摘要
提出了一种基于激光雷达的车辆跟踪与识别方法.该方法采用Kalman滤波实现目标跟踪,运用坐标变换法消除不同数据帧中物体形状差异对跟踪中心的影响,结合车辆矩形投影及速度特征识别车辆.针对由于遮挡或某部分反射率低而引起的目标分割现象,提出了一种结合车辆外形与轮廓特征的聚类合并算法.仿真实验表明,该算法精度高、实时性好且鲁棒性强.
A method for vehicle tracking and recognition based on scanning laser radar was presented. Kalman filter is utilized to track objects, and a method of coordinate transformation to eliminate the impact of shape change on tracking center. Vehicle's rectangle projection along with velocity feature is introduced to recognize vehicle. Since some object will be separated into multiple parts due to occlusions or bad reflectivity, a cluster fusion algorithm according to the shape and contour characteristics of vehicle was presented. The experiment shows this algorithm has the advantage of high precision, fast processing and strong robustness.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2009年第6期923-926,共4页
Journal of Shanghai Jiaotong University
基金
上海科委登山计划资助项目(062107035)
上海科委浦江人才计划资助项目(07pj14055)
关键词
激光雷达
车辆识别
目标跟踪
卡尔曼滤波
scanning laser radar
vehicle recognition
object tracking
Kalman filter