采用全局扫描思路所引起的系统复杂、运算量大以及精度不高的缺点。同时又有别于静态的重点区域检测的思路,使得系统更加灵活,运算更加快捷。通过实验验证,该系统可以快速准确地提取交通车辆的移动轨迹。Along with the development of traffic simulation,it requires that we can afford the real moving processes of vehicles.So that traffic simulation systems can realistically model traffic systems and at the same time it helps to validate the results of traffic simulation systems.This paper introduces an algorithm of extracting the trace of moving vehicles by image processing techniques.The approach proposed by the writer adopts the technique of dynamic target region detection.Its core is that the region to be detected is dynamic and is changing with the target's change.This algorithm effectively improves the velocity and the precision of the system.It avoids some disadvantages,such as,complex frames,enormous data and lower precision,which caused by the process based on full frame processing.At the same time this algorithm is different from the approach that processes based on the key region processing.So the system introduced by this paper is more flexible and the operation is faster.The proposed algorithm has been implemented and it can accurately and quickly extract the trace of moving vehicle in the traffic system.

关 键 词：图像处理 标记 轨迹 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]