首页 | 官方网站   微博 | 高级检索  
     

移动机器人的运动目标实时检测与跟踪
引用本文:王璐,崔益安,苏虹,蔡自兴.移动机器人的运动目标实时检测与跟踪[J].计算机工程与应用,2005,41(15):30-33.
作者姓名:王璐  崔益安  苏虹  蔡自兴
作者单位:中南大学信息科学与工程学院智能所长沙410083;中原工学院计算机系郑州450007;中南大学信息科学与工程学院智能所长沙410083
基金项目:国家自然科学基金重点项目(编号:60234030)
摘    要:运动目标检测及跟踪是机器视觉领域备受关注的前沿课题之一。该文针对移动机器人导航领域对检测与跟踪的实时性要求,基于Kalman滤波器实现了驱动单目摄像头的反馈控制系统。采用简单的三帧差背景剪除策略检测运动目标,合并运动估计和背景补偿以加快系统反应速度。系统误差保存在协方差阵中,以增益的形式参与控制。该文还详细分析了运动补偿对检测的影响及误差的变化趋势。实验表明,系统能够保持对运动目标稳定偏差的平滑跟踪,在480320的图像分辨率下控制速度达到20Hz(fps)。

关 键 词:运动目标检测及跟踪  实时  Kalman  滤波器  背景补偿
文章编号:1002-8331-(2005)15-0030-04

Real-time Moving Object Detection and Tracking of Mobile Robots
Wang Lu,Cui Yian,Su Hong,Cai Zixing.Real-time Moving Object Detection and Tracking of Mobile Robots[J].Computer Engineering and Applications,2005,41(15):30-33.
Authors:Wang Lu  Cui Yian  Su Hong  Cai Zixing
Affiliation:Wang Lu1,2 Cui Yian1 Su Hong1 Cai Zixing11
Abstract:Moving object detection and tracking is one of advanced problems and has attracted much more attention.This paper presents an implementation of the feedback control system to drive the camera based on Kalman filter for satisfying real-time requirement of mobile robots navigation.A simple strategy named 3 frames difference background subtraction is adopted to detect moving object,and movement estimation and compensation are combined to improve processing speed.System errors are saved in a covariance matrix and are attached to control system as increment.This paper also analyzes in detail the influence of movement compensation on detection and the trend of errors.Experiments indicate that this system can keep tracking moving object smoothly with fixed deflection and the control rate can be 20Hz with the resolution of 480320.
Keywords:moving object detection and tracking  real-time  Kalman filter  background compensation
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司    京ICP备09084417号-23

京公网安备 11010802026262号