基于粒子滤波框架Catmull-Rom样条的多车道侦测及跟踪算法研究  

Study of multi-lane detection and tracking algorithm based on Catmull-Rom spline and particle filter framework

作　　者：陈孟元[1] 

机构地区：[1]安徽工程大学安徽省电气传动与控制重点实验室,芜湖241000

出　　处：《电子测量与仪器学报》2015年第10期1520-1528,共9页Journal of Electronic Measurement and Instrumentation

基　　金：安徽省高校自然科学研究重大(KJ2014ZD04)项目

摘　　要：车道线识别是智能交通系统(intelligent transport system,ITS)实现的重要组成部分。为了解决道路图像中光源干扰、阴影遮挡、车道线破损、标志线不连续等因素导致道路识别方法鲁棒性较差的问题,提出一种基于粒子滤波框架的Catmull-Rom样条的车道线侦测及跟踪算法,这种算法利用Catmull-Rom样条数学模型能够灵活且精确的标定车道线,且基于粒子滤波框架的滤波器保障了系统的稳定性及鲁棒性。实地测试结果显示在9种路况环境下算法侦测及跟踪性能准确稳定;该方法与近远景分割算法比较,路况侦测及跟踪边界更清晰;进行基于影像序列的量化稳定性评估,正确识别率接近98%。Lane detection is an important part of the intelligent transportation system.In order to solve the problem that light interference,shadow,worn out lane resulting in the robustness of lane detection bad,even worse,a spline-based multi-lane detection and tracking system is proposed in this paper.The major novelty of the proposed approach is the usage of the so-called Catmull-Rom in combination with the particle filtering framework.The new spline-based model enables an accurate and flexible modeling of the lane markings.At the same time the application of the filter based on particle filtering framework contributes significantly to the system robustness and stability.The test results under nine kinds of road environment show the stability and accuracy of the detection and tracking algorithm.Compared with so-called close-range and vision algorithm,the algorithm proposed in the paper detects boundary more clearly and gets nearly 98%correct identification rate in stability assessment based on video sequence.

关 键 词：弯道侦测 车道识别 样条曲线 多车道侦测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]