摘要
该文针对智能监控的需求,提出基于特征的多运动目标跟踪算法。通过自适应Marr小波核函数背景建模算法,在冗余离散小波域进行多运动目标识别。运动跟踪采用SIFT特征粒子滤波算法,并采用队列链表法记录多运动目标之间的数据关联,在提高识别准确率的同时降低了运算的复杂度。实际测试表明,该算法对于多运动目标识别跟踪具有更优越的实时性和抗遮挡性,在智能监控领域具有较广泛的应用前景。
For the widely demanding of adaptive multiple moving targets tracking, a type of feature based multi-target tracking algorithm is presented. Background is adaptively modeled by Marr wavelet kernel function and a background subtraction technique based on redundant discrete wavelet transforms is used to detect multiple moving targets. A type of particle filtering combined with SIFT key points is used for tracking, and a queue chain method is used to record data association among different targets, which can improve the detection accuracy and reduce the complexity. Actual tests show that the algorithm can track multi-target with a better performance of real time and mutual occlusion robustness; it can be used in intelligent monitoring with extensive application prospect.
出处
《电子与信息学报》
EI
CSCD
北大核心
2010年第5期1111-1115,共5页
Journal of Electronics & Information Technology
基金
国家自然科学基金(60772080)
天津市智能交通"十一五"发展规划科研基金资助课题
关键词
多运动目标跟踪
运动识别
智能监控
Multiple moving targets tracking
Motion detection
Intelligent monitoring
