摘要
为同时保证运动目标检测与跟踪的稳定性与准确性,提出一种基于高斯模型和卡尔曼预测的检测与跟踪方法。在检测中,先采用分块拼接方式初始化背景,再利用动态权值完成高斯背景模型自适应更新,使得目标检测能够持续有效。在跟踪中,Kalman滤波器利用目标检测结果完成预测跟踪,并且对观测噪声矩阵进行自适应更新,使得跟踪的稳定性得到加强。实验结果表明,该算法能够良好地保证其有效性。
In order to guarantee the stability and accuracy of moving target detection and tracking, a algorithm for detection and tracking based on Gaussian model and Kalman prediction is presented. In detection, the proposed method first initializes the background by blocks splicing, and then achieve adaptive background updating by using dynamic weights, so objects detection can be sustained and effective. In tracking, a Kalman filter is used to track the object region centroid by the objects detection re- sult and a adaptive algorithm is presented to adjust the observation noise to strengthen the stability of the tracking. The experi- mental results show that the algorithm can guarantee the effectiveness of moving objects detection and tracking.
出处
《计算机工程与设计》
CSCD
北大核心
2013年第1期247-251,共5页
Computer Engineering and Design
基金
国家自然科学基金项目(11071266)
重庆市教委科研基金项目(KJ100505)
关键词
目标检测
目标跟踪
高斯模型
背景差分
二次帧差
KALMAN预测
object detection
object tracking
Gaussian model
background subtraction
dual subtraction
Kalman prediction
