摘要
在应用摄像头进行目标跟踪的过程中,由于图像信号的采集、传输和处理时延的影响,使得目标不能处在摄像头的最佳观测位置,从而产生数据缺失,造成分析结果滞后,由此可能导致摄像头运动控制误差较大.本文提出基于卡尔曼滤波的摄像头预测跟踪模型,充分利用Kalman滤波方程递推预估计能力对运动目标位置进行跟踪,及时调整摄像头偏转角度,使得摄像头始终超前运动目标,解决了由于摄像头运动惯性产生的数据缺失现象,对后续的图像分析与识别提供了保证,最后仿真结果显示了该模型的正确性.
The pickup camera is usecally lagged behind the moving object while pickup camera is used to trace an object. This phenomenon causes the lose of useful image data and leadsto the error in image analysis and identification. Aim at this problem,this paper brings out a track model based on Kalman filter equation to predict the moving object location. The prediction and estimation are used to adjust the movement of pickup camera and make pickup camera trace the object ahead. The way has sloveed image infor mation lose because of pickup camera lagging behind object. The result of simulation proves the correction of the model.
出处
《西安工业学院学报》
2006年第1期1-4,共4页
Journal of Xi'an Institute of Technology
关键词
卡尔曼滤波
摄像头
目标跟踪
二维空间
kalman filter
pickup camera
tracing object
two-dimensional space
